Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Economic Bulletin May 2017 Lisbon, 2017 • www.bportugal.pt Economic Bulletin | May 2017 • Banco de Portugal Av. Almirante Reis, 71 | 1150-012 Lisboa • www.bportugal.pt • Edition Economics and Research Department • Design and printing Communication Directorate | Image and Graphic Design Unit • Print run 25 • ISSN 0872-9794 (print) • ISSN 2182-0368 (online) • Legal Deposit no. 241772/06 Index I The Portuguese economy in 2016 | 5 1. Overview | 7 Box 1 | Assessment of projections for 2016 | 10 2. International environment | 13 3. Monetary and financial conditions | 21 3.1. Euro area | 21 Box 3.1.1 | The impact of the Corporate Sector Purchase Programme | 24 Box 3.1.2 | Dispersion of interest rates in the euro area money market | 27 3.2. Portugal | 30 Box 3.2.1 | Market dynamics and the deleveraging of Portuguese firms | 39 4. Fiscal policy and situation | 43 Box 4.1 | Structural developments in tax revenue in 2016 | 50 Box 4.2 | Analysis of deviations in budget execution in 2016 | 52 5. Supply | 54 Box 5.1 | Capital per worker and productivity | 59 6. Demand | 62 Box 6.1 | Developments in unit values of Portuguese exports of goods | 71 7. Prices | 75 8. Balance of payments | 80 Box 8.1 | Direct investment flows into the Portuguese economy | 86 II Special issue | 91 Distribution mechanisms of monetary policy in the Portuguese economy | 93 I The Portuguese economy in 2016 1. Overview 2. International environment 3. Monetary and financial conditions 4. Fiscal policy and situation 5. Supply 6. Demand 7. Prices 8. Balance of payments The Portuguese economy in 2016 1.Overview In 2016 output growth in the Portuguese econ- of firms has been important, especially for small omy stood at 1.4 per cent, which compares with firms and for those with activity in the construc- 1.6 per cent in the previous year. Albeit mod- tion sector (Box 3.2.1. ‘Market dynamics and the erate, these GDP developments had a marked deleveraging of Portuguese firms‛). Although the intra-annual profile, accelerating strongly in the exit of firms is part of a natural process of restruc- second half of the year, signalling that the recov- turing and reallocation of productive factors in ery process of the economy is expected to con- the economy, the potential non-recovery of debts tinue. In 2016 Portuguese output growth stood raises difficulties to credit institutions, especially in 0.3 percentage points below that observed in terms of profitability and impact on own funds. the euro area, and its level is still 4 per cent below that registered in 2008, the year that marked the outbreak of the latest international economic and financial crisis. The continued reduction of the indebtedness levels of the Portuguese economy requires the creation of favourable conditions to economic growth, against a background of social cohe- In terms of the fundamental macroeconomic bal- sion and sustained improvement of citizens’ ances, there was a slight increase in the current living conditions. Investment is one of the key and capital account surplus (0.5 p.p.) and a signifi- variables for Portuguese economic growth. In cant decline in the total general government defi- 2016 this variable declined in volume by 0.8 per cit (2.3 p.p.), against a background of stabilisation cent, after an increase of 4.7 per cent in 2015, of the structural balance and decline in the struc- although with an intra-annual recovery profile tural primary balance (0.3 p.p.). The improvement and an acceleration in the business compo- in these balances is a prerequisite to ensure nent. However, investment developments, are macroeconomic stability and the credibility of the insufficient, not only because the fall during the Portuguese economy among international inves- crisis period was very sharp, but also because tors, and should therefore be a guideline for eco- capital per employee in the Portuguese econ- nomic policy decisions adopted at the national omy is still at low levels, when compared with level. In effect, the high indebtedness levels pre- the euro area average (Box 5.1 ‘Capital per vailing in the different institutional sectors make worker and productivity‛). The low capital levels the Portuguese economy particularly vulnerable contribute to the under-performance of labour to any developments implying an increase in the productivity in the Portuguese economy, which interest rates applied to external financing. declined again in 2016. The deleveraging process of the Portuguese The resident population and the labour force con- economy continued in 2016, reflecting a decline in households’ and corporations’ indebtedness ratios, which nonetheless continued to be very high. Public debt ratio, in turn, virtually stabilised in net terms, in parallel with a reduction of external indebtedness in the total economy. In this context, there was an improvement in the international investment position of 7 percentage points (p.p.) of GDP, which nonetheless stands at a negative level of 105 per cent of the product. tinued to decline in 2016. These declines maintain the trend observed since 2011, although slightly less marked than in previous years. In 2016 resident population in the 25-34 age group declined by 2.5 per cent, while the labour force declined by 2.7 per cent. The sharp demographic trends in the Portuguese economy are not easily revertible and are aggravated by migration phenomena. Although the ageing of the population is evident in several European countries, this situ- In the case of non-financial corporations, the con- ation is an important constraint to Portuguese tribution to the decline in debt due to the exit economic growth and puts greater pressure on 7 8 BANCO DE PORTUGAL • Economic Bulletin • May 2017 the dynamics of public expenditure with pensions non-resident entities, almost exclusively focus- and health. ing on the issue of debt securities. Therefore, In 2016 labour market developments were characterised by an increase in employment, exceeding that of Gross Value Added and maintaining the recovery profile observed since the second it was possible to offset the new decline of the stock of bank loans granted by resident banks, which indicates an alternative means to finance investment projects. quarter of 2013. Moreover, although remain- The less buoyant economic activity in 2016 also ing at very high levels, the unemployment rate reflected the deceleration in exports of goods declined by 1.3 p. p., in a context of stronger and services, against a background of lower wage dynamics than in past years. This may have growth of the external demand directed towards contributed to the slight increase in the inflation the Portuguese economy. The deceleration in rate in the course of 2016, via an increase in unit exports of goods and services in 2016 reflected labour costs. The inflation rate in Portugal, meas- a lower growth of exports of non-tourism ser- ured by the change in the Harmonised Index of vices and chiefly the fall in exports of energy. Consumer Prices, stood at 0.6 per cent in 2016, with zero price growth in goods and 1.5 per cent price growth in services. The fall in investment contributed to less robust growth of domestic demand than that observed in 2015. Also, private consumption increased slightly less than in the previous year, as a result of less buoyant consumption of non-food current goods and durable goods. In turn, public consumption, determined by the dynamics of compensation of employees and consumption of goods and services in the general government, increased slightly less in 2016 than in 2015. Developments in investment and consumption took place in a context of gradual improvement of the financing conditions of private economic agents, in line with the maintenance of an expansionary stance of the monetary policy in the euro Excluding this category, growth of exports of goods was 1.8 percentage points higher than in 2015. In turn, the buoyancy of exports of tourism services continued to be high, in tandem with the significant increase in nights spent by non-residents in Portuguese hotels. The performance of Portuguese exports in 2016 translated into a new gain in market share, which underlines the competitive capacity of Portuguese firms in international markets. The stance of the Portuguese economy, gradually turning towards the production of goods and services tradable abroad is a very positive feature of the adjustment in the productive structure, which was started before the latest international economic and financial crisis. Only by strengthening its export capacity may the Portuguese economy accommodate growth of imports resulting from a desirable strengthening of the invest- area. This improvement, visible in the course ment rate and an increase in consumption lev- of 2016, was reflected in a decline of the inter- els. The re-emergence of current account defi- est rate spread between Portugal and the euro cits, even if financed via foreign investment, may area, for both corporations and households. New be perceived as a return to imbalances, which loans to households reinforced the growth trend will tend to adversely affect refinancing condi- observed in early 2016, especially in consump- tions of existing debt. The relevance of exports tion. Notwithstanding the increase in new credit, performance for developments in the Portu- the households’ deleveraging process continued. guese economy and the prevailing high indebt- The annual rate of change of total credit to edness levels highlight the importance of devel- corporations was also positive at the end of opments in the international economy. the second half of 2016. Relevant, however, In 2016 world economic activity decelerated, was the important contribution of credit from reaching the lowest growth since the latest The Portuguese economy in 2016 international economic and financial crisis. In in goods and services, shall be complementary turn, growth in the euro area was slightly lower to internal debate on strengthening growth con- than in 2015, with an increase in output more ditions. The future of the Portuguese economy based on domestic demand than in the past. In depends heavily on its external partners, either 2016 financial market developments were con- as destination of exports and source of produc- ditioned by the UK referendum in June and the tive investment, or within the scope of the ongo- US presidential election in November, with vola- ing financial integration process. This stresses tility peaks and an increase in uncertainty follow- the need to increase national agents’ standards ing both events. Stock market indices, however, regarding the improvement of the country’s com- registered gains, chiefly in the second half of the petitiveness. Only in this vein shall it be possible to year and in the banking sector, partly reflecting resume convergence towards European Union’s the strong liquidity creation via non-standard average levels of well-being. monetary policies. In a context of low economic growth, the existence of high volatility and strong valuations in financial markets may lead to disturbances in the international economy, which will tend to be amplified in most indebted economies. In addition, the rise in inflation in the euro area from mid-year onwards, triggered by energy prices, as well as the increase in the federal funds rate target contribute to maintain concerns regarding the development of the interest rate at which the Portuguese economy is financed. The Portuguese economy has shown a remarkable macroeconomic adjustment capacity and a sectoral restructuring based on the internationalisation of Portuguese firms. Continuing economic growth and increasing citizens’ wellbeing require the strengthening of firms’ productive investment and the availability of skilled human resources, in a context of financial stability and legal and fiscal certainty. In this context, the obstacles created by the low savings rates in the Portuguese economy and the unfavourable demographic developments, as well as cyclical difficulties associated with the high indebtedness levels and the elevated uncertainty in the international framework, require a consistent and forward-looking approach to economic policies. The discussion on the deepening of European integration, which has always been a catalyst for reforms in the Portuguese economy, allowing for gains through intensified capital flows and trade 9 10 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Box 1 | Assessment of projections for 2016 This box analyses the gap between Banco de Portugal’s projection for GDP in 2016 and the actual figures, trying to identify the main factors explaining that gap. The analysis is based on the projection published in March 2016, which contains for the first time comprehensive information for 2015.1 The above-mentioned projections for the Portuguese economy are based on Banco de Portugal’s quarterly econometric model (‘M’ Model). In addition to this main model, other satellite models are considered, such as the MIMO model for inflation projection, and bridge models of shortterm activity developments.2 Forecasts depend on a range of assumptions regarding the external environment and a set of rules for the incorporation of fiscal variables. In the first case, as usually mentioned in the projection articles of the Economic Bulletin, assumptions correspond to those of the Eurosystem’s projection exercises. Fiscal variables follow the rule used within the scope of these exercises, i.e., incorporating the policy measures that have already been approved (or that are highly likely to be approved) and that have been specified in detail. As mentioned in the note published in March 2016, the projection incorporates most measures included in the State Budget for 2016, which are sufficiently detailed. Projection errors are the result not only of the fact that assumptions incorporated in the forecast exercise have not materialised, but also of factors related to the model and judgement elements incorporated in the projection. Comparing the annual growth rate of real GDP in 2016 published in the Quarterly National Accounts by Institutional Sector of 24 March with the projection published by Banco de Portugal in early 2016, it may be concluded that the projection error was -0.1 p.p., i.e., Banco de Portugal projected GDP growth at 1.5 per cent, while actual GDP growth was 1.4 per cent. Based on elasticities implied in the ‘M’ model, it is possible to analyse the contributions to projection errors made by changes in assumptions, in particular financial conditions, foreign exchange rates, and the international scenario (external demand, oil prices and external prices). In this context, Table 1 presents the revisions of the external environment assumptions – defined as the difference between the actual value and the value underlying the March exercise – and their impact on the annual GDP growth rate in 2016. Table 1 • Revisions on the external environment and its impact on the real annual growth rate of GDP in 2016 | Revision on the annual growth rate in p.p., except interest rate, compared to the projection of March 2016 Revisions on the external environment Impact on GDP 0.0 0.0 Euro effective exchange rate -0.3 0.0 Competitors' prices in national currency -1.0 0.0 Oil prices in euros 18.2 -0.1 External demand -1.9 Interest rate Impact of revisions on the external environment -0.4 -0.5 Source: Banco de Portugal. Note: A ‘-‛ sign in the effective euro exchange rate corresponds to a higher depreciation or a lower euro appreciation than expected in March. The Portuguese economy in 2016 Expectations in early 2016 pointed to a more favourable international scenario than that subsequently observed. In particular, external demand for Portuguese goods and services was anticipated to have an annual average growth of 3.9 per cent, i.e., 1.9 p.p. above that actually registered in 2016. This significant downward revision of external demand was due to both intra-euro area countries and third countries (Section 2: ‘International environment’) and implied a downward revision of the GDP growth rate projected at 0.4 p.p. for 2016. In addition, the technical assumption for developments in the effective exchange rate of the euro, which assumes the maintenance, over the projection horizon, of the average levels observed in the two weeks prior to the cut-off date, was revised slightly downwards, implying an appreciation of the euro slightly below that expected in March 2016. This was reflected in a slight improvement in the price competitiveness of the Portuguese economy against the main competitors located outside the euro area. There was also a slight decline in competitors’ prices in external markets, with a negative impact on the price competitiveness of Portuguese exports. The impact on the GDP projection of the revision of exchange rates and external prices is negligible. The deterioration of the international scenario in the course of 2016 was offset by a more favourable than anticipated behaviour of the export market share in 2016. The very significant gains of the Portuguese export market share were registered in both goods and services, particularly in the second half of the year (Section 6: ‘Demand’). Therefore, contrary to any conclusion that might be withdrawn from the broadly downward revision of the international environment assumptions, exports of goods and services were higher than projected in the March exercise. As regards monetary and financial conditions, the technical assumption for the 3-month Euribor rate remained virtually unchanged from March projection (Section 3.1: ‘Monetary and financial conditions in the euro area’). The increase in oil prices in the course of 2016 was higher than that expected and implied an upward revision from the assumption considered in the March exercise, with a negative impact of approximately 0.1 p.p. on the GDP growth rate. Therefore, taking into account the multipliers implied in the ‘M’ model, revisions of the external environment led to a significant downward revision of GDP by around 0.5 p.p. Table 2 presents the projection errors in GDP and its components, as well as the respective gross and net contributions of the imported content. Analysing the components net of imports, it may be concluded that the projection error of -0.1 p.p. in GDP was due to the overestimation of the public finance variables, in particular to the lower growth of public consumption and to the stronger fall of public investment (Box 4.2: ‘Analysis of deviations of budget execution in 2016‛). These effects were partly offset by the underestimation of exports, largely reflecting the already mentioned unanticipated gains in market share. In spite of an underestimation of private consumption of 0.3 p.p. in gross terms, the contribution of this component net of the imported content for the GDP error was virtually zero, given that the projection error of this variable largely reflected car purchases. Moreover, the annual average growth rate of GDP and its components in 2016 implied a sharp intra-annual profile from the first to the second half of the year. The first half of the year was characterised by low growth, with quarter-on-quarter rates of change of 0.2 per cent, whereas in the second half there was an acceleration in economic activity, with quarter on quarter rates of change of 0.9 and 0.7 per cent in the third and fourth quarters respectively. The projection published in March 2016 did not anticipate these intra-annual GDP dynamics, and projected a smoother profile, with a slight acceleration in the second half of the year. In this 11 12 BANCO DE PORTUGAL • Economic Bulletin • May 2017 context, there was an overestimation of GDP and its main components in the first half of the year and an underestimation in the following half year. Table 2 • Projection error in real GDP growth rate and in major components in 2016 | Observed– projected in March 2016 GDP Weights 2015 in % GDP Projection error Gross contributions to GDP growth Net contributions to GDP growth -0.1 100.0 -0.1 -0.1 Private consumption 65.6 0.5 0.3 0.0 Public consumption 18.2 -0.6 -0.1 -0.1 Investment 15.5 -1.0 -0.2 -0.2 Exports 40.6 2.2 0.9 0.3 Imports 39.8 2.3 -0.9 Source: Banco de Portugal. Note: The demand aggregates net of imports are obtained by subtracting an estimate of the imports needed to meet each component. The calculation of import contents was based on data for 2005. For more information, see the Box entitled ‘The role of domestic demand and exports in economic activity developments in Portugal’, in the June 2014 issue of the Economic Bulletin. The Portuguese economy in 2016 2.International environment World economic activity and trade decelerated further in 2016, dropping to the lowest level since the last international economic and financial crisis Taking the pre-crisis period as reference (2002-07), international trade recorded an average growth much higher than global activity (around 8 and 5 per cent respectively), while world trade and GDP grew by around 3 per cent, on average, from 2012 to 2016, i.e. a decline in world trade elasticity of around 0.7 p.p. (Chart 2.2). This decrease in In 2016 world economy decelerated, recording the lowest growth since the last international economic and financial crisis (Table 2.1 and world trade elasticity is expected to be associated, inter alia, to slower developments in global value chains and a geographical change in trade benefiting emerging market economies, which have Chart 2.1). This slow recovery from the crisis shown a slower response of trade to changes in has been different across regions. In 2016 as a output. whole, contrary to the previous year, advanced economies slowed down and emerging market and developing economies decelerated slightly. In 2016 developments in financial markets were affected by the referendum in the United Kingdom in June and the presidential election in the After recovering from its major collapse, world United States in November, which were followed trade growth has been weakening since 2012. by volatility peaks and increased uncertainty. Table 2.1 • Gross Domestic Product – Real year-on-year rate of change | Percentage 2002-07 2013 2014 2015 2016 World 4.8 3.4 3.5 3.4 3.1 Advanced economies 2.6 1.3 2.0 2.1 1.7 USA 2.7 1.7 2.4 2.6 1.6 Japan 1.4 2.0 0.2 1.2 1.0 Euro area 2.0 -0.2 1.2 1.9 1.7 Germany 1.4 0.6 1.6 1.5 1.8 France 1.8 0.6 0.7 1.2 1.1 Italy 1.1 -1.7 0.2 0.7 1.0 Spain 3.5 -1.7 1.4 3.2 3.2 2.7 1.9 3.1 2.2 1.8 Emerging and developing economies 7.2 5.1 4.7 4.2 4.1 Emerging and developing Europe 5.9 4.9 3.9 4.7 3.0 Commonwealth of Independent States 7.6 2.1 1.1 -2.2 0.3 7.1 1.3 0.7 -2.8 -0.2 United Kingdom Russia Emergind and developing Asia China Latin America and the Caribbean Brazil Sources: Eurostat, IMF and Thomson Reuters. 9.0 6.9 6.8 6.7 6.4 11.2 7.8 7.3 6.9 6.7 4.1 2.9 1.2 0.1 -1.0 3.9 3.0 0.5 -3.8 -3.6 13 14 BANCO DE PORTUGAL • Economic Bulletin • May 2017 In June, the result of the referendum raised con- from July onwards, reflecting reduced concerns cerns about the UK’s economic outlook, and, in about the world economy outlook, the percep- November, the result of the US elections sig- tion that monetary authorities might be less nalled shifting expectations regarding the new reactive to the first signs of an inflation hike, administration’s policies. In the first half of the and the anticipation of a more expansionary fis- year, long-term interest rates followed a down- cal policy and an increase in investment in infra- ward trend in the United States, the United King- structures after the US elections. dom and Germany, amid doubts about growth At the start of the year, equity indices record- in global economic activity, the situation in the ed losses, given the more negative outlook for euro area financial sector and continued (or the world economy, particularly in China, which strengthening) accommodative monetary poli- were followed by a recovery, amid rising oil pric- cies by major central banks (Chart 2.3). In con- es. After reaching a trough of 28 USD/barrel in trast, long-term interest rates began increasing January, oil prices increased gradually, before 12.0 Chart 2.1 • Gross Domestic Product - annual growth between 2010 and 2016 | Percentage 10.0 8.0 6.0 4.0 2.0 0.0 -2.0 World Advanced economies Euro area US Emerging economies China Pre-crisis growth (2002-2007) | Percentage and percentage points 13.0 2.6 7.0 2.1 1.0 1.6 -5.0 1.1 -11.0 1980 1984 1988 1992 GDP growth Pre-crisis GDP average growth Trade elasticity (rhs) 1996 2000 2004 2008 2012 Trade growth Pre-crisis trade average growth Sources: IMF and Banco de Portugal calculations. Note: Trade elasticity is calculated over 5-year periods. Pre-crisis period: 2002-2007. 2016 0.6 Percentage points Chart 2.2 • Growth of world GDP and trade volumes Percentage Sources: Eurostat, IMF and Thomson Reuters. The Portuguese economy in 2016 stabilising at around 50 USD/ barrel at the end Advanced economies grew by 1.7 per cent of the summer. At the end of the year, the stock in 2016, compared with 2.1 per cent in 2015, market in the United States reached historical improving in the second half of the year. For highs due to the anticipation of expansionary most economies, domestic demand and, in par- policies, which influenced the various sectors, ticular, private consumption continued to be the in particular banks (Chart 2.4). drivers of growth. Investment, as measured by 15 gross fixed capital formation (GFCF), continued to grow moderately and accelerated in the sec- Advanced economies grew less strongly, while emerging economies accelerated slightly, despite differences across countries ond half of 2016, making a marginal contribution to GDP. In the United States, economic activity slowed down in 2016, with GDP growing by 1.6 per cent, i.e. 1 p.p. less than in the previous year. The US 6 Chart 2.3 • 10-year sovereign debt rates 5 4 | Percentage 3 2 1 0 -1 Jan. 11 Jan. 12 Jan. 13 Euro area Jan. 14 US Jan. 15 Jan. 16 UK Japan Sources: Thomson Reuters and ECB. 200 175 125 Chart 2.4 • Major stock markets indexes, total and banks 100 | Index Jan. 2014=100 150 75 50 25 Jan. 11 Jan. 12 Eurostoxx Banks: euro area Source: Thomson Reuters. Jan. 13 Jan. 14 S&P500 Banks: US Jan. 15 Jan. 16 Footsie Banks: UK 16 BANCO DE PORTUGAL • Economic Bulletin • May 2017 economy recorded weak growth during the first to reinforce the pass-through of monetary policy half of the year, largely due to developments in to the real economy. private business investment, recovering in the Japan gradually recovered throughout 2016, second half of the year owing to private consumption. At the same time, labour market conditions remained favourable, with the unemployment rate remaining below 5 per cent, for the first time since 2007. Inflation increased, in particular from the middle of the year onwards, due to an increase in commodity prices, but continued below the Federal Reserve’s target. In terms of the average annual rate of change, inflation reached 1.3 per cent (0.1 per cent in 2015), while inflation excluding food and energy remained slightly above 2 per cent. Against this background, in December 2016, the Federal Reserve decided to increase the target for the federal funds interest rate to a 0.50-0.75 per cent range. The United Kingdom grew above the forecasts released after the June referendum, which resulted in the United Kingdom leaving the European Union. Economic activity grew by 1.8 per cent in 2016, i.e. 0.4 p.p. less than in 2015, supported by private consumption, which, contrary to expectations, did not show signs of slowing down. However, in the next few years, real household income may decelerate, negatively affecting private consumption, amid moderating wage growth and higher import prices associated with the depreciation of the pound sterling. The year-on-year rate of change in the Harmonised Index of Consumer Prices (HICP) remained below 2 per cent, but increased over the year, reaching 1.6 per cent (0.2 per cent in December 2015), as a result of the depreciation of the pound sterling, which was more pronounced immediately after the referendum’s unexpected result. In this owing to domestic demand and supported by accommodative monetary and fiscal policies and favourable financial conditions. The Japanese GDP grew by 1.0 per cent, compared with 1.2 per cent in 2015. In 2016, the Bank of Japan made two changes to its Quantitative and Qualitative Monetary Easing programme, by introducing a negative interest rate to banks’ current accounts in January, and by conditioning the purchase of Japanese government bonds, in September, so that 10-year interest rates would remain close to zero. In addition, the Bank committed itself to expanding the monetary base until the observed inflation rate exceeds the target (inflation-overshooting commitment). The situation of emerging market and developing economies continued to be quite mixed. In China, economic activity remained robust in 2016, with GDP growing by 6.7 per cent yearon-year (6.9 per cent in 2015), benefiting from the economic policy stimulus. Nevertheless, economic activity was weaker than expected in a number of Latin American countries, which have been in a recession since 2015, specifically Brazil. In Russia, economic growth was slightly better than expected, partly owing to developments in oil prices. Euro area economic activity continued to recover gradually, with a greater support from domestic demand than in the past context, in August, the Bank of England adopt- Euro area GDP grew at a relatively stable rate ed a package of measures to support the econ- throughout the year. Growth of 1.7 per cent omy, which included a cut in the Bank Rate, an reflected a positive contribution from domes- increase in the stock of purchased UK govern- tic demand and a negative contribution from ment bonds, a programme to purchase non- net exports. Above all, this pattern was driven financial investment-grade corporate bonds and by consumption, associated with favourable a Term Funding Scheme for eligible institutions developments in households’ real disposable The Portuguese economy in 2016 income, amid growing employment and declin- Unemployment continued to decline in the euro ing oil prices. The recovery in activity in the cur- area, with some cross-country differences. The rent economic cycle has a greater support from larger economies experienced declines, with domestic demand than in the previous cycle, the exception of Italy. In parallel, employment with services playing a greater role (Chart 2.5). continued to grow moderately, above the lev- The recovery seen after 2013 is also more broad- els observed in 2015. 17 ly-based across countries, with particularly positive developments in Spain (3.2 per cent growth) (Chart 2.6). GDP grew by 1.8 per cent in Germany and slightly above 1 per cent in France. In External demand for Portuguese goods and services decelerated turn, GDP growth stood at 1.0 per cent in Italy, External demand for Portuguese goods and standing considerably below pre-crisis levels. services decelerated in 2016, growing by 2 per Chart 2.5 • Euro area two latest economic recoveries – contributions by sector | Index 2009 Q2=100 Contributions to GDP Contributions to GVA 110 110 110 108 108 106 106 104 104 108 110 106 108 102 102 100 100 9898 2009 2010 2012 2013 2009 Q2Q2 2010 Q4Q4 2012 Q2 Q2 2013 Q4 Q4 20152015 Q2 Q220162016 Q4 Q4 Domestic demand Domestic demand NetNet Exports Exports GDPGDP cumulated growth sincesince 20092009 Q2 Q2 cumulated growth 104 106 104 102 102 100 100 98 98 Q2 2009 2009 Q2 2011 Q2 2011 Q2 Industry 2013 Q2 2013 Q2 2015 Q2 2015 Q2 Industry Services Services Construction Construction Gross value added cumulated growth since 2009Q2 Gross value added cumulated growth since 2009Q2 Sources: Eurostat, CEPR and Banco de Portugal calculations. Note: Two latest economic recoveries - since the 2009 Q2 and 2013 Q1 troughs, according to the CEPR Dating Commitee. 120 Chart 2.6 • GDP evolution since trough 115 110 | Index 2009 Q2=100 105 100 95 90 2009 Q2 2010 Q4 Euro area 2012 Q2 Germany Sources: Eurostat, CEPR and Banco de Portugal calculations. 2013 Q4 France 2015 Q2 Italy 2016 Q4 Spain 18 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Increase in euro area inflation from the middle of the year cent year-on-year, after around 4 per cent in 2015 (Table 2.2). These developments resulted from lower growth in demand from euro area trade partners (3.5 per cent, compared with 5.6 per cent in 2015), and a contraction in external demand from economies outside the euro area (-0.4 per cent, after 1.1 per cent in 2015). When external demand from Angola In 2016 developments in HICP inflation mostly reflected the path of energy prices (Chart 2.7). This influence translated into a low or nega- is taken into account, the effect is even more tive inflation in the first months of the year and pronouced.3 External demand for goods and subsequent increases, which were heightened services is expected to have grown by 1.3 per by the base effect of the falls observed in 2015 cent in 2016, compared with 2.6 and 5.4 per and related to an increase in energy prices. cent in 2015 and 2014 respectively. Table 2.2 • External demand of goods and services – Real year-on-year rate of change | Percentage Weights(b) 2013 2014 2015 2016 100.0 1.9 5.0 3.9 2.0 66.3 0.9 5.2 5.6 3.5 External demand (ECB)(a) Intra euro area external demand of which: Spain 27.1 -0.5 6.5 5.6 3.3 Germany 13.7 3.2 4.0 5.0 3.6 France 12.5 2.2 4.8 6.4 3.6 3.9 -2.3 3.1 6.7 3.1 33.7 3.7 4.6 1.1 -0.4 Italy Extra euro area external demand of which: United Kingdom 5.6 3.4 2.5 5.5 2.8 USA 3.5 1.1 4.4 4.6 1.1 Goods and services imports from Angola (c) 9.0 11.4 -21.8 -27.8 Adjusted external demand (d) 2.3 5.4 2.6 1.3 Comércio mundial de bens e serviços (FMI) 3.7 3.7 2.7 2.2 Importações mundiais de mercadorias (CPB) 2.0 2.6 1.9 1.2 Memo: Sources: ECB, CPB, IMF, Thomson Reuters and Banco de Portugal calculations. Notes: (a) External demand is computed as weighted average of the imports volume of Portugal's main trading partners. Each country/region is weighted by its share in Portuguese export. (b) Shares computed using 2012-2014 data. (c) The weight refers to the weight of nominal goods and services exports to Angola on portuguese exports. (d) External demand indicator adjusted for the importance of the foreign trade with Angola. Corresponds to the weighted average (by the exports weight) between the external demand indicator calculated by the ECB and the volume of the goods and services imports of the Angolan economy. The IMF forecasts (Word Economic Outlook) for the growth of the volume of imports of Angola in 2016 are used. The year-on-year rate of change in HICP exclud- countries. As in the past, differentials between ing energy and food remained stable at relatively groups of countries reflected divergences in low levels, ranging between 0.7 and 1.0 per cent inflation excluding energy and food (Charts 2.8 throughout the year (Chart 2.7). Since mid-2016, and 2.9). At the end of the year, the year-on-year the increase in euro area inflation was boost- rate of change in HICP excluding energy and ed by energy prices and broadly-based across food stood at 1.4 per cent in Germany, 0.9 and The Portuguese economy in 2016 19 1.5 Chart 2.7 • Contributions to year-on-year HICP rate of change in the euro area 1.0 0.5 0.0 | Percentage points -0.5 -1.0 -1.5 Jan. 14 Jul. 14 Jan. 15 Jul. 15 Excl. food and energy Energy Jan. 16 Food Jul. 16 Total Source: Eurostat. 3 Chart 2.8 • Inflation differentials between high-rated and low-rated countries 2 1 0 | Percentage points -1 -2 Jan. 99 Jan. 03 Jan. 07 Total Jan. 11 Jan. 15 Excluding food and energy Sources: Eurostat, ECB and Banco de Portugal calculations. Note: High rated countries: Austria, Belgium, France, Finland and Germany. Low rated countries: Cyprus, Greece, Ireland, Italy, Portugal and Spain. 4 Chart 2.9 • Euro area HICP 3 | Year-on-year rate of change 2 1 0 Jan. 99 Jan. 03 Jan. 07 High rated countries Jan. 11 Jan. 15 Low rated countries Sources: Eurostat, ECB and Banco de Portugal calculations. Note: High rated countries: Austria, Belgium, France, Finland and Germany. Low rated countries: Cyprus, Greece, Ireland, Italy, Portugal and Spain. BANCO DE PORTUGAL • Economic Bulletin • May 2017 0.7 per cent in Spain and Italy respectively, and in services prices in high rated countries and a 0.4 per cent in France. slight acceleration in lower rated countries. Long- Since 2013, the inflation differential between term inflation expectations increased from September onwards, standing above the levels seen tries has been negative, in contrast to the pre- at the start of 2015, but below historical aver- crisis period. This reflects lower growth in pric- ages. Long-term inflation expectations seem to es in the services component in lower-rated have become rather more sensitive to changes countries, associated with wage adjustments. in short-term expectations and oil prices both in Since the second half of 2015, the differential the euro area and in the United States and the has decreased, reflecting continued low growth United Kingdom (Chart 2.10). Chart 2.10 • Euro area inflation expectations and oil prices | Percentage and USD dolars per barrel 3.0 175 2.5 150 2.0 125 1.5 100 1.0 75 0.5 50 0.0 Jan. 06 25 Jan. 08 1-y, 1-y ahead Jan. 10 Jan. 12 Jan. 14 5-y, 5-y ahead Sources: Thomson Reuters and Banco de Portugal calculations. Note: Expectations implicit in inflation swaps. Vertical line indicates the APP anouncement. Jan. 16 Oil price (rhs) US dollars high rated countries and lower rated coun- Percentage 20 The Portuguese economy in 2016 3.Monetary and financial conditions 3.1. Euro area The Eurosystem maintained an accommodative monetary policy in 2016 Euro area financial markets were influenced by the accommodative monetary policy stance. The downward path in euro area government bond yields reflected the effect of the APP purchases and other monetary policy measures taken by the ECB. Overall effects were also relevant, in 2016 saw the adoption of non-standard mon- particular an increase in uncertainty about the etary policy measures on the part of the ECB in outlook for global growth at the start of the year March and December. At the start of the year, and the result of the UK referendum in June. economic and financial conditions deteriorated From the summer onwards, a decline in con- and downside risks to inflation increased, lead- cerns about developments in the global econ- ing the Governing Council to reduce the rates omy and the incorporation of an inflation risk on the main refinancing operations, the deposit premium led to a partial reversal in the decline facility, and the marginal lending facility to 0 per seen in interest rates. cent, -0.4 per cent and 0.25 per cent respectively, Euro area equity indices improved slightly through- and to adopt a number of other measures. This package included expanding the monthly purchases under the Expanded Asset Purchase Program (APP) from €60 billion to €80 billion, and extending it until March 2017, and expanding the purchase programme to debt securities issued by non-bank corporations established in the euro area (Box 3.1.1: ‘The impact of the Corporate Sector Purchase Programme‛). The ECB also announced a new series of targeted longer-term refinancing operations (TLTRO II), to support lending to the economy. These operations are different from previous TLTROs mostly because the bank will pay a rate equal to the deposit facility rate if it grants loans above a specific limit. At the end of the year, the Governing Council considered there was a need for further substantial monetary stimulus to the economy in order to ensure a sustained adjustment in the path of inflation towards the inflation aim. The Governing Council therefore decided to extend the APP until December 2017, but to reduce the purchases from €80 billion to €60 billion. Monetary and financial conditions in the euro area continued to evolve favourably out the year, although the intra-annual pattern showed marked movements, consistent with the previously mentioned factors. Equity prices in the European financial sector fell by around 8 per cent in 2016, amid prospects of lower yields in the future and the negative impact of non-performing loans on bank profitability. In foreign exchange markets, the euro appreciated slightly, in effective terms, in the course of the year, reflecting in particular the result of the UK referendum, the world economic outlook and monetary policy differences across jurisdictions. The effective appreciation of around 2 per cent was the result of an appreciation of 17 per cent against the pound sterling and 4 per cent against the Chinese renminbi, and a depreciation of 3 per cent against the US dollar and 6 per cent against the Japanese yen. Money market interest rates continued to decline, but, since the summer of 2016, the differential has been widening between the EONIA and deposit facility rates vis-à-vis the secured money market rates and rates on government debt, which might indicate a less effective transmission of monetary policy decisions (Box 3.1.2: ‘Dispersion of interest rates in the euro area money market‛). 21 22 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Firms’ financing costs continued to decline in Against this background, the net issuance of euro- 2016. In particular, the negative deposit facil- denominated bonds increased considerably from ity rate and the new TLTRO II announced in the second quarter of 2016 onwards (Chart 3.2). March are expected to have helped reduce the The cost of bank loans to firms decreased by interest rates on bank loans to firms. In addition, around 30 b.p. throughout the year, slightly the introduction of a purchase programme for below the decline seen in loans to households investment-grade bonds issued by corporations for house purchase, standing at around 1.8 per established in the euro area might also partly cent in both cases. In addition, the differential explain the decline in the respective interest between the cost of loans to firms in lower rated rates, which have stood at levels similar to those of and high rated countries returned to the levels highly-rated banks’ bonds since June (Chart 3.1). observed at the start of 2011. 6 Chart 3.1 • NFC's financing costs | Percentage 5 4 3 2 1 0 Jan. 11 Jan. 12 Jan. 13 Jan. 14 High rated NFCs bonds NFC loans below 1 million Jan. 15 Jan. 16 NFC loans over 1 million High rated Banks' bonds Sources: ECB, Bank of America Merrill Lynch, Blooomberg and Banco de Portugal calculations. Note: Vertical line indicates the announcement of CSPP and TLTRO II. Chart 3.2 • External financing of NFC in the euro area 275 15 225 10 175 5 | Billion euros 125 Jan. 11 0 Jan. 12 Jan. 13 Bank loans (new operations) Source: ECB. Note: 12-month moving average. Jan. 14 Jan. 15 Jan. 16 Net issues of bonds in euros (rhs) The Portuguese economy in 2016 Bank lending to firms grew considerably in the improvement in lending conditions for firms and first half of the year both for high rated countries households in 2016. Banks have also reported and lower rated countries (Chart 3.3). In annu- an increase in demand for loans to both firms al average terms, loan growth stood at 2.3 per and households, which was broadly-based across cent in December 2016, compared with 0.5 per all loan categories. Banks have used the liquidity cent in December 2015, accelerating 1.4 p.p. up obtained from the APP for granting loans, for refi- to June. Loans to households recovered more nancing purposes and, to a lesser extent, for pur- slowly, particularly in high rated countries, from chasing assets. As regards the negative deposit 1.4 to 2.0 per cent during the same period. The facility rate, banks have indicated a positive impact improvement in financial and monetary condi- on lending volumes, but a negative impact on net tions is in line with the results of the euro area interest income and loan margins. 23 Bank Lending Survey, which have indicated an 6.0 Chart 3.3 • Anual rate of change of bank loans in the euro area 3.0 0.0 | Percentage -3.0 -6.0 -9.0 Jan. 11 Jan. 12 Jan. 13 High rated countries Jan. 14 Low rated countries Jan. 15 Jan. 16 Euro area Source: ECB. Note: Loans adjusted for sales and securitization. Full lines represent firms and dashed lines represent families. 24 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Box 3.1.1 | The impact of the Corporate Sector Purchase Programme The Corporate Sector Purchase Programme (CSPP) was announced by the Governing Council of the European Central Bank on March, 10 2016 as an extension of the Expanded Asset Purchase Programme (APP)4 to the purchase of debt securities issued by non-financial corporations established in the euro area. The programme aims to further strengthen the pass-through of the Eurosystem’s asset purchases to the financing conditions of the economy, to provide further monetary policy accommodation, and to contribute to a return of inflation rates to levels below, but close to, 2 per cent in the medium term. As the CSPP is part of the APP, these purchases are expected to continue until the end of 20175 or beyond, if necessary, and in any case until the Governing Council of the ECB sees a sustained adjustment in the path of inflation consistent with its inflation aim. This box describes the CSPP and analyses its impact on the funding costs of euro area non-financial corporations and on their funding structure. Following the announcement of the programme, the technical parameters were published on April, 21 2016 and purchases began on June 8. Eligible bonds under the CSPP have a minimum remaining maturity of 6 months and a maximum remaining maturity of 30 years and 364 days at the time of the purchase. The lower limit was established in order to ensure that bonds issued by smaller firms, usually with shorter maturities, were included in the purchases. Moreover, bonds must be denominated in euros and they must have high credit quality (rating of BBB or higher), obtained from an external credit assessment institution. Finally, bonds must be marketable instruments eligible as collateral for Eurosystem liquidity-providing operations. Bonds must be issued by an institution established in a euro area Member State which is not a credit institution, does not have a parent undertaking which is a credit institution, and is not an eligible issuer under the public sector purchase programme (PSPP) in secondary markets. When the technical parameters of the programme were announced, the purchase of bonds with negative yields was permissible provided these were above the deposit facility rate. At the December 2016 meeting, the Governing Council of the ECB announced that purchases below the deposit facility rate would also be permitted. In order to limit risk exposure and to ensure the diversification of the Eurosystem’s portfolio, a number of restrictions are applied on the amounts and on issuer entity’ groups. Regarding the amount, the purchase of any type of bond may not exceed 70 per cent of its outstanding amount. Nevertheless, this limit may be lower for bonds issued by public undertakings. Lastly, the programme is coordinated by the ECB, but only six Eurosystem central banks carry out the purchases in both primary and secondary markets. Each central bank is responsible for purchases of a particular geographical allocation of the euro area and is responsible for publishing a list of all purchased bonds at the end of each week, as these will become available in the securities lending market. At the end of 2016, the Eurosystem had purchased a total of €51,069 million under the CSPP, of which 86.4 per cent were purchased in the secondary market and 13.6 per cent in the primary market. At that time, the CSPP accounted for 3.3 per cent of total purchases under the APP. Even though the programme is small in volume, its importance within the APP has been increasing since its implementation. This is particularly relevant in a context where, historically, firms tend to fund themselves through bank loans, as is the case of the euro area. Overall, the CSPP seems to have helped reduce debt issuance financing costs of non-financial corporations and to increase the use of this type of funding. The Portuguese economy in 2016 A preliminary assessment shows that apparently the programme had some impact on corporate bond yields and on the yields of bonds issued by other sectors. Economic literature assessing the APP has identified the impact of its announcement as the most significant event. The same seems to apply to the CSPP. The first vertical line in Chart 1 marks the date of the CSPP announcement and it is clear that, at that moment, bond yields dropped considerably for both non-financial corporations and banks. In the case of non-financial corporations, this downward trend continued, being only interrupted by some volatility in May and June, as a consequence of the outcome of the UK referendum on leaving the European Union. Following the start of the programme (second vertical line in Chart 1), developments in yields resumed the downward trend. Simultaneously, interest rates on long-term government bonds remained fairly constant, contributing to the perception that the implementation of the programme has helped reduce bond yields. High-yield bonds and yields on bonds issued by financial corporations also declined immediately after the announcement of the programme, particularly in peripheral countries (Table 1). Table 1 • Change in yields following the CSPP announcement | Basis points Investment-grade bonds Non-financial corporations High yield bonds issued by non-financial corporations Banks Euro Area -14 -8 -58 Germany -14 -4 -56 France -13 -8 -72 Italy -19 -17 -63 Spain -19 -16 -73 Portugal -65 - -64 Sources: Bloomberg, Bank of America Merril Lynch and Banco de Portugal calculations Notes: Event study: change in yields within a 3-day window following the CSPP announcement on 10 March 2016. Chart 1 • Euro area bond yields 5 400 350 4 Percentage 250 2 200 150 1 Basis points 300 3 100 0 50 -1 Jan. 12 0 Jan. 13 Asset swap spread (rhs) Jan. 14 Jan. 15 Jan. 16 Non-financial corporations Banks Sources: Bloomberg, Bank of America Merril Lynch and Banco de Portugal calculations. Note: Effective yields on euro-denominated investment-grade bonds. Asset swap spread: difference between the bond yield and a low-risk swap rate of the same maturity. Vertical lines: CSPP announcement; start of the program. 25 26 BANCO DE PORTUGAL • Economic Bulletin • May 2017 The programme must also be assessed in terms of volume. Financing of non-financial corporations through debt issuance has been increasing since the first quarter of 2016. In addition, after a slow down that began in 2013, the average gross issuance of bonds between June and December recovered in 2016 despite remaining at levels below those of previous years (Chart 2). This period is particularly interesting as it compares the issuance during the months when the programme was already being carried out with the same period in previous years. This suggests that the CSPP may have led firms to be more likely to prefer funding via debt securities. Although other factors may have influenced the financing costs of non-financial corporations and were possibly taken into consideration when choosing between debt issuance and bank loans, the developments described above seem to indicate that the CSPP has had the intended effect on the financing conditions of the economy. On the one hand, the financing costs of debt issuance have declined and remain low by historical standards. On the other hand, since the start of the programme, activity in this market seems to have gained some momentum. Chart 2 • Gross issuance of debt securities by euro area non-financial corporations | Millions of euros 100,000 80,000 60,000 40,000 20,000 0 Jan.10 Jan.11 Jan.12 Gross issuance Sources: Thomson Reuters and Bloomberg. Jan.13 Jan.14 Jan.15 June to December average Jan.16 The Portuguese economy in 2016 Box 3.1.2 | Dispersion of interest rates in the euro area money market In a money market operating perfectly, interest rates on all financial instruments should fully reflect the central bank’s reference rate – perceived as a risk-free interest rate – plus premia associated, inter alia, with credit and liquidity risk. Money market interest rates should therefore stand at a level close to the central bank reference rate and, ceteris paribus, any change in this rate resulting from monetary policy decisions should mostly pass through to negotiated rates for granting and obtaining short-term funding. At present, in the euro area, the deposit facility rate, which stands at -0.4 per cent since March 2016, would be expected to represent the effective lower bound of money market interest rates, and assets perceived to be safe and liquid would be expected to provide a yield close to this benchmark. However, since the start of 2015, the downward trend seen in the interest rates of short-term secured instruments and sovereign bonds with a high credit quality has been more marked than the declines in the deposit facility rate and the EONIA (Chart 1). This spread widened from the summer of 2016 onwards, contributing to greater dispersion around the ECB’s monetary policy interest rate, which should provide an anchor for euro area money market interest rates, under normal circumstances. High dispersion of interest rates in the money market may indicate market segmentation, which, as a result of the lower importance given to the central bank’s rate, translates into difficulties in the monetary policy transmission. Against this backdrop, it is crucial to analyse the reasons behind the evidence illustrated in Chart 1. First, the downward trend in interest rates on short-term instruments with a high credit quality may reflect anticipation by financial market participants of further cuts in the ECB’s policy rates. Nonetheless, since the summer of 2016, agents’ expectations about further reductions in Chart 1 • Evolution of representative euro area money market rates since 2015 | Percentage 0.25 1 0.00 0 -0.25 -1 -0.50 -2 -0.75 -3 -1.00 -4 -1.25 Jan. 2015 -5 Apr. 2015 Jul. 2015 Oct. 2015 Jan. 2016 Apr. 2016 Deposit facility rate EONIA Germany 3-month sovereign debt rate Euro area 3-month GC Polling rate Germany 3-month repo rate (rhs) Sources: Thomson Reuters and Bloomberg. Jul. 2016 Oct. 2016 Jan. 2017 27 28 BANCO DE PORTUGAL • Economic Bulletin • May 2017 the deposit facility rate were rather contained and consequently insufficient to explain the movement verified in some short-term interest rates in the euro area market. Furthermore, the increase in the dispersion of euro area money market interest rates occurred at the time of the referendum on whether the UK should remain in the European Union. Once the results were known, behaviours typical of a decline in risk appetite were observed, such as a search for safe and liquid assets (safe havens), which exerted downward pressure on their interest rates. However, most of these movements were reversed the following weeks, which was not the case with euro area money market interest rates. In addition, the misalignment of short-term interest rates and the ECB’s deposit facility rate may signal the existence of arbitrage opportunities that agents are not exploiting. In effect, to carry out a direct arbitrage operation on the basis of the current levels of short-term interest rates in the euro area, at least two conditions must apply: (i) accessing to the deposit facility rate6 and (ii) holding securities with yields lower than the deposit facility rate. This fact leads to two considerations. On the one hand, the persistence of potential arbitrage opportunities may indicate that monetary financial institutions (hereinafter called banks) in the euro area, which have access to the ECB’s deposit facility rate, face constraints in their capability to explore the apparent opportunities for profit persisting in the euro area money market. For example, the need to hold securities which may be used as collateral may justify why arbitrage operations are not conducted. Similarly, the fact that banks face non-negligible costs in adjusting their balance sheets may discourage changes to the composition of their portfolios. On the other hand, the importance of non-bank institutions in the euro area has increased over the past few years. Likewise, a significant share of Eurosystem purchases under the expanded asset purchase programme (APP) comes from the non-bank sector and non-resident investors. Taking into account that, in the euro area, only banks may make deposits (without risk) with the ECB and assuming agents who do not belong to the banking sector may be unwilling to keep large amounts deposited (among other reasons, large amounts are not covered by a deposit insurance), these investors will tend to look for safe alternatives for their funds. In this context, an investment possibility would be to purchase instruments with high credit quality, which exerts downward pressure on their yields and may help explain the dispersion of interest rates in the euro area money market. In parallel, in the recent past, there has been an increasing scarcity of short-term securities with low credit risk and high liquidity, which adds to the downward pressure on the level of their interest rates. This trend may be justified by at least three factors: (i) agents’ concerns arising from the considerable uncertainty patterns at international level; (ii) lesser availability of securities owing to the implementation of non-standard monetary policy measures on the part of the Eurosystem, namely the APP; and (iii) growing demand for liquid and safe securities, driven by regulatory requirements. The scarcity of short-term securities is more visible in instruments associated with jurisdictions which are perceived as having a high credit quality. Specifically, in the past few years, German sovereign bonds have strengthened their status as a safe haven, are considered special,7 and therefore strongly sought after by financial market participants. In addition to these arguments, the dispersion of short-term interest rates in the euro area may also reflect regulatory issues. In particular, the Basel III requirements may be affecting the smooth functioning of the money market. Admittedly, in the current regulatory framework, banks must The Portuguese economy in 2016 comply with a set of requirements which have an impact on liquidity deposited with the central bank. These guidelines tend to influence money market activity as a whole, thereby influencing short-term interest rates, and ultimately monetary policy transmission. Moreover, a stricter regulatory framework also leads to higher balance sheet adjustment costs.8 Consequently, the reversal of positions taken during times of lower risk appetite is harder than in the past and, in this sense, the correction of certain atypical phenomena, such as those resulting for the UK referendum, tends to be more gradual. To sum up, greater dispersion of interest rates in the euro area money market may be the result of a set of conditions interacting with each other, such as (i) greater uncertainty in global financial markets privileging low-risk assets; (ii) lower interbank activity on the part of euro area banks, which, in a demanding regulatory framework, do not take advantage of potential arbitrage opportunities and face higher costs in adjusting their portfolios, which, in turn, may lead to temporary shocks having a more prolonged effect in time; (iii) growing importance of non-bank institutions which do not have access to the deposit facility rate and seek to allocate their resources to assets with a high credit quality; and (iv) increasingly lower availability of short-term securities perceived as safe havens. Although all of these factors may contribute to downward pressure on short-term interest rates, additional analysis is needed to assess their relative importance. In this vein, it is relevant to continue to tackle the scarcity of liquid and safe securities, which might be accomplished, for example, through the Security Lending Scheme by Eurosystem central banks. Similarly to the initiative for reverse repurchase agreement operations, proposed by the US Federal Reserve, broadening the base for counterparties with access to the deposit facility rate might, in a context of ample liquidity, strengthen the role of the ECB’s policy interest rate. Hence, from the perspective of the stance and transmission of monetary policy and aspects related to financial stability, it is crucial to ensure the smooth functioning of the euro area money market, by avoiding distortions in price formation. 29 30 BANCO DE PORTUGAL • Economic Bulletin • May 2017 3.2. Portugal levels seen in interbank reference rates, which continued to decline throughout 2016. Monetary and financial conditions continued to improve throughout 2016 Monetary and financial conditions in the nonfinancial private sector improved in 2016, helped by the impact of the non-standard monetary policy measures adopted by the ECB (Section 3.1: ‘Monetary and financial conditions in the euro area‛). In the course of 2016, the differential between interest rates in Portugal and the euro area decreased both for firms and households. The current differentials are comparable to those seen in the period before 2009 (Chart 3.4). New loans to households reinforced the growth trend seen at the start of 2016, particularly in consumer credit The growth rate of new bank loans to households increased in the second half of 2016. Monthly volumes of new loans for house purchase continued to increase, despite remaining at levels far from those seen in the period before the financial crisis (Chart 3.6). The share of loans with an initial rate fixation over 1 year continued to grow, accounting for more than 40 per cent of new loans granted in this Interest rates on new loans to households remained on a downward trend Interest rates on new loans to households for house purchase and consumption remained on the downward path observed since 2012 (Chart 3.5). In addition, December 2016 figures for these segments correspond to minimum lev- segment at the end of 2016. This trend occurs amid very low interest rates – with the accommodative conditions expected to continue for an extended period – and the existence of longer-term financing instruments for banks. The recovery in new loans for consumption continued both in personal loans and loans for car purchase, with loans for the purchase of used cars gaining importance in this segment (Chart 3.7). els in the series (which started in 2003). These On the basis of the results of the Bank Lending developments are associated with the negative Survey throughout 2016, the recovery in loans 5.0 Chart 3.4 • Interest rate differencials between Portugal and the average of the euro area | In percentage 4.0 3.0 2.0 1.0 0.0 Dec. 07 Jun. 08 Dec. 08 Jun. 09 Dec. 09 Jun. 10 Dec. 10 Jun. 11 Dec. 11 Jun. 12 Dec. 12 Jun. 13 Dec. 13 Jun. 14 Dec. 14 Jun. 15 Dec. 15 Jun. 16 Dec. 16 -1.0 -2.0 Differencial (Housing) Differencial (Consumption) Sources: Consesus Economics, Thomson Reuters and Banco de Portugal. Differencial (non financial corporations) 31 The Portuguese economy in 2016 to households for house purchase and consumption was in line with a sustained increase in demand, supported, in particular, by improvements in housing market prospects and the general level of interest rates. According to the survey, although credit standards are relatively stable, competitive pressure between banks and Household deleveraging continued, particularly in the housing segment At the end of 2016, households’ total indebtedness stood at 75.4 per cent of GDP, compared with 79.4 per cent one year earlier, continuing the improvements in housing market prospects con- deleveraging process that started in mid-2011. tributed to an easing of credit standards for loans The annual rate of change in loan stocks for to households. In addition, several banks reported house purchase increased slightly, but never- a positive effect of TLTROs on an increase in loans theless remained at negative levels. In consumer to households across segments. credit, the upward trend continued, with a rate 14.0 Chart 3.5 • Interest rates on new loans granted by resident banks to households | Percentage and 12.0 10.0 8.0 6.0 percentage points 4.0 2.0 0.0 Dec. 07 Dec. 08 Dec. 09 EAPR – housing Dec. 10 Dec. 11 Spread – housing Dec. 12 Dec. 13 Dec. 14 EAPR – consumption Dec. 15 Dec. 16 Spread – consumption Sources: Thomson Reuters and Banco de Portugal. Notes: Average interest rates are based on new loans by initial fixation period and weighted by new loan amounts in each period. In the case of loans for consumption, the 6-month Euribor, the 1-year Euribor and the 5 year swap rate were considered as reference interest rates for loans with initial fixation period of less than 1 year, 1 to 5 years and more than 5 years, respectively. In the case of housing, the reference interest rate is the 6-month Euribor. 2,500 Chart 3.6 • New loans granted by resident banks to households 2,000 1,500 | 3-month moving average, EUR millions 1,000 500 0 Dec. 07 Dec. 08 Dec. 09 Housing (<1 year) Dec. 10 Dec. 11 Dec. 12 Housing (>=1 year) Dec. 13 Dec. 14 Consumption Source: Banco de Portugal. Note: In the case of housing, new loan amounts are disaggregated by interest rate fixation period. Dec. 15 Dec. 16 32 BANCO DE PORTUGAL • Economic Bulletin • May 2017 of change of 5.7 per cent at the end of 2016. In loans for other purposes, the loan stock continued to decline at a rate of around 2.5 per cent for the year as a whole (Chart 3.8). Nevertheless, the differential against the EURIBOR stabilised over the course of the year. Together, these factors show the considerable impact of the ECB’s monetary policy on the reduction in bank interest rates. Interest rates on new loans to firms reached new lows at the end of 2016 The distribution of interest rates on new loans shows a gradual increase in risk differentiation Interest rates on new loans to firms remained on a downward trend in 2016, reaching a new his- On the basis of estimates on the probability of torical low of 2.8 per cent in December (Chart 3.9). default for Portuguese non-financial corporations 500 Chart 3.7 • New loans to households for consumption by credit category | 3-month moving 375 250 average, EUR millions 125 0 Dec. 09 Dec. 10 Dec. 11 Dec. 12 Personal credit Dec. 13 Dec. 14 Car loans – new Dec. 15 Dec. 16 Car loans – second hand Source: Banco de Portugal. Notes: New loan amounts for consumption granted by financial institutions. The analysis excludes credit cards, current accounts, and overdraft facilities. 15.0 Chart 3.8 • Loans granted by resident banks to households | Annual rate of change, percentage 10.0 5.0 0.0 -5.0 -10.0 -15.0 Dec. 2007 Dec. 2008 Dec. 2009 Total Dec. 2010 Dec. 2011 Consumption Dec. 2012 Dec. 2013 Housing Dec. 2014 Dec. 2015 Dec. 2016 Other purposes Source: Banco de Portugal. Notes: Annual rates of change are based on the relation between end-of-month outstanding amounts (adjusted for securitisation operations) and monthly transactions. Monthly transactions correspond to the difference in the end-of-month outstanding amounts adjusted for reclassifications, write-offs/write-downs, exchange rate and price revaluations, and any other variations that do not correspond to financial transactions. Whenever relevant, figures are additionally adjusted for sales of credit portfolios. 33 The Portuguese economy in 2016 in 2016, the distribution of interest rates on new the average. This analysis may suggest the market loans to lower-risk firms continued to move to is close to stabilising, with no further reductions lower levels, particularly through the reduction in expected in interest rates. the volume of loans with high interest rates for The annual rate of change in total credit to firms recorded positive levels in the second half of 2016 the low risk profile of this class (Chart 3.10).9 Panel b) in Chart 3.10 shows that interest rates on new loans to risky firms continue to show a considerable dispersion. Nevertheless, a comparison of the distribution of interest rates in the last quarter of 2016 with those of the same period in 2015 in the The annual rate of change in total credit to non- segment for risky firms shows a greater concen- financial corporations recorded a positive level tration of new loans with interest rates close to of 0.5 per cent at the end of 2016 (Chart 3.11).10 8.0 Chart 3.9 • Interest rates on new loans granted by resident banks to households 6.0 4.0 2.0 | Percentage and percentage points 0.0 -2.0 Dec. 07 Dec. 08 Dec. 09 Dec. 10 Dec. 11 Dec. 12 Average interest rate Dec. 13 Dec. 14 Dec. 15 Euribor (3-month) Dec. 16 Differencial Sources: Consesus Economics, Thomson Reuters and Banco de Portugal. Notes: Average interest rates are based on new loans by initial fixation period, weighted by new loan amounts in each period. Chart 3.10 • Density distribution of interest rates on new loans granted by banks to private corporations by credit risk profile Panel b) – high risk 0.6 0.6 0.4 0.4 Density Density Panel a) – low risk 0.2 0.0 0.2 0.0 0 1 3 5 6 8 9 11 12 14 0 1 3 5 2014 Q4 2015 Q4 6 8 9 11 12 14 Interest rate Interest rate 2016 Q4 2014 Q4 2015 Q4 2016 Q4 Source: Banco de Portugal. Note: Interest rates weighted by loan amounts. The sample includes pro-profit corporations. High (low) risk firms lie in the first (last) two deciles of the credit risk distribution. Credit risk is measured by the Z-score estimated according to Antunes, Gonçalves and Prego, ‘Firm default probabilities revisited‛, Banco de Portugal Economic Studies, Vol. 2, No 2, April 2016. 34 BANCO DE PORTUGAL • Economic Bulletin • May 2017 These developments were affected by a con- Survey, which indicate a slight increase in loans siderable increase in credit from non-resident to large firms and a narrowing of margins on entities, which contributed with 3.2 p.p. to the average loans. annual rate of change. This increase in credit Conversely, in the second half of 2016, differ- was concentrated almost exclusively in debt ing loan dynamics observed across econom- securities issuance. Conversely, the annual rate ic sectors diminished, mainly explained by a of change in the stock of credit granted by resi- stabilisation in the rate of change in lending dent banks remains negative, worsening slight- to retail and wholesale trade, which fell in line ly in the last quarter of 2016. In the resident with manufacturing figures (Chart 3.13). Finally, financial sector, lending experienced a par- lending to construction and real estate contin- ticularly marked contraction, with holdings of ued to experience a contraction, with an annu- debt securities by resident entities recording a al rate of change of almost -7 per cent. slightly positive rate of change in 2016. Developments in lending by resident and nonresident institutions are also asymmetric when breaking down developments in credit by firm size (Chart 3.12). This asymmetry became more Bank loans continue to be channelled to firms with a better risk profile marked in the last quarter of 2016. During this Bank loans by resident banks to firms contin- period, lending to large firms and holding com- ue to show mixed developments, according to panies posted a rate of change of 0.4 per cent, risk profile (Chart 3.14). Only firms in the lower as these entities have greater access to non-res- risk quartile continue to show positive rates of ident financial institutions, and lending to micro, change of loans. However, although experienc- small and medium-sized firms recorded a rate ing a decline, the total stock of non-performing of change of -1.8 per cent. This is consistent with loans accounts for 18 per cent of the total stock of the results of the January 2017 Bank Lending loans in 2016 (Chart 3.15). Chart 3.11 • Total credit granted to non-financial corporations | Annual rate of change and contributions, percentage and percentage points 6 6 4 4 2 2 0 0 -2 -2 -4 -4 -6 -6 -8 -8 2010 Q1 2010 Q3 2011 Q1 2011 Q3 2012 Q1 2012 Q3 2013 Q1 2013 Q3 2014 Q1 2014 Q3 2015 Q1 2015 Q3 2016 Q1 2016 Q3 Loans and debt securities held by resident financial institutions Loans and debt securities held by other resident sectors and trade credit from residents Non-residents (loans, debt securities and trade credit) Total credit (rhs) Loans and debt securities – resident banks (rhs) Source: Banco de Portugal. Notes: Annual rates of change are based on the relation between end-of-month outstanding amounts (adjusted for securitisation operations) and monthly transactions. Monthly transactions correspond to the difference in the end-of-month outstanding amounts adjusted for reclassifications, write-offs/write-downs, exchange rate and price revaluations, and any other variations that do not correspond to financial transactions. Whenever relevant, figures are additionally adjusted for sales of credit portfolio. 35 The Portuguese economy in 2016 Different credit standards according to risk pro- loans (Chart 3.16). By contrast, other services con- file have helped reduce the concentration of loan tributed with nearly 0.2 p.p. stocks in certain sectors. For example, the share of Contrary to 2015 and in particular 2014, loans (performing) loans to construction and real estate continued to decline (from a peak of 40.4 per cent of the total in 2007 to 24.6 per cent in 2016). granted to large firms with a credit event in the past again posted a negative rate of change during 2016 (Chart 3.17). As for loans granted to In terms of the stock of non-performing loans, con- micro, small and medium-sized firms with a cred- struction and real estate contributed with around it event in previous years, the decline in loans remained stable during the course of 2016. 0.6 p.p. in 2016 to the decline in non-performing 2.0 Chart 3.12 • Credit granted by resident banks to non-financial corporations by firm size 0.0 -2.0 -4.0 | Annual rate of change, percentage -6.0 -8.0 2010 Q1 2010 Q3 2011 Q1 2011 Q3 2012 Q1 2012 Q3 2013 Q1 2013 Q3 2014 Q1 2014 Q3 2015 Q1 2015 Q3 2016 Q1 2016 Q3 Micro, small, and medium sized firms Large firms and holdings Source: Banco de Portugal. Notes: Credit includes bank loans and debt securities held by resident banks. Figures are adjusted for securitisation operations, reclassifications, write-offs/write-downs, and exchange rate and price revaluations. Whenever relevant, the figures are additionally adjusted for sales of credit portfolio. 20.0 15.0 Chart 3.13 • Loans granted by resident banks to non-financial corporations by sector of activity | Annual rate of 10.0 5.0 0.0 -5.0 change, percentage -10.0 -15.0 Dec. 07 Jun. 08 Dec. 08 Jun. 09 Dec. 09 Jun. 10 Dec. 10 Jun. 11 Dec. 11 Jun. 12 Dec. 12 Jun. 13 Dec. 13 Jun. 14 Dec. 14 Jun. 15 Dec. 15 Jun. 16 Dec. 16 Total Manufacturing Construction and real estate activities Trade Source: Banco de Portugal. Notes: Annual rates of change are based on the relation between end-of-month outstanding amounts (adjusted for securitisation operations) and monthly transactions. Monthly transactions correspond to the difference in the end-of-month outstanding amounts adjusted for reclassifications, write-offs/write-downs, exchange rate and price revaluations, and any other variations that do not correspond to financial transactions. Whenever relevant, figures are additionally adjusted for sales of credit portfolio. 36 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Average risk for firms declined further in 2016 line with the recovery in the Portuguese economy. Although the decline in 2016 was smaller in size, all economic sectors made a positive contribution, in particular wholesale and retail trade, and electricity, gas and water. Average risk for firms with performing loans declined further in 2016, with an aggregate probability of default of 5.9 per cent, slightly below Corporate indebtedness decreased slightly, in parallel with a recovery in profitability the level observed in 2007 (Chart 3.18). After growing in 2012 and 2013, average risk posted a decline, in 2016, for the third year in a row, in 20.0 Chart 3.14 • Loans granted by resident financial institutions to private non-financial corporations by credit risk quartile | Year-on-year rate of change, percentage 15.0 10.0 5.0 0.0 -5.0 -10.0 -15.0 Dec. 07 Dec. 08 Dec. 09 Dec. 10 st Dec. 11 Dec. 12 nd 1 quartile (less risk) Dec. 13 Dec. 14 rd 2 quartile Dec. 15 Dec. 16 th 3 quartile 4 quartile (more risk) Source: Banco de Portugal. Notes: Credit risk is measured by the Z-score estimated according to Antunes, Gonçalves and Prego, ‘Firm default probabilities revisited‛, Banco de Portugal Economic Studies, Vol. 2, No 2, April 2016. The year-on-year rate of change is the annual rate of change of outstanding amounts in each month. 120 100 120 80 60 40 20 0 Chart 3.15 • Performing and nonperforming loans stock | Thousands millions euros 100 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 80 60 40 20 0 2006 2007 2008 2009 2010 Trade e Utilities – performing Other services – performing Tourism, transports and consulting – performing Construction and real estate activities – performing Manufacturing, Mining and Agriculture – performing 2011 2012 2013 2014 2015 2016 Trade e Utilities – non-performing Other services – non-performing Tourism, transports and consulting – non-performing Construction and real estate activities – non-performing Manufacturing, Mining and Agriculture – non-performing Source: Banco de Portugal. Notes: Credit risk is measured by the Z-score estimated according to Antunes, Gonçalves and Prego, ‘Firm default probabilities revisited‛, Banco de Portugal Economic Studies, Vol. 2, No 2, April 2016. The year-on-year rate of change is the annual rate of change of outstanding amounts in each month. Annual rates of change are based on the relation between end-of-month outstanding amounts and monthly transactions. 37 The Portuguese economy in 2016 The deleveraging process of non-financial cor- Statistics on non-financial corporations in the porations continued in 2016, with the level of Central Balance Sheet Database, for 2015 as a indebtedness standing at 105.5 per cent of GDP whole, confirm that their indebtedness ratio con- at the end of the year, compared with 109.6 per tinued to improve (Chart 3.19). With the excep- cent in the previous year. Although internation- tion of services and accommodation, this decline al comparisons of indebtedness ratios face sta- was broadly-based across economic sectors tistical challenges, Portuguese firms have delev- and is partly explained by firms exiting the eraged more strongly than their euro area coun- market (Box 3.2.1: ‘Market dynamics and the terparts. deleveraging of Portuguese firms‛). In parallel, 11 4 Chart 3.16 • Non-performing loans evolution by industry sector 3 2 |Percentage points annual contributions 1 0 -1 2007 2008 2009 2010 Tourism, transports and consulting Other services 2011 2012 2013 Construction and real estate activities Trade e Utilities 2014 2015 2016 Manufacturing, Mining and Agriculture Source: Banco de Portugal. Notes: Credit risk is measured by the Z-score estimated according to Antunes, Gonçalves and Prego, ‘Firm default probabilities revisited‛, Banco de Portugal Economic Studies, Vol. 2, No 2, April 2016. The year-on-year rate of change is the annual rate of change of outstanding amounts in each month. Annual rates of change are based on the relation between end-of-month outstanding amounts and monthly transactions. 30 Chart 3.17 • Credit to firms with a past default event 20 10 | Percentage points 0 -10 -20 -30 Sep. 07 Sep. 08 Sep. 09 Sep. 10 Sep. 11 Micro and SMEs Sep. 12 Sep. 13 Sep. 14 Sep. 15 Sep. 16 Large firms Source: Banco de Portugal. Note: This analysis includes the evolution of banking loans to firms that registered a credit default event during the period of 2006 to 2016. 38 BANCO DE PORTUGAL • Economic Bulletin • May 2017 a broadly-based increase was observed in the cor- construction and real estate continue to show porate profitability ratio. This improvement was high levels of indebtedness and low levels of prof- experienced by all economic sectors, with prof- itability. By contrast, manufacturing, mining and itability levels in 2015 standing at levels above quarrying and agriculture posted the lowest lev- those of 2007 in wholesale trade and transpor- el of indebtedness and the highest level of prof- tation, and services and accommodation. Devel- itability. In addition, the level of indebtedness in opments in the ratio of indebtedness to profita- services and accommodation grew consider- bility continued to differ across economic sectors. ably, without a significant change occurring in the Despite improvements compared with the past, sector’s profitability. 1.5 Chart 3.18 • Average risk for firms with only performing loans and industry contributions | Annual rate of change and contributions, percentage and percentage points 8% 1.0 6% 0.5 0.0 4% -0.5 2% -1.0 -1.5 0% 2006 2007 2008 2009 2010 2011 2012 2013 Tourism, transports and consulting Manufacturing, mining and agriculture Trade e utilities 2014 2015 2016 Construction and real estate activities Other services Average risk (rhs) Source: Banco de Portugal. 50 | In percentage 2011 40 2015 Indebtedness Chart 3.19 • Firm's indebtedness and profitability 2007 2011 2015 2011 30 2015 2015 2011 2007 2007 2007 20 0 2 4 6 8 10 12 14 16 18 Profitability Manufacturing, Mining and Agriculture Construction and Real estate activities Construction and Real estate activities Trade and transports Source: Banco de Portugal. Note: Firm's indebtedeness ratio calculated as the sum of total debt over total assets. Firm's profitability ratio calculated as EBITDTA (which stands for earnings before depreciations and amortizations, interest expenses, and income tax) divided by the sum of equity and total debt. The Portuguese economy in 2016 Box 3.2.1 | Market dynamics and the deleveraging of Portuguese firms The economic and financial crisis of 2008 highlighted the negative effects of the excessive indebtedness of Portuguese firms on financial stability and economic growth. The Economic and Financial Assistance Programme identified the private sector deleveraging, including the financial system, as an important factor for the adjustment of the Portuguese economy. In this context, the indebtedness of Portuguese non-financial corporations as a proportion of total assets, is estimated to have declined by approximately 4 p.p. to 49 per cent in the period between 2012 and 2015, accounting for 118 per cent of GDP by the end of this period.12 Simultaneously to the deleveraging process observed in the past few years in the private sector, the number of firms ceasing their activity increased.13 Therefore, it is important to understand to what extent market dynamics contributed to explain the deleveraging process of Portuguese firms. The firm balance sheet information reported in the survey Simplified Corporate Information (Informação Empresarial Simplificada – IES) was considered in the analysis.14 This dataset allows the identification of firms exiting and entering the market in the period between 2010 and 2015. A firm exit in year t is defined by its absence from the IES in that year, provided that this absence does not constitute a reporting gap.15 The entry of firms in the market is identified by the first year reporting to the IES. Chart 1 shows the number of new firms and firms that have ceased their activity in the period between 2010 and 2015. In the analysis, debt includes the funding obtained (loans and debt securities) by the firm and reported in the IES. According to this information, the total debt of non-financial corporations declined by around 3.4 percent in 2014 and 6.0 percent in 2015 (Chart 2). Chart 3 depicts the contributions to the change in the debt of non-financial corporations, considering the set of new firms, the set of firms that have ceased their activity, and the set of firms that remained active in the period between 2011 and 2015. The contribution of the set of new firms to the aggregate change in debt is small, in contrast to the substantial contribution of firms that have Chart 1 • Entry and exit of firms 20,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 2011 2012 Entrants (1ˢᵗ report in t) 2013 2014 2015 Exits in t (do not report in t) Source: Banco de Portugal. Note: The sample includes private non-financial corporations of all economic activity sectors except insurance and financial activities. 39 BANCO DE PORTUGAL • Economic Bulletin • May 2017 ceased their activity. The negative contribution of firms exiting the market to changes in debt was particularly evident in 2015.16 In turn, the debt of firms that remained active increased considerably in 2011 and less significantly in 2012, and decreased in the following years. Nevertheless, the exit of firms from the market does not necessarily mean that their debt amounts are removed from the debt outstanding amounts recorded in the balance sheets of financial institutions, which are the main source of information when analysing the growth of credit granted by the financial system. In fact, the existence of assets and personal guarantees allows the repayment of part of the debt even after the firm ceases its activity. This box does not analyse this difference between information reported to the IES and information in the balance sheets of financial institutions. Chart 2 • Total debt of private non-financial corporations | Annual rate of change 220,000 7 200,000 5 180,000 160,000 3 140,000 1 120,000 100,000 -1 Percentage EUR millions 40 80,000 -3 60,000 40,000 -5 20,000 -7 0 2010 2011 Total debt 2012 2013 2014 2015 Annual rate of change (right-hand scale) Source: Banco de Portugal. Note: Total debt includes loans and debt securities. The sample includes private non-financial corporations of all economic activity sectors except insurance and financial activities. Chart 3 • Contributions to debt change by firms' dynamics | EUR millions 10.000 5.000 0 -5.000 -10.000 -15.000 -20.000 2011 Entrants (1ˢᵗ report in t) 2012 2013 Stayers (report in t-1 and t) 2014 2015 Exits (do not report in t) Source: Banco de Portugal. Note: Total debt includes loans and debt securities. The sample includes private non-financial corporations of all economic activity sectors except insurance and financial activities. The Portuguese economy in 2016 Chart 4 shows the contributions to changes in debt by firm size. The debt reduction resulting from the exit of firms is concentrated in smaller-sized firms. Large firms that remained in the market in 2014 and 2015 also registered a reduction in debt. The decomposition of the contributions to changes in debt by sector of economic activity shows that the exit of firms in the construction sector had a substantial impact on debt reduction, especially in the last year (Chart 5). Firms in the trade sector that remained in the market in 2012 and 2013 significantly reduced their debt. The debt reduction resulting from the exit of firms in this sector of economic activity is particularly relevant in 2015. In the manufacturing sector, it is particularly relevant the debt reduction of firms that remained in the market in 2014 and firms that ceased their activity in 2015. Chart 4 • Contributions to debt change by firms' dynamics | By firm size, EUR millions SME Large corporations 6.000 4.000 2.000 0 -2.000 -4.000 -6.000 -8.000 -10.000 -12.000 5.000 4.000 3.000 2.000 1.000 0 -1.000 -2.000 -3.000 -4.000 -5.000 2011 Entrants (1ˢᵗ report in t) 2012 2013 Stayers (report in t-1 and t) 2014 2015 2011 Exits (do not report in t) 2012 Entrants (1ˢᵗ report in t) 2013 2014 Stayers (report in t-1 and t) 2015 Exits (do not report in t) Source: Banco de Portugal. Note: Total debt includes loans and debt securities. The sample includes private non-financial corporations of all economic activity sectors except insurance and financial activities. Chart 5 • Contributions to debt change by firms' dynamics | By sector of economic activity, EUR millions Construction Trade 2.000 2.000 1.000 0 -1.000 -2.000 -3.000 -4.000 -5.000 -6.000 1.000 0 -1.000 -2.000 -3.000 -4.000 -5.000 -6.000 2011 Entrants (1ˢᵗ report in t) 2012 2013 Stayers (report in t-1 and t) 2014 2015 Exits (do not report in t) 2011 Entrants (1ˢᵗ report in t) 2012 2013 Stayers (report in t-1 and t) 2014 2015 Exits (do not report in t) Manufacturing 2.000 1.000 0 -1.000 -2.000 -3.000 -4.000 -5.000 -6.000 2011 Entrants (1ˢᵗ report in t) 2012 2013 Stayers (report in t-1 and t) 2014 2015 Exits (do not report in t) Source: Banco de Portugal. Note: Total debt includes loans and debt securities. The sample includes private non-financial corporations of all economic activity sectors except insurance and financial activities. 41 Neste contexto de desalavancagem das empresas, importa aprofundar de que forma A análise empírica decompõe a evolução do end A análise empírica abaixoabaixo decompõe a evolução do endividam 42 a dinâmica de mercado contribuiu para o endividamento e rendibilidade das empresas. das empresas portuguesas no período compreendido das empresas portuguesas no período compreendido entre e BANCO DE PORTUGAL • Economic Bulletin • May 2017 A análise empírica abaixo decompõe a evolução do endividamento e da rendibilidade a dinâmica de mercado e idade. A análise catego com acom dinâmica de mercado edas idade. A análise categoriza das empresas portuguesas no período compreendido entre 2010 e 2015, de acordo Neste contexto de desalavancagem das empresas, importa aprofundar de Neste contexto de desalavancagem das empresas, importa aprofundar de qu Neste contexto de desalavancagem empresas, importa aprofundar de qq In this context of corporate deleveraging, it is important to know to what extent the market dynamics com atodinâmica mercado e idade. A análise categoriza as empresas de contributed corporatede indebtedness and profitability. The empirical analysis presented below adinâmica dinâmica de mercado contribuiu para oendividamento endividamento eque rendibilidade das a adinâmica de mercado contribuiu para oo endividamento eacordo das em mercado contribuiu para erendibilidade rendibilidade das eeo ade idade: empresas novas são asdetêm ou dois com acom idade: empresas novas são as que têm dois me and 2015, according to market dynamics and firm age. of InPortuguese the analysis, classified to composes changes in indebtedness and profitability firmsfirms in theare period betweenaccording 2010 com a idade: empresas novas são as que têmdecompõe dois ou menos anosdo no mercado, A análise empírica abaixo decompõe a evolução do endividamento eda da ren A análise empírica abaixo a evolução endividamento e eda rendi A análise empírica abaixo decompõe a evolução do endividamento rend their age: young firmsaccording are those have been in the market for years ortêm less, vintage firms are and 2015, to that market dynamics and firm age. Insão the two analysis, firms are classified empresas vintage são as que têm according mais dethose dois anos e empresas vintage as que mais de dois anos e meno to their age: young firms are those that have been inde thedois market for years or less, vintage firms with more than two years and less than five, and firms are those aged more than five years. empresas vintage são as que têmestablished mais anos etwo menos de cinco, e empresas das empresas portuguesas no período compreendido entre 2010 2015, das empresas portuguesas no período compreendido entre 2010 e ee 2015, de das empresas portuguesas no período compreendido entre 2010 2015, dd are thoseequation with morewas thanestimated: two years and less than five, and established firms are those aged more Then, the following estabelecidas são astêm quemais têm de mais de cinco De s estabelecidas são as que cinco anos.anos. De seguida sãofollowing as que têm mais deestimated: cinco anos. De a seguinte com dinâmica demercado mercado eidade. idade.AAestimou-se Aanálise análisecategoriza categoriza asempresas empresasde com a aadinâmica de e eseguida, asas com dinâmica de mercado idade. análise categoriza empresas dd than estabelecidas five years. Then, the equation was equação: com idade:empresas empresasnovas novassão sãoasas asque quetêm têmdois doisou oumenos menosanos anosno nom a aaidade: com idade: empresas novas são que têm dois ou menos anos no equação: �� � �� � �� � �� �equação: ��com ���� ��� � �� � �� ��������� � �� ������� � �� ��������� � ��� empresas vintage são as que têm mais de dois anos menos de cinco, empresas vintage são as que têm mais de dois anos e ee menos de cinco, e ee em empresas vintage são as que têm mais de dois anos menos de cinco, e where ݕ௧two assumes indicators: of debt total to assets) profitability ൌ ߚindicators: ߛ௦ ߚindebtedness ܽ݀ܽݎݐ݊ܧ ߚଷassets) ܰܽݒ ߚand ݑ௧ of ߙtwo ߜ ௧ indebtedness ଵ ܵܽÀ݀ܽ(ratio ௧ ߚof ଶ(ratio ௧ debt ௧ ସ ܸ݅݊݁݃ܽݐ ௧ (ratio where ��� assumes total to and profitability ݕ ൌߙߚ ߜߙ ߛߜ ߚߛܵܽÀ݀ܽ ߚ ܵܽÀ݀ܽ ߚ ܽ݀ܽݎݐ݊ܧ ߚ ݕൌߚ ߚ ܽ݀ܽݎݐ݊ܧ ߚ ܰܽݒ ௧. The Exit ௧mais ௦ de ଵ ௧anos. variables Entry are binary and ଶ௧ equal toଶ (ratio of EBIT to assets) of firm ௧in year as ௧que ௦and ଵde ௧ ଷ௧ estabelecidas são as que têm mais de cinco anos. De seguida, estimou-se estabelecidas que têm cinco De seguida, estimou-se a asaଷ estabelecidas são as têm mais cinco anos. De seguida, estimou-se EBIT17 to assets) of firm i in year t. The variables Exitsão and Entrant are binary variables equal to one for one for firms exiting and entering the market respectively.The variables Exit and Entrant are binary ementering que ݕ௧ assume dois respectively. indicadores: endividamento (rácio and da dívida total ativo) e firms exiting and the market, The variables Young Vintage arenoalso binary equação: equação: variables equal to one for firmsequação: exiting and entering the market, respectively. The variables Young variables equal rendibilidade to one for firms less two old firms with two to fivevariáveis years, respectively. 17 ݕyears em ݕfor assume indicadores: endividamento em que assume dois indicadores: endividamento (rácio(rá d ௧to dothan EBIT noque ativo) daand empresa idois notwo ano t. As Saída e and Vintage are also (rácio binary variables equal one ௧ firms less than years old and firms with Finally, the terms �� , �� , and �� denote a set of firm, time, and sector of economic activity fixed effects, two to five years, respectively. Finally, the , a set time, and ܵܽÀ݀ܽ ܽ݀ܽݎݐ݊ܧ ܰܽݒ ܸ݅݊݁݃ܽݐ ൌ ߚterms ߙߙ ߜ௧ߜߜand ߛ௦ߛߛ௦ ߚଵߚߚܵܽÀ݀ܽ ߚfirm, ߚଷߚߚܰܽݒ ߚସߚߚ ܸ݅݊݁݃ܽݐ ݕ௧ ൌൌ ߚߚ ߚߚܽ݀ܽݎݐ݊ܧ ܽ݀ܽݎݐ݊ܧ ݕݕ௧௧ ௧ ௦denote ଵܵܽÀ݀ܽ ௧of ௧sector ௧ ߙ ௧ ௧ ଶ ௧no ௧ ௧ ௧ ௧ ଵ ଶଶentram ௧ ଷଷܰܽݒ ସସܸ݅݊݁݃ܽݐ Entrada binárias e effects iguais aaccount um para empresas que mercado, 17 respectively. Firm and são sector fixed for their specific (observed rendibilidade (rácio dosaem EBIT noand ativo) da empresa in rendibilidade (rácio do EBIT noe17 ativo) daunobserved) empresa i no ano 17 of economic activity fixed effects, respectively. Firm and sector fixed effects account for their spe- characteristics, provided they remain constant over time. Insão turn, time effects attempt to capture the cific (observed and unobserved) characteristics, provided they remain constant over time. In(rácio turn, respetivamente. As variáveis Nova eassume Vintage também binárias e assumem o valor em que assume dois indicadores: endividamento (rácio da dívida total n em que ݕ௧ dois indicadores: endividamento da dívida total no em que ݕݕ௧௧assume dois indicadores: endividamento (rácio da dívida total no cyclical position of the economy. In this case, the comparison group are the established firms that Entrada são binárias e iguais a um para empresas que Entrada são binárias e iguais a um para empresas que saem time effects attempt to capture the cyclical position of the economy. In this case, the comparison 17 1717 de umIn no caso de rendibilidade empresas novas eobservations com idades entre dois e percentile cinco anos, rendibilidade (rácio do EBIT no ativo) da empresa no ano As variávei ativo) da empresa i no ano t. t.t. As variáveis (rácio do EBIT (rácio do ativo) da empresa i i no ano As variáveis remained ingroup activity. order to avoid extreme values, above the 99th and are the established firms rendibilidade that remained in activity. InEBIT order no tono avoid extreme values, obser- st As variáveis below the 1st percentile notrespetivamente. considered. The was performed using the fixed-effects respetivamente. As variáveis Nova eanalysis Vintage são també e Vintage são bin vations above thewere 99th percentile and below the 1analysis were notNova considered. The respetivamente. Finalmente, os termos ߙpercentile , ߜe௧ eee ߛiguais um conjunto de efeitos datambém Entrada são binárias um para empresas que saem entram no Entrada são binárias iguais a aa um para empresas que saem e ee entram no m Entrada são binárias iguais um para empresas que saem entram no ௦ denotam estimator for panel data. using the fixed-effects estimator for panel data. was performed empresa, tempo e setor, respetivamente. Os efeitos da empresa e do setor respetivamente. As variáveis Nova eVintage Vintage são também binárias eassume assume respetivamente. As variáveis Nova e eVintage são também binárias e eassumem respetivamente. As variáveis Nova são também binárias deinno um no caso de empresas novas e com idade de um caso de empresas novas epermitem com idades ent of the estimation are reported in Table According to estimates, firms exiting the The resultsThe of results the estimation are reported Table 1.1.According tothe the estimates, firms exiting the are the most and the least profitable. In turn, firmsthat that have have recently entered market are market the controlar most indebted and least profitable. In firms recently entered thedois paraindebted as the respetivas características específicas (observadas e entre não de umno no casode deturn, empresas novas comidades idades entre doise eecinco cin de um caso empresas novas e ee com de um no caso de empresas novas com idades entre dois cin respetivamente. Finalmente, termos e ߛ௦ denot respetivamente. Finalmente, termos ߙ , firms ߜexiting e ,ߛ௦ߜthe um are less and less profitable older incumbent firms. Inos particular, ௧denotam ௧ߙ market are the lessmarket indebted andindebted less profitable than olderthan incumbent firms. os In particular, firms observadas), desde que sejam constantes ao longo do tempo. Por sua vez, os efeitos respetivamente. Finalmente, os termos ߙ ,௧ߜߜprofitable e௦older denotam um conjunto de respetivamente. Finalmente, osos termos ,ߜ um conjunto de efe respetivamente. Finalmente, termos ,than ߛߛ௦denotam um conjunto de the market33.9 are approximately 33.9 p.p.and more indebted and 4.93 p.p.ߙߙ than ௦ denotam less ௧e௧ eߛ market are exiting approximately p.p. more indebted 4.93 p.p. less profitable incumbent older incumbent firms, on average. This selection dynamics suggests that the in the empresa, tempo e da setor, respetivamente. Os efeitos da empresa, tempo ethe setor, respetivamente. Os efeitos da empr firms, on average. This selection suggests that improvement inimprovement the aggregate financial temporais têm emdynamics consideração atempo posição cíclica economia. Note-se que neste empresa, esetor, setor, respetivamente. Os efeitos da empresa edo do setor empresa, tempo e esetor, respetivamente. Os efeitos da empresa e edo setor pe empresa, tempo respetivamente. Os efeitos da empresa setor aggregate financial position of Portuguese firms is partly due to the fact that the most indebted position of Portuguese firms is partly due to the fact that the most indebted and less profitable firms and less have exited the market. casoprofitable o grupofirms de comparação são as empresas estabelecidas que permaneceram no específicas controlar para as respetivas características espe controlar para as respetivas características controlar para as respetivas características específicas (observadas controlar para asas respetivas características específicas (observadas controlar para respetivas características específicas (observadas have exited the market. mercado. Para evitarobservadas), a presençadesde de valores extremos, nãoao se consideraram observadas), desde que sejam constantes ao longo do tempo. Por sua vez, que sejam constantes longo do tempo. Por sua vez, oso observadas), desde que sejam constantes ao longo do tempo. Por sua vez, Table 1 • Fixed effects estimates observadas), que sejam constantes ao longo do t observadas), desdedesde que sejam constantes ao longo do tempo 12. See Box “Developments in corporate indebtedness in Portugal and the euro area”, Economic Bulletin, May 2016, Banco de Portugal. The analysis considers temporais têmem em consideração posição cíclica daeconomia. economia.Note-se Note-sequ temporais têm consideração cíclica temporais têm em consideração aaposição posição cíclica da economia. Note-se qq data from the national financial accounts and the measure of indebtedness considered is the ratio of debt (loansaand debt securities) to totalda assets. Indebtedness Profitability temporais têmconsideração emde consideração a posição da e temporais a posição cíclicacíclica da econo 13. Félix, S. (2017) “Firm creation and survival in Portugal”, Banco de Portugaltêm Economicem Studies, Banco Portugal. caso ogrupo grupo de comparação são as empresas que permane caso o ogrupo de comparação são asas empresas estabelecidas permanece caso de comparação são empresas estabelecidas que permane 14. The survey IESEntrant includes economic, financial, and accounting information of-0.0332*** virtually all the Portuguese non-financial corporations and isestabelecidas available for the que -0.0413*** period between 2005 and 2015. Nevertheless, changes in the accounting standards system in 2010 have led to a loss of data comparability. caso oPara grupo de são as empresas estabel caso o grupo deevitar comparação são as empresas estabelecidas (0.0031) (0.0018) mercado. Para evitar presença devalores valores extremos, nãose seconsid con Para a aacomparação presença de extremos, não mercado. de valores extremos, não se cons 17 EBIT é o acrónimo para earnings before interest and taxes. is absent 15. A reporting gap is observed when the firm reports to themercado. IES in year t-1 and year t+1,evitar but in presença year t. Exit 0.3386*** -0.0493*** 16. Some of the absences that are considered exits from the market in 2015 may constitute reporting gaps. This can only be verified when the IES for 2016 65 mercado. Para evitar a (0.0024) presença de valores extre Paragaps evitar aunderpresença de valores extremos, becomes available. However, only about 0.2 per cent ofmercado. the observations are reporting in the period analysis. (0.0047) 17. EBIT stands for Earnings Before Interest and Taxes. Young -0.0603*** -0.0543*** (0.0037) (0.0026) Chart3.2.1Entryandexitoffirms 17 1717 EBIT é o acrónimo para earnings before interest and taxes. EBIT é o acrónimo para earnings before interest and taxes. EBIT é o acrónimo para earnings before interest and taxes. -0.0223*** -0.0067*** Entrants(1stVintage yearint) Exitsint(absentint) (0.0025) (0.0016) Source:BancodePortugal. Observations 2,001,156 1,063,059 Note:Thesampleincludesprivatenon‐financialcorporationsofalleconomicactivitysectorsexcept 17 EBIT é o acrónimo para earnings before interest and taxes. 17 EBIT é o acrónimo para earnings before interest and taxes. insuranceandfinancialactivities. Note: *, **, *** denote statistical significance at 1, 5, and 10 percent, respectively. Both specifications include firm, sector of economic activity, and time fixed effects. Robust standard errors clustered at firm level in parentheses. Chart3.2.2Totaldebtofprivatenon‐financialcorporations|Year‐on‐yearrateofchange Totaldebt Year‐on‐yearrateofchange(right‐handscale) Source:BancodePortugal. Note:Totaldebtincludesloansanddebtsecurities.Thesampleincludesprivatenon‐financialcorporations ofalleconomicactivitysectorsexceptinsuranceandfinancialactivities. The Portuguese economy in 2016 4.Fiscal policy and situation The general government deficit declined significantly, against a background of a stabilisation in the structural balance and a reduction in the primary structural balance In 2016, the general government deficit stood at 2.0 per cent of GDP, which is a historically low level (Table 4.1). This figure stands below the official target included in the State Budget for 2016 (2.2 per cent of GDP), the revised estimate included in the State Budget for 2017 (2.4 per cent of GDP), and the limit laid down by the European Union Council in August 2016 (2.5 per cent of GDP). the impact of temporary measures (Chart 4.1). In 2015 the general government balance was negatively affected by the resolution of BANIF (-1.4 per cent of GDP). In 2016 the effect of temporary measures corresponds to revenue obtained from full payments made within the scope of a special programme for the settlement of overdue tax and social contributions (Programa Especial de Redução do Endividamento ao Estado – PERES) and the refund by the European Financial Stability Facility (EFSF) of amounts previously paid by Portugal under the Economic and Financial Assistance Programme. 18 Overall, these effects contributed to increase the fiscal balance in 2016 by 0.4 per cent of GDP. Therefore, the decline in the fiscal deficit adjusted for temporary measures The deficit improved by 2.3 p.p. vis-à-vis the pre- stood at 0.5 p.p. of GDP, 0.3 p.p. of which refer vious year, chiefly as a result of the change in to a drop in interest expenditure. Table 4.1 • Main fiscal indicators | In percentage of GDP 2010 Overall balance (1) Temporary measures and special factors (a) (2) Overall balance excluding temporary measures and special factors (3 = 1-2) Cyclical component (4) Structural balance(b) (5 = 3 - 4) Interest expenditure (6) 2011 2012 2013 2014 2015 2016 Change 2016-2015 (a) -11.2 -7.4 -5.7 -4.8 -7.2 -4.4 -2.0 2.3 -2.8 -0.2 0.0 0.3 -3.6 -1.4 0.4 1.8 -8.4 -7.2 -5.6 -5.2 -3.6 -3.0 -2.4 0.5 1.5 0.6 -1.7 -2.3 -1.4 -0.6 -0.1 0.5 -9.9 -7.9 -3.9 -2.9 -2.2 -2.4 -2.3 0.0 2.9 4.3 4.9 4.9 4.9 4.6 4.2 -0.3 Structural primary balance (7=5+6) -7.0 -3.5 1.0 2.0 2.7 2.2 1.9 -0.3 Structural revenue (percentage of trend GDP) (b) 39.9 42.0 43.1 44.9 44.9 44.1 42.5 -1.6 Structural primary expenditure (percentage of trend GDP) (b) 47.1 45.6 42.2 43.0 42.3 42.0 40.6 -1.3 Public debt 96.2 111.4 126.2 129.0 130.6 129.0 130.4 1.4 Change in public debt (in p.p.) 12.6 15.2 14.8 2.8 1.6 -1.6 1.4 – (-) Primary balance 8.2 3.1 0.8 0.0 2.3 -0.2 -2.2 – Differential between the effects of interest and of GDP growth 0.8 6.4 10.0 3.5 2.8 -0.1 0.5 – Deficit-debt adjustments 3.5 5.8 4.0 -0.6 -3.5 -1.3 3.1 – Sources: INE and Banco de Portugal. Notes: (a) Special factors are operations with an one-off effect on the general government deficit which cannot be treated as temporary measures according to the definition adopted in the Eurosystem. (b) Structural figures are adjusted for the impacts of the cycle, temporary measures and special factors. The cyclical component and temporary measures are computed by Banco de Portugal according to the methodologies adopted in the Eurosystem. 43 44 BANCO DE PORTUGAL • Economic Bulletin • May 2017 According to the cyclical adjustment methodol- terms, both the tax and contributory burden and ogy adopted in the Eurosystem, economic activ- the non-tax revenue declined as a ratio to trend ity gave a positive contribution to fiscal develop- GDP (by -0.9 p.p. and -0.7 p.p. respectively). ments, accounting for 0.5 p.p. of the deficit improvement. Therefore, the structural balance is estimated to have stabilised in 2016.19 As regards the structural primary balance, which excludes the impact of interest expenditure, it is estimated to have declined for the second year in a row (by 0.3 p.p. of trend GDP). The drop in structural revenue largely stemmed from tax revenue developments In actual terms, tax and contributory revenue rose by 2.2 per cent in 2016, reflecting an The decline in the structural primary balance increase in revenue from taxes on production in 2016 reflects a drop in structural revenue by and imports and social contributions. Econom- 1.6 p.p. of trend GDP, which more than offset the ic activity is estimated to have given a positive 1.3 p.p. contraction recorded by structural primary contribution to these developments. They also expenditure (Chart 4.2). In particular, in structural reflect revenue collected within the scope of 6.0 Chart 4.1 • Breakdown of the change in the general government balance 4.0 | In percentage points -2.0 of trend GDP 2.0 0.0 -4.0 -6.0 2011 2012 2013 Structural primary expenditure ( - ) Temporary measures Structural balance (p.p. of GDP) 2014 2015 Interest expenditure ( - ) Structural revenue 2016 Cyclical component Overall balance (p.p. of GDP) Sources: INE and Banco de Portugal. 50 Chart 4.2 • Structural revenue and primary expenditure developments | In percentage of 45 40 trend GDP 35 30 1995 1997 1999 2001 2003 Structural revenue Sources: INE and Banco de Portugal. 2005 2007 2009 2011 2013 Structural primary expenditure 2015 The Portuguese economy in 2016 The decrease in general government expenditure chiefly reflected a decline in investment PERES, part of which, as previously mentioned, has a temporary nature. The structural tax and contributory burden, which is adjusted for the impact of the cycle and temporary measures, stabilised in 2016, implying a decline by 0.9 p.p. of trend GDP (Chart 4.3). These developments are chiefly explained by a drop in revenue from the taxes on income and wealth (-0.8 p.p.), reflecting the impact of policy measures implemented not only in 2016, but also in previous years (Box 4.1: ‘Structural developments of tax revenue in 2016’). 45 The structural decline in primary expenditure was chiefly driven by a decrease in capital expenditure. Adjusting for the effect of temporary measures with an impact on 2015, the decline in capital expenditure was 0.9 p.p. of trend GDP.20 These developments were largely due to the decline in public investment (Chart 4.4). In 2016, in particular, public investment declined by 30.9 per In 2016 non-tax revenue declined by 6.3 per cent in nominal terms, which partly mirrored a cent, resulting from a drop in other current rev- base effect resulting from the recording of real enue (including sales of goods and services) and estate acquisitions by Oitante in 2015, within the capital revenue. The 3.7 per cent decline in oth- scope of the resolution of BANIF, and the record- er current revenue reflects decreases in most ing of military material deliveries in 2016.21 The of its items, particularly as regards interest and change in public investment adjusted for these dividends. Capital revenue, in turn, declined by factors stands at -24.8 per cent. This decline may 27.9 per cent in actual terms. This occurred in have partly reflected a decrease in funds received spite of the refund of the margins previous- from the European Union and it was particularly ly paid to the EFSF, which, as previously men- relevant in the Autonomous Services and Funds tioned, has a temporary impact. Therefore, the subsector. Turning to other capital expenditure change in capital revenue adjusted for tempo- adjusted for the impact of temporary measures, rary measures would be more negative in 2016, there was also a decline as a ratio to trend GDP at -51.0 per cent, as a result of the significant (0.3 p.p.), reflecting several of base effects that fall in capital transfers from the European Union contributed to a one-off growth of this expendi- matched by expenditure in the same year. ture item in 2015.22 3.0 Chart 4.3 • Breakdown of the change in structural revenue 2.0 1.0 | In percentage points of trend GDP 0.0 -1.0 -2.0 2012 2013 Taxes on households' income VAT Social contributions 2014 2015 2016 Taxes on firms' income Other taxes on production and imports Other revenue (a) Sources: INE and Banco de Portugal. Note: (a) Other revenue encompasses ‘other current‛ revenue, including sales, and ‘capital revenue‛. 46 BANCO DE PORTUGAL • Economic Bulletin • May 2017 In 2016, expenditure on social benefits in cash pension expenditure increased by 1.5 per cent. increased by 1.3 per cent, both in actual and Whereas pension expenditure paid by CGA (pub- structural terms, and declined by 0.2 p.p. as a lic employees’ pension system) virtually stabi- ratio to trend GDP. Expenditure on unemploy- lised due to a lower growth of new retirements, ment benefits decreased by 14.3 per cent in in the general pension system expenditure grew actual terms (14.1 per cent in structural terms). by 2.4 per cent. This is partly explained by the This was due to a significant decline in the num- increase in the number of old-age pensioners ber of subsidised unemployed, which was more and the update of the lowest pensions, for the marked than as regards the overall number of first time since 2010. Social benefits in kind, in turn, unemployed individuals in the economy, and to declined by 0.7 per cent, as a consequence of a slight increase in the average benefit. In 2016, lower expenditure in the health sector. 2.0 Chart 4.4 • Breakdown of the change in structural primary expenditure | In percentage points of trend GDP 1.0 0.0 -1.0 -2.0 -3.0 -4.0 2012 2013 2014 2015 Investment Intermediate consumption Wage bill 2016 Other primary expenditure (b) Pension expenditure Change in structural primary expenditure Sources: INE and Banco de Portugal. Notes: (a) Other primary expenditure includes social payments excluding pensions, general government social contributions, subsidies and other current and capital expenditure. 30.0 Chart 4.5 • Breakdown of the change in public debt-to-GDP ratio | In percentage points of GDP 25.0 20.0 15.0 10.0 5.0 0.0 -5.0 -10.0 2012 2013 Primary balance effect Deficit-debt adjustments Sources: INE and Banco de Portugal. 2014 2015 GDP growth effect Change in debt 2016 Cumul: 2012-2016 Interest effect The Portuguese economy in 2016 In 2016 compensation of employees and inter- in advance in 2016 (2.4 per cent of GDP). This mediate consumption remained unchanged as repayment contributed to reduce the share of a ratio to trend GDP. Compensation of employ- debt held by non-residents, from 66.5 per cent ees increased by 2.8 per cent, after two years of at the end of 2015 to 58.2 per cent at the end decline. This is explained by the growth in wages of 2016 (Chart 4.6). This effect was reinforced by and salaries, as a result of the full reinstatement the increased importance of Banco de Portugal of the wage cuts in force since 2011 and, to a as a holder of Portuguese public debt, reflecting lesser extent, of developments in the number the widening of the Eurosystem’s asset purchase of general government employees.23 Indeed, programme to include sovereign bonds.24 More in 2016, against a backdrop of a decrease in the number of weekly hours worked in general government, from 40 to 35, the number of employees recorded, for the first time since 2011, a positive quarterly average change. In the case of intermediate consumption, the 2.9 per cent increase was strongly influenced by the rise in expenditure related to public-private partnerships in the road sector, which was again significant. Excluding these outlays, this expenditure item decelerated. specifically, between 2014 and 2016 the percentage of public debt held by the Central Bank increased from 0.7 to 8.8 per cent, implying an increase in the Central Bank’s exposure to sovereign risk. This occurred against the decline in the share of non-residents and, to a lesser extent, of other MFIs as debt holders. In the course of 2016, the Portuguese State participated regularly in sovereign debt markets. Against a backdrop of a widening in the differential between the cost of financing 10-year Por- The public debt ratio increased, reflecting a significant accumulation of deposits tuguese and German debt, the allotment rates At the end of 2016, the general government 2.3 in the previous year. As regards short-term debt stood at 130.4 per cent of GDP, 1.4 p.p. above the figure recorded at the end of the previous year. In a context where the primary balance contributed to reduce the debt ratio (by -2.2 p.p. of GDP), the change was chiefly explained by deficit-debt adjustments (3.1 p.p. of GDP). The fact that the implicit interest rate on debt exceeded the nominal GDP growth rate has also contributed to increase the ratio (by 0.5 p.p. of GDP) (Chart 4.5). As regards deficit-debt adjustments, the figure recorded in 2016 is mostly explained by the increase in deposits held by the general government by an amount equivalent to 2.4 p.p. of GDP. This deposit accumulation was likely related to the need to finance the capital injection by the State in Caixa Geral de Depósitos, which occurred at the beginning of 2017. Therefore, in 2016 the debt ratio net of general government deposits declined from 118.9 to 118.1 per cent of GDP. of long-term securities were higher than those recorded in 2015 (Chart 4.7). In 2016, in particular, the average rate on 10-year Treasury Bond auctions stood at 3.1 per cent, up from issues, the average interest rate on Treasury Bill auctions declined slightly from 0.04 to 0.02 per cent. General government debt servicing expenditure declined from 4.6 to 4.2 per cent of GDP in the year under review, which is underpinned by a reduction in the implicit interest rate from 3.6 per cent in 2015 to 3.3 per cent in 2016.25 These developments stem, on the one hand, from a decline in the interest burden associated with loans under the Economic and Financial Assistance Programme and, on the other hand, from the repayment of Treasury Bonds with higher rates than those of the new issues. The euro area structural primary balance declined for the second year in a row Similarly to developments in 2015, part of the According to the European Commission Winter loan granted by the IMF under the Economic forecasts, published in February 2017, the struc- and Financial Assistance Programme was repaid tural primary balance in the euro area declined 47 48 BANCO DE PORTUGAL • Economic Bulletin • May 2017 100 Chart 4.6 • Breakdown of public debt by holder | Percentage 90 80 70 60 50 40 30 20 10 0 2011 2012 Non-residents 2013 Central Bank 2014 Other monetary financial institutions 2015 2016 Other financial institutions Other residents Source: Banco de Portugal. 20.0 Chart 4.7 • 10 year government bond yields | Percentage 15.0 10.0 5.0 0.0 Germany Spain Sep. 16 Jan. 16 May 16 Sep. 15 Jan. 15 May 15 Sep. 14 Jan. 14 May 14 Sep. 13 Jan. 13 May 13 Sep. 12 Jan. 12 May 12 Sep. 11 Jan. 11 Portugal May 11 Sep. 10 Jan. 10 May 10 Sep. 09 Jan. 09 May 09 Sep. 08 Jan. 08 May 08 Sep. 07 Jan. 07 May 07 -5.0 Italy Source: Thomson Reuters. Chart 4.8 • Fiscal policy and cyclical position in the euro area in 2016 | In percentage points of potential GDP iscal stance (change in the structural primary balance) 6 Pro-cyclical tightening of fiscal policy Counter-cyclical tightening of fiscal policy 5 4 PT 12 3 2 EA 12 PT 14 1 EA 13 PT 13 0 -1 EA 14 PT 16 EA 16 Counter-cyclical loosening of fiscal policy EA 15 PT 15 Pro-cyclical loosening of fiscal policy -2 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 Cyclical position (change in the output gap) Source: European Commission (Winter 2017 Forecasts). Note: The cyclical position of the economy is assessed by the change in the output gap, which broadly corresponds to the difference between actual GDP and potential GDP growth rates. The Portuguese economy in 2016 slightly in 2016 (Chart 4.8). This reflects declines Pact, 2016 was the deadline for the correction in the structural primary balance in most Mem- of the excessive deficit in Portugal and Greece. ber States and particularly in Spain, Cyprus, Aus- In both cases, the nominal deficit stood below tria and Italy. In turn, Greece, Malta, the Neth- 3 per cent of GDP (Table 4.2). Spain and France, erlands, Latvia and Estonia followed fiscal con- whose correction deadlines are 2018 and 2017 solidation strategies in 2016. Regarding eco- respectively, recorded nominal deficits above nomic activity, the Commission estimates that the reference value. GDP growth in the euro area exceeded the In 2016 public indebtedness remained at a very change in potential GDP, although the difference between these aggregates remained negative. This implies that fiscal policy was pro-cyclical in 2016, for the second year in a row. 49 high level in most euro area countries (Chart 4.9). In particular, the public debt ratio stood below the 60 per cent of the GDP reference value in only 6 of the 19 Member States. Compared Within the European fiscal surveillance frame- with the previous year, there were important work, the excessive deficit procedures for Ireland, declines in Slovenia, Ireland, Germany and the Slovenia and Cyprus were abrogated in 2016. Netherlands. Nevertheless, in 2016 only Mal- As regards the euro area countries subject to ta recorded a public debt ratio below that of the corrective arm of the Stability and Growth 2007, before the outset of the financial crisis. Table 4.2 • Euro area countries in excessive deficit procedure | As a percentage of GDP Country European Council decision date Current deadline for correction Last change in the correction year Fiscal balance in 2016 Portugal 02-12-2009 2016 08-08-2016 -2.0 Spain 27-04-2009 2018 08-08-2016 -4.5 Greece 27-04-2009 2016 04-12-2012 0.7 France 27-04-2009 2017 10-03-2015 -3.4 Sources: Eurostat (Excessive Deficit Procedure – April notification) and European Commission. 200 180 Chart 4.9 • Public debt-to-GDP ratio in the Euro Area in 2016 160 140 120 | As a percentage of GDP 100 80 60 40 20 Source: Eurostat. Greece Italy Portugal Cyprus Belgium Spain France Euro area Austria Slovenia Ireland Germany Finland Netherlands Malta Slovakia Lithuania Latvia Luxembourg Estonia 0 50 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Box 4.1 | Structural developments in tax revenue in 201626 In 2016 the ratio of structural revenue from taxes and social contributions to trend GDP declined by 0.9 p.p. vis-à-vis the previous year. This stemmed from a decline in taxes on income and wealth and, to a smaller extent, in taxes on production and imports (Chart 1). In the year under review, structural revenue from social contributions as a ratio to trend GDP remained unchanged. Note that the decline in structural tax revenue occurred notwithstanding the positive contribution of the permanent component of receipts obtained within the scope of PERES (0.05 p.p. of GDP). In 2016 structural revenue from taxes on income and wealth declined by 0.8 p.p. of trend GDP, driven by the effect of legislative changes. In particular, revenue from taxes on households declined by 0.6 p.p. in structural terms, largely reflecting the elimination of the Personal Income Tax surcharge introduced in 2013 and the effect of the reform of this tax in 2015 (which impacted 2016 outturn via an increase in net refunds). The impact of the Personal Income Tax reform exceeded the initial estimates, which negatively affects the residual component. This component also reflects the decline in receipts from final withholding taxes on capital revenue resulting from the decrease in interest rates on financial investments. As regards the structural revenue from taxes on corporate income, there was a decline by 0.2 p.p. of trend GDP, reflecting measures approved in previous years. Indeed, the collection of the Corporate Income Tax was negatively affected by the decline in the statutory tax rate from 23 to 21 per cent (within the scope of the reform initiated in 2014), as well as by the change, in 2015, in the regime applicable to investment funds, which brought forward to that year revenue that would have been collected in 2016. These effects more than offset the overall positive impact of the change in the asset assessment regime included in the State Budget for 2016 and the permanent part of the Corporate Income Tax revenue collected under PERES (Chart 2). Chart 1 • Breakdown of the overall change in structural fiscal burden | In percentage points of trend GDP 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 2012 2013 2014 2015 Legislation changes Decoupling between the macro base and GDP (in nominal trend terms) Impact of the tax elasticity (in structural terms) Residual Change in structural tax burden Sources: INE and Banco de Portugal. 2016 The Portuguese economy in 2016 In structural terms, the ratio of revenue from taxes on production and imports to trend GDP declined by 0.1 p.p. in 2016. This was chiefly due to the fall in VAT structural revenue (-0.2 p.p.), which more than offset the increase in the collection of other indirect taxes (0.1 p.p.). In the case of VAT revenue, special reference should be made to the negative effect associated with the cut in the rate applicable to some restaurant services. The residual component of this tax is negative, reflecting the unusually high growth of refunds in the first months of 2016. However, this may have been partly offset by the possible overestimation of the effect of the above mentioned VAT rate cut. As regards the other indirect taxes, the structural increase in the ratio of revenue to trend GDP was largely due to several legislative changes included in the State Budget for 2016 (yielding an overall impact of 0.3 p.p. of trend GDP). Due to the magnitude of their estimated impacts, reference should be made to the contributions of the rises in taxes on oil products and on tobacco (0.2 p.p. and 0.1 p.p. of trend GDP respectively). A negative residual (-0.2 p.p.) possibly indicates the overestimation of these effects.27 Finally, the ratio of structural revenue from social contributions to trend GDP remained virtually unchanged in 2016, reflecting two symmetrical effects. On the one hand, the fact that the private sector trend wage bill grew less than nominal GDP should have given a negative contribution to structural revenue from social contributions. On the other hand, similarly to previous years, the residual component net of the effect of social contributions related to the civil servants regime was positive.28 This result may suggest gains from increased efficiency in the collection of contributions. Chart 2 • Breakdown of the change in structural taxes and social contributions in 2016 | In percentage points of trend GDP 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 Taxes on households' Taxes on firms' income income VAT Other taxes on production and imports Legislation changes Decoupling between the macro base and GDP (in nominal trend terms) Impact of the tax elasticity (in structural terms) Residual Total Sources: INE and Banco de Portugal. Social contributions 51 52 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Box 4.2 | Analysis of deviations in budget execution in 2016 In 2016 the general government deficit stood below the first official estimate for the year, presented in the State Budget for 2016, and the revised estimate included in the State Budget for 2017. However, in the course of the year, different fiscal projections released by international institutions suggested that the official target would be missed. Against this background, an analysis of the fiscal deviations gains particular relevance, allowing not only for a better understanding of the 2016 outcome, but also for the identification of some risks in budget execution in the current year. The State Budget for 2016 was published in early February and entered into force only in April. The estimate for the deficit in the National Accounts included in this document was thus based on information for 2015 with a lower than usual degree of uncertainty and appears to have benefitted from some information on revenue developments at the beginning of the year. The differential between budget outturn and this estimate turned out to be very small (0.2 per cent of GDP), but is underpinned by a significant negative deviation in total expenditure that is almost fully offset by a similar and also negative differential on the revenue side (Chart 1). Indeed, total general government revenue outturn fell significantly short of the amount envisaged when the State Budget for 2016 was prepared, reflecting, in particular, an overestimation of the collection of non-tax current revenue (including sales) and capital revenue. Growth of tax collection was also lower than expected, particularly as regards taxes on production and imports. In the latter case the deviation may reflect some overestimation of the impact of indirect taxation hikes included in the State Budget for 2016. Note that the decrease in tax revenue occurred in spite of the fact that developments in economic activity and, in particular, in private consumption were close to the projection included the State Budget for 2016. In addition, the deviation might have been more expressive if the PERES had not been implemented. Chart 1 • Budget execution in 2016 in national accounts: deviations vis-à-vis the State Budget for 2016 | In million euro Total revenue Taxes on income and wealth Taxes on production and imports Social contributions Other current revenue Capital revenue Total expenditure Social payments Compensation of employees Intermediate consumption Interest Subsidies Other current expenditure Investment Other capital expenditure Overall balance -2,500 Sources: INE and Ministry of Finance. -2,000 -1,500 -1,000 -500 0 500 1,000 The Portuguese economy in 2016 Total expenditure outturn stood also substantially below budgeted. A determining factor contributing to this result was the fact that the State Budget for 2016 stipulated a very significant amount of expenditure in the form of provisions, reserves and spending appropriations. This raises significantly the scope for discretion in the implementation of the budget and, given the limited nature of information available, hinders the monitoring of budget execution in the course of the year. For further details on the utilisation of budget control instruments in 2016, see Box 3 of the report ‘Analysis of the General Government Account 2016’, published by the Portuguese Fiscal Council. Both intermediate consumption expenditure and other current expenditure posted in 2016 much lower figures than those considered in the State Budget for 2016. Due to its magnitude, reference should also be made to the negative deviation in public investment, which was likely be affected by the limited execution of revenue from structural funds received from the European Union. Finally, similarly to previous years, expenditure on debt servicing stood below forecasts. In fact, although the budget pointed to the near stabilisation of this expenditure item in 2016, it actually declined vis-à-vis the previous year. 53 54 BANCO DE PORTUGAL • Economic Bulletin • May 2017 5.Supply Moderate growth of economic activity in 2016 path observed since early 2014. This increase As in the past few years, in 2016 Portuguese GDP accommodation and food services (Chart 5.3). per capita continued the convergence towards The dynamics observed in these sectors was the average EU level and currently stands at associated with very favourable developments around 71 per cent, still lower than observed in tourism exports (9.7 per cent in real terms, fol- at the onset of the international economic and lowing an 8.9 per cent increase in 2015) and the financial crisis (Chart 5.1). dynamics of domestic demand. The services con- In 2016 gross value added (GVA) at basic prices fidence indicator maintained a positive stance in mainly reflected 3.4 per cent activity growth in the subsectors trade and repair of motor vehicles and increased by 0.8 per cent year-on-year, following 2016 (Chart 5.4). 1.2 per cent growth in 2015. This increase consol- There was a 0.7 per cent year-on-year increase idated the moderate recovery path started at the in industrial activity, in contrast to a 2.0 per cent end of 2013, with overall developments consist- rise in 2015. However, this was not accompa- ent with the coincident indicator for activity and nied by an increase in industrial confidence over the economic sentiment indicator (Chart 5.2). the year. This notwithstanding, there was a deceleration vis-à-vis the previous year and, in this context, the GVA level was still around 3 per cent lower than the value recorded in 2008. In 2016 activity in the construction sector declined by 1.8 per cent, after a virtual stabilisation in 2015. In annual average terms, activity in this sector remained on a downward trend. This GVA in the services sector rose by 1.1 per cent trend reflected this sector’s structural adjust- year-on-year in 2016, as in the previous year. ment to a lower level of activity, following high Growth in this sector continued on the recovery investment in construction in previous decades. Chart 5.2 • GVA, Coincident indicator of activity and economic sentiment indicator 76 3 Year-on-year rate of change, in per cent 75 74 73 72 71 70 69 68 120 1 100 -1 80 -3 67 -5 Source: European Comission (Ameco). Note: EU15 refers to the initial 15 members of the European Union. 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 66 60 2008 2009 2010 2011 2012 2013 2014 2015 2016 Real GVA Coincident indicator of activity Economic sentiment indicator (r.h.s.) Sources: European Comission, INE and Banco de Portugal. Index 1990-2013 = 100 Chart 5.1 • GDP per capita at current prices, PPP adjusted, Portugal as a percentage of the European Union (EU15) The Portuguese economy in 2016 GVA in construction at the end of 2016 account- declined by 377 and 345 thousand individuals ed for little over half of the value recorded respectively (Chart 5.5). in 2008. The demographic trends in the Portuguese Downward trend of the total population and labour force economy cannot be easily reversed, and emigration phenomena tend to compound them. This is a significant drag on Portuguese economic growth, given that the labour factor is an input for production, and the population In 2016 the resident population and the labour dynamics itself stimulates activity in various force continued to decline, by 0.3 per cent (Ta- sectors. In addition, a more aged society puts ble 5.1). These developments continued to fol- pressure on the sustainability of social security low the downward trend observed since 2011, schemes. although slightly less markedly. Since 2008 the resident population and the labour force have decreased by approximately 250 and 350 thousand individuals respectively. Resident popula- Improvement in labour market conditions tion in the 25-34 age group declined by 2.5 per Labour market developments in 2016 continued cent in 2016, while the labour force fell by to be characterised by a rise in employment and 2.7 per cent. These falls are quite relevant, but a decline in the unemployment rate, maintaining less sharp than those seen in the past few years. the improvement seen from the second quarter In cumulative terms, since 2008 the popula- of 2013 onwards, in a context of more marked tion and the labour force in this age group have wage dynamics. Table 5.1 • Labour market indicators | Year-on-year rate of change, in per cent, unless otherwise stated Thousands of individuals in 2016 Population Population 25-34 years Years 2014 2015 Semesters 2016 S2 2015 S1 2016 S2 2016 10 306 -0,6 -0,5 -0,3 -0,5 -0,3 -0,3 1180 -3,6 -2,8 -2,5 -2,7 -2,7 -2,3 Labour force 5178 -1,1 -0,6 -0,3 -0,5 -0,7 0,1 Labour force 25-34 years 1055 -3,8 -3,1 -2,7 -3,1 -2,9 -2,4 73,2 73,4 73,7 73,5 73,4 74,0 Participation rate 15-64 years (in % of the population) Total employment 4605 1,6 1,1 1,2 0,9 0,6 1,8 Employees 3787 4,4 2,8 2,1 1,9 1,7 2,4 Self-employment 789 -8,2 -5,7 -3,2 -3,6 -5,0 -1,3 Total unemployment 573 -15,1 -11,0 -11,4 -9,7 -10,0 -12,8 Unemployment rate (in % of the labour force) 13,9 12,4 11,1 12,1 11,6 10,5 Unemployment rate 25-34 years (in % of the labour force) 15,5 13,1 12,5 12,5 13,4 11,6 Long-term unemployment (in % of total unemployment) 65,5 63,5 62,1 62,7 61,5 62,7 5,2 5,0 4,6 5,2 4,5 4,7 Discouraged inactives (in % of the labour force) Source: INE. Notes: Long-term unemployment includes the unemployed individuals that have been actively seeking employment for 12 months or more. The discouraged inactives include the inactive individuals who were available for work but had not looked for a job during the period. 55 56 BANCO DE PORTUGAL • Economic Bulletin • May 2017 According to data released by the Ministry of cent in 2015), maintaining the downward trend Labour, Solidarity, and Social Security, average started in 2013 and standing close to the level wages declared to Social Security in 2016 grew observed in 2010. In 2016 the decline in the by 1.6 per cent compared to the same period total number of unemployed year-on-year was a year earlier, which reflects a 1.0 p.p. accelera- particularly marked in the 35-44 age group. The tion from the previous year. This greater wage share of unemployed receiving unemployment dynamics should partly reflect the increase in benefits amounted to 25.5 per cent in 2016, the national minimum wage. The wage dynam- against 28.5 per cent in 2015 (Chart 5.6). ics plays a particularly important role in a con- In addition, the number of discouraged work- text where collective bargaining instruments, particularly of a sectoral nature, are still a few, although there has been a slight increase from the minimum level recorded in 2012. In Portugal the wages of around 90 per cent of employees are defined within the scope of these instruments. This greater wage dynamics seems to have also influenced the trend of the inflation rate in 2016 (Section 7: ‘Prices’). ers, i.e. individuals not actively seeking a job but who are available for work, accounted for around 4.6 per cent of the labour force in 2016, slightly below the value recorded in 2015 (5.0 per cent). The measure of the unemployment rate that also includes discouraged workers continued to evolve in parallel with that of the official unemployment rate, after having grown more rapidly than the latter up to 2013 (Chart 5.7). Nevertheless, this group is formed by around The unemployment rate declined, but remained at very high levels 238 thousand discouraged workers. One of the most negative aspects of Portuguese labour market developments in the past few years was the very high level of long-term unem- According to Statistics Portugal’s Labour Force ployment, which tends to cause a sharp depreci- Survey, in 2016 the total number of unem- ation of human capital, having an adverse impact ployed declined by 11.4 per cent year-on-year, on the economy’s potential growth. In 2016 the after a 11.0 per cent fall in 2015 (Table 5.1). The number of unemployed seeking a job for more unemployment rate was 11.1 per cent (12.4 per than 12 months fell by 13.4 per cent (13.7 per Chart 5.3 • GVA by main sectors of activity | Index 2008 Q1 =100 120 Chart 5.4 • Confidence indicators | Balances (quarterly mean) 20 10 0 -10 -20 -30 -40 -50 -60 -70 -80 110 100 90 80 70 60 50 2008 2009 2010 2011 2012 2013 2014 2015 2016 Total GVA Services Construction Source: INE. Agriculture, forestry, and fishing Manufacturing 2008 2009 2010 2011 2012 2013 2014 2015 2016 Manufacturing confidence indicator Construction confidence indicator Services confidence indicator Source: European Comission. Note: Seasonally adjusted figures. The Portuguese economy in 2016 cent fall in 2015). Nevertheless, the weight of decline in 2015). The share of these short-term these individuals in total unemployment has unemployed in the labour force already stands remained at a very high level (62.1 per cent, cor- at levels close to those observed prior to the responding to a total of approximately 356 thou- international financial crisis (Chart 5.8). sand individuals). In turn, very long-term unemployment, covering those seeking a job for more Employment recorded a positive trend, although remaining at historically low levels group of unemployed (47.5 per cent in 2016 on average).29 In the same vein, in 2016 the number of unemployed seeking a job for less than 12 months fell by 7.8 per cent (5.7 per cent Chart 5.5 • Population, labour force, and employment – Total and age group (25-34 years) | Index 2008 =100 Unemployed receiving unemployment benefits (as a percentage of total unemployment) 105 Chart 5.6 • Number of unemployed receiving unemployment benefits and coverage rate 100 95 90 85 80 75 70 2009 2010 2011 2012 2013 Population Employment Labour force (25-34 years) 2014 2015 45 350 40 300 35 250 30 25 200 20 150 15 100 10 50 5 0 0 2008 2009 2010 2011 2012 2013 2014 2015 2016 2016 Labour force Population (25-34 years) Employment (25-34 years) Coverage rate of unemployment benefits Number of unemployed receiving unemployment benefits (r.h.s.) Source: INE (Labour Force Survey). Source: INE (Labour Force Survey). 22 5 20 4.5 18 4 16 3.5 14 3 12 2.5 10 2 2011 2012 Official rate Source: INE (Labour Force Survey). Number of unemployed (Thousands) than 24 months, accounted for almost half of the 2008 57 2013 2014 Including discouraged workers 2015 2016 Difference (r.h.s.) Chart 5.7 • Official unemployment rate and unemployment rate including discouraged workers | Percentage 58 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Reduction in apparent labour productivity According to the Labour Force Survey, total employment increased by 1.2 per cent in 2016, after a 1.1 per cent increase in 2015. This reflects a higher number of employees (2.1 per cent The current recovery phase of the Portuguese more), given that self-employment fell markedly economy brings together moderate activity growth (3.2 per cent). In sectoral terms, the main con- and relatively stronger employment growth. In this tribution to the year-on-year rate of change in context, apparent labour productivity based on employment was recorded in tradable servic- the quarterly national accounts of Statistics Por- es, which include tourism-related activities.30 tugal declined in 2016, in the context of lower In spite of the upward trend, employment and atypical dynamics compared to prior phas- remained at historically low levels, in the wake es of economic recovery.31 However, from 2011 of an unprecedented fall observed between the to 2013 apparent labour productivity growth end of 2008 and early 2013, which according to was strong, also in contrast to prior recessive quarterly national accounts corresponded to phases of the cycle (Chart 5.9). The explana- around 650 thousand individuals. Approximate- tions for the reduction in productivity are com- ly 290 thousand jobs seem to have been recov- plex and may relate to changes in the produc- ered by the end of 2016. With respect to general tive structure, in a context where the levels of government employment, according to informa- capital per worker remained low, after several tion from the Directorate-General for Adminis- years of low investment levels (Box 5.1: ‘Capital tration and Public Employment, the number of per worker and productivity’). employees appears to have increased in 2016, as in 2015, which confirms the interruption of the downward trend observed in previous years. Chart 5.8 • Unemployment rate by duration of unemployment | Months 20 18 16 14 12 10 8 6 4 2 0 Chart 5.9 • Apparent labour productivity in the last three crisis and recoveries | Biannual figures; Minimum level of activity = semester t=100 108 106 104 102 100 98 96 94 2008 2009 2010 2011 2012 2013 2014 2015 2016 Unemployment rate Short-term unemployment (less than 12 months) Long-term unemployment (12 months or more) Source: INE (Labour Force Survey). t-4 t-3 t-2 t-1 t t+1 1993S1 t+2 2003S1 t+3 t+4 t+5 t+6 t+7 2013S1 Sources: INE and Banco de Portugal. Note: Labour productivity was computed dividing GVA (volume chain-linked data, millions of euro) by total (full-time equivalent) employment, except for the crisis in 1992, in which GDP was considered instead. The Portuguese economy in 2016 Box 5.1 | Capital per worker and productivity The low growth of apparent labour productivity in the Portuguese economy in the past few decades causes concern as it limits the ability to sustain higher levels of consumption without causing external imbalances. A significant catching-up of output per worker to the euro area average is a particularly important challenge, in a context where there are still weaknesses in the Portuguese economy. Developments in the level of output per worker relate to various structural issues whose interaction is complex. In its simpler version, the economic growth theory points to the accumulation of inputs and the evolution of the technological framework as determinants of output growth. In this context, the level of output per worker depends on technology and increases with the level of the available capital stock for each worker, with lower variations expected for higher capital intensities. Economic literature has addressed this issue, identifying more elaborate formulations that include, for example, the quality of inputs and more flexible assumptions on their substitutability. In addition, from an empirical viewpoint, the literature has shown weaknesses in the statistical information needed for this analysis, for example as regards capital stock measurement.32 The existing estimates for the capital stock level in economies typically result from the accumulation of annual investment flows, less an annual depreciation rate resulting from assumptions on the lifetime of the different capital goods. This box presents some statistical information to place the Portuguese economy in the context of the euro area in the 1995-2016 period in terms of the capital-labour ratio and how it relates to productivity, measured as the level of output per worker. Panel (a) of Chart 1 shows the cumulative trend of the capital-labour ratio since 1999 in Portugal, jointly with the cumulative trend of each component. The demographic contraction trends experienced by the Portuguese economy, and especially the sharp cyclical developments stemming from the macroeconomic imbalances’ adjustment process, led to a strong reduction in Chart 1 • Net capital stock per worker (2010 prices) b) Euro area countries 300 140 250 Thousand euros 150 130 120 110 100 200 150 100 90 50 80 0 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Index 1999=100 a) Portugal Capital stock Capital per worker Source: Ameco – European Commission. Employment LV LT SK SI MT EE PT CY DE GR IT BE ES FI IE LU FR AU NL 2016 1999 59 BANCO DE PORTUGAL • Economic Bulletin • May 2017 employment after 2008, only partly reversed from 2013 onwards. In turn, European Commission estimates show a contraction of the capital stock after 2011, as a result of insufficient investment levels to offset the depreciation of installed capital. In this context, the capital-labour ratio in the Portuguese economy started to decline after 2013, maintaining a downward path in 2016. Panel (b) of Chart 1 shows the capital-labour ratios in euro area countries in 1999 and 2016, with the Portuguese economy positioned in the lower segment of the distribution. Although there may be reservations as to the accuracy of capital stock estimates, especially in the countries that have most recently joined the European Union and the euro area, the Portuguese economy’s position is unfavourable. The relationship among euro area economies in terms of capital intensity and labour productivity levels for the average of the period 2007-2016 is shown in Chart 2. As was to be expected, the chart suggests a positive relationship between the two variables, and Portugal’s position is far behind the more advanced euro area countries. Hence, although it is important to stress that sectoral distribution and quality of investment are important determinants of growth, the reality seems to show a need for greater aggregate investment levels to increase the productivity of Portuguese workers. The analysis of the quality of the capital stock existing in the economies is difficult, and the statistical problems mentioned above remain. An imperfect assessment of the capital stock quality can be made from its structure. Chart 3 is based on data from the Penn World Tables 9.033 to show the composition of the capital stock by type of capital goods for euro area countries in the 2012-2014 period, according to four categories: (residential and non-residential) structures; machinery (including computers, communication equipment and other machines); transport and other equipment (including software, other intellectual property goods and cultivated assets). In this very elementary analysis, Portugal appears as the country with the greatest proportion of Chart 2 • Capital and GDP per worker in the euro area (average 2007-2016) | Thousand euros 120 LU 100 IE GDP per worker 60 80 DE 60 CY MT 40 SK LV 20 LT ES IT BE FI FR NL AU GR PT SI EE 0 0 50 100 150 Capital per worker Source: Ameco – European Commission. 200 250 300 The Portuguese economy in 2016 structures in the capital stock, which necessarily implies a lower share of equipment in machines and intangible goods. These latter categories tend to be associated with higher profitability rates, and thus there seems to be room for improving the composition of the national capital stock. Portuguese economic growth will benefit from higher levels of investment in the firms, sectors and types of goods with higher profitability. This investment effort should be primarily based on higher public and private saving rates, to avoid jeopardising the current balance in the external accounts, which is a key element of the Portuguese economy’s adjustment process of the past few years. However, creating favourable conditions for attracting foreign direct investment, especially targeted at creating new productive capacity, is also a way to increase the levels of capital per worker without compromising macroeconomic stability. The structural conditions needed to attract foreign direct investment are largely also needed for a correct allocation of investment towards more productive purposes. Efficient market functioning, a level playing field, the reduction of the so-called framework costs, and the stability of the regulatory and tax environments are all part of this set of conditions. Chart 3 • Structure of capital stock per type of investment (average 2012-2014) | Percentage 100 95 90 85 80 75 70 65 60 55 50 MT SK IE BE Structures Source: Penn World Tables. EE AU LU SI Machinery FI DE NL IT ES Transport equipment CY LT FR LV Other GR PT 61 62 BANCO DE PORTUGAL • Economic Bulletin • May 2017 6.Demand Moderate GDP growth based on the dynamics of domestic demand and especially of exports of the prior recession. In this context, the GDP In 2016 the Portuguese economic activity grew economic agents and the need to adjust their level at the end of 2016 was around 4 per cent lower than at the start of 2008. The recovery of economic activity takes place in the context of high indebtedness levels of public and private by 1.4 per cent in volume, following a 1.6 per balance sheets. cent growth in 2015 (Table 6.1). In 2016 as a The important demographic changes that oc- whole GDP growth in Portugal was once again curred in Portugal in the past few years ren- lower than in the euro area, the differential der the analysis of per capita GDP develop- standing at -0.3 p.p. (Chart 6.1). GDP growth in ments particularly relevant. In 2016 this varia- 2016 showed a markedly intra-annual profile, ble grew by 1.7 per cent (2.1 per cent in 2015), accelerating to 1.9 per cent in the second half which exacerbated the gap vis-à-vis GDP de- of the year, above the value observed on aver- velopments (Chart 6.2). As in 2014 and 2015, age in the euro area in the same period (1.7 per cent). GDP growth was once again higher than that of GVA, which in 2016 grew by 0.8 per cent (Chart Developments in Portuguese economic activity 6.3). The discrepancy between GDP growth and in the most recent period continued to be char- GVA growth related to the evolution of taxes less acterised by moderate growth, particularly tak- subsidies, which grew by 4.8 per cent in volume ing into account the seriousness and duration in 2016, compared to 5.1 per cent in 2015. Table 6.1 • GDP and main components | Year-on-year change in percentage, unless otherwise stated % of GDP in 2016 2015 2016 2016 GDP Domestic demand 2015 2016 H1 H2 Q1 Q2 Q3 Q4 100.0 1.6 1.4 1.0 1.9 1.0 0.9 1.7 2.0 98.8 2.5 1.5 1.1 1.8 1.5 0.8 1.0 2.5 Private consumption 65.8 2.6 2.3 2.1 2.5 2.5 1.6 1.9 3.0 Public consumption 18.0 0.7 0.5 0.8 0.3 1.2 0.5 0.2 0.3 Investment 14.9 4.7 -0.8 -2.1 0.6 -2.1 -2.1 -1.8 3.0 14.8 4.5 -0.1 -2.4 2.2 -2.5 -2.2 -0.1 4.5 0.0 -0.1 0.0 -0.2 0.1 0.0 -0.3 -0.2 GFCF Change in inventories (a) Exports 40.3 6.1 4.4 2.8 6.0 3.7 1.9 5.6 6.4 Imports 39.1 8.2 4.4 3.1 5.6 4.8 1.5 3.9 7.3 Contribution of domestic demand net of imports (a) 1.1 0.5 0.4 0.6 0.4 0.4 0.4 0.9 Contribution of exports net of imports (a) 0.5 0.9 0.6 1.2 0.7 0.6 1.3 1.1 0.5 1.3 0.2 0.2 0.9 0.7 2.5 1.6 1.1 2.0 1.4 0.8 1.3 2.7 Memo item: GDP – change over the previous period Domestic demand (exc. change in inventories) 98.7 Sources: INE and Banco de Portugal calculations. Note: (a) Contributions to the real growth of GDP in percentage points. The demand aggregates net of imports are obtained by subtracting an estimate of the imports needed to meet each component. The calculation of import content was based on data for 2005. For more information, see the Box entitled ‘The role of domestic demand and exports in economic activity developments in Portugal’, in the June 2014 issue of the Economic Bulletin. The Portuguese economy in 2016 Intra-annual acceleration path of economic activity, reflecting a recovery of both domestic demand and exports of goods and services in the second half of the year The less buoyant economic activity in 2016 reflects less robust domestic demand growth, largely due to a fall in investment, in parallel with a deceleration in exports of goods and services, in line with lower domestic demand growth. In this context, the reduction of trade flows with Angola had an impact on the Portuguese economy, with a particularly negative influence on exports of goods.34 In turn, private consumption grew slightly less than in the previous year, which was accounted for by the less buoyant consumption of non-food current goods and durable goods. Public consumption, determined by the dynamics of compensation of employees and expenses on goods and services in general government, grew by 0.5 per cent in 2016, i.e. slightly less than in 2015. The deceleration in some demand components with a greater import content translated into lower import growth in 2016 as a whole. 63 In spite of the lower dynamics over the year as a whole, developments in economic activity in 2016 were characterised by a clearly upward intra-annual profile. Hence, following 1.0 per cent growth in the first half of 2016, GDP increased by 1.9 per cent in the second half. The recovery of both domestic demand and exports of goods and services contributed to this evolution. The more buoyant domestic demand in the second half of 2016 was particularly evident in the last quarter, when both private consumption and investment accelerated. In a context of rapid expansion of the consumption of dura- Considering demand components less imports ble goods over the whole year, the consump- (deducting from each component an estimate tion of current goods accelerated in the sec- of the imports needed to meet said demand), ond half of the year. In turn, investment grew the contribution from domestic demand to slightly in the second half, after a fall in the first GDP growth is estimated to have amounted half of the year. This development had a major to 0.5 p.p. in 2016 and the contribution from contribution from the significant recovery in exports to have stood at 0.9 p.p. investment in machinery and equipment. Chart 6.1 • GDP growth in Portugal and in the euro area | Year-on-year change in percentage 4 Chart 6.2 • GDP and GDP per capita in Portugal | 2011Q1=100 102 3 100 2 1 98 0 -1 96 -2 -3 94 -4 -5 -6 2000 2002 2004 2006 Portugal 2008 2010 2012 2014 Euro area Sources: Eurostat, INE and Banco de Portugal calculations. 2016 92 2011 2012 2013 GDP 2014 2015 2016 GDP per capita Sources: INE and Banco de Portugal calculations. 64 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Private consumption followed an intra-annual recovery path over the year, with emphasis on consumption of current goods which accelerated considerably in the last quarter In 2016 private consumption grew by 2.3 per cent, i.e. 0.3 p.p. less than in 2015. Despite the lower dynamics, private consumption continued to grow more than GDP. The behaviour of consumption over the whole year was chiefly accounted for by a deceleration in consumption of current goods and services, in particular the non-food component, in a context where consumption of durable goods grew once Intra-annual developments in private consumption showed an upward profile, as a result of particularly pronounced growth in the last quarter of the year, based on an acceleration in the consumption of current goods and services and the maintenance of the high growth of durable goods, particularly sharp in the last quarter of the year. The more favourable developments in private consumption in the second half of the year occurred in an environment of sharp improvement in consumer confidence, which remained at historically high levels, and a higher household disposable income. The improvement in the consumer confidence indicator at the end of 2016 was particularly noticeable in the assessment of unemployment expectations, which lowered considerably. again considerably, though less than in 2015 In 2016 the savings rate stood at 4.4 per cent of (Chart 6.4). In 2016 the consumption of current disposable income, i.e. 0.1 p.p. below the rate goods and services grew by 1.6 per cent, after recorded in 2015 (Chart 6.6), reflecting slightly 1.8 per cent in 2015, while the consumption of higher nominal growth of private consumption durable goods grew by 9.5 per cent (11.9 per than that of disposable income (3.4 and 3.2 per cent in 2015). In particular, passenger car sales cent respectively). The increase in household dis- grew by 16.1 per cent (Chart 6.5). posable income benefited from an improvement Chart 6.3 • Recent behavior of GDP and GVA in Portugal | 2011Q1=100 Chart 6.4 • Main contributions for the growth of private consumption in volume | Year-on-year change in percentage, and contributions in percentage points; quarterly values 4 100 2 98 0 96 -2 -4 94 -6 92 2011 -8 2012 2013 2014 2015 GDP Sources: INE and Banco de Portugal calculations. 2016 GVA 2011 2012 2013 2014 2015 2016 Durables (exc. vehicles) Durables (vehicles) Non-durables Private consumption (growth in percentage) Sources: INE and Banco de Portugal calculations. 65 The Portuguese economy in 2016 in the labour market situation, especially in terms In 2016 the volume of investment fell by 0.8 per of employment, in parallel with a rise in the cent following a 4.7 per cent increase in 2015. national minimum wage and the impact of meas- This development involves a contraction of gross ures included in the State Budget for 2016 relat- fixed capital formation of 0.1 per cent, following ing to the reversal of wage cuts and the person- 4.5 per cent growth in the previous year. al income tax surcharge (Section 4: ‘Fiscal policy and situation’).35 Reference should also be made to the potential effect of the reduction of the debt service in the most recent period, stemming from the stabilisation at low levels of the interest rates prevailing in the household sector, After a sharp fall observed during the economic and financial crisis, the necessary recovery of GFCF in Portugal has been sluggish. This is also observed in other economies, namely the euro area economy, where the investment level is still jointly with a decline in indebtedness.36 Against lower than seen in 2008 (Chart 6.8). Although this background, households’ financing con- there is no consensus on the motives for these ditions improved, notably as regards consumer developments, they are particularly negative for credit (Section 3.2: ‘Monetary and financial condi- Portugal, not only because the fall in investment tions in Portugal’). The weight of total consumer during the crisis period was relatively sharper, credit in private consumption has been following but also because capital per employee in the an upward path, standing at values higher than Portuguese economy remains at low levels com- seen prior to the sovereign debt crisis (Chart 6.7). pared with the euro area average (Box 5.1: ‘Capi- It is an expectable development that nevertheless needs to be monitored with caution.37 tal per worker and productivity’). In this context, the higher uncertainty, both domestic and external, in late 2015 and early 2016, Fall in investment largely reflecting a decline in GFCF in construction in parallel with the maintenance of a high level of corporate indebtedness, appears to have contributed to restrain investment decisions, Chart 6.5 • Sales of light passenger vehicles | Thousands of vehicles; annual values Chart 6.6 • Savings rate and households’ disposable income | Percentage change and savings rate as a percentage of households’ disposable income 12 10 8 6 4 2 0 -2 -4 -6 300 250 200 150 100 50 2001 2003 2005 2007 2009 2011 2013 Sales of light passenger vehicles (sum of the previous 12 months) Average 2001-2016 Sources: INE and Banco de Portugal calculations. 2015 2000 2002 2004 2006 2008 2010 Private consumption deflator Nominal disposable income Private consumption (nominal) Sources: INE and Banco de Portugal calculations. 2012 2014 2016 Private consumption (real) Savings rate 66 BANCO DE PORTUGAL • Economic Bulletin • May 2017 namely in the first half of the year (Section 2: ‘In- the past few years, only interrupted in 2015, ternational environment’ and Section 3: ‘Mon- is part of a stabilisation trend of the Portu- etary and financial conditions’). However, in- guese housing stock, following a considerable vestment recovered in the second half of 2016, increase in the 1990s. This seems to have been particularly in the last quarter, with gross capital reinforced in the most recent period by a need formation growing by 4.5 per cent year on year. to reduce household indebtedness levels. By According to the Investment Survey released by contrast, corporate GFCF grew by 6.6 per cent Statistics Portugal, in January 2017, 39.4 per cent in volume in 2016, following a rate of change of of firms claimed to have had their investment re- 2.3 per cent in the previous year and showing strained in 2016. This was a historically low val- a markedly upward intra-annual profile. ue that reflected a considerable fall vis-à-vis the The fall in GFCF in 2016 was largely accounted share computed in July 2016 (50.2 per cent). In addition, although the deterioration of sales prospects continued to be the main constraint on investment, their weight decreased substantially backed by factors such as self-financing capacity or the difficulty to obtain credit (Chart 6.9). for by the construction component. In the wake of the declines observed since 2002, followed by a significant recovery in 2015, this investment component, whose weight in total GFCF was around 50 per cent, declined once again sharply. This may partly reflect the effect of the By institutional sector, the reduction in gross conclusion of a number of major public works fixed capital formation in 2016 as a whole at the end of 2015. GFCF in construction recov- largely reflected a considerable decline in pub- ered slightly at the end of the year (Chart 6.10), lic investment, which fell by 29.2 per cent in vol- in line with the growth of cement sales to the ume, after having increased by 16.4 per cent in domestic market. In addition, the number and the previous year. Residential investment also amount of tenders for public works grew signifi- declined in 2016 (-1.9 per cent), with an impact cantly in the second half of 2016 (Chart 6.11). on activity in the construction sector (Section After major falls at the end of 2015 and in 5: ‘Supply’). The fall in residential investment in the first half of 2016, GFCF in machinery and Chart 6.7 • Share of private consumption in new loans granted by banks to consumption | New operations Chart 6.8 • Developments in GFCF in selected euro area countries | 2008Q1=100 in percentage 110 8 7 100 6 5 90 4 80 3 2 70 1 0 2009 2010 2011 2012 2013 2014 2015 2016 60 2008 2009 2010 2011 2012 2013 2014 2015 2016 Euro area Italy Sources: INE and Banco de Portugal calculations. Germany Spain Sources: INE and Banco de Portugal calculations. France Portugal The Portuguese economy in 2016 equipment recovered strongly in the third and from the value recorded in 2015. The posi- fourth quarters. This investment component tive performance of exports is structural, rely- grew by 2.2 per cent in the year as a whole ing on a corporate restructuring that started (6.0 per cent in 2015). In turn, GFCF in trans- before the international financial crisis, and is port equipment, although decelerating from not accompanied by systematic declines in unit 2015, continued to grow considerably in 2016 prices (Box 6.1: ‘Developments in unit values of (8.6 per cent). Portuguese exports of goods’). As regards ser- 67 vices, exports of services not related to tourism Deceleration in exports of goods and services reflecting a fall in energy exports and the lower growth of exports of services not related to tourism decelerated considerably. In turn, exports of In 2016 exports of goods and services grew by The lower growth of the international economy 4.4 per cent, compared to an increase of 6.1 per contributed to the less favourable performance cent in 2015. This reflects the behaviour of the of exports in 2016 as a whole compared to the goods component, which grew by 4.3 per cent, previous year. In this context, exports to some after 6.3 per cent in 2015, and of the services emerging economies, especially Angola, contin- component, which decelerated by 1.1 p.p. to ued to make a negative and significant contri- 4.6 per cent (Chart 6.12). In particular, energy bution to the change in total exports. However, exports fell, after growing considerably in 2015. exports of goods and services showed a clear Excluding this component, exports of goods upward intra-annual profile, widespread across grew by 5.1 per cent, i.e. accelerating by 1.8 p.p. all the main components. tourism services continued to be highly buoyant, in line with developments in their nominal revenue and the considerable growth of the number of overnight stays of non-residents in national hotels. In 2016 exports of tourism services grew by 9.7 per cent in volume (8.9 per cent in 2015). Chart 6.9 • Main limitation to investment | As percentage of total firms that reported limitations Chart 6.10 • Developments in GFCF in Portugal by type of investment | 2008Q1=100 to investment Other 120 Level of interest rates 100 Access to bank credit 80 Self-financing capacity 60 Investment profitability 40 Deterioration of sales prospects 0 10 20 30 40 50 60 70 20 2008 2009 2010 2011 2012 2013 2014 2015 2016 GFCF Total GFCF – Construction GFCF – Transport equipment GFCF – Machinery and equipment Share of firms (percentage) 2016 Source: INE. 2015 2014 2013 Source: INE. 68 BANCO DE PORTUGAL • Economic Bulletin • May 2017 The deflator of goods exports fell in 2016 (-3.4 per The growth of the volume of exports of goods cent) though less markedly when excluding the and services in 2016 was higher than that of fuels component (-2.4 per cent). Based on inter- the external demand indicator normally used national trade data, exports of goods grew by by Banco de Portugal, calculated on the basis 1.0 per cent in nominal terms in 2016 (3.7 per of the information used within the Eurosys- cent in 2015). This was largely due to a fall in car tem, leading to a further gain in the export exports, particularly to Germany, and a decline in exports of fuels and base metals. The reduction in car exports reflected the temporary closure of an important production unit in early 2016. By destination country, there was a sharp fall in exports of goods to non-EU markets, specifically Angola and China, and to the Netherlands. Angola’s contribution to the growth of nominal exports of goods in 2016 was -1.2 p.p., which corresponds to a drop of 28.4 per cent in share. However, this external demand indicator does not consider the relative importance of external trade with Angola. When taking into account the actual weight of Angola and the trend of its imports, external demand for Portuguese goods and services grew less in the most recent period, which implied a slightly higher gain in the share of exports than that obtained with the indicator normally used (Chart 6.14). the year as a whole. This adverse development Considering the main intra-EU trading part- faded away over the course of 2016 (Chart 6.13), ners, exports of Portuguese goods and servic- contributing to the intra-annual acceleration es to these destinations have been experiencing profile of exports. overall higher nominal growth than that of total 38 Chart 6.11 • Number and amount of tenders for public work | Year-on-year change in percentage; half-yearly values Chart 6.12 • Main contributions for the growth of goods and services exports in volume | Year-on-year change in percentage, and contributions in percentage points; quarterly values 60 10 40 8 6 20 4 0 2 -20 0 -40 -2 H1 H2 2014 H1 H2 2015 Call for tenders - number Source: AECOPS. H1 H2 2016 Call for tenders - value -4 2012 2013 2014 2015 2016 Other services Tourism Exports of energy goods Exports of goods (exc. energy) Exports of goods and services (year-on-year change, percentage) Sources: INE and Banco de Portugal calculations. 69 The Portuguese economy in 2016 imports of goods and services from those coun- cent, after 15.0 per cent growth in 2015) and a tries, understood as part of external demand for less buoyant imports of services, which grew Portuguese goods and services (Chart 6.15). For by 2.6 per cent in 2016, after a 6.7 per cent example, in 2016 nominal exports of goods and change in 2015 (Chart 6.16). Imports of goods services from Portugal to Spain grew by 5.8 per excluding energy grew by 6.5 per cent, decel- cent, while total Spanish imports grew by 1.6 per erating slightly from the value observed in 2015 cent. Although expressed in nominal terms, this (7.1 per cent). behaviour suggests that Portuguese exports performed well in the Spanish market. The deceleration in imports of goods in 2016 is consistent with the less buoyant behaviour of The deceleration in imports of goods and services continued, linked to a fall in exports of energy and services not related to tourism a number of demand components with a high import content, with a particular emphasis on a fall in exports of fuels and a significant deceleration in GFCF in machinery and equipment. Following the path observed since early 2013, the deflator of goods imports fell further in 2016 (-3.9 per cent). This reduction is less sig- Imports of goods and services grew by 4.4 per nificant when excluding fuel exports, standing cent in 2016 (8.2 per cent in 2015). This chief- at -2.4 per cent. Based on international trade ly reflected a fall in energy imports (-4.0 per data, imports of goods recorded a 1.3 per cent Chart 6.13 • Contributions from main trading markets for the nominal growth of goods exports excluding fuel | Year-on-year change in percentage, and contributions Chart 6.14 • Exports of goods and services and external demand | Year-on-year change in percentage in percentage points; quarterly values 12 15 9 10 5 6 0 3 -5 0 -10 -3 -15 -20 -6 2012 2013 2014 2015 2016 China Angola Extra-EU exc. Angola and China Intra-EU Nominal exports exc. energy (growth in percentage) Source: INE (International Trade) and Banco de Portugal calculations. 2008 2010 2012 2014 2016 External demand Portuguese exports External demand including Angola Sources: INE, IMF, ECB and Banco de Portugal calculations. Note: The indicator of external demand adjusted by the importance of foreign trade with Angola corresponds to the weighted average (weights based on exports) between the indicator of external demand computed by the ECB and Angola’s total volume of imports of goods and services. These calculations use the IMF’s projections (Word Economic Outlook) for the growth in volume of Angola’s imports in 2016. 70 BANCO DE PORTUGAL • Economic Bulletin • May 2017 nominal growth in 2016. By product, in addi- 3.9 per cent, with imports originating in Angola tion to a strong reduction in fuel purchases, dropping by 29.1 per cent. also imports of base metals dropped. Similar- In 2016 imports of goods and services grew ly to exports, the lower dynamics of imports more than the import-weighted overall demand, of goods was particularly evident in non-EU which led to an increase in import penetration, markets. In 2016 imports of goods from these as in the three previous years. countries recorded a nominal reduction of Chart 6.15 • Exports of goods and services for the main EU trading partners in 2016 and total imports from these countries | Year-on-year change Chart 6.16 • Main contributions for the growth of goods and services imports in volume | Year-on-year change in percentage, and contributions in percentage points; quarterly values in percentage; nominal values Growth of Portuguese nominal exports of goods and services by country of destination; percentage 16 45º 16 12 12 Italy UK 8 4 France 8 Spain 4 0 Germany 0 -4 Netherlands -4 -8 -12 -8 -8 -4 0 4 8 12 16 Growth of total nominal imports of goods and services; percentage Sources: Eurostat, INE and Banco de Portugal calculations. Note: The circles represent the locus of combinations between the growth of nominal imports of goods and services in each of the 6 countries considered (X-axis) and the growth of nominal exports of goods and services from Portugal to each of those countries (Y-axis). The size of the circles reflect the importance of each country in the Portuguese exporting structure. 2012 2013 2014 2015 2016 Other services Tourism Imports of energy goods Imports of goods (exc. energy) Imports of goods and services (year-on-year change, percentage) Sources: INE and Banco de Portugal calculations. The Portuguese economy in 2016 Box 6.1 | Developments in unit values of Portuguese goods exports This box complements the analysis of developments in Portuguese goods exports, focusing on changes in their unit values from a predominantly microeconomic viewpoint. The analysis presented here is based on unit values, which approximate the behaviour of exported goods prices. Nevertheless, prices of internationally traded goods depend on a complex set of factors, including the quality of products and the ability to exercise market power. All calculations are based on both Comext data and detailed reporting on international trade transactions conducted by firms, within the scope of Intrastat and Extrastat surveys. In addition, considering the volatility associated with energy prices, their unit values were excluded from the analysis based on corporate information. Chart 1 presents the annual rate of change in the unit values of exported goods for the 2003-15 period. This growth rate has always been positive, except for 2003, 2009 and 2013-14. The most exported goods over the whole period and, thus, contributing the most to the total rate of change – namely ‘vehicles’ and ‘machinery, equipment and electrical components’ –, account on average for 17 per cent of the total value of goods exports. The growth rates of these two categories of the Standard International Trade Classification (SITC) show that their unit values have increased after 2010, especially in the machinery and equipment component. Analysing the growth of exports’ unit values is more meaningful if compared to developments in a reference region. Hence, taking into account the existing institutional and economic environment, the natural benchmark is the average rate of change in the unit values of goods exported by the European Union (EU).39 The difference between the growth rates recorded in Portugal and in the EU was centred around zero for the 2003-15 period, which seems to exclude a scenario where Portuguese exporters sought to systematically compete on price (Chart 2). Chart 1 • Unit-values growth rates of Portuguese goods exports (including energy goods) | Percentage 8 4 0 -4 -8 -12 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Total (incl. Energy goods) Road vehicles (including air-cushion vehicles) Electrical machinery, apparatus and appliances, n.e.s., and electrical parts thereof Souce: Banco de Portugal calculations based on Comext dataset. 2013 2014 2015 71 72 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Among the products whose unit value growth rates surpassed those of the EU are, consistently, ‘articles of apparel and clothing accessories’, ‘crude animal and vegetable materials, n.e.s.’ and ‘organic chemicals’ (Chart 3). Finally, ‘footwear’ is the only category whose unit value growth rate was always positive in the period under analysis, and in most years higher than the EU average rate (Chart 4). Corporate information offers an analysis of the relationship between the differences of unit values and firm characteristics, namely in terms of sector of activity and age. The share of goods exporters by sector whose unit value growth rates are higher than the EU weighted average rate Chart 2 • Differences in unit-values growth rates (Portugal vs EU) | Percentage points 20 15 10 5 0 -5 -10 -15 -20 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: Banco de Portugal calculations based on Comext dataset. Note: The interval delimited by ‘+‛ signs represents the sample range after eliminating outliers – following Tukey (1977)'s method. The interval defined by the vertical line corresponds to the interquartile range and the circle to each distribution's median. Chart 3 • Unit-values growth rate of articles of apparel and clothing accessories | Percentage 8 6 4 2 0 -2 -4 -6 2002 2003 2004 2005 2006 2007 2008 Portugal Source: Banco de Portugal calculations based on Comext dataset. 2009 2010 2011 EU 2012 2013 2014 2015 The Portuguese economy in 2016 is shown in Chart 5. In the period under review, on average, this share stood at 50 per cent both for manufacturing and services firms. In 2011 and 2012 there was a decline in the proportion of goods exports by service firms whose unit value rate was higher than the EU’s. However, in the 2013-15 period as a whole, this percentage surpassed the previous, which may indicate an increase in these firms’ competitiveness. Service firms that export goods may be international trade wholesalers or may represent the commercial arm of corporate manufacturing groups. As regards exporters’ age, there is volatility regarding prevalent changes in unit values above the EU average. However, the share of exports associated with firms aged less than five years and Chart 4 • Footwear unit-values growth rate | Percentage 6 5 4 3 2 1 0 -1 2003 2004 2005 2006 2007 2008 2009 2010 Portugal 2011 2012 2013 2014 2015 EU Source: Banco de Portugal calculations based on Comext dataset. Chart 5 • Exports' share with positive difference in unit-values growth rate (Portugal vs EU) by sector of activity | Percentage 100 90 80 70 60 50 40 30 20 10 0 2003 2004 2005 2006 2007 Manufacture 2008 2009 2010 2011 2012 Services Sources: Banco de Portugal calculations based on Intrastat, Extrastat surveys and Comext dataset. 2013 2014 2015 73 74 BANCO DE PORTUGAL • Economic Bulletin • May 2017 between five and ten years whose growth rate in exported goods’ unit values was above that seen in the EU on average is considerable, often surpassing that of firms aged more than ten years, as was the case in the 2013-15 period (Chart 6). Chart 6 • Exports' share with positive difference in unit-values growth rate (Portugal vs EU) by age cohort | Percentage 100 90 80 70 60 50 40 30 20 10 0 2003 2004 2005 age<=5 2006 2007 2008 2009 5<age<=10 2010 2011 2012 2013 age>10 Sources: Banco de Portugal calculations based on Intrastat, Extrastat surveys and Comext dataset. 2014 2015 The Portuguese economy in 2016 7.Prices Slight increase in the inflation rate, based on services and unprocessed food prices although to a lesser extent, to the contribution In 2016 the inflation rate in Portugal as meas- by 0.3 p.p. to 1.6 per cent in 2016, while the rate ured by the rate of change in the Harmonised In- of change in processed food prices was only dex of Consumer Prices (HICP) was 0.6 per cent, 0.3 per cent, i.e. 0.9 p.p. below the 2015 value. compared to 0.5 per cent in 2015 (Table 7.1 and Chart 7.1). In the year as a whole, services and unprocessed food prices were more buoyant. By contrast, prices of industrial goods, in particular energy, recorded a negative change. An analysis of contributions from the different categories of goods to the inflation rate in 2016 (Chart 7.2) shows a positive contribution from the change in services prices (0.7 p.p.) and to a less- of telephone and telefax services, financial services, and also restaurants and cafés. The rate of change in unprocessed food prices declined Non-energy industrial goods maintained a negative rate of change (-0.3 per cent), although less sharp than in the previous year (-0.7 per cent). In turn, energy prices remained on the downward trend started in November 2013, declining by 1.8 per cent in 2016, after a 3.7 per cent fall in the previous year, in a context of recovery in oil prices throughout 2016. er extent in unprocessed food prices (0.2 p.p.). Energy prices contributed to the rise in inflation, in a context of recovery in oil prices and an increase in the tax on oil products In turn, energy and non-energy industrial goods continued to make a negative contribution in 2016 (-0.1 and -0.2 p.p. respectively), while processed food made a virtually nil contribution. In 2016 services prices grew by 1.5 per cent, accounting for a 0.1 p.p. rise from 2015. This The pattern of pre-tax fuel prices largely reflec- was essentially due to accommodation services, ted a decline in oil prices, and that of final prices which are part of the very dynamic Portuguese was mainly influenced by a rise in the tax on oil tourism sector. Reference should also be made, prices in February 2016 (Charts 7.3 and 7.4). Table 7.1 • HICP – Main components | As a percentage Weights Year-on-year rate of change Annual rate of change 2016 2014 2015 2016 16 Q1 16 Q2 16 Q3 16 Q4 100.0 91.9 -0.2 0.0 0.5 0.8 0.6 0.9 0.4 0.8 0.5 0.9 0.7 1.0 0.8 0.7 81.5 0.2 0.7 0.8 0.8 0.9 0.8 0.6 57.2 24.0 10.4 13.5 33.3 25.2 8.1 42.8 -1.1 -0.7 -2.1 0.4 -1.4 -1.4 -1.5 1.1 -0.1 1.5 1.9 1.2 -1.3 -0.7 -3.7 1.4 0.0 0.8 1.6 0.3 -0.7 -0.3 -1.8 1.5 -0.2 0.4 0.1 0.5 -0.7 0.2 -3.0 1.4 -0.3 0.6 1.4 0.1 -1.1 -0.2 -3.6 1.7 0.1 1.5 3.3 0.2 -0.8 -0.2 -2.6 1.5 0.3 0.8 1.5 0.2 -0.1 -0.8 2.0 1.6 – – 0.3 0.1 0.1 0.2 0.1 -0.1 0.1 0.2 0.1 0.3 0.2 -0.4 0.2 -0.4 Consumer Price Index (CPI) – -0.3 0.5 0.6 0.5 0.5 0.7 0.8 HICP – Euro area – 0.4 0.0 0.2 0.0 -0.1 0.3 0.7 Total Total excluding energy Total excluding unprocessed food and energy Goods Food Unprocessed food Processed food Industrial Non-energy Energy Services Memo items: Contribution of administered prices (in p.p.) Contribution of taxes (in p.p.) Sources: Eurostat and Statistics Portugal. 75 76 BANCO DE PORTUGAL • Economic Bulletin • May 2017 This contrasts with developments in 2015, when profit margins and refining costs made a positive contribution. Positive inflation differential between Portugal and the euro area continued In 2016 the inflation rate in Portugal was higher than the euro area’s, which stood at 0.2 per cent, maintaining the differential recorded in 2015 (Chart 7.5). Although the inflation rate in the euro area increased markedly in the second half of the year, only in November did it surpass inflation in Portugal. Energy industrial goods accounted for most of the inflation differential between Portugal and the euro area in 2016 (Chart 7.6). In this component, the differential moved from 3.1 p.p. in 2015 to 3.3 p.p. in 2016. The services differential rose to 0.4 p.p. in 2016, after 0.2 p.p. in 2015. Conversely, non-energy industrial goods prices grew at a slower pace than in the euro area, as seen since early 2012. Unit labour costs made the main contribution to the inflation rate Chart 7.7 breaks down the year-on-year inflation rate into the contribution of the main HICP determinants based on Banco de Portugal’s 5.0 Chart 7.1 • Harmonized index of consumer prices | In per cent 4.0 3.6 3.0 2.8 2.0 1.0 1.4 0.6 0.4 -0.2 0.0 0.5 -1.0 2010 2011 2012 2013 YoY rate of change 2014 2015 2016 Annual average rate of change Source: Statistics Portugal. 4.0 Chart 7.2 • Contributions to the annual growth rate of the HICP | In percentage points 3.0 2.0 1.0 0.0 -1.0 2010 2011 2012 Unprocessed food Energy Services Source: Statistics Portugal. 2013 2014 2015 Processed food Non-energy industrial goods Total 2016 The Portuguese economy in 2016 Chart 7.3 • Contributions to the year-on-year rate of change of petrol prices | In percentage points 77 Chart 7.4 • Contributions to the year-on-year rate of change of diesel prices | In percentage points 24.0 16.0 18.0 8.0 12.0 0.0 6.0 0.0 -8.0 -6.0 -16.0 -12.0 2011 2012 2013 2014 Oil price Taxes 2015 2016 2011 Profit margins and refining costs Final price 2012 2013 2014 Oil price Taxes Sources: ECB and Directorate General for Energy and Geology. 2015 2016 Profit margins and refining costs Final price Sources: ECB and Directorate General for Energy and Geology. 5.0 Chart 7.5 • HICP inflation in Portugal and in the euro area 4.0 3.0 2.0 |Year-on-year rate of change in percentage and percentage points 1.0 0.0 -1.0 -2.0 Jan. 10 Jul. 10 Jan. 11 Jul. 11 Jan. 12 Jul. 12 Jan. 13 Jul. 13 Differential Jan. 14 Jul. 14 Portugal Jan. 15 Jul. 15 Jan. 16 Jul. 16 Euro area Sources: Eurostat and Statistics Portugal. 1.5 Chart 7.6 • Inflation differential between Portugal and the euro area | Contributions, 1.0 0.5 0.0 -0.5 in percentage points -1.0 -1.5 -2.0 2008 2009 2010 2011 2012 Unprocessed food Energy Services Sources: Eurostat and Statistics Portugal. 2013 2014 2015 2016 2016 Q1 2016 Q2 2016 Q3 2016 Q4 Processed food Non-energy industrial goods Total 78 BANCO DE PORTUGAL • Economic Bulletin • May 2017 inflation analysis and forecasting model, the so- (1.4 per cent and 9.2 per cent in the fourth quar- called MIMO (Monthly Inflation Model). ter respectively). The analysis suggests that the rise in the infla- Inflation expectations released by Consensus tion rate over the course of 2016 essentially Economics remained stable at around 0.6 per reflected an increase in unit labour costs, whose cent in 2016, in line with actual average infla- contribution went up over the year, consistently tion (Chart 7.8). This contrasts with the succes- with an increase in compensations per employ- sive upward revisions of inflation expectations ee (Section 5: Supply). By contrast, non-ener- observed over the course of 2015. Inflation gy goods import prices contributed to mitigate expectations for 2017 also remained relative- the rise in the inflation rate throughout 2016. ly stable throughout 2016, at around 1.2 per Fuels, excluding taxation, also made a nega- cent, i.e. there is an implicit expectation of price tive but progressively lower contribution in the inflation between 2016 and 2017 (Chart 7.9). 40 recent period, mainly reflecting the time-path of oil prices. The Value Added Tax (VAT) applied to restaurants declined in the second half of 2016, generating a negative contribution from indirect taxation to inflation dynamics. In the third quarter of 2016, the item Others – which covers other unidentified factors contributing to infla- GDP deflator growth higher than HICP growth, in a context of gradual decline in terms of trade over the year tion – assumed a considerable role, which may Throughout 2016 the GDP deflator continued be associated with restaurants and cafés not to record a year-on-year rate of change above fully reflecting the decline in VAT on the pric- that of the HICP and of the domestic demand es charged. This is reinforced by the dynamics deflator. This pattern started in the second of the HICP and the Harmonised Index of Con- half of 2012. Since Portugal is a net impor- sumer Prices at Constant Tax Rates (HICP-CT) ter of energy goods, the recovery of oil prices in this item. While the former index recorded a throughout 2016 gradually reduced the positi- year-on-year rate of change of 1.1 per cent in ve effect of terms of trade, making it negative the third quarter, the latter grew by 8.8 per cent in the fourth quarter (Chart 7.10). 6.0 Chart 7.7 • Inflation rate disaggregation using MIMO | Contributions, in percentage points 5.0 4.0 3.0 2.0 1.0 0.0 -1.0 -2.0 -3.0 -4.0 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 2008 2009 2010 ULC Fuels and lubricants Inflation 2011 2012 PMX IT & ADM 2013 2014 2015 2016 Food commodities Others Sources: Eurostat and Banco de Portugal. Notes: ULC – Unit Labour Costs; PMX – Import prices excluding energy goods; IT – Indirect taxes; ADM – Administered prices. Obs: The effect of taxes in fuels and lubricants is included in IT & ADM. The Portuguese economy in 2016 79 5.0 Chart 7.8 • Inflation expectations and observed inflation 4.0 3.0 2.0 | In per cent 1.0 0.0 -1.0 -2.0 2008 2009 2010 2011 2012 Current year forecast 2013 2014 2015 2016 Observed inflation (12-month average rate of change) Sources: Consensus Economics and Statistics Portugal. 5.0 Chart 7.9 • Inflation expectations for current and next year 4.0 3.0 2.0 | In per cent 1.0 0.0 -1.0 -2.0 2008 2009 2010 2011 2012 2013 Current year forecast 2014 2015 2016 Next year forecast Source: Consensus Economics. 5,0 4,0 Chart 7.10 • GDP deflator, domestic demand deflator and terms of trade | In per cent 3,0 2,0 1,0 0,0 -1,0 -2,0 -3,0 -4,0 2010 Q1 2011 Q1 2012 Q1 Change in terms of trade Source: Statistics Portugal. 2013 Q1 2014 Q1 GDP deflator 2015 Q1 2016 Q1 Domestic demand deflator 80 BANCO DE PORTUGAL • Economic Bulletin • May 2017 8.Balance of payments In 2016 the net lending of the economy increased to 1.5 per cent of GDP, up from 0.3 per cent in 2015. This improvement reflected a sharper reduction in investment than in savings, in a context of relative stabilisation of net capital transfers. In sectoral terms (Chart 8.1), households had a slight decline in net lending, due to the reduction in capital transfers received. Current savings in this sector, as a percentage of GDP, remained at a lower level than investment, and both aggregates declined slightly in 2016. Non-financial corporations, in turn, moved from net borrowing in 2015 to net lending in 2016, as a result of the increase in current savings, which more than offset the slight increase in investment. Increase in the current and capital account surplus, based on net exports of goods and services In 2016 the current and capital account surplus increased to 1.7 per cent of GDP, up from 1.2 per cent in 2015 (Table 8.1). This increase in the balance was the result of a reduction in the deficit of the goods and primary income accounts (both by 0.3 p.p. of GDP) and an increase in the surplus of the services account (by 0.2 p.p. of GDP), associated with the performance of the tourism sector (Chart 8.2). In addition, the capital account surplus declined to 0.9 p.p. of GDP, down from 1.2 p.p. in 2015, strengthening the trend observed in recent years, which are related to the decline in European fund transfers for investment purposes. The buoyancy of the tourism sector continued to contribute decisively to the behaviour of the goods and services account The goods account deficit declined by 0.3 p.p. for the second year in a row. This decline was a result of the more pronounced growth of goods’ exports (0.8 per cent) than imports (0.3 per cent). The services account surplus increased again, to stand at 7.1 per cent of GDP in 2016. Exports of services grew by 4.4 per cent, decelerating from the previous year, when the rate of change had stood at 7.5 per cent. Growth in 2016 was chiefly due to the contribution of travel and tourism. This component continued to post very significant buoyancy, reflecting a 10.7 per cent change in 2016, up from 10.2 per cent in the previous year. The main contribution to the deceleration in exports of other services was made by the behaviour of services provided by corporations, which had a rate of change of -3.3 per cent, compared with 15.4 per cent growth in 2015. Table 8.1 • Current and capital accounts | As a percentage of GDP Current and capital accounts Current account Goods and services account Goods Services 2013 2014 2015 2016 3.2 1.5 1.2 1.7 1.6 0.1 0.1 0.8 1.9 1.1 1.8 2.2 -4.7 -5.5 -5.2 -4.9 6.6 6.6 6.9 7.1 of which: Travel and tourism 3.6 4.1 4.4 4.8 Primary income account -1.2 -2.0 -2.5 -2.2 0.9 0.9 0.8 0.8 Secondary income account of which: 1.4 1.5 1.6 1.5 Capital account Emigrants / immigrants remittances 1.6 1.4 1.2 0.9 Financial Account 3.3 1.6 1.2 1.7 Errors and omissions 0.1 0.1 0.0 0.0 Sources: INE and Banco de Portugal. The Portuguese economy in 2016 Imports of services had similar developments, with a change of 3.1 per cent, down from 5.8 per cent in 2015. The main contributions to growth in 2016 were travel and tourism, intellectual property rights not included elsewhere and other services. 81 Income paid abroad declined and the subsidies received from the European Union increased The primary income account deficit decreased by 0.3 p.p. of GDP as a result of the 0.4 p.p. decline The trade account balance in 2016 benefitted in income paid abroad and the maintenance of from a positive volume effect, contrary to devel- income received from non-resident entities. The opments in the two previous years (Chart 8.3). decline in income paid abroad was mainly asso- The largest contribution, however, was made by ciated with the decrease in the stock of exter- the terms of trade effect, albeit less expressive nal liabilities by resident sectors, in particular the than in the previous year. This was associated general government and other monetary finan- with the development of energy prices (Section 2: cial institutions (other MFI). The stabilisation of ‘International environment’ and Section 7: ‘Prices’). primary income received from abroad was the 10 Chart 8.1 • Net lending/net borrowing of the economy 5 0 -5 | As a percentage of GDP -10 -15 -20 2009 2010 2011 2012 2013 2014 Non-financial corporations Financial corporations General Government Households 2015 2016 Net lending/net borrowing Source: INE. 10 8 Chart 8.2 • Decomposition of current and capital account balance 6 4 2 0 -2 | As a percentage of GDP -4 -6 -8 -10 2013 2014 Capital account Services Goods Sources: INE and Banco de Portugal. 2015 Secondary income account Primary income account Current and capital accounts 2016 82 BANCO DE PORTUGAL • Economic Bulletin • May 2017 result of mixed developments. On the one hand, i.e., the purchase of external financial assets there was a decline in interest received associat- by residents was higher than the purchase of ed with long-term debt securities held in portfo- national assets by non-residents (Chart 8.4). lio by Banco de Portugal, insurance corporations and pension funds and, to a lesser extent, households. The reduction by insurance corporations and pension funds and households was associated with a decline in the holdings of long-term debt securities issued by non-residents, whereas in the case of Banco de Portugal it was due to a decline in yields on securities. On the other hand, there was an increase in the subsidies from the European Union. The secondary income account components stabilised as a percentage of GDP. Emigrants’ remittances decelerated, which was reflected in a 0.8 per cent increase, down from 8.3 per cent in 2015. Behaviour differed across home countries. Remittances from Switzerland, Angola and Germany declined, but there was an increase in transfers from France, the United States, the United Kingdom and, to a lesser extent, Spain and Luxembourg. Net lending of the Portuguese economy was reflected in a fall in the negative balance of the international investment position and net external debt These net outflows corresponded to 1.7 per cent of GDP and contributed to the improvement in the International Investment Position (IIP) of the Portuguese economy. At the end of 2016, the IIP stood at 105.1 per cent of GDP, which corresponded to an improvement of 6.9 p.p. of GDP (Table 8.3). In addition to the effect of financial account transactions, the total change in the IIP, as a percentage of GDP, may be broken down into changes in prices of the financial instruments, exchange rate changes in instruments denominated in foreign currency, other changes, such as reclassifications, and also the denominator effect, i.e. the GDP nominal change. Therefore, in addition to the financial account effect, the balance of the IIP as a percentage of GDP was positively influenced by the effect of the nominal GDP change, estimated at 3.3 p.p., and price changes, which correspond to 2.7 p.p. of GDP (Chart 8.5). In this latter item, both the gains in financial assets held by residents and the losses in national liabilities to non-residents contributed to a positive effect on the IIP balance. In particular, rises in sovereign debt yields are reflected in a decline in security prices with a positive impact on the IIP. Conversely, exchange-rate changes and other chang- The current and capital account surplus recor- es made a negative contribution, equivalent ded in 2016 implied a financial account surplus, to 0.7 p.p. of GDP. Table 8.2 • Financial Account | As a percentage of GDP 2013 Financial Account Direct investment 2014 2015 2016 3,3 1,6 1,2 1,7 -1,7 -1,5 -0,6 -2,2 Net acquisition of assets 3,0 4,2 0,5 1,9 Net incurrence in liabilities 4,7 5,7 1,1 4,1 8,0 Portfolio investment 2,6 -0,7 0,2 Net acquisition of assets -0,1 4,2 0,6 0,9 Net incurrence in liabilities -2,7 4,9 0,3 -7,1 0,6 1,1 0,2 0,3 Financial derivatives Other investment 1,6 1,7 0,6 -6,9 Net acquisition of assets -6,2 2,3 -5,4 -2,0 Net incurrence in liabilities -7,8 0,5 -5,9 4,9 0,3 1,0 0,8 2,5 Reserve assets Sources: INE and Banco de Portugal. 83 The Portuguese economy in 2016 Banco de Portugal and nonfinancial corporations purchased foreign assets, whereas nonresidents disinvested from public debt and other-MFI’s securities. In 2016 Banco de Portugal was the main agent purchase of debt securities, either as reserve assets, or, in particular, as portfolio investment, and deposits with non-resident entities. This initiative by Banco de Portugal corresponds to the participation in the implementation of Eurosystem’s monetary policy decisions. Notwithstanding these transactions, the IIP of Banco de Portugal remained virtually unchanged, also due to a investing in external assets, especially via the significant increase in liabilities to non-residents, 4,000 Chart 8.3 • Decomposition of goods and services account 3,000 2,000 1,000 | Million of euros 0 -1,000 -2,000 -3,000 2013 Volume effect 2014 Price effect 2015 Terms of trade effect 2016 Crossed effect Change in balance Sources: INE and Banco de Portugal. Note: A positive change (negative) implies an increase (decrease) in the overall balance of the goods and services account. The change in the overall balance of goods and services account can be decomposed in four effects: – volume effect – effect of the change in quantities imported and exported; [Xt-1.vxt]-[Mt-1.vmt] – price effect – effect of the average price growth of external trade; [Xt-1. pt]-[Mt-1.pt] – terms of trade effect – effect of the relative change in exports and imports prices; [Xt-1.(pxt--pt)]-[Mt-1.( pmt-pt)] – crossed effect – effect of the interaction between the change in quantities and in prices of exports and imports; [Xt-1.vxt.pxt]-[Mt-1. vmt.pmt] The following notation applies: Xt-1 and Mt-1 are the exports and imports in year t-1 at current prices; vxt and vmt are the change rates in volume of exports and imports in t; pxt and pmt are the change rates of exports and imports prices in t; pt is the average change rate of the prices of external trade in year t ((pxt+pmt)/2). 15.0 Chart 8.4 • Net change in financial assets and liabilities and overall financial balance 10.0 5.0 | As a percentage 0.0 of GDP -5.0 -10.0 2013 2014 Overall balance Sources: INE and Banco de Portugal. 2015 Net acquisition of assets 2016 Net incurrence of liabilities 84 BANCO DE PORTUGAL • Economic Bulletin • May 2017 particularly in the form of liabilities to the Euro- increase in liabilities than in assets and in the de- system. Monetary gold valuation also contribut- terioration of the IIP balance of this sector. ed to the increase in financial assets by 0.9 p.p. Contrary to the increase in external assets by of GDP (Charts 8.6 and 8.7). Banco de Portugal and non-financial corpora- Non-financial corporations were also engaged tions, non-residents disinvested significantly from in investments abroad, albeit to a lesser extent. national public debt securities (Section 4: ‘Fiscal These investments occurred mainly with corpo- policy and situation’) and, to a lesser extent, from rations with which they had established direct securities issued by other MFI. investment links and from which they had ob- In the case of the general government, and in tained funds (Box 8.1 ‘Direct investment flows addition to transactions in debt securities issued, into the Portuguese economy’). The balance of reference should be made to the repayment of these cross transactions was reflected in a higher loans obtained under the European and Financial 112 Chart 8.5 • Change in International investment position | As a percentage of GDP 1.7 110 2.7 108 106 3.3 112.0 0.7 104 105.1 102 100 IIP 2015 (symmetrical) Transactions Changes in prices Change in GDP Exchange rate and IIP 2016 (symmetrical) other adjustments Sources: INE and Banco de Portugal. Chart 8.6 • Net acquisiton of assets (1) | As a percentage of GDP Chart 8.7 • Net incurrence of liabilities(2) | As a percentage of GDP 8 8 6 6 4 4 2 2 0 0 -2 -2 -4 -4 -6 -6 -8 Total CB OMFI 2015 GG NMFI NFPS 2016 -8 Total CB OMFI 2015 GG NMFI NFPS 2016 Sources: INE and Banco de Portugal. Sources: INE and Banco de Portugal. Note: (1) The net acquisiton of assets corresponds to buys less sells of foreign assets by residents. A plus sign represents a net outflow of funds from the portuguese economy. Note: (2) The net incurrance of liabilities corresponds to the increase less redemptions of national liabilities with non-resident entities. A plus sign corresponds to a net inflow of funds in the portuguese economy. The Portuguese economy in 2016 Assistance Programme. These transactions were an also declined. The volume of these transactions important contribution to the decline of 12.5 p.p. was more significant than the decline in liabili- of GDP in the stock of external liabilities in this ties, leading to a deterioration of the negative sector. This result also had the important contri- balance of the IIP in this sector by 1 p.p. of GDP. bution of both GDP and the devaluation of Treas- Taking into account the significant transactions ury bonds, in a context of an increase in public debt yields. 85 in debt securities issued by residents, net external debt, which is obtained by excluding capital As regards the liabilities of other MFI, in addition instruments and financial derivatives from the to the decline in liabilities in the form of debt IIP, declined by 7 p.p. of GDP, down from 101.5 securities, there was a decrease in deposits by per cent of GDP in 2015 to 94.4 per cent of GDP non-residents. In the same vein, investments in 2016 (Chart 8.8). abroad in these type of securities by other-MFI Table 8.3 • International Investment Position | As a percentage of GDP International Investment Position Direct investment Direct investment - Assets Direct investment - Liabilities 2013 2014 2015 2016 -116.3 -117.7 -112.0 -105.1 -27.3 -31.1 -30.8 -32.0 42.1 42.7 42.6 43.5 69.4 73.9 73.4 75.5 -10.2 -10.6 -9.2 0.8 Portfolio investment - Assets 64.8 68.8 66.7 65.9 Portfolio investment - Liabilities 75.0 79.3 76.0 65.1 -84.5 -84.3 -82.0 -86.6 68.2 69.1 61.9 58.1 Other investment - Liabilities 152.7 153.4 143.9 144.7 Reserve assets and derivatives 5.7 8.3 10.0 12.8 Portfolio investment Other investment Other investment - Assets Sources: INE and Banco de Portugal. 0 Chart 8.8 • International investment position, net external debt and position in capital instruments | As a percentage -20 -40 -60 -80 of GDP -100 -120 -140 2005 2006 2007 2008 Net External Debt (symmetrical) Sources: INE and Banco de Portugal. 2009 2010 2011 Net Capital Instruments 2012 2013 2014 2015 2016 International Investment Position 86 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Box 8.1 | Direct investment flows into the Portuguese economy41 The current situation of the Portuguese economy is characterised by moderate economic growth, relatively weak investment rate and unfavourable developments in labour productivity, against a background of high indebtedness of the different institutional sectors. In this context, the acceleration of economic activity and the increase in productivity require an investment effort that will be easier to implement if there is the possibility to attract external capital. Foreign direct investment in Portugal reflects the Portuguese economy capacity to attract capital, also resulting in positive externalities, especially in terms of sharing technology and business knowledge. Therefore, it is important to analyse in more detail the nature and composition of these flows. Balance of payments statistics on foreign direct investment in Portugal include financial transactions on assets and liabilities of resident corporations with their non-resident direct investors, as well as all financial transactions between resident and non-resident corporations, held (directly or indirectly) by the same non-resident entity.42 Diversified financial flows are thus recorded, comprising reinvested earnings, investment in real estate, debt instruments and other capital instruments. Reinvested earnings correspond to corporations’ results that are retained, and reflect their allocation to foreign holders of the respective capital. Their registration in the balance of payments is therefore made under investment income and in the financial account, corresponding to an increase in investment. Reinvested earnings are recorded in the period in which they are earned. Investment in real estate corresponds to investment by non-residents in property and residences in the national territory, either for personal use or for rental. Debt instruments cover funds lent by non-resident entities – in the form of loans or debt securities – whereas capital instruments represent the property rights on the resident entity. In an open economy as the Portuguese, foreign direct investment is an important source of financial inflows. Compared with other European countries, Portugal is among the main recipients of direct investment as a percentage of GDP (Charts 1 and 2). For many years, inflows into the Portuguese economy occurred mainly in the form of portfolio investment and other investment. In the 1996-2016 period, foreign direct investment, on average, accounted for annual net inflows into the economy of 2.4 per cent of GDP. In most recent years, however, and in a context of declining external indebtedness, direct investment flows have been gaining relative importance in the Portuguese economy. In 2013 and 2015 they were in fact the main source of external financing (Chart 3). Since 2008, this investment represented, on average, 2.1 per cent of GDP. It was higher in 2011 and 2012, when it reached 4.1 per cent of GDP, and in the last two years, when it stood, on average, at 3.2 per cent of GDP. Financing requirements of the Portuguese economy were partially met with recourse to privatisations, which translated into the purchase of assets by non-residents. Therefore, in 2011 and 2012, the composition of foreign direct investment in Portugal primarily took the form of capital instruments, and was associated with the non-financial corporations sector (Charts 4 and 5). After a decline in inflows in 2013 and 2014, financial transactions intensified again in the last two years. In terms of instrument, however, there was a decline in the amounts of investment in corporations’ capital, against an increase in transactions in debt instruments. I.e., resident corporations that received foreign direct investment used that channel to obtain financing from the firms in the The Portuguese economy in 2016 Chart 1 • Foreign Direct Investment – transactions – 2015 | As a percentage of GDP 5 4 3 2 1 0 -1 SK CR EE EL FI CZ DE DK ES IT UK SE AT FR LT RO LV PO PT IS BG SL BE Source: Eurostat. Chart 2 • Direct investment – stocks – 2015 | As a percentage of GDP 120 BE Inward flows, as a percentage of GDP 100 EE BG 80 60 SK 40 RO LV IS PT CZ CR UK ES PO LT FI SL 20 DE IT AT SE DK NO FR EL 0 0 20 40 60 Outward flows, as a percentage of GDP 80 100 120 Source: Eurostat. Chart 3 • Foreign direct investment and other inward funds in portuguese economy – transactions | As a percentage of GDP 30 25 20 15 10 5 0 -5 -10 -15 -20 -25 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 FDI Sources: INE and Banco de Portugal. Portfolio investment Other investment 87 88 BANCO DE PORTUGAL • Economic Bulletin • May 2017 group, thus circumventing more restrictive financing conditions in Portugal. This may be related either to loans or to issues of debt securities taken by firms in the group. In the last three years, investment in real estate also reinforced its share as component of direct investment.43 In 2016, flows associated with these transactions exceeded flows of investment in other capital instruments (0.7 per cent versus 0.4 per cent of GDP). These developments reflected demand for land and property by non-residents, in line with the profile of residence permits for investment purposes. Chart 4 • Foreign direct investment in Portugal, by sector – transactions | As a percentage of GDP 6 4 2 0 -2 2008 2009 2010 2011 2012 Non-financial corporations Non-monetary financial instituitions except ICPF General Government and Households 2013 2014 2015 2016 Other Monetary Financial Institutions Insurance Corporations and Pension Funds (ICPF) Sources: INE and Banco de Portugal. Chart 5 • Foreign direct investment in Portugal, by type of transaction – transactions | As a percentage of GDP 5 4 3 2 1 0 -1 -2 -3 2008 2009 2010 Capital Sources: INE and Banco de Portugal. 2011 Reinvested earnings 2012 2013 Real estate 2014 2015 Debt instruments 2016 The Portuguese economy in 2016 Notes 1. Available at https://www.bportugal.pt/sites/default/files/anexos/pdf-boletim/projecoes_mar_pt_0.pdf 2. For further details, see Félix et al. (2007) ‘MIMO – A Monthly Inflation Model’, winter 2007 issue of the Economic Bulletin of Banco de Portugal. 3.Angola is not included in the set of countries used under the Eurosystem’s procedure to calculate external demand for the goods and services of each individual country. 4.The APP is composed of four programmes: the Covered Bond Purchase Programme (CBPP), which is currently in its third programme; the Asset-backed Securities Purchase Programme (ABSPP), the Public Sector Purchase Programme (PSPP), and the CSPP. The first two programmes started in 2014 and the third programme in 2015. 5.In its December 2016 meeting, the Governing Council of the ECB announced its decision to extend the APP until the end of 2017. However, purchases were reduced from €80 billion to €60 billion from March onwards. 6.The Eurosystem’s architecture only allows monetary financial institutions to have access to the ECB’s deposit facility rate. 7.In this respect, see the speech by Yves Merch, ‘Ructions in the repo market – monetary easing or regulatory squeezing?‛, 26 January 2017. 8.This fact tends to be visible in the repo market, particularly on dates close to accounting reporting dates and for rates associated with jurisdictions perceived as safe. On the last business day of December 2016, a number of repo rates reached historical lows. 9.See Antunes, Homero and Prego (2016) ‘Firm default probabilities revisited‛, Banco de Portugal Economic Studies, Vol. 2, No. 2, Banco de Portugal. 10. Total credit to non-financial corporations includes loans by resident banks, loans by non-residents, debt securities issuance (held by residents and non-residents), trade credits (granted by residents and non-residents), household loans and Treasury loans (in the case of loans to public undertakings). 11. This definition of indebtedness includes the sum of bank loans, debt securities and trade credits. 12. See Box “Developments in corporate indebtedness in Portugal and the euro area”, Economic Bulletin, May 2016, Banco de Portugal. The analysis considers data from the national financial accounts and the measure of indebtedness considered is the ratio of debt (loans and debt securities) to total assets. 13. Félix, S. (2017) “Firm creation and survival in Portugal”, Banco de Portugal Economic Studies, Banco de Portugal. 14. The survey IES includes economic, financial, and accounting information of virtually all the Portuguese non-financial corporations and is available for the period between 2005 and 2015. Nevertheless, changes in the accounting standards system in 2010 have led to a loss of data comparability. 15. A reporting gap is observed when the firm reports to the IES in year t-1 and year t+1, but is absent in year t. 16. Some of the absences that are considered exits from the market in 2015 may constitute reporting gaps. This can only be verified when the IES for 2016 becomes available. However, only about 0.2 per cent of the observations are reporting gaps in the period under analysis. 17. EBIT stands for Earnings Before Interest and Taxes. 18. These factors meet the criteria for classification as temporary measures, as defined in the Eurosystem context, which does not necessarily coincide with the classification adopted by other institutions. 19. In August 2016, the EU Council recommended a stabilisation of the structural balance. Compliance with this recommendation shall be assessed on the basis of the cyclical adjustment methodology and the definition of temporary measures adopted by the European Commission. 20. The impact of the resolution of BANIF in 2015 is treated as a temporary measure, according to the definition adopted by the Eurosystem. This operation had a significant impact on capital transfers, chiefly as a result of the capital increase by the State. To a lesser extent, it also affected the public investment aggregate via the recording of real estate acquisitions by Oitante, a special-purpose entity created in the context of the resolution process. 21. Military material delivered in 2016 relates to the sale of F-16 aircrafts to Romania, which negatively affects the general government investment aggregate. In spite of their one-off impact on the balance, these operations are not treated as temporary measures in the context of the definition adopted by the Eurosystem. 22. In 2015 a number of operations contributed to an increase in general government capital expenditure: the conversion into an equity increase of loans granted by Wolfpart to its holding Caixa Imobiliário; the recording of requests for budget increases by Instituto de Financiamento da Agricultura e Pescas (Agricultural and Fisheries Financing Institute); capital increases in public corporations that are not included in the general government. 23. In 2015 the general government wage cuts in force since 2011 started to be reversed, with 20 per cent of the cuts being reinstated. In 2016 the remainder was gradually reinstated in the course of the year, by 25 per cent each quarter. 24. In early 2015, the European Central Bank announced the widening of the asset purchase programmes then in force, to also include the purchase of bonds issued by euro area central governments, as well as by European organisations and institutions. 25. The implicit interest rate is computed as the ratio of interest expenditure to the simple average of the debt stock at the end of the current and the previous year. 26. The analysis in this box is based on the methodology developed in the context of the Eurosystem which allows the breakdown of structural revenue from taxes and social contributions into the following components: i) impact of legislative changes; ii) contribution of the discrepancy between the trend nominal change in the relevant macroeconomic base and GDP; iii) effect of the structural fiscal elasticity in structural terms and iv) residual component. The residual component corresponds to the part of the structural change that is not explained by the previous effects. For further details, see Kremer et al. (2006), A disaggregated framework for the analysis of structural developments in public finances, Working paper ECB 579, and Braz, C. (2006), ‘The calculation of cyclically adjusted balances at Banco de Portugal: an update’, winter 2006 issue of the Economic Bulletin of Banco de Portugal. 89 90 BANCO DE PORTUGAL • Economic Bulletin • May 2017 27. In the case of the tax on oil products, the estimate for the impact of an increase in taxation reflects an ex post reassessment based on the final outturn for the year. 28. Total revenue from social contributions includes actual and imputed social contributions referring to civil servants, which are also recorded on the expenditure side, under compensation of employees. 29. See Box 5.2 of the October 2016 issue of the Economic Bulletin for a characterisation of very long-term unemployment in Portugal. 30. According to Eurostat's definition, tradable services include accommodation and food services, trade and repair, transportation and storage, and information and communication activities. 31. Apparent labour productivity was calculated dividing VAT (volume chain-linked series, € millions) by total employment (full-time equivalent), except in the 1992 recession, when it was calculated on the basis of GDP. 32. For a discussion, see for example F. Caselli (2005) ‘Accounting for Cross-Country Income Differences‛, Handbook of Economic Growth, volume 1, Chapter 9. 33. Information available at https://fred.stlouisfed.org/release?rid=285. 34. Taking into account the weight of trade with Angola in Portuguese nominal exports, it is possible to estimate the mechanical impact of the fall in exports to Angola on nominal GDP. Hence, in 2016 total exports to Angola (goods and services), excluding an estimate for import contents, appear to have made a contribution of around -0.6 p.p. to 3.0 per cent growth of nominal GDP (in 2015 the contribution was around -0.7 per cent for 3.7 per cent growth of nominal GDP). 35. In 2016, in particular, remuneration received by households grew by 3.6 per cent, after 2.4 per cent in 2015. 36. In 2016 the decline in interest expenses of households contributed 0.4 p.p. to the 3.2 per cent growth of nominal disposable income. 37. See the special issue ‘Recent developments in consumer lending: A macroprudential approach‛, Financial Stability Report, Banco de Portugal, November 2016. 38. Following 1.5 per cent year-on-year falls in the first two quarters of 2016, nominal exports of goods grew by 1.9 and 5.1 per cent respectively in the third and fourth quarters of 2016. This was also due to the stabilisation of exports of fuels in the second half of the year, after the break observed in the first half. 39. Information on the aggregation and calculation of rates of change in European export prices in Banco de Portugal, 2016, ‘Portuguese international traders: some facts about age, prices and markets‛, Economic Bulletin, October 2016, special issue. 40. For further details, see Félix et al. (2007) ‘MIMO – A Monthly Inflation Model ‛, Winter 2007 issue of the Economic Bulletin of Banco de Portugal. 41. This Box was based on information on Foreign Direct Investment in Portugal, according to the directional principle. I.e., all transactions with non-residents regarding resident direct investment firms are considered, irrespective of whether they are an asset or a liability of such corporations. 42. For further details on the statistical treatment of information on direct investment, see Banco de Portugal (2015), ‘Statistics of the balance of payments and the international investment position – Methodological notes’, Supplement to the Statistical Bulletin No. 2/2015. 43. For a more detailed analysis of real-estate investment, see Box 7.1 of the May 2016 issue of the Economic Bulletin. II Special issue Distribution mechanisms of monetary policy in the Portuguese economy Special issue Distribution mechanisms of monetary policy in the Portuguese economy 1. Introduction models with heterogeneous agents, together In the wake of the global economic and financial with the dissemination of microeconomic data crisis, the link between the implementation of monetary policy and developments in inequality has re-entered the public debate. This is set against a background where inequality has once again been considered inseparable from economic policy. The perception that the reduction in inequality seen throughout most of the last century had been reversed, together with the fact that advanced economies have failed to return to the activity levels that would have been reached if the trend before the crisis had continued, contributed to the greater importance of issues related to the distribution of economic resources. It is well known that inequality is not, and should not be, an objective pursued by central banks. In the case of the euro area, the primary objective of the European Central Bank (ECB) is price stability and, secondarily, support to the general economic policies in the European Union. The maintenance of price stability helps prevent the arbitrary redistribution of income. In any case, it is potentially important to understand the link between the implementation of monetary policy and inequality for two sets of reasons. On the one hand, inequality levels prevailing in an economy may affect monetary policy transmission to economic agents, thereby hampering its expected results. On the other hand, monetary on an individual basis, has helped focus academic discussion on distribution issues. There is a new consensus in economic literature that monetary authorities must be aware of the redistribution effects of monetary policy, as well as the unobserved heterogeneity in the economy. This conclusion stands even when the redistribution impact of monetary policy is relatively small over long periods, which are the most suitable when analysing structural developments in inequality: over such time horizons, structural factors become more important, including inter alia the openness of economies, technological progress, institutional changes in the output and labour markets, the tax regime and the social protection network. The purpose of this article is to look into some of the main monetary policy transmission channels impacting on the distribution of monetary resources among households. Such channels are illustrated on the basis of microeconomic data on the Portuguese economy. However, the analysis will necessarily compare modestly with the title of this article. In fact, the impact of monetary policy on the Portuguese economy goes well beyond the channels analysed in this article. For instance, the reduction in the ECB’s intervention rates following the international financial crisis made a key contribution to relieving the debt service of Portuguese households policy may affect inequality levels prevailing in and to reduce liquidity constraints. This channel the euro area and individual economies and, had a major impact on the Portuguese econo- as such, influence a considerable part of public my as a whole, but it will not be reviewed here. debate and the Eurosystem’s accountability. Studying the distribution effects of monetary The redistribution impact of monetary policy policy would involve describing a counterfac- has also been increasingly reviewed in econom- tual scenario that can only be created using a ic literature (see e.g. Auclert, 2016, or Coibion et proper model. al., 2016, and respective references). The use of The Special Issue is structured as follows. Section increased computing power to solve economic 2 looks into the methodology and the microdata 93 94 BANCO DE PORTUGAL • Economic Bulletin • May 2017 bases used in the analysis. Section 3 describes between its assets and its liabilities. This net some of the main monetary policy redistribution wealth, by construction, is predetermined and channels, illustrating their potential impact based generates some return, including capital gains, on microeconomic data on the Portuguese econ- over time. On the other hand, there is the so- omy. Section 4 includes an exercise that analyses called human wealth, i.e. the sum of current the quantitative impact on the Portuguese econ- and future income resulting from household omy of the standard and non-standard mone- decisions as to how to spend time in the labour tary policy measures implemented over the past market. few years. Section 5 presents and discusses the Over the lifetime of each household, human main conclusions of this exercise. wealth will depend on exogenous conditions – such as the labour market environment and 2. Methodology and databases structural changes – and idiosyncratic features In this article, the household is the unit of analy- of the household under review. Among house- sis. By definition, this is the unit where inequal- hold-specific features are the (accumulated or ity resulting from its links with the market is of inherent) efficiency with which household mem- greater interest. Although in Portugal it is still bers that participate in the labour market are not possible to track wealth trends for a repre- able to turn their time into income, and the start- sentative sample of the population over time, ing position in the labour market, which, togeth- measures implemented over the past few years. Section 5 presents and discusses main conclusions er withthe age, partially determines future returns results are now available from two rounds of the exercise. on that market. The transformation of initial er the past few years. Section 5 presentsFinance and discusses the main conclusions Household and Consumption Survey, a Eurosystem initiative that Banco de Portugal conditions into returns in the labour market is and Statistics Portugal (Portuguese acronym: anything but deterministic. As is well known, luck INE) have implemented domestically. The inforethodology and databases plays its part, which means that idiosyncratic mation gathered from these surveys will make shocks help determine, inter alia, which house- ses members will resulting be employed, the quality of it possible to start tracking the impact article, the household is the unit of analysis. By definition, this isunder the unithold where inequality thepossible match between employees and workers BancoAlthough de Portugal and Statistics ts links with the market is ofreview greater(see interest. in Portugal it is still not to track wealth the unit of analysis. By definition, this is the unit where inequality resulting (according to two theirrounds skills) and 2016, Costa andtime, Farinha, 2012, s for a representative samplePortugal, of the population over results are and now available from of developments in of greater interest. Although in Portugal it is still not possible to track wealth thede sector of activity where a household memCosta, 2016). ousehold Finance and Consumption Survey, a Eurosystem initiative that Banco Portugal and Statistics ple of the population over time, results are now available from two rounds of ber will work. Labour gal (Portuguese acronym: INE) have implemented domestically. The information gathered from income these results from the In this article, household inequality assessed sumption Survey, a Eurosystem initiative that Banco de Portugalis and Statistics combination of these opportunities. This clearys will make it possible to start the impact underofreview (see Banco de Portugal and Statistics on tracking the basis of information two concepts wealth: wealth NE) have implemented domestically. The gathered from these ly shows that, in contrast to net wealth, human gal, 2016, Costa and Farinha,in 2012, and Costa, 2016). sense net wealth. tart tracking the impact underbroad review (seeand Banco de Portugal and Statistics wealth depends on a great deal of unavailable sa, article, household inequality is assessed on the basis of concepts of wealth: wealth in broad 2012, and Costa, 2016). Wealth in broad sense assesses, at two a given moinformation. At best, information is only avail- and net wealth. mentofintwo time,concepts the rangeof of wealth: decisionswealth that houseality is assessed on the basis in broadable on probability distributions. holds may make in the course of their h in broad sense assesses, at a given moment in time, the range oflifetime decisions that households may make At a given moment in time, the sum of the expect- and, therefore,provides provides aagood of the course of their lifetime and, therefore, goodsummary summary of the resulting well-being. This at a given moment in time, the range of decisions that households may makeed value of labour income in the course of a liferesulting well-being. This aggregate may be brogate may be broken down into two very different components, as illustrated by the following equation: nd, therefore, provides a good summary of the resulting well-being. Thistime, discounted back to that moment, deterken down into two very different components, n into two very different components, as illustrated by the following equation:mines the household’s human wealth. This aggreas illustrated by the following equation: gate includes pension income, given its close link to labour market participation. ܹ݈݁ܽ ݁ݏ݊݁ݏ݀ܽݎܾ݄݊݅ݐൌ ݄ܰ݁ݐ݈ܽ݁ݓݐ ݄ݐ݈ܽ݁ݓ݊ܽ݉ݑܪ (1) In practical terms, (1) ܹ݈݁ܽ ݁ݏ݊݁ݏ݀ܽݎܾ݄݊݅ݐൌ ܰ݁ ݄ݐ݈ܽ݁ݓݐ ݄ݐ݈ܽ݁ݓ݊ܽ݉ݑܪ each household was assumed to have a wide hori- (1) zon when making its decisions, as would be the On the one hand, the net wealth (or non-human case when using an infinite-horizon agent mod- wealth) of each household when the period el. Assuming that labour income corresponds to e one hand, the net wealth (or non-human wealth) of each household when the period under review s results from the difference between its assets and its liabilities. This net wealth, by construction, is (or non-human wealth) of each when the from period reviewa perpetuity, and taking into account a discount underhousehold review begins results theunder difference termined and generates some return, including capital gains, over time. On the other hand, there is ce between its assets and its liabilities. This net wealth, by construction, is o-called human wealth, i.e. the sum of current and future income resulting from household ome return, including capital gains, over time. On the other hand, there is ons as to how to spend time in the labour market. .e. the sum of current and future income resulting from household he in lifetime of each household, human wealth will depend on exogenous conditions – such as the labour me the labour market. et environment and structural changes – and idiosyncratic features of the household under review. old, human wealth will depend on exogenous conditions – such as the labour Special issue rate of 4%, this would imply that human wealth for wealth in broad sense and net wealth com- may be proxied by multiplying current labour prise different households. Indeed, there is a income by 25. This proxy is used in the exercises relatively small overlap between households below, although results are qualitatively robust to across distribution quintiles, as illustrated in the use of different discount rates. Equation (1) Table 2. For instance, only 48 per cent of house- can therefore be re-written as follows: holds in the lowest quintile of wealth in broad 1 95 sense are also included in the lowest net wealth ܹ݈݁ܽ ݁ݏ݊݁ݏ݀ܽݎܾ݄݊݅ݐൌ ܰ݁ ݄ݐ݈ܽ݁ݓݐ ݄ݐ݈ܽ݁ݓ݊ܽ݉ݑܪ ൎ ݄ܰ݁ݐ݈ܽ݁ݓݐ ʹͷ ൈ ݁݉ܿ݊݅ݎݑܾܽܮ quintile. This different household composition (2) in both measures of wealth distribution plays a (2) key role when interpreting the results obtained ݁ݏ݊݁ݏ݀ܽݎܾ݄݊݅ݐൌ ܰ݁ ݄ݐ݈ܽ݁ݓݐ ݄ݐ݈ܽ݁ݓ݊ܽ݉ݑܪൎ ܰ݁ ݄ݐ݈ܽ݁ݓݐ ʹͷ ൈ ݁݉ܿ݊݅ݎݑܾܽܮ in Section 3.4 (2) Knowing the distribution sum of these two types of wealth among households representing a given ݄ݐ݈ܽ݁ݓ݊ܽ݉ݑܪ ൎ ݄ܰ݁ݐ݈ܽ݁ݓݐ ʹͷof ൈ the ݁݉ܿ݊݅ݎݑܾܽܮ Table 1 illustrates the high inequality levels in net (2) population provides a good measure of inequality among that population. The way in which this distribution wealth and wealth in broad sense in Portugal. distribution the sum–ofincluding these reacts to Knowing structuralthe changes in theofeconomy a persistent policy change – provides a measure In particular, the average istribution ofofthe sum of these two types of wealth among households representing a givenvalue of net wealth in two types of wealth among households repthe distribution impact of that change. first distribution quintile is close to zero. This ovides a good measure of inequality among that population. The waythe in which this distribution wo types of wealth among households representing a given resenting a given population provides a good To simplify the analysis, in this article households are broken down into wealth in broad sense and net wealth explains the high inequality tural changes the economy –ofincluding aamong persistent policy change – provides a measurein net wealth as measity among thatinpopulation. The way in which this distribution measure inequality that population. 2 quintiles, assuming that the average of each quintile summarises theQ5/Q1 corresponding population group. ured by the ratio. Turning to wealth in The tion impact of that change. ncluding a persistent policy provides a measure The way change in which– this distribution reacts to ratios used to gauge inequality are the ratio of the average population to its thedispersion intermediate quintile (Average/Q3), broad sense, is much lower, given structural changes in the economy – including a broad sense and net wealth analysis, in this article households are broken down into wealth in the ratio of the highest quintile to the intermediate quintile (Q5/Q3) and the ratio of the highest quintile to that human wealth’s dispersion is relatively low. persistent policy change – and provides a measure 2 ming thethe average of each quintile summarises the corresponding population group. The s arethat broken down into wealth in broad sense net lowest quintile (Q5/Q1). The higher thewealth ratios, the more unequal the distribution is. These Overall, the heterogeneity prevailing in theratios distri-should of the distribution impact of that change. 2 auge inequality are the ratio of the average population to the intermediate quintile (Average/Q3), intile summarises the corresponding population group.while Thethe Average/Q3 ratio is not very sensitive to small changes in be interpreted with caution. For instance, bution of net wealth and wealth in broad sense eaverage highestpopulation quintile totothe intermediate quintile (Q5/Q3) and theifratio of the are highest quintile to 3 the intermediate quintile (Average/Q3), To simplify analysis, inreacts this article the distribution, the the Q5/Q1 ratio very households strongly Q1 figures very suggests that thelow. potential redistribution impact ntile higher the ratios, more unequal the distribution is. These ratios should diate(Q5/Q1). quintileThe (Q5/Q3) and the ratio ofinto the highest quintile to are broken downthe wealth in broad sense and of monetary policy on the sense Portuguese economy Table 1 shows descriptive statistics on the level and distribution of wealth in broad and net wealth in the with For instance, while the Average/Q3 ratio is not very sensitive to small changes in os, thecaution. more unequal thewealth distribution is. assuming These ratios net quintiles, thatshould the average cannot be ignored. Portuguese economy, on the basis of the Household Finance and Consumption Survey for 2010 (to see a n,the theAverage/Q3 Q5/Q1 ratioratio reacts very strongly ifsummarises Q1 to figures very low. is each not very sensitive smallare changes in 3 of quintile the correspondcomparison between the outcome of the Household Finance Consumption Survey 2010exercisand that for Thisand article includes a number of for stylised 3 2 ngly if Q1 figures are very low. The ratios used to ingthe population descriptive statistics on level2016). andgroup. distribution of wealth in gauge broad sense and wealthininbroad the sense and net wealth 2013, see Costa, The various distribution quintiles fornet wealth es discussing in what way monetary policy chang- inequality aresense the ratio the average popula- Survey for 2010 (to see a onomy, on the basis of the Household Finance and Consumption nd distribution of wealth in broad andof net wealth in the comprise different households. Indeed, there is a relatively small overlap between households es can be transmitted to household wealth. across tion to the intermediate quintile (Average/Q3), tween Finance the outcome of the Household Finance and Consumption Survey for 2010 and that for ehold and Consumption Survey for 2010 (to see a distribution quintiles, as illustrated in Table 2. For instance, only 48policy per has centdistribution of households Monetary effects in via the sev-lowest the ratio of Survey the highest quintile to the intermedista, 2016). distribution quintiles in broaderal sense and net wealth hold FinanceThe andvarious Consumption for 2010 and that quintile of wealth in broad sensefor arewealth alsoforincluded in the lowest wealth quintile. This different channels. Thenet unexpected impact on price ate quintile (Q5/Q3) and the ratio of the higherent households. Indeed, there is a relatively small overlap between households across tion quintiles for wealth in broad sense and net wealth household composition in both measures of wealth distribution a key role when interpreting levels isplays one of the most studied channels inthe lit- results estoverlap quintile to the lowest quintile as illustrated in 2. 3. For only 48 (Q5/Q1). per centThe of households the lowest 4 instance, eintiles, is a relatively small between households across obtained in Table Section erature. Thisin‘Fisher channel‛ has been analysed higher the ratios, the more unequal the distribuealth in broad also in the lowest net wealthin quintile. This different or instance, onlysense 48 perare cent of included households in the lowest literature for a long time, starting with Fisher Table 1 illustrates the high inequality levels in net with wealth and wealth in broad sense in Portugal. In particular, tion is. These ratios should be interpreted mposition of wealth distribution plays a key role when interpreting the results included in inboth the measures lowest net wealth quintile. This different (1933). Briefly, an shock leading to the average value of net wealth in the first distribution quintile is close to zero.unexpected This explains the high inequality in 4 caution. For instance, while the Average/Q3 ratio ction 3. alth distribution plays a key role when interpreting the results an increase in the consumer price index has a net wealthisas measured by the ratio. Turning wealth in broad sense, its dispersion is much lower, given not very sensitive toQ5/Q1 small changes in the to disredistribution effect by changing the real value tes the highthat inequality levels in net wealth and wealth in broad sense in heterogeneity Portugal. In particular, human wealth’s dispersion is relatively low. Overall, the prevailing in the distribution of net tribution, the Q5/Q1 ratio reacts very strongly if of household exposures.inAn increase in prices uewealth of net wealth in the first distribution quintile issuggests close to that zero.the This explains the high inequality et and wealth in broad sense in Portugal. In particular, wealth and broad sense potential redistribution impact of monetary policy on the 3 Q1wealth figuresinare very low. redistributes from households with measured by the Q5/Q1 Turning to wealth in broad sense,inits dispersion is much resources lower, given ution quintile is close to ratio. zero. This explains the inequality Portuguese economy cannot be high ignored. Table 1 descriptive statistics theprevailing level positive nominal wealth positions to households ealth’s dispersion is relatively low. Overall, heterogeneity in the distribution of net ng to wealth in broad sense, itsshows dispersion isthe much lower,on given This article includes a number of stylised exercises discussing in whatnominal way monetary policy changes can be and distribution of wealth in broad sense and net with negative positions, which means alth in broad sense suggestsprevailing that the in potential redistribution . Overall, the heterogeneity the distribution of netimpact of monetary policy on the transmitted to household wealth. Monetary policy has distribution effects via several channels. The unexpected wealth in the Portuguese economy, on the onomy cannot be ignored. the potential redistribution impact of monetary policy on the basis that it generates a reduction in real wealth for impact on levels is one of the studied channels in literature. This “Fisher channel” has been of price the Household Finance andmost Consumption creditors and an increase in real wealth for debtludes a number of stylised exercises what way policy Briefly, changes be analysed in literature for a discussing long time, in starting withmonetary Fisher (1933). ancan unexpected shock leading to an Survey for 2010 (to see a comparison between ors, if assets held at the time of the policy shock household wealth. policy has distribution effects The unexpected cises discussing inMonetary what wayconsumer monetary policyindex changes beseveral channels. increase in hascan a via redistribution by changing real value of household areeffect denominated in cashthe units. thethe outcome of theprice Household Finance and Conlevels is exposures. one effects of the via most studied channels in literature. This “Fisher channel” has been ycehas distribution several channels. The unexpected An increase in prices redistributes resources from households with positive nominal wealth sumption Survey for 2010 and that for 2013, see raturechannels for apositions longintime, with Fisherchannel” (1933). Briefly, an unexpected shockmeans leadingthat to an udied literature. This “Fisher has been tostarting households negative nominal positions, which it generates a reduction in Costa, 2016). The with various distribution quintiles eith consumer price index has a redistribution effect by changing the real value of household Fisher (1933). Briefly, an unexpected shock leading to an real wealth for creditors and an increase in real wealth for debtors, if assets held at the time of the policy increase ineffect prices redistributes households with positive nominal wealth distribution by changing theresources real valuefrom of household shock are denominated in cash units. ouseholds nominal which means resources with fromnegative households with positions, positive nominal wealth that it generates a reduction in Another redistribution channel attributable to monetary policy results from developments in nominal interest creditors and an increase in real wealth for debtors, if assets held at the time of the policy al positions, which means that it generates a reduction in rates. Persistently increasing the average inflation rate, as well as the short-term nominal interest rate, results in indebtors, cash units. alominated wealth for if assets held at the time of the policy a relatively larger increase in taxes for cash-holding households – with approximately zero nominal profitability 96 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Table 1 • Distribution of wealth in broad sense and net wealth in Portugal | Values in euros Wealth in broad sense Net wealth Q1 291,814 1,282 Q2 386,690 34,010 Q3 495,483 78,982 Q4 698,424 146,449 Q5 1,375.553 529,285 649,441 157,933 Total Q5/Q3 2.8 6.7 Q5/Q1 4.7 412.8 Average/Q3 1.3 2.0 Note: Calculations based on HFCS 2010. Average values. Table 2 • Distribution of the population by quintiles of wealth in broad sense and of net wealth | In percentage Quintiles of wealth in broad sense Net wealth quintiles <20 20-40 40-60 60-80 >=80 Total <20 48 31 16 4 0 100 20-40 26 28 25 18 4 100 40-60 14 21 27 26 11 100 60-80 9 15 20 29 27 100 100 >=80 3 5 11 23 58 Total 100 100 100 100 100 Note: Calculations based on HFCS 2010. Average values. Another redistribution channel attributable to Consequently, the redistribution impact of mone- monetary policy results from developments in tary policy via both channels described above has nominal interest rates. Persistently increasing probably been negligible in the recent past and, the average inflation rate, as well as the short- therefore, is not analysed in this article. By con- term nominal interest rate, results in a rela- trast, the following section looks into the channels tively larger increase in taxes for cash-holding that may have had a significant impact on the dis- households – with approximately zero nomi- tribution of net wealth and human wealth among nal profitability – compared with households Portuguese households in the recent past. These holding assets with positive returns. Given that channels result from the implementation of stand- relatively poorer agents hold a larger share of ard monetary policy measures – more specifically, their wealth in cash, they are negatively affected through changes in the ECB’s intervention rates – by the aforementioned policy change compared as well as the adoption of non-standard monetary with richer agents. This effect is described in policy measures, namely through the purchase of Erosa and Ventura (2002) and Adão and Correia assets by the Eurosystem. (2016). On the one hand, as regards net wealth, the Since the beginning of the euro area, the Por- article analyses the transmission channel via tuguese economy has been characterised by a asset prices and interest rates in various mar- monetary regime where nominal interest rates kets, i.e. the channel through which monetary and the inflation rate are low and not very volatile. policy values (positively or negatively) household Special issue assets and liabilities. This channel is analysed unchanged at a given moment for a specific in detail in Sub-sections 3.1 and 3.2, through a household. Developments in net wealth over 97 natureza tenderá diminuir as due taxas de juro innominais partial equilibrium exercise,desta which adopts a simtime a are exclusively to changes the path (ativas e passivas) e a desta natureza tenderá a diminuir as taxas de juro nominais (ativas e passivas) e a ilar approach to that of Doepke and Schneider followed by prices and returns on assets and aumentar o emprego, a atividade, os preços dos ativos e os lucros das empresas. Estes desta desta natureza natureza tenderá tenderá a diminuir a diminuir as taxas as taxas de juro de juro nom aumentar o emprego, a atividade, os preços dos ativos e os lucros das empresas. Estes liabilities. The question, therefore, is how, due (2006). são os choques estilizados a analisar nas subseções seguintes. desta natureza tenderá a diminuir as taxas de juro aumentar o emprego, a atividade, os preços dos ativos e nominais (ativas e passivas) e a aumentar o emprego, a atividade, os preços dos ativo to monetary policy, developments in taxas thesede juro nominais (ativas e p desta natureza tenderá a diminuir as On the other hand, são os choques estilizados a analisar nas subseções seguintes. as regards human wealth, aumentar o emprego, a atividade, os preços dos ativos e os lucros das empresas. Estes prices or returns são os choques estilizados a analisar nas subseções seguin change over time and how são os choques estilizados a analisar nas subseções se Sub-section 3.3 describes the redistribution aumentar o emprego, a atividade, os preços dos ativos e os lucros das emp A estratégia de identificação adotada mantém fixa a composição da carteira de ativos this leads to different developments wealth e a desta natureza tenderá a diminuir as taxas de juro nominais (ativas e inpassivas) são os choques estilizados a analisar nas subseções seguintes. impact of monetary policy via the labour income A estratégia de identificação adotada mantém fixa a composição da carteira de ativos são os choques estilizados a analisar nas subseções seguintes. A estratégia de identificação adotada mantém fixa a com A estratégia de identificação adotada mantém fixa a for each household, under the assumption of A evolução do valor da num para uma determinada família. channel. This channel arises duedeterminado to changes in momento aumentar o emprego, a atividade, os preços dos ativos e os lucros das empresas. Estes num determinado momento para uma determinada família. A evolução do valor da unvarying portfolios. This exercise is along the para labour income,A estratégia de identificação adotada mantém fixa a composição da carteira de ativos associated with monetary polnum num determinado determinado momento momento para uma uma determinada determinada fam A estratégia de identificação adotada mantém fixa a composição da carte riqueza líquida ao longo do tempo é devida unicamente a alterações nas trajetórias de são os choques estilizados a analisar nas subseções seguintes. lines as that developed byevolução Domanski et valor al. icy measures, num with adeterminado varying impact on differriqueza líquida ao longo do tempo é devida unicamente a alterações nas trajetórias de momento para same uma determinada família. A do da riqueza líquida ao longo do tempo é devida unicamente riqueza líquida ao longo do tempo é devida unicamen num determinado momento para uma determinada família. A evolução preços e retornos dos ativos e responsabilidades. A questão será então como, devido à (2016), taking Doepke and Schneider (2006) as ent segments of the population. This exercise A estratégia de identificação adotada mantém fixa a composição da carteira de ativos riqueza líquida ao longo do tempo é devida unicamente a alterações nas trajetórias de preços e retornos dos ativos e responsabilidades. A questão será então como, devido à preços e retornos dos ativos e responsabilidades. A quest preços e retornos dos ativos e responsabilidades. A q riqueza líquida ao longo do tempo é devida unicamente a alterações nas t is based on Household Finance and Consumpa starting point. política monetária, se altera a evolução ao longo do tempo destes preços ou retornos e num determinado momento para uma determinada política monetária, se altera a evolução ao longo do temp família. A evolução do valor da preços e retornos dos ativos e responsabilidades. A questão será então como, devido à política monetária, se altera a evolução ao longo do tempo destes preços ou retornos e política monetária, se altera a evolução ao longo do t tion Survey data, together with microeconomic preços e retornos dos ativos e responsabilidades. A questão será então com como daí decorre uma evolução diferente da riqueza de cada família, sob a hipótese de 3.1. Transmission channel via asset prices: como daí decorre uma evolução diferente da riqueza de cada família, sob a hipótese de política monetária, se altera a evolução ao longo do tempo destes preços o stock valuation carteiras invariantes. Trata‐se de um exercício na mesma linha do desenvolvido por preços e retornos dos ativos e responsabilidades. A questão será então como, devido à como daí decorre uma evolução diferente da riqueza de cada família, sob a hipótese de carteiras carteiras invariantes. invariantes. Trata‐se Trata‐se de um de um exercício exercício na mesm na m como daí decorre uma evolução diferente da riqueza de cada família, sob a carteiras invariantes. Trata‐se de the um exercício na mesma do desenvolvido por One of transmission channels oflinha mone3. Distribution channels of monetary Domanski et al. (2016), partindo de Doepke e Schneider (2006). política monetária, se altera a evolução ao longo do tempo destes preços ou retornos e carteiras invariantes. Trata‐se de um exercício na mesma linha do desenvolvido por Domanski et al. (2016), partindo de Doepke e Schneider ( Domanski et al. (2016), partindo de Doepke e Schneid carteiras Trata‐se de um inexercício tary policyinvariantes. to the distribution of wealth broad na mesma linha do desen policy in Portugal Domanski et al. (2016), partindo de Doepke e Schneider (2006). política monetária, se altera a evolução ao longo do tempo destes preços ou retornos e data from the riqueza líquida ao longo do tempo é devida unicamente a alterações nas trajetórias de Labour Survey. como daí decorre uma evolução diferente da riqueza de c como daí decorre uma evolução diferente da riqueza como daí decorre uma evolução diferente da riqueza de cada família, sob a hipótese de Domanski et al. (2016), partindo de Doepke e Schneider (2006). sense stems from valuation changes in the stock Domanski et al. (2016), partindo de Doepke e Schneider (2006). How does a monetary policy change affect the 3.1 Canal de transmissão pelos preços dos ativos: valorização do stock 3.1 3.1 Canal de transmissão pelos preços dos ativos: va Canal de transmissão pelos preços dos ativo of assets. At first sight, given the carteiras invariantes. Trata‐se de um exercício na mesma linha do concentration desenvolvido por 3.1in broad Canal de transmissão pelos preços dos ativos: valorização do stock distribution of wealth sense and net 3.1 Canal de transmissão pelos preços dos ativos: valorização do stock of assets in the top wealth quintile (wealth in Um dos canais de transmissão da política monetária à dist Um dos canais de transmissão da política monetária à Um dos canais de transmissão da política monetária à distribuição da riqueza em sentido 3.1 Canal de transmissão pelos preços dos ativos: valorização do sto Domanski et al. (2016), partindo de Doepke e Schneider (2006). wealth of households in Portugal? This section Um dos canais de transmissão da política monetária à distribuição da riqueza em sentido broad sense and,lato decorre de alterações na valorização do stock de at in particular, net wealth), it Um dos canais de transmissão da política monetária à distribuição da riqueza em sentido lato decorre de alterações na valorização do stock d aims to answer this question. The answer is proUm dos canais de transmissão da política monetária à distribuição da riquez lato decorre de alterações na valorização do stock de ativos. À primeira vista, poderia could be expected that an increase in the value lato decorre de alterações na valorização do stock de ativos. À primeira vista, poderia 3.1 Canal de transmissão pelos preços dos ativos: valorização do stock lato decorre de alterações na valorização do stock de ativos. À primeira vista, poderia vided within a partial equilibrium framework and parecer que, dada a concentração dos ativos no último q parecer que, dada a concentração dos ativos no últim lato decorre de alterações na valorização do stock de ativos. À primeira v parecer que, dada a concentração dos ativos no último quintil da riqueza (em sentido of assets would lead to an increase in inequalunvarying composition of investment portfoUm dos canais de transmissão da política monetária à distribuição da riqueza em sentido parecer que, dada a concentração dos ativos no último quintil da riqueza (em sentido parecer que, dada a concentração dos ativos no último quintil da riqueza (em sentido lato e, em particular, em termos líquidos), um aumento lato e, em particular, em termos líquidos), um aume parecer que, dada a concentração dos ativos no último quintil da riqueza ity. However, this is not necessarily the case. lato e, em particular, em termos líquidos), um aumento do valor dos ativos implicaria lios, thereby excluding agents’ reaction to polilato decorre de alterações na valorização do stock de ativos. À primeira vista, poderia lato e, em particular, em termos líquidos), um aumento do valor dos ativos implicaria um aumento da desigualdade. Esta conclusão não é, no e um aumento da desigualdade. Esta conclusão não é, lato e, em particular, em termos líquidos), um aumento do valor dos ativos implicaria Indeed, a specific increase in the value of assets lato e, em particular, em termos líquidos), um aumento do valor dos ativ cy changes. A stylised approach is used on the um aumento da desigualdade. Esta conclusão não é, no entanto, correta. De facto, um parecer que, dada a concentração dos ativos no último quintil da riqueza (em sentido um aumento da desigualdade. Esta conclusão não é, no entanto, correta. De facto, um impacts on wealthdeterminado aumento do valor dos ativos impacta a distri (in broad sense and net) disdeterminado aumento do valor dos ativos impacta a d redistribution impactum aumento da desigualdade. Esta conclusão não é, no entanto, correta. De facto, um of monetary policy via valum aumento da desigualdade. Esta conclusão não é, no entanto, correta. D tribution, not only directly on the assets per se, determinado aumento do valor dos ativos impacta a distribuição da riqueza (em sentido uation changeslato e, em particular, em termos líquidos), um aumento do valor dos ativos implicaria in the stockdeterminado aumento do valor dos ativos impacta a distribuição da riqueza (em sentido of net wealth (Sub- lato e líquida) não só diretamente nos próprios ativos, ma lato e líquida) não só diretamente nos próprios ativos determinado aumento do valor dos ativos impacta a distribuição da riqueza determinado aumento do valor dos ativos impacta a distribuição da riqueza (em sentido but also via the leverage of the various populasection 3.1), income flowslato e líquida) não só diretamente nos próprios ativos, mas também através do grau de generated by this um aumento da desigualdade. Esta conclusão não é, no entanto, correta. De facto, um lato e líquida) não só diretamente nos próprios ativos, mas também através do grau de alavancagem dos diferentes quintis da população. Esta re alavancagem dos diferentes quintis da população. Es lato e líquida) não só diretamente nos próprios ativos, mas também atravé tion quintiles. This link is illustrated by equation lato e líquida) não só diretamente nos próprios ativos, mas também através do grau de stock of net wealth (Sub-section 3.2) and labour determinado aumento do valor dos ativos impacta a distribuição da riqueza (em sentido alavancagem dos diferentes quintis da população. Esta relação encontra‐se explicitada alavancagem dos diferentes quintis da população. Esta relação encontra‐se explicitada na equação (3) para o caso da riqueza líquida, definida co na equação (3) para o caso da riqueza líquida, definida alavancagem dos diferentes quintis da população. Esta relação encontra‐s (3) for net wealth, which is defined as the differincome changes (Sub-section 3.3). alavancagem dos diferentes quintis da população. Esta relação encontra‐se explicitada lato e líquida) não só diretamente nos próprios ativos, mas também através do grau de na equação (3) para o caso da riqueza líquida, definida como a diferença entre os Ativos ence between assets ( as ) and ( ) of giv-um (�) (�) e e Dívidas as debt Dívidas (�) (�) de aum de dado dado quintil quintil da distribuição da distribui na equação (3) para o caso da riqueza líquida, definida como a diferença entre os Ativos na equação (3) para o caso da riqueza líquida, definida como a diferença en The exercise focuses on stylised monetary polna equação (3) para o caso da riqueza líquida, definida como a diferença entre os Ativos en distribution quintile ( �).). The equationrevela showsque o alavancagem dos diferentes quintis da população. Esta relação encontra‐se explicitada A equação (�) e as Dívidas (�) de um dado quintil da distribuição (� aumento percentual da riqueza líquida de um determina aumento percentual da riqueza líquida de um determ icy changes towards greater monetary accome as dado Dívidas (�) de um dado quintil (� da � ). distribuição (�revela (�) e as Dívidas (�) (�) de um quintil da distribuição A equação que o r � ). A equação that the percentage increase in net (� wealth of a (�) e as Dívidas (�) de um dado quintil da distribuição ). A equação revela que o � � na equação (3) para o caso da riqueza líquida, definida como a diferença entre os Ativos aumento percentual da riqueza líquida de um determinado quintil, decorrente de um � modation, through a reduction in interest rates �� ��� � aumento percentual da riqueza líquida de um determinado quintil, decor , será igual ao , será igu aumento percentual do valor dos ativos aumento percentual do valor dos ativos aumento percentual da riqueza líquida de um determinado quintil, decorrente de um given quintile, resulting from a percentage rise � �� � �� � and the implementation of non-standard mon- (�) e as Dívidas (�) de um dado quintil ��da � , será igual ao produto entre esse aumento distribuição (�� ). A equação revela que o aumento percentual da riqueza líquida de um determinado quintil, decorrente de um ���� aumento percentual do valor dos ativos � � ���� � �� aumento percentual do valor dos ativos , será igual ao produto entre es etary policy measures. Although assessing all � percentual do valor dos ativos e um fator percentual do valor dos ativos e um fator que depe que d in the value of assets , is obtained by�mul� aumento percentual do valor dos ativos ��� ��� aumento percentual da riqueza líquida de um determinado quintil, decorrente de um ���� ��� , será igual ao produto entre esse aumento �� ��� � � aumento percentual do valor dos ativos , será igual ao produto entre esse aumento � � � � transmission mechanisms poses a number of � percentual do valor dos ativos e um fator � que depende positivamente do grau de � �� �this percentual do valor dos ativos e um fator �� � percentage � tiplying increase in the value �of�� que depende positivament ���� � � aumento percentual do valor dos ativos , será igual ao produto entre esse aumento percentual do valor dos ativos e um fator que depende positivamente do grau de challenges, one can argue that such a monetary � �� � alavancagem desse quintil. � alavancagem desse quintil. � �� � ��� percentual do valor dos ativos e um fator assets by� a factor impulse will tend to reduce (lending and deposalavancagem desse quintil. � �� que depende positivamente do grau de ��� that ���positively depends �� � alavancagem desse quintil. percentual do valor dos ativos e um fator � que depende positivamente do grau de � � � � � � � � � � � �� � �� � � �� � � � � ��1 � � it) nominal interest rates alavancagem desse quintil. and foster employ����� ������ � � ��� �� 1 �� � on leverage � � � � � � in that quintile. �� � �� �� � �� ��� ����� � � �� �� � �� ��� ��� � �� � �� �prices �and �� �� �� � � � � � � � � � ment, activity, asset business profits. alavancagem desse quintil. �� �� 1 �� ��� � � � �� � �� �� �� � alavancagem desse quintil. � ��� �� �� � � � ��� � 1 1 �1 � �� �� � � �� analysed � � fol� � These stylised shocks are ��� � ��� � �� ��� in ��the ��� ��� ��� � � �� �� � �� � � � � � � � �� ����� � ��� � ���� ����1���� 1 � 1 � �� � �� �� �� �� lowing sub-sections. ����� � � �� � � ��� �� 10 � �� ��� �� ��� � � �� �1 � ���� �� ��� � �� � 10 � ����� � ��� � ���� 1 � � � � � ��� (3) 1 � �� �� identification �� � �� �� ��� � �strat�� �� �� According to the �� ��� 10 ��� � �� �� ��adopted � �� � � � � � � ���� � � � � egy, the asset portfolio composition remains 10 10 10 10 1� ��� 1� ��� ಲ ο ೂ e value of assets ଵ ( ), is obtained by multiplying this percentage increase in therelated value to leverage in each quintile but also negativ increase in assets is not only positively ೂ ೂ ೂವ aଵವfactor (ଵି ) depends that positively on leverage that quintile. ο ratio depends ofinhuman wealth to totalinassets in each quintile: ೂ ) that positively on leverage that quintile. ೂ assets ( ି ೂಲ ), is obtained by multiplying this percentage increase in the value ಲೂ ೂ ೂ ೂ ೂ ೂ οሺ ି ሻ ο ο t positively depends leverage in that quintile. BANCO DE PORTUGAL Economic ൌೂ ೂ ೂBulletin ೂ• May 98 on ାுௐ ೂοሺൌ ೂೂ 2017 ି ሻ ೂ ο ( ೂ•ି ೂ ି ൌ ( οሺ ೂ ିೂ ሻ ିೂ ሻ ൌ ೂ ିೂ ೂ ο ೂ ( ೂ ο ೂ ଵ ೂ ೂ ೂൌ ೂ ೂ ି ି ሻ ೂ ο ವೂ ೂοሺ ଵି ೂ ಲ ೂ ೂ ೂ ି ൌ ) ೂ ( ൌ ೂ ାுௐ ೂ ିೂ ο ೂ ଵ ೂ ೂ ೂ ) ವೂ ିin In the case of wealth broadsense, ଵିthe impact ಲೂ of changes in assets is illustrated by equation (4). ೂ ೂ ೂ ଵ ೂವೂ ) ൌ οೂ ೂ ଵି ೂ ೂ ାுௐಲೂೂ ି (3) ଵ ಹೈೂ ವೂ ଵା ೂ ି ೂ ಲ ಲ ) In the case of net wealth, the Portuguese econo- (3) my is characterised by a population where grow- of wealth ο ଵ in broad sense, the impact of changes in assets is illustrated by equatio Inοthe case ൌ sense, ൌ In this h in changes is illustrated by equation (4). corresponds In the wealth aggregate includes human ing wealth to ೂ broad ೂ ೂ the ೂimpact ೂofcase, ) in assets (3)a decreasing indebtವೂ ି ொ ொ this As toincludes assess the impact transmission it is necessary case, the wealth aggregate human wealth, ( ܹܪ ),is highly and, therefore, the impact of a per , and, therefore, the impact of a perܹܪ wealth, ggregatethis includes human wealth, (ଵି ), and, therefore, thedistribution impact of a edness percentage ratio.via This significant and channel, implies ಲೂsuch, of wealth in related broadtosense, the impact of in assets illustrated equation (4). In househ centage increase inquintile assets the isbut notchanges onlynegatively positively ot only positively leverage in only each also to is the thatand a similar increase in the value link between asset aggregate the amount of by liabilities heldnegatively by each increase in assets is not positively related to related leverage inpercentage each quintile but also relate ொ negarelated to leverage in each quintile but also h wealth to total assets in eachincludes quintile: of assets across net wealth distribution quintiles ܹܪ aggregate human wealth, ( ), and, therefore, the impact of a percentage distribution quintiles. This link is shown in Charts 1 and 2, for the distribution of wealt ratio of human tively wealth to total assets in each quintile: related to the ratio of human wealth to total always improves the relative of the lowsense, the only impact of changes is illustrated by equation (4). Inposition and net in wealth respectively. Thequintile slope of the lines each quintile corresponds ssets is not positively related toassets leverage in each but also for negatively related to the to levera assets ೂin each quintile: est quintiles vis-à-vis the average quintile and the ொ debt-to-assets ο ଵ the ratio. the The charts show the highest wealth quintiles (both net and ܹܪ cludes human wealth, ( ο ), and, therefore, impact of athat percentage ൌ ൌ an wealth to total assets in each quintile: ) highest quintiles. This means that higher asset ಹೈೂ ವೂ ௐ ೂ ିೂ ೂ ೂ ାுௐ ೂ ିೂ ೂ ି ଵା ೂ ೂ lowest ಲೂೂ have the οሺ ೂ ାுௐ ି ሻ ο ೂleverage. ೂ negatively ο ೂincrease ଵ relatively sitively related to leverage inಲeach quintile but also related to the values more the net wealth ൌ ൌ ) ( ೂ ାுௐ ೂ ିೂ ಹೈೂ ವೂ ೂ ೂ ାுௐ ೂ ିೂ ೂ (4) ି ೂ regardless of poorerଵା households, of asset con- by a populatio is characterised ssets ೂin each quintile: ೂ In theೂ case of netೂ wealth, the Portuguese ಲೂeconomy ಲ οሺ ାுௐ ೂ ିೂ ሻ ο ο ଵ centration. Similarly, in the case of a decrease in corresponds a decreasing ൌ ೂ to ಹೈ ೂ ವೂ ) indebtedness ratio. This is highly significant and imp ( ೂ ାுௐ ೂ ିೂ ൌ ೂ ೂwealth ାுௐ ೂ ି ೂ ுௐ ೂ ିೂ ሻ ೂ ೂ the value of assets, more leveraged households e distribution impact via this transmission channel, it is ଵା necessary to look into the (4) ೂ ି ೂ percentage increase in ಲthe of assets across net wealth distribution quintiles alw value ಲ would post a larger decline in net wealth. set the amount held by each household typifying the ೂ οaggregate ೂand ο ೂ of liabilities ଵ relative position of the lowest quintiles vis-à-vis the average quintile and the (4)highest qui ൌThisೂlinkೂis shown ൌ ೂ1 and ೂ distribution 2,ೂforವthe of wealth in broad sense ) In the case of the distribution of wealth in broad ಹೈ ାுௐ ೂ ିinೂCharts ି ೂ asset values ଵା that increase relatively more thechannel, net wealthit of households, the distribution impact via this transmission is poorer necessary to lookre higher ಲೂ ಲ corresponds ctively. As The such, slope ofto theassess lines for each quintile to leverage, measured by sense across quintiles, the link between leverage concentration. Similarly, in case of asense) decrease in the value assets, more leveraged h tio. Thelink charts show that the highest wealth quintiles (both net and in broad (4) and wealth is not so clearly monotonic (exclud-household typif between the asset aggregate andimpact thethe amount of liabilities held byofeach As such, to assess the distribution via post a larger decline in nettowealth. age. assess the distribution impact via this transmission it is necessary look the ing the highest Therefore,to in the caseinto of this transmission channel, is necessary look channel, distribution quintiles. This link is itshown in Charts 1 and 2,quintile). for the distribution of wealth in broa a homogeneous increase in the value of assets, into the linkcharacterised between the asset aggregate and where ealth, Portuguese economy is by adistribution population growing n thetheasset aggregate and the amount of liabilities byin each household the In the case of slope the of held wealth broad sense acrosstypifying quintiles, the link meas betwe and net wealth respectively. The of the lines for each quintile corresponds to leverage, the final impact on inequality will result from the the amount of liabilities held by each household o a decreasing indebtedness ratio. This is highly significant and implies that a similar tion impact vialink thisistransmission it is2,necessary to(excluding look into the highest wealth ischannel, not so clearly monotonic quintile). Therefore, in quintiles. This shown in Charts 1 and for the distribution of the wealth in broad sense weighing of relative share of human wealthnet and in broad ratio. The charts show highest wealth quintiles (both typifying the distribution quintiles. Thisthat link isthe in the the valuedebt-to-assets of assets across net wealth distribution quintiles always improves thethe increase in the value of assets, the impact measured on inequalityby will result fro gate and the amount of inhomogeneous liabilities held bydistribution each household typifying lth respectively. The slope of the each quintile to final leverage, inThis assets and leverage, asthe illustrated by equation Charts 1lines and and 2, for forthe the of corresponds e lowest quintiles the average quintile highest quintiles. means have the vis-à-vis lowestshown leverage. the relative share ofwealth human wealth inChart assets andthat leverage, as illustrated by equation (4). C (4). 3broad shows the ratio of in human wealthsense) is in Charts 1theand 2, for the distribution of wealth sense wealth in broad sense and net respeces shown increase relatively net wealth of poorer households, regardless ofin asset assets ratio. Themore charts show that the highest wealth quintiles (both net and broad ுௐ ೂ to assets (the term in )equation the ratio of human wealth tohouseholds assets (the term ( in equation (4)) is lower for the h tively. The slope the lines forleveraged each quintile Incase the of net wealth, the Portuguese economy is characterised by(4))ais lower population where in the of case a lines decrease ineach the value of of assets, more would ey,slope of the for quintile corresponds to leverage, measured ೂ by west leverage. for the highest quintile of wealth in broad sense corresponds to leverage, measured by the debtn net wealth. wealththat corresponds towealth a decreasing indebtedness ratio. This issense) highly significant and leverage implies inthat wealth in broad sense than for the other in contrast to the lower this harts show the highest quintiles (both net and in quintiles, broad than for by the other quintiles, in contrast to the to-assets ratio. The charts show that the highest of net wealth, the Portuguese economy is characterised a population where growing tribution of wealth in broad sense quintiles, linkreveals between leverage and is predominant. empirical simulation which effect Chart 4 shows thisimpro ratio percentage increase inacross the(both value ofthe across netleverage wealth distribution always lower in this quintile. Only anquintiles empirical wealth quintiles net and inassets broad sense) early monotonic (excluding the highest quintile). Therefore, inhighly the case of a esponds to a decreasing indebtedness ratio. Thisinisthe significant and implies that a similar quintiles, but has no role simulation reveals which effect is predominant. lowest leverage. relative positionhave ofthe the lowest quintiles vis-à-visanalysis. the average quintile and the highest quintiles. Thi se in the value of economy assets, the final impact on inequality will result from the weighing of growing Portuguese characterised bywealth a population where increase in the value of isassets across net distribution quintiles always improves For that purpose, Table 3 (first column) presents a of stylised simulation of thethe impact of cha that higher asset values increase relatively more the net wealth poorer households, regardless uman wealth in assets and leverage, as illustrated by equation (4). Chart 3 shows that asing indebtedness ratio. is highly significant and implies that acase similar tion of the lowest quintiles vis-à-vis the average quintile and inthe This means ுௐ ೂ This ratios, thehighest of quintiles. bothmore the wealth in broad sense an ealth toconcentration. assets (the term ( Similarly, )ofin assets equation (4))some is lower for the highest quintile of in thefor case of ainterquintile decrease in the value of assets, leveraged household ೂ ue of values assets across net wealth distribution always improves theHousehold distributions. Calculations made on the basis of the and Consum asset increase relatively more the net quintiles wealth of poorer households, regardlessFinance of asset than for the a other quintiles, in contrast the lower leverage inwere this quintile. Only an post larger decline in nettowealth. The table confirms that similar increasehouseholds in the value ofwould assets across dis quintiles vis-à-vis average quintile and highest quintiles. This means which is predominant. Chart 4in shows this ratio for net percentage wealth n.reveals Similarly, ineffect the the case of 2010. a decrease thethe value ofaassets, more leveraged Chart 1 • Relation between assets and liabilities, Chart 2 • Relation between assets and liabilities, In analysis. the case ofof the distribution of wealth inbybroad sense across quintiles, linkinequality between lever role in the fosters inequality in the distribution of wealth in broad sense and the reduces in net we by net quintiles wealth in broad sense net wealth quintiles edecline relatively more the net wealth of poorer households, regardless of asset in wealth. is not so clearly (excluding the quintile). thebyca le 3 (first presents stylised of the impact of changes inhouseholds the value Thesimulation table also breaks down the impact ofhighest the valuation of the Therefore, various assetsinheld ho ase ofwealth acolumn) decrease in athe value of monotonic assets, more leveraged would of the distribution of wealth in broad sense across quintiles, the link between leverage and nterquintile ratios, in the case of both the wealth in broad sense and the net wealth homogeneous increase in the broken value of assets, theassets final impact on inequality will result from the we are typically down into real and financial assets. Real assets chiefly include re h. were ons on the basis of the Household Finance andhighest Consumption Survey for ot so made clearly monotonic (excluding the quintile). Therefore, in the case of a 5 employment businesses , and a large share in net wealth households Portugal an the relative shareincrease of human wealth in assets andhold leverage, as illustrated by of equation (4).inChart 3 sh ms that a similar percentage in the value of assetsthe across distribution quintiles of wealth in broad sense across quintiles, link between leverage and ೂ us increase in the value of assets, the final impact on ுௐ inequality will result from the weighing of e distribution of wealth broad sense and reduces inequality net wealth the ratio of in human wealth to assets (theinterm ( distribution. ) in equation (4)) is lower for the highest qu ೂ in the case of a notonic (excluding thein assets highestand quintile). hare of human wealth leverage,Therefore, as illustrated by equation (4). Chart 3 shows that down wealth the impact of the valuation of the various assets held by households. Assets to the lower leverage in this quintile. ೂother in broad sense than for the quintiles, in contrast ுௐ value of assets, the final impact on inequality will result from thelower weighing of highest quintile of human wealth to assets (the term ( ) in equation (4)) for the own into real assets and financial assets. Real assets ೂchiefly include real estate andisself empirical which effect is predominant. Chart alth assets andsimulation leverage, asreveals illustrated byin equation 3 shows that 4 shows this ratio for net ses5,in and hold a large share in net wealth of households Portugal and(4). mostChart euro area oad sense than for the other quintiles, in contrast to the lower leverage in this quintile. Only an ೂ no role in the analysis. quintiles,Note: but has ுௐ Based on the Household Finance and Consumption Survey for Note:highest Based on the Household Finance and Consumption Survey for ssets (the term ( ) in equation (4)) is lower for the quintile of mulation reveals which effect is predominant. Chart 2010. ೂ 2010. 4 shows this ratio for net wealth For that purpose, Table 3to(first presents a stylised simulation other quintiles, contrast thecolumn) lower leverage in this quintile. Only anof the impact of changes in t uthe has no role in thein analysis. assetsisforpredominant. some interquintile ratios, in the both thewealth wealth in broad sense and the ne whichof effect Chart 4 shows thiscase ratiooffor net pose, Table 3 (first column) presents a stylised simulation of the impact of changes in the value distributions. Calculations were made on the basis of the Household Finance and Consumption Su analysis. r some interquintile ratios, in the case of both the wealth in broad sense and the net wealth 2010. The table confirms that a similar percentage increase in the value of assets across distribution Special issue Chart 4 shows this ratio for net wealth quintiles, made on the basis of the Household Finance but has no role in the analysis. and Consumption Survey for 2010. The table For that purpose, Table 3 (first column) presents confirms that a similar percentage increase in a stylised simulation of the impact of changes in the value of assets across distribution quintiles the value of assets for some interquintile ratios, fosters inequality in the distribution of wealth in the case of both the wealth in broad sense and in broad sense and reduces inequality in net the net wealth distributions. Calculations were wealth distribution. 99 Table 3 • Impact of changes in the value of assets on the distributions of wealth in broad sense and net wealth | Percentage changes in initial ratios, excluding ratio values Ratios based on quintiles of wealth in broad sense Initial ratios 10% increase in real estate, tradable assets and businesses (1) (2) (3) (4) Q5/Q3 3.6 0.305 -0.277 0.063 0.533 Q5/Q1 11.9 0.343 -0.292 0.060 0.589 Average/Q3 1.4 0.150 -0.157 0.034 0.279 10% increase in real estate 10% increase in tradable assets 10% increase in the values of businesses 10% increase in real estate 10% increase in tradable assets 10% increase in the values of businesses Ratios based on net wealth quintiles Initial ratios 10% increase in real estate, tradable assets and businesses (1) (2) (3) (4) Q5/Q3 6.7 -1.411 -3.221 0.120 1.877 Q5/Q1 412.4 -48.701 -49.625 -0.015 1.686 2.0 -0.761 -1.967 0.076 1.255 Average/Q3 Note: Calculations based on HFCS 2010. Chart 3 • Relation between human wealth and assets, by quintiles of wealth in broad sense Note: Calculations based on the Household Finance and Consumption Survey for 2010. Chart 4 • Relation between human wealth and assets, by net wealth quintiles Note: Calculations based on the Household Finance and Consumption Survey for 2010. 100 BANCO DE PORTUGAL • Economic Bulletin • May 2017 The table also breaks down the impact of the val- from the impact of high leverage on the first net uation of the various assets held by households. wealth quintile, as explained above (see Chart 2). Assets are typically broken down into real assets Similarly, self-employment businesses also con- and financial assets. Real assets chiefly include tribute, as expected, towards increased inequality real estate and self-employment businesses5, and hold a large share in net wealth of households in in wealth distribution. 3.2. Transmission channel via return on net Portugal and most euro area countries. In turn, Por seu turno, os ativos financeiros são fundamentalmente compostos por depósitos e Por seu turno, os ativos financeiros são fundamentalmente compostos por depósitos e financial assets mainly comprise deposits and assets: income flows títulos transacionáveis, que incluem ações, obrigações e unidades de participação em títulos transacionáveis, que incluem ações, obrigações e unidades de participação em marketable securities, including shares, bonds o, os ativos financeiros são fundamentalmente compostos por depósitos e fundos de investimento. fundos de investimento. and mutual fund investment units. A monetary policy change may lead to transformations in the distribution of household wealth, cionáveis, que incluem ações, obrigações e unidades de participação em In the case of a percentage increase in a comnot only due to changes in asset valuation (SubPara o caso de um aumento percentual de uma componente do ativo total das famílias, Para o caso de um aumento percentual de uma componente do ativo total das famílias, ponent of total household assets, e.g. housing, vestimento. section 3.1), but also due to changes in income �� por exemplo a habitação, designada por � , a equação (3), relativa à distribuição da equation (3) on net wealth represented as ��,, a equação (3), relativa à distribuição da por exemplo a habitação, designada por � flows stemming from policy changes. This subdistribution would be re-written as follows (and e um aumento percentual de uma componente do ativo total das famílias, section analyses the issue, once again by simuriqueza líquida, seria reescrita como segue (e analogamente a equação (4) para o caso riqueza líquida, seria reescrita como segue (e analogamente a equação (4) para o caso analogously for equation (4) on wealth in broad �� a habitação, designada por � , a equação (3), relativa à distribuição da da riqueza em sentido lato): da riqueza em sentido lato): sense): lating stylised shocks applied to microeconomic data available in the Household Finance and Con- da, seria reescrita como segue (e analogamente a equação (4) para o caso Survey for 2010. More specifically, the �� �� �� ���� �� � ������ �� �� � ��sumption 1 �� � � �� 1��� ����� � ��� � �� �� � ���� m sentido lato): � � ������������������������������������������� � � ������������������������������������������� consists in assessing the impact ��� �� ���� ����� � � ��� stylised ��exercise � �(5) ��� � � �� � �� ��� ������ � �� ����� ������ �� 1 � �� 1 � �� on the � distribution of net wealth and wealth in � �� � �� �� � �� ��� 1 �� �� � �� broad sense of: (i) a 4 p.p. reduction in the inter� � � � � � � ������������������������������������������� � ��� � � � � � � � ��� Deste modo, um mesmo aumento percentual nos preços da habitação nos diferentes � � � � Deste modo, um mesmo aumento percentual nos preços da habitação nos diferentes 1 � �� est rate on time deposits, (ii) a 4 p.p. increase in � the return rate on self-employment businesses, quintis impactará a distribuição da riqueza líquida em função do aumento percentual quintis impactará a distribuição da riqueza líquida em função do aumento percentual As such, a similar percentage increase in housing and (iii) a 4 p.p. reduction in interest rates on con um mesmo aumento percentual nos preços da habitação nos diferentes nos preços da habitação de cada quintil, do peso da habitação no total dos ativos de nos preços da habitação de cada quintil, do peso da habitação no total dos ativos de prices across quintiles will impact on net wealth sumer loans and housing loans. ctará a distribuição da riqueza líquida em função do aumento percentual cada quintil, e do respetivo grau de alavancagem. cada quintil, e do respetivo grau de alavancagem. distribution according to the share of housing in This exercise is based on a number of assumpa habitação de cada quintil, do peso da habitação no total dos ativos de total assets and leverage for each quintile. tions. First, it is once again assumed that no deciAs colunas 2, 3 e 4 do Quadro 3 caracterizam o efeito na riqueza de aumentos de 10% As colunas 2, 3 e 4 do Quadro 3 caracterizam o efeito na riqueza de aumentos de 10% e do respetivo grau de alavancagem. Columns 2, 3 and 4 of Table 3 illustrate the effect sions have been made on wealth composition. nos preços da habitação, no valor dos ativos financeiros transacionáveis e no valor dos nos preços da habitação, no valor dos ativos financeiros transacionáveis e no valor dos on wealth of 10% increases in housing prices, on It only varies due to changes in interest/return 3 e 4 do Quadro 3 caracterizam o efeito na riqueza de aumentos de 10% the value of marketable financial assets and on negócios por conta própria. Esta análise é particularmente interessante dada a perceção negócios por conta própria. Esta análise é particularmente interessante dada a perceção rates. Second, it takes into account a change in the value of self-employment businesses. This a habitação, no valor dos ativos financeiros transacionáveis e no valor dos de que a política monetária — em particular a política monetária não convencional — de que a política monetária — em particular a política monetária não convencional — interest rates from 5% to 1% and a change in analysis is particularly useful given the perception conta própria. Esta análise é particularmente interessante dada a perceção return on self-employment businesses from 1% teve um impacto significativo sobre a valorização destes ativos. teve um impacto significativo sobre a valorização destes ativos. that monetary policy – in particular, non-standard to 5%, although results are relatively robust to tica monetária — em particular a política monetária não convencional — monetary policy – had a sizeable impact on the other rate levels compatible with a similar 4 p.p. Os resultados do exercício revelam que um aumento nos preços da habitação contribui Os resultados do exercício revelam que um aumento nos preços da habitação contribui increase in value of these assets. acto significativo sobre a valorização destes ativos. change. Third, it is assumed that, among depospara diminuir a desigualdade da distribuição da riqueza, visto que este ativo tem uma para diminuir a desigualdade da distribuição da riqueza, visto que este ativo tem uma The exercise demonstrates that an increase in its, only time deposits bear interest. Fourth, the s do exercício revelam que um aumento nos preços da habitação contribui housing prices helps reduce inequality in wealth importância relativamente homogénea nos diferentes quintis de riqueza, mas menor no importância relativamente homogénea nos diferentes quintis de riqueza, mas menor no reduction in interest rates affects all time depos- distribution, given that housing has a relatively r a desigualdade da distribuição da riqueza, visto que este ativo tem uma caso do quintil mais elevado de riqueza (Gráficos 5 e 6). Por seu turno, um aumento no caso do quintil mais elevado de riqueza (Gráficos 5 e 6). Por seu turno, um aumento no its and housing loans, taking into account a residhomogenous importance across wealth quin- relativamente homogénea nos diferentes quintis de riqueza, mas menor no ual maturity of onetipicamente year in deposits and the fact valor das ações, e fundos de investimento contribui para valor das ações, obrigações fundos de obrigações investimento contribui tipicamente para tiles,e but lower in the case of the highest wealth that housing loans typically have a floating rate. til mais elevado de riqueza (Gráficos 5 e 6). Por seu turno, um aumento no quintile (Charts 5 and 6). In turn, an increase in aumentar a desigualdade, em linha com a elevada desigualdade na distribuição destes aumentar a desigualdade, em linha com a elevada desigualdade na distribuição destes Fifth, it is assumed that the reduction in inter- theinvestimento value of shares, bonds tipicamente and mutual funds ões, obrigações e fundos ativos, muito concentrados e com um peso relativamente elevado no último quintil da de contribui para ativos, muito concentrados e com um peso relativamente elevado no último quintil da est rates affects 40% of consumer loans, with a typically contributes to an increase in inequality, in desigualdade, em linha com a elevada desigualdade na distribuição destes distribuição (Gráficos 7 e 8). A única exceção é o rácio Q5/Q1 na distribuição de riqueza distribuição (Gráficos 7 e 8). A única exceção é o rácio Q5/Q1 na distribuição de riqueza residual average maturity of 2.5 years. Finally, line with the high inequality in the distribution of concentrados e com um peso relativamente elevado no último quintil da the exercise that time deposits these assets, which are greatly concentrated and alavancagem líquida, o que decorre do impacto da no elevada no primeiro quintil de mature líquida, o que decorre do impacto da elevada alavancagem primeiro quintil de assumes hold a relatively high share in the last distribution over this period and that interest is reinvested. Gráficos 7 e 8). A única exceção é o rácio Q5/Q1 na distribuição de riqueza riqueza líquida, tal como atrás explicado (ver Gráfico 2). Analogamente, o contributo dos riqueza líquida, tal como atrás explicado (ver Gráfico 2). Analogamente, o contributo dos quintile (Charts 7 and 8). The sole exception is the e decorre do impacto da elevada alavancagem no primeiro quintil de Q5/Q1 ratio in net wealth distribution, stemming 15 15 da, tal como atrás explicado (ver Gráfico 2). Analogamente, o contributo dos Similarly, it is assumed that debt maturing over this period is refinanced, including interest. Special issue Table 4 shows results for this exercise, simu- Q1, Q3 and Q5 quintiles – is lower than in net lating shocks that will tend to follow an expan- wealth (Chart 9). Second, an increase in return sionary monetary policy shock. Its main conclu- on businesses leads to greater inequality, due sions are as follows. First, on the asset side, a to their greater share of the highest wealth (net decrease in return on time deposits leads to and in broad sense) quintiles (column 2). Third, an increase in net wealth inequality (column 1), a decline in loan rates tends to reduce inequali- reflecting the combination of greater leverage ty, given that leverage is relatively greater in the for poorer households and a relatively similar lowest wealth quintiles, particularly in the case share of time deposits in total assets among of net wealth. This conclusion stands both for the Q1, Q3 and Q5 net wealth distribution quin- housing loans and consumer and other loans tiles (Chart 10). In the case of wealth in broad (columns 3 and 4). The smaller reduction in sense, the decrease in return on deposits helps inequality in the case of wealth in broad sense reduce inequality. In wealth in broad sense, is due to the lower dispersion in leverage as both the dispersion in the share of deposits well as the fact that loans are more unevenly in total assets and the dispersion in leverage distributed across quintiles in this case. 101 across household quintiles – in particular, the Chart 5 • Relation between real estate and total assets, by quintiles of wealth in broad sense Note: Calculations based on the Household Finance and Consumption Survey for 2010. Chart 7 • Relation between tradable assets and total assets, by quintiles of wealth in broad sense Note: Calculations based on the Household Finance and Consumption Survey for 2010. Chart 6 • Relation between real estate and total assets, by net wealth quintiles Note: Calculations based on the Household Finance and Consumption Survey for 2010. Chart 8 • Relation between tradable assets and total assets, by net wealth quintiles Note: Calculations based on the Household Finance and Consumption Survey for 2010. 102 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Table 4 • Impact of changes in income flows on the distributions of wealth in broad sense and net wealth | Percentage changes in initial ratios, excluding ratio values Ratios based on quintiles of wealth in broad sense Initial ratios 4 pp decrease in the interest rate on deposits 4 pp increase in the rate of return of businesses 4 pp decrease in the interest rate on housing loans 4 pp decrease in the interest rate on consumer loans (1) (2) (3) (4) Q5/Q3 3.6 -0.004 0.158 -0.004 -0.002 Q5/Q1 11.9 -0.006 0.175 -0.004 -0.003 Average/Q3 1.4 -0.001 0.083 -0.002 -0.001 4 pp decrease in the interest rate on deposits 4 pp increase in the rate of return of businesses 4 pp decrease in the interest rate on housing loans 4 pp decrease in the interest rate on consumer loans (1) (2) (3) (4) Ratios based on net wealth quintiles Initial ratios Q5/Q3 6.7 0.041 0.558 -0.042 -0.011 Q5/Q1 412.4 0.702 0.502 -2.242 -1.433 2.0 0.027 0.373 -0.024 -0.005 Average/Q3 Note: Calculations based on HFCS 2010. 3.3. Labour income channel: the redistribution impact on human wealth economy, based on microeconomic data from Monetary policy may also have redistribution and Consumption Survey. The exercise first effects due to the heterogeneous impact on labour remuneration developments. This heterogeneity stems from the combination of two factors: on the one hand, cyclical labour market transitions are known to differ across demographic groups; on the other hand, the distribution of the various demographic groups is uneven across wealth (net and in broad sense) quintiles. Several authors have documented that expansionary monetary policy disproportionally reduces unemployment for young people and less qualified workers. If these population segments are predominant in the lowest wealth distribution quintiles, an expansionary monetary policy will tend to reduce inequality in the respective wealth distribution. the Labour Survey and the Household Finance shows that, comparing the 2011–13 recession with the recovery phase in 2013–16, a considerable heterogeneity emerges in transitions between employment and unemployment according to age and qualification. Afterwards, evidence is presented on how individuals are unevenly distributed across quintiles of wealth in broad sense and net wealth. Given the two time periods under review, in this stage, the exercise used data from the Household Finance and Consumption Survey for 2013, which provides a better picture of the 2011–16 period. Finally, the exercise looks into the redistribution impact of a 1% increase in net employment growth on the Portuguese economy. Since the first quarter of 2011, there have been two distinct stages in the Portuguese economy.6 Up to the second quarter of 2013, an This section analyses the potential importance unprecedented recession in recent times took of this channel in the case of the Portuguese hold. Afterwards, a sustained recovery began, Special issue although more gradual compared with previous more frequent for younger and more qualified economic upturn periods. The two periods are individuals. Comparing the two periods under used as a reference to illustrate heterogeneity in review, i.e. 2011 Q2-2013 Q2 and 2013 Q3-2016 flows between states in the labour market in the Q4, the frequency of flows between employ- Portuguese economy. More specifically, the anal- ment and unemployment declined by approxi- ysis focuses on the percentage change in the mately 25% and the frequency between unem- frequency of transitions between employment ployment and employment increased by around and unemployment, and vice-versa, by break- 18%, both in terms of average quarterly transi- ing down the labour force into four demographic groups according to two predetermined characteristics of individuals: age and qualification. 103 tions. According to microeconomic data from the Labour Survey, the decrease in flows to unemployment was relatively broad-based across age Unsurprisingly, transitions from employment to groups and relatively more concentrated in less unemployment are more frequent for younger qualified groups, while the increase in flows to and less qualified individuals. Moreover, transi- employment was relatively higher in younger tions from unemployment to employment are and less qualified groups (Charts 11 and 12). Chart 9 • Relation between savings deposits and total assets, by quintiles of wealth in broad sense Chart 10 • Relation between savings deposits and total assets, by net wealth quintiles Note: Calculations based on the Household Finance and Consumption Survey for 2010. Chart 11 • Change in the frequency of transitions from unemployment to employment, between 2011 Q3-2013 Q2 and 2013 Q3-2016 Q4 | In percentage 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 -5.0 -10.0 Note: Calculations based on the Household Finance and Consumption Survey for 2010. Chart 12 • Change in the frequency of transitions from employment to unemployment, betwee 2011 Q3-2013 Q2 and 2013Q3-2016 Q4 0.0 -5.0 -10.0 -15.0 -20.0 -25.0 -30.0 -35.0 Total Age < 45 and Age < 45 and Age >= 45 (working-age education education and <65 and persons) below equal to or education secondary above below secondary secondary Source: Labour Survey. Age >= 45 and <65 and education equal to or above secondary Total (working-age persons) Age < 45 and Age < 45 and education education below equal to or secondary above secondary Source: Labour Survey. Age >= 45 and <65 and education below secondary Age >= 45 and <65 and education equal to or above secondary 104 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Charts 13 and 14 confirm that this heteroge- data from the Labour Survey and the Household neity affects differently the various quintiles of Finance and Consumption Survey for 2013. the distributions of wealth in broad sense and The exercise assessed the distribution impact net wealth. In terms of qualifications, the charts show that the less qualified labour force is relatively more concentrated in the lowest quintiles of both distributions. However, in terms of age groups, the two distributions differ. Indeed, in the distribution of wealth in broad sense, the share of individuals under 45 in total labour force is relatively stable across distribution quintiles, while as regards net wealth distribution the share of individuals under 45 is relatively higher in the lowest distribution quintiles. Furthermore, the share of individuals aged 65 or more in total population is relatively higher in the lowest quintiles of the distribution of wealth in broad sense, while as regards net wealth distribution the opposite applies, i.e. older age groups are more concentrated in the highest net wealth quintiles. of a net increase in employment of 1%, taking into account the circumstances characterising the economic recovery seen since 2013. Based on the Labour Survey, a 1 p.p. growth in employment was broken down into the four aforementioned age groups (for the workingage population), similarly to that seen between 2011 Q2–2013 Q2 and 2013 Q3–2016 Q4. Next, the assumption was that this growth occurred regardless of the wealth distribution quintile of the individuals. Under the assumption that, for each demographic group in each wealth quintile, job creation implies an increase in labour income corresponding to remuneration per capita of workers in that population segment, it is possible to estimate income changes in each segment stemming from the 1% aggregate shock on net The combination of these data suggests that employment. To take into account the possible policies contributing to a rise in employment will existence of unemployment benefits, as well as tend to affect the distributions of wealth in broad the fact that income arising from a re-entry to sense and net wealth differently. To quantify the labour market is typically below the average, these effects on the Portuguese economy, a styl- it was assumed that the increase in income for ised exercise was conducted, by cross-checking an individual that finds a job is only 50% of the Chart 13 • Population structure, by quintiles of wealth in broad sense Chart 14 • Population structure, by net wealth quintiles 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 Age >= 65 Age >= 45 and <65 and education equal to or above secondary Age >= 45 and <65 and education below secondary Age < 45 and education equal to or above secondary Age < 45 and education below secondary Note: Calculations based on the Household Finance and Consumption Survey for 2013. Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 Age >= 65 Age >= 45 and <65 and education equal to or above secondary Age >= 45 and <65 and education below secondary Age < 45 and education equal to or above secondary Age < 45 and education below secondary Note: Calculations based on the Household Finance and Consumption Survey for 2013. Special issue income per worker calculated for each popula- monetary policy that causes a continuous reduc- tion segment. This factor affects the magnitude tion in the short-term nominal interest rate for a but not the sign of the results. long period of time. This monetary impulse leads Finally, it was assumed that the increase in labour income fosters both wealth in broad sense and net wealth. In the case of wealth in broad sense, monetary policy is assumed to have no permanent real effects and, therefore, labour income to changes in factors affecting wealth distribution. The ultimate goal is to compare the relative importance of the various channels to wealth distribution changes, while keeping in mind the merely stylised nature of this exercise. increases only have an impact during the peri- Although it can be challenging to pinpoint all od under review. In the case of net wealth, it mechanisms at play, models can be used in order 7 was assumed that the increase in income is fully invested in assets, with no changes to the composition of household portfolios. This choice is not straightforward, but takes into account that additional income expands the range of options open to households at the end of the period. Given the assumptions made, a 1% increase in net employment helps reduce inequality in net wealth and, to a lesser extent, inequality in wealth in broad sense. In the case of net wealth, the ratio of the top to the bottom quintile (Q5/ Q1) falls by 3.8% and the Q5/Q3 ratio by 0.1%.8 In the case of wealth in broad sense, percentage decreases in these ratios are lower. The smaller redistribution impact on wealth in broad sense results, inter alia, from the fact that: (i) the bottom net wealth distribution quintile is associated with very low wealth, (ii) a relatively larger share of population aged over 65 – assumed, in this exercise, to not be working – is included in the bottom quintiles of the distribution of wealth in broad sense, and (iii) younger groups account for a relatively significant share in the top quintiles of the distribution of wealth in broad sense. 4. Stylised quantification of the distribution impact of monetary policy in Portugal after 2010 The previous sections looked into the impact of stylised changes in prices and other relevant factors on wealth inequality indicators. This section to rationalise such effects. For instance, in a NeoKeynesian model, an unexpected reduction in the nominal interest rate leads to a decrease in savings – given that saving has become less appealing – and the resulting immediate increase in consumption. This direct effect produces a temporary rise in economic activity leading to upward pressure on labour demand by firms. The resulting increase in wages and employment is an indirect effect of monetary policy changes, via general equilibrium mechanisms in this type of economy. This indirect mechanism will positively affect human wealth of households and, possibly, net wealth too, as a result of the investment of additional resources in various assets. A further example is the effect of economic activity on the profits of privately-owned firms. Indeed, if, during the monetary stimulus period, GDP grows more rapidly than without the stimulus, corporate profits will also be greater, thus freeing up additional resources to the households that own these firms. The accumulated additional resources correspond to an increase in non-human wealth at the end of the period under review. To calculate realistic paths for the main aggregate variables associated with monetary stimulus that may help in the exercise, the PESSOA model was estimated for the Portuguese economy (see Almeida et al., 2013 and Júlio and Maria, 2017). This stylised simulation calculates the effect of additional monetary accommodation aggregates these channels in their entirety as a on any baseline scenario. response to monetary policy changes. The start- To make this exercise more illustrative of the Por- ing point for this exercise is a stylised change in tuguese experience in the recent past, monetary 105 106 BANCO DE PORTUGAL • Economic Bulletin • May 2017 accommodation is analysed from the end of euro area level. The PESSOA model also takes 2010 onwards, immediately after the Household into account changes in demand for Portuguese Finance and Consumption Survey for 2010, with goods and services by euro area countries, given this stimulus being extended up to the end that they are also affected by reductions in the of 2015. nominal interest rates. During this period, euro area short-term nom- A 3 p.p. monetary accommodation leads to a inal interest rates fell markedly. From 2014 cumulative increase in GDP during the five-year onwards, standard monetary policy was large- period of approximately 5.2%, compared to a ly exhausted given that ECB key interest rates, scenario where no additional accommodation is particularly the deposit facility interest rate, were close to their lower bound.9 Following the introduction in 2015 of non-standard monetary policy measures based on the purchase of sovereign debt securities and other assets, interest rates impacting on households and firms kept on decreasing. When the nominal interest rate is close to the lower bound, the concept of the ‘shadow‛ interest rate becomes useful. In an economy where no lower bound is applied to interest rates and non-standard monetary policy measures have not been implemented, this interest rate would have accommodative monetary effects similar to those in an economy with an interest rate set at the lower bound and nonstandard monetary policy measures in place. The calculation of such shadow interest rates (see, e.g. Wu and Xia, 2017, as well as the references in that study) shows that they continued to decrease up to late 2016. This exercise gauges the impact of interest rate cuts in a benchmark case, e.g. a less aggressive reaction by the ECB in terms of monetary accommodation. implemented. Employment increases by around 3.9% and disposable income by approximately 8.4%. The chart also illustrates developments in EBITDA as a percentage of assets in private Portuguese firms. These developments result from the impact on this corporate profitability measure of the increase in economic activity and, in particular, the increase in GDP during the period under review.11 The monetary stimulus leads to an increase in EBITDA, which, cumulatively over the period, results in a 1.9% increase in value for firms, under the assumption of a reinvestment of earnings.12 These additional resources add, therefore, to initial non-human wealth. Monetary policy also has an indirect impact on the prices of a number of assets. In particular, macroeconomic developments illustrated in Chart 15 lead to changes in housing prices, which, in turn, result in a 1.2% increase from the initial situation.13 Table 5 shows the outcome, in terms of the distributions of wealth in broad sense and net wealth, of the path followed by a number of variables compared to a benchmark case where there is no monetary stimulus. The exercise Chart 15 illustrates possible developments in a employs the same methods and assumptions number of aggregate variables of importance to regarding the increase in value of stocks and the exercise. The nominal interest rate declined flows as described in Section 3. Given that this throughout the five-year period and was nor- is a five-year period, labour income increases malised to a 3 p.p. reduction from the initial val- are accumulated over the five years taking into ue. This cannot be interpreted as an estimate account the gradual rise in employment. For of total monetary accommodation in the course the sake of simplicity, only the Q5/Q3 and Q5/ of the period, but merely as a quantitative Q1 ratios are presented. 10 benchmark allowing a comparison of the relative importance of the various channels analysed in this article. This accommodation is exogenous to the Portuguese economy, as it is decided at Special issue 107 Table 5 • Impact of a prolonged monetary stimulus on wealth distribution | Initial and final ratios and changes stemming from several components of wealth in broad sense and of net wealth Ratios based on quintiles of wealth in broad sense 3 pp decrea- 3 pp decrea1,9% 3 pp decrea- se in the se in the All changes 1,24% increase in se in the interest rate interest rate 3,9% Initial Final except increase in the value of interest rate on housing on consu- increase in ratios ratios All changes employment real estate businesses on deposits loans mer loans employment Q5/Q3 3.636 3.629 -0.20 -0.18 -0.068 0.10 -0.018 -0.18 -0.010 -0.025 Q5/Q1 11.86 11.83 -0.24 -0.19 -0.071 0.11 -0.027 -0.19 -0.014 -0.050 Ratios based on net wealth quintiles 3 pp decrea- 3 pp decrea1,9% 3 pp decrea- se in the se in the All changes 1,24% increase in se in the interest rate interest rate 3,9% Initial Final except increase in the value of interest rate on housing on consu- increase in ratios ratios All changes employment real estate businesses on deposits loans mer loans employment Q5/Q3 6.707 6.490 -3.2 -2.1 -0.83 0.36 0.18 -1.8 -0.056 -1.1 Q5/Q1 412.4 106.4 -74 -61 -20 0.32 3.1 -50 -6.7 -34 Note: Figures result from a simulated change in the nominal interest rate using the PESSOA model and satellite models estimated for the Portuguese economy. Methodological assumptions for the calculation of the impact of the various simulated changes on the distribution are detailed in Section 3. Calculations based on the Household Finance and Consumption Survey for 2010 and 2013 and the Labour Survey. The results indicate that an increase in mone- from a rise in employment. The effect via hous- tary accommodation, as illustrated by Chart 15, ing prices – with housing in Portugal being the under the various methodological assump- main form of non-human wealth of households tions adopted, leads to a sizeable reduction in and being comparatively well distributed – also net wealth distribution inequality, chiefly due plays a key role in that reduction, although the to cuts in interest rates on housing loans and effect of monetary accommodation on housing the increase in resources available stemming prices is low. Difference vis-à-vis the absence of stimulus (percentage points) 10 Chart 15 • Macroeconomic developments in a number of variables following monetary stimulus 8 6 4 2 0 -2 -4 End of 2010 GDP Nominal interest rate Employment EBITDA End of 2015 Disposable income House prices Note: Figures result from a simulated change in the nominal interest rate using the PESSOA model and satellite models estimated for the Portuguese economy. 108 BANCO DE PORTUGAL • Economic Bulletin • May 2017 In turn, improvements in corporate profitabi- in asset prices, business profitability and net job lity stemming from increased GDP growth con- creation, and declines in interest rates on depos- tribute to an increase in inequality, given that its and loans. households in the top two quintiles benefit the Distribution effects via an increase in asset pric- most from this channel. However, households es depend on leverage and the asset portfolio in the bottom quintiles also benefit indirectly composition for each population group. In the from improvements in corporate profitability due case of Portugal, it is shown that households in to a rise in employment, stemming from firms the top wealth quintiles are, on average, less lev- distribution inequality. eraged than those in the bottom quintiles. This Overall effects on wealth in broad sense are lower than those in net wealth. Turning to the impact of labour income, the underlying reasons are analysed in Section 3 and include the relatively larger share of population aged over 65 in the bottom quintiles of the distribution, together with a higher share of younger and more qualified individuals in the top quintiles. As such, the boost given by monetary policy to employment is broadly shared across population groups. The outcome of the exercise is broadly in line with empirical and quantitative literature on the impact of monetary policy on net wealth (see Coibion et al., 2016, Deutsche Bundesbank, 2016 or Furceri et al., 2016). This study shows that, in the case of Portugal and with certain caveats, the outcome can also be applied to wealth in broad sense, although in this case the impact is lower. 5. Conclusions implies that increases in value of assets tend to benefit relatively more households in the bottom quintiles, which, in turn, tends to reduce net wealth inequality. Leverage also decreases with quintiles of wealth in broad sense, but less so than in the case of net wealth. The ratio of human wealth to total assets also decreases in parallel with wealth quintiles. As a result, the distribution effect of increases in the value of assets leads to an increase in inequality in wealth in broad sense. The data suggest that increases in housing prices will tend to reduce inequality in net wealth and wealth in broad sense. In the case of an increase in self-employment businesses and marketable financial assets, it will lead to greater distribution inequality. Reductions in interest rates tend, in turn, to affect the accumulation of wealth via changes in income flows from financial investments and loan instalments. In the case of deposit rates, their reduction leads to a decline in net wealth inequality. This Special Issue looks into the distribution Turning to loans, they reduce inequality in the dis- effects of monetary policy using exercises based tribution of both types of wealth. on empirical data on the Portuguese econo- Net job creation potentially resulting from mon- my. First, wealth is characterised using House- etary policy chiefly benefits younger and less hold Finance and Consumption Survey data, as qualified workers. This effect tends to reduce net well as Labour Survey microdata, and the various wealth inequality and, to a lesser extent, inequal- channels affecting two types of wealth – wealth ity of wealth in broad sense, due to differences in broad sense and net wealth – are analysed. It in population quintile distribution for both types is also demonstrated that households in quintiles of wealth. of wealth in broad sense are not necessarily those The factors that help reduce wealth distribution of the corresponding net wealth quintiles. Exer- inequality are not separable from those that con- cises conducted here assess in a stylised man- tribute to its increase. For instance, the rise in ner the impact of increased monetary accom- employment (which helps reduce inequality) is modation, which will tend to lead to increases largely due to higher corporate profitability (which Special issue contributes to an increase in inequality). There- distribution of various wealth components by fore, it is necessary to analyse the total effect of population groups also leads to a reduction in the various channels on wealth distribution. inequality, although lower. The Special Issue also includes a stylised exercise As a whole, the results describe monetary policy where, using various macroeconomic and statis- as an instrument influencing not only the gener- tical models adapted to the Portuguese econo- al state of the economy, but also cyclical devel- my, all channels are simultaneously simulated. opments in inequality, through several chan- Results are similar to those described above, nels. However, a number of important channels although the relative magnitude of the various in terms of the impact of monetary policy on the channels can now be analysed more realisti- economy as a whole are not included in the anal- cally. In the case of net wealth, the total effect ysis. More detailed research into the distribution leads to a reduction in inequality. In the case of impact of these mechanisms would, therefore, wealth in broad sense, the more homogeneous be an interesting topic for future study. References Adão, B. and I. H. Correia, 2016, ‘Changes in Infla- Doepke, M. and M. Schneider, 2006, ‘Inflation and tion Regimes‛, manuscript. the Redistribution of Wealth‛, Journal of Political Almeida, V., G. Castro, R. M. Félix, P. Júlio and J. R. Economy, 114(6): 1069–97. Maria, 2013, ‘Inside PESSOA – A detailed descrip- Domanski, D., M. Scatigna and A. Zabai, 2016, tion of the model‛, Working Paper No 16/2013, Ban- ‘Wealth Inequality and Monetary Policy‛, BIS Quar- co de Portugal. terly Review (March). Auclert, A., 2016, ‘Monetary Policy and the Redistri- Erosa, A. and G. Ventura, 2002, ‘On Inflation as bution Channel‛, manuscript, Stanford University. a Regressive Consumption Tax‛, Journal of Mon- Banco de Portugal and Statistics Portugal, 2016, etary Economics, 49(4): 761–795. ‘Household Finance and Consumption Survey for Fisher, I., 1933, ‘The Debt-Deflation Theory of Great 2013‛, Press Release, 26 October. Depressions‛, Econometrica Vol. 1, No. 4: 337-357. Coibion, O., Y. Gorodnichenko, L. Kueng, and J. Furceri, D., P. Loungani and A. Zdzienicka, 2016, Silvia, 2016, ‘Innocent Bystanders? Monetary Policy ‘The Effects of Monetary Policy Shocks on Inequal- and Inequality in the U.S.‛, University of Texas-Austin, ity‛, IMF Working Paper WP/16/245. manuscript. Júlio, P and J. F. Maria, 2017, ‘Output in the Por- Costa, S., 2016, ‘Financial situation of the house- tuguese post-2008 period: A general equilibrium holds in Portugal: an analysis based on the HFCS narrative‛, Banco de Portugal Economic Studies, 2013‛, Banco de Portugal Economic Studies, 2(4), 3(2), Banco de Portugal. pp. 15–55, Banco de Portugal. Wu, J. C. and F. D. Xia, 2017, ‘Time Varying Lower Costa, S. and L. Farinha, 2012, ‘Inquérito à situa- Bound of Interest Rates in Europe‛, Working Paper. ção financeira das famílias: metodologia e principais resultados‛, Occasional Paper No 1-2012 of Banco de Portugal. Deutsche Bundesbank, 2016, ‘Distributional effects of monetary policy‛, Monthly Report, September. 109 110 BANCO DE PORTUGAL • Economic Bulletin • May 2017 Notes 1. The 25x multiplier factor applied to labour income estimates the current value of that flow over a long period of time. For instance, a constant income flow updated at a 3.5% nominal rate over 40 years has a current value of approximately 22 times the value of the flow. In practice, this factor does not influence the outcome in the case of reasonable values in the update rate and the remaining period for the flow. As such, a 25x multiplier factor was adopted. 2. The quintiles for a given set of elements characterised by an attribute are the five groups such that each comprises a fifth of the whole, grouped in ascending order of the attribute. Therefore, the first wealth quintile for a population of households corresponds to the 20% least wealthy households, the second quintile corresponds to the 20% of households that come immediately after in terms of wealth, and so on. 3. If households in the intermediate quintile are very differently affected from those in the bottom quintile, the Q5/Q3 and Q5/Q1 ratios may follow different paths. 4.Furthermore, by definition, the first quintile of the distribution of wealth in broad sense largely coincides with the first income distribution quintile. Indeed, approximately 75% of households in the first income distribution quintile are also included in the first quintile of the distribution of wealth in broad sense. By contrast, only one-third of households in the first income distribution quintile is included in the first net wealth distribution quintile. Therefore, the analysis of redistribution effects on the first quintile of wealth in broad sense is a better proxy for the situation of poor households, as typically defined within the European Union (i.e. the lack of monetary income, measured in relative terms against average population income). 5. Self-employment businesses are those where a household member owns and actively participates as a worker/manager. 6. In the first quarter of 2011, the Labour Survey underwent major methodological changes. 7. That is, the increase in labour income is not multiplied by 25 in the estimate for wealth in broad sense developments at the end of the period. 8. In the Household Finance and Consumption Survey for 2013, the first net wealth quintile is negative. Calculations assume that the first quintile posts positive net wealth, such that the Q5/Q1 ratio is close to 412, similarly to the Household Finance and Consumption Survey for 2010 (see Table 1). 9. See the Special Issue entitled ‘Life below zero: monetary policy transmission under negative interest rates‛, December 2016 Economic Bulletin, Banco de Portugal. 10. The path followed by the nominal interest rate is due to the magnitude of the reduction in shadow interest rates between the end of 2010 and the end of 2015 as implied in Wu and Xia (2017), although other references would result in similar values. 11. The impact of an increase in GDP on EBITDA of private firms was gauged using a model where the dependent variable is EBITDA as a percentage of assets of Portuguese small and medium-sized firms, selected among the class of autoregressive models of order zero, one and two, estimated in terms of levels, quarterly differences and year-on-year differences, with up to two explanatory variables of a group that includes: (i) logarithmic values of current GDP, GDP at constant prices, HICP and employment; (ii) the interest rate on the stocks of loans to non-financial corporations and the unemployment rate. The specification was selected using the Akaike information criterion. The chosen model implies a 0.37 elasticity of quarterly changes in EBITDA as a percentage of corporate assets vis-à-vis the year-on-year GDP growth rate. 12. This assumption implies that income flows from self-employment businesses remain constant compared with the initial value. 13. This impact is estimated on the basis of an econometric model that links housing prices in Portugal with variables such as employment, the real interest rate and household disposable income.