Download Deflation and real estate

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

United States housing bubble wikipedia , lookup

Debt wikipedia , lookup

Greeks (finance) wikipedia , lookup

Business valuation wikipedia , lookup

Interest wikipedia , lookup

Real estate broker wikipedia , lookup

Pensions crisis wikipedia , lookup

Financialization wikipedia , lookup

Present value wikipedia , lookup

Stock selection criterion wikipedia , lookup

Hyperinflation wikipedia , lookup

Financial economics wikipedia , lookup

Monetary policy wikipedia , lookup

Inflation wikipedia , lookup

Transcript
Deflation and real estate
Real estate capital gains are muted in a deflationary economy
Name
Email
Andrea van Buren
[email protected]
Thesis MSRE
Assessor 1
Assessor 2
Completion Date
Hans op 't Veld
Arthur Marquard
February 2016
2
Foreword
This thesis is written at the end of my Master of Science of Real Estate (MSRE) studies at the Amsterdam School of Real
Estate. With the completion of this thesis, I look back with satisfaction on the period in which I followed this master.
Not only gives the completion of this thesis, but also the completion of a second university degree a lot of satisfaction.
Starting to write this thesis felt initially as a tough job. In fact, there was little research available on the topic deflation
and real estate from which I could draw, or which I could build on. In addition the writing of a thesis depends for a
large part on self-discipline; there is no social component, which there always was during my study.
In spite of this, actually writing this thesis was a little bit easier than initially thought it would be. This was partly due to
the topicality of the subject. The effect of QE on the development on the inflation rate in European markets last year I
have followed closely. One of the scenarios that MN currently has, is a deflation scenario. MN asks the question what
the impact of deflation would be on real estate returns. With this thesis there is now a comprehensive reply, although
there are still enough related topics that are well suited for continued research. And it is precisely the fact that the
combination of deflation and real estate is a fairly 'virgin' topic, which made the writing of this thesis extra interesting.
Perhaps the writing of this thesis became a little bit easier as well, because I could keep the flow. This is partly due to
my employer MN. MN allowed, if obligations at work would not suffer too much, that I could spend one day per week
to work on this thesis. I am MN very grateful for that. In addition I would like to thank my colleague Marissa
Maradona; she was helpful with the quantitative analyses. Last, but not least, I am grateful to my supervisor Hans op 't
Veld, who gave me constructive feedback during the writing of this thesis.
Andrea van Buren
Amsterdam, February 2016
3
Management Summary
Japan, as the only country in the world, was facing a long time longer periods of deflation. Since the beginning of 2015
also Europe has shown short periods of deflation, which makes deflation(fear) also a hot topic in Europe currently.
With regard to the topic inflation and real estate many studies are published; it is generally accepted that real estate
provides protection against inflation. About the relationship between deflation and real estate less research was done.
This survey deals with deflation and its impact on real estate returns, approached from the pension investor. This is a
reviewing and exploratory research thesis. On the basis of literature I distilled the following hypothesis. Real estate
capital gains are muted in a deflationary economy. This hypothesis is tested on the basis of empirical research on the
Japanese real estate market, given the long period of deflation. The objective of this research is not only understanding
the impact of deflation on real estate returns, but also how a real estate portfolio is performing during deflation. This
research leads to insight into the impact of deflation on real estate returns.
Deflation is "a continued fall in the average price level, measured on the basis of the development of a price index".
Deflation may be caused by supply effects or by demand effects. Deflation is detrimental when having debt. Because
real interest rate increases, the value of debt is increasing. The providers of long-term loans see the value, as well as
their real return on their capital, increase. Another consequence of deflation is that it is more attractive to save,
because real money is worth more later. This results in less expenditure. Also for banks deflation may have a negative
impact. At the beginning the effects of deflation may particularly be positive, with rising real purchasing power and
rising interest margins. However, after expiry there can be problems regarding payment caused by the downward
pressure on wages. In the event of deflation house prices can decrease and the value of the collateral can fall below
the mortgage loan. The bank is therefore facing a greater credit risk. Deflation also leads to the postponement of
consumption expenditure. In the most disastrous scenario the economy is in a downward spiral, with persistent
economic shrinkage and a recession. For pension funds, deflation is particularly serious for the coverage ratio, if the
nominal interest rate is (very) negative caused by deflation. A general measure is monetary easing. In this context, it is
important to distinguish asset price inflation and consumer price inflation. QE has the goal to increase consumer price
inflation. The result is often purely an increase of asset price inflation. This brings risks.
During the lost decade Japan was in prolonged periods of deflation. Causes include demographic developments and
inefficient fiscal and monetary policy. Not only they saved more, also expenditure was postponed. In addition some
believe that the main cause for deflation in Japan has been the currency policy; prices needed to fall enough to reduce
real wages. The Abenomics policy aims to push inflation, but still the inflation target of 2% is not there. The effect of
quantitative easing is questionable in general, not only because the inflation rate is still low (both in Japan and in
Europe), but also because markets devalue their currencies because of QE. This is at the expense of other markets,
which in their turn use QE to improve their competitive position. A risk is a “race to the bottom".
The similarities between Europe and Japan are remarkable: Europe suffers from a very low inflation and some
countries suffer even with deflation, the economic growth disappoints, interest rates break new low records and
solving problems is difficult. In both Japan and Europe inflation is still very low despite QE. Nevertheless, both regions
show significant asset price inflation. Finally, many European countries have, although to a lesser extent than Japan, an
aging population.
The results of the examination with the question whether property provides protection against inflation are not
unambiguously. It depends on whether a long or a short period is considered, and also on what period is considered,
however opinions of scientists are divided. In addition, it appears that the type of data is strongly influencing the
results. Also the one sector responds differently on inflation than the other. It is thus difficult to be able to say welladvised that real estate offers an inflation hedge.
Then an exploratory investigation follows by using regression analyses to analyze the potential impact of deflation on
4
rental prices. The variable to explain is the twelve months moving average of the change in the rental price. The
explanatory variables are the change in vacancy rates and the change in the price level, whether or not with a lag.
Lagging is of great influence in this analysis. The change in the price level shows a significant and positive impact on
the change in the rental price. Vacancy has more impact on the change of the rental price than inflation or deflation.
Adding new office space to the real estate stock in a deflationary environment can therefore be a problem. This is also
shown in Japan, where the increase in supply has contributed to a continued decline of prices. Especially in the light of
the current low interest rate environment it is recommended to be vigilant with many new construction projects. The
low interest rate may be attractive to borrow, in case of deflation there is a large risk behind many new construction
projects in relation to the development of the rental prices.
The impact of deflation in the required return on real estate is twofold. Deflation is often associated with economic
recession. Less demand for real estate on the underlying market ensures falling rents, decreasing values and a rise in
vacancy rates; the risks for the investor are increasing. The investor requires to be compensated for this, leading to a
higher risk premium. On the other hand, in relation to for example fixed income, real estate generates attractive
returns, even in periods of deflation: in the event of a decline in interest rates the real return rises and so does the
spread. A lower risk-free rate, which is plausible in a deflationary environment, provides an environment where
required returns are lower. Because it is difficult to operationalise required returns, the hypothesis is based on the
performance as a result of change in value (capital gain). The hypothesis is as follows: Real estate capital gains are
muted in a deflationary economy. The hypothesis is tested. The results from the various analyses showed that this
hypothesis is supported by statistics. The regression analysis, showing the relationship between the return on capital
gain/change in value and the expected inflation on the one hand and expected economic growth on the other hand, is
significant. The relationship between the return on capital gain/change in value and expected inflation is positive. This
is also evident from the correlation matrix.
When putting these results in a broader context, one could elaborate that a positive relationship between inflation and
returns not always makes sense in an inflationary scenario. Also, the current increase in value of core real estate in
Europe is explained. The main reason for this increase in value can be found in the low interest rates and the
expectation that the interest rate will remain low provisionally because of the further enlarged monetary policy
programme of the ECB. As long as the interest rates remain low and no correction takes place, it is likely that the
relative attractiveness of real estate will remain because of the large spread. In case Europe would face a deflation
scenario, then this will have an adverse effect on returns on real estate.
It is quite conceivable that more surveys about deflation will follow, given the current situation in Europe, the general
expectation that the interest rate will remain low provisionally, the as yet little visible effects of QE on the rate of
inflation, (the inflation target of 2% is currently not there yet) and hence also the increasing social relevance of this
topic.
5
Table of Contents
Foreword ................................................................................................................................................................................ 3
Management Summary.......................................................................................................................................................... 4
Table of Contents ................................................................................................................................................................... 6
1
Introduction ................................................................................................................................................................ 8
1.1 Provocative ................................................................................................................................................................. 8
1.2 The impact of deflation on shares and bonds ............................................................................................................ 9
1.3 The impact of deflation on real estate........................................................................................................................ 9
1.4 Hypothesis .................................................................................................................................................................. 9
1.5 Objective..................................................................................................................................................................... 9
1.6 Research methods ...................................................................................................................................................... 9
1.7 Delimitation and generalizability .............................................................................................................................. 10
2
Deflation ................................................................................................................................................................... 12
2.1 Introduction .............................................................................................................................................................. 12
2.2 What is deflation?..................................................................................................................................................... 12
2.3 Reasons for deflation ................................................................................................................................................ 12
2.4 General effects deflation .......................................................................................................................................... 13
2.5 Impact deflation for pension funds .......................................................................................................................... 14
2.6 Measures against deflation....................................................................................................................................... 15
2.7 Historic development deflation in Japan .................................................................................................................. 15
2.8 Japan scenario in Europe? ........................................................................................................................................ 18
2.9 Conclusion ................................................................................................................................................................ 19
3
Inflation and real estate............................................................................................................................................ 20
3.1 Introduction .............................................................................................................................................................. 20
3.2 Real estate as inflation hedge? ................................................................................................................................. 20
3.3 Investment horizon................................................................................................................................................... 21
3.4 Listed versus non listed real estate........................................................................................................................... 23
3.5 Sectors ...................................................................................................................................................................... 24
3.6 Conclusion ................................................................................................................................................................ 24
4
Deflation and real estate .......................................................................................................................................... 25
4.1 Introduction .............................................................................................................................................................. 25
4.2 Deflation, build up method and real estate returns ................................................................................................. 25
4.3 Deflation, risk-free rate and real estate efficiency ................................................................................................... 27
4.4 Deflation, interest and real estate efficiency............................................................................................................ 27
4.5 Nominal and real performance................................................................................................................................. 27
6
4.6 Deflation and the Japanese real estate market ........................................................................................................ 29
4.7 Conclusion ................................................................................................................................................................ 30
5
Empirical research I: explanation rent by vacancy and/or inflation?....................................................................... 31
5.1 Introduction .............................................................................................................................................................. 31
5.2
Expected Links.......................................................................................................................................................... 31
5.3 Description data........................................................................................................................................................ 31
5.4 Correlation and regression analyzes ......................................................................................................................... 32
5.5 Conclusion ................................................................................................................................................................ 35
6
Empirical research II: review hypothesis .................................................................................................................. 36
6.1 Introduction .............................................................................................................................................................. 36
6.2 Description data........................................................................................................................................................ 36
6.3 Discussion ................................................................................................................................................................. 39
6.4 Conclusion ................................................................................................................................................................ 40
7
Conclusion and reflection ......................................................................................................................................... 41
7.1 Conclusion ................................................................................................................................................................ 41
7.2 Reflection.................................................................................................................................................................. 42
Bibliography ......................................................................................................................................................................... 44
Appendix .............................................................................................................................................................................. 46
7
1
Introduction
1.1
Provocative
Deflation was something unique for a long time. Only one country in the world has had longer periods of deflation and
that is Japan. In the early nineties the Japanese stock crashed followed by a protracted period of deflation. This period
is also called the lost decade: a long ongoing combination of falling prices, low growth and low interest rates. Currently,
Japan still suffers with the ‘deflation ghost’. Deflation means that the price level drops. Long-term deflation is harmful
for economic growth. Large purchases are being postponed because they can be purchased later at a cheaper price.
Also, long-term deflation makes debts untenable.
Fear of deflation is also a hot topic in Europe now. Since the beginning of 2015 Europe also has shown short periods of
deflation. With the outbreak of the global financial crisis, Europe is contending with low economic growth or even
shrink. That’s why the average inflation rate fell gradually since the end of 2011 (see Figure 1.1). Lower oil prices have
contributed to low inflation. These developments made the deflation scenario in Europe more actual. Deflation is now
not associated with Japan only anymore. The ECB is trying with the help of monetary easing to stimulate the economic
growth and to push inflation. The program means that the ECB buys government bonds, for an amount of EUR 60
billion per month, from banks and financial institutions, to stimulate the economy and push the inflation. The program
has recently been extended to the end of March 2017. The idea is that the low interest rates and a weak euro ensure
more spending and export, stimulating economic growth. Whether this policy will succeed and inflation will actually be
stimulated to 2%, remains to be seen.
Figure 1.1
Development inflation Euro Area (% on an annual basis)
Source: Eurostat, October 2015, Oxford Economics October 2015, LaSalle november 2015
*Excluding energy, food and tobacco
MN, the pension provider for “Pensioenfonds Metaal en Techniek”, abbreviated PMT, and for “Pensioenfonds
Metaalelectro”, abbreviated PME, uses different scenarios with corresponding expected yields of various asset classes.
One of these scenarios is a deflation scenario, also known as "Japanification". In this scenario there is a risk that the
excess capacity in the EMU is even greater than is assumed in the basic scenario. It is also possible that the
overcapacity puts more downward pressure on inflation than expected. In the past, sometimes, the disinflatoire trend
returned only when the overcapacity had completely disappeared. Because then there arise opportunities for business
and workers to increase prices. And because inflation is already very low, and it will take several years before the
overcapacity will be disappeared, the low inflation could turn into deflation. Of course the ECB will do everything they
can to prevent deflation.
8
1.2
The impact of deflation on shares and bonds
In theory, deflation is detrimental for shares. In fact, a period of decreasing prices, reducing wages, declining
employment and weak demand is causing pressure on the sales and profits of enterprises and therefore also the share
prices. Government bonds on the long-term performance are performing usually well at deflation. Bonds are loans. At
the end of the term investors get their investment back. In the event of deflation prices fall and the money is worth
more. Then, over three years, more can be bought than now is the case. Cash is also appealing because the money
increases in value. Often long-term deflation is not good for high-risk investments. This survey deals with the impact of
deflation on real estate.
1.3
The impact of deflation on real estate
There are many studies regarding the topic inflation and real estate. Inflation means that the price rises. It is generally
accepted that real estate provides protection against inflation, by leasing contracts where rents increase with the
annual inflation at least in times of an economic upswing. This assumption is not correct: the results of investigations
whether property provides protection against inflation are not unambiguously. In many surveys it makes sense
whether a long or a short period is considered, and what period is considered. However opinions of scientists are
divided. In addition, it appears that the type of data strongly influence the results. Also, one sector responds differently
on inflation than the other. It is thus hard to be able to say well-advised that real estate offers an inflation hedge.
About the relationship between deflation and real estate less research is done. Previous research (J.P. Hildering, 1999)
has shown that it is not straightforward what the effect of deflation is on real estate returns. This thesis is of added
value, because of the limited amount of published studies with regard to deflation and real estate. There is a tension
between unfavourable macro economic conditions which are having a negative impact on real estate on the one hand.
This brings risks, such as higher vacancy rates, falling rents and a higher chance of defaulting tenants. For these risks
the investor want compensation with a higher risk premium. On the other hand, a lower risk-free rate, which is
plausible in a deflationary environment, provides an environment where required returns are lower. A reduction in the
risk-free rate leads to a larger spread and a higher real performance.
1.4
Hypothesis
This survey deals with deflation and its impact on real estate returns. The question is approached from the pension
investor. This is a reviewing research. On the basis of literature the following hypothesis is distilled: Real estate capital
gains are muted in a deflationary economy. This hypothesis is supported from the literature. This hypothesis is tested
on the basis of empirical research on the Japanese real estate market, given the long period of deflation over there.
1.5
Objective
The objective of this research is not only to obtain knowledge and a better understanding of the impact of deflation on
real estate returns, but also to know how a real estate portfolio is performing during deflation. This research leads to
insights into the impact of deflation on real estate returns.
1.6
Research methods
This study consists of two quantitative analyses; the first is exploratory and the second is reviewing. Both analyses
relate to the Japanese real estate market. In the first quantitative analysis is examined what the effect is of deflation
and vacancy rates on the change in rents of office space in Tokyo. A correlation analysis and multiple regression
analysis show this. Also, in this analysis is examined what the effect is of various time lags of the variable vacancy rate
and of the variable deflation. The lag runs from 3 months to a lag of three years, where in one case the lag applies only
to the variable vacancy rate and in the other case only applies to the variable deflation. Also the effect of a lag that is
the same for both variables is taken into account in this analysis. This relates to exploratory research.
9
In the second quantitative analysis the hypothesis Real estate capital gains are muted in a deflationary economy is
assessed; this is reviewing research. This is done with the help of various regression analyses. It includes the
explanatory variables expected inflation and expected economic growth in Japan and the variable to explain is the
return on capital gain/change in value. In addition to the hypothesis, also is looked at the impact of deflation on the
direct return using a regression analysis. Indeed, the IPD Japanese performance data make a distinction between direct
performance and capital gain/change in value.
1.7
Delimitation and generalizability
In this research inflation is based on CPI, what stands for consumer price index. This consists among other things the
components energy, food, alcohol and tobacco. The current low inflation is caused partly by the sharply lower energy
prices. Core inflation is the rate of inflation which is measured without the categories food and energy. What type of
inflation a pension fund tries to keep up with can be different for each fund; this is dependent on what agreements
have been made. It is important to note that at the current rate of the coverage ratio it is not feasible to keep up with
inflation. This is more extensive described in paragraph 2.5.
The likelihood that the Japan-scenario will be seen in Europe is not included in this thesis. It is in paragraph 2.8 where
the similarities between Japan and Europe are described: Europe faces like Japan low inflation/deflation, the economic
growth disappoints, interest rates break new low records and solving problems is difficult. The QE programme of the
ECB which started in March 2015, will be extended because the inflation rate is still at a low level. The effectiveness of
the European QE is often the subject of discussion. In Japan QE is still going on and the inflation target of 2% is still not
there. Finally, another similarity is that many European countries, although to a lesser extent than Japan, have an aging
population.
1.8 Content
Chapter 2 deals with deflation in general. In paragraph 2.2 definitions are described, not only the term deflation, but
also inflation, hyper-inflation, disinflation and stagflation are defined. Paragraph 2.3 concentrates on the two main
causes of deflation. Section 2.4 describes the consequences of deflation, whilst paragraph 2.5 addresses specifically
the effects of deflation for Dutch pension funds. In paragraph 2.6 the measures against deflation are described. The
following paragraph is about the lost decade in Japan; Japan was facing long periods of deflation. The last paragraph
contains the conclusions from this chapter.
Chapter 3 is about inflation. Paragraph 3.2 is dedicated to various scientific studies which examine the question
whether real estate offers an inflation hedge. This includes literature studies of Van Gool (2013), Geltner (2014) and
Eichholtz (1997). Paragraph 3.3 looks at the impact of the investment horizon regarding this question. Literature
examination of Eichholtz (2000), Hoesli (1997), Matysiak (1996) and Barber and White (1995) are quoted here. The
following section focuses the difference between listed and not listed real estate and their capacities with regard to
protection against inflation. The examinations of Hoesli (1996), Matysiak (1997) and Yobaccio (1995) are included here.
Paragraph 3.5 focuses on investigations on the inflation hedging capacity of various real estate sectors. The literature
research provided for in this paragraph contains those of Miles (1996), Barber and White (1995) and the White (2007).
This chapter is closed with paragraph 3.6, in which the conclusion is described.
Chapter 4 deals with deflation and real estate. More specifically, this section is about the question what the effect is of
deflation on the return that an investor requires from his/her real estate investment. Paragraph 4.2 begins on the
‘built up method’. Paragraph 4.3 focuses on the risk-free approach and paragraph 4.4 looks at the influence of the level
of interest rates. Then in paragraph 4.5 there is more detail on the difference between nominal and real estate returns
at an inflation scenario and deflation scenario. Paragraph 4.6 zooms in on the impact of deflation in the Japanese real
estate market. This section also describes the development of the rents for offices in Tokyo. These data are used in
chapter 5, where the first quantitative analysis is based on. This chapter is closed with the conclusion that contains the
key findings. These findings also lead to the hypothesis: Real estate capital gains are muted in a deflationary economy.
10
Section 5 includes the first quantitative analysis. The question is whether the change of the rental price can be
explained by the change in the price level and/or a change in the vacancy rate. Data are coming from the Japanese
office market. This relates to exploratory research. Paragraph 5.2 describes the plausible links. Paragraph 5.3 is gives
more detail about which data are used. Then in paragraph 5.4 is defined which analyses will be done. By running
regression analyses with different time lags, it becomes clear what the effect of lagging is. This chapter is closed with
paragraph 5.5, containing the main conclusions from the analyses.
In chapter 6 the hypothesis is tested in the second quantitative analysis. This affects reviewing research. This is done
on the basis of return data coming from the Japanese real estate market. The hypothesis is Real estate capital gains
are muted in a deflationary economy. In paragraph 6.2 a description is given of the data used in the quantitative
analysis. Paragraph 6.3 will look at the mutual correlations. Then there are the various regression models. The results
from paragraph 6.3 shall be placed in a wider context in section 6.4 This chapter is closed with paragraph 6.5: the
conclusions from this quantitative analysis.
In chapter 7 I will discuss the main conclusions from each chapter. Finally, paragraph 7.2 relates to a short reflection of
this thesis.
11
2
Deflation
2.1
Introduction
This chapter is about the concept of deflation. In paragraph 2.2 definitions are described. Not only the definition of
deflation is defined, also the terms inflation, hyper-inflation, disinflation and stagflation will be summarized. Paragraph
2.3 concentrates on the two main causes of deflation. Section 2.4 describes the consequences of deflation, whilst
paragraph 2.5 addresses specifically the effects of deflation for Dutch pension funds. In paragraph 2.6 measures
against deflation are described. With regard to the current ECB policy, it is important to emphasize the difference
between asset price inflation and consumer goods/price inflation. This is also part of paragraph 2.6. The following
paragraph is about the lost decade in Japan, as Japan faced deflation for many years. In the last paragraph there are
briefly the conclusions from this chapter.
2.2
What is deflation?
According to Boonstra W. and M. Verduijn (2014) the definition of deflation is "a continued fall in the average price
level, measured on the basis of the development of a price index." Consumer prices are often used. Depending on the
development of the demand for products, the price can be lower or higher. Inflation is the opposite: “a sustained
increase in the average price level, measured on the basis of the development of a price index." Only if there is a
continuing increase in the price index and therefore of the overall price level, one can speak of inflation. In the event of
very high inflation there is hyperinflation. There is not a hard limit when inflation is called hyperinflation. Stagflation is
the case when there is a combination of a lack of economic growth and inflation. Suppose the consumer price index
drops from 155 in December 2013 to 150 in December 2014. In this case there is deflation of 3.33%. Prices are, on
average, 3.33% lower than one year earlier. The real purchasing power is 3.33% higher in 2014.
2.3
Reasons for deflation
Deflation may be caused by supply effects and demand effects (Boonstra W. and M. Verduijn, 2014). In the event of
supply effects there is a sharp decrease in costs. Deflation on the demand side is called ‘cost deflation’. This cost
reduction can for example be a result a decrease of prices of imported products, for example oil. It is also possible that
the cost of products fall by a rapid increase in productivity. There is disinflation, in case the rate of inflation decreases
as a result of these developments. Deflation caused by supply effects is usually1 as result of technological progress. The
business productivity increases. That’s why prices decline and why real incomes increase. The rise in real incomes can
result in more demand. In general this type of deflation is experienced as beneficial to the economy (Hildering, 1999).
A good example of this is the price of a Giga-byte digital storage: this decreased from USD 500,000 to USD 700,000
early eighties to USD 0.03 now. Every few years the price dropped 50% as a result of technological progress. Because
of the fall in prices, other products could be produced more cheaply and more jobs could be created, which stimulated
economic growth (Mauldin, 2014).
Deflation can also result from less demand. This is also referred to as ‘spending deflation’. This can happen when
consumers have less confidence in the future. Then consumers rather save, or decide to amortize their debts or to
increase their reserves. The government can discourage spending of consumers or companies, for example by austerity
measures or by burden. The demand for products then decreases. In order to maintain sales, often the price is
reduced. This seems to be positive for consumers on the short term. This will be detrimental for employers: profit
margins disappear. They will try to be compensated for this by reducing costs. Often this leads to dismissal of workers.
The purchasing power of the employees drops. The rising unemployment creates a downward pressure on wages. This
results in a lower purchasing power in general. Loss of demand can also be the result of demographic developments.
Demographic shrinkage and aging play an important role in deflation in Japan. In paragraph 2.7 there is more detail.
1
A strong decrease on a broad front of commodity prices can also lead to deflation through the supply side and has nothing to do with
technological progress.
12
2.4
General effects deflation
Before addressing the consequences of deflation it is important to elaborate on the distinction between nominal and
real interest rates. The nominal interest rate corresponds to the rate of remuneration that a depositor must be paid or
received by the person who is borrowing. The real interest rate is the nominal interest rate which is corrected for
inflation. Paragraph 4.5 shows two examples with an investment property with this distinction. Nominal interest rates,
real interest rates and inflation are linked to each other. The relationship is as follows:
RN = RR + i
RN = nominal interest
RR= real interest
I = inflation
Deflation is detrimental to have debt, because the real interest rate increases, the debt value gets higher. With deflation,
debts are repaid by an amount which in real terms has become worth more than at the time of the commitment of the
debt. That is not only detrimental to people who have debts, but also for the government debt. The net capital power
position is deteriorating. The providers of long-term loans see the value, as well as their real return on their capital,
increase.
Another consequence of deflation is that it is more attractive to save, because real money is worth more. This leads to
less expenditure. On the other hand borrowing is discouraged. The real interest rate is the nominal interest rate
corrected for inflation. In the event of deflation the real interest rates are higher than the nominal interest rate. One
consequence of this, is that companies are less likely to borrow money for investments. This is not favourable for
economic growth. Due to deflation, the real value of money increases; with the same amount of money one can buy
more later. Often consumers delay their purchases, because of the expectation that products will be cheaper. It usually
consists of non-urgent products, such as certain durable consumer goods (white goods, cars or an apartment). A
negative spiral can occur: less is sold, resulting in less production, making that fewer employees are needed. The
unemployment rate is increasing and consumer spending will continue to fall.
Another consequence is that income must be redistributed. In the event of deflation employees with a weaker job
position will often be affected first by dismissal, with a decrease in the purchasing power as a result. Employees with a
strong position in the job market can take advantage of the cheaper products (Boonstra W. and M. Verduijn, 2014). It
is difficult to interpret price signals from the market by inflation or deflation. The functioning of the price mechanism is
disturbed. If for example the price of a given product in a certain period drops by 3%, than the producer may conclude
that there is little demand for his product and he needs to adapt his production downwards. If in the same period
there was a decrease in the average price level of 6%, then he has made the wrong conclusion (Boonstra W. and M.
Verduijn, 2014).
Also for banks there can be a negative impact due to deflation. In the event of deflation and declining interest rates on
the capital market, the interest charges will fall. The interest rate on the assets (long-term mortgages) are rated
nominally and come under pressure. The interest margin of the bank will increase in the short term. First the effects of
deflation may be particularly positive, with a rising real purchasing power and rising interest margins. However, after
expiry problems for payment can arise by the downward pressure on wages (by the wider labour market or job loss). In
the event of deflation of house prices, the value of the collateral can fall below the mortgage loan. The bank is
therefore exposed to a greater credit risk. This is also called 'debt deflation', that goes together with bank crises
(Fisher, 1933). In the most disastrous scenario the economy is in a downward spiral, in which there is persistent
economic shrinkage and there is a recession.
13
2.5
Impact deflation for pension funds
For pension funds, deflation is particularly serious for the height of the coverage rate, if the nominal interest rate
becomes (strongly) negative by the deflation. In fact, a declining nominal interest ensures that the value of the
liabilities rise dramatically. Pension funds are technically bankrupt. They will have a coverage ratio of approximately
90% and below and are being forced to reduce the pension payments. This can further strengthen deflation, because
retirees have less money to spend. At the macroeconomic level this means even less demand for goods and services,
so that the prices may fall (further).
More specifically, long-term deflation may make it more difficult to make fulfil the promises of a pension system based
on "defined benefit", and "defined contribution" will possibly be more widespread. In the current pension system
pension payments are often based on 'defined benefit’. That means that someone builds an amount of pension money
per year, what shall be paid each year, when the participant has reached the retirement age. Pension funds have the
ambition to increase the payment with inflation, to maintain the purchasing power of the pension payment, but this is
not an obligation. With regard to what kind of inflation pension funds try to keep up with, this can be different
measure, like wage inflation or consumer price inflation, but pension funds can also prepare an inflation measure their
selves. In the event of ‘defined benefit’ the final salary scheme and the medium salary scheme are the most well
known. In the event of a final salary scheme the pension is based on the last earned salary, before the retirement age
is reached. This system offers the best possible protection against inflation. This scheme is removed from many/all
pension funds in the period 2002 to 2005. The medium salary scheme bases the pension on the average salary. Per
year one can build up a pension amount, based on the salary of that year. All of these pension amounts together form
the pension. The counterpart of “defined benefit” is “defined contribution”: the annual pension payment will not be
fixed in advance. The employer and employee close a pension agreement, in which the pension premium is to be fixed.
The height of the final pension depends on the premiums, the cost and the returns. In here, no promise is made for
pension payments, this is in contrast to “defined benefit”.
The ambition of the pensions to keep up with inflation is a difficult task in a low interest rate environment, as is
currently the case. The Nederlandsche Bank proposes strict rules in respect of the degree of the coverage ratio. The
coverage ratio must be at least 110%, to be able to increase pension payments with inflation partly. Full indexation is
only possible on a coverage ratio of 125% to 130%. Such rules have been drawn to protect the young participant. In the
event of a declining interest rates, the returns may be higher, because the value of fixed income increase due to a
lower interest rate, but the present value of the future pension payments increase much more. In the last few years
the liabilities increased in size significantly. The declining interest here is the main cause, but the increase in life
expectancy also plays a role.
In a deflation scenario it is the question whether the nominal interest rate either is around the 0% or negative. If this is
at around 0%, then the purchasing power of the pension fund increases 'obviously', because the general price level
drops. From a technical point of view, a pension fund could achieve negative indexation, so that the purchasing power
of the pension payment maintains constant, but that is highly unlikely. If the nominal interest rate is (strong) negative,
this is disastrous for the coverage ratio; liabilities will rise very high, at (strong) negative interest rates the present
value of future cash flows is greater than the total size of the cash flows themselves. Not any investment can
compensate for this.
If it is not “defined benefit”, but "defined contribution" becoming more popular, then the pension fund management
does not need to take the degree of the coverage ratio into account, because "defined contribution"-schemes don’t
have coverage ratios, but all shocks are absorbed directly in the pension power of the participants. However, the effect
on the pension income of participants is perhaps stronger in a "defined contributions' scheme than the effect would be
in "defined benefit schemes. In fact, in a "defined contribution" scheme is a large piggy bank, being used for
purchasing an annual pension payment, when reaching of the pensionable age. The amount of the annual payment
which can be obtained, depends on the amount of money that has been built up in the piggy bank. So poor returns
during the active life lead to a lower pension. Also the position of the interest rate at the time of the annual payment is
14
purchased is of importance: if the net interest rate is very low, then the payment is also relatively low. In the event of a
high rate of interest, the payment is also relatively high. In a "defined benefit” scheme the effects of poor returns or
low interest rate are less profound, because the fund does not need to reduce claims directly, but deficits may spread
over a period of ten years to eliminate them. As a result, the effect is less one-on-one in the pension payments. In the
event of a "defined contribution" scheme pension funds do no commitments and will therefore not come back on
promises that were previously made. It is quite conceivable that, if the interest rate remains low for a longer period,
"defined contribution" will be the norm.
2.6
Measures against deflation
It is stated that the fight against inflation with tight monetary policy is always effective, but that fight deflation with
monetary policy is more difficult. In the event of a fall in prices interest rates can be reduced, but normally this drops
not below zero. However, this did happen in Japan. In the event of further falling prices, the increase the real interest
costs go on, which is negative for economic growth. Central Banks may extend quantitative easing (QE), but this is no
guarantee for higher spending or for economic growth. More public spending could prevent a negative spiral, but this
also brings a rising public debt. Finally, a government spending financed monetary impulse can be given: a growth of
the money supply which directly results in more expenditure. This policy brings great risks for price stability (Boonstra
W. and M. Verduijn, 2014).
In the context of QE it is important to consider the difference between asset price inflation and consumer price
inflation. Asset price inflation means the nominal increase in the value of shares, bonds, real estate, derivatives and
other asset classes. Consumer goods are not included here. Inflation is usually measured without taking account of
asset price inflation. In here, the focus is purely at consumer goods, to consumer goods/price inflation, such as the CPI
(Consumer Price Index). Economic growth is usually measured by the change in the gross domestic product, corrected
for inflation, which refers to the CPI and not on asset price inflation.
Monetary or quantitative easing has the goal to increase the CPI. The result of the QE however is often asset price
inflation. Nevertheless, central banks consider asset price inflation also as favourable as the so-called wealth effect
(people have net more power as a result of increased stock quotes, if the profit is taken) can ensure that more is spent,
stimulating economic growth. The danger is that it can take a long time before effects are appreciable and the
economy as a result of that grows. Asset price inflation can be a misleading signal of economic growth. In this way the
stock exchange and real estate prices rise although, yet this does not lead directly to real economic growth. It should
also be noted that the confidence of consumers often increases, when stock prices and house prices rise. Consumers
will then be tempted to consume, stimulating economic growth. In an indirect way this contributes to economic
growth.
A risk in the event of such an asset price inflation is that prices will continue to rise, without justifying this high prices
with regard to the underlying market. When the prices are shooting too far up, a correction may lead to a lasting
decline in prices, causing asset price deflation. Asset price deflation goes usually together with a recession. Sometimes
asset price deflation is even the cause of a recession. Also in Japan recession and asset price deflation are associated
with each other. When the central bank cuts interest rate, than borrowing becomes attractive. Thus, in theory the
demand is stimulated. In Japan this did not work. There was no demand at all, despite an interest level of almost 0%.
The high level of debt, withhold people to further increase their indebtedness. Paragraph 2.7 deals extensively with
deflation in Japan.
2.7
Historic development deflation in Japan
Japan faced a long period of deflation. In the period 1985 to 1990, the Japanese economy grew by an average rate of
5% per year, mainly driven by investments. Sharply rising share prices and land prices have resulted in immense credit
provision by banks. The availability of collateral was enormous. In the 1980s land prices rose by more than 400% and
stock quotes with more than 600% (Haar, van der E. 2009). After the burst of the financial bubble in the early nineties,
15
the necessary reforms were retained by star policy, resulting in a long period of stagnation: declining interest rates,
inflation and economic growth, as well as a decrease in profitability to very low levels, appreciation of currencies, a
strong rise in public debt and financial repression. After the financial impact the economy grew even with 1% and in
1995 there was deflation. This period is also called the lost decade: a long ongoing combination of falling prices, low
growth and low interest rates. Since 2000, the Japanese economy is recovered a little bit. Nevertheless the economy
balances still on the edge of deflation. At the end of 2012 a new policy, "Abenomics", was announced, in order to
ensure stimulate economic growth, stop deflation and achieve restructuring. Later in this section there is more detail.
In Japan aging and demographic shrinkage have a large role in deflation. Currently, 23% of the Japanese population is
aged over 65. It is expected that this will rise to 33% in 2025 and to 40% in 2060 (Op Den Brouw, 2014). Since 1950 the
number of new babies dropped in Japan. Currently this compensates for only 35% of the number of deaths. In
addition, the average life expectancy increases. The peak of the Japanese population was reached in 2007 with 128
million people. After this year the population decreased each year with a million people. It is expected that this trend
will continue. The population decreases from 127,77 million people in 2005 to 89,93 million in 2055. This represents a
decrease of around 30%. See Figure 2.1.
The decline in the working population is even at a faster rate: from 8,41 million in 2005 to 4,6 million in 2055. This
represents a decrease of 45%. In 2055 there is 1,3 working on one senior. This is a sharp decrease; in 1960, this still
11.2 workers on one senior. This decreased sharply, to 3.3 workers already on one senior in 2005 and will therefore
continue to decrease. A higher retirement age is not enough to beat against this shrinking labour market. The shrinking
labour market provides a shrinking ability to generate revenue; the gross national product drops and there is a further
slowdown in productivity growth per head of the population. The shrinking labour market and the social and economic
burden of aging impede economic growth. Aging is particular a problem in urban areas. The proportion of elderly
people in large cities such as Tokyo is one to three times greater than the Japanese average (Op Den Brouw, 2014).
Figure 2.1 Demographic shifts in the Japanese population
Source: Op Den Brouw, 2014
It is not only the demographic developments that play a role in deflation in Japan. It is stated that in two respects
errors have been made in the fiscal and monetary policy in Japan. In the first place the policy was very reticent. The
financial bubble burst in the early nineties, but only in 2001 the fight against deflation was an explicit objective of the
Bank of Japan, resulting in a QE program. Secondly, the importance and the urgency of debt reorganization and
restructuring of business was underestimated. Despite the fact that many loans had become worthless, banks continue
to lend to inefficient enterprises. Public and business had great reluctance to companies to go bankrupt. Companies
were experiencing high debts and banks had a lot of unprofitable loans outstanding. Ten years after the crash, the
16
total volume of the problems is not clear, because of the closed nature of the banking sector (Haar van der van E.,
2009). According to Koo (2008) it is the public policy that ensured that the Japanese economy did not fully collapse.
The largest economic shrinkage in Japan amounted to 3% and took place in 1998, the year in which there was a deep
recession (Heltner H.A., 2001). The government implemented such tax incentive measures that the balance of the
private sector could recover again (Koo R., 2008).
The business wanted to have debts as little as possible and focused on maximizing profit. When debts are paid off and
households and businesses save more in order to improve the financial position or to improve credit ratings, then
there is a "balance sheet recession". Monetary policy will lose its effectiveness, because companies or households with
debts are not interested to borrow more, despite a very low interest rate. This is exactly what happened in Japan. In
1995, the Bank of Japan dropped the interest rate to almost zero, while the interest rate in 1991 was more than 8%.
The desired impact was not there: there was absolutely no reaction in the economy or asset prices, which decreased
only further (Koo R., 2008).
Some economists say that Japan struggles with a "liquidity trap". With short term interest rates close to zero percent,
monetary action (including the purchase of long-term government securities, shares or foreign currency) is not an
effective way to influence price levels or to push economic growth (Heltner H.A., 2001). However, the concept of
"balance sheet recession" was not known yet in 1990; as well as the fact that the policy would be ineffective. The
result was that at any time when the economy showed first signs of recovery by fiscal incentives, that this was
considered to be sufficient for further economic recovery. The incentive measures were therefore stopped, to reduce
the budget deficit. But no structural recovery is possible in a "balance sheet recession" without the actually recovery of
the balances in the private sector. The early stop of the tax incentives resulted in an economic shrinkage, which
resulted in new tax incentive measures, which directly stopped at the first signs of economic recovery. This "stop and
go" policy continued years (Koo R., 2008).
In the public sector the tax incentives had an opposite effect. The average propensity to consume decreases as the
debt as a percentage of GDP increases. According Heltner (2001) the savings behaviour of Japanese have a direct link
with aging. In fact, the aged population has earned relatively much for its own pension, invested in real estate and
shares. This capital was hit by the considerable decrease of share prices and real estate prices. While the government
promised to keep enough cash for pension and health care, there was a rising public debt. Because of fear that the
public would partly fail in this promise, the Japanese saved more. So, the fiscal incentives were not effective. The more
there was saved, the less the impact was of the tax incentives. The reaction of the government of this was increasing
the fiscal incentives, and thus increasing the debt, and as a result the Japanese saved more. All in all, the additional
public spending contributed little or nothing to the potential growth. Not only one saved more, also expenditures were
postponed. Deflation caused the expectation that products could be purchased cheaper, increasing deflation further.
In particular, the purchase of durable consumer goods were postponed (Hori and Shimizutani, 2005).
Heltner (2001) suggests that the exchange rate policy is the principal reason for deflation in Japan. In the nineties the
cost of production increased in Japan, in relation to the cost of production in the USA and depreciation of the nominal
yen/dollar was not allowed by the Ministry of Finance. The objective of the currency policy was deflation in Japan.
Prices needed to be reduced enough to reduce real wages. Due to deflation the Japanese households saved more and
companies increased their balance sheet. Later in this section the term 'currency war' is briefly discussed.
As said, the battle against deflation became an explicit objective of Bank of Japan in 2001 and a QE program was
started. Although this resulted in some recovery of the economy, still Japan is, almost fifteen years later, having
troubles with the 'deflation ghost'. At the end of 2012, a new policy was announced, "Abenomics", in order to
stimulate economic growth, beat deflation and achieve restructurings. The Abenomics policy rests on three pillars,
namely 1. An immense monetary impulse, 2. Public investments and 3. Structural reforms. The main objective of the
Bank of Japan is the fight against deflation. With a large-scale purchase program of assets as bond yields, corporate
17
bonds and shares (the first pillar of Abenomics), the Bank of Japan tries to get inflation to a rate of 2%, to suppress the
rate of the yen and to stimulate the export. This should ensure that the economy is improving, that the confidence of
consumers will rise, in order to increase expenditure. In 2013 and 2014 large buy-back programs were launched,
respectively EUR 360 and 570 billion. In the first instance, some effects seemed visible, until a new tax policy resulted
in the fact that consumers had to pay more tax to cover the deficit. This put a brake on spending and on economic
growth. Japan moved back in the direction of deflation. The second buy-back program did turn the tide, while it should
be noted that the inflation target of 2% is still not there. Rather one could speak of asset price inflation than of
consumer price inflation. The second pillar focuses on public investment. Japan expands this and spends more on
pensions, babysitting, wages and other facilities. The third pillar focuses on the reform of the tax system and the
labour market and the corporate governance in Japanese companies. Also immigration and the number of working
women must increase drastically to help the shrinking labour market.
Despite some positive effects, there remain uncertainties whether Abenomics will succeed structurally. The inflation
target of 2% is still not visible. Initially, the demographic shrinkage remains an unresolved and structural problem. Also
there is still no convincing structural recovery of the Japanese economy. In addition, a detrimental effect of Abenomics
is a growing public debt. The government debt has increased to 230% of the gross domestic product. This is the highest
public debt in the world. This brings high costs. An increase in interest rates would be devastating for the government.
The central bank is therefore obliged to continue to buy bonds, in order to keep the interest rates low and the yen
cheap. At the same time the debt becomes even bigger. The longer the Bank of Japan continues with QE, the more
difficult it will be to stop. If the Bank of Japan would stop with QE, interest rates will increase. But because the price of
bonds are determined for the most part by the central bank, it is difficult to predict how quickly the interest rate will
increase if the bank will purchase less bonds. In addition, QE results in a weak currency which is favourable for the
international competitive position. A strong increase of the currency is not favourable for economic growth.
With this regard, it is relevant to consider the term 'currency war' briefly: a costly currency harms the economy. If one
country devaluates its currency, the other (neighbouring) countries or trading partners devalue their own currencies as
a reaction, to make their position competitive and to boost their economy via export. This is also called the “beggarthy-neighbor" principle. This means that devaluation of a single currency is at the expense of another. Each country is
trying to attract more demand, due to a strengthened competitive position (by a weak currency as a result of QE) at
the expense of other countries. A risk is a 'race to the bottom', because there is devaluation in response to previous
devaluation(s) successively. Currently the US (although the Fed recently announced for the first time in ten years time
the first interest hike) and Europe maintain expansionary monetary policy. This is the reason why QE is controversial.
2.8
Japan scenario in Europe?
The likelihood that a Japan scenario will arise in Europe is disregarded in this thesis. It can be noted that there are a
number of important similarities between Japan and Europe: Europe suffers from a very low inflation and some
countries suffer even with deflation, the economic growth disappoints, interest rates break new records troughs and
solving problems is difficult. In both Japan and Europe QE is not effective so far, as the inflation rate is not even close
to the objective of 2%. Both regions show significant asset price inflation. Finally, many European countries, although
to a lesser extent than Japan, are facing an aging population.
QE in Japan has not proved yet to be effective. The incentives in the public sector had the opposite effect. The reaction
of the government was to increase the fiscal incentives, thereby increasing the debt, and as a result even more savings.
In Europe this is not the case. However, in both Europe and Japan, one could discuss the effectiveness of QE. The
question is whether QE ultimately will ensure the inflation at a level of 2%.
An important difference is considered to be in the cultural sphere. With regard of openness Japan and Europe are less
similar to each other. Japan has a long history of secrecy. European countries don’t close their borders for migration
and most of them have an open economy and culture. The result is that our society adapts faster to changing
18
circumstances and is therefore less vulnerable to shocks. Another difference is that the Japanese debt is financed
internally. The system of savings and pensions is a closed system in order to finance the debt of Japan. In Europe this is
not the case and many countries have their debt financed by other countries. Another difference concerns the level of
public debt. In Japan this has now risen to 230% of GDP. In the Euro zone this is with 91% significantly lower.
Figure 2.2 Does QE lose effectiveness?
2.9
Figure 2.3 Inflation developed economies
Conclusion
In this chapter different aspects of deflation are described. Deflation is defined as "a continued fall in the average price
level, measured on the basis of the development of a price index". Deflation may be caused by supply effects or by
demand effects. The effects of deflation are described in detail. Also, the impact of deflation on pension funds became
clear: deflation is dramatic for the level of the coverage ratio, if the nominal interest rate is (strong) negatively due to
the deflation. One of the measures against deflation is monetary easing. In this context, the distinction between asset
price inflation and consumer price inflation is emphasized. QE has the goal to increase consumer price inflation.
However, the result is often purely asset price inflation. This brings risks.
That deflation has negative long-term consequences for the economy, became clear in the last paragraph in which the
lost decade in Japan is described. The demographic developments in Japan have negative impact on demand. Fiscal
and monetary policies were not efficient, allowing Japan faced with deflation for years and turned in a negative spiral.
Not only the Japanese saved more, also expenditures were postponed. Incentive measures in the public sector were
even counterproductive. In addition, some believe that the main cause for deflation in Japan has been the currency
policy: depreciation of the nominal yen/dollar was not allowed by the Ministry of Finance in the 1990s, when the cost
of production in Japan increased in relation to the cost of production in the USA. Prices needed to fall enough to
reduce real wages. The Abenomics policy aims to stir inflation, but still the inflation target of 2% is not there. The effect
of quantitative easing is questionable in general, not only because the inflation target of 2% is still not there (both in
Japan and in Europe), but also because markets devalue their currencies, due to QE, at the expense of other markets,
which in their turn use QE to improve their competitive position. A risk is a “race to the bottom".
In this chapter also the similarities between Europe and Japan are considered: Europe suffers from a very low inflation
and some countries suffer even with deflation, the economic growth disappoints, interest rates break new records
troughs and solving problems is difficult. In both Japan and Europe QE doesn’t succeed so far in pushing inflation
toward the objective of 2%. Both regions showed significant asset price inflation. Finally, many European countries,
although to a lesser extent than Japan, have an aging population.
In what way deflation impact on real estate will be described in Chapter 4. The following section deals with the
question whether real estate provides protection against inflation.
19
3
Inflation and real estate
3.1
Introduction
There are many studies known about inflation and real estate. Often is assumed that property provides protection
against inflation. This chapter shows that this argument is controversial. In paragraph 3.2 research is shown, that
comment on the question whether real estate offers actually an inflation hedge, with examination of Van Gool (2013),
Geltner (2014) and Eichholtz (1997). Paragraph 3.3 looks at the impact of the investment horizon in this issue, with
literature of Eichholtz (2000), Hoesli (1997), Matysiak (1996) and Barber and White (1995). The following section
focuses on the real estate capacities of inflation hedging regarding the difference between listed and non listed real
estate. Here literature of Hoesli (1996), Matysiak (1997) and Yobaccio (1995) is quoted. Paragraph 3.5 focuses on
investigations on the inflation covering capacity concerning various real estate sectors, with literature of Miles (1996),
Barber and White (1995) and De Wit (2007). In paragraph 3.6 finally the conclusion is described.
3.2
Real estate as inflation hedge?
One can speak of an inflation hedge when an investment offers protection against inflation. Assets offering returns
with a positive correlation inflation can be seen as inflation hedges (Bruine, 2009). The association that real estate
offers an inflation hedge is mainly caused by the indexation of the rents of many lease contracts, the increase in the
construction costs, often increasing more than the rate of inflation and the continuing increase in the value of land.
According to Van Gool et al. (2013) returns of direct real estate usually correlate with inflation only to a limited extent.
The relationship will become stronger if the return is measured over a long period. Where the correlation between
inflation and returns is limited, cash flows from real estate do correlate highly with inflation. If rental prices are linked
to the inflation index, rents will grow with inflation. Linked rents are also beneficial for capital gain/change in value. A
higher the rental income, as a result of inflation linked rents, will be reflected in a higher value of the property. In fact,
rental income determinates for a large part the value of the property.
However, if a lease is agreed for a long period and no agreements have been made on the annual indexing of the rent,
than real estate offers, in terms of rental income, no protection against inflation. Nevertheless, inflation still impacts
the rent level with a time lag. When the rental contract expires, a new price needs to be negotiated, which will be in
line with the market price. Contracts may also be constructed in a way that the owner of the property is protected
against an increase in the operating expenses. In addition building costs of real estate will increase when inflation is
rising.
In order to know the impact of inflation on real estate, it is of great importance whether the economy is bullish or
bearish. In the event of a booming economy rents should at least grow in line with inflation. A lot of demand for
property provides a strong negotiating position for the owner. In fact, the tenant has little alternative possibilities. A
downside with this regard is that structurally high inflation brings significant increases of rents. In the event of poor
economic growth, it may be difficult for tenants to pay such high rents. This could ultimately lead to a rise in vacancy.
In the event of a downturn, the situation can arise that contract rents are higher than market rents. Lessees have a
strong position when they have a lot of choice where to rent space. If the lease is renewed then the contract rent has
to be adjusted to market rent. This phenomenon is known as saw tooth effect (‘zaagtandeffect’). In this situation real
estate provides little protection against inflation. Break options in leases or short lease contracts will ensure that the
rents adjust to the market rents sooner.
Also the return that an investor requires is important. The required return is made up of the risk-free rate, plus a
premium for the investment risk minus the expected growth of the rental income. With regard to the growth of the
rental income often the expected inflation is used. In addition, reference is made to a self fulfilling prophecy: because
of the association that real estate offers a good inflation hedge, it is possible that there is more demand for real estate
is in a period of high inflation. This will ensure higher prices (Van Gool et al. 2013).
20
Geltner e.a. (2014) points out that in the 1970s and early 1980s there were many concerns about high inflation. Real
estate investments, a capital-intensive asset class, with a high leverage was considered as an attractive asset class,
because borrowing is attractive with high inflation: “get rich quick”. After all, the money that the investor lends, is less
worth at the time that the debt must be repaid. This is compensated by the interest to be paid on the borrowed
amount. The interest rate would have been sufficient to compensate for the level of the expected inflation including
some return. If a fixed interest rate is agreed, and the actual inflation is higher than expected, than the borrower is
better off. This will, in general, have a positive impact on the value of the property. According to Geltner et al. (2013)
rent prices and values will grow with inflation. When the rate of inflation is higher than expected, then rents and also
values will increase more than expected, measured in nominal values. In that respect it is a hedge against inflation.
Nevertheless he points out that real estate is a risky asset class, even measured in real values, i.e. corrected for
inflation. Real estate is not a perfect protection against inflation. An unexpected change of inflation is usually
accompanied by an event in the macro economy or in monetary policy. Often such events have an impact on the real
value of real estate.
The ‘Herengracht index’ (Eichholtz, 1997) shows a historic series of indexed house prices at the Herengracht in
Amsterdam for the period 1628 to 1973. This index shows that the value of the house can be preserved and that the
real value has not or hardly risen. This index is based on transaction prices. The index shows real and nominal values.
The index starts in 1628-1629 with a value of 100.
Figure 3.1 The Herengracht Index 1628-1973
Herengracht index
2600,0
2100,0
1600,0
1100,0
600,0
1628
1638
1648
1658
1668
1678
1688
1698
1708
1718
1728
1738
1748
1758
1768
1778
1788
1798
1808
1818
1828
1838
1848
1858
1868
1878
1888
1898
1908
1918
1928
1938
1948
1958
1968
100,0
nominal index
real index
Source: Eichholtz, 1997
The index varies from 100 to 700 until the 1950s. Until the Second World War there is hardly any substantial increase
in value. After the Second World War there is a structural price increase. This increase is mainly caused by inflation.
The nominal value between 1628 and 1973 is more than a twentyfold increase. The index has also experienced periods
with reductions in value, like at the end of the 18th and the start of the 19th century. Inflation had little impact on the
value. It is noteworthy that the increase in the value disappears completely, if the value is measured in real terms.
Corrected for inflation prices move roughly within the same bandwidth. The real increase is limited; in 1972-1973 the
index measures 218,7. However, if 1632-1633 was selected as starting year, then there would be no real increase in
the index. It can be concluded that the value has been preserved in the long term and that the fair value has not or
hardly risen (Eichholtz, 1997). This shows that the assumption that a house is always increasing in value (measured in
real terms) on the long term is not true.
3.3
Investment horizon
What horizon is taken into account is of great importance. The most pension investors are looking for high correlations
on the shorter term. This plays a major role in the age of the participants of the pension fund. An 'old' pension fund
21
usually has a horizon of 0 to 10 years, because an old pension fund must pay obligations at a relatively short notice. A
young pension fund has a horizon of approximately 20 years. Examination of Eichholtz ea (2000) compares various
lengths of investment horizons with each other in his research about the inflation hedge capacity of residential. The
data used in this study are coming from a residential index. This is an index of total returns which is based on sales
figures from the NVM and a rental index from the Ministry of Housing, Spatial Planning and Environment (VROM).
Firstly, the inflation hedging capacity of residential properties is compared with that of shares and bonds for the period
1965 to 1999. He compares correlations between inflation and the average performance of this asset classes. The
asset class residential performs best with a correlation coefficient of 0,20. This analysis shows even a negative
correlation coefficient (-0,15) for shares. Bonds hardly show a connection with inflation, with a coefficient of 0.06.
Residential can be considered to be the least bad protection against inflation risks, in the short term. Eichholtz also
looks to the longer term; a horizon of one year is too short for a pension fund, with regard to the liability structure.
Due to data availability for a time period of 35 years, he looked to overlapping periods of one, five, ten and fifteen
years. For these periods he examined how large the chance was on a negative real performance. The results are shown
in table 3.1. This analysis showed that the risk of a negative real return for residential is the smallest (14.7%) with a
horizon of a year. The risk of a negative real return for all asset classes decreases as the horizon is longer. This chance
becomes smaller at fastest for residential and shares and the slowest for bonds. In the case of bonds the chance of a
negative real return is still present at a period of 15 years. In the event of a horizon of ten years, the asset class
residential did not show a negative real performance. In the event of a period of 15 years, the same applies also to
shares.
Table 3.1 Inflation protection and investment horizon
1
5
10
15
Residential
32.4
41.2
14.7
13.3
30.0
13.5
4
24
0
0
10
0
Average negative real return (%)
Shares
Bonds
-12
-3.9
-8
-4
-1.4
-5
-1.4
-1
0
0
-1
0
20.7
8.3
9.3
10
4.1
6.3
7.1
3.2
3.5
5.5
2.6
1.8
34
30
25
20
Horizon (years)
Risk of negative real return (%)
Shares
Bonds
Residential
Standard deviation of all real yields (%)
Shares
Bonds
Residential
Observations
Source: Eichholtz ea (2000)
Then, he shows the average negative real return per asset class is in the years that this occurs. In the event of a
horizon of a year this is most negative for shares (-12%) and the least negative for bonds (-3,9%). The returns improve
as the horizon is longer. On an annual basis residential offers a better protection against inflation than shares and
bonds. In addition the chance of a positive real performance increases with a longer horizon. In the last part he shows
how the horizon influences the standard deviations of real returns. The standard deviations decrease as the horizon is
longer, but not equally quick for each asset class. Residential at a horizon of 15 years shows the lowest standard
deviation; this is even lower than that of bonds. It can be concluded that the asset class residential offers a reasonable
protection against inflation, especially in the long term.
This conclusion is supported by several studies showing that direct real estate offers an inflation hedge on the longterm (like Barkham ea, 1996, Hoesli ea, 1997, Matysiak ea, 1996). The study of Hoesli ea (1997) considers inflation
hedging characteristics of real estate in the United Kingdom in the short term, in relation to other asset classes. The
22
real estate data he used in this study are coming from valuation data from IPD. It is evident that real estate has less
good inflation hedging characteristics in respect of total return and return on capital gain/change in value and with
regard to the change in income compared with shares, but performs better in this area than bonds. Also, it is noted
that the ratios can change with another economic climate. The results can thus change over time. Matysiak ea (1996)
examines the inflation covering capacity of commercial real estate in the United Kingdom in the long term. Real estate
data in this investigation originate from IPD for direct returns, shares ranges are derived from Datastream. The
forecasting period is 1964-1993. This shows that direct commercial real estate has no inflation hedging capacity on the
short term. However, in the long term there is a positive correlation between real estate returns and expected and
unexpected inflation.
However, the examination of Barber and White (1995) shows that real estate offers not a good protection against
inflation, regardless whether a long or short term is considered. In this research they use data of the IPD Long Term All
Property Index. Firstly, the average inflation rate in the period 1971- 1995 is with 8.5% higher than the average real
commercial real estate returns (3%) in the United Kingdom. The average real return on residential real estate was
almost 6% per year. So, over a period of 25 years, the real return on both the commercial real estate and residential
real estate were not able to exceed the inflation rate. Secondly, in the investigation of Barber and White evidence was
found via a vector regression analysis. This analyzes short-term and long-term effects of commercial real estate returns
on inflation shocks. This shows that even after three years less than 40% of the real value that is lost through inflation
is restored. Thirdly, for the period 1985 to 1995 the reaction of unexpected inflation on real estate returns is examined
on a monthly basis. It is evident that real estate returns on the short term do not respond to unexpected inflation.
3.4
Listed versus non listed real estate
There are also several studies regarding the question whether real estate offers an inflation hedge, with the focus on
the difference between listed and not listed real estate. Examination of Matysiak ea (1996) shows listed property
shares do not appear to be as effective as a long-term inflation hedge compared with direct real estate. Real estate
data in this investigation originate from IPD for direct returns, shares ranges are derived from Datastream. The
investigated period is 1964-1993. Despite the fact that the impact of leverage on the returns is taken into account, the
research found little supporting evidence for inflation protection characteristics in listed property shares.
According Hoesli ea (1997) there is only limited evidence that supports the argument that property provides protection
against inflation, or that real estate is a better protection against inflation than shares. The type of data series is a
determining factor for the outcome. If one uses performance series based on values, then the general conclusion is
that property provides protection against inflation (Heart Zell, Hekman and Miles, 1987 for the United States and
Limmack and Ward, 1988 for the United Kingdom). If unlisted performance series are used, the opposite conclusion
can be seen: real estate has a negative correlation with inflation (Park, Millineaux and Chew, 1990, for the United
States and Liu, Heart Zell and Hoesli 1997 for Australia, France, Japan, South Africa, Switzerland, the United Kingdom
and the United States). From these studies no unambiguous conclusion can be drawn whether real estate has or has
not inflation protective capacities. The studies which use series on the basis of values might suffer from the fact that it
can be the case that valuers correct the value with a certain inflation factor. If this is the case, it is not strange that
there have been positive coefficients in a regression analysis with regard to inflation. Listed real estate series have a
strong correlation with shares, which in many countries show a negative relationship with inflation.
According to research by Yobaccio ea (1995) REITs offer a moderate protection against any form of inflation (current
inflation, expected inflation and unexpected inflation). Real estate data concerning REIT returns originate from the
NAREIT share price index. In particular, the protection against unexpected inflation is weak. It has to be noted that this
is an old exam. The characteristics of listed real estate companies in the US have changed a lot in the past twenty
years.2
2
Better quality real estate, internal instead of external management, less leverage, less development and more industry focus.
23
3.5
Sectors
Many studies focus on the comparison between a certain real estate sector and other asset classes such as shares and
bonds. It is also relevant to make the distinction between different real estate sectors. The one sector responds
differently to inflation than the other. Miles (1996) examined how a change in the rate of inflation has an impact on
returns of commercial real estate. He used the IPD index for commercial real estate. There is no strong relationship
between the IPD series of commercial real estate and inflation for the period 1971-1995. The correlation over this
period is -0.05. There is no direct relationship between nominal returns of commercial real estate and inflation. In line
with the earlier research cited from Barber and White (1995), he showed that both returns on residential and
commercial real estate were not able to exceed the inflation rate over a period of 25 years.
Examination of De Wit (2007) assesses the inflation hedge capacity of direct and indirect international real estate. The
results show that in most countries there is a positive correlation between the office properties and the expected and
unexpected inflation. Economic growth and a higher level of inflation both have a positive effect on the total return
(Bruine, 2009).
3.6
Conclusion
It can be concluded that the results of the examination whether property provides protection against inflation is not
unambiguous. In many surveys it depends on whether a long or a short period is considered, and also what period is
considered, but also the opinions of scientists are divided. In addition, it appears that the type of data is strongly
influencing the results. Also one sector responds differently to inflation than the other. Fully informed, it is thus
difficult to say that real estate offers an inflation hedge.
Currently, pension funds are more concerned about the level of the coverage ratio than on keeping up with inflation.
As defined in paragraph 2.5, the coverage ratio needs a certain level to allow indexation. The current low interest rate
press the coverage ratio down, making that investments are not able to keep track with the liabilities. Keeping up with
inflation is an ambition, but at this moment it is the level of the coverage ratio being the main concern.
Now that it has become apparent that the inflation protective capacity of real estate is controversial, in the next
paragraphs the focus will be on the impact of deflation on required returns on real estate. Different methods have
been applied.
24
4
Deflation and real estate
4.1
Introduction
This chapter concentrates on the question what the effect of deflation is on the return that an investor requires on
his/her real estate investment. Paragraph 4.2 starts with the built up method. Paragraph 4.3 focuses on the risk-free
rate and paragraph 4.4 looks at the influence of the interest rate. Then in section 4.5 gives details on the difference
between nominal and real estate returns in an inflation and in a deflation scenario. Paragraph 4.6 zooms in on the
impact of deflation in the Japanese real estate market. This chapter is closed with the conclusion, including the
hypothesis which logically results from the literature and findings from this chapter. The hypothesis will be tested in
chapter 6.
4.2
Deflation, build up method and real estate returns
There are multiple ways to determine the required return on an investment. One of the approaches is the so-called
‘summation approach’ or ‘built up method’, which is often also called ‘risk premium method' (Van Gool P., 2013, e.a.).
With on the basis of a risk-free rate and various premiums the market discount rate / the risk premium can be
approached. Van Gool appoints as an example of a new investment in German retail shopping centre:
4.5%
1.0%
1.0%
0.5%
the risk-free rate, for example the return on 10-year German Government bonds, plus
General risk premium for investment property, plus
General risk premium for retail, plus
Asset-specific risk premium for shopping centres in German cities, results in
7.0%
required return
It is relevant to note that the risk premium varies over time. The recent development of the 10-year German
Government bonds for example has helped to ensure that initial yields in the prime/core-segment of many European
real estate markets have been on a downward trend and at historically low levels. See Figure 4.1.
Figure 4.1 Prime Property yields core Europe, the yield on German 10-year government bonds and the spread
Prime property yields en German bund yield (10yr)
8,00
7,00
6,00
5,00
4,00
3,00
2,00
1,00
0,00
Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1
2008
2009
2010
2011
2012
2013
spread vs office
industrial + logistics prime yields
retail prime yields
10 yr bond yield Germany
2014
2015
2016
office prime yields
Source: CBRE and Datastream
This figure also shows the 'spread'. The return on real estate is often compared to the return on risk-free rate (German
government bonds). The difference is called the 'spread'. In the example of the investment in the German shopping
centre the difference between the required return and the risk-free rate is 2.5%. A reduction in the risk-free rate will
result in a lower required return. The difference between the required return and the risk-free rate will be bigger in
25
the event of a reduction in the risk-free rate; the spread is wider. This is a development that is currently visible on the
European property market. Often, in the comparison, also is looked at the return on high yield corporate bonds, for
example the BBB rated bonds. In terms of credit, these are more comparable instruments and give a better feeling of
the spread of real estate. Geltner (2013) makes the link with the bond market. He distinguishes the following
components in the "contractual yield component stack": illiquidity premium, yield degradation, default risk, yield
curve, component, inflation premium and real risk-free rate. Later on in this section the above example of the required
return of a German shopping centre investment comes back again in a deflation scenario.
It is clear that a lower risk-free rate will ensure a lower required return. However, as the built up method shows, there
are more premiums to be distinguished than purely the risk-free rate, to determine the required return. This is
important, because these premiums can demonstrate another development than the risk-free base in an event of
deflation. Indeed, deflation is usually accompanied with economic recession. It is quite possible that investors require
a higher premium because of the risks that a recession brings. One can think of defaulting tenants, who are no longer
in a position to meet their rental obligations, problems to solve vacancy, reductions in the level of the rent, etc. Risks
like these increase with an economic recession. It is likely that the investor wants to be compensated for these risks. It
is the question how these higher premiums relate to the lower risk free rate. This tension makes it difficult for
unambiguous statements about the required returns in a deflationary environment.
To indicate this, the example of the German shopping centre investment, as presented in the beginning of this section, is
shown again, but then in a deflation scenario. In a deflationary environment the risk-free rate is not 4.5%, but -1%. The
general risk premium for real estate is assumed to be higher, as a result of the unfavourable macro-economic conditions
that involve risks, as defined, for which the investor wants to be compensated. In this example an overall risk premium
for investment in real estate is assumed to be 3%. The general risk premium for shops and the asset-specific risk
premium remain unchanged in this example. This results in a lower required return of 3.5%. It is clear that the lower
required in this example in particular is caused by the negative risk-free rate, which is partially compensated by a higher
risk premium. The required return is therefore lower with this approach, but the spread is wider: the difference between
the required return and the risk-free rate is now 4.5%. So there is a field of tension. On the one hand ensure
unfavourable macro-economic conditions for a negative impact on real estate. Risks are for example increasing vacancy,
rent decreases and more chance of defaulting tenants. For such risks the investor would be compensated with a higher
risk premium. On the other hand a low or negative risk-free rate makes the required return lower. A reduction in the
risk-free rate leads to a wider spread and a higher real performance.
-1%
3.0%
1.0%
0.5%
the risk-free foot, for example the return on 10-year German Government bonds, plus
General risk premium for investment property, plus
General risk premium for shops, plus
Asset-specific risk premium for shopping centres in German cities, results in
3.5%
required duration performance
Previous research (J.P. Hildering, 1999) has shown that the effect of deflation on real estate returns is not a
straightforward one. This research analyzes the initial yield, built from a risk-free rate plus a risk premium on real estate
minus the expected growth. An investor will require a higher return, as the return on other investments is also higher
(risk-free rate) and the specific risk of the real estate is higher (the risk premium). In this study, the risk-free rate shall be
the same as the average considered investment property. Because monetary deflation is not conducive to the economy
and thus not for real estate, the risk premium will increase. Regarding growth expectations, because of the fall in prices
also the replacement costs of buildings will decrease. This causes pressure on the rental prices. It is relatively attractive
to buy. All in all it is concluded that a combination of the above mentioned effects of deflation do not always provide a
clear effect on the return.
26
4.3
Deflation, risk-free rate and real estate efficiency
Another, more theoretical formal method to determine the required return is the capital asset pricing model (CAPM).
The required return is a function of the risk-free return, the systematic risk and the expected return of the market. It is
the question what happens with the required return if the risk-free rate falls. As risk-free return often long-term
government bonds are used. The CAPM builds on the Modern Portfolio Theory of Markowitz. CAPM believes that all
investors would have to invest the market portfolio, in combination with a risk-free investment. The model states that
due to this combination portfolios are possible with a more favourable risk return ratio. CAPM also maintains that an
investor receives purely return for the systematic performance risk or market risk. This is the risk that remains after full
diversification. This risk cannot be influenced and is produced by general developments in the market. An investor
would be compensated for this by means of a risk premium. The higher the risk, the higher the risk premium required.
The systematic risk or market risk is measured on the basis of beta. Beta measures the extent to which the investment
responds to general developments in relation to the market. When beta is equal to 1, then the systematic risk of the
investment is equal to the systematic risk of the total market. If beta is greater than 1, the systematic risk is greater
than the risk of the total market and at a beta less than 1 this is smaller (Gool P., e.a., 2013). The formula reads as
follows:
ERi = Rr + betai ∗ (ERm − Rr)
ERi= estimated required return on investment i
Rr = risk-free return
Betai = systematic risk of investment i
ERm = the expected market return
The required return for an investment can be approached, because, following CAPM, systematic risk must be rewarded
(Gool P., 2013). As long as the market returns are higher than the risk-free return (an investment will have a beta greater
than 1), a return is required that should be above return of the market. What happens with the required return as the
risk-free rate fall? The required performance decreases when the risk-free rate fall. The following example illustrates
this. Suppose a risk-free rate of 4%, an expected market rate of return of 8% and a beta of 0.5. The required return on
this investment amounts to 6%. If the risk-free rate would be not 4% but 2% and the rest of the variables would remain
the same, than the required return would be 5%. Investors can therefore be sufficient with a lower return at a lower riskfree rate. The formula shows that a lower risk-free rate results in a lower required return.
4.4
Deflation, interest and real estate efficiency
It is important to note that inflation/deflation is a given and that the money market interest rate is an instrument of
the bank. Each month the Board of Directors of the European Central Bank (ECB) decides on the official interest rates
in the euro area. This is the ‘money market interest rate’ or the 'short-term interest rate'. In this way the ECB tries to
have influence on the rate of inflation. The ‘capital market interest rate’/'long-term interest rates' is determined by
investors.
In particular the capital market interest rate has an effect on real estate, because real estate loans have often a
duration of a long period of time. Often, long-term interest rates and the short-term interest rates move in the same
direction. In the last few years the interest showed a downward trend. A lower interest rate leads to a lower required
return in the short term because of attractive financing. However, it is important to note again that having debts is
very disadvantageous in the long term in the event of persistent deflation.
4.5
Nominal and real performance
Returns can be measured at a nominal and at a real level. For the majority of investors it is the real return that
matters. This is the net return, which is corrected for inflation or deflation. A real return is a return that includes the
current purchasing power factor constantly. Many returns are reported as nominal returns performance, where
inflation or deflation makes part of (Geltner et al. 2013).
27
Suppose there is an inflation rate of 3.33%. Suppose that an asset is worth € 1,000,000 per T=0 and that asset would
show a 2% rise in value, being worth € 1,020,000 per T=1. The net rent income amounts to € 80,000. If rent is paid on
the end of the year, the nominal return would be:
RN = (80,000 + (1,020,000 - 1,000,000))/1,000,000 = 10%
RN = nominal return
This 10% is made up of 8% direct return (return coming from rental income) and for 2% from capital gain (return from
rise in value). In order to obtain the real return rates, the return should be corrected for the inflation rate of 3.33%. The
real rate of return is:
RR = (77.422 + (987,129 - 1,000,000)) /1,000,000 = 6.46%
RR = real return
If the rent is not paid on the end of the year, but every month (€ 6.667), than the nominal direct return would be 8.3%
and the total return would be 10,3%. Corrected for inflation, this represents a real total return of 6.7%.
Looking at the first example, the difference between the nominal and real return therefore would be around 3.55%.
The difference is bigger than the inflation rate of 3.33%, because inflation has not only effect on the value of the
property at the beginning of the period (€ 1,000,000), but also on the rental income(€ 80,000) and the increase in
value (€ 20.0000). In real terms there is a decrease in value of 1.29%. The property should have a value of € 1,033,300
to keep pace with inflation. The effect of inflation on the capital gain/change in value, is greater, in terms of
percentage, than it is on the direct return. This is explained by the size of the annual rental income being less than the
value of the property. Inflation is often ignored when looking at the income component of an investment property
(Geltner et al. 2013).
We show an example where we assume deflation. Because deflation often goes hand in hand with an economic
downturn, the assumptions are adjusted. It is assumed that the rent is not € 80,000 per year anymore, but € 60,000
per year, so € 5,000 per month. In addition it is assumed that the value of the property after a year is not increased to
€ 1,020.000, but fell to 2% to € 980,000. Table 4.1 shows the distinction between nominal and real effectiveness in an
inflation and deflation in a scenario.
It is made clear that the real total return increases in a deflation scenario. The difference between nominal and real
total returns, as seen in the event of inflation, is also the case in the event of deflation, and for the same reason as
already described in the example of inflation. Deflation is particularly detrimental to the performance when this occurs
if the asset was purchased in an inflationary environment. When a loan is agreed with a fixed interest rate is and the
actual inflation is lower than expected, it is the bank person who benefits from it. Having a debt is not attractive.
Table 4.1 Nominal and real return on an investment property by inflation and deflation
Nominal
Value asset T=0
Value asset T=1
Rental Income
Direct return
Indirect return
Total return
Inflation
€1,000,000
€1,020,000
€80,000
8,00%
2,00%
10,00%
Real
Deflation
€1,000,000
€980,000
€60,000
6,00%
-2,00%
4,00%
28
Inflation 3.33%
€1,000,000
€987,129
€77,422
7,74%
-1,29%
6,46%
Deflation 3.33%
€1,000,000
€1,013,758
€62,067
6,21%
1,38%
7,58%
4.6
Deflation and the Japanese real estate market
This section deals with the Japanese real estate market in relation to deflation. Kawaguchi (2009) investigates the
Japanese housing market in the period 1972 to 2006. The Japanese housing market is down after the nineties.
Nationally the average house price dropped by an average of 3.7% per year in the period 1992- 2006. Reasons are the
decline in population size, the lost decade in the period 1992- 2001 and the deflation period of 1998-2005. This
decrease was also caused by the oversupply of homes, despite the ongoing period of economic stagnation. Too many
newly built homes in Japan during the lost decade and the deflation period can be explained by government policy.
The government took anti-cyclical measures to stimulate demand, including an interest rate of 0%, the easing of loan
conditions and less transfer tax. This incentive measures and monetary policy (easy and cheap money to borrow)
caused more supply on the housing market. The problem is that as long as more homes were built, the Japanese house
prices fell further (Mr Kawaguchi, 2009).
It is relevant to look at the investigation on inflation of De Bruine (2009), in which the inflation hedging quality of
Japanese real estate is analyzed. In Japan there is a 'flat rent' regime: the rents remain generally the same for the
duration (3-5 years) of the contract. This seems not to be a hedge against inflation. Examination of Ganesang and
Chiang (1998) and Chu and Sing (2004) show that in Asian countries with a comparable rent regime real estate is not a
good protection against inflation. De Bruine (2009) notes that there are little significant results in his study. The
analyzed period is 2003-2007, a particular economic situation in Japan. From 2003 to 2005 there was deflation,
followed by minimum inflation. In the event of minimal inflation and deflation flat rent contracts can offer protection,
but the question is whether that also applies if there is a high rate of inflation. It is concluded that the return of capital
gain contributes stronger to the inflation hedge capacity than the direct return (De Bruine, 2009).
In theory a flat rent regime could provide some protection in the event of deflation. A landlord is in the advantage, if
rental prices can remain at the same level in the event of long-term deflation. If the contract rent can remain stable,
because it was agreed in the lease, whilst the market rent fall as a result of deflation and economic downturn, then
such contracts offer a deflation hedge in terms of rental income. In the event of prolonged deflation it is plausible that
rent prices are adjusted downwards after the expiration of the rental contract. Data from the Japanese office market
show that rent prices indeed fall for a long time. The Miki offices index is a monthly series with the average rent for
existing office stock in Tokyo. The rental prices are in Yen per Tsubo, the Japanese surface area measurement. The
rental prices in Tokyo have noticeably decreased. Where in the early nineties still on average around 45,000 Yen per
Tsubo was paid, this is the end of 2013 only 16,000 Yen on average; a factor 2.8 less. See Figure 4.2. To what extent
can the development of the rental price be determined by vacancy and inflation? This question is empirical examined
in the following section.
Figure 4.2 Average rent office space Tokyo
50.000
45.000
40.000
35.000
30.000
25.000
20.000
15.000
10.000
5.000
0
Source: Miki offices index
29
4.7
Conclusion
The impact of deflation in the required performance is twofold. Deflation is often associated with economic recession.
Less demand for real estate on the underlying market ensures falling rents, decreasing values and a rise in vacancy.
Also on the investors market demand for properties can decrease, because it is not attractive to borrow. Real estate is
a capital intensive asset class. This will have a negative impact on the value of real estate. In addition, a high leverage
press the value, because the debt on the repayment will be more than at the time of the commitment of the debt. For
such risks the investor would be compensated in a higher risk premium.
On the other hand it is quite conceivable that real estate in relation to for example fixed income can offer an attractive
return, even in periods of deflation. In addition, real estate offers an advantage because leasing contracts (temporary)
can provide protection against deflation. This can lead to a lower risk premium. In addition, a lower risk-free rate, which
is plausible in a deflationary environment, results in a lower required return, as has been demonstrated in accordance
with the "built up method" and the capital asset pricing model. As a result of a declining risk-free rate the spread is wider
and the real returns are up. Because it is difficult to operationalise required returns, the hypothesis is about the return
on capital gain/change in value. The hypothesis is as follows: Real estate capital gains are muted in a deflationary
economy. This hypothesis is examined quantitatively in chapter 6.
The calculation example has illustratively demonstrated what happens with the nominal and real total return in an
inflation and deflation scenario. The impact in this example shows in particular large effect on the return on capital
gain/change in value.
In the following chapter I examine quantitatively whether there is a link between the changes in rent on the one hand
and the change in the price level and the vacancy rate on the other hand. More specifically, Japanese data are used,
because of the long period of deflation in Japan.
30
5
Empirical research I: explanation rent by vacancy and/or inflation?
5.1
Introduction
In this chapter a quantitative analysis is done to examine whether the change in rents can be explained by the change in the
price level and/or a change in the vacancy rate. This analysis includes also the impact of lagging. This relates to exploratory
research. Te data originate from the Japanese office market (see also paragraph 4.6). Paragraph 5.2 sets out the plausible
links. In paragraph 5.3 is described what data may be used. Paragraph 5.4 then focuses on the analysis and in the following
section 5.5 the conclusions become clear.
5.2
Expected Links
As stated, this chapter considers the impact of a change in the vacancy rate and a change in the price level on the change
in the rental price. Rent levels originate from rent levels of office space in Tokyo. These variables are selected, partly
because Geltner (2013) indicates that the rental price is a reflection of the balance between demand and supply, with
macroeconomic conditions playing an important role too. It is likely that there is:
1.
A negative relationship between the change in the vacancy rate and the change in rents;
2.
A positive relationship between the change in the price level (inflation/deflation) and the change in rents.
Ad. 1 Change in vacancy rate
It is expected that there is a negative relationship between the change in the vacancy rate and the change in the rental
price. In the case of declining vacancy tenants have a weak starting position; they have less choice where to rent space.
Such a situation is to the advantage of the owner. The owner is in a strong position to ask a higher rent. It is possible
that a change of the vacancy rate works with a delay in the change in the rental price. Lease contracts have a certain
duration, making adjustments in the rental price generally with a delay. Previously, the so-called saw tooth effect is
mentioned, where this refers to. For example, in the case of an economic downturn the situation can arise that
contract rents are at a level above the market rent. If the lease would be renewed then the contract rent has to be
adjusted to the market rent. In this situation real estate provides little protection against inflation. It is likely that the
vacancy rate would have an effect on the change of the rental price both with and without a delay. A negative
relationship is assumed. An effect without lag is plausible, because it is expected that the degree of supply/choice for
the tenant influences the negotiating position of the tenant. An effect with lag can be explained due to the previously
defined saw tooth effect.
Ad. 2 Change price level
It is expected that there will be a positive relationship between the change in the price level and the change in the
rental price. As defined in paragraph 3.2, it is of great importance whether there is an economic cyclical upturn or
downturn. If there is an economic upturn the market rent should increase at least in line with inflation, because there
is a lot of demand for real estate. The owner has a strong negotiating position. In fact, the tenant has little alternative
possibilities. A downside of this is that structurally high inflation brings significant increases of the rent. In the event of
poor economic growth it may be difficult for tenants to pay such high rents, what ultimately can result in a rise in
vacancy. In the event of a downturn the situation can arise that contract rents are above the market rent. Lessees have
a strong position when they have a lot of choice where to rent space. Inflation can influence the level of the rent with a
delay. When the rental contract expires, a new price has to be negotiated, which will be in line with the market rent. It
is expected that there is both a direct effect as an effect with time lag. The direct effect can be a result of the extent to
which rental prices can be increased; often inflation is used. The effect with time lag is expected, because of the above
mentioned saw tooth effect. As a result of long-term leases only at the renewal of the lease rental price adjustment
can be made.
5.3
Description data
In order to carry out the analysis the following data were collected.
31
3.
The change of the rent, originate from Miki offices index. This is monthly basis series. 12 months moving
4.
5.
averages are calculated. The development of the rental price is shown in figure 4.2 in section 4.6.
The change in the price compared to a year earlier (inflation/deflation).
The change in the vacancy rate.
The forecasting period runs from 31 December 2000 to 31 May 2015. In figure 5.1 and 5.2 these variables are
displayed graphically, always together with the change of the rental price, because this the variable to explain. To see
the relationship between rent and vacancy rate more clearly, it is a conscious choice to display graphically the
occupancy rate instead of the vacancy rate. In the analysis the change of the vacancy rate is included. Figure 5.1 shows
quite clearly that there is a positive correlation between the development of the occupancy rate and the change in the
rental price. This is in line with the expectation. If the occupancy rate increases, the rental price increases and vice
versa. The link in figure 5.2 seems at first sight to be positive, but is slightly less straightforward. The expectation was a
positive relationship.
Figure 5.1 Rental Prices and occupancy rate Tokyo
Figure 5.2 Rental prices and inflation
change rental price (L) and inflation (R)
Change rental price(L) occupancy rate(R) offices
2,0%
100
1,5%
98
3
-1,5%
90
-1,5%
1-1-2015
1-5-2014
1-9-2013
1-1-2013
1-5-2012
1-9-2011
1-1-2011
1-5-2010
1-9-2009
1-1-2009
1-5-2008
1-9-2007
1-1-2007
1-5-2006
1-9-2005
1-1-2005
1-5-2004
1-9-2003
-1,0%
1-1-2003
-0,5%
92
1-5-2002
94
2
0,0%
1-9-2001
1-1-2015
1-5-2014
1-9-2013
1-1-2013
1-5-2012
1-9-2011
1-1-2011
1-5-2010
1-9-2009
1-1-2009
1-5-2008
1-9-2007
1-1-2007
1-5-2006
1-9-2005
1-1-2005
1-5-2004
1-9-2003
1-1-2003
1-5-2002
1-9-2001
1-1-2001
0,0%
0,5%
1-1-2001
96
0,5%
-1,0%
4
1,5%
1,0%
1,0%
-0,5%
2,0%
1
0
-1
-2,0%
-2,0%
88
-2,5%
-3,0%
86
change rental price (%)
5.4
-2
-2,5%
-3,0%
-3
change rental price (%)
occupancy rate (%)
inflation (%)
Correlation and regression analyzes
Because the expected relationships are described and the data are defined, it is possible to examine whether there is a
significant link between the change of the rent and the change in vacancy, and the change of the rent and the change
in the price level by using a multiple regression analysis. The strength and the direction of the connection can be
demonstrated. It is assumed that:
ΔH = α + βΔP + γΔL
ΔH = the 12 months moving average of the rental price
ΔP = The change in the price level compared to a year earlier (inflation/deflation)
ΔL = The change of the vacancy rate on a monthly basis
Since real estate responds delayed to changing market conditions, I will also give a statement on different time lags. It
was therefore decided to run multiple regression models for which either the vacancy rate, either the inflation variable
has a lag. In some models for both variables the same lag was applied. The extent to which the X variables (change of
the price level and the change of the vacancy rate) explain (the Rsquare) the Y variable (12 months moving average of
the change of the rental price) diverges. It is also clear from some models that there is a significant relationship, but
not in all models. The degree of lagging plays a role. The results are summarized in order to get a feeling of the impact
of the various lags. A confidence level of 95% applies to all models.
•
•
At no time lag the model shows a Rsquare of 28%. The variable vacancy rate is not significant, but inflation
is. For inflation there is a slightly positive relationship (see Appendix: Model 1).
In the event of a three months lag for both inflation and vacancy rate the model shows again a Rsquare of
32
•
•
•
•
•
28%. The two variables have significant P values. The coefficient for the vacancy rate is negative; this is in
line with the expectation. The coefficient for inflation is positive, what was also assumed (see Appendix:
Model 2).
When inflation is delayed only with three months and the vacancy rate has no lag, then it appears that the P
value for the vacancy rate variable is not significant. The Rsquare comes down to 27% in this model (see
Appendix: Model 3).
When only the vacancy rate is delayed with three months, then there appears to be no significant link
between the 12 months moving average of the rent and the change in the vacancy rate. The Rsquare is 29%
(see Appendix: Model 4).
What happens with a lag of 12 months at both vacancy rate and inflation? This improves the fit of the
model significantly, the Rsquare is 40%. The P values of the two variables are also significant. The negative
coefficient for the vacancy rate variable and the positive coefficient for the inflation variable are both in line
with expectations (see Appendix: Model 5).
The model with a lag of inflation with 1 year and no lag for vacancy rate provides no significant P value of
the vacancy rate variable. The Rsquare is only 4% (see Appendix: Model 6).
In the event of a lag of 1 year for the vacancy rate variable and no lag for inflation the model demonstrates
the best fit with a Rsquare of 49%. The two variables show significant P values. The negative coefficient for
the vacancy rate variable and the positive coefficient for the inflation variable are both in line with
expectations (see Appendix: Model 7).
Apparently, an increase of the vacancy rate, in particular after one year, allows a price reduction. Tenants can enforce
a lower price, for example at the end of the rental contract, or by the use of break options. Because rental leases or
break options usually are committed in the time unit of years, it is logical that the models with a three-months lag
show less optimal results.
What would happen with the results in case there is a lag of more than one year ? It has to be noted that the greater
the time lags are the more data is lost; the number of observations decreases rapidly. I test the model with a two-year
lag and a three-year lag for both variables. In both models (see Appendix: Model 8 and 9) we see very low Rsquares,
1% and 2%. In both models the P values for the two variables are not significant. Apparently, a change of the vacancy
rate or price level, taking place two years ago or longer, has no impact on the change in the rental price. The lag is
probably too large. The real estate market is lagging, but changes of these variables that happened two or three years
ago, have no significant impact anymore. Changes of two or three years ago might already be 'overtaken’ by changes
of one year ago.
Table 5.1 gives an overview of the results of the different models. The models with the highest model fit are model 5
and model 7. In model 5 there is a lag of 12 months for both X variables; in model 7 is there is lag of 12 months for the
vacancy variable. The model fit in these models is significantly higher in relation to the other models. Also in these
models the P values of the two variables are significant. The relationship with the change of the vacancy rate is
negative and the relationship with the change in the price level is positive, consistent with the expectations. The
coefficients show that the relationship with the vacancy rate is stronger than the connection with the price level.
Apparently, according to this analysis, vacancy rate has more effect on the change of the rental price than
inflation/deflation.
Since model 7 gives the best results, we look more in detail to this model. Table 5.2 shows the regression statistics.
The Rsquare is 49%. If the Rsquare 1, it would mean that there is a 100% model fit. In this case almost half of the
moving average of the change of the rental price is explained by the change in the vacancy rate with a 1 year lag and
no lag for the change in the price level. This means a fairly high model fit. The number of observations in the model is
161. As said, there are observations lost as the time lag becomes greater.
33
Table 5.1 Summary Results Regression Models with various time lags (95% confidence level)
model
1
2
3
4
5
6
7
8
9
lag in months
delta vacancy
Delta price level
0
0
3
3
0
3
3
0
12
12
0
12
12
0
24
24
36
36
P value
R sq
28,3%
28,2%
26,5%
28,9%
40,4%
4,3%
48,9%
1,0%
1,9%
coefficient
vacancy
Price level
vacancy
Price level
0,7541
0,0500
0,8567
0,2113
0,0000
0,4074
0,0000
0,4168
0,5553
0,0000
0,0000
0,0000
0,0000
0,0246
0,0118
0,0000
0,4453
0,1699
0,0030
-0,0186
-0,0018
-0,0120
-0,0841
-0,0092
-0,0673
-0,0109
-0,0082
0,0027
0,0025
0,0027
0,0026
0,0009
0,0013
0,0018
-0,0005
-0,0009
Table 5.3 summarizes the F test results. The F test tests the entire model. By dividing the sum of the squared
differences (sum of squares) by the number of degrees of freedom (degrees of freedom) the mean square deviation
(Mean Square) is obtained. This can be distinguished in an explained part (Regression) and a non-explained part
(Residual). The significance of F is smaller than 0.05. This F is significant: if the probability of overrunning would be less
likely than 0.05, then the value of F is statistically significant with 95% certainty. The calculation of F shows that the
regression equation is statistically significant.
Table 5.2 Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations
0.6995
0.4893
0.4829
0.0040
161
Table 5.3 F Test
df
Regression 2
Residual
158
Total
160
SS
MS
F
Significance F
0.0024 0.0012 75.7023 0.0000
0.0025 0.0000
0.0049
The results from the regression analysis are summarized in table 5.4. The regression looks as follows.
ΔH= α + βΔP + γΔL , so
ΔH = −0,0005 + (0,0018 ∗ ΔP)– (0,0673 ∗ ΔL)
In other words, each percentage that the change of the vacancy rate increases, after a year, results in a decrease in the
rental price. The percentage of reduction of the twelve months moving average of the rental price is 0,0673%. Each
percentage of price level increase has a positive effect on the rental price. The percentage increase in the twelve months
moving average of the rental price is 0.0018%.
Table 5.4 Regression results
Intercept
vacancy
Price level
Coefficients
-0.0005
-0.0673
0.0018
Standard Error
0.0003
0.0085
0.0003
t Stat
-1.5591
-7.8860
5.6982
P-value
0.1210
0.0000
0.0000
Based on the data series used in model 7 a correlation matrix is created (Table 5.5). This confirms the negative link
between the change in the vacancy rate (with a lag of one year) and the change of the rental price, as previously
assumed. It also shows a positive relationship between change in the price level (without time lag) and the change in
the rental price. This is in line with the previously expected relationships, as defined in paragraph 5.2.
34
Table 5.5 Correlation matrix
Change rental price
Change vacancy
Change price level
Change rental price
1
-0.62
0.536979
Change vacancy
Change price level
1
-0.38356
1
5.5
Conclusion
This chapter considers the possible impact of deflation on rental prices of office space in Tokyo. The data used are on a
monthly basis. I did several regression analyses with different time lags for every model. The variable to explain is the
twelve months moving average of the change in the rental price. The explanatory variables are the change of the
vacancy rate and the change in the price level, whether or not with a lag. Lagging is of great influence in this analysis.
It is striking that the model fit significantly improves, where there is a lag of 12 months for both X variables. The best
model fit follows from regression model with the change of the vacancy rate with a lag of one year (Rsquare of 49%). It
shows a negative relationship between the change in the vacancy and the rental price (with a time lag of one year)
and a positive relationship with the change in the price level compared to a year earlier. The two variables are
significant. A decline in the general price level, according to this analyses, therefore results in a decrease in the rental
price. Inflation is less significant compared with the vacancy rate.
The fact that the best model fit is the model with a lag of 12 months, can be explained by tenants who can enforce a
lower rental price, after the expiry of the lease, or by using break options, which are usually committed in the time
unit of years. Logically, the models with a three months lag show less optimal results. Apparently, a change of the
vacancy rate or the price level, taking place two years ago or longer, has no impact on the change in the rental price.
The real estate market is lagging, but changes two or three years ago, are already 'overtaken’ by changes of one year
ago.
So, it can be concluded that inflation or deflation have a significant positive effect on the change of the rental price.
Vacancy has more impact on the change of the rental price than inflation or deflation. Adding new supply to the real
estate stock in a deflationary environment can therefore be a problem. This is also shown in Japan, where the increase
in supply has contributed to continue declining prices (see paragraph 4.6). Especially in the light of the current low
interest rate environment it is recommended to be vigilant with many new construction projects. The low interest rate
may be attractive to borrow, at the event of deflation there is a large risk behind many new construction in relation to
the development of the rental prices.
The following chapter assesses the hypothesis Real estate capital gains are muted in a deflationary economy in a
quantitative review. This is done on the basis of performance data of the IPD Japanese real estate market and macroeconomic data.
35
6
Empirical research II: review hypothesis
6.1
Introduction
This chapter assesses the hypothesis Real estate capital gains are muted in a deflationary economy; this concerns
quantitatively reviewing research. This is done on the basis of return data sets from the Japanese real estate market.
Using regression and correlation analyses will be examined to what extent there is a link between deflation and real
estate returns of non listed Japanese real estate. This is to focus on the difference between the direct return and the
return coming from capital gain/change in value. For the justification of the hypothesis, I refer to section 4. The
chapter starts with a description of the data used in the quantitative analyses. Paragraph 6.3 will look at the mutual
correlations of the X and Y variables. And then there are several regression models analyzed. The results from
paragraph 6.3 shall be placed in a wider context in section 6.4. This chapter ends with the conclusions from these
quantitative analyses in paragraph 6.5.
6.2
Description data
To explain variables: not listed return data
For non listed real estate I used data of the IPD index, with returns on a monthly basis, from 31 January 2002, and
including a breakdown between direct returns and returns on capital gain/change in value. The considered sectors are
offices, retail and miscellaneous. I selected 'all properties', because this will result in the largest amount of data. The
returns series is in local currency in order to eliminate currency effects. The returns are shown graphically in figure 6.1.
This shows a relatively stable direct return (performance resulting from rental income) and a more volatile return on
capital gain (return from the development value). Together they form the total return. The dip of the total return in
the period 2008 to 2010, is mainly due to the sharp decline in the development in value. In this period Japan was facing
a strong economic shrink. The IPD index with non listed return series from the Japanese real estate market go back
until January 2002. Because the analysis is based on performance data on an annual basis, the first observation is in
January 2003. The analysis is therefore based on the period of January 2003 to October 2014, on a monthly basis with
an annual change.
Explanatory: Inflation data and GDP data
For the X variables is looked initially at two sets of inflation and a series of GDP data: actual inflation, expected inflation
and expected economic growth on an annual basis. These series are derived from Datastream. The ranges are shown
graphically in figure 6.2. The graph shows clearly the decline in economic growth in the period 2008 to 2010. In the
graph is also clearly visible that the inflation rate for a long period is around the zero point. Remarkable is the inflation
peak at the end of 2013 to approximately 2%. The explanation for this can be found in a temporary effect of the
Abenomics policy.
-6%
-8%
-15%
IPD total
IPD income
GDP YOY
IPD capital
36
Inflation YOY (actual)
Inflation expectation
1-9-2014
1-2-2014
1-7-2013
1-5-2012
1-12-2012
1-3-2011
1-10-2011
1-8-2010
1-1-2010
1-6-2009
1-4-2008
1-11-2008
-4%
1-9-2007
1-8-2014
1-1-2014
1-6-2013
1-4-2012
1-11-2012
1-9-2011
1-2-2011
1-7-2010
1-5-2009
1-12-2009
1-3-2008
1-10-2008
1-8-2007
1-1-2007
1-6-2006
1-4-2005
1-11-2005
1-9-2004
-5%
-10%
1-2-2004
-2%
1-7-2003
0%
0%
1-12-2002
5%
1-2-2007
2%
1-7-2006
10%
1-5-2005
4%
1-12-2005
15%
1-3-2004
6%
1-10-2004
20%
1-8-2003
Figure 6.2 Inflation and expected inflation and Expected GDP
Japan
1-1-2003
Figure 6.1 Returns IPD Japanese Real Estate
6.1
Correlation and regression analyzes
Before starting the regression analyzes, first a correlation analysis is carried out looking at the mutual correlations
between all variables. In table 6.1, the correlation matrix is displayed. This analysis shows that:
•
The total return and the return from capital gain/change in value show a high correlation; the total return is
driven by the return from capital gain/change in value. This is consistent with the expectation and also well
visible in figure 6.1;
•
There is a negative relationship between inflation and the direct return. This is not consistent with the
expectation. I come back to this later in this section;
•
Expected inflation and actual inflation are highly correlated with each other. This is logical and very visible in
figure 6.2.
In particular the latter finding is relevant for the regression analysis. Because of the high correlation between expected
inflation and actual inflation, I decided to choose one of the two variables and to eliminate the other. I selected
expected inflation; actual inflation is disregarded in the further analysis. When I would have been chosen for the
variable actual inflation, then a lower model fit follows in all three regression models compared with the model fit
where the variable expected inflation is used. This justifies the choice for the variable expected inflation.
Table 6.1 Correlation Matrix X and Y variables
IPD total
IPD income
IPD capital
Inflation YOY (actual)
Inflation expectation
GDP YOY
IPD total
1,000
0,186
0,993
0,404
0,427
0,590
IPD income
IPD capital
1,000
0,068
-0,197
-0,255
0,305
1,000
0,435
0,464
0,563
Inflation
(actual)
1,000
0,871
0,324
YOY Inflation
expectation
1,000
0,276
GDP YOY
1,000
Based on the data as described, it is by using regression analyses to be examined whether:
1.
A significant link exists between the direct return on the one hand and expected inflation and expected
2.
economic growth on the other hand;
A significant link exists between the return coming from capital gain/change in value on the one hand and
expected inflation and expected economic growth on the other hand.
It is expected that the relationship between the return coming from capital gain/change in value on the one hand and
expected inflation and expected economic growth on the other hand is positive. Using regression analyses this will be
tested. The focus is therefore on the inflation variable.
Total return
Starting with the regression analysis, the Y variable is the total return. The formula reads as follows:
Rt = α + βiv + γGDPv
Rt = total return
iv = expected inflation
GDPv = expected economic growth
This results in a model fit of 42%. The P values of both X variables are significant and the coefficients are positive. The
coefficient of the expected inflation rate is higher than that of expected economic growth. This means that expected
inflation has more effect on the total return than expected economic growth.
37
Table 6.2 regression results indirectly efficiency (R sq=42%)
Coefficients
Intercept
0,033
inflation expectation 1,739
GDP expectation
1,239
Standard Error
0,004
0,400
0,161
t Stat
8,307
4,348
7,695
P-value
0,000
0,000
0,000
Direct return
Then the regression analysis is done with as Y variable direct return. The formula reads as follows:
Rd = α + βiv + γGDPv
Rd = direct return
iv = expected inflation
GDPV = expected economic growth
The results show a model fit of 22%. In this model, it is also the case that both X variables have significant relationship
with the direct return. The direct return remains relatively stable. The direct return shows relatively little change in the
reviewed period. The explanation for this is that the rental contract is agreed with a certain maturity, with rent
remaining the same throughout the duration of the rental contract ('flat rent regime'). As a result, no effect with inflation
could be well explained. If there is any change of inflation, it has no effect on the rental price, because this remains 'flat'
for the duration of the rental contract. In theory, a flat rent regime offers a temporary protection against deflation. Such
a rental contract however turns out to be detriment, if there is (unexpected) inflation at the time of the current rental
contract.
Remarkable is the positive relationship between GDP expectation and the direct return, while the relationship
between expected inflation and the direct return is negative. It would be reasonable to assume that the effect of
expected economic growth and inflation on the direct return would point out in the same direction. This is however
not the case. Apparently, a positive expected GDP leads to a higher direct return, but higher expected inflation leads to
a lower direct return. The positive relationship between economic growth and the direct return is logical to explain,
due to the expectation of future rental growth.
Table 6.3 regression results direct return (R sq=22%)
Coefficients
Intercept
0,054
inflation expectation -0,254
GDP expectation
0,119
Standard Error
0,001
0,055
0,022
t Stat
97,585
-4,603
5,362
P-value
0,000
0,000
0,000
Return from capital gain/change in value
Finally the analysis follows to explain the return coming from capital gain/change in value on the basis of expected
inflation and expected economic growth. The formula reads as follows:
Ri = α + βiv + γGDPv
Ri = return coming from capital gain or return coming from a change in value
iv = expected inflation
GDPV = expected economic growth
The regression model shows a model fit of 42%. Both X variables show a significant link. The link is for both variables
positive. As is the case for the total return, this also applies to the return coming from capital gain/change in value: the
coefficient of the expected rate of inflation is higher than that of expected economic growth. That means that inflation
38
has more effect on the total return than expected economic growth. The difference between the coefficients of the X
variables in model looking at return coming from capital gain/change in value is greater than the model looking at total
return.
Table 6.4 regression results indirectly efficiency (R sq=42%)
Coefficients
Intercept
-0,020
inflation expectation 1,901
GDP expectation
1,072
Standard Error
0,004
0,377
0,152
t Stat
-5,226
5,036
7,051
P-value
0,000
0,000
0,000
This outcome supports the hypothesis that the return coming from capital gain/change in value drops in case of
deflation. The relationship between return coming from capital and expected inflation is positive and significant. This
means that if there is any negative price change (deflation) the return coming from capital will also drop. This is clear
from both the results of the correlation matrix and the regression analyses. The hypothesis is therefore supported by
statistics in these analyses. Deflation leads to a muted return coming from capital gain/change in values. If the price
level drops, also the return coming from a change in value decreases.
6.3
Discussion
The results from paragraph 6.3 shall be placed in a wider context in section 6.4. Deflation results in a decline in the
value of real estate. What would the relationship be in an inflationary environment? In an inflationary environment a
positive relationship between inflation and return on capital gain/change in value is not always obvious. After all,
economic growth is usually accompanied with a rising inflation and a rising interest rate. A rising interest rate will often
lead to higher initial yields, implying a decrease in the value. It has to be noted that economic growth is often
accompanied by more demand for property, what may bring an increase in rents. Higher rents have a positive effect
on the value of real estate. However, if in the current low interest environment there will be an increase in interest
rates, than there must be such a rent increase, which either need to be significantly, or for a long period, to
compensate a fall in the value of the current high valued real estate, as a result of higher interest rates.
How do the results relate to the current developments in the European property market? Currently, Europe is in a very
low interest rate environment and inflation is very low. On the European property market the initial yields of core real
estate are falling, which implies a rise in value. Would it be fair to say a low inflation entails a rise in value? That is
rather simplistic. Hence a bit of nuance is needed. The current rise in value of core real estate in Europe is for a large
part driven by the low interest rates and the expectation that the interest rate remain low provisionally. It is likely that
the interest rate will not increase, until the rate of inflation will reach an acceptable level. The development of the
inflation factor has therefore an important role, but it is mainly the low interest rate that causes the current yield
compression.
It seems to be that the interest rate will remain low provisionally, given the further extended policy of the ECB. The
rate of inflation in Europe remains low and in November 2015 inflation was only to 0.1%. This is mainly due to the low
oil price, but without the prices of energy the inflation would not be higher than 1%. In December 2015 the ECB has
therefore extended the buy-back program of 60 billion Euros per month from September 2016 to March 2017 (this
means an additional EUR 360 billion in the system). The objective is to reduce interest rates, making investments more
attractive and pushing the economic growth and inflation. Also the deposit rate was reduced from -0.2% to -0.3%:
banks must now pay a higher fee for liquidity that they put away to the balance of the ECB. That would offer incentives
to the individual banks to borrow the money and to let the many ‘work’. Finally, the ECB will invest again the
redemption of terminating bonds. The result of this is that the liquidity, which is pumped into the system, will remain
present pressing interest rates for a long time after the expiry of the buy-back program.
Certain underlying real estate markets improve slightly. Yet it is the question whether the underlying market (in certain
39
markets) justify the rising prices sufficiently. A correction can lead to continued price decreases causing asset price
deflation. That is a risk. As long as that does not happen and as long as the interest rate remains low, it is likely that
investors in European real estate will continue to invest, partly because the spread remains attractive (see also
paragraph 4.2). Investing in real estate is an attractive alternative compared to invest in fixed income. This translates
as yet in rising volumes of investment in European real estate, which is now approaching the record levels of 2007.
6.4
Conclusion
In this chapter, the hypothesis Real estate capital gains are muted in a deflationary economy was investigated
empirically. The results from of the various analyses showed that this statement can statistically be assumed. The
regression analysis that explains the relationship between the return on capital gain/change in value on the one hand
and the expected inflation and expected economic growth on the other hand is significant. The relationship between
return on capital gain/change in value and expected inflation is positive. This is also apparent from the correlation
matrix.
As applies for the total return also for the return on capital gain/change in value the coefficient of the expected rate of
inflation is higher than that of expected economic growth. That means that inflation has more effect on the total
return than expected economic growth. The difference between the coefficients of the X variables in the return on
capital gain/change in value is greater than on the total return.
The variable expected economic growth has a significant and positive impact on both the direct return as on the return
on capital gain/change in value. This is a logical relationship to explain. When the expectations of economic growth are
positive, then there will be generally a high level of consumer confidence, more expenditures, more employment and
more demand leading to higher returns and rental income. A higher economic growth leads to expectation of future
rental growth. Remarkable however is the positive relationship between expected GDP and direct return, while the
relationship between expected inflation and direct return is negative. It would be reasonable to assume that the effect
of expected economic growth and inflation on the direct return would point out in the same direction. This is not the
case. Apparently, a positive expected GDP leads to a higher direct return but a higher inflation expectations results in a
lower direct return.
When putting these results in a broader context, we have seen that:
• A positive relationship between inflation and return on capital gain/change in value is not always logical in an
inflationary scenario;
• The current rise in value of core real estate in Europe is for a large part driven by the low interest rates and
the expectation that the interest rate remains low provisionally. The ECB has recently broadened its program;
• It is likely that, as long as there is no correction and the interest rate remains low, investors will continue to
invest in European real estate, partly because the spread remains attractive.
40
7
Conclusion and reflection
7.1
Conclusion
Deflation was something unique for a long time. Japan was always the only country in the world with longer periods of
deflation. Fear of deflation is now also a hot topic in Europe. Since the beginning of 2015 also Europe has been facing
short periods of deflation. Regarding inflation and real estate many studies have been published. Generally, it is
accepted that real estate provides protection against inflation. About the relationship between deflation and real
estate less research is done. This survey deals with deflation and its impact on real estate returns. This is a reviewing
research. On the basis of literature the following hypothesis is distilled: Real estate capital gains are muted in a
deflationary economy.
In the first place the definition of deflation was described: “a sustained decrease in the average price level, measured
on the basis of the development of a price index." Deflation may be caused by the supply effects or by demand effects.
The effects of deflation are defined, as well as the specific consequences for pension funds: deflation is particularly
dramatic for the height of the coverage ratio. One of the measures against deflation is monetary easing. In this context
it is important to make the distinction between asset price inflation and consumer price inflation. QE has the goal to
increase the consumer price inflation. The result however is often purely asset price inflation. This brings risks.
That deflation can have negative consequences for the long-term economy, was shown during the lost decade in
Japan. The demographic developments in Japan have negative impact on demand. Fiscal and monetary policies were
not efficient. As a result Japan was facing deflation for years and turned in a negative spiral. Not only the Japanese
saved more, also expenditures were postponed. Incentive measures in the public sector were even counterproductive.
Some believe that a major cause for deflation in Japan has been the currency policy; prices had to fall enough to
decrease real wages. The Abenomics policy, with the aim to push inflation, shows little effects so far. The effect of
quantitative easing is questionable in general, not only because of the limited effect on the level of inflation but also
because markets devalue their currencies, due to QE, at the expense of other markets, which in their turn use QE to
improve their competitive position. A risk is a 'race to the bottom". In this section I also analysed the similarities
between Europe and Japan: Europe suffers from a very low inflation; some countries suffer even with deflation, the
Economic growth disappoints, interest rates break new records troughs and solving problems is difficult. In both Japan
and Europe QE does not succeed so far to get the inflation level toward the objective of 2%. Both regions showed
significant asset price inflation. Finally, many European countries, although to a lesser extent than Japan, have an aging
population.
Before dealing with deflation and real estate, I put the question under the microscope whether the general association
that real estate offers inflation hedging qualities is true. That is not the case. The impact of deflation is on real estate is
examined in multiple ways. In the first place the impact on the required returns is analyzed. In this question there is a
tension. Often deflation is associated with economic recession. Less demand for real estate on the underlying market
provides more risks, like falling rents, decreasing values and a rise in vacancy. For such risks the investor would be
compensated in a higher risk premium. On the other hand it is quite conceivable that real estate in relation to for
example fixed income can offer an attractive return, even in periods of deflation. The real return increases and the
spread widens on a declining risk-free rate. Because it is difficult to operationalise required returns, the hypothesis
focuses on return coming from capital gain/change in value. The hypothesis is as follows: Real estate capital gains are
muted in a deflationary economy.
Then the possible impact of deflation on rent prices was examined in an exploratory research, using different
regression analyses. Data were from office rents in Tokyo. The variable to explain is the twelve months moving average
of the change in the rental price. The explanatory variables are the change of the vacancy rate and the change in the
price level, whether or not with a lag. The two main conclusions are:
41
•
Lagging is of great influence in this analysis. There is a negative relationship between the change in the
vacancy and the rental price (with a time lag of one year) and a positive relationship with the change in the
price level compared to a year earlier. The two variables are significant. A decline in the general price level,
according to this analysis, therefore leads to a decrease in the rental price.
•
The relationship with inflation is less significant compared with the relationship with the vacancy rate.
Adding many new supply to the real estate stock, as also happened on the Japanese housing market, had a
negative effect to the rental price development. Especially in the light of the current low interest rate
environment it is recommended to be vigilant with many new construction projects. The low interest rate
may be attractive to borrow, in the event of deflation there is a large risk coming with many new
construction projects in relation to the development of the rental prices.
The hypothesis is Real estate capital gains are muted in a deflationary economy. This is quantitatively reviewing
research. This is done on the basis of return data sets from the IPD Japanese real estate market and macro economic
data. The results of the various analyses proved that this hypothesis may be supported by statistics:
• The regression analysis, explaining the relationship between the return coming from capital gain/change in
value on the one hand and the expected inflation and expected economic growth on the other hand, is
significant.
• The relationship between the return coming from capital gain/change in value and expected inflation is
positive. This is also apparent from the correlation matrix.
• Deflation leads to a low return coming from capital gain/change in value on real estate. If the price level
drops, also the return coming from capital gain/change in value decreases.
When putting these results in a broader context, I elaborated that a positive relationship between inflation and return
coming from capital gain/change in value does is not always makes sense in an inflationary scenario. Also the principal
statement of the current increase in the value of core real estate in Europe is described, namely the low interest rates
and the expectation that the interest rate will remain low provisionally, because of the further enlarged monetary
policy of the ECB. Recently the ECB monetary policy broadened, because of the current low inflation. As long as the
interest rates remains low and there no correction takes place, it is likely that real estate will remain an attractive asset
class on the investment market because of the wide spread. If Europe is would face a deflation scenario, then this will
have a negative effect on the value of real estate.
7.2
Reflection
If Europe would actually be in a deflationary environment, real estate returns would hit hard: deflation will impact the
value and so the returns. In the event of a decline of the general price level, also the value of real estate drops. Also
deflation has effect on the rental price development. From that perspective it might be advisable to be reticent with
regard to real estate investments, anticipating a deflationary environment. The extent to which the impact on real
estate returns stands in relation to the effect on the performance of other asset classes would therefore be an
interesting subject for further research. Other interesting continued research could focus on the difference in the
decrease in value of prime assets versus non prime assets in a deflation scenario. It is probable the case that in a
deflation scenario prime assets will be in a better position to retain their value or show a less revaluation decrease,
than non prime assets. The differences between real estate sectors and their reactions on deflation is another
potential subject for a follow-up study. Finally, relevant research could focus on the question what the effects are of
QE on real estate returns in different markets and sectors.
Despite the fact that the analyses relate only to the Japanese real estate market, it is very likely that the results from
this investigation can be considered as a generally valid. It is important to note that the opportunities in relation to
data availability are limited. For quantitative analyses of deflation, the number of countries with a long history of
deflation is (luckily) (yet) limited to Japan. In addition, the underlying data of the analyses presented are subject to
many adaptations, because the raw data in the first instance showed no interesting results. This involved adjustments
42
in the form of unsmoothing, change on an annual basis or on a monthly basis, moving averages, varying degrees of
lagging, etc. Not all data adjustments appeared to be useful. This was a trial and error process and the analyses
presented ultimately led to the final result.
It is interesting to recognize the similarities between Japan and Europe. Japan showed that stopping too early with QE
resulted in a drop of economic growth. At the same time is a very pertinent question how long to continue with QE
being still tenable and also how effective QE is. QE is controversial; QE in one market ensures QE in another market.
How long will this race to the bottom go on? And when will there really be a hike in inflation? These questions were
highly topical at the time of writing this thesis.
That deflation is negative for the economy and for the asset class is clear. But how much is daily life actually affected
by deflation? Of course, deflation is undesirable but the world does not stop when there is deflation. In Europe in
particular there is fear for the unknown, while the Japanese did find a way how to deal with, and how to live in a
deflationary environment: they have continued to work, continued to consume and continued to pay tax and adjust to
this situation.
It is quite conceivable that more surveys about deflation and real estate will follow, given the current situation in
Europe, the general expectation that the interest rate will remain low provisionally, as yet the little visible effects of QE
on the level of inflation and hence also the increasing social relevance of this topic.
43
Bibliography
•
Baarda, D.B., e.a. (2012), Basisboek Methode en Technieken: Kwantitatief praktijkgericht onderzoek op
wetenschappelijke basis. Noordhoff Uitgevers.
Barber, C en White M (1995), Property and inflation: The hedging characteristics of Commercial
•
Property, Barber White Property Economics, London
•
Barkham, R. en C. Ward (1996), The inflation-hedging characteristics of UK property, Journal of Property
Finance, Vol 7, nr 1, pag 62-76.
Boonstra, W. en M. Verduijn (2014), Deflatie in de Eurozone, is het waarschijnlijk?, Rabobank, Kennis en
Economisch
Onderzoek,
consulted on 5
april 2015,
https://economie.rabobank.com/publicaties/2014/juli/deflatie-eurozone-is-het-waarschijnlijk/
Brouw P. op den (2014), Innovatie Attaché Netwerk Tokio, Snelle vergrijzing in Japan vraagt om actie,
http://www.rvo.nl/snelle-vergrijzing-japan-vraagt-om-actie
Bruine, M.A. (2009), Vastgoed en inflatie: een analyse van de inflatie hedge kwaliteiten van Nederlands en
internationaal onroerend goed, MSRE scriptie.
Chaplin R. (1996), Unsmoothing commercial property performance indices, Bristol, University of the West of
England.
Chatrath, A en Y. Liang (1998), REITs and inflation : a long-run perspective, Journal of Real Estate Research,
1998, nr 3, pagina's 311-325.
Chu, Y., en T.F. Sing (2004). Inflation Hedging Characteristics of the Chinese Real Estate Market.
Journal of Real Estate Portfolio Management, vol. 10, (2), 145-154.
•
•
•
•
•
•
•
•
•
•
•
Eichholtz, P.M.A. (1997), A Long Run House Price Index, The Herengracht Index, 1628-1973, Real Estate
Economics 1997, V25, pag. 175-192.
e
Eichholtz, P., H. Hilverink en M.A.J. Theebe (2000), Woningen in de pensioenportefeuille, ESB, 85
jaargang, nr. 4275, pagina 812, 13 oktober 2000.
•
Ganesan, S., en Y.H. Chiang (1998). The inflation-hedging characteristics of real and financial assets in Hong
Kong. Journal of Real Estate Portfolio Management, 4, (1), p55, 13p.
•
Geltner, D.M., Miller, N.G., Clayton, J., & Eichholtz, P. (2013). Commercial real estate investment and analysis
(3rd ed.). Mason OH: Thompson South-Western.
Geltner, D.M. , J.D. Fisher (2000), De-lagging the NCREIF index: transaction prices and reverse-
•
engineering, Real estate finance : the quarterly review of commercial finance techniques, pag 7-22.
•
•
•
•
•
•
•
Geurts, T.G. en M.L. Gagne (2003), Real estate in Japan : analyzing an investment decision, Journal of real
estate and education, pag 283-291.
Gool P. van, P. Jager, M. Theebe en R. Weisz (2013), Onroerend goed als belegging, Noordhoff Uitgevers
Groningen.
Fisher I., (1933), The debt deflation theory of great depressions, Econometrica, consulted at 5 april 2015, van
https://fraser.stlouisfed.org/docs/meltzer/fisdeb33.pdf
Hildering J.P. (1999), Effect van deflatie op vastgoed, 20 jaar vastgoedonderzoek in Nederland, 1992- 2011,
pagina 95-99.
Hoesli M.E., G.A. Matysiak, N. Nanthakumaran, B.D. MacGregor (1997), The short term inflation- hedging
characteristics of UK real estate, Journal of real estate finance and economics, 1997, nummer 1, pagina's 2757.
Hori, M. en S. Shimizutani (2005), Price Expectations and Consumption under Deflation: Evidence from
Japanese Household Survey Data, Hitotsubashi University Institute of Economic Research, Discussion Paper
Series No.98
Kawaguchi Y. (2009), Japan’s Real Estate Crisis, What went wrong? Why? What lessons can be learned?,
44
•
Waseda University, pag 1-28.
Koo R. (2008), The Holy Grail of Macroeconomics: Lessons from Japan’s Great Recession.
•
Matysiak G., M. Hoelsi, B. MacGregor en N. Nantakumaran (1996), The long term inflation-hedging
characteristics of UK commercial property, Journal of Property Finance, Vol 7, nr 1, pag 50-61.
•
Mauldin J (2014), The Last Argument of Central Banks: Asset Inflation Versus Price Inflation., article internet,
http://www.valuewalk.com/2014/11/central-banks-asset-vs-price-inflation/99999/
Meltzer H.A. (2001), Monetary Transmission at Low Inflation: Some Clues from Japan in the 1990s, Carnegie
Mellon University, 2-2001.
Miles D. (1996), Property and inflation, Journal of property finance, 1996, nummer 1, Pagina's 21-32.
Pagliari J.P., K.A. Scherer en R.T. Monopoli (2004), Public versus Private Real Estate Equities : A more refined,
long-term comparison, Real Estate Economics, 2005, v.33, P. 147-187.
Haar, van der E. (2009), Sociaal Economisch Plan 2009, Wat leert het ‘verloren decennium’ van Japan over de
huidige crisis?, CPB publication, page 41, having been consulted on 4 April 2015, of
Http://www.cpb.nl/sites/default/files/cep2009_kader_p41.pdf
Wit de, I. (2007), performance and performance drivers in global property markets, Thesis,
Maastricht University.
Yobaccio, E, H.J. Rubens and D.C. Ketcham (1995),The inflation-hedging properties of risk assets: the case of
REITs, Journal of Real estate research, 1995, no 3, pag 279-296
•
•
•
•
•
•
•
45
Appendix
Model 1 : no lag (Rsq: 28%)
Coefficients
Standard Error
t Stat
P-value
Intercept
-0.0008
0.0004
-2.0337
0.0436
vacancy
0.0030
0.0097
0.3137
0.7541
Price level
0.0027
0.0003
7.8635
0.0000
Model 2: 3 months lag vacancy rate and 3 months lag inflation (Rsq: 28%)
Coefficients
Standard Error
t Stat
P-value
Intercept
-0.0006
0.0004
-1.5590
0.1210
vacancy
-0.0186
0.0094
-1.9747
0.0500
Price level
0.0025
0.0004
7.2031
0.0000
Model 3: 3 months lag inflation (Rsq: 27%)
Coefficients
Standard Error
t Stat
P-value
Intercept
-0.0006
0.0004
-1.6774
0.0954
vacancy
-0.0018
0.0097
-0.1809
0.8567
Price level
0.0027
0.0004
7.5170
0.0000
Model 4: 3 months lag vacancy (R sq: 29%)
Coefficients
Standard Error
t Stat
P-value
Intercept
-0.0007
0.0004
-1.9269
0.0558
vacancy
-0.0120
0.0096
-1.2551
0.2113
Price level
0.0026
0.0004
7.3436
0.0000
t Stat
P-value
Model 5: 12 months lag inflation and vacancy rate(R sq: 40%)
Coefficients
Standard Error
Intercept
-0.0002
0.0004
-0.6795
0.4978
vacancy
-0.0841
0.0086
-9.8305
0.0000
Price level
0.0009
0.0004
2.2694
0.0246
Model 6: 12 months lag inflation (R sq: 4%)
Coefficients
Standard Error
t Stat
P-value
Intercept
-0.0005
0.0005
-1.1990
0.2323
vacancy
-0.0092
0.0111
-0.8306
0.4074
Price level
0.0013
0.0005
2.5468
0.0118
46
Model 7: 12 months lag vacancy rate(R sq: 49%)
Coefficients
Standard Error
t Stat
P-value
Intercept
-0.0005
0.0003
-1.5591
0.1210
vacancy
-0.0673
0.0085
-7.8860
0.0000
Price level
0.0018
0.0003
5.6982
0.0000
t Stat
P-value
Model 8: 24 months lag vacancy and inflation (R sq: 1%)
Coefficients
Standard Error
Intercept
-0.0012
0.0005
-2.2259
0.0276
Vacancy
-0.0109
0.0133
-0.8142
0.4168
Price level
-0.0005
0.0006
-0.7654
0.4453
Model 9: 36 months lag vacancy and inflation (R sq: 2%)
Coefficients
Standard Error
t Stat
P-value
Intercept
-0.0010
0.0006
-1.8378
0.0683
Vacancy
-0.0082
0.0139
-0.5913
0.5553
Price level
-0.0009
0.0007
-1.3800
0.1699
47