Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
ECONOMIC GROWTH IN AN ENLARGED EUROPE: THE HUMAN CAPITAL AND R&D DIMENSIONS ALBERTO BUCCI Working Paper n. 2004-22 GIUGNO 2004 JEAN MONNET CHAIR Economics of European Integration Università degli Studi di Milano Dipartimento di Economia Politica e Aziendale Via Conservatorio 7 -- 20122 Milano tel. ++39 02 503 21501 (21522) http://www.economia.unimi.it fax ++39 02 503 21450 (21505) E Mail: [email protected] Third Milan European Economy Workshop, 28-29 Maggio 2004 Pubblicazione depositata presso gli Uffici Stampa della Procura della Repubblica e della Prefettura di Milano ECONOMIC GROWTH IN AN ENLARGED EUROPE: THE HUMAN CAPITAL AND R&D DIMENSIONS* Alberto Bucci† 1. INTRODUCTION During the 1950s and 1960s, many European Union (EU, henceforth) countries have grown relatively faster than the United States (US), leading to a gradual convergence in per-capita income levels across these two economic areas. This convergence process seems to have reversed during the 1990s and the beginning of the XXIst century, when economic growth in the US has been steadily above that observed in the major EU nations. The US growth revival and the related reversal in the convergence process have led to a renewed interest in analysing the relative contributions of policies and institutions to the economic growth performance across countries and over time. In this respect, in July 2002 a high-level Group of independent experts was invited by the European Commission and its President (Mr. Romano Prodi) to review and analyse the entire system of EU economic policies and to advance proposals in order for the EU to become in the next future a dynamic knowledge-based economy with sustainable economic growth, greater social cohesion and higher stability among all the Member States. This Report (An Agenda for a Growing Europe), also called the “Sapir Report” (2004), represents the final output of this working-group. In this paper I will focus on human capital accumulation, the intensity of R&D activity, and the pace of technological progress. These factors are now considered among the most important in explaining the sluggish economic performance (low per-capita income levels and growth rates) of the EU with respect to the US in the last few decades. I will also comment on some of the policies (different from, but complementary to, the ones proposed by the group of experts) that can be pursued by the European Commission in each of these two inter-related areas (R&D and technology, on the one hand, and human capital, on the other) in order to boost economic growth in an enlarged EU, highlighting at the same time some open questions, which probably require further reflection by both the economic profession and the policy-makers. This paper is organized as follows. In the next section I review the main changes in the patterns of growth that have recently occurred in many EU-countries with respect to the US experience. In sections 3 and 4, I move to the analysis of the policies that can be implemented (respectively in the fields of human capital accumulation and R&D) with the objective of increasing the growth rate in an enlarged Europe. As usual, the last section concludes. 2. LEVELS AND GROWTH RATES OF PER-CAPITA GDP OVER SELECTED PERIODS AND COUNTRIES: SOME STYLIZED FACTS Starting from the ‘80s, the pattern of growth of many EU countries has radically changed, leading most of them to have, at the beginning of the XXIst century, levels of per-capita GDP (at current prices and purchasing power parities) similar to those shown at the beginning of the ’70s (and around the 70% of the US level). It is widely maintained that behind this broad picture lies an * I wish to thank, without implicating, G. Bognetti, M. Florio, O. Garavello and A. Missale for comments on an earlier draft of this paper. The usual disclaimer applies. † University of Milan (Department of Economics and Business - Milan, Italy). Contact Address: Department of Economics and Business, University of Milan - via Conservatorio, 7 I-20122 Milan, Italy. Tel.: ++39-(0) 2-50321.463; Fax: ++39-(0) 2-50321.505; E-mail: [email protected]. 1 evident divergence between EU and US performance on employment and labor productivity. The following Table summarises these facts: GDP 1970 per Capita GDP 2000 1970 per Hour Hours 2000 1970 per Capita 2000 US 100 100 100 100 100 100 EU-15 69 70 65 91 101 77 Table 1 : PPP per-capita GDP; PPP GDP per hour and per-capita Hours (total hours worked divided by total population) for two selected years (1970 and 2000): US vs. EU-15 (US=100). Source: based on Blanchard (2004), Table 1, p.4. The first two columns in Table 1 confirm the first fact mentioned above: in terms of per-capita GDP, the gap between the EU-15 and the US has remained roughly constant over the last 30 years (1970-2000). This means that in 2000 the EU-15 countries showed on average a standard of living (per-capita GDP) almost unchanged with respect to the one they had at the beginning of the ’70. However, the evolution of the GDP per Hour and of the Hours per Capita across the two areas over the considered period is evidently different. The second two columns show that labour productivity (GDP per Hour Worked) has increased much faster in Europe than in the US, even though the European labour productivity level was still lower (91%) than the US one in 2000. Finally, the last two columns show that in Europe hours worked per capita have drastically decreased in the past 30 years. Since relative EU labour productivity increased and relative EU hours worked decreased in roughly the same proportion, the EU GDP per-capita level has remained roughly constant with respect to the US in the period 1970-2000 (see Blanchard, 2004, pp. 4-5). Behind these numbers, there are clearly many other relevant differences between EU and US. First of all, differences in growth patterns between the two sides of the ocean conceal differences at the sectoral level, too. In the US, aggregate productivity growth has been mainly driven by hightech manufacturing industries. In Europe, instead, aggregate productivity growth was mostly driven by manufacturing industries with low or medium level technologies and by some service industries such as telecommunications and finance (Nicoletti and Scarpetta, 2003). Secondly, in many European countries the unemployment rate is much higher than in the US, and high unemployment naturally leads low productivity/low skilled workers out of employment. In turn, this tends to increase measured labour productivity in Europe. Lastly, one important determinant of the labour productivity is the state of technology that depends in an important way on the R&D intensity of a country and the “quality” of its education system. All this points to the existing differences in (R&D and education) institutions between the two economic systems as major factors potentially able to explain why the gap between the EU and the US per-capita GDP has remained quite constant over the last thirty years.1 The following table summarises the average trend growth rates in GDP per capita for the United States and the EU-15 countries over two selected periods: 1 In this paper I will abstract from differences in the labor market institutions between the EU and the US, leaving to other contributions in this special issue of the journal such analysis. 2 COUNTRIES 1980/1995 1995/2002 Austria Belgium Luxembourg Netherlands 2.0 1.9 4.2 1.7 1.9 2.0 3.7 2.2 Change between first and second period -0.1 +0.1 -0.5 +0.5 Denmark Finland Sweden 1.8 1.6 1.3 2.0 3.5 2.5 +0.2 +1.9 +1.2 France Germany(a) 1.5 1.6 1.9 1.4 +0.4 -0.2 Greece Italy Portugal Spain 0.6 2.0 3.1 2.3 3.0 1.5 2.5 2.7 +2.4 -0.5 -0.6 +0.4 Ireland United Kingdom 3.8 2.1 7.1 2.2 +3.3 +0.1 EU-15 US 1.8 2.0 2.0 2.3 +0.2 +0.3 Czech Republic 2.0 Hungary 3.9 Poland 4.1 Slovak Republic 1.7 Table 2 : Average Trend Growth Rates in GDP per capita over selected countries in two time periods: 1980-1995 and 1995-2002. Source: based on OECD (2003a), Economic Outlook, No.73, Ch.5, May 2003 (Table V.I.). Note: a) West Germany before 1991. For 1980-1995 average excludes 1991. The Table above clearly shows that the convergence process appears to have reversed between the 1980s and the 1990s with trend GDP per capita growing faster in the US than in the largest EU member countries. In more detail, from the last two columns of Table 2 we may infer that after 1995 the convergence process towards the US GDP per capita was particularly successful for some smaller economies (in particular for Finland, Greece, Ireland and Sweden), where economic growth accelerated very much and has been steadily higher than the US rate, and particularly unsuccessful for other countries (mainly Italy, Portugal and Germany). From this discussion we may draw the following stylised facts: STYLISED FACT #1: Over the 1970-2000 period, the EU-15 countries did not do much progress in catching-up with the level of the US GDP per capita. In 2000 the PPP per-capita GDP in the EU-15 was still at the 70% of the US PPP per-capita GDP in the same year. The situation was not that 3 different thirty years before (in 1970). However, the relative contribution to the convergence process to the US per-capita GDP levels has not been homogeneous across the EU-15 member States. Indeed… STYLISED FACT #2: …In the last few years (1995-2002) some small EU-15 countries have shown average trend growth rates in GDP per capita much higher than the US (in particular Finland, Greece, Ireland and Luxembourg). At the same time, most of the major EU-15 countries (France and, in particular, Germany and Italy) have greatly under-performed with respect to the United States. STYLISED FACT #3: In some of the ten new EU member-countries the trend growth rate of per-capita GDP has been higher than the EU-15 average in the last few years (see Table 2 above for Hungary and Poland), even though in these countries the average per-capita income level is still less than half the EU-15 average (Sapir et al., 2004, p.186 and Table 8.1). All these three stylised facts together call for policy changes in order to stimulate growth and convergence and to increase per-capita GDP levels in an enlarged Europe. In the remainder of this paper I will focus on two important areas of intervention for European policy-makers: human capital and R&D. 3. POLICIES IN FAVOUR OF (HIGHER) HUMAN CAPITAL ACCUMULATION IN AN ENLARGED EUROPE. Considered in a broad sense, human capital accumulation is one of the most important determinants of individuals’ earning capacity, and then plays a relevant role in affecting the level of per-capita income. At the same time, because human capital accumulation improves the “quality” of the labor force, it also influences output growth. Recent empirical estimates (Bassanini and Scarpetta, 2001) suggest that one extra year of average education (roughly equivalent to a 10 per cent rise in human capital2 ) has in the past raised the long-run output per capita by around 4 to 7 per cent on average across OECD countries. This seems to imply that persistent differences in the accumulation of human capital are important sources of differences in growth paths across countries. In this paragraph I will focus on the role of tertiary education in growth. The reason is twofold. First of all, one of the two strategic economic goals set by the European Union for the decade ending in 2010 is to become the most dynamic knowledge-based economy with sustainable economic growth (Lisbon Agenda). Tertiary human capital formation may facilitate the achievement of this goal since it is likely to ease the adoption of new technologies and/or the process of innovation, leading to an acceleration of technical progress and economic growth. Secondly, it is widely recognized that primary and secondary-level schooling (compulsory education) brings benefits to the society as a whole, whereas the returns from tertiary education generally accrue to the single individual. As shown in Blöndal et al. (2002), this fact raises a lot of 2 Given the absence of direct measures, human capital is usually assessed in terms of educational attainment. In this respect, two among the most important human capital indicators are respectively the average number of years of education and the percentage of population that has reached a certain level of education. See Wößmann (2003) for a detailed survey about the different measures of human capital most often employed in empirical studies of economic growth. 4 issues and policy questions (such as the appropriate role of government in the provision of such education) that it is more difficult to raise for compulsory education. Concerning tertiary education, the Sapir Report recommends increasing public spending on this form of human capital: “…At the moment, in Europe we invest 1.4% of our GDP in tertiary education. The EU would need to invest much more […] in order to obtain the level of knowledge required to reach a higher growth path. Targets of close to 3% […] seem appropriate in view of international comparisons” (the Sapir Report, 2004, p.244). The main objective of this paragraph is to show that other human capital-based policies (different from simply increasing public spending) may also be needed for old (and new) EU members to grow higher and to enjoy better standards of living in the future. In order to make this point, the following two tables (Tables 3 and 4, respectively) show public spending on tertiary education (as a percentage of total GDP) and the private internal rates of return to tertiary education for men and women in the largest EU countries in recent years. COUNTRIES Austria Belgium Netherlands PUBLIC EXPENDITURE ON TERTIARY EDUCATION (% GDP - year 1999) 1.7 1.5 1.3 Denmark Finland Sweden 2.4 2.1 2.1 France Germany 1.1 1.1 Greece Italy Portugal Spain 1.1 0.8 1.0 0.9 Ireland UK 1.2 1.1 EU-Countries Mean 1.39 United States 1.4 Table 3 : Public Spending on Tertiary Education as a percentage of GDP in the EU and the US (1999). Source: based on OECD (2002a), Education at a Glance. The Table above shows that there are EU-countries (such as Austria, Denmark, Finland and Sweden) that in the recent past have spent (as a percentage of their own total GDP) much more than the US on tertiary education (notice that in 1999 the countries that still lagged behind in this respect in Europe were Italy, Spain and Portugal). Moreover, the same Table shows that on average the US do not seem to have recently over-spent in tertiary education with respect to the average of EU5 countries. These two facts together suggest that the better performance of the US in terms of educational attainment of the population aged 25-64 years3 is to be explained by the higher private incentives to accumulate tertiary human capital in that country than in Europe. The following table takes a further step in this direction and reports the private internal rates of return to tertiary education for the United States and some European countries over the period 1999-2000.4 Earnings differentials and the length of studies A. TERTIARY EDUCATION - MEN (%), 1999/2000 US Germany France Italya UK Denmark NLb 18.9 7.1 13.3 8.0c 18.1 7.9 11.7 Sweden 9.4 Taxes -2.3 -1.5 -1.6 - -2.1 -2.1 -2.0 -1.5 Unemployment Risk 0.9 1.1 2.4 0.3 1.6 1.0 0.0 1.2 Tuition Fees -4.7 -0.3 -1.1 -0.8 -2.7 -0.1 -0.6 -0.7 Public Student Support 2.1 2.7 1.3 0.0 3.6 4.8 2.9 3.0 Comprehensive Rate 14.9 9.1 14.3 7.5 18.5 11.5 12.0 11.4 3 It is estimated that 37.3% of the US population aged 25-64 years possessed a tertiary education degree in 2000, against a EU-15 average of 23.8% (in Italy, in the same year this percentage was 10.0%). See the Sapir Report (p.60, Table 4.4). 4 Unfortunately, I did not succeed in finding comparable data on the private internal rates of return to university education for Austria, Belgium, Finland, Greece, Ireland, Portugal and Spain. 6 Earnings differentials and the length of studies B. TERTIARY EDUCATION - WOMEN (%), 1999/2000 US Germany France Italy UK Denmark NLb 18.8 7.0 12.1 16.4 6.0 9.4 Sweden 7.4 Taxes -2.0 -1.6 -1.7 - -2.3 -1.1 -1.0 -0.7 Unemployment Risk 1.4 0.6 4.8 - 1.3 0.7 0.7 1.6 Tuition Fees -6.0 -0.6 -1.7 - -2.5 -0.1 -0.7 -0.8 Public Student Support 2.7 3.0 1.9 - 3.2 5.6 4.1 3.3 Comprehensive 14.9 8.4 15.4 16.1 11.1 12.5 10.8 Rate Table 4 : Private Internal Rates of Return to Tertiary Education for Men and Women in some European Countries and the US (1999/2000). Source: based on Blöndal et al. (2002), p. 23, Table 3. Note: The rates of return are calculated by comparing t he benefits and costs with those of upper secondary education. In Italy reliable data on earnings for women were not available. a) 1998. b) 1997. c) After-tax earnings. The private internal rate of return to tertiary education is defined as the “discount rate that equalises the real costs of tertiary education during the period of study to the real gains from education thereafter” (Blöndal et al., 2002, p.21). The costs include tuition fees, foregone earnings net of taxes and adjusted for the probability of being in employment, minus the resources made available to students in the form of grants and loans. The benefits are the gains in post-tax earnings adjusted for higher employment probability minus the repayment, if any, of public support during the period of study. The calculations assume that the student is in full-time education and has no working activity, and hence no earnings while studying. Finally, the probability of course drop-out is not taken into account. Then, the reported internal rates of return are conditional on successful completion of the relevant education programmes.5 In sum, the Table above highlights the fact that private incentives to invest in tertiary education are affected by a number of policy factors: 1) First of all we find the differences in earnings between tertiary-level and lower-level human capital and the length of study periods . Clearly, the larger the wage gap between workers with tertiary education degrees and workers with lower than tertiary education degrees, the higher the return from tertiary education, the higher the incentive for people to invest in tertiary human capital. At the same time, very long study periods will tend to discourage investment in tertiary education; 2) The second policy factor that enters into the computation of the internal rate of return is the tax system. A progressive income tax system discourages higher education as it implies an 5 See Blöndal et al. (2002), p.22 for a more detailed discussion on the assumptions made and the methodology used for the construction of this index. 7 (implicit) tax on human capital. Indeed, by taxing the earnings of the better-educated at a higher rate than the one applied to the earnings of the less-educated, the after-tax earnings differential of the two categories of workers mentioned above shrinks and the gains from higher human capital investment lowers; 3) Thirdly, we have unemployment risk. As long as the risk of remaining unemployed is lower for a worker with a tertiary education degree, he/she will have a major incentive to invest in higher education. In other words, the higher the ex-post probability of being employed, the higher the return accruing to an individual upon the completion of his/her tertiary-level studies, the higher the ex-ante incentive to invest in tertiary education; 4) Moreover, we have the tuition fees. In most countries these fees do represent the major cost component at the tertiary education level; 5) Finally, we have public student support. Many countries support educational activity by offering subsidies to individuals during their tertiary studies in the form of grants and personal loans. The Table above conveys information that may be relevant for a more effective human capitalbased economic policy at the (enlarged) EU level. First of all, from the Table we notice that the private real internal rates of return to university education differ across countries. The country with the highest return (both for men and women) is the UK. However, the United States, France, Denmark, the Netherlands and Sweden are also characterised by relatively high internal rates of return. Germany and Italy (for men) represent, across those considered, the countries with the lowest rates of return to higher education and, hence, the places where it seems to be less rewarding to invest in tertiary human capital (at least from a private point of view). In the remainder of this section I use some of the main conclusions one may draw from Table 4 in order to highlight alternative policies (different from those suggested in the Sapir Report) that could be pursued in favour of higher education in an enlarged European Union. HC Policy 1): Reduce the length of tertiary education programmes. In all the considered countries, the prime determinant (in terms of relative importance) of the internal private rates of return to tertiary education is represented by the earnings differentials (to be exploited upon completion of the university studies) and the length of education programmes, jointly considered. This is true both for men and women. Apart from differences in earnings differentials across countries (that could themselves be due to differences in the respective labor market institutions and, as such, are outside the scope of this paper), one policy prescription that seems to arise from Table 4 would be to reduce the length of university degrees in order to increase the financial rewards from (tertiary) human capital accumulation. The following Table seems to support the belief that private returns to tertiary education are higher in those countries where the theoretical (or legal) length of degrees is shorter: 8 Countries Netherlands Theoretical (or legal) length 4 Typical length 5-6 Denmark Sweden 3 3-4 3 - France Germany Italy 3 5 4-6 3-4 6.5 4-6 UK 3-4 4 US 4 Table 5 : The length of standard first-degree tertiary education (in years). Source: based on OECD (1999a). 5 In 1999 in US, UK, France and the Netherlands (some of those countries with the highest returns to tertiary education), the theoretical (legal) length of first-degree university education did not exceed four years. It was on average equal to five years in Germany and Italy (where the returns to tertiary education were the lowest). Moreover, in Germany (and to a lesser extent in the Netherlands) there exists a quite large gap between the legal and the typical length of first-degree tertiary programmes. In the last few years this legal length has been reduced in Italy (where now it is generally equal to three years). But many concerns arise for the new EU-members, where “…the university curricula still need updating and modernization” (Sapir Report, 2004, p. 192). To which extent can in these countries a reduction in the length of first-degree tertiary education programmes (due to harmonization of undergraduate degrees) be pursued without compromising their “quality”? Actually, the EU policy-maker seems to face a key trade-off. On the one hand, extending the period of formal education brings about an important dividend in terms of long-run output per capita growth (Bassanini and Scarpetta, 2001); on the other, instead, this policy could seriously harm private incentives to accumulate (especially tertiary) human capital. With this possible trade-off in mind, the policy action that the European Commission will decide to take in this field in the near future is going to be crucial. HC Policy 2): Boost public financial support to students involved in tertiary education programmes. Although private tuition costs at the tertiary level tend to be low in all European countries (except for UK), public financial support to students in the form of grants and favourable loan arrangements are particularly generous in some English-speaking and Nordic countries. Table 4 suggests that this form of public support gives (both for men and women) a significant boost to private incentives to tertiary education, especially in Denmark, UK, Sweden, Netherlands and Germany. Indeed, in these countries the public support to students at the tertiary level represents the second most important determinant of the private return to university education. It is disappointing that for Italy (men) the impact of this policy factor is completely absent, whereas it is weak in France. Since public financial assistance to individuals during their tertiary studies seems to have relevant and positive effects on private incentives to acquire higher education, a harmonization of the public intervention across EU countries in this area could in principle generate huge benefits. At the same time, as long as students and their families face liquidity constraints in financing their own tertiary education, the amount of loans and grants publicly made available to them could represent a significant incentive mechanism. This could be particularly true for the ten new EU accession 9 members, where the per-capita GDP is markedly lower than the EU-15 average and the presence of liquidity constraints is likely to play a major role in the private decisions to invest in (tertiary) human capital due to the existence of a less-developed banking system. HC Policy 3): Boost investment in adult tertiary education. Acquiring tertiary human capital does not bring about the same return at all ages. The reason is threefold. First of all, wages are likely to increase with age and, as a consequence, the opportunity cost of foregone earnings will be higher for older adults. Secondly, often adults do not have access to public financial support. In some countries, e.g. Germany and Sweden, there exist age limits to apply for student grants and loans (see Blöndal et al., 2002, p. 30). This can substantially contribute to a further increase of the adults’ costs of tertiary education. Thirdly, the shorter the remaining length of the working life, the shorter the period over which the benefits from higher education can be reaped and the lower the incentives (and, hence, the return) to acquiring formal tertiary education by an older worker with respect to his/her younger counterpart. The next Table reports stylized private internal rates of return to standard first-degree university education for a male starting studies at the ages of 40, 45 and 50 respectively. This table is directly comparable with Table 4 above: Countries Sweden Age 40 3.9 Age 45 0.6 Age 50 -7.5 France Germany Italy 7.3 -1.5 0.4 1.9 -9.7 -4.1 -11.4 -23.0 -21.6 UK 11.1 8.8 5.5 US 8.9 6.7 3.5 Table 6 : Stylised private internal rates of return to tertiary education for older adults (men) in selected countries. Source: based on Blöndal et al. (2002), p. 31, Table 5. Note: The internal rates of return to tertiary education are calculated by comparing the benefits and costs with those of upper-secondary education. All the calculations assume that the wage premia at the end of study are identical to those received by a young male finishing his degree as a part of initial education, and that they evolve over time in line with those for a young graduate. The length of study is assumed to be identical to that for young persons, and adult students are assumed to have no public grants or loans. According to this Table: a) private incentives to invest in tertiary human capital diminish quickly with age; b) by the age of 40, the internal rate of return to tertiary education is very low in most of the considered countries (except for UK, US and France) and is negative in Germany (largely because of long education periods and a low average retirement age). In Italy this return is almost zero; c) by the age of 50, the return to tertiary education is positive only in the US and the UK, whereas it is (strongly) negative elsewhere. Adults rarely participate to formal tertiary education programmes. Table 6 suggests that this can be mainly due to the fact that, after a threshold age (that can vary across countries), adult men do not have much incentive to engage in this kind of activity. In this respect, many policies can be adopted (at the single country or the EU level) in order either to postpone this threshold age or to increase adults’ motivation to acquire higher education. A first group of policies could be aimed at reducing the total cost of acquiring tertiary education for older adults, through, for example, an easier access to public grants. On the other hand, a second set of policies that can be considered are 10 those that increase the total benefits accruing to older adults from the participation to formal tertiary education programmes. Among these, one may think for example of pension, health and family policies that can jointly contribute to influence both the length of people’s (working) life and their facility to face jointly family and work responsibilities. Such policies are already heterogeneous across EU-15 countries and are going to be much more heterogeneous across EU-25 countries. The role the European Commission can play in the future in order to boost adult individuals’ participation to formal tertiary education programmes still remains an open policy question. Again, the problem seems to be particularly relevant for the ten new accession EU member countries, characterised by large masses of workers with lower educational levels than the rest of the Union. HC Policy 4): Boost investment in training activity by employers and firms. Human capital accumulation is not limited to compulsory or post-compulsory education. Training activity by firms may also play an important role. The available empirical evidence indicates that, through such activity, it is possible to obtain a substantial increase of productivity in the firms and sectors involved. Dearden et al. (2000), using industry-level data for the UK, conclude that a 5 percent increase in training incidence brings about an increase in the level of labour productivity by 4 percent. They also show that both employers and employees win from more intensive training activity. Indeed, labour productivity gains are larger than wage gains, so that, after training has taken place, employers’ profit is higher. Notwithstanding the positive effect exerted by training on wages and profits, an employer survey for the EU countries in the mid1990s6 suggests that the average time spent on training by those who receive any training at all amounted to one to two weeks per year in most countries reviewed, the equivalent of 2 to 4 percent of average annual hours worked. This may represent a sub-optimal investment in training (especially if compared with the potential ex-post wage and profit benefits mentioned above). Many scholars believe that the main reason behind this under-investment is to be found in the existence of opportunism problems between employers and employees. Since at least part of training is general in nature, it can be used by employees in other firms and sectors that pay higher wages. In order to avoid such a problem, employers keep artificially low the supply of training activity in their own firms/sectors. From this we infer that there exists a clear failure in training activity market and, as a consequence, there is sufficient room for active intervention by governments and public authorities in this field. Policies aimed at reducing the cost of training for firms and increase the incentive to be trained for employees already exist in some European countries (such as France, Belgium and Denmark). What still needs to be understood is to which extent these policies will represent a priority for the European Commission even in the framework of an enlarged Europe. HC Policy 5): Skill-biased technological progress: school-to-work transition and organizational change at the firm level. Which role for the EU policy? Investment in human capital (both through formal education and training-on-the-job activities) is essential in the context of rapid technological change. There is now large evidence suggesting that in the last few years technical progress has become more and more skill-biased, in the sense that it has considerably increased the demand for “knowledge-intensive” employment (scientists, engineers, ICT specialists, etc.). This evidence is summarised in the following table: 6 See OECD (1999b). 11 Workers’ Type Employment Growth Knowledge workers +3.3% Service workers +2.2% Management workers +1.6% Data workers +0.9% Goods-producing workers -0.2% Table 7 : Employment growth by group of occupations - Average annual percentage change, 1992-1999. Source: OECD (2001), p. 14, Figure 5. As can be seen from Table 7, in recent years, and especially due to diffusion of Information and Communication Technologies (ICTs), the demand for a particular type of workers (knowledge workers) has grown much faster than that for other types of workers in OECD countries. This pattern in the employment growth calls for two main policy issues. The first one has to do with the school-to-work transition process and the policy question is: how should responsibilities in the education of the prospective “knowledge workers” be shared between schools, trainees and employers? The second issue concerns the design of the higher education system and the organization of firms. Indeed, there is no doubt that the new technologies are changing very fast both the nature of the links between universities and the labour market and the internal organization of firms (with more and more emphasis being placed on concepts such as that of teamwork, among others). How should the structure and the content of university courses be changed in order to guarantee an easier access of students to the new professions? Moreover, how can the EU policy encourage the process of work re-organization at the firm level requested by the diffusion of new technologies? Things are increasingly going to become more difficult in the light of the enlargement process and for this reason these issues deserve a careful analysis by European policy-makers. 4. POLICIES IN FAVOUR OF R&D AND INNOVATION IN AN ENLARGED EUROPE. It is widely recognized that R&D activity is an important engine of economic growth in the long run. In this connection, it is estimated (OECD, 2003b, p.85, Table 2.6.) that the elasticity of the output growth rate with respect to total R&D intensity (total R&D spending/GDP) is 0.14 in a regression that covers 14 to 17 OECD countries (depending on the specification) over the period 1981-1998. The next table presents some data about government spending on R&D as a share of total public spending in some selected countries in two years (1985 and 1995): 12 Countries 1985 1995 Austria Belgium Netherlands 1.2 0.9 1.8 1.4 … … Change between 1985 and 1995 +0.2 ... … Denmark Sweden 1.2 1.7 1.2 1.7 0 0 France Germany 2.3 2.2 1.8a 1.8 -0.5 -0.4 Italy Portugal Spain 1.2 0.5c 0.7 1.0 0.9b 0.9 -0.2 +0.4 +0.2 Ireland UK 0.8 2.0 0.8a 1.5 0 -0.5 US 4.1 2.8 -1.3 Table 8 : Government spending on R&D as a share of total public spending: selected years and countries. Source: OECD (2003b), p. 69, Table 2.2. Notes: a) 1993 instead of 1995. b) 1992 instead of 1995. c) 1984 instead of 1985. The table yields two major stylised facts: STYLISED FACT #1: Over the period 1985-1995, as a share of total public spending, government expenditure on R&D has increased solely in smaller EUcountries (Austria) and/or those countries (notably Portugal and Spain) that in 1985 devoted less resources than any other to public R&D; STYLISED FACT #2: Except for Denmark, Ireland and Sweden (where it remained roughly constant), the share of public R&D on the total government spending has decreased in all the other considered countries, with the most evident reduction having occurred in the US. Indeed, behind the second “stylised fact” reported above there is a clear economic reason. Park (1995) has already showed that most of the variability of growth rates across OECD countries is due to variability of private sector R&D expenditure. This is also confirmed by a recent finding of OECD (2003b, p.84), according to which it is mainly business-performed R&D (rather than the R&D carried out by other institutions, above all public research institutes) to account for the positive correlation between total R&D intensity and output growth. 13 Notwithstanding the fact that it is primarily private R&D to boost output growth,7 the potential benefits from product/process innovation may not accrue fully to the innovators themselves, due to the existence of important spillover effects.8 The presence of such spillovers implies that, without policy intervention, the private sector would sub-optimally spend on R&D activity. As a consequence, some public (indirect) involvement in R&D continues to seem somewhat desirable. There are many ways through which governments can (indirectly) contribute to an increase of the private sector innovation. In this respect, and in the remainder of this section, I will focus on two of the main policy instruments that can be put into use: 1) product market competition and regulation; 2) protection of intellectual property rights (patents). R&D Policy 1): Boost product market competition, but let the enforcement of EU competition policy and sectoral regulation be centrally driven. Many academics now believe that one possible reason for the poor recent performance (low productivity levels and growth rates) of many EU countries (especially if compared with US accomplishment) is to be found in the lack of competition in many product and factor input markets.9 As an example we can mention the case of ICTs. Insufficiently low levels of competition in the product market have led many industries in the EU countries to delay further the adoption of such technologies in the recent past. On the contrary, the US have benefited most from the ICT sector, since in the 1980s it had a stronger product market regulation and competition legal system. Textbook theory suggests that there are at least two fundamental channels through which product market competition (PMC) may influence economic efficiency. On the one hand, more intense PMC results in less slack in the use of inputs and better resource allocation. This should lead to higher total factor productivity growth rates in the long run. On the other hand, since Schumpeter (1942) it is argued that more competition, by eroding the monopolistic rents that can be appropriated by the successful innovator, induces fewer incentives to R&D activity, so harming technological progress and economic growth in the future. The composition of these two effects (the positive static resource allocation effect and the negative dynamic Schumpeterian or profit incentive effect) would seem to imply that the relationship between PMC and aggregate productivity growth might be inverse U-shaped.10 For low initial levels of competition, more PMC is beneficial to growth since it allows a substantial better use of resources, without hampering that much innovation incentives (the resource allocation effect outweighs the profit incentive effect and the correlation between competition and growth would be positive). On the other hand, when PMC is too tough, more competition is likely to reduce drastically technological progress, improving only marginally the allocation of resources across economic activities (the profit incentive effect prevails over the resource allocation effect and the correlation between competition and growth would be negative). Indeed, there exists some evidence that, for given level of protection of intellectual property rights, the relationship between product market competition, on the one hand, and innovation and productivity growth, on the other, might be bell-shaped (Aghion et al., 2002). The 7 Bassanini and Scarpetta (2001) find that a 0.1 percentage point increase in the share of business-sector R&D spending in GDP, at a minimum, increases the level of per-capita GDP by about 1.25% in the long run. 8 According to Griliches (1980) there exist two main kinds of positive spillovers: 1) rent spillovers and 2) knowledge spillovers. On the contrary, Jones and Williams (1998) maintain that there are at least three possible types of negative spillovers from R&D activity: 1) the fishing out effect; 2) congestion externalities and 3) the creative destruction effect. 9 See, among others, Baily and Gersbach (1995) and Borsch-Supan (1998). A recent paper by Bayoumi et al. (2004) confirms this belief. Using a general-equilibrium simulation model featuring nominal rigidities and monopolistic competition in product and labor markets, these authors find that greater competition produces large effects on macroeconomic performance. In particular, they find that differences in competition can account for over half of the current gap in GDP per capita between the euro area countries and the US and that structural policies aimed to increase competition toward US levels may increase output in the euro area by about 12.5%. 10 See Bucci (2004) for a Romerian-flavor model where these two effects (the resource allocation and the profit incentive effects) are combined in order to obtain a bell-shaped relationship between product market competition and aggregate economic growth. 14 data used by Aghion et al. (2002) come from Datastream and include all firms quoted on the London Stock Exchange between 1968 and 1997. Product market competition is measured by one minus the Lerner index (ratio of operating profits minus financial costs over sales), controlling for capital depreciation, advertising expenditures, and firm size. The long time series on firms in each industry allows the authors to control for industry level effects as well as common time effects. The inverted-U relationship between PMC and innovation is found to be robust to many alternative specifications and remains true in the data for many individual industries. This means that if competition is too tough, firms will never be able to recoup the innovation costs, so productivity growth (both at the single firm and sector level) may stagnate. On the other hand, in the absence of any risk of being replaced by outside potential competitors, a pure monopolist may also have no incentive to innovate, and again productivity growth may stagnate. The existence of an (empirically based) bell-shaped relationship between product market competition and innovation (and growth) leads one to conclude that: 1) innovation and growth could be highest at an intermediate level of competition; 2) this growth-maximizing level of product market competition could differ across sectors. This is where the recommendation that the enforcement of the EU competition policy and the regulation of specific sectors be assigned to independent European Authorities is of the highest relevance.11 This suggestion becomes imperative in the view of an enlarged Europe, with some of the new accession member countries having no or less experience in competition/regulation policy matters. R&D Policy 2): Harmonize efforts in product market regulation policies across EU countries in order to catch-up with the US level of market liberalization. Competition policy (the elimination of anti-competitive agreements or abuse of dominant position; the control of mergers between firms and the monitoring of state-aids) represents only one specific component of product market regulation policy. In turn, this includes, among others, 1) privatisation policies; 2) those policies aimed at liberalizing potentially competitive markets; and 3) policies that regulate natural monopoly markets. In principle, all of these public policies are able to influence private incentives to invest in R&D. Accordingly, they may play a strategic role within the project of boosting economic growth in the dynamic, knowledge-based EU economic system. Unfortunately, in the last few years countries did not reform at the same pace12 and the variance of product market regulation approaches has increased across most of them. Nowadays it is maintained that the divergence in such policies is especially large within the EU, as the following table clearly demonstrates: 11 According to the Sapir Report (2004, pp. 298-99): “…Responsibility for enforcing EU competition policy (except state aids) should be assigned to a European Competition Authority accountable to Council and Parliament, whose decisions could be subject to formal override by the Commission. Sectoral regulation should also be assigned to independent authorities”. 12 “…Countries that moved early to liberalize telecommunications have much lower communication costs and a wider diffusion of ICT than countries that were late to take action. Telecommunication monopolies have almost disappeared in the OECD area, but incumbent firms are still dominant in many countries, which contributes to keeping costs high” (OECD, 2001, p.9). 15 Countries Netherlands 1975 4.4 BE 1990 5.2 France Germany 6.0 5.3 5.1 4.3 1998 2.3 3.3 1.9 1975 5.6 PO 1990 5.6 1998 4.0 6.0 4.6 5.8 3.9 4.9 3.0 U.S. 5.5 2.4 1.5 1.7 1.5 1.5 Table 9 : Indexes of product market regulation in France, Germany, the Netherlands and the US over the period 1975-1998. Source: Blanchard (2004), Table 4, p. 19 (based on Nicoletti and Scarpetta, 2003). Notes: BE= “Barriers to Entrepreneurship”. PO= “Public Ownership”. Each index ranges from 0 (no barriers/ no public ownership) to 6 (highest barriers/ maximum public ownership). Table 9 is based on a data set (being described in Nicoletti and Scarpetta, 2003) that gives an overview of the evolution of two synthetic indexes of regulation in seven sectors from 1975 to 1998 for the US and other OECD countries. The first of such indexes is called “barriers to entrepreneurship”, and the second is called “public ownership”.13 Table 9 allows us to draw three major conclusions: 1) product market regulation has steadily decreased in Europe (especially during the ‘90s); 2) Europe has been (and still is on average) more regulated than the US; 3) the regulation framework continues to be heterogeneous across countries. In 1998 the barriers to entrepreneurship were very high in France (3.3) and low in Germany (1.9), where they were almost at the same level of US (1.5). At the same time, the presence of the State in the economy was still high in France and the Netherlands and lower in Germany.14 Since the array of industry-specific regulations that may affect product market competition, innovation and growth is large enough, in the next future it would be optimal to let (de) regulation policy be driven by the European Commission itself, in order to guarantee the maximum level of harmonization of the business environment at the industry level across the entire Single Market. However, it remains the problem of establishing in which specific sectors the Commission should concentrate more its own (de) regulation efforts at the European level in the next future. R&D Policy 3): Patents are a powerful growth engine and their analysis deserves further research effort and policy attention. A patent is an exclusive right to exploit (use, sell, or import) an invention over a limited period of time within the country where the application is made. The three most important characteristics that a new idea must have in order to be patented are novelty (it has to be original), non-obviousness 13 According to Nicoletti and Scarpetta (2003, p. 26): “…In building the indicators, we focus on the two main areas of regulation that are likely to have an impact on governance and/or product market competition: state control and barriers to entry. By state control we mean provisions that aim at establishing partial or full control over resources or economic activities that could, in principle, be managed by private agents. Examples include public ownership and/or control. …By entry barriers we mean provisions that create barriers to entrepreneurship in domestic markets where fixed costs, technology and demand conditions make competition viable. …Examples include: laws or regulations limiting the number of competitors or providing an unfair advantage to some of them (e.g. antitrust exemptions); structural arrangements that make it difficult for competitors to access fixed networks (e.g. vertical integration); regulatory and administrative burdens that impose fixed costs on businesses; and policies that create impediments to international trade and investment (such as foreign investment restrictions and tariff and non-tariff barriers). 14 In 1998, two other EU countries being notably known as very much regulated were Italy and Greece (Nicoletti and Scarpetta, 2003, p. 34). 16 (it has to be innovative) and usefulness (it must have an industrial application). Since Schumpeter (1942), the need to provide the successful innovator with some form of patent protection in order to stimulate investment in R&D has been widely recognized. As patent protection helps innovators to benefit from their research efforts, it creates or increases private incentives to innovate. It also seeks to avoid duplication of research efforts and to promote technological progress. On the other hand, however, patents have an important drawback: they reduce for a given period of time product market competition and the speed of technology diffusion. Notwithstanding the crucial influence patents may have on social welfare, technological progress and economic growth, it is surprising that recommendations on how to reform the patent regime in the future Europe are almost entirely absent from the Sapir Report (2004). Before turning to patent policy issues, it is worth giving a synthetic outline of recent trends of patenting activity across sectors and countries in Europe.15 Between 1992 and 2002 there has been a doubling of the number of patent applications at EPO (the European Patent Office). Among the EU countries, the leading ones in patenting activity over the second half of 90’s have been Germany, Finland and Sweden. Across sectors, the most dynamic in patenting were biotechnology and ICT. The share of biotechnology in EPO filings climbed from 4.3% in 1994 to 5.5 in 2001. During the same period the share of ICT climbed from 28% to 35%. The increase in the EPO share of the countries mentioned above (Germany, Finland and Sweden) can be traced essentially to ICT patents. The ratio of patents to business-funded R&D (patents per dollar of R&D) has increased by 50% for European patentees at the EPO between 1994 and 2000 (it is estimated that such an increase was primarily due to Germany as a country and to ICT as an industry). All these numbers confirm that the surge in patenting activity in Europe is mainly due to invention and further development of new technology areas.16 At the same time, they call for a new patent policy in an enlarged Europe. There are at least four dimensions of a patent system that policy-makers could potentially use in order to enhance innovation and economic growth: • Patent Subject Matter: is the domain of knowledge that can be patented, once a new idea has met the criteria of novelty, non-obviousness and usefulness. Scientific discoveries and/or too abstract ideas are, for example, generally not patentable. A precise definition of the patent subject matter may help to understand which discoveries really improve a society’s welfare and, as such, merit (public) attention; • Patentability Requirement: is a minimum innovation size required to receive a patent. A too low patentability requirement increases the risk of having many patented inventions with a very little social value. On the other hand, a too high patentability requirement increases the risk of discouraging those innovations that, although not radical, may importantly contribute to advance technological progress; • Lagging/Leading Patent Breadth: Lagging breadth limits imitation by specifying inferior products that other firms cannot produce. Leading breadth limits future innovators by specifying superior products that other firms cannot produce. Both too broad and too narrow patents may discourage innovation; • Patent Life: is the length of time for which a patent is valid. On the one hand, patent lifetime should be long enough to provide sufficient incentives to innovation; on the other hand, it should be short enough to allow for a proper diffusion of fundamental knowledge. There is no simple solution to this delicate balance.17 15 See OECD (2004) for major details. A similar pattern can also be documented for the United States, where the number of software-related patents has grown much faster than total patents granted in the near past, and now accounts for between 4% and 10% of all patents (OECD, 2002b, p. 6). 17 In Bucci and Saglam (2000) an attempt in this direction is provided. 16 17 Other two legal aspects of a patent that have a potential bearing on private incentives to innovate (and to patent the innovation) concern respectively the enforcement of the exclusive rights conferred to patent holders and the amount of damages attributed by courts in case of infringement. All together, all these characteristics of a patent define the strength of a patent system. Despite the efforts towards harmonization that have been made in recent years across countries (longer patent lifetime and expansion of the subject matter of patents, for example), the US and the European patent regimes continue to display important differences in two respects. First of all, since the creation of a centralised court system in 1982 (the Court of Appeal of the Federal Circuit, CAFC), rights of patent holders are more frequently and strongly enforced in court in the United States, whereas the implementation of a centralised patent litigation system is still under discussion in Europe. At the same time, damage awards in patent litigation trials have substantially increased in recent years in the US. Secondly, the US patent system still appears as more flexible than the European one, as it allows “…the final grant to be different (usually narrower) than from the initial application” (OECD, 2004, p.18). All this points to the need of increasing the harmonization efforts between these two international patent systems. Harmonization has to go especially in the direction of improving consistency of patent regimes and practices (evaluation procedures, standards and enforcement). Simultaneously, serious administrative reforms throughout the enlarged Europe are also needed in order to speed patent awards and reduce patenting costs for firms (in particular the Small and Medium-sized Enterprises, SMEs). I am completely aware that until now modern economic theory did not pay the due level of attention to the analysis of the role patents may have in growth (Keely, 2001).18 However, with the arrival of new countries and new sectors (e.g. biotechnology and ICT) on the technological scene, Bruxelles has to take the lead in stimulating and initiating a deep discussion even on this important research and policy topic. 5. CONCLUDING REMARKS. Achieving the objective of making the European Union the most competitive and dynamic knowledge-based economy is not an easy task, as it requires increasing levels of collaboration between academics and policy-makers. The Sapir Report represents an important step in this direction, showing how a solid, current-research-based European economic policy can be built. In this paper I took the following approach. After recognizing that human capital (especially tertiary education) and technological capital (R&D) accumulation are (and will be more and more in the near future of Europe) two fundamental growth engines, I tried to highlight those factors that are potentially able to have a bearing on the private incentives to accumulate these two forms of intangible assets. As for human capital, we saw that other policies (complementary to the ones suggested in the Sapir Report) that can be put into use in order to accomplish the above-mentioned objective include those that: 1) reform and harmonize the length of tertiary education programmes throughout the enlarged Europe; 2) financially sustain (through public support) students involved in tertiary education programmes; 3) introduce reforms in the fields of health care, pension-systems and family in order to give stronger incentives to older men to invest in (adult) tertiary education; 4) increase human capital accumulation by training-on-the-job (and, then, reduce the cost of training for firms and rise the financial reward from training for employees); 5) help firms and schools to pool resources in order to improve the school-to-work transition process of students. I showed that 18 A notable exception is the recent work by O’Donoghue and Zweimüller, 2004. 18 most of these policies may importantly contribute to increase the private returns to tertiary education and, then, may play a relevant role in spurring human capital accumulation and growth. On the other hand, and concerning the R&D policy, our analysis allowed concluding that product market competition and regulation are fundamental factors of innovation. However, since in the last few years EU-countries did not reform at the same pace and the variance of product market regulation approaches has increased among most of them, efforts must be made in order to harmonize competition and regulation policies across European countries. In this context the lead should be definitely taken by European Commission itself in order to avoid uncertainty and heterogeneity in innovation incentives between countries. Finally, particular attention must be devoted in the future to the design of an European-based patent system that strikes a balance between the need of spurring private R&D incentives and that of ensuring a rapid diffusion of technological knowledge. The Sapir report does not provide any clear recommendation in this respect and does not highlight the problem of the possible links between the patent and competition policies.19 My impression is that thinking seriously of these issues can be strategic to reach the objective of a future (enlarged) Europe with increasing levels of economic well-being. REFERENCES Aghion, P., Bloom, B., Blundell, R., Griffith, R. and P. Howitt (2002): “Competition and Innovation: An Inverted-U Relationship”, NBER Working Paper No.9269. Bassanini, A. and S. Scarpetta (2001): “The driving forces of economic growth: panel data evidence for the OECD countries”, OECD Economic Studies, 0(33), pp.9-56. Baily, M.N. and H. Gersbach (1995): “Efficiency in manufacturing and the need for global competition”, Brookings Paper on Economic Activity: Microeconomics, pp.307-47. Bayoumi, T., Laxton, D. and P. Pesenti (2004): “Benefits and Spillovers of Greater Competition in Europe: A Macroeconomic Assessment”, NBER Working Paper No.10416. Blanchard, O. (2004): “The economic future of Europe”, NBER WP No.10310, February. Blöndal, S., Field, S. and N. Girouard (2002): “Investment in human capital through postcompulsory education and training: selected efficiency and equity aspects”, OECD Economics Department WP No.19/2002. Borsch-Supan, A. (1998): “Capital’s contribution to productivity and the nature of competition”, Brookings Paper on Economic Activity: Microeconomics, pp.205-44. Bucci, A. (2004): “An Inverted-U Relationship between Product Market Competition and Growth in an Extended Romerian Model”, University of Milan, Department of Economics, mimeo. Bucci, A. and C.H. Saglam (2000): “Growth-Maximizing Patent Life-Time”, Université catholique de Louvain, Department of Economics, Louvain-la-Neuve (Belgium), mimeo. Dearden, L., Reed, H. and J. van Reenen (2000): “Who gains when workers train? Training and corporate productivity in a panel of British industries”, CEPR Discussion Paper No.2486. Griliches, Z. (1980): “Returns to research and development expenditures in the private sector”, in “New Developments in Productivity Measurement and Analysis”, by J. Kendrick and B. Vaccara (eds.), Chicago: Chicago University Press, pp. 419-54. Jones, C.I. and J.C. Williams (1998): “Measuring the social return to R&D”, Quarterly Journal of Economics, 113(4), pp.1119-35. Keely, L.C. (2001): “Using Patents in Growth Models”, Economics of Innovation and New Technology, 10(6), pp. 449-92. 19 Maurer and Scotchmer (2004) analyze some basic principles that can reconcile patent and antitrust law. 19 Maurer, S.M. and S. Scotchmer (2004): “Profit Neutrality in Licensing: The Boundary between Antitrust Law and Patent Law”, NBER Working Paper No.10546. Nicoletti, G. and S. Scarpetta (2003): “Regulation, productivity and growth: OECD evidence”, Economic Policy, 0(36), pp.9-51. O’Donoghue, T. and J. Zweimüller (2004): “Patents in a Model of Endogenous Growth”, Journal of Economic Growth, 9(1), pp. 81-123. OECD (1999a), Classifying educational programmes. Manual for ISCED-97 implementation in OECD countries, 1999 Edition, Paris. OECD (1999b), Employment Outlook, Paris. OECD (2001), The New Economy: Beyond the Hype - Executive Summary, Paris. OECD (2002a), Education at a glance - OECD indicators, Paris. OECD (2002b), OECD Information Technology Outlook - Highlights, Paris. OECD (2003a), OECD Economic Outlook, No.73 (May), Paris. OECD (2003b), The sources of economic growth in the OECD countries, Paris. OECD (2004), Patents and Innovation: Trends and Policy Challenges, Paris. Park, W.G. (1995): “International R&D spillovers and OECD economic growth”, Economic Inquiry, 33(4), pp. 571-91. Sapir, A. et alii (2004), Europa, un’agenda per la crescita: Rapporto Sapir, Il Mulino, Bologna. Schumpeter, J. (1942): Capitalism, socialism and democracy, London: Allen and Unwin. Wößmann, L. (2003): “Specifying Human Capital”, Journal of Economic Surveys, 17(3), pp.23970. 20