European Integration and Inequality among Countries: A Lifecycle Income Analysis

1. Introduction

With the entry of Romania and Bulgaria in 2007, the European Union (EU) reached the figure of 27 members and culminated, for the moment, a process of enlargement begun half a century earlier when a group of six countries took the first steps in the project of European unity, from which substantial economic and social benefits were expected. However, the process of integration raises a series of questions that deserve an adequate response. In particular: (1) Do all member countries benefit equally? (2) Can the more developed countries better exploit the advantages of a wider market? Or, on the contrary, (3) Can the less developed countries, owing to their lower labor costs, benefit to a greater extent from the entry of capital, foreign investment, and the transfer of technology? In order to answer these questions, the economic convergence of EU member countries has to be examined. The fundamental question is whether European integration gives rise to a pattern of growth that generates convergence or not.

Most of the convergence studies (Barro and Sala-i-Martin, 1995) are based on current per capita income. Although this methodology provides useful information, this approach could be enriched with a methodology that also takes into account the whole life cycle dimension. Some recent studies have tried to consider this issue by using alternative measures. Dowrick et al. (2003) proposed their own index based on consumption and life expectancy; Becker et al. (2001, 2005) analyzed welfare inequality by giving an economic value to the gains achieved in terms of life expectancy; likewise Philipson and Soares (2001) proposed and examined the properties of a measure of total income (full income measure of human development).

This study aims to analyze European integration using an approach that is also different from the conventional one, but complementary to it (Serrano, 2006; Pastor and Serrano, 2008). The main characteristic of the method is that it considers the lifecycle income of individuals (present value of future current income), rather than only the income of a specific period. By using the present value (permanent income or lifecycle income), we can continue to take into account the level of current income of the period, while also considering factors such as the different life expectancies in each country and the different pace of future per capita income growth (which may converge much, little or not at all).

A country's entry into the EU is a structural change in its economy and its effects can only be valued from a long term perspective. For this reason, it is appropriate to use a lifecycle approach to analyze the effects of integration on the lifecycle income of individuals, rather than only on the income of specific periods.

2. The Process of European Integration: Expected Effects on Economic Growth and Empirical Evidence

The sources of the potential benefits associated with integration are very diverse (Viner, 1950). In addition to permitting more extensive exploitation of economies of scale (Harris, 1984; Gasiorek et al., 1992; Francois et al., 1994), integration fosters greater specialization and better use of the comparative advantages of the economies. Also, the opening-up of markets among the member countries (elimination of legal and customs barriers) increases competition, exerting further pressure to increase the efficiency of production (MacDonald, 1994; Caves and Barton, 1990) and countries can purchase raw materials and intermediate goods on better terms, with the consequent increase in productive efficiency (Lee, 1992). Finally, changes in the quantity and quality of the factors of production used can also be expected, because of their greater mobility within the area and increased technical progress (Maudos et al., 1999).

These potential advantages have been a continued stimulus towards an ever greater degree of integration among European countries. It has been a long and many-staged process whose ultimate outcome is the European Union of 27 members, and whose successive phases it would be appropriate to review before undertaking the empirical analysis. France, the Federal Republic of Germany, Italy, Holland, Belgium, and Luxemburg made up the six countries that were signatories of the Treaty of Rome in 1957, founding the European Economic Community (EEC). In 1973 the EEC expanded to nine members with the entry of Denmark, Ireland, and the UK. In the 1980s, the process of enlargement continued with the entry of Greece in 1981, and Spain and Portugal in 1986. At the end of 1990, the reunification of Germany took place and thus the länder belonging to the former GDR came to form part of the EEC. At the start of 1995, three other countries joined (Austria, Finland, and Sweden), raising the total membership to 15. In 2004 there was a new enlargement with the entry of 10 new members (Malta, Cyprus, Czech Republic, Slovakia, Hungary, Poland, Slovenia, Estonia, Latvia, and Lithuania). Finally, in early 2007, the last enlargement of the EU to date occurred, with the entry of Romania and Bulgaria into the European Union of 27 members.

As we can see there are a number of different stages of economic integration. Benefits could be expected from the free trade effect, customs union effect, common market effect or economic and monetary union effect. In this paper, we look at their combined effect on inequality. Also, it must be stressed that, as different countries became members at different time periods, the economic effects are not to be fully realized within the time span considered for the most recent members. Moreover, for some countries, such as those not sharing the euro as a currency (i.e. UK, Denmark, etc.), only part of those effects can be expected.

The empirical literature on the EU countries indicates that convergence is not steady over time. Various factors seem to act in opposing directions, some generating convergence and others divergence. Furthermore, these factors seem to play a different role in different countries, so the effect of integration can vary significantly among countries.

Henrekson et al. (1997) suggested that European integration may affect not only static efficiency through changes in resource allocation, but also long term growth rates. Their basic result is a fairly robust association between European integration and growth. The growth effect would be of the order of 0.6–0.8 percentage points per annum. The results also suggest that technology transfer is the main mechanism through which EC and EFTA membership affect growth. Delgado-Rodriguez and Alvarez-Ayuso (2008) analyzed the evolution of labor productivity among EU-15 countries over the period 1980–2001. Using β-convergence techniques, they identified periods of non-significant convergence (1980–85 and 1993–96), as well as others of rapid and significant convergence (1986–92 and 1997–2001). For the whole period results are not statistically significant. Physical and human capital accumulation appeared to be the main driving force behind the process. On the contrary, technological progress tended to contribute to divergence, although a change in the trend was observed at the end of the period.

Maudos et al. (1999) analyzed the evolution of the EU countries and the impact on efficiency and productivity of the successive enlargements during the period 1965–90. The results show that until 1990 integration was beneficial for all the participants. The countries that joined experienced substantial relative gains in efficiency. In addition, the growth rate of TFP in the founder countries received a positive impulse with each new enlargement. Kaitila (2004) analyzed both σ- and β-convergence and discussed the impact of EU membership. According to the results, the EU15 countries' real per capita gross domestic product (GDP) levels, adjusted for purchasing power, converged in 1960–2001. Convergence occurred over two periods, in 1960–73 and 1986–2001, with an interim period of stagnation. Abiad et al. (2007) stressed that, owing to increasing financial integration, capital in Europe has travelled “downhill” from rich to poor countries accelerating income convergence. In Reza and Zahra (2008) real convergence of the 10 new members' economies to the EU average income was tested by using quarterly real GDP per capita data from 1995 to 2005. They concluded that the 10 new members of the EU in 2005 tend to converge towards the EU average income.

In any case, these studies are devoted to analyzing convergence and inequality in current per capita income and/or labor productivity. However, given that the effects of European integration are long term, it is natural to use also a lifecycle approach such as the one described in the next section.

3. General Formulation

As we have already been indicated, this paper analyzes inequality within the EU throughout its existence, using as the key variable the per capita permanent income or lifecycle income. The per capita permanent income (PI_i0) of economy i at time 0 is the discounted value of the present and future per capita current income (y_it), taking into account the survival rate in each period. We define S_i(t, 0) as the probability of a person of being alive in period t. Thus, the permanent income at time 0 is defined as indicated in the following expression:

(1)

in which we assume a common and constant interest rate r, and that the maximum life of an individual is 120 years (an unrestrictive assumption).

Note that, ceteris paribus, according to expression (1), countries will have higher levels of permanent income: (i) the higher their initial per capita incomes (y_i0); (ii) the higher their rates of growth (g_i); (iii) the greater the survival rates, S_i(t, 0); (iv) the lower the rate of discount (r), since this increases the present value of future incomes. A more formal analysis of this approach and its relationship to theoretical models of growth is offered in Serrano (2006).

At empirical level, three factors will influence the inequality in permanent income: the initial levels of per capita current income, the per capita future income flows and the survival rates of individuals. Bearing this in mind, in the next section we will consider different counterfactual scenarios to evaluate separately the role of each of these determining factors. This will enable us to value the effect of each of these factors on inequality and convergence in the EU.

4. Data and Results

All the data (GDP per capita and survival rates) are taken from World Bank Development Indicators 2006. Table 1 offers the detailed data regarding life expectancies and per capita current incomes in 1960 and 2005. The data are shown relative to the benchmark economy (USA). Additionally, we present the relative positions of each country in terms of current and permanent per capita income in both periods.

With respect to life expectancy at birth, the data show modest but significant differences. Thus, in 1960, a country such as Holland had a life expectancy 5.2% longer than the USA, while Portugal ranked 9.1% below that level. This is a substantial difference, of the order of 15%. Among the rest of the countries the differences were smaller but appreciable. In 2005, the greatest difference was between Sweden (4% above the USA) and Romania (7.9% below the USA), a difference of 12%. There were significant changes during this period, such as the relative improvements of countries such as France, Italy, Portugal, Spain, Austria, and Finland; on the contrary, the evolution was less satisfactory in other cases (Holland, Denmark, and nearly all the eastern European countries). These differences in terms of life expectancy and the changes occurring during the period make this an aspect to be taken into account when valuing the levels of inequality in the EU and convergence among countries from a long term perspective.

In terms of per capita income, the differences are greater. In 1960, Denmark and Bulgaria were the extreme cases with per capita incomes equivalent respectively to 88.7% and 4.5% of the per capita income in the USA. In 2005, these two countries still showed the maximum and minimum values within the group of countries currently forming the EU27, Denmark with a per capita income of 84.1% that of the USA and Bulgaria barely reaching 5.5%. In addition to the size of the differences, the changes occurring during the period should also be taken into account. The extreme values show a stability which it would be deceptive to consider as something general. Thus, countries like Ireland, Belgium, Italy, Greece, Portugal, Austria, Finland, Hungary, Malta or Slovenia achieved substantial improvements in relative terms. Other countries like Denmark, Holland, UK, Czech Republic or Sweden, however, showed a less satisfactory evolution. This disparity of behaviors over time makes it important to take into account not only the per capita incomes at each moment, but also the present value of the per capita incomes expected in the future. With such disparate growth rates of per capita income, inequality in terms of lifecycle income can vary significantly from inequality in current income of the period.

The estimations of per capita permanent income based on the “historical scenario” in relative terms to the US per capita permanent income are offered in the last two columns of Table 1. In this scenario, according to expression (1), the actual initial per capita income and the actual survival rates have been used for each country. Also, we have use a long term growth rate based on the assumption that per capita income grows at the mean rate achieved during the period 1960–2005. Finally, to convert future incomes to present values we have used a discount rate of 2%. This procedure (expression (1)) has been used to estimate the per capita permanent income for all the European countries as well as for the benchmark economy (USA).

In 1960, permanent income varied between the value for Denmark (87.6% of the USA's) and that for Bulgaria (5.4% of the USA's). It would seem, therefore, that considering the lifecycle is not too important given that the results for both extreme cases are similar to those obtained using simply the current income for 1960. However, when we observe what happens in the other countries and not only in the extreme cases, significant changes are seen. Although all the countries are below the USA, their relative positions change substantially if current income or permanent income is considered. Among others, Belgium, France, Italy, Greece, Portugal, Spain, Austria, Finland, or especially Ireland, improve considerably when permanent income is considered. Thus, for example, Ireland goes from 34.6% of the USA to 79.5%; Spain from 26.6% to 40.5%, or Italy from 41.6% to 50.3%. On the contrary, there are countries such as the UK or Denmark where the opposite occurs and others where the improvement is unimportant (e.g. Germany, Holland, or Poland). In 2005, something similar occurs and it is easy to see that the ranking of countries would change considerably if permanent income were considered instead of current income.

The above results show the existence of differences in terms of life expectancy and economic growth rates among European countries and, therefore, justify the interest in adopting a permanent income approach to analyze inequality in the EU. Also, the estimations of permanent income indicate that the results can differ in many countries from those habitual in exercises based on current income. For this reason, we analyze the evolution of inequality among countries within the EU throughout the period, together with the possible changes associated with the various enlargements, with this lifecycle perspective always in mind.

Our analysis of inequality in the EU is based on the use of the coefficient of variation of per capita income, a dispersion statistic habitual in this type of inequality analysis. Figure 1 shows the level of inequality in different periods (including those years when enlargements of the EU occurred) for the countries forming the EU at that time, both in terms of current income and of permanent income. Let us first examine the results in terms of current income. The coefficient of variation of current income of the EU6 was 0.161 in 1960. Following the entry in 1973 of Denmark, the United Kingdom, and Ireland, the inequality increased to 0.254. In 1981, the year of Greece's entry, it was moderated to 0.249, growing substantially to 0.359 in 1986 (entry of Portugal and Spain). From then onwards there was a gradual descent, to 0.315 in 1990 (reunification of Germany) and 0.282 in 1995 (enlargement of the EU15 to 15). It has to be said that the effect of the most recent enlargements has been a very significant increase in inequality among countries in the EU. In 2004 (enlargement to 25 member countries), the coefficient reached 0.591 and if we include Bulgaria and Romania (which joined in 2007), the coefficient would be at levels around 0.653.

In sum, we observe that there is a progressive increase in inequality until the mid 1980s, a reduction over the next 20 years and a sharp rise as a consequence of the latest enlargements towards Eastern Europe. Thus, within this much more heterogeneous 27-member EU, the current levels of relative inequality multiply several times those existing among the original members of the EU.

Let us now consider inequality from a lifecycle perspective, using once again the coefficient of variation, but in this case that of permanent per capita income. The results appear in the same Figure 1. The temporal profile now shows a continued growth which, indeed, accelerates quite visibly with the latest enlargements. Thus, the coefficient starts at a level of 0.101 in 1960 (EU6), increases to 0.141 in 1973, 0.168 in 1981, 0.218 in 1986, 0.236 in 1990, 0.258 in 1995, 0.702 in 2004, and 0.770 in 2005. When permanent income is used we observe that levels of inequality are at maximum values for the EU, much higher than the initial ones.

When the results obtained from these two perspectives (current and permanent income) are compared, we observe some interesting differences. During the period prior to the latest enlargement, inequality was much lower if permanent income is considered, especially for the years before 1995. That is to say, taking into account the future rather than only the income of the period, the levels of inequality within the EU were substantially less than those indicated by the current per capita income of the period. However, after the latest enlargements taking place this century, the image is quite the contrary. Inequality is greater in terms of permanent per capita income. The differences of life expectancy and of incomes foreseeable in the future tend to magnify the inequality among the members of the present EU, whereas in the past the opposite occurred. This means that, unless European cohesion policies contribute more actively to changing this panorama, the levels of inequality in the EU will continue to be high.

Once we have analyzed inequality and convergence in permanent income (scenario 1 or historical scenario), our next step is to evaluate separately the role of each of their determining factors, such as survival rates, per capita current income, and rate of convergence. In order to do this, the effect of each of them will be isolated step by step, i.e. allowing for changes only in one of these variables each time. These correspond to different assumptions, such as survival rates, the initial levels of per capita current income, and the existence or otherwise of convergence among countries in terms of per capita current income. On the basis of these scenarios, we simulate the inequality of permanent income of the countries by building some counterfactual scenarios.

Table 2 shows the coefficients of variation of permanent per capita income corresponding to these new counterfactual scenarios. As we have been shown above, one of the potentially important factors in determining permanent income is the future growth rate of current per capita income. To estimate its impact on inequality we define scenario 2 in which we have obtained the permanent incomes of each country on the assumption of a common future growth rate. Specifically, the growth rate of the US per capita income from 1960 to the present has been used for all countries (g_i = g_US). The initial levels of per capita income and life expectancies, on the contrary, continue to be those of each country. The results of scenario 2 show increasing inequalities until the creation of the EU12, followed by slight reductions in the 1990s and a sharp rise with the latest enlargements. What is more interesting is the comparison between these results and those of scenario 1, as the differences between the two indicate the part of the inequality in permanent income associated with the different long term growth rates of each country's current per capita income. As can be observed, the different growth rates of per capita income during the period 1960–2005 helped to reduce inequality, with the exception of 1960 and 2005.

In the successive enlargements from 1973 to 1995, we can observe that the inequality in permanent income of the historical scenario (scenario 1) is always less than that which would have been obtained with a common rate of growth. Thus, in 1986 the coefficient of variation of the EU12 countries was 0.218, but applying the common growth rate (scenario 2) means that this coefficient would be 0.313, indicating nearly 50% more inequality in permanent income. The reason is that, when the member countries of the EU9, EU10, EU12, and EU15 are considered as a whole, the countries with lowest per capita income showed faster long term growth during the period 1960–2005. However, the latest enlargements again show differential characteristics. For the EU25 inequality stands at 0.702 and for the EU27 at 0.77, according to scenario 1. When using a common growth rate, inequality falls to 0.603 and 0.667 respectively because the latest enlargements have brought in countries that are less developed and which have, in the past, shown less capacity for long term growth.

The second important factor for explaining inequality among countries is the difference in life expectancy. Scenario 3 has been defined for the purpose of evaluating the importance that differences in life expectancy have had for the levels of inequality among the countries of the EU throughout its history. The results of this scenario were obtained under the assumption that all the countries had a common life expectancy, specifically that of the US (S_i = S_US). Therefore, the differences between the inequality levels of scenario 3 and those of scenario 1 (historical scenario) have to be attributed to the differences in life expectancy of each country. In 1960, the differences in life expectancy among the member countries of the EU6 explain a significant part of the inequality in permanent income (with a common life expectancy such as that of the USA the coefficient of variation of permanent income would have been 0.089 instead of 0.101). During the 1980s and 1990s, on the contrary, the effect was the opposite, helping to reduce slightly the inequality in permanent income (thus, in 1995 the coefficient of variation among the countries of the EU-15 with a common life expectancy would have been 0.267 instead of 0.258). The impact of life expectancy is currently very low, and the levels of inequality would barely change even if the differences in life expectancy disappeared.

Scenario 4 corresponds to the estimations of permanent income obtained by assuming that all EU countries start with the same initial per capita income, specifically that of the USA (y_io = y_USo), but maintain the life expectancies and long term growth rates of each country. This allows us to analyze the effect of these last two variables by comparisons with scenario 1. In 1960, inequality of permanent income was 0.075, clearly below the inequality in current per capita income and also below the inequality observed in the historical scenario (scenario 1), confirming that differences in life expectancy and, particularly, in growth rates, contribute significantly to the inequality among the member countries of the original EU. From that moment onwards, the impact of these two factors becomes increasingly significant. Inequality in permanent income is greater than that obtained in scenario 1 in 1973, 1981, 1986, and 1995. With the latest enlargement this trend has been broken, given that most of the new member countries present a substantial gap between their initial current per capita incomes and those of the existing member countries. The impact of growth rates and of differences in life expectancy is substantial (thus, the inequality in scenario 4 rises from 0.294 for the EU15 in 1995 to 0.507 for the EU25 in 2004), but the differences are smaller than in scenario 1 (thus, for the EU25 in 2004 inequality in permanent income is 0.507 in scenario 4 and 0.702 in scenario 1).

The evidence for the growth trajectories of the European countries after each enlargement is not especially encouraging. The results of scenario 5 were obtained by estimating in each year the future current incomes on the basis of the growth rates experienced by each country since that year (g_i = g_iPOST). Let us recall that the estimations of scenario 1 are based always on the growth rates measured from 1960 to the present. The comparison between scenario 5 and scenario 1 is clear. We can leave aside the result for 1960 which, naturally, has to coincide. In the rest of the years, the inequality estimated in permanent income is significantly greater because post-enlargement growths are used. This already occurs in 1973 (0.241 and 0.141) and the phenomenon persists with increasing intensity. In 1995, the level of inequality would be multiplied by four (1.078 and 0.258) and with the latest enlargement the result is similar (2.85 and 0.77). Naturally, we must bear in mind that the more recent the year analyzed, the shorter the period that serves to calculate the post-enlargement growth rates, which may be more subject to immediate factors than to long term ones. In sum, the results indicate that if the current post-integration growth rates are maintained, substantial levels of inequality will remain.

Let us recall that the base estimation (scenario 1) was obtained from the rates of growth of per capita long term income (1960–2005) of each country in the past. A different pattern of growth in the future could have a substantial impact. In scenario 6 these growth rates have been replaced by others that correspond to a situation of convergence in per capita current incomes among the countries of Europe. Under this hypothesis, the countries with lowest per capita income would grow most and would do so faster owing to their relative backwardness. Three different annual convergence rates have been posited: 2%, 3%, and 5%. These three convergence scenarios correspond to the hypothesis that each country manages to reduce the gap with the benchmark economy in current per capita income by 2%, 3%, or 5% (respectively) each year.

The results indicate that this would lead to a steep reduction of the inequality in permanent income among EU countries. The results of scenario 1 indicate that the inequality in permanent income in the EU is currently 0.77. However, if instead of maintaining the past growth rates of each country we assume a future annual convergence of 2%, the estimated inequality in permanent income would currently be only 0.223, less than one third. If a somewhat greater annual convergence in current per capita income (3%) were achieved, it would be only 0.167. Finally, with a convergence rate of 5%, the inequality in permanent income would be barely 0.116.

In other words, if a convergence rate of 2% were achieved, as a result of the process of economic integration itself or to the EU's cohesion policies, then the inequality in current income in 2005 of 0.653 would be compatible with a lifecycle inequality two thirds lower (0.223). It should be emphasized that rates of convergence between economies of 2% are perfectly feasible. Numerous studies have estimated similar convergence rates among the countries of the Organisation for Economic Co-operation and Development (OECD), the states of the USA, the prefectures of Japan, the regions of Germany, Spain, etc. (Barro and Sala-i-Martin, 1995). Furthermore, any increase in that rate of convergence would have its reward in an appreciable reduction of the inequality in permanent income.

Finally, scenario 7 allows structural change as a result of integration through two different growth rates for the per capita income of each new member economy: one for the period before the year of integration of the country (historical growth rate from 1960 until then), and another for the period after its integration in the EU (historical growth rate from membership until 2005). The results are very similar to those of scenario 5 (using the post integration growth rate until 2005). The levels of inequality are significantly greater than those of scenario 1. When we take into account the possibility of changes in steady-states for the countries joining the EU, substantial levels of inequality seem to remain. Furthermore, the inequality in permanent income of this scenario is clearly higher than the one observed in current income from the 1980s onwards.

After analyzing the influence of the determinants of permanent income on inequality, we may wonder about the evolution of inequality following the successive enlargements, both for the “old” member countries, and for the countries that are joining. Table 3 permits us to observe the phenomenon of inequality in this multiple dimension, both in terms of current per capita income (panel a) and in terms of permanent per capita income (panel b).

The rows of the upper panel permit us to see the evolution of the inequality in current per capita income of the successive groups of countries that have come to form the EU throughout the period 1960–2005. In the first row, we can observe that the inequality among the founding countries at the start of the EU (EU6) decreased progressively from 0.161 to less than half that in 1995 (0.078), rising slightly thereafter (0.085 in 2005), though remaining well below the initial levels. For the expanded group of countries that formed the EU9 in 1973 something similar occurred, with inequality decreasing from 0.254 in 1973 to 0.152 in 2005. With the entry of Greece in 1981 the EU10 was formed, although the reduction of inequality for this group was weaker (from 0.249 in 1981 to 0.235 in 2005). The experience of the EU12 was more positive with the entry of Spain and Portugal in 1986: inequality was reduced from a level of 0.359 in that year to 0.309 in 2005. Slight reductions were also observed for the expanded collectives EU15 and EU25 from the moment of these enlargements. That is to say, the successive enlargements were accompanied by reductions in inequality among the “old” members.

The columns of the table indicate that inequality for all countries of the EU would have increased as a consequence of the entry of new, more heterogeneous, countries. Thus, if we observe the levels of inequality in current per capita income in 2005 for the different subgroups (last column), we notice that it increases with the subgroup considered: 0.085 for the EU6, 0.152 for the EU9, 0.235 for the EU10, 0.309 for the EU12, 0.278 for the EU15 (only exception, logical given the characteristics of the three new members), and a steep rise to 0.583 for the EU25 and 0.653 for the EU27.

The lower panel shows the results in terms of permanent per capita income. There are many similarities, but also some significant differences from the estimates regarding current per capita income. Thus, observing the level of inequality in 2005 we can see that the various enlargements have meant a clear increase in inequality for the EU, from the level of 0.074 for the founding countries (EU6) to levels of around 0.50 for the EU9, EU10, and EU12 and, finally, levels of 0.699 for the EU25 and 0.77 for the EU27. No increase in inequality is observed for 2005, when the EU9 expanded to EU10 or EU12, something which did happen when current income was examined.

If we now focus attention on the impact of the successive enlargements over time, we will see that only in the case of the founding countries (EU6) can we observe a reduction of inequality in permanent income: the coefficient of inequality went from 0.101 in 1960 to 0.074 in 2005. For the rest of the groupings that have formed the EU at each time, the trend is the opposite: the EU9 countries go from 0.141 in 1973 to 0.500 in 2005, the EU10 countries from 0.168 in 1981 to 0.509 in 2005, the EU12 countries from 0.218 in 1986 to 0.499 in 2005, and the EU15 countries from 0.258 in 1995 to 0.430 in 2005. The general tendency is that the successive enlargements have been accompanied by substantial increases in inequality of permanent incomes among the “old” members, except in the case of the six founding countries.

5. Conclusions

The usual analyses of inequality focus on the evolution of the current per capita income of the period. In this study, we have adopted a complementary approach to analyze the EU experience from a life-cycle, or permanent income, perspective.

The results obtained indicate that inequality in terms of permanent income was substantially less than that shown by current per capita income for the EU until the most recent processes of enlargement. However, the opposite occurs with the enlargements of 2004 and 2007. The inequality in current income at present underestimates the inequality in permanent income, a somewhat unsatisfactory situation.

Nor does the temporal evolution of inequality in permanent income permit us to be too optimistic. If we focus on the fixed groups of countries that have formed the different European Unions, we observe convergence until the mid 1980s and divergence from then onwards. Therefore, all member countries do not seem to have benefited equally from the enlargement process. The less developed countries seem to have benefited to a greater extent in terms of permanent income during the enlargements until the mid 1980s (convergence period). However, the more developed countries seem to have been able to better exploit the advantages of integration from then onwards (divergence period), although it is too soon to evaluate the final results of the most recent enlargements. On the contrary, the behavior of the changing group of countries that have formed the EU in the course of time shows an almost permanent tendency towards divergence.

Differences in life expectancy would have helped to increase inequality in permanent income in the initial phases of the European integration project. However, at present they have very little impact. The overall level of inequality would now be practically the same even if all the countries had the same life expectancy. The differences in the starting level of per capita current incomes are a more important factor of inequality in permanent income, though the sign of their effect varies in the course of the period analyzed. In the 1960s and also at the present time, they contribute substantially to generating greater inequality in permanent income. We should also point out the influence of the different economic growth rates of each country. This factor would have clearly contributed to reducing inequality in permanent income systematically since 1960. However, with the latest enlargement its effect has been inverted and it becomes a source of greater inequality in permanent income. Finally, the analysis of the effects of the enlargements on the different groups of countries indicates that the successive enlargements have been accompanied by substantial increases of inequality in permanent income among the “old” members, except with regard to the case of the six founding countries.

All these results point to the key role that economic growth plays in achieving further reductions in inequality in the EU, given that the contribution of other variables such as life expectancy seems, at present, to be rather limited. Policies that stimulated greater growth of the least developed countries would have considerable effect. The simulations carried out show that with an annual convergence rate of 2% (i.e. the countries reduce the gap existing in current per capita income by 2% every year), the inequality in permanent income would be less than one third of what it is now. This rate of convergence is ambitious, but not impossible, as it is consistent with that recorded by the OECD countries in the past and with those habitually obtained when analyzing convergence among the regions of a single country.