Abstract
The external imbalances that accompanied Latvia’s integration into the global economy have generated mixed effects. Stabilizing net foreign liabilities will require an improvement in the goods and services balance. Outward migration from Latvia has raised concerns about social and economic stresses that could intensify in the near future if labor flow increases. Several steps have to be taken to mitigate any adverse consequences of outward migration. The purpose of this paper is to examine macroeconomic and prudential aspects of the rapid growth of lending in Latvia.
II. Labor Outflows, Capital Inflows and Income Convergence in Latvia12
A. Introduction and Overview
1. Latvia, along with many other emerging market economies, is concerned with the economic and social consequences emigration. This paper addresses the causes and implications of labor emigration in the context of a converging economy with a much lower capital-to-labor ratio than its EU partners.
Section B examines the extent of Latvian emigration and compares Latvia’s experience to two historical episodes of migration—the so-called age of mass migration during 1870-1910 and the migration that accompanied the reunification of Germany starting in 1989.
Section C briefly summarizes macroeconomic growth and economic geography theories which view migration as a natural part of economic convergence. They suggest a number of possible outcomes, ranging from orderly convergence to permanent divergence. Unfortunately, economic theory is not yet mature enough to put forward specific policy advice, but it does offer an analytical approach for evaluating developments in Latvia, as well as other emerging economies.
Using this analytical approach section D takes a look at recent trends in economic convergence (labor and capital flows) in Latvia and compares this experience to several current and historical episodes of convergence. The performance is mixed: While Latvia appears to be heading in the right direction, these trends must be interpreted very cautiously, especially since available data ends just as Latvia joined the EU in 2004. With open borders with the UK, Ireland, and Sweden and possibly open borders with other EU members in the near future, both capital and labor flows might look very different in a few years’ time.
Section E reviews the economic and social consequences of outward migration. While the impact of emigration is poorly understood and there are no easy answers for mitigating its economic impact, there are several factors that should be considered in formulating a long-term immigration policy.
Finally, section F offers some short- and longer-term policy options for coping with labor outflows.
B. How Large is the Migration Problem?
2. Official statistics reveal that the population of Latvia has declined markedly in recent years, owing both to net outward migration and to a falling birthrate (Table 1). According to these data, Latvia has almost 400,000 fewer residents in 2005 compared to the population in 1990. As shown in the last two columns of the table, the population decrease is about equally explained by net outward migration and by a natural decrease in the population (as a fall in the fertility rate led to deaths outnumbering births). The official statistics also indicate that more than one-half of the residents that left Latvia between 1990 and 2005 emigrated to CIS countries. The fall in the population due to natural causes—which is not discussed in this paper—is also quite worrisome, as it continues unabated and will begin to exacerbate a shrinking labor market in a few years’ time.
Latvia: Population Changes, 1990-2005
Latvia: Population Changes, 1990-2005
| Year | Population (in thousands) |
∆ in Population (in thousands) |
∆ in Population (per 1,000) |
Net Migration (per 1,000) |
Natural Increase (per 1,000) |
|---|---|---|---|---|---|
| 1990 | 2,689 | ||||
| 1995 | 2,501 | -189 | -73.4 | -53.1 | -20.3 |
| 2000 | 2,377 | -123 | -43.6 | -14.3 | -29.3 |
| 2005 | 2,306 | -71 | -29.8 | -4.0 | -25.8 |
| 1990-2005 | -383 | -147 | -71 | -75 |
Latvia: Population Changes, 1990-2005
| Year | Population (in thousands) |
∆ in Population (in thousands) |
∆ in Population (per 1,000) |
Net Migration (per 1,000) |
Natural Increase (per 1,000) |
|---|---|---|---|---|---|
| 1990 | 2,689 | ||||
| 1995 | 2,501 | -189 | -73.4 | -53.1 | -20.3 |
| 2000 | 2,377 | -123 | -43.6 | -14.3 | -29.3 |
| 2005 | 2,306 | -71 | -29.8 | -4.0 | -25.8 |
| 1990-2005 | -383 | -147 | -71 | -75 |
3. The official statistics measure only long-term migration (CSB, 2006). According to the practices of Latvia’s Citizenship and Migration Board, long-term migration refers to individuals who leave Latvia with the intent of settling in a new country for at least one year and who report this intention to an appropriate authority. Therefore, the data in Table 1 do not capture several categories of migration:
illegal migrants who move permanently to another country without proper authorization from the receiving country,
permanent migrants to another EU country who do not report their change of residency, and
temporary migrants to another EU country who are not required to report their arrival in the receiving country.
The latter two categories are especially important in Latvia’s case: Since EU accession, many Latvians have been moving to Ireland and the U.K.—either permanently or temporarily—to seek and obtain work.
4. Indeed, anecdotal evidence suggests that unreported outward migration is significant and has picked up substantially since EU accession. The Latvian authorities estimate that at least 50,000 additional Latvians have moved abroad—about twice the official number and about 2 percent of the 2005 population—since Latvia joined the EU in 2004. Other (unofficial) estimates suggest that this number may be as large at 100,000 individuals. Moreover, the number of emigrants could increase in the near future, as more EU countries open their borders to EU-member workers.13
5. Several recent surveys suggest that potential outward migration is substantial. For example, a survey conducted by SKDS (2006) in January 2006 revealed that about 22 percent of Latvian residents are “very likely” or “somewhat likely” to go to another country for work “in the next two years”. Based on the current estimated population, this translates into between 350 and 450 thousand residents (between 15 and 20 percent of the 2005 population). The survey also indicated that these respondents were significantly skewed toward the relatively young (15-35), which would significantly reduce the working-age population and labor force in the near future. These respondents were also slightly more likely to be male, less educated, low-income, employed in the private sector, or non-Latvian.14
6. How does Latvia’s emigration experience compare with other historical periods of migration? To summarize the preceding analysis, it is reasonable to assume that between 5 and 10 percent of Latvia’s residents are currently living (and possibly working) abroad. Moreover, surveys suggest that an additional 15 to 20 percent of current residents are thinking of leaving in the next few years. How do these numbers compare with experiences of other countries?
7. There was a massive flow of migrants from the old world to the New world between 1870 and 1910. This era represents one of the most open periods in terms of labor mobility. It also represents an era with great economic disparities: In 1870, real wages in Ireland, for example, were about 40 percent of real wages in the New World, and disparities were even larger in Denmark, Italy, Norway, and Sweden (see Taylor and Williamson (1994)). Laborers and their families responded en masse to these incentives (Table 2).15 The largest outflows were from Ireland, where over 40 percent of the labor force and about a third of the population left between 1870 and 1910. There were smaller but sizeable outflows from Italy and the Nordic countries, and these outflows were of similar magnitude to the outflows (actual plus potential) for Latvia that were discussed previously.
Demographic Impacts During the Age of Mass Migration, 1870-1910
Per 1,000
Demographic Impacts During the Age of Mass Migration, 1870-1910
| Average Annual Net Migration Rate 1 | Cumulative Percent Change | |||
|---|---|---|---|---|
| Country | Population | Labor Force | Population | Labor Force |
| Denmark | -2.4 | -3.2 | -9 | -12 |
| Ireland | -10.1 | -13.3 | -33 | -41 |
| Italy | -6.8 | -8.5 | -23 | -29 |
| Norway | -4.8 | -6.2 | -17 | -22 |
| Sweden | -3.8 | -5.0 | -14 | -18 |
Per 1,000
Demographic Impacts During the Age of Mass Migration, 1870-1910
| Average Annual Net Migration Rate 1 | Cumulative Percent Change | |||
|---|---|---|---|---|
| Country | Population | Labor Force | Population | Labor Force |
| Denmark | -2.4 | -3.2 | -9 | -12 |
| Ireland | -10.1 | -13.3 | -33 | -41 |
| Italy | -6.8 | -8.5 | -23 | -29 |
| Norway | -4.8 | -6.2 | -17 | -22 |
| Sweden | -3.8 | -5.0 | -14 | -18 |
Per 1,000
8. Migration from East to West Germany after the fall of the Berlin Wall in 1989 also provides an interesting comparison (Table 3). In order to speed up convergence between the two regions, West Germany supplied massive amounts of capital to update East Germany’s infrastructure, and labor market policies entailed equalization of nominal wages that led to a significant jump in real wages for East Germans. Nevertheless, despite these efforts, East Germany’s population dropped almost 9 percent between 1989 and 1999. Similar to Latvia, about half of this decline can be attributed to net outward migration—amounting to 4.7 percentage points—with a falling birth rate accounting for the rest. In addition, gross migration was also quite sizeable, with more than 7 percent of East Germany’s population moving to West Germany.16
Demographic Impacts in East Germany, 1989-1999
Per 1,000
Demographic Impacts in East Germany, 1989-1999
| Average Annual Net Migration Rate 1 |
Percent Change in Population |
|
|---|---|---|
| Between East and West | -6.6 | -7.2 |
| From Abroad | 2.3 | 2.5 |
| Total Net Migration | -4.3 | -4.7 |
Per 1,000
Demographic Impacts in East Germany, 1989-1999
| Average Annual Net Migration Rate 1 |
Percent Change in Population |
|
|---|---|---|
| Between East and West | -6.6 | -7.2 |
| From Abroad | 2.3 | 2.5 |
| Total Net Migration | -4.3 | -4.7 |
Per 1,000
C. What Causes Migration?
9. There are four main economic and non-economic categories of migration. First, there are refugees—migrants who move for a number of non-economic reasons: those attempting to escape political, social, or religious repression and those fleeing war or natural disasters. Second, there are migrants who move temporarily to another region for educational or training purposes. Third, some workers temporarily relocate to another region, in response to economic shocks in either the sending or the receiving country. For example, a construction worker might seek temporary work in another region in response to a short-term construction boom in that region. Finally, there are workers who permanently relocate. These workers tend to view their long-term, economic prospects as relatively better in host locations compared to their own country.17 This paper focuses on the latter two categories—those who migrate for economic reasons, either temporarily or permanently.
10. For those who migrate for economic reasons, expected income or wage differentials are the primary determinants, although cultural factors also play a role. Economic theory predicts that labor will move from a region with low wages and incomes and high unemployment to regions with relatively higher wages and incomes and lower unemployment. However, several factors complicate this prediction. First, there are cost barriers to moving. This indicates that the wage rate or the income level must exceed some threshold level before outward migration begins to occur. Second, there are also informational costs regarding the labor market in the destination country. This leads to an inverted U-shaped emigration pattern, whereby migration begins slowly, increases rapidly when information about employment conditions flows back to the sending country from previous migrants, and then tapers off once labor markets in the two countries converge. Third, there are important cultural and language differences to account for. This is consistent with Decressin and Fatas (1995), who show that labor mobility among European countries is much lower than among U.S. states and, therefore, that most region-specific shocks in Europe are absorbed by changes in participation rates rather than by labor migration. Finally, as demonstrated by Fidrmuc (2004), region-specific shocks may induce both inward and outward migration. For example, a sector-specific shock in one region may induce an inflow of high-skilled workers from other regions, while producing an outflow of low-skilled workers.
11. Growth and economic geography theories offer broader explanations for movements of goods and factors. To illustrate the basic economic concepts, consider two hypothetical regions—East and West (Figure 1). The two regions were effectively separated for some time, either because transport costs were prohibitively high or because trade in goods and factors was explicitly prohibited. Both regions have access to the same technology. However, the two regions have different endowments of labor (shown on the horizontal axis) and capital (shown on the vertical axis). More specifically, labor is relatively abundant in the East, while capital is relatively scarce (point A). Similarly, wage rates should be relatively low in the East, and the return on capital should be relatively high.
Capital, Labor, and Convergence Between East and West
Citation: IMF Staff Country Reports 2006, 354; 10.5089/9781451824599.002.A002
12. Now consider what happens when trade barriers are removed. The neoclassical growth models—of Solow (1956) and others—predict that convergence will occur between East and West, with capital flowing toward the East and labor moving to the West until capital-to-labor ratios are equalized across countries. Thus, the convergence process is depicted by a movement toward the “K-L line” in Figure 1, and total convergence results in a point on the K-L line. The models also predict convergence of factor prices, with average wage rates increasing and average rental rates for capital falling in the East, until they reach Western levels.18 The degree to which factors of production are mobile and the relative mobility between factors will determine how long the convergence process will take and how each factor will adjust. For example, suppose that capital is perfectly mobile and that labor is not. In this case, capital from the West will relocate instantly to the East, and convergence will take place at point B. Under this scenario, there would be no movement of labor between East and West.
13. In reality, capital is not perfectly mobile and labor is able to relocate. Setup costs and capital market imperfections imply that capital is less-than-perfectly mobile, although capital is likely more mobile than labor. In this case, there will be some shedding of labor, with net migration from East to West. That is, convergence will take place at a point like C. The more that capital is immobile or the more that labor is mobile, the further that point C will lie below and to the left of point B. In addition, other factors, such as government policies and endowments of human capital, land, and other natural resources, will affect where point C lies on the K-L line. Nevertheless, the neoclassical models predict that—if technologies are identical in both regions—convergence will eventually take place, with capital-to-labor ratios, wages, rates of returns on capital, and incomes becoming equal.
14. Newer theories of economic growth suggest the possibility of permanent economic divergence. The endogenous growth theories suggest that returns to capital—see Romer (1986)—do not have to be diminishing and that economic convergence is not assured. That is, under certain conditions, there is a possibility of permanent divergence in wages, incomes, and labor productivity.19 In Lucas (1988), human capital has increasing returns to scale and offers the possibility that highly-skilled or highly-educated workers might accumulate in some regions (at the expense of other regions), leading to permanent divergence. Similarly, as in Romer (1990), there are models where increasing returns are achieved via research and development, and they also suggest the possibility of permanent—and, perhaps, ever-widening—divergence.
15. The “new economic geography” (NEG) models offer an even richer set of possibilities.20 In general, NEG models emphasize the existence of two strong, opposing forces. First, there are agglomeration or locational effects that tend to pull capital and labor (together) to central locations. These agglomeration forces come from a number of sources: (i) ports, roads, and other infrastructure, (ii) efficiency gains from firms locating near other firms that provide them with intermediate goods and vice versa, (iii) efficiency gains from firms locating near large markets or large pools of labor, (iv) utility gains from consumers finding a larger variety of goods in large population centers, and (v) reduced search costs for laborers locating near “thick” labor markets. On the other hand, these agglomeration tendencies are moderated by poor business climates, congestion and other decreases in quality of life, and transport costs of goods to distant markets. Which force is stronger for a given region—agglomeration or congestion effects—is a complex function of technology and tastes, historical accident, government policies, and institutions.
16. The NEG theory offers a range of possible outcomes for East and West. Returning to Figure 1, it is reasonable to assume that firms and populations are initially concentrated in the West.21 This suggests two possible extremes for changes in capital and labor locations. On the one hand, agglomeration forces could continue to keep capital in the West—near supply and demand centers—once trade barriers are removed. This is depicted by a point like D, where very little capital flows into the East, and convergence is achieved mostly by labor flows from East to West. A much less likely outcome—point E—is where both capital and labor flee the East. On the other hand, NEG theories also offer the possibility that the West is relatively overcrowded, making the East look attractive for capital (and, possibly, labor) inflows. Good government policies (such as low taxes, ease of doing business, good public infrastructure) would also help to make capital inflows more likely, and, thereby, reducing the need for labor outflows. This outcome is represented by a point like C.
17. To summarize, the endogenous growth and NEG models suggest a much broader range of economic outcomes relative to the neoclassical growth models. Unfortunately, these literatures are still too underdeveloped to make economic predictions or to provide specific policy advice. Nevertheless, the next section of the paper takes a look at recent trends in Latvia with an eye toward understanding whether Latvia is headed toward convergence—with or without major labor outflows—or divergence. It also compares Latvia’s recent experiences with those of other countries.
D. Trends in Latvia with Some International Comparisons
18. Where is Latvia headed and when will emigration end? While the theoretical literature discussed in the previous section does not offer clear answers to these questions, it does lead to others that can be explored empirically: Is convergence or divergence taking place in Latvia? If wage and income convergence is taking place, what is the nature of this convergence—how much capital is flowing in relative to labor outflows? Is convergence on a sustainable path; that is, what is the nature of capital inflows? Finally, how long is the convergence process likely to take? These questions essentially relate back to issues raised in Figure 1: Is Latvia heading toward the K-L line, and, if so, is it headed to point B, C, D, or E?
19. Before turning to the empirical evidence in Latvia, it is interesting to take a quick look at convergence during the age of mass migration discussed in section B. Despite the massive number of emigrants from many Old World countries between 1870 and 1910, there was little convergence in terms of either real wages or labor productivity (Table 4). In Ireland, for example, despite a loss of more than 40 percent of its workforce, the average wage rate rose very little compared to the average wage in the United States. The same is true for the 5 largest sources of migrants, as relative wages rose only modestly, on average, from 29 percent to 48 percent compared to the United States. Taylor and Williamson (1994) do not attribute these developments to agglomeration effects, where capital and labor are seen as complements. Instead, they argue that land and other natural resources were relatively abundant in the New World, while capital and labor were relatively scarce. Thus, it was natural that both capital and labor would migrate to the New World.22 Indeed, Taylor and Williamson used a calibrated model to construct a counterfactual case (that of no migration) and to calculate the effects of migration on convergence (Table 5). They conclude that if migration had not occurred, real wages would have been about 30 percent less in Ireland, for example, and labor productivity would have been 20 percent less.
Convergence During the Age of Mass Migration, 1870-1910
(ratio, relative to the United States)
Convergence During the Age of Mass Migration, 1870-1910
(ratio, relative to the United States)
| Real Wages | GDP per Worker | |||
|---|---|---|---|---|
| 1870 | 1910 | 1870 | 1910 | |
| Denmark | 0.31 | 0.58 | 0.50 | 0.55 |
| Ireland | 0.43 | 0.54 | — | — |
| Italy | 0.23 | 0.29 | 0.39 | 0.37 |
| Norway | 0.24 | 0.41 | 0.47 | 0.44 |
| Sweden | 0.24 | 0.59 | 0.47 | 0.47 |
| Average | 0.29 | 0.48 | 0.46 | 0.46 |
Convergence During the Age of Mass Migration, 1870-1910
(ratio, relative to the United States)
| Real Wages | GDP per Worker | |||
|---|---|---|---|---|
| 1870 | 1910 | 1870 | 1910 | |
| Denmark | 0.31 | 0.58 | 0.50 | 0.55 |
| Ireland | 0.43 | 0.54 | — | — |
| Italy | 0.23 | 0.29 | 0.39 | 0.37 |
| Norway | 0.24 | 0.41 | 0.47 | 0.44 |
| Sweden | 0.24 | 0.59 | 0.47 | 0.47 |
| Average | 0.29 | 0.48 | 0.46 | 0.46 |
Impact of Migration on Convergence Measures, 1870-1910
(percent of total change)
Impact of Migration on Convergence Measures, 1870-1910
(percent of total change)
| On Real Wages | On GDP per Worker | |
|---|---|---|
| Denmark | 9 | 5 |
| Ireland | 31 | 20 |
| Italy | 23 | 15 |
| Norway | 12 | 8 |
| Sweden | 10 | 6 |
Impact of Migration on Convergence Measures, 1870-1910
(percent of total change)
| On Real Wages | On GDP per Worker | |
|---|---|---|
| Denmark | 9 | 5 |
| Ireland | 31 | 20 |
| Italy | 23 | 15 |
| Norway | 12 | 8 |
| Sweden | 10 | 6 |
20. Returning now to Latvia, this paper uses information on advanced and emerging market countries to estimate the level of capital stock. Unfortunately, comparable measures of the capital stock are not readily available for the New Member States. Panel regressions on a broad set of countries are used to infer a value for the capital stock in 1994, and then investment data is used to construct estimates of the capital stock for 1995-2004. The aggregate production function can be written as follows:
The second equation shows that labor productivity (output per worker) is a function of technology (which includes the effects of human capital) and the capital-to-labor ratio. Table 6 shows the results of estimating this relationship for a cross-section of 62 countries in 1994. The capital-to-labor ratio was obtained from Bosworth and Collins (2003). Two proxies for technology were used—the share of agriculture in GDP was obtained from the World Bank’s Social Development Indicators, and the average number of years of schooling was obtained from Barro (2000). Regional dummy variables were included to capture common geographic, social, and economic factors not captured by other explanatory variables. The capital stocks for the NMS in 1994 were then inferred, using known values for output, labor, agricultural shares, educational attainment, and the estimated coefficients in Table 6.23 Estimates of the capital stock for 1995-2004 were then constructed using national accounts data obtained from the World Bank’s Social Development Indicators.
Estimation Results
(Dependent Variable is log of GDP per Worker in 1994)
Bolded coefficients and p-values indicate that the estimated coefficient is significantly different from zero at the 5 percent level.
Estimation Results
(Dependent Variable is log of GDP per Worker in 1994)
| Variable | Coefficient | p-value1 |
|---|---|---|
| Constant | 1.037 | 0.05 |
| Log (Agr Share) | -0.082 | 0.14 |
| Log (Human Capital) | 0.132 | 0.26 |
| Log (K/L) | 0.783 | 0.00 |
| OECD | 0.303 | 0.03 |
| Latin America | 0.204 | 0.04 |
| East Asia | -0.017 | 0.87 |
| South Asia | -0.121 | 0.29 |
| Middle East & N. Africa | 0.070 | 0.48 |
| N = 62 R2 = .99 | ||
Bolded coefficients and p-values indicate that the estimated coefficient is significantly different from zero at the 5 percent level.
Estimation Results
(Dependent Variable is log of GDP per Worker in 1994)
| Variable | Coefficient | p-value1 |
|---|---|---|
| Constant | 1.037 | 0.05 |
| Log (Agr Share) | -0.082 | 0.14 |
| Log (Human Capital) | 0.132 | 0.26 |
| Log (K/L) | 0.783 | 0.00 |
| OECD | 0.303 | 0.03 |
| Latin America | 0.204 | 0.04 |
| East Asia | -0.017 | 0.87 |
| South Asia | -0.121 | 0.29 |
| Middle East & N. Africa | 0.070 | 0.48 |
| N = 62 R2 = .99 | ||
Bolded coefficients and p-values indicate that the estimated coefficient is significantly different from zero at the 5 percent level.
21. The capital-to-labor ratio in Latvia has increased steadily over the past 9 years (Table 7). The estimated capital-to-labor ratio—which includes residential and business structures, in addition to stocks of machinery and equipment—increased 125 percent between 1995 and 2004. Moreover, as shown in the last column of the table, the gap between the K/L ratio in Latvia and the EU-15 average narrowed modestly over the same period. The lower rows of the table decompose developments in the K/L ratio. The K/L ratio has increased at an average, annual rate of almost 14 percent. This increase comes entirely from an expanding capital stock—at a 15.3 percent rate per year—which was slightly offset by a modest increase in the number of workers over the nine-year period. It is also interesting to note that the increase in workers came from an increase in the employment rate, as the Latvian labor force shrank, mirroring the net outward migration rates discussed earlier.24
Convergence in Latvia, 1995-2004
Convergence in Latvia, 1995-2004
| Latvia | EU Average | Ratio | |||
|---|---|---|---|---|---|
| (in 2000 US$) | |||||
| Capital per Worker in 1995 | 9,859 | 119,742 | 0.1 | ||
| Capital per Worker in 2004 | 22,237 | 141,147 | 0.2 | ||
| (average annual growth rate) | |||||
| Capital per Worker | 13.9 | 2.0 | 7.0 | ||
| Capital | 15.3 | 2.7 | 5.7 | ||
| Workers | 0.6 | 0.6 | 1.0 | ||
| Labor Force | -0.5 | 0.3 | -1.9 | ||
| Employment Rate | 1.2 | 0.3 | 4.0 | ||
Convergence in Latvia, 1995-2004
| Latvia | EU Average | Ratio | |||
|---|---|---|---|---|---|
| (in 2000 US$) | |||||
| Capital per Worker in 1995 | 9,859 | 119,742 | 0.1 | ||
| Capital per Worker in 2004 | 22,237 | 141,147 | 0.2 | ||
| (average annual growth rate) | |||||
| Capital per Worker | 13.9 | 2.0 | 7.0 | ||
| Capital | 15.3 | 2.7 | 5.7 | ||
| Workers | 0.6 | 0.6 | 1.0 | ||
| Labor Force | -0.5 | 0.3 | -1.9 | ||
| Employment Rate | 1.2 | 0.3 | 4.0 | ||
22. While Latvia appears to be moving in the right direction in terms of catching up by raising its capital-to-labor ratio and accumulating tangible capital, there are reasons to be cautious in viewing these trends.
First, these estimates indicate that Latvia is a long way from reaching convergence with the EU-15. To put these numbers in perspective, consider the amount of capital needed to reach convergence: Latvia would have needed about $164 billion more in capital in 2004 to match EU-average capital-to-labor ratios, or about 1,600 percent of GDP.25 Alternatively, Latvia would have had to shed about 80 percent of its workforce to reach the EU-average capital-to-labor ratio in 2004. At the current pace of convergence, it will take 30 to 40 years for Latvia to approach EU capital-to-labor ratios.
Second, the estimates are based on data through 2004. As discussed earlier, there is evidence that the rate of outward migration has picked up sharply since EU accession. This would tend to promote further convergence.
Third, the recent pace of capital absorption may not be sustainable. The capital stock has grown at double-digit rates over the past few years, a pace exceeding all other New Member States (Table 8).26 This rate of growth also exceeds the rate of capital accumulation in the “newer” Old Member States (upper part of Table 9). It is also striking to observe that capital growth rates in Latvia also surpass those of Korea and Singapore during a period of unprecedented, rapid, and oft-studied periods of factor accumulation (lower part of Table 9). Such a rapid pace of capital accumulation raises concerns about the ability of the economy—households, businesses, and the banking system—to absorb and efficiently allocate such large amounts of capital.27
Finally, the type of capital that is being accumulated has longer-term consequences for output growth and labor productivity. As Martin and Sanz (2003) discuss in their study of convergence in Greece, Ireland, Portugal, and Spain, the contribution to growth from international spillovers can be quite important. The most direct contribution is through contracts for transfer of technology, but the economic growth literature also stresses two indirect sources of spillovers—foreign direct investment and imports of goods, especially imports of machinery and equipment. Indeed, Martin and Sanz find that the knowledge spillovers through trade account for much of the recent catch-up in the “newer” Old Member States. In this respect, Latvia does not score well, as its imports of machinery and equipment appear to be lagging behind its European neighbors (Table 10).
Comparisons to Other New Member States, 1995-2004
Comparisons to Other New Member States, 1995-2004
| Average Annual %∆ in | |||||||
|---|---|---|---|---|---|---|---|
| K/L in 1995 | K/L in 2004 | Ratio | K/L | K | L | ||
| Latvia | 9,859 | 22,237 | 2.3 | 13.9 | 15.3 | 0.6 | |
| Estonia | 11,960 | 25,083 | 2.1 | 12.2 | 11.5 | -0.3 | |
| Slovak Republic | 13,885 | 26,665 | 1.9 | 10.2 | 10.2 | 0.0 | |
| Czech Republic | 20,977 | 37,026 | 1.8 | 8.5 | 7.8 | -0.4 | |
| Poland | 15,337 | 25,549 | 1.7 | 7.4 | 6.9 | -0.3 | |
| Lithuania | 11,958 | 18,897 | 1.6 | 6.4 | 7.3 | 0.5 | |
| Hungary | 19,851 | 30,099 | 1.5 | 5.7 | 6.6 | 0.6 | |
| Slovenia | 42,279 | 62,989 | 1.5 | 5.4 | 5.8 | 0.2 | |
Comparisons to Other New Member States, 1995-2004
| Average Annual %∆ in | |||||||
|---|---|---|---|---|---|---|---|
| K/L in 1995 | K/L in 2004 | Ratio | K/L | K | L | ||
| Latvia | 9,859 | 22,237 | 2.3 | 13.9 | 15.3 | 0.6 | |
| Estonia | 11,960 | 25,083 | 2.1 | 12.2 | 11.5 | -0.3 | |
| Slovak Republic | 13,885 | 26,665 | 1.9 | 10.2 | 10.2 | 0.0 | |
| Czech Republic | 20,977 | 37,026 | 1.8 | 8.5 | 7.8 | -0.4 | |
| Poland | 15,337 | 25,549 | 1.7 | 7.4 | 6.9 | -0.3 | |
| Lithuania | 11,958 | 18,897 | 1.6 | 6.4 | 7.3 | 0.5 | |
| Hungary | 19,851 | 30,099 | 1.5 | 5.7 | 6.6 | 0.6 | |
| Slovenia | 42,279 | 62,989 | 1.5 | 5.4 | 5.8 | 0.2 | |
Selected Countries: Changes in the Capital-to-Labor Ratio, 1960-2000
Selected Countries: Changes in the Capital-to-Labor Ratio, 1960-2000
| Average Annual %∆ in | |||||||
|---|---|---|---|---|---|---|---|
| K/L in 1995 | K/L in 2004 | Ratio | K/L | K | L | ||
| Greece | 13,044 | 71,259 | 5.5 | 4.4 | 5.2 | 0.8 | |
| Ireland | 23,524 | 103,721 | 4.4 | 3.8 | 4.7 | 0.9 | |
| Portugal | 14,084 | 59,254 | 4.2 | 3.7 | 4.7 | 1.0 | |
| Spain | 16,455 | 87,514 | 5.3 | 4.3 | 5.3 | 1.0 | |
| Korea | 2,646 | 58,137 | 22.0 | 8.1 | 11.0 | 2.7 | |
| Singapore | 4,880 | 138,803 | 28.4 | 8.8 | 12.4 | 3.3 | |
Selected Countries: Changes in the Capital-to-Labor Ratio, 1960-2000
| Average Annual %∆ in | |||||||
|---|---|---|---|---|---|---|---|
| K/L in 1995 | K/L in 2004 | Ratio | K/L | K | L | ||
| Greece | 13,044 | 71,259 | 5.5 | 4.4 | 5.2 | 0.8 | |
| Ireland | 23,524 | 103,721 | 4.4 | 3.8 | 4.7 | 0.9 | |
| Portugal | 14,084 | 59,254 | 4.2 | 3.7 | 4.7 | 1.0 | |
| Spain | 16,455 | 87,514 | 5.3 | 4.3 | 5.3 | 1.0 | |
| Korea | 2,646 | 58,137 | 22.0 | 8.1 | 11.0 | 2.7 | |
| Singapore | 4,880 | 138,803 | 28.4 | 8.8 | 12.4 | 3.3 | |
Imports of Machinery 1/
(percent of total imports)
Excluding cars.
Imports of Machinery 1/
(percent of total imports)
| 1995 | 2000 | 2004 | |
|---|---|---|---|
| Hungary | 23.2 | 43.4 | 44.5 |
| Czech Republic | 29.4 | 31.7 | 32.9 |
| Estonia | 22.1 | 34.1 | 28.8 |
| Slovakia | 23.9 | 23.6 | 25.4 |
| Poland | 24.3 | 26.8 | 25.2 |
| Slovenia | 20.7 | 22.1 | 21.5 |
| Lithuania | 13.9 | 15.8 | 19.3 |
| Latvia | 17.6 | 20.8 | 18.8 |
Excluding cars.
Imports of Machinery 1/
(percent of total imports)
| 1995 | 2000 | 2004 | |
|---|---|---|---|
| Hungary | 23.2 | 43.4 | 44.5 |
| Czech Republic | 29.4 | 31.7 | 32.9 |
| Estonia | 22.1 | 34.1 | 28.8 |
| Slovakia | 23.9 | 23.6 | 25.4 |
| Poland | 24.3 | 26.8 | 25.2 |
| Slovenia | 20.7 | 22.1 | 21.5 |
| Lithuania | 13.9 | 15.8 | 19.3 |
| Latvia | 17.6 | 20.8 | 18.8 |
Excluding cars.
E. The Consequences of Migration
23. Net outward migration can have important positive effects on emerging market economies. Following the discussion of the previous section, outward migration can be an integral part of economic development and convergence. First, immigration can reduce the pool of unemployed workers, which has an immediate positive effect on the fiscal budget. Second, on average, the loss of labor raises the capital-to-labor ratio and overall labor productivity, thereby raising average real wages. Third, if workers eventually return to their home country, they can bring back important knowledge and skills gained while working abroad.28 Finally, worker remittances are an important source of international capital for developing countries - often surpassing other sources of capital inflows (see World Bank (2006a)). However, since remittances may also discourage labor force participation and the number of working hours of recipients, the initial economic benefits of the remittances can be offset over the longer term.29
24. There are also a number of important undesirable impacts of migration on the originating country. First, migration of employed workers can lead to a significant loss of tax revenue, which can create pressures to slow or reduce expenditures on infrastructure, education, health care, and pensions. Second, the net economic benefits of migration are not so clear for higher-skilled workers, since there are many possible externalities created by the loss of high-skilled workers and highly educated citizens.30 On one hand, economic growth is reduced if: (i) other workers lose the opportunity of interacting with higher-skilled workers, either in the workplace or in the educational system, (ii) economies of scale in skill-intensive activities cannot be exploited, or (iii) society misses out on the returns from using public funds to finance education. On the other hand, well-educated workers living abroad can enhance network opportunities for the originating country, leading to better access to technology, capital, and other business advantages.
25. Although there is a plethora of studies on the consequences of migration, the literature is fairly inconclusive and is hampered by a number of difficulties. First, as in the case of Latvia, there is a dearth of quality data on migration patterns, both within countries and across countries. OECD countries generally keep track of legal immigrants by originating countries and emigrants by destination country. Therefore, there are limited data on legal migration flows and stocks to and from OECD countries. However, this approach underestimates migratory flows to the extent that there is illegal migration to and from OECD countries. In addition, this approach completely ignores migration from developed countries to developing countries and migration among developing economies. Likewise, cross-country data on wages, incomes and unemployment are fraught with mismeasurement and definitional differences. Second, most econometric studies of migration are backward looking, whereas the decision to migrate is almost surely one based on assessing future prospects. Finally, most theories and empirical studies of migration are partial-equilibrium analyses, and they typically do not account for the role of other factors of production, such as the availability of physical capital, human capital, and land. In other words, these studies offer no basis for understanding the underlying macroeconomic causes of migration, and they do not offer any broader policy solutions for addressing the consequences of migration. Conversely, recent advancements in the theory of regional and international development discussed in section C have important implications for migration, and they offer a much richer set of explanations for factor movements.
26. At the end of the day, the effects of migration are poorly understood, and there are no easy answers for mitigating its economic impact. The World Bank (2006a) recently observed that “[m]igration is as complex as it is diverse, so predicting the impact of policy changes will be problematic until more research is done and better data is obtained.” Nevertheless, that report discusses several factors that should be considered in formulating migration policy.31 Those factors can be summarized as follows:
Since the decision to migrate is both risky and costly, governments should provide information on migration opportunities and risks in order to avoid poor decisions.
Given that the migration of low-skilled workers is generally a benefit to the originating country, governments might also consider managed migration programs—designed jointly by the origin and destination countries—that lower migration burdens for the poorest of low-skilled workers.
Many countries with high rates of high-skilled emigration also have poor investment climates. While it is not clear whether poor investment climates cause emigration (or vice versa), it is important to consider whether public infrastructure, government policies, or the business climate more generally is limiting the productive employment of high-skilled workers.
Post-graduation service requirements in return for access to publicly-financed education can be easily evaded including through emigration, and likely discourage return to the originating country.
Origin countries can encourage educated emigrants to return by identifying job opportunities, permitting dual nationality, and facilitating the portability of social insurance benefits.
F. Conclusions and Policy Options
27. Outward migration from Latvia has raised concerns about social and economic stresses, that could intensify in the near future if labor flows increase. However, as argued above, net outward migration can be viewed as a natural part of economic convergence, since labor shedding raises both the capital-labor ratio and labor productivity. So far, Latvia does not appear to be heading toward a scenario where both capital and labor are exiting the country. Rather, capital inflows have been strong, wages have been increasing, and the employment rate has been rising, which has more than offset losses in the overall labor force. Nevertheless, complete economic convergence is a distant prospect, and Latvia faces a number of challenges to achieve a desirable outcome, where convergence is reached in an orderly fashion and without significant loss of population.
28. With emigration an integral part of economic convergence, measures to stem the tide are unlikely to be successful, while overheating can exacerbate outflows. Full convergence of capital-labor ratios and wage levels is expected to take many years. To preserve competitiveness in the interim—and avoid the missteps made during German reunification—it is important that wage growth remain anchored to improvements in productivity. Raising local wages—unsupported by productivity gains—in order to deter emigration is unlikely to be effective, especially if cross-country differences in wages are large. Moreover, it should be recognized that an overheated economy or a real estate price boom can facilitate additional outflows (especially those at the lower end of the economic ladder) because higher consumer prices erode the purchasing power of local wages, while steep increases in house prices put home ownership out of reach of those earning local wages.
29. Several steps can be taken to mitigate any adverse consequences of outward migration. First, facilitating greater internal mobility would allow workers to shift to sectors where labor shortages are most pronounced. This could be achieved by releasing excess labor in the public sector through productivity improvements and reducing economy-wide hiring and firing costs.32 Second, increasing the effective labor supply by encouraging greater labor force participation (which is low for males compared with EU15 countries) could be realized through a reduction of the tax wedge on labor income of low-wage earners. Increasing the PIT exemption or granting tax credits for lower income workers could deliver this result. Third, consideration could be given to easing restrictions on the immigration of higher-skilled workers from non-EU countries in order to ease growth bottlenecks while also raising the average skill level of the work force. And fourth, improving the investment climate could encourage capital accumulation to support sustainable wage growth. Doing so will require restoring macrostability, and utilizing EU funds to their best advantage to build human capital, facilitate absorption of new technologies, integrate local firms into multinational production chains, and improve infrastructure for viable sectors.
30. Latvia should also give consideration to designing a longer-term migration plan in conjunction with other countries in order to ease outflows and promote future return. This could include: (i) Providing information on employment opportunities abroad and risks associated with migration in order to avoid poor decision making; (ii) Considering managed migration programs—designed jointly by the origin and destination countries—that lower migration burdens for the poorest of low-skilled workers; (iii) Resolving problems that could be discouraging the productive employment of high-skilled workers abroad; and (iv) Identifying employment opportunities at home to encourage the return of educated emigrants.
References
Barro, Robert J., and Jong-Wha Lee, 2002, “International Data on Educational Attainment, Updates, and Implications”, Working Paper #7911, National Bureau of Economic Research, Cambridge, MA.
Bosworth, Barry P., and Susan M. Collins, 2003, “The Empirics of Growth: an Update”, Brookings Papers on Economic Activity, 2:113-206.
Brunovskis, Anette, Anne Britt Djuve, and Hilde Maria Haualand, 2003, “Facing a Baltic Invasion? Mobility of Baltic Labor Towards the Nordic Countries,” Report 432, Fafo, Oslo.
Burda, Michael C., and Jennifer Hunt, 2001, “From Reunification to Economic Integration: Productivity and the Labor Market in Eastern Germany”, Brookings Papers on Economic Activity, 2: 1–92.
Central Statistics Burea of Latvia, 2006, “Migration of Population of Latvia During 2005”, press release dated April 21, 2006.
Co, Catherine Y., Ira N. Gang, Myeong Su Yun, 2000, “Returns to Returning”, Journal of Population Economics, 13: 57–79.
Chami, Ralph, Thomas Cosimano, and Michael T. Gapen, 2006, “Beware of Emigrants Bearing Gifts: Optimal Fiscal and Monetary Policy in the Presence of Remittances”, Working Paper #61, International Monetary Fund, Washington, DC.
Decressin, Jörg, and Antonio Fatas, 1995, “Regional Labor Market Dynamics in Europe”, European Economic Review, 39: 1627–55.
Fidrmuc, Jan, 2004, “Migration and Regional Adjustment to Asymmetric Shocks in Transition Countries”, Journal of Comparative Economics, 32: 230–47.
Giuliano, Paula, and Marta Ruiz-Arranz, 2005, “Remittances, Financial Development, and Growth”, Working Paper #234, International Monetary Fund, Washington, DC.
Hazans, Mihails, 2006, “Potential Emigration of Latvia Labor Force after Joining the EU and its Impact on Latvian Labor Market”, manuscript, University of Latvia and Baltic International Centre for Economic Policy Studies (BICEPS), Riga, Latvia.
Krugman, Paul, 1981, “Trade, Accumulation, and Uneven Development”, Journal of Development Economics, 8: 149–61.
Krugman, Paul, and Anthony Venables, 1995, “Globalization and the Inequality of Nations”, Quarterly Journal of Economics, 110: 857–80.
Lucas, Robert E., 1988, “On the Mechanics of Economic Development”, Journal of Monetary Economics, 22: 3–42.
Martin, Carmela, and Ismael Sanz, 2003, “Real Convergence and European Integration: The Experience of the Less Developed EU Members”, Empirica, 30: 205–36.
Mishra, Prachi, 2006, “Emigration and Brain Drain: Evidence from the Caribbean”, Working Paper #25, International Monetary Fund, Washington, DC.
Mundende, D. Chongo, 1989, “The Brain Drain and Developing Countries”, in The Impact of International Migration on Developing Countries (edited by Reginald Appleyard), pp. 183–95, Development Centre of the Organisation for Economic Cooperation and Development, Paris.
Romer, Paul M., 1986, “Increasing Returns and Long Run Growth”, Journal of Political Economy, 94:1002–37.
Romer, Paul M., 1990, “Endogenous Technological Change”, Journal of Political Economy, 98:5–21.
SKDS (Market and Public Opinion Centre), 2006, “Public Opinion on Manpower Migration: Opinion Poll of Latvia’s Population” manuscript, Riga, Latvia.
Solow, Robert M., 1956, “A Contribution to the Theory of Economic Growth”, Quarterly Journal of Economics, 70:65–94.
Taylor, Alan M., and Jeffrey G. Williamson, 1994, “Convergence in the Age of Mass Migration”, Working Paper #4711, National Bureau of Economic Research, Cambridge, MA.
World Bank, 2006a, Global Economic Prospects: Economic Implications of Remittances and Migration, Washington, DC.
World Bank, 2006b, “Doing Business in Latvia”, downloadable at www.doingbusiness.org, Washington, DC.
Prepared by Allan Brunner.
By 2011, all EU countries are required to accommodate intra-EU labor migration.
Other surveys show similar results—see Brunovskis, Djuve, and Haualand (2003) and Hazans (2006).
The effects of this mass migration on economic convergence between the Old World and the New World will be discussed in section D.
See Burda and Hunt (2001) for a very extensive analysis of this episode.
With respect to the fourth category, the literature tends to focus on the causes and consequences of high-skilled migration and the so-called “brain drain” and “brain gain” phenomena.
Since the West is much larger than the East, the impact of convergence on factor prices in the West is negligible.
In the context of Figure 1, the long-term equilibrium does not lie on the K-L line, and there will be permanent differences in capital-to-labor ratios, wages, incomes, and returns to capital.
See Krugman (1991) for an analysis of uneven development, and Krugman and Venables (1995) for the importance of economic geography.
For expositional reasons, Figure 1 indicates a sizeable population in the East. But, this is not necessary. What is important is that East is relatively better endowed with labor; that is, point A lies below the K-L line.
Nevertheless, their explanation remains somewhat unconvincing, as it does not explain why convergence of real wages or labor productivity were not complete after 40 years of migration.
A further assumption was made about where the NMS stand vis-a-vis the regional dummy variables. It was assumed that they have technologies similar to those in Latin American countries. However, assuming that technologies were closer to those in industrial countries does not change the qualitative nature of the subsequent analysis.
It is unlikely that employment rates will increase much further. The unemployment rate has fallen sharply in recent months, and signs of overheating suggest that unemployment may be near or below its natural rate.
By comparison, Singapore (discussed below) accumulated nearly $250 billion of capital between 1970 (when its capital-to-labor ratio was similar to Latvia’s in 2004) and 2000 (when convergence with the West was largely complete).
It is interesting to note that three NMS (Estonia, Poland, and the Czech Republic) have seen a reduction in their labor forces, which contributes to an increase in their capital-to-labor ratios. As with Latvia, the other NMS have seen their labor forces increase slightly.
These issues are discussed further in two companion papers—”Integration, External Imbalances and Adjustment: The Latvian Experience” and “Latvia-Aspects of Rapid Credit Growth”.
Co, Gang, and Yun (2000), for example, found that Hungarian women earned a wage premium for work experience abroad, while the results were not conclusive for men.
Nonetheless, Giuliano and Ruiz-Arranz (2005) and Chami, Cosimano, and Gapen (2006), for example, argue that remittances are on balance a net economic gain.
See World Bank (2006a) for a more complete discussion. Also, see Mishra (2006), who concludes “that the losses due to high-skill migration (ceteris paribus) outweigh the official remittances to the Caribbean region”.
Similar factors were discussed by Mundende (1989).
According to the World Bank’s “Doing Business 2007” assessment of countries’ business environments, Latvia scores relatively highly overall (24th of 175 countries—up from 31st place in the previous assessment), but is rated poorly (123rd) on employing workers owing to weak grades on difficulties in hiring and firing.
