Saving public pensions: Labor migration effects on pension systems in European countries

Abstract

European countries have experienced population aging and consequent pressure on public pensions. Some European countries, therefore, have welcomed migrants, expecting that the inflow of people will ease the demographic and fiscal problems. It is important to ask if this policy approach has had the intended effects. This paper examines the effects of labor migration on public pension systems. Using error correction models (ECMs) with cross-country time-series data on European countries from 1981 to 2009, this analysis demonstrates that labor migration has deterred the reduction of public pension benefit levels and government expenditure on pension as well as the expansion of private pensions. This implies that labor migration eases the pressure on public pension systems. Migration contributory effects have been larger in countries with Bismarckian pension systems because those countries have experienced greater pressure on public pension systems than other countries.

Keywords:

• Labor migration
• Pension

Notes

1 The multiple links are well summarized in Soroka, Banting, and Johnston (2006). According to them, migration can increase welfare spending levels, at least in the short term, because migrants are more dependent on welfare benefits, particularly unemployment insurance and child-care benefits. Migration may decrease the levels, particularly in the long run, because it reduces public support for generous welfare systems and/or increases support for right-wing parties that are relatively more tightfisted on welfare programs.

2 For example, highly skilled migrants are more likely to be net contributors than are unskilled migrants because of their higher wage rates and lower unemployment risk (Razin, Sadka, & Nam, 2005). In addition, differences in fiscal environments, such as whether foreign workers pay the same rates of taxes as native workers and whether migrants have access to welfare benefits, create variation in the fiscal effects of migration (Hanson, 2005).

3 Although public pensions are just one of many welfare programs in advanced economies, they usually are the largest program. More than 30% of social spending was used for the old-age pension program in 2002 in European developed countries, compared to 26% for health care and 6% for unemployment insurance.

4 There are more reasons than their relatively younger ages for migrants’ low participation rates in pension systems. Old migrants who have resided in receiving countries for a long time and have retired sometimes return to their home countries without receiving pension benefits from the host countries. Migrants are also less likely to be eligible for pension benefits than native people because of their short working history in receiving countries (Gott & Johnston, 2002).

5 For example, the arrival of new foreign workers decreases wage rates, and the decrease can reduce revenues for pension contributions. Kemnitz (2003), however, find that pensioners still gained from migration even though he took into account the reduced wage rates caused by migration. In addition, Krieger (2004) argue that if migrants have the same fertility rates as native people and immigrants’ children are less educated than natives’ children, the contribution will be reduced and the positive effect of migration will be offset. One survey, however, shows that migrants, particularly those from non-European countries, have more children than native people do in every European developed country (Boeri, Hanson, & McCormick, 2002). In addition, the educational gap between the children of immigrants and natives is not clear, and there is huge variation between countries: while immigrants’ children are less educated than natives’ children in some countries, they are more educated in others (Dustmann, Frattini, & Lanzara, 2012).

6 For example, while cutting pension benefit levels has a direct and imminent effect on alleviating the pressure on public pension funds, its political cost is high. Regarding expanding private pensions, while it reduces the tax wedge in the voluntary system, the mandatory system increases overall labor costs.

7 A replacement rate is the percentage of original income workers can receive in the event of unemployment, sickness, or retirement.

8 Contribution rates were never changed from 1988 to 2004 in Austria and Belgium. They were increased in Finland, Ireland, and the Netherlands, but annual changes occurred very rarely. Only marginal changes are found in most of the other countries. Retirement ages were never changed in most of the countries. Regarding the qualifying period, while frequent changes have been observed in some Western European countries, many other countries have never changed it.

9 The social democratic regime represents welfare systems where benefits are less related to labor market experience and the whole population is covered by the systems. Thus, the regime has the highest benefit levels among the three welfare regimes. The conservative regime relates welfare benefits to labor market participation, and thus benefits are earnings-related. Finally, the liberal regime focuses on means-tested benefits to assist poor people. This type of regime usually has the lowest benefit levels.

10 The different pension systems actually converged to a degree as they went through the pension crises after the 1990s. Bonoli (2000), however, states that “important differences exist with regard to benefit formulas, source of financing, financing method, and in the roles played by private and occupational provision. In general, the initial choice in terms of the Bismarck or the Beveridge model still affects the current shape of a pension system.”

11 However, it does not mean that the Beveridgean system is free from the pressure on the pension system. One aspect of the aging problem is extended life expectancy. As people live longer after retirement, the period of benefit receipts also becomes prolonged. Thus, even private pension systems in which current benefits are funded by previous contributions come under pressure from the aging problem (Bonoli & Shinkawa, 2005).

12 The countries are Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Italy, the Netherlands, Norway, Portugal, Spain, Sweden, and the U.K.

13 Scruggs’ data can be obtained at his website: http://www.polisci.uconn.edu/people/faculty/faculty.php%3Fname=scruggs. Standard pension single person replacement rates are used. Minimum pension single person replacement rates are also used for a robustness check. The robustness check, however, does not substantially change the analysis results.

14 The jus solis system draws more migrants than the jus sanguinis system because the children born to foreign parents can automatically acquire the nationality of the host country in the jus solis system (Leblang, Fitzgerald, & Teets, 2009). The description of the nationality laws, the characteristics of the jus soli system in each nationality law, and the coding of the nationality law variable can be obtained from the author.

15 When countries are poor, economic growth increases emigration because it enables people to finance migration costs. However, when countries become wealthier, further economic growth reduces emigration incentives because home countries provide as decent economic payoffs as destination countries.

16 The literature surveyed is: Bonoli (2000); Bonoli and Shinkawa (2005); Hering (2005); Schludi (2005).

17 The scales are from Castles and Mair (1984), Huber and Inglehart (1995), Laver and Hunt (1992), and Warwick (1994).

18 The population aging variable is the share of people who are 65 years of age or older (World Bank, World Development Indicators). The deindustrialization variable is the percentage of employees in sectors other than manufacturing (OECD, Labor Force Statistics). The budget deficit data are from International Monetary Fund, International Financial Statistics. The GDP per capita data are from the Penn World Tables, v6.3. The trade openness variable indicates the volume of imports plus exports as a percentage of GDP (World Bank, World Development Indicators). The capital inflow data are from World Bank, World Development Indicators.

19 It is the value of stocks traded as a percentage of GDP (World Bank, World Development Indicators).

20 The single-equation model is employed not only because it is considered theoretically superior to the two-step method, but also because the data satisfy the restriction for using the model (Beck, 1992).

21 There is no theoretical reason for using the 20-year-lag except that such phasing-in periods are usually 15–25 years. It is the longest lag that the data permits. However, it is presumed that the specific length of lag is not critically significant because the over-time variance of the foreign labor variable is much smaller than the cross-country variance. Using a 10-year-lag does not substantially change the result.

22 Scrugg's replacement rate data end in 2002, so only the period of 1981–2002 is tested for the benefit level models. In addition, the period of the 1980s is excluded from the analyses for private pension systems because of data availability.

23 The coefficients imply that the one standard deviation increase of foreign labor stock explains more than one standard deviation change of the benefit level.

24 The results of the first-stage regression for pension spending and private pension are not reported, but can be obtained from the author on request. Regarding the first-stage regression on the pension system, a cross-sectional time-series probit model was used because the pension system variable is a binary one.

25 The correlation between the predicted values from the first-stage regression and the real values is 0.7662 (foreign labor) and 0.6384 (pension system).

26 The education level is used as a proxy for the skill level (O’Rourke & Sinnott, 2006). The data on the education level of foreign workers is from OECD, International Migration Outlook.