1. Introduction
The equity-efficiency quandary of the welfare state is usually attributed to perverse incentive
effects in the labor market. The welfare state includes ‘unproductive’ government spending
which reduces the return to work and is financed by distortionary taxes. Apparently, little
evidence exists on the effect of welfare state arrangements on investment in human capital.
The welfare state can be seen as a social insurance mechanism, see for example Sinn (1995).
When the insurance terms for the insured improve, her incentives to invest to avoid capture
are weakened. In a macroeconomic context, this moral hazard problem may have detrimental
effects on investment in human capital, saving, and, ultimately, economic growth.1 Ehrlich
and Zhong (1998) and Ehrlich and Kim (2007) find a negative effect of old-age pension
benefits on secondary school enrolment rates, particularly for developed countries.2 In
addition, Bj0rnskov, Dreher and Fischer (2007) find that higher government consumption
spending is related to less subjective well-being, perhaps through misallocation of resources,
the inefficiencies generated through modern taxation schemes or other transmission channels
described above.
To our best knowledge, we are among the first to empirically investigate to what extent
government involvement in the economy through public goods creation and income
redistribution affects individual investment in human capital. In this study, we approximate
the former by both government consumption, social expenditures, and progressivity of the tax
system, but the latter by international student test scores, which we adjust to ensure
comparability across countries and time. Most of the existing empirical analyses on economic
growth include as a proxy of human capital some measure of quantity of education in the
population. This is obviously a crude measure, and we follow Wossmann (2003a) who argues
that the number of quality-education-years varies across countries stronger than the mere
duration of education, with which it might even be uncorrelated. Indeed, Hanushek and
Kimko (2000) find that average student achievement in compulsory schooling is a much more
sizable determinant of economic growth than years of education in the population. The strong
effect of student achievement is confirmed by Hanushek and Wossmann (2007) and Jamison,
Jamison and Hanushek (2007).
1 Welfare state arrangements may also be seen as interventions in imperfect markets, working in the opposite
direction. Indeed, the empirical cross-country literature indicates that there is no relationship between
government expenditures and growth, although the results vary somewhat across studies, see for example Folster
and Henrekson (2001) and Agell, Ohlsson and Thoursie (2006). Kneller, Bleaney and Gemmell (1999)
distinguish between different types of taxes and spending categories, and find that overall government
expenditures induce growth, but with welfare expenditures having significantly lower effects compared to what
they call “productive expenditures”.
2 Zhang and Zhang (2004) find the opposite relationship. Ehrlich and Kim (2007) report that - not unsurprisingly
in a growth context - the estimates are sensitive to whether the models condition on initial GDP or not.