system. Indeed, in the OECD country sample, different components of social welfare impact
educational investment in different ways. Arguably, our results even suggest in tendency that
student effort would increase when means-tested family allowance programs targeted at poor
households with children were extended. However, we also have to acknowledge that our
findings are for high- and middle-high income countries only - leaving the question
unanswered to what extent least developed countries are affected by such redistributive
government involvement in the economy.26 For the least developed countries which are stuck
in poverty traps and to which additional economic growth may be most beneficial, empirical
research on the welfare state effects is still lacking, and different policies may appear as
optimal.
Acknowledgement
We thank Fredrik Carlsen, Urs Fischbacher, Stefan Hupfeld, Thomas Lange, Bjarne Stem,
Ludger Woβmann, and seminar participants at the CESifo Area Conference on Employment
and Social Protection, 2008, Economics Departments at the University of Konstanz and the
Norwegian University of Science and Technology for helpful comments and suggestions.
Justina Fischer acknowledges a EU Marie Curie experienced researcher fellowship (RTN
ENABLE) while visiting the Stockholm School of Economics.
26 Indeed, also Bjornskov et al. (2007) argue that their well-being lowering influence of size of government
sector may be driven by exclusion of least developed countries, possible disguising an inverted U-shaped
relation.
27