Short‐run distributional effects of public education transfers to tertiary education students in seven European countries

Abstract

Direct provision of public services can alter the balance of resources across income groups. We focus on the issues arising when taking account of the impact of publicly provided education services across the income distribution. We combine OECD information on spending per student in particular levels of the education system with micro data from nationwide income surveys to track the allocation of resources. We pay particular attention to the role of third‐level education, and provide comparable results for seven European countries (Belgium, Germany, Greece, Italy, Ireland, the Netherlands and the United Kingdom).

Keywords:

• inequality
• in‐kind transfers
• tertiary education
• Europe

Acknowledgements

The research was carried out in the framework of the EU‐supported research project ‘Accurate Income Measurement for the Assessment of Public Policies (AIM‐AP)’. The authors would like to thank the project coordinator Holy Sutherland, as well as Joachim Frick, Markus Grabka, Tim Goedemé, Olaf Groh‐Samberg, Christos Koutsambelas, Mattia Makovec, Gerlinde Verbist, Klaas de Vos and Francesca Zantomio for providing a number of estimates reported in the paper as well as offering comments and suggestions on earlier versions of the paper, thank the guest editor of this volume Milena Bevc and, particularly, thank an anonymous referee for useful comments and suggestions.

Notes

1. That is, school is free in the sense that there are no tuition fees; ancillary costs of participation (such as books or school uniforms, etc.) may fall on parents in some countries and may be subsidised in part by governments or schools.

2. The treatment of loans merits further attention, but the subsidy arising from loans is limited to the gap between the interest rate charged and the opportunity cost of funds. See Dearden et al. (2008) on income‐contingent loans employed for paying tuition in Australia and, recently, in the United Kingdom, and other nations, but not in the nations studied here at the time period covered by our study.

3. Where different forms of schooling are on offer to an age group, with differing costs of provision, differences can emerge depending on the pattern of participation across school types.

4. The exception is Belgium, where the pre‐school phase has also been included in the OECD figures. Here, however, we focus on the benefits deriving from public expenditure on primary, secondary and tertiary education only.

5. It may be argued that external benefits are rather small and are mainly limited to the primary years of schooling where basic literacy and civil behaviour are taught. See Wilson, Lambright, and Smeeding (2006) for a discussion of spillover benefits, who calculate that about 10% of public spending for elementary and tertiary schooling creates external benefits to society at large in the United States. These spillovers should be pure public goods and therefore may be considered distributionally neutral.

6. US studies indicate that public school spending per pupil may differ by up to 50% between rich and poor districts (for example, Wilson, Lambright, and Smeeding 2006). In the absence of similar studies, we are unable to consider this factor here; but this is an area warranting further research.

7. Estimates reported in Callan, Smeeding, and Tsakloglou (2007) suggest that the findings of the paper do not change substantially when public transfers to tertiary education students include R&D spending. It should be noted that the data in Table 4 are influenced by the particular method used by the OECD to calculate the average number of years of tertiary education studies and, consequently, the number of full‐time equivalent students in each country.

8. Both indices satisfy the desirable properties for an inequality index (anonymity, mean independence, population independence, transfer sensitivity). Higher values of e make the Atkinson index relatively more sensitive to changes closer to the bottom of the distribution, while, in practice, the Gini index is relatively more sensitive to changes around the median of the distribution (Lambert 2001).