University efficiency and public funding for higher education in Bulgaria

Abstract

The recent pressure on public finances in Bulgaria has exposed the need for a performance-based system of public funding for higher education. This article estimates the relative technical and cost efficiency of Bulgarian universities and explores the correlation between public funding and efficiency levels. In particular, a recent government proposal to use university rankings for the allocation of funds is evaluated with regard to efficiency. The results indicate that public universities are less efficient than private institutions, especially in teaching-related aspects. A larger share of the education market, fewer fields of study and more science-related majors result in efficiency gains. Efficiency is not a significant determinant of the amounts of subsidy allocated to a university, while the rankings of efficiency and funding are found to be negatively correlated. However, the rankings to be used under the proposed policy are positively related to cost efficiency, suggesting that the reform effort is a step in the right direction.

Acknowledgements

The authors would like to thank Walter McMahon and Boyan Zahariev as well as the participants at the Eurasian Business and Economics Society conference in Zagreb in October 2011 and the International Symposium on the Economics of Education at the Chinese University of Hong Kong in December 2011 for helpful comments and suggestions. Financial support from the Research and Creative Activities Fund (Grant #60609) at Texas Christian University and the International Studies Grant at Texas A&M University – Commerce is gratefully acknowledged.

Notes

 1. We also test the possibility that efficiency results from (rather than being the determinant of) the funding provided but the results indicate that this is not the case.

 2. According to the Bulgarian Ministry of Education, Youth and Science, about 10% of all graduates from secondary school left the country to study at foreign universities in 2009–10. This information is based on the number of school diplomas notarised by the Ministry for admission to institutions of higher education abroad.

 3. Bulgaria has a currency board, which pegs the Bulgarian currency (leva) to the euro. This arrangement drastically limits the scope of monetary policy and imposes requirements for strict fiscal discipline, which has led to painful cuts in public spending during the recent global crisis.

 4. The advantage of SFA over DEA in this regard is that it takes into account stochastic noise; however, it also makes assumptions about the distributional properties of the components of the stochastic term which are often violated (Greene 1999).

 5. In the output-oriented DEA model, inputs have to be positively related with outputs, and therefore the unemployment rate, which is a negative quality indicator, is included in the model in the form of its reciprocal.

 6. It is of little use for a university to churn out graduates who cannot secure jobs. In a competitive market, the reputation of the institution suffers, leading to a decrease in its share of the education market. However, Bulgarian universities are generally not involved in job placement, and career services on campuses have been slow to emerge.

 7. In fact, we initially included two variables to control for the impact of macroeconomic factors on efficiency in the regression model below. However, neither economic growth nor the general unemployment rate in the cities where the universities were located had a significant impact. Therefore, these two variables were dropped from the final regression analysis.

 8. We chose to drop the Academy of the Ministry of the Interior, the Naval Academy, the National Defence Academy and the National Military University from the sample owing to incomplete data. Furthermore, these institutions are financed by the Ministry of Defence and the Ministry of the Interior, in contrast to all other public institutions of higher learning in Bulgaria, which receive funding from MEYS.

 9. Professional schools in Bulgaria are labeled either ‘academy’ (especially in the arts) or ‘institution of higher learning’ (vishe uchilishte). But many also carry the word ‘university’ in their names, which is misleading owing to their narrow academic focus, and should not be confused with the university category we use in this article.

10. The data were released in November 2010 and are available at http://rsvu.mon.bg/.

11. SCOPUS includes journals in languages other than English only if the articles they contain have an abstract in English.

12. We estimated all possible combinations of the outputs listed, but the resulting efficiency levels are very similar to those reported in the article, which prompted us to exclude them from the analysis to save space. These results are available from the authors upon request.

13. The alternative specification of the price of capital as capital expenditure per square metre of floor area produced very similar estimates of the levels of cost efficiency which are not reported here and are available from the authors upon request.

14. The various specifications of our DEA model take into account several variables from each category of criteria except for campus life and prestige, which are based on surveys and are thus less preferable than qualitative measures such as citations of published work or income after graduation. Moreover, campus life and prestige represent only 20% of the overall rank.

15. Regular Bulgarian students and ethnic Bulgarian students from abroad enter the calculation in absolute numbers, while doctoral students are assigned a weight of 2 and long-distance students a weight of 1/3.

16. Art academies and medical schools benefit from larger subsidies per student compared with most other institutions. Treating these as outliers and dropping them from the sample did not alter the correlations with the efficiency rankings significantly.

17. The three measures of technical efficiency include the average of the two teaching-related estimates, the research-related estimate and the average of the three comprehensive estimates from Table 3.

18. Science-related fields are defined as those in the realm of natural sciences, engineering and technology and medicine.

19. In this case the regression is censored from below at zero, which corresponds to the Tobit model specification.

20. The only reason for the coefficient to lack significance is that the government uses only six broader categories of academic fields while we include 51 possible fields of study.