The Impact of Direct Democracy on Public Education(b)


2 Education Quality and Direct Democracy
2.1 Direct Democracy

In modern (semi)direct democracies, a representative democracy is complemented by direct democratic institutions. The most prominent cases are Switzerland and the United States,which are both also shaped by a very strong fiscal decentralization, with each level having its own sources of tax revenue. Therefore, there exists a direct institutional link between the power to tax and the power to spend, as direct legislative institutions provide citizens with the political means to influence both sides of the budget equally. In Switzerland, popular rights can be exerted at all three levels of the state (federal, cantonal, and communal).

Cantonal constitutions differ with respect to the degree of direct democracy, which is exerted through initiatives and referenda. In addition, Switzerland is one of the most politically decentralized countries in the world, and the organization and execution of public education is among the core competencies of Swiss cantons (Germann, 2002; Lijphart, 1999: 38). The fundamental regulations of public education, school organization and the financial contributions of each state level are laid down in various cantonal laws on public education, which are subject to voters’ direct influence through statutory initiatives and referenda. In consequence, since all 26 cantons differ with respect to the degree of direct democracy in their constitutions, it is possible to analyze the impact of a change in citizen empowerment on the provision of public schooling (Feld & Kirchgässner, 2001).
Swiss cantons are not only responsible for the organization of (compulsory) public
education but bear the financial burden for its provision (Swiss federal constitution, art. 62).
Concerning the overall costs of compulsory education, the federal government contributes
only 0.2%, whereas the cantons bear 38.8% and the communes 61.1%.2 With respect to the communes, they mostly finance primary schools. In general, in all 26 Swiss cantons, two  types of advanced education can be distinguished: basic education and education to meet advanced requirements (e.g. university preparation). Usually, the second type can only be entered on a selective basis, either through passing an entry examination or having obtained a certain average grade in the prior school year. However, as public education is in the authority of cantons, virtually 26 different school systems coexist within Switzerland, differing in their financial structure, organization and school curricula (Freitag & Bühlmann, 2003; Meunier,2004). For example, in some cantons secondary I education commences after completion of the sixth year of schooling, while in others it starts after the forth or fifth year. However, in all cantons compulsory education, that includes primary and secondary I education, finishes with the ninth grade, excluding secondary II education. This distinction is important as the dataset which is employed in this study is based on student performance of ninth graders, at the end of compulsory education.

At the time of the study, there was a vast heterogeneity across cantons in the organization
of teacher education.3 Training of primary and secondary I teachers took place at over 150
cantonal teacher seminaries, and the regional validity of the earned teacher certificate
impeded mobility of teaching personnel across Swiss cantons. In Switzerland, the duration of teacher training in these teacher seminaries varied regionally between one and four years. In  contrast, teachers for secondary II schools were educated at universities, where an at least four-year lasting study led to a subject-specific master’s degree, e.g. a M.Sc. in Biology (supplemented by pedagogical courses). Whereas admission to university required a secondary II degree obtained after twelve to thirteen years of schooling, a primary or
secondary I teacher in spe started her training after completion of compulsory education,
namely after just nine years of schooling. Overall, the educational gap between these two
types of teachers in Switzerland amounted to approximately four to five years of education.
And even qualitatively, those who qualified for university studies had passed at least two
selection processes during schooling, in contrast to those attending teacher seminaries (see
EDK, 2001).

2.2 The literature and hypotheses

Institutions of direct legislation, argue their many supporters, serve as a means to discipline politicians and bureaucrats, who are assumed to behave in a Niskanen-like manner rather than as benevolent dictators (Niskanen, 1975). Specifically, these bureaucrats exercise monopoly power and aim at maximizing their budgets.4 In consequence, means of direct legislation are thought to limit the government spending without negatively affecting the quality of the provided public goods, indicating an allocation of goods and resources closer to the median voter’s preferences through reducing waste (Feld & Kirchgässner, 2001; Pommerehne, 1978,1983; Besley and Coate, 2001).5 In the US, people in favor of the introduction of school budget-restraining tax limits actually did believe that these budget cuts would lead to such efficiency gains in the provision of schooling (Temple, 1996). That, in principle, in economic reality such potential for improvement might exist can be concluded from the beneficial impact of competition among public schools, which lowered per pupil spending, but equally raised student test scores (Hoxby, 2000).

Local revenue and thus school budgets, in particular, appeared considerably reduced in the
U.S. through the introduction of property tax limits, often through statutory initiatives (Card & Payne, 2002; Bradbury, Mayer, & Case; 2001; Shadbegian, 1999, 2003).6 As regards direct democracy, for Switzerland various studies show that it leads to both smaller revenue and smaller expenditure of (combined) cantonal and communal budgets (Kirchgässner, 2002; Feld and Matsusaka, 2003; Hug, 2004). This effect is stronger for those policy areas in which Swiss cantons are granted political autonomy by the Swiss constitution (art. 3 of the Swiss federal constitution), specifically, the health system, culture, and education (Germann, 2002;Schaltegger, 2001).7 In particular, in Switzerland government’s educational expenses for several school types were found to be lowered by direct democracy (Fischer, 2005, 2005a). In contrast, Freitag & Bühlmann (2003) find no linkage between effectively used direct democratic institutions and educational expenses of Swiss cantons, while for the US Santerre (1989) identifies even an educational spending increasing influence in town-meeting type communities among 90 jurisdictions in Connecticut. On the other hand, Megdal (1983) reports ambiguous results for the effect of the existence of school budget referenda in 177 New Jersey school districts.8 Some of these contradicting findings can, however, be explained by incomparability of institutions or level of analysis.9

In response to such budget constraints, in jurisdictions with stronger popular rights a
reliance on user charges was observed that makes the quality of the public good more
independent of the financial resources of the government (Feld & Matsusaka, 2000;
Matsusaka, 1995). However, in the case of compulsory and free public schooling, this
solution is (politically) not an option. Consequently, if the school administration were already working efficiently prior the decision to cut its budget, a decline in the quality of the public service should be revealed even though the school administrator in charge acted like a ‘benevolent dictator’.

Modern economic theories of bureaucracies such as that of Niskanen, however, assume a
selfish, namely budget-maximizing behavior of the administrator (e.g. Inman, 1979).
Extending these theories, limitation of the budget is suggested to give rise to two different
adaptive strategies. The first is a substitution of budget maximization with a re-allocation of  means between budget components such that administrative staff is increased at the expense of the resources available for production of the public good (Williamson, 1964). The rationale for this strategy is that a large personal staff gives the bureaucrat a feeling of importance and power (Downs, 1967). Thus, when facing the decision to cut either administrative or instructional spending, a Leviathan-like school administrator is expected to choose the latter.
Indeed, for U.S. school districts, Figlio (1998) shows that the above mentioned tax limits are associated with a spending shift from the teaching component to the administration
component of school budgets (see also Dye & McGuire, 1997; Figlio, 1997).10 Equally, such
spending shift was then found to be mirrored by larger class sizes, higher pupil-teacher ratios,lower beginning teachers’ wages, worse teacher qualification, unchanged level of school service, but, on the other hand, ongoing overstaffing of administrations (Downes, 1996;Figlio, 1998, 1997, 1997a; Shadbegian, 2000, 2003; Poterba & Rueben, 1995).
Alternatively, Figlio & O’Sullivan (2001) propose a manipulative bureaucratic behavior
in which the administrator deliberately allows the quality of the public good to decline by
allocating (relatively) fewer financial resources to its provision. Being persuaded that budget limitation has a deleterious impact on the quality of the public good, the administrator expects the electorate to overrule (‘override’) the previous tax limit vote in the next election. Indeed,Figlio & O’Sullivan (2001), using expense data for police, fire protection, and general administration from 5,150 U.S. cities, show that in those cities with a so-called ‘override option’, the deterioration in public service was larger than in cities without this option.
Moreover, they observe the same phenomenon with respect to teacher-administrator ratios in school districts with an override option. 11


Justina A.V. Fischer*
Stockholm School of Economics
SSE/EFI Working Paper Series in Economics and Finance
No 688-December 2007


1. The average test score of 494 for Switzerland was below the international mean of 500 for the PISA study.
2. Appropriate information on this issue can be found at, or the Federal Statistical Office,
3. In contrast to these organizational differences across cantons, school curricula in primary and secondary I stages of education have been harmonized to a great extent.
4. See Romer & Rosenthal (1978, 1979, 1982, 1983) and Romer, Rosenthal, & Munley (1992).
5. Efficiency gains serve as explanation for the growth-improving impact of direct democracy detected by Feld & Savioz (1997).
6. The source ACIR (1995) provides a catalogue of existing tax limits. In contrast, Poterba (1997) does not report any significant influence of property tax limits on per pupil K-12 school spending by US states. Besides missing some political determinants in his model, it might well be that analysis at the aggregate level prevented identification of a significant impact.
7. Available at (10. Nov 2006)
8. For Switzerland, also Grob & Wolter (2005) analyze determinants of educational spending across cantons, but with a focus on socio-demographic characteristics solely. In their analysis, political institutions are missing.
9. Freitag & Bühlmann (2003) employ the number of held fiscal referenda and initiatives (in place of using the mere presence of institutions) with spending at the cantonal level as regressor, thus neglecting the financial contribution of local jurisdictions. Similarly, the data used by Santerre (1989) and Megdal (1983) are only crosssectional and obtained from local jurisdictions, among which fierce competition might prevent any differential impact of political institutions. Moreover, Sass (1991), using the identical dataset as Santerre (1989), detects no institutional impact once endogeneity of government structure is accounted for.
10. Dye & McGuire (1997) identify a mitigating effect of strong competition between jurisdictions.
11. It is, however, questionable whether a change in the ratio of administration to production costs provides sufficient evidence for one of the two theories. If instruction costs are more variable than administrative costs, in the short term only a cut in instruction costs might be practical.

About sooteris kyritsis

Job title: (f)PHELLOW OF SOPHIA Profession: RESEARCHER Company: ANTHROOPISMOS Favorite quote: "ITS TIME FOR KOSMOPOLITANS(=HELLINES) TO FLY IN SPACE." Interested in: Activity Partners, Friends Fashion: Classic Humor: Friendly Places lived: EN THE HIGHLANDS OF KOSMOS THROUGH THE DARKNESS OF AMENTHE
This entry was posted in News and politics and tagged , , , , . Bookmark the permalink.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.