There are a few options. Try, please, look at phenomenon of local government services in terms of delivery. H. Wollmann wrote quite recently a lot on this, e.g. http://amor.cms.hu-berlin.de/~h0598bce/docs/HW-2012-Local-Government-Reforms.pdf, http://en.iju.hr/ccpa/ccpa/downloads_files/001-Wollmann.pdf, http://www.ub.edu/graap/Final%20Papers%20PDF/Wollmann%20Hellmut.pdf ... or try to find an answer in: H. Wollmann - G. Marcou (2010): The Provision of Public Services in Europe: Between State, Local Government and Market. Edward Elgar. But, please, keep in mind that efficiency is influenced also by other factors, e.g. financial and functional capacities or levels of decentralization. From this perspective you can find many different outcomes between the countries in the field of quality of local government services (including their delivery) - please, look for instance at: P. Swianiewicz (2010): Territorial Consolidation Reforms in Europe. OSI/LGI (you can download if for example from here: http://147.232.5.162/krvam/files/clanky/klimovsky/klimovsky_02.pdf).
You may calculate efficiency scores using stochastic frontier analysis or data envelopment analysis.
Another way is to get satisfaction data, i.e. surveys asking respondents how satisfied they are with the provision of local (or national) public services.
In a number of current research papers we match individual senators’ voting behavior on legislative proposals with real referendum decisions on exactly the same issues with identical wording. Thus, we measure directly what constituents want and what representatives actually do (at least for the national level but with differences between districts). This setting allows us to analyze which factors influence the behavior of politicians towards their voters. Moreover we may evaluate theoretical predictions such as the median voter model’s quality.
I hope these suggestions are helpful.
Article Evaluating the median voter model's explanatory power