Abstract
This paper investigates the relationship between fiscal federalism and the sizes of local governments. While many empirical studies emphasized that grants encourage the growth of local public spending and local taxes constrain it, they are more silent regarding the effects of different types of tax autonomy. The paper addresses this issue by arguing that tax decentralization as organized on tax bases used only by local governments (tax-separation), rather than on tax-base sharing, would restrain local public expenditures. Using an unbalanced panel of OECD countries, the key finding is that only property taxes—mostly based on a “tax-separation” scheme—seem to favor smaller local governments. Thus, while tax decentralization is a necessary condition for limiting the growth of local governments, it does not appear sufficient, as tax-separation schemes among government levels would in fact be required.
Similar content being viewed by others
Notes
However, this theoretical prediction is not clear-cut. More accountability of office-holders could lead voters to support more spending under decentralization than under centralization.
Even though “tax separation” between different tiers of government appears to be a more orderly approach (Musgrave 1983), in practice, models of tax-base sharing—where the same tax base is taxed by more than one level of government—seem to prevail.
Tax sharing identifies a system wherein the central government redistributes a fraction of the taxes that originate from the local tax bases to local government. Revenue sharing, instead, identifies an alternative mechanism wherein the central government transfers a certain amount of given central tax collections on the basis of indicators of local governmental needs.
It is worth noting that among the studies focusing on the effects of decentralization on the sizes of sub-central governments (LG in Table 1), some refer to aggregate local governments (LG1), while others refer to specific government levels (e.g., counties, provinces, and cantons, denoted LG2). The logic of these sub-categories is different. In the first case, the total effect of decentralization is measured without checking for possible different impacts on specific sub-central governments, which is typical in cross-country studies. In the second case, which is adopted frequently in single-country studies, the sign and the explanatory power of the decentralization impact cannot be extended to all sub-central units.
“Overall intrusion of government into the economy may be smaller, the greater the extent to which taxes and expenditures are decentralized” (Brennan and Buchanan 1980). In general, in the public choice perspective of “excessive government,” tax competition reduces rents, and hence a smaller public sector enhances overall welfare (Rodden 2003).
With regard to tax revenue composition, it is instead found that income taxes are negatively affected by fragmentation while capital taxes would be positively influenced.
Moreover, Keen and Marchand (1997) show that fiscal competition may lead not only to inefficient levels of aggregate public expenditure, but also to systematic inefficiencies in the composition of public expenditure.
In Iceland, one’s own taxes amount to 90 % of the total, while in The Netherlands they cover just about 13 %. Moreover, federal countries allocate a slightly larger tax share to sub-central governments than unitary countries.
In general, the main difficulty in assigning GST to sub-national level is that the debiting and crediting of VAT is likely to differ across sub-central jurisdictions, which usually implies an arbitrary apportionment of VAT revenues across jurisdictions. Another difficulty is that, in the European Union, VAT cannot be differentiated regionally.
This definition excludes local non-tax revenues and local capital revenues which are recorded irregularly.
This corresponds to subcategory LG1 in Table 1.
This estimator is thought to be more efficient, especially when having 10 to 20 panels and 10 to 40 time periods, which is basically the case for our dataset. In this case a Prais–Winsten estimator can also be used, which—according to Beck and Katz (1995)—may perform more satisfactorily.
The null of no panel-level heteroscedasticity and the null of no cross-sectional correlation are both rejected (at the 1 % significance level). The null of no autocorrelation within panels is accepted for Austria only.
Note, however, that fixed effects have a more problematic interpretation in our case, as the sample does not represent a closed and exhaustive set of units. The Hausman test rejects the random effect estimator.
In order to verify whether the result depends on the United States where sales taxes are levied only by states, the same regression has been run excluding that country. The results do not change. With regard to GST, Musgrave (1983) suggested that taxes on production and consumption over a large tax base—particularly as the general sales tax—should be treated differently depending on the production stage to which they are applied. In this vein, the assignment to sub-national levels is recommended only for taxes in the end-stage of the production/distribution process, unless local governments are too fragmented. Indeed, local authorities should be large enough to overcome the mobility issue (as in the United States and Canada). Indeed, where applied, state/regional governments rely more on taxes on goods and services (mainly sales taxes) than local governments do (OECD 2009a).
Note, however, that the model of Keen and Kotsogiannis (2002) is a framework wherein intergovernmental transfers are absent and the total tax revenue that is obtained by the central government when taxing the shared tax base is given back to local governments on an equal sharing basis. We cannot assume that this particular form of intergovernmental relationship applies in the countries considered here.
Estimates are not reported in the paper; they are available upon request.
Some points are worth noting. First of all, the share of local spending over total spending on education, communications and housing is on average above 60 %; the same share is above 50 % for transport, about 46 % for health and, as expected, around 20 % for social security and welfare. Secondly, the within-country variability of these shares is rather low, with few exceptions (Austria for communications and housing, Greece for social security and welfare, Iceland for health). This means that in our dataset, the within-country task assignment is significantly stable over time. Thirdly, cross-country variability is instead relatively greater, in particular for health, social security, and welfare, but seems to be driven by specific countries (e.g., Iceland and Switzerland for health). Indeed, and this is the fourth point, cross-country variability is smaller for education, transport, communications, and housing, which are also the functions that are more likely to be decentralized and with, on average, larger shares of local expenditure.
It is, however, worth noting that the availability of data for sub-categories of local spending is not as wide as in the previous case; thus, this new regression entails some loss of observations. We have used 303 observations in the case of health and welfare spending over GDP as dependent variables and 237 observations considering the last two spending groups.
Control variables are not reported in Table 9 but they are included in the estimation.
As suggested by Roodman (2009), the difference-GMM instruments differences with levels, while the system-GMM would add the possibility of instrumenting levels with differences, provided that the first-differences of the instrumental variables are uncorrelated with fixed effects. This may be particularly useful when the dependent variable and the explanatory variables are persistent, causing lagged levels to be weak instruments, and when it is difficult to obtain “external” instruments for the variables that are included in the benchmark specification. Hence, as our panel dataset suffers from lack of suitable exogenous candidates, we rely upon the lag-structure of our dependent and independent variables to build a matrix of suitable instruments in order to gain identification (and over-identification) of the parameters of interest.
The results are obtained by imposing the second lag of the endogenous variables as an instrument, consistently with the fact that the Sargan test may weaken with a larger number of instruments when the number of countries is relatively small. Furthermore, while the first lag is correlated with the current error-term, the second is not.
As for the robustness of these findings, the Sargan test does not reject the validity of instruments; while the Arellano and Bond (1991) tests for autocorrelation of the first and second order also give the right outcome; this suggests that the structure of lags is consistent.
This transformation implies that the average of all future available observations is subtracted from the current observation and it is particularly useful with unbalanced panels. It is computable for all observations, excepting the last, with a minimization of the data loss.
The long-run coefficient is obtained from the long-run equation, implying that λy=−[θ 1 x+θ 2 z+θ 3 k+θ 4 g]. The non-stationarity of the series and the stationarity of their first differences are proved by a panel version of the Augmented Dickey-Fuller test as developed by Im et al. (1997). The null of non-stationarity cannot be rejected for variables in levels, while it can be rejected for their first differences (on average). The results of the test are not reported in the table.
References
Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. Review of Economic Studies, 58(2), 277–297.
Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics, 68(1), 29–51.
Ashworth, J., Galli, E., & Padovano, F. (2009). Government decentralization and public spending programs. Paper presented at the XXI conference of the Italian Society of Public Economics (SIEP).
Bahl, R., & Bird, R. (2008). Subnational taxes in developing countries: the way forward. Public Budgeting and Finance, 28(4), 1–25.
Bahl, R. W., & Linn, J. (1992). Urban public finance in developing countries. New York: Oxford University Press.
Banerjee, A., Dolado, J. J., Hendry, D. F., & Smith, G. W. (1986). Exploring equilibrium relationship in econometrics through static models: some Monte Carlo evidence. Oxford Bulletin of Economics and Statistics, 48(3), 253–277.
Beck, N., & Katz, J. N. (1995). What to do (and not to do) with time-series cross-section data. American Political Science Review, 89(3), 634–647.
Besley, T., & Case, A. (1995). Incumbent behaviour: vote-seeking, tax-setting and yardstick competition. American Economic Review, 85(1), 25–45.
Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1), 115–143.
Boetti, L., Piacenza, M., & Turati, G. (2010). Decentralization and local governments’ performance: how does fiscal autonomy affect spending efficiency? Working Paper No. 11. Department of Economics and Public Finance.
Bordignon, M., & Minelli, E. (2001). Rules transparency and political accountability. Journal of Public Economics, 80(1), 73–98.
Borge, L. E., & Rattso, J. (2008). Property taxation as incentive for cost control: empirical evidence for utility services in Norway. European Economic Review, 52(6), 1035–1054.
Brennan, G., & Buchanan, J. (1978). Tax instruments as constraints on the disposition of public revenues. Journal of Public Economics, 9(3), 301–318.
Brennan, G., & Buchanan, J. (1980). The power to tax: analytical foundations of a fiscal constitution. Cambridge: Cambridge University Press.
Breton, A. (1987). Towards a theory of competitive federalism. European Journal of Political Economy, 3(1-2), 263–329.
Brülhart, M., & Jametti, M. (2006). Vertical versus horizontal tax externalities: an empirical test. Journal of Public Economics, 90(10–11), 2027–2062.
Cassette, A., & Paty, S. (2010). Fiscal decentralization and the size of government: a European country empirical analysis. Public Choice, 143(1–2), 173–189.
Crowley, G. R., & Sobel, R. S. (2011). Does fiscal decentralization constrain Leviathan? New evidence from local property tax competition. Public Choice, 149, 5–30.
De Mello, L. (2001). Fiscal federalism and government size in transition economies: the case of Moldova. Journal of International Development, 13, 255–268.
Dollery, B., & Worthington, A. (1996). The empirical analysis of fiscal illusion. Journal of Economic Surveys, 10(3), 261–297.
Ehdaie, J. (1994). Fiscal decentralization and the size of government, an extension with evidence from cross-country data. Policy Research Working Paper 1387. World Bank.
Fausto, D. (1996). Il federalismo fiscale: un falso rimedio nel caso italiano. Economia delle scelte pubbliche, 2–3, 175–196.
Feld, L. P., Kirchgässner, G., & Schaltegger, C. A. (2003). Decentralized taxation and the size of government: evidence from Swiss state and local governments. CESifo Working Paper Series 1087.
Fischel, W. A. (2001). The homevoter hypothesis: how home values influence local government taxation. School of Finance and Land-Use Policies. Cambridge: Harvard University Press.
Fiva, J. H. (2006). New evidence on the effect of fiscal decentralization on the size and composition of government spending. FinanzArchiv/Public Finance Analysis, 62(2), 250–280.
Forbes, K. F., & Zampelli, E. M. (1989). Is Leviathan a mythical beast? American Economic Review, 79(3), 568–577.
Glaeser, E. L. (1996). The incentive effects of property taxes on local governments. Public Choice, 89(1–2), 93–111.
Grossman, P. J. (1989). Fiscal decentralization and government size: an extension. Public Choice, 62(1), 63–69.
Grossman, P. J., & West, G. (1994). Federalism and the growth of government revisited. Public Choice, 79(1–2), 19–32.
Im, K. S., Pesaran, M. H., & Shin, Y. (1997). Testing for unit roots in heterogeneous panels. Working Paper. University of Cambridge.
Jin, J., & Zou, H. (2002). How does fiscal decentralization affect aggregate, national, and subnational government size. Journal of Urban Economics, 52(2), 270–293.
Keen, M., & Kotsogiannis, C. (2002). Does federalism lead to excessively high taxes? American Economic Review, 92(1), 363–370.
Keen, M., & Kotsogiannis, C. (2004). Tax competition in federations and the welfare consequences of decentralization. Journal of Urban Economics, 56(3), 397–407.
Keen, M., & Marchand, M. (1997). Fiscal competition and the pattern of public spending. Journal of Public Economics, 66(1), 33–53.
King, D. (1984). Fiscal tiers—the economics of multi-level government. London: Allen and Unwin.
Marlow, M. (1988). Fiscal decentralization and government size. Public Choice, 56(3), 259–269.
Martinez-Vazquez, J. (2008). Revenue assignments in the practice of fiscal decentralization. In N. Bosch & J. M. Durán (Eds.), Fiscal federalism and political decentralization: lessons from Spain, Germany, and Canada (pp. 27–55). Cheltenham: Edward Elgar.
McKenzie, R. B., & Staff, R. J. (1978). Revenue sharing and monopoly government. Public Choice, 33(3), 93–97.
McLure, C. E. (1998). Tax assignment problem: conceptual administrative considerations in achieving sub-national fiscal autonomy. Paper presented at the Intergovernmental Fiscal Relations and Local Financial Management Course, Vienna, Austria.
Musgrave, R. (1959). The theory of public finance. New York: McGraw-Hill.
Musgrave, R. (1983). Who should tax, where and what. In C. E. McLure (Ed.), Tax assignment in federal countries (pp. 2–19). Centre for Research on Federal Financial Relations. Canberra: Australian National University.
Nelson, M. A. (1986). An empirical analysis of state and local tax structure in the context of the Leviathan model of government. Public Choice, 49(3), 283–294.
Oates, W. E. (1972). Fiscal federalism. Aldershot: Gregg Revivals.
Oates, W. E. (1985). Searching for Leviathan: an empirical study. American Economic Review, 75(4), 748–758.
Oates, W. E. (2001). Property taxation and local government finance. Essays in Honor of C. Lowell Harriss. Cambridge: Lincoln Institute of Land Policy.
Oates, W. E. (2005). Toward a second-generation theory of fiscal federalism. International Tax and Public Finance, 12, 349–373.
OECD (2009a). The fiscal autonomy of sub-central governments: an update. Paris: Network on Fiscal Relations across Levels of Government.
OECD (2009b). Tax and grants: on the revenue mix of sub-central governments. Paris: Network on Fiscal Relations across Levels of Government.
Pereira, P. T. (2000). Fiscal decentralization, public sector size and the wealth of nations. In Economia Pùblica Regional e Local. Coimbra: APDR.
Prohl, S., & Schneider, F. (2009). Does decentralization reduce government size? A quantitative study of the decentralization hypothesis. Public Finance Review, 37(6), 639–664.
Rodden, J. (2003). Reviving Leviathan: fiscal federalism and the growth of government. Industrial Organization, 57(4), 695–729.
Rodrik, D. (1998). Why do more open economies have bigger governments? Journal of Political Economy, 106(5), 997–1032.
Roodman, D. (2009). How to do xtabond2: an introduction to “difference” and “system” GMM in Stata. Stata Journal, 9(1), 86–136.
Salmon, P. (1987). Decentralization as an incentive scheme. Oxford Review of Economic Policy, 3(2), 24–43.
Shadbegian, R. J. (1999). Fiscal federalism, collusion, and government size: evidence from the states. Public Finance Review, 27(3), 262–281.
Shah, A. (1999). Fiscal federalism and macroeconomic governance: for better or worse? Policy Research Working Paper No. 2005. World Bank.
Stein, E. (1999). Fiscal decentralization and government size in Latin America. Journal of Applied Economics, 11(2), 357–391.
Zax, J. S. (1988). The effects of jurisdiction types and numbers on local public finance. In H. S. Rosen (Ed.), Fiscal federalism: quantitative studies (pp. 79–106). Chicago: Chicago University Press.
Zax, J. S. (1989). Is there a Leviathan in your neighborhood? American Economic Review, 79(3), 560–567.
Wallis, J. J. (1991). The political economy of new deal fiscal federalism. Economic Inquiry, 29(3), 510–524.
Wallis, J. J., & Oates, W. E. (1988). Decentralization in the public sector: an empirical study of state and local government. In H. S. Rosen (Ed.), Fiscal federalism: quantitative studies (pp. 5–32). Chicago: Chicago University Press.
Weingast, B. R. (2009). Second generation fiscal federalism: the implication of fiscal incentives. Journal of Urban Economics, 65(3), 279–293.
Wickens, M. R., & Breusch, T. S. (1988). Dynamic specification, the long run and the estimation of transformed regression model. Economic Journal, 98(309), 189–205 (Supplement Conference Papers).
Winer, S. L. (1983). Some evidence on the effects of the separation of spending tax decisions. Journal of Political Economy, 91(1), 126–140.
Acknowledgements
We would like to thank B. Bises, G. Messina, three anonymous referees and the editor-in-chief for their discerning comments. Suggestions from participants at the 18th Symposium on Public Economics held in Malaga on 3–4 February 2011 are gratefully acknowledged. Remaining errors are our responsibility.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Liberati, P., Sacchi, A. Tax decentralization and local government size. Public Choice 157, 183–205 (2013). https://doi.org/10.1007/s11127-012-9937-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11127-012-9937-9