The stimulative effect of an unconditional block grant on the decentralized provision of care


Understanding the impact of central government grants on decentralized healthcare provision is of crucial importance for the design of grant systems, yet empirical evidence on the prevalence of flypaper effects in this domain is rare. We study the decentralization of home care in the Netherlands and exploit the gradual introduction of formula-based equalization to identify the effect of exogenous changes in an unconditional block grant on local expenditure and utilization. A one euro increase in central government grants raises local expenditure by twenty to fifty cents. Adjustments occur through the number of hours as well as through substitution between basic and more advanced types of assistance. These findings suggest that conditioning of grants is not required for the central government to retain a moderate degree of control over the decentralized provision of care.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4


  1. 1.

    Chernew and Newhouse (2012) provide an overview and discussion.

  2. 2.

    Since World War II, many European countries have decentralized healthcare policy to lower levels of government (Tediosi et al. 2009; Mosca 2006). For instance, in the Nordic countries administrative, managerial and substantial political and fiscal responsibilities have been decentralized to municipalities. In Spain and Italy, healthcare responsibilities have been decentralized to the regional level, yet fiscal responsibility has not been decentralized fully (Tediosi et al. 2009; Rico and León 2005).

  3. 3.

    The externalities that arise when redistributive policies are decentralized are discussed in, e.g., Pauly (1973), Boadway and Wildasin (1984) and Wildasin (1991). Essentially, by underproviding redistributive services, subnational governments externalize costs through two channels: needy households may move to other places in which provision is more generous and even if they stay, residents in other places may experience a utility cost when preferences for redistribution transcend jurisdictional boundaries. These externalities may be addressed through matching grants, where the matching rate corresponds to the magnitude of the externality.

  4. 4.

    Brueckner (2000) reviews the debate on funding welfare provision in the USA. See Blank (2002) for a broad evaluation of the reform.

  5. 5.

    Dilger and Boyd (2014) provide an overview of the political debate on block granting Medicaid in the USA.

  6. 6.

    Block grants are used to fund decentralized healthcare provision in Denmark, Finland, Italy, Norway, Spain and Sweden (Rico and León 2005; Kim et al. 2009).

  7. 7.

    As health care is redistributive in nature, decentralization may also induce underprovision through other channels, as discussed in footnote 3.

  8. 8.

    On the basis of a questionnaire, Allers et al. (2013) report that the majority of respondents view intermunicipal differences in the level of social services, such as health care, as socially undesirable and that tasks should remain a central government responsibility if decentralization gives rise to such differences.

  9. 9.

    See Hines and Thaler (1995) and Inman (2008) for an overview of the early literature, as well as a discussion of potential explanations for the flypaper effect.

  10. 10.

    So, for example, basic ADHA could consist of cleaning the fridge, but a provider of advanced ADHA would also check expiry dates of products in this fridge and make an overview of stocks that need replenishing.

  11. 11.

    EMEA also covered other health services, such as specialized hospital services.

  12. 12.

    Personal care consists of assistance in daily activities such as preparation of meals or assistance in taking a shower. People who give personal care sometimes perform activities that typically fall under advanced ADHA.

  13. 13.

    For instance, a household requiring cleaning of their home would receive between one and 1.5 h of ADHA. Typical activities that should be done in this timeframe are cleaning up after meals, doing the dishes, dusting and changing bedsheets. Households with children older than twelve receive fewer hours of ADHA, as teenagers are expected to clean their own room. Households who need assistance in grocery shopping would receive 1 h of ADHA per week for this task. This could be extended to two times 1 h per week when the household has children younger than thirteen or when the household is large. Households are eligible for an additional half an hour of ADHA if the distance to shops is large. See CIZ (2006).

  14. 14.

    Geographical variation in healthcare utilization is a well-known phenomenon—see Skinner (2012) for a discussion of causes and consequences. One potential explanation in the context of ADHA provision is that regional purchasing agencies acquired ADHA largely on the basis of uptake in the past, rather than an estimate of current needs (CEBEON 2005).

  15. 15.

    Often the largest private healthcare insurer within an EMEA-region acts as healthcare purchasing agency.

  16. 16.

    Long-term care provided under EMEA has been the fastest growing type of health care as total public expenditure on healthcare costs in the Netherlands has been growing from 6 to 8.2% of GDP between 1991 and 2005 (CPB 2005).

  17. 17.

    Hence, in the OECD classification, it corresponds to the definition of a block grant (Bergvall et al. 2006).

  18. 18.

    This mechanism is discussed in for instance Tirole (1994).

  19. 19.

    As municipalities gained experience with ADHA provision, their knowledge about local expenditure needs has likely improved and the budgeting process may have become less dependent on information embedded in the grant amount.

  20. 20.

    An alternative explanation, put forward by Baicker (2001) for budget stickiness of state spending on Medicaid in the USA, is fear of a reduction in future grants if decentral governments do not spend current grants on the corresponding item—see also Brennan and Pincus (1996). The fact that the grant allocation for ADHA is formula based and not discretionary limits the applicability of this explanation.

  21. 21.

    Based on a survey collected from 391 municipalities, Van der Torre et al. (2011) report an average price of ADHA of 20.71 per hour for basic assistance and 23.25 euro per hour for advanced assistance, yielding a weighted average about 23 euro per hour.

  22. 22.

    These supplementary grants also smoothed the loss of general grant money due to a change in the rules for the municipal subsidy on VAT and the revision of money received for municipal expenditures on public order and safety (Department of the Interior 2007).

  23. 23.

    See for instance APE (2008), BMC (2008) and Pommer et al. (2009).

  24. 24.

    We scale by dividing through the aggregate counterfactual grant amounts over all municipalities and then multiplying by the total budget in 2007.

  25. 25.

    See for example Barrow and Rouse (2004), Gordon (2004), Brooks et al. (2011) and Allers and Vermeulen (2016). Each of these papers addresses this concern in a different way.

  26. 26.

    It should be borne in mind, though, that changes in needs indicators and vote shares are potentially endogenous and may create a bad controls problem (Angrist and Pischke 2008).

  27. 27.

    The number of municipalities decreased from 443 in 2007 to 408 in 2013 due to municipal amalgamations. We treat municipal amalgamations in retrospect. Thus, if municipalities A and B amalgamated into municipality C in 2008, we treat municipalities A and B as if they had amalgamated in 2007 already. We verify that results are robust to leaving the amalgamated municipalities out of our sample. These results are available from the authors upon request.

  28. 28.

    The reliability of these data may be gauged by comparing them to the administrative records on the number of hours of ADHA provided per capita. The correlation equals 0.80 in 2007, 0.73 in 2010 and 0.66 in 2013.

  29. 29.

    These municipalities moved from a system based on entitlements to inputs (hours of ADHA) to a system based on entitlements to outcomes. Hence, in 2007 they assigned clients a certain number of ADHA, yet in 2013 they assigned clients the right to ‘a clean house.’ As a result, CAO no longer registers the provided hours of ADHA in these municipalities. These municipalities are Alblasserdam, Dordrecht, Emmen, Hendrik-Ido-Ambacht, Papendrecht, Rotterdam, Sliedrecht and Zwijndrecht.

  30. 30.

    Due to a change in their administrative system, the CAO could not deliver the user fees collected in 2007.

  31. 31.

    We collect 2005 values of all variables in the allocation formula, except average income and information on beds in nursing homes and hospitals. The former is first measured at the municipal level in 2006, and the latter is only observed for 2004. For three municipalities average income has not been reported by Statistics Netherlands. For these municipalities, the relative income indicator is set to the municipal average.

  32. 32.

    We define votes for left-wing parties as the votes going to Groen Links, PvdA or SP. We consider D66 and VVD to be right-wing parties. Votes to Christian democratic parties are votes to CDA, CU or SGP. Finally, we group together votes going to local parties and combinations of parties. At the local level, it sometimes happens that right-wing and left-wing parties work together and provide one list of candidates. Such combinations are not separately controlled for as it would inflate the number of indicators and often parties decide to work together because they are expected to collect a very small share of the votes on their own.

  33. 33.

    Assuming an income elasticity of demand for ADHA of one, an increase in the grant for ADHA by one euro should result in an increase in ADHA expenditures by 0.3 cent. See e.g., Inman (2008).

  34. 34.

    For completeness, the first column of Appendix Table 7 shows results based on a reduced form specification in which the change in expenditure is directly regressed on both reforms. We cannot reject the hypothesis that the effect of the first and second reform is equal to our preferred estimate in Table 2.

  35. 35.

    The third and fourth columns of Appendix Table 7 show results based on a reduced form specification in which the change in hours per capita of basic and advanced ADHA is directly regressed on both reforms, which are consistent with results reported in Table 5.

  36. 36.

    The allocation formulas were constructed using the so-called analysis of differences. The idea behind this method is that municipal costs for the provision of ADHA can be divided into costs that are the result of exogenous or endogenous cost drivers. An allocation formula is constructed based on the exogenous costs and exogenous cost drivers only. Municipalities have been grouped according to income (high, low) and share of elderly (high, low) and relevant cost drivers and their weights have been selected by their ability to explain cost differences within and between these groups of municipalities (see Huigsloot 2007). The allocation formulas have been presented to municipalities before officially being published.

  37. 37.

    As an example, the weight for population size was 0.30 in 2008 and 0.31 in 2010.


  1. Allers, M. A., Steiner, B., Hoeben, C., & Geertsema, J. B. (2013). Gemeenten in perspectief, COELO

  2. Allers, M. A., & Vermeulen, W. (2016). Capitalization of equalizing grants and the flypaper effect. Regional Science and Urban Economics, 58, 115–129.

    Article  Google Scholar 

  3. Angrist, J. D., & Pischke, J. S. (2008). Mostly harmless econometrics: An empiricist’s companion. Princeton: Princeton University Press.

    Google Scholar 

  4. APE. (2008). Analyse van de negatieve herverdeeleffecten van het verdeel-model voor de WMO. Den Haag: APE.

    Google Scholar 

  5. Baicker, K. (2001). Government decision-making and the incidence of federal mandates. Journal of Public Economics, 82(2), 147–194.

    Article  Google Scholar 

  6. Baicker, K., & Staiger, D. (2005). Fiscal shenanigans, targeted federal health care funds, and patient mortality. The Quarterly Journal of Economics, 120(1), 345–386.

    Google Scholar 

  7. Barrow, L., & Rouse, C. E. (2004). Using market valuation to assess public school spending. Journal of Public Economics, 88(9), 1747–1769.

    Article  Google Scholar 

  8. Bergvall, D., Charbit, C., Kraan, D.-J., & Merk, O. (2006). Intergovernmental transfers and decentralized public spending. OECD Journal on Budgeting, 5(4), 1–36.

    Article  Google Scholar 

  9. Blank, R. M. (2002). Evaluating welfare reform in the United States. Journal of Economic Literature, 40(4), 1105–1166.

    Article  Google Scholar 

  10. BMC. (2008). Onderzoek naar de uitkomsten van het objectieve verdeelmodel WMO. Emmen/Zwartewaterland: BMC.

    Google Scholar 

  11. Boadway, R. W., & Wildasin, D. E. (1984). Public sector economics. Brown: Little.

    Google Scholar 

  12. Botter, F. M. P. (2010). Maatschappelijke Ondernemingen, naar een andere benadering van maatschappelijk ondernemen. Delft: Eburon.

    Google Scholar 

  13. Brennan, G., & Pincus, J. J. (1996). A minimalist model of federal grants and flypaper effects. Journal of Public Economics, 61(2), 229–246.

    Article  Google Scholar 

  14. Brooks, L., Philips, J., & Sinitsyn, M. (2011). The cabals of a few or the confusion of a multitude: The institutional trade-off between representation and governance. American Economic Journal: Economic Policy, 3(1), 1–24.

    Google Scholar 

  15. Brueckner, J. K. (2000). Welfare reform and the race to the bottom: Theory and evidence. Southern Economic Journal, 66(3), 505–525.

    Article  Google Scholar 

  16. CBS. (2014). Statline electronic database. Visited September 2014.

  17. CEBEON. (2005). Verdeelsleutel decentralisatie eerste tranche middelen Wet maatschappelijke ondersteuning. CEBEON.

  18. Chernew, M. E., & Newhouse, J. P. (2012). Handbook of health economics II, volume 2, chapter 1: Health care spending growth (pp. 1–43). North Holland: Elsevier.

  19. CIZ. (2006). Wmo richtlijn indicatieadvisering voor hulp bij het huishouden, Centrum Indicatiestelling Zorg.

  20. Craig, S. G., & Howard, L. L. (2014). Is Medicaid crowding out other state government expenditure? Internal financing and cross-program substitution. Regional Science and Urban Economics, 64(1), 164–178.

    Article  Google Scholar 

  21. CPB. (2005). Macro-Economische Verkenningen 2005, Centraal Planbureau.

  22. Dahlberg, M., Mörk, E., Rattsø, J., & Ågren, H. (2008). Using a discontinuous grant rule to identify the effect of grants on local taxes and spending. Journal of Public Economics, 92(12), 2320–2335.

    Article  Google Scholar 

  23. Department of the Interior. (2006). Circulaire gemeentefonds van 30 mei 2006. Ministerie van Binnenlandse Zaken en Koninkrijksrelaties.

  24. Department of the Interior. (2007). Circulaire gemeentefonds van 29 juni 2007. Ministerie van Binnenlandse Zaken en Koninkrijksrelaties.

  25. Department of the Interior. (2010). Junicirculaire gemeentefonds 2010, Ministerie van Binnenlandse Zaken en Koninkrijksrelaties.

  26. Department of the Interior. (2014). Septembercirculaire gemeentefonds 2014. Ministerie van Binnenlandse Zaken en Koninkrijksrelaties.

  27. Dilger, R. J., & Boyd, E. (2014). Block grants: Perspectives and controversies. CRS report R40486, Congressional research service.

  28. Electoral Council. (2014). Electronic database municipal elections. visited September 2014.

  29. Evans, W. N., & Owens, E. G. (2007). Cops and crime. Journal of Public Economics, 91(1), 181–201.

    Article  Google Scholar 

  30. Gordon, N. (2004). Do federal grants boost school spending? Evidence from Title I. Journal of Public Economics, 88(9), 1771–1792.

    Article  Google Scholar 

  31. Hines, J. R., & Thaler, R. H. (1995). Anomalies: The flypaper effect. The Journal of Economic Perspectives, 9(4), 217–226.

    Article  Google Scholar 

  32. Huigsloot, P. (2007). H4 The equalised allocation of local expenditure needs in the Netherlands: An optimised mixture of objectivity and politics. In J. Kim & J. Lotz (Eds.), The Copenhagen workshop 2007 meausuring local government expenditure needs (pp. 106—130).

  33. Inman, R. P. (2008). The flypaper effect. Working paper 14579, National Bureau of Economic Research.

  34. Kim, J., Lotz, J., & Mau, N. J. (Eds.) (2009). The Copenhagen workshop 2009 general grants versus earmarked grants theory and practice. The Korea Institute of Public Finance and the Danish Ministry of Interior and Health.

  35. Knight, B. (2002). Endogenous federal grants and crowd-out of state government spending: Theory and evidence from the Federal Highway Aid Program. The American Economic Review, 92(1), 71–92.

    Article  Google Scholar 

  36. Lundqvist, H. (2015). Granting public or private consumption? Effects of grants on local public spending and income taxes. International Tax and Public Finance, 22(1), 41–72.

    Article  Google Scholar 

  37. Lutz, B. (2010). Taxation with representation: Intergovernmental grants in a plebiscite democracy. The Review of Economics and Statistics, 92(2), 316–332.

    Article  Google Scholar 

  38. Mosca, I. (2006). Is decentralisation the real solution? A three country study. Health Policy, 77(1), 113–120.

    Article  Google Scholar 

  39. Pauly, M. V. (1973). Income redistribution as a local public good. Journal of Public Economics, 2(1), 35–58.

    Article  Google Scholar 

  40. Pommer, E., Van der Torre, A., & Eggink, E. (2009). Definitief advies over het Wmo-budget huishoudelijke hulp voor 2009. Den Haag: Sociaal en Cultureel Planbureau.

    Google Scholar 

  41. Rico, A., & León, S. (2005). Health care devolution in Europe: Trends and prospects. Working paper 1, Health Organization Research Norway.

  42. Singhal, M. (2008). Special interest groups and the allocation of public funds. Journal of Public Economics, 92(3), 548–564.

    Article  Google Scholar 

  43. Skinner, J. (2012). Handbook of health economics II, volume 2, chapter 2: Causes and consequences of regional variations in health care (pp. 45–93). North Holland: Elsevier.

  44. Tediosi, F., Gabriele, S., & Longo, F. (2009). Governing decentralization in health care under tough budget constraint: What can we learn from the Italian experience? Health Policy, 90(2), 303–312.

    Article  Google Scholar 

  45. Tirole, J. (1994). The internal organization of government. Oxford Economic Papers, 46(1), 1–29.

    Article  Google Scholar 

  46. Tweede Kamer. (2004). Nieuwe regels betreffende maatschappelijke ondersteuning (Wet Maatschappelijke ondersteuning), 30 131, nr. 3, memorie van toelichting.

  47. Van der Torre, A., Jansen, S., & Pommer, E. (2011). Advies over het Wmo-budget huishoudelijke hulp voor 2012. Sociaal en Cultureel Planbureau.

  48. VNG. (2011). Het PGB in de Wmo, nu en straks, Position paper Vereniging van Nederlandse Gemeenten.

  49. Wildasin, D. E. (1991). Income redistribution in a common labor market. The American Economic Review, 81(4), 757–774.

    Google Scholar 

Download references


We thank two anonymous referees for detailed feedback. We are also indebted to Maarten Allers, Thomas Brändle, Peter Egger, Brice Fabre, Steve Gibbons, Christian Hilber, Jørgen Lotz, colleagues at CPB Netherlands Bureau for Economic Policy Analysis and at Social and Cultural Planning Office, as well as participants of the 2016 meeting of the European Public Choice Society, Freiburg, of the 2015 annual congress of the International Institute of Public Finance, Lugano, of the 2015 Journées Louis-André Gérard-Varet, Aix-en-Provence, and of the 2015 Annual Conference of the Spatial Economics Research Centre at the London School of Economics for insightful comments.

Author information



Corresponding author

Correspondence to Mark Kattenberg.

Appendix: The grant allocation formulas

Appendix: The grant allocation formulas

Tables 7, 8, 9, 10 and 11 show further robustness checks that are discussed in the main text. The remainder of this Appendix provides more detail on the grant allocation formulas used in 2010 and 2013.

Table 7 Reduced form equations
Table 8 Effect on expenditures and service provision conditional on change control variables
Table 9 Effect on expenditures and service provision, asymmetric response
Table 10 Effect on user rate of cash benefits
Table 11 Effect on user fee per hour of ADHA, 2008–2013
Table 12 Allocation formula for ADHA
Table 13 Correlation coefficients main indicators allocation formula and control variables

In Appendix Table 12, we list the indicators and the weights that were used in the ADHA allocation formulas in 2010 and 2013 (Columns 1 and 3).Footnote 36 Column 2 reports the average share of the total grant for ADHA that each indicator allocated in 2010. Column 4 does so for 2013. The indicators of the allocation formulas are measured in levels. Multiplication of the weights with the realized indicators yields the size of the grant for ADHA municipalities receive. As the national budget the central government distributes using the grant for ADHA varies from year to year, the weights in 2008 and 2009 are very similar to the 2010 weights presented in Appendix Table 12, but not exactly identical.Footnote 37 The same holds for the weights in 2011, 2012 and 2013.

In 2011, the allocation formula was substantially revised. Four new indicators were added to the allocation formula. These were the number of people in a municipality who are chronically ill and three indicators that interact the relative income of a municipality with the number of households in various age groups. These new variables allocated more than half of the grant for ADHA on average (see Column 4). Note that before 2011, more than half of the ADHA grant was allocated based on indicators measuring the number of people with low incomes and households with heads aged 75 or older, the number of people on social support, and the number of single person households with heads aged between 75 and 84. Municipalities with many people on social support and many old single person households thus lost from the reform in 2011, whereas municipalities with many people who are chronically ill or with a relatively low average income gained.

Table 13 presents pairwise correlation coefficients for our control variables (measured in 2005) and the main indicators of the allocation formula in 2010 and in 2013. They indicate that the most important indicators of the allocation formula (in both 2010 and 2013) correlate strongly with especially the share of the population that was older than 75, average personal income and the mortality rate in 2005. For these controls, the correlation coefficients often exceed fifty percent. Surprisingly, the share of the population that belongs to a minority group in 2005 correlates strongly (\(\rho >0.4\)) with the indicators that are a function of households with low income and the number elderly households in the municipality. Population density does not correlate strongly with the main indicators of the allocation formula in 2010 or 2013. Nonetheless, we use it as a control because provision of ADHA is believed to more expensive in remote areas with low population density because of travel costs.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kattenberg, M., Vermeulen, W. The stimulative effect of an unconditional block grant on the decentralized provision of care. Int Tax Public Finance 25, 166–199 (2018).

Download citation


  • Intergovernmental transfers
  • Flypaper effect
  • Decentralization of health care

JEL Classification

  • H42
  • H51
  • H71
  • H75