What is the appropriate balance of taxes and spending between the central government of a country and the subcentral governments--in the US, state and local government? Countries vary, and there's no one-size-fits-all mode. But Kass Forman, Sean Dougherty, and Hansjörg Blöchliger provide an overview of how countries differ and some standard tradeoffs to consider in "Synthesising Good Practices in Fiscal Federalism Key recommendations from 15 years of country surveys" (OECD Economic Policy Paper #28, April 2020).
Here are a couple of figures to give some background on the underlying issues. On this figure, the horizontal axis is the share of spending done by subcentral governments, while the vertical axis is the share of taxes collected by subcentral governments. Being on the 45-degree line line would mean that these were the same. However, every country falls below the 45-degree line, which means that for every country, some of the revenues spent by subcentral govenments are collected by the central government.
Its interesting to note the different models of fiscal federalism that prevail in various countries. At the far right, Canada is clearly an outlier, with nearly 70% of all government spending happening at the subnational level, and half of all taxes collected at the subnational level. Other countries where about half or more of government spending happens at the subnational level include the US, Sweden, Switzerland (CHE) and Denmark.
Mexico is an interesting case where 40% of government spending happens at the subnational level, but tax revenues collected at that level are very low. Germany (DEU) and Israel are countries with a substantial level of subnational spending that is also nearly matched by the level of subnational taxes--and thus a relatively low redistribution of revenue from central to subcentral governments. Many countries huddled in the bottom left of the figure are low both subnational spending and subnational taxes.
Here's a figure showing the change in these patterns across countries from 1995-2017.
The crossing point of the horizontal and the vertical lines means relatively little change: for example, the US had a small rise in the share of spending happening at the subnational level and a small drop in the share of revenues raised at the subnational level.
Some countries with a big rise in the share of spending happening at the subnational level include Spain (ESP), Belgium, and Sweden. Some countries with a big rise in the share of subnational taxes collected include Spain, Belgium, and Italy. Clearly, Spain stands out as a country that has been decentralizing both government revenue and spending. Conversely, Denmark (DNK) stands out as a country that has been decentralizing government spending, but centralizing the collection of tax revenue. Hungary and Netherlands stand out as countries that have moved toward centralizing their spending, and Hungary in particular seems to be both increasing subnational taxes while decreasing subnational spending.
What are the key tradeoffs here? Forman, Dougherty, and Blöchliger write (citations omitted):
Fiscal federalism refers to the distribution of taxation and spending powers across levels of government. Through decentralisation, governments can bring public services closer to households and firms, allowing better adaptation to local preferences. However, decentralisation can also make intergovernmental fiscal frameworks more complex and risk reinforcing interregional inequality unless properly designed. Accordingly, several important trade-offs emerge from the devolution of tax and spending powers. ...
[D]ecentralised fiscal frameworks allow for catering to local preferences and needs, while more centralised frameworks help reap the benefits of scale. Another key trade-off derives from the effect of decentralisation on the cost of information to different levels of government. While greater decentralisation implies that sub-national governments can access more information about the needs of a constituency at lower cost, it simultaneously increases the informational distance between central and sub-national government. In turn, this may make information more costly from the perspective of the central government, impeding its co-ordination and monitoring functions.
Decentralisation could also engender a costly misalignment of incentives. For example, a “common pool” problem may arise when decentralisation narrows the sub-national revenue base and raises the vertical fiscal gap. In this case, the necessary reliance on revenue sharing with central government to ensure SCG [subcentral government] fiscal capacity may also distort the cost/benefit analysis of sub-national governments—particularly in situations where an SCG realises a payoff without bearing the entirety of the associated cost. Rigid arrangements that entrench fiscal dependence on the central government may drive SCGs to manipulate tax-sharing agreements in order to increase their share while undermining their motivation to cultivate the local tax base. Therefore, the possible efficiency and equity gains from decentralisation are closely linked to mitigating the pitfalls of poorly designed revenue sharing.
What does this mean in practical terms? Their survey of the cross-country research has a number of intriguing findings:
OECD research has found a broadly positive relationship between revenue decentralisation and growth, with spending decentralisation demonstrating a weaker effect ...[D]ecentralisation appears to reduce the gap between high and middle-income households but may leave low incomes behind, especially where jurisdictions have large tax autonomy ...In healthcare, research suggests costs fall and life expectancy rises with moderate decentralisation, but the opposite effects hold once decentralisation becomes excessive (Dougherty et al., 2019). With respect to educational attainment, Lastra-Anadón and Mukherjee (2019) find that a 10 percentage point increase in the sub-national revenue share improves PISA scores by 6 percentage points ...Decentralisation has also been linked to greater public investment, with a 10% point increase in decentralisation (as measured by both SCG spending and revenue share of government total) “lifting the share of public investment in total government spending from around 3% to more than 4% on average”. The investment driven by decentralisation appears to accrue principally to soft infrastructure, that is human capital as measured by education.
In the US version of fiscal federalism, states and local governments face constraints on their borrowing, while the federal government does not. In the case of disruptions from a national recession or pandemic, when a surge of government borrowing is needed, as in the case of a pandemic, it will thus be natural for subnational governments to turn to the US federal government for support. However, it's worth remembering that in more normal times, having state and local governments bear a substantial responsibility for their there own tax and sending levels can have real benefits for accountability and government services.