Revisiting Oates After Five Decades

The federal share of government spending has inched up steadily since 2002. This shift has concrete implications: when the federal government funds local infrastructure projects, communities may get roads they don’t prioritize while neglecting the parks or schools they actually want. Understanding whether this centralization trend reflects economic efficiency or political drift has direct bearing on debates over everything from education funding to climate policy.

I evaluate increased centralization using Wallace Oates’ seminal work on fiscal federalism and ask whether a rising federal share is consistent with optimal public finance.Oates, Wallace E. Fiscal Federalism. Harcourt Brace Jovanovich, 1972. The short answer: probably not, but there are important caveats.

How Should We Measure Centralization?

The traditional Oates framework assumes a clean division of labor: local governments provide locally-consumed public goods, while the federal government provides uniform national public goods. This assumption underlies the use of federal expenditure share as a measure of centralization. However, American governance has never fit this tidy model, and the mismatch has grown more pronounced over time.Even in 1972, Oates recognized this tension, worrying that “economic theory is becoming increasingly irrelevant to the problems of fiscal federalism.”

The federal government has long been in the business of place-based policy. On the spending side, it channels federal money to particular locations through grants (like the Community Development Block Grant), infrastructure projects (like the Interstate Highway System), and education (through community colleges and universities).Hanson, Gordon H., Dani Rodrik, and Rohan Sandhu. “The U.S. Place-Based Policy Supply Chain.” Working Paper. Working Paper Series. National Bureau of Economic Research, February 2025. https://doi.org/10.3386/w33511. On the tax side, the progressive income tax effectively imposes a higher tax burden on households in highly productive or unattractive locationsAlbouy, David. “The unequal geographic burden of federal taxation.” Journal of Political Economy 117, no. 4 (2009): 635-667., and the SALT deduction has a large effect on the net-of-tax price of local public goods in high-tax states.Ambrose, Brent W., and Maxence Valentin. “Federal tax deductions and the demand for local public goods.” Review of Economics and Statistics (2024): 1-39.

This complexity raises a fundamental question: Does the federal expenditure share meaningfully capture centralization trends? There are two good reasons to put some stock in the federal share as a benchmark, but they come with important caveats.

First, even when federal spending is geographically targeted, it often reflects federal rather than local priorities. Federal infrastructure grants may build highways communities don’t want while neglecting the parks or schools they actually need. This substitution of federal for local judgment is precisely what the centralization concern captures.

Second, if federal place-based spending is roughly constant, we can still learn something from changes in the federal share. Place-based federal policy affects the level of federal current spending, not the trend, so a rising federal share appears to indicate a shift towards centralization.

The Oates Framework Reconsidered

Oates’ framework emphasized that decentralized provision of public goods generates efficiency gains under two conditions: when households have heterogeneous preferences over public goods; and when households are sorted geographically on the basis of those preferences. In the absence of spillovers and economies of scale, his “decentralization theorem” showed that variable and decentralized provision dominates uniform central provision.Oates, Wallace E. Fiscal Federalism. Harcourt Brace Jovanovich, 1972.

The past five decades have witnessed dramatic changes in the economic, technological, and social landscape. Despite these changes, the federal government’s share of current expenditures has followed a simple path: after gradually declining for three decades, it began to inch upwards steadily in 2002. The share is now at its highest level since 1988.This increase is not caused by rising spending on entitlement programs: transfer payments are not counted as production and are therefore excluded from the NIPA measures of consumption expenditures. Intergovernmental transfers and interest payments are also excluded. “Chapter 9: Government Consumption Expenditures and Gross Investment.” NIPA Handbook. Bureau of Economic Analysis, December 2024. https://www.bea.gov/resources/methodologies/nipa-handbook/pdf/. The federal share of current expenditures

Is a larger federal share good or bad? Oates described the key economic fundamentals that determine the optimal balance between federal and local spending. I revisit these fundamentals and ask if they’ve changed since 2002 in a way that would justify a rising federal share.

Factor Mobility and Sorting

Labor mobility is likely to increase the efficiency gains from decentralization. When households “vote with their feet” and sort themselves into jurisdictions that match their preferences for public goods, there will be a closer alignment between the provision of public goods and the residents’ preferences. Think of the vast difference in demand for green energy between California and Oklahoma, or the different preferences for clean public spaces among New Yorkers and retirees in Florida.

Internal migration rates in the US have declined over time. These declines are pervasive across demographic groups and household types and appear to be driven by two forces: an aging population and an increased attachment to one’s place of origin.Mangum, Kyle, and Patrick Coate. “Fast locations and slowing labor mobility.” (2019). https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3843805. Declining mobility suggest that households may be less likely to move to jurisdictions with their preferred level of public goods, reducing the efficiency gains from decentralization.

However, one reason households may be less likely to move is that they already live in their preferred location. Indeed, between 1980 and 2017 individuals have become increasingly segregated by education.Diamond, Rebecca, and Cecile Gaubert. “Spatial Sorting and Inequality.” Annual Review of Economics 14, no. 1 (2022): 795–819. https://doi.org/10.1146/annurev-economics-051420-110839. Greater segregation by education suggests higher efficiency gains from decentralization, indicating that the recently increased federal share is unlikely to improve the efficiency of public goods provision.

Factor mobility generally present a challenge for subnational governments. If localities cannot impose a pure benefit tax, perhaps because of legal or technological constraints, then they will struggle to finance public goods without causing distortions in capital and labor markets. This “flight to the bottom” problem has drawn particular attention in the context of corporate taxation, where local governments compete to attract businesses through subsidies or low tax rates.Slattery, Cailin. “Bidding for Firms: Subsidy Competition in the U.S.” (April 23, 2024). Available at SSRN: https://ssrn.com/abstract=3250356.

A second challenge is that factor mobility enlarges the scope for spillovers across jurisdictions. If households are highly mobile, then children educated in one jurisdiction will often work elsewhere, providing benefits to other jurisdictions through their human capital.Of course, this is not a concern if parents fully internalize the benefits of their children’s education and are willing to finance it through local taxes. Similarly, start-ups that incubate in San Francisco may move to Austin or New York, reducing the returns to local public goods that support innovation.

While changes in capital mobility could motivate a larger federal share of spending to reduce tax competition and internalize, I am not aware of any evidence that capital mobility inside the US has changed significantly in the past two decades. On the whole, then, increased sorting since 2000 suggests the federal share of spending ought to be smaller, not larger.

Baumol’s Cost Disease

Most state and local spending is on services like education and safety. These public goods are subject to Baumol’s cost disease: as the rest of the economy becomes more productive, the cost of hiring teachers and police rises relative to providing physical goods.

Federal spending, on the other hand, is more concentrated in capital-intensive goods like defense and infrastructure, where we expect productivity gains over time. Even in the administration of federal programs, the introduction of information technology should have reduced the cost of paperwork, record-keeping, and other cognitive routine tasks, and economies of scale in the provision of public goods should have reduced federal spending relative to state and local.A good example of how IT has driven greater productivity is the Social Security Administration, which has driven down administration costs to less than 0.5% of benefits. These gains stand in contrast to the fixed labor cost of educating primary and secondary students, where classroom sizes have changed little over time.

The conclusion, again, is tentatively that the federal share of spending likely should be smaller, not larger. The exception is if new economies of scale in the provision of public goods have emerged since 2002, leading to a larger optimal federal share.

Crises and Federal Capacity

Federal governments tend to take on additional responsibilities during national crises. This pattern has played out repeatedly in US history, most obviously during the Great Depression and World War II, but also in recent years during the Great Recession and COVID-19 pandemic. The step-up in the federal share during the latter two is clearly visible on the figure. However, while such crises may justify a level change in the federal share that sticks long after its original motivation, it is not clear why a crisis would cause a steady increase.

The Future of Federalism

The steady increase in federal expenditure share since 2002 appears to be a slow-motion policy mistake, unsupported by changes in underlying economic fundamentals. In an era of increased flexibility due to remote work and substantial scope for policy experimentation with AI, offering more local control over public goods provision could yield large efficiency gains by better matching public goods to local preferences.

Nevertheless, federalism has few friends in national politics.Nagourney, Adam. “Trump’s Power Grab Defies G.O.P. Orthodoxy on Local Control.” The New York Times. https://www.nytimes.com/2025/02/07/us/politics/donald-trump-california-newyork-executive-orders.html. While budgetary pressures from rising entitlement spending and interest payments are likely to divert some federal resources away from current expenditures, they are also likely to depress intergovernmental transfersSanger-Katz, Margot and Sarah Kliff. “G.O.P. Targets a Medicaid Loophole Used by 49 States to Grab Federal Money.” The New York Times. https://www.nytimes.com/2025/05/06/upshot/medicaid-hospitals-republicans-cuts.html. and increase demands on taxpayers in high-tax cities, reducing the ability of local governments to finance their own public goods. The future of federalism in the US appears bleak.