The architecture of modern fiscal systems rests on an uncomfortable tension. Central governments possess superior stabilization capacity and redistributive tools, while subnational governments hold informational advantages about local preferences and service delivery. Reconciling these comparative advantages requires sophisticated interbudgetary machinery that most federations implement poorly.

Consider the empirical regularity: across OECD countries, subnational governments execute roughly 40 percent of public expenditure but collect only 20 percent of tax revenue. This vertical fiscal imbalance is neither accidental nor pathological—it reflects the Mirrleesian logic that broad-based taxation belongs at higher jurisdictional levels where mobility-induced distortions are minimized, while expenditure follows the subsidiarity principle toward populations whose preferences it serves.

The design problem, then, is not whether to centralize revenue and decentralize spending, but how to engineer the intergovernmental transfers, borrowing frameworks, and coordination mechanisms that bridge this structural gap without eroding subnational accountability or undermining macroeconomic discipline. Get the mechanism design wrong, and you produce the familiar pathologies: soft budget constraints, procyclical subnational fiscal behavior, and tax competition races toward inefficient equilibria. Get it right, and fiscal federalism delivers what no unitary system can—allocative efficiency responsive to heterogeneous preferences, experimentation across jurisdictions, and redistributive capacity calibrated to national social welfare functions.

Optimal revenue assignment follows a tractable hierarchy derived from the interaction of mobility elasticities, administrative economies of scale, and benefit-tax linkages. Taxes on mobile factors—capital income, corporate profits, and increasingly high-skilled labor—belong at the central level, where horizontal tax competition cannot erode the base. Taxes on immobile factors, particularly land and real property, are natural candidates for subnational assignment because their incidence falls on resident owners who also consume local public goods.

The middle tier presents harder choices. Consumption taxes exhibit administrative complementarities that favor centralization—witness the near-universal failure of subnational retail sales taxes to handle cross-border digital commerce—yet revenue-sharing arrangements can preserve subnational fiscal autonomy without fragmenting the tax base. The Canadian HST and the German VAT-sharing mechanism represent divergent but defensible solutions to this design problem.

Personal income taxation admits an underappreciated solution: piggyback surtaxes on a centrally-administered base. Nordic countries demonstrate that subnational rate-setting authority can coexist with unified collection, preserving both accountability at the margin and administrative efficiency. The marginal tax price visible to local voters disciplines expenditure choices in ways that block grants fundamentally cannot.

Administrative capacity imposes binding constraints often ignored in normative analyses. The theoretical case for subnational property taxation dissolves in jurisdictions lacking cadastral systems, valuation expertise, and enforcement infrastructure. Assignment principles must therefore be dynamic, contemplating capacity-building sequences rather than steady-state optima alone.

The accountability dividend of own-source revenue is empirically substantial. Subnational governments financed primarily through transfers exhibit measurably weaker expenditure discipline, lower responsiveness to local preferences, and greater susceptibility to the flypaper effect. Revenue assignment is therefore not merely a technical allocation problem but a governance mechanism.

Takeaway

Who taxes whom shapes who answers to whom. The political accountability of subnational governments erodes precisely to the extent that they spend money they did not raise.

The structural gap between subnational expenditure responsibilities and own-source revenues—the vertical fiscal imbalance—is typically closed through some combination of revenue sharing, conditional grants, and equalization transfers. Each instrument carries distinct incentive properties that determine whether the resulting system promotes or undermines efficient subnational behavior.

Formula-based equalization, pioneered in Australia and Canada, addresses the horizontal dimension of the imbalance by compensating jurisdictions for differences in fiscal capacity and expenditure needs. The mechanism design challenge is ensuring that equalization does not implicitly tax subnational tax effort at confiscatory rates. When a dollar of additional own-source revenue triggers nearly a dollar of equalization clawback, the optimal subnational tax effort collapses toward zero.

Conditional grants serve the correction of interjurisdictional externalities—education, environmental protection, infrastructure with spillover benefits—but their proliferation produces a familiar pathology. Categorical restrictions create expenditure distortions, administrative compliance costs consume substantial shares of transferred resources, and the political economy of grant design systematically favors visible, ribbon-cuttable projects over maintenance and operational excellence.

The soft budget constraint problem looms over all transfer design. When subnational governments anticipate central bailouts for fiscal distress—whether through explicit rescue packages or through the accommodation of arrears—the ex ante incentive to maintain fiscal discipline disintegrates. Credible no-bailout commitments require institutional scaffolding: constitutional debt limits, transparent reporting, market-based borrowing with genuine default risk, and political costs to central governments from subnational rescues.

The empirical evidence suggests that lump-sum, formula-driven transfers with stable rules outperform discretionary or negotiated arrangements across virtually every dimension of performance. Predictability enables multi-year fiscal planning, while rule-based allocation insulates transfers from the political economy distortions that plague discretionary systems.

Takeaway

Every transfer mechanism is simultaneously a subsidy and a tax. The question is not whether it distorts behavior, but whether the distortions it introduces are the ones you intended.

Subnational fiscal behavior can amplify or dampen national macroeconomic fluctuations, and the default pattern absent coordination mechanisms is unambiguously procyclical. Balanced budget requirements, common at the subnational level, force expenditure contraction during recessions when revenues fall—precisely when countercyclical policy would be welfare-enhancing.

The coordination problem is fundamentally one of assigning stabilization responsibility. The Musgravian consensus places stabilization with the central government, whose monetary sovereignty and borrowing capacity dwarf subnational capabilities. Yet when subnational spending constitutes 40 to 50 percent of total government expenditure, central countercyclical policy alone is insufficient; it must be complemented by mechanisms that either allow subnational smoothing or compensate for its absence.

Rainy-day funds represent the most straightforward smoothing mechanism, but their calibration requires sophistication. Optimal reserve levels depend on revenue volatility, the correlation of subnational shocks with national cycles, and access to alternative smoothing instruments. The Chilean fiscal rule, indexed to structural rather than headline balances, offers an instructive template that has been adapted with varying success across Latin America.

Subnational borrowing frameworks present the tightest policy knot. Complete prohibition eliminates borrowing-driven fiscal indiscipline but also eliminates intertemporal smoothing and infrastructure financing. Unrestricted market access produces the debt accumulation pathologies observed in Brazilian states during the 1990s and Spanish autonomous communities during the eurozone crisis. The middle path—market discipline supplemented by numerical rules, debt brakes, and transparent reporting—requires institutional machinery most federations are still building.

Coordination councils and fiscal responsibility frameworks have emerged as the institutional response to these challenges. Their effectiveness depends less on formal authority than on information architecture: shared fiscal projections, harmonized accounting, and credible enforcement mechanisms that transform coordination from exhortation into binding constraint.

Takeaway

A fiscal system decentralized in authority but unified in economic consequences requires coordination mechanisms that are formal enough to bind and flexible enough to function.

The design of interbudgetary relations is not a peripheral technical exercise but a central determinant of how effectively government translates social objectives into outcomes. Revenue assignment shapes accountability, transfer mechanisms shape incentives, and coordination frameworks shape macroeconomic performance.

The Mirrleesian insight that mechanism design matters more than first-best targets applies with particular force here. Optimal fiscal federalism is not the configuration that would emerge from a benevolent planner's calculation but the one that aligns the incentives of self-interested subnational actors with socially desirable behavior under realistic information constraints.

The practical implication for institutional reformers is humbling: there is no universal optimum. The appropriate architecture depends on administrative capacity, political economy, shock structure, and the existing distribution of functions. What travels across contexts are the design principles—accountability through own-source revenue, rule-based rather than discretionary transfers, credible no-bailout commitments, and information systems capable of sustaining coordination.