The double distortion: Income-based traffic fines, labor supply, and the tax-benefit interaction

Introduction¶

How should traffic fines vary with income? Finland has linked speeding penalties to earnings since 1921, producing headlines when executives receive six-figure tickets. In the United States, San Francisco launched the first major income-based traffic fine pilot in 2025, and several other jurisdictions---including New York, which tested day-fines in Staten Island as early as 1988 Hillsman (1990)---have experimented with similar systems. The motivation is straightforward: a $200 fine that devastates a minimum-wage worker barely registers for a high earner. Income-based fines appear to solve this problem by scaling penalties with ability to pay Harris (2016)Polinsky & Shavell (1991).

This paper identifies a countervailing force that complicates this reasoning. When fines depend on earned income, they function as an implicit marginal tax---but one that applies selectively to individuals who speed. This selective taxation distorts labor supply in a way that flat fines do not, compounding pre-existing deadweight losses from the income tax. We develop a model that captures this trade-off, calibrate it to US data using PolicyEngine microsimulation PolicyEngine (2024), and propagate parameter uncertainty through forward Monte Carlo simulation.

The double distortion¶

The central mechanism is what we call the double distortion. The first distortion is intentional: fines deter speeding by raising its cost. The second is an unintended byproduct: by conditioning penalties on income, the fine system creates an implicit tax on earning, reducing labor supply among those who speed.

To see why this matters, consider a US worker whose marginal tax rate (MTR)---combining federal income tax, state income tax, Federal Insurance Contributions Act (FICA) payroll taxes, and benefit phase-outs---is already 35%. Under a flat fine, her labor supply decision is unaffected by her speeding---the fine is a lump sum. Under an income-based fine with rate $\phi$, her effective marginal tax rate becomes $\text{MTR}_i + \phi s$, where $s$ is her speeding intensity. Standard public finance results imply that deadweight loss rises with the square of the tax rate Harberger (1964), so adding even a small increment to an already-high marginal rate generates disproportionate efficiency costs.

Crucially, US marginal tax rates are heterogeneous across the income distribution, ranging from near-zero to over 50%. Workers on the Earned Income Tax Credit (EITC) phase-out face effective marginal rates near 40% Maag et al. (2012), while some middle-income workers face rates below 25%. This heterogeneity---captured by the per-person marginal tax rates we compute from PolicyEngine’s Enhanced Current Population Survey (CPS) microsimulation---means that income-based fines interact very differently with the tax-benefit system at different income levels.

This interaction between fines and labor taxation connects to the tax-interaction effect identified in the environmental tax literature. Bovenberg & Mooij (1994) showed that Pigouvian taxes on pollution should be set below marginal external damage when they interact with distortionary labor taxes. Jacobs & Mooij (2015) qualified this result, showing the interaction vanishes at the second-best optimum when the income tax is set optimally. Our setting differs in a key respect: income-based fines apply only to offenders, creating heterogeneous effective tax rates that cannot be replicated by adjusting the income tax schedule.

Contributions¶

This paper makes three contributions to the intersection of public economics and law and economics.

First, we develop a tractable model of joint labor supply and speeding decisions under alternative fine structures. Agents maximize utility over consumption, speeding, and leisure, subject to a budget constraint that embeds the fine structure. Speeding generates private benefits (time savings, utility from speed) but raises mortality risk following the Nilsson power model Nilsson (2004), where fatality risk scales as $(1+s)^n$ with $n \approx 4$. Under flat fines, labor and speeding decisions separate; under income-based fines, they are coupled through the effective tax rate. We solve for mean-field equilibrium where fine revenue is redistributed as a universal transfer.

Second, we calibrate the model to the United States using real microdata from PolicyEngine’s Enhanced CPS PolicyEngine (2024)US Census Bureau (2024). Each simulated agent is drawn from the actual US income distribution and assigned their empirically estimated marginal tax rate---capturing federal and state income taxes, FICA payroll taxes, the EITC, and benefit phase-outs. This approach replaces the stylized single-tax-rate assumption of earlier work with dramatically heterogeneous tax treatment across the income distribution. We specify informative priors for behavioral parameters---drawing on meta-analyses for labor supply elasticities Chetty (2012), traffic safety parameters Nilsson (2004)Elvik (2019), and US regulatory estimates for the value of statistical life (VSL) US Environmental Protection Agency (2024)---and propagate uncertainty through Monte Carlo simulation.

Third, we decompose the welfare difference between fine systems into three components: a deterrence gain from more uniform penalties across the income distribution, a labor distortion loss from the implicit tax on earnings, and a revenue effect from differences in equilibrium fine and tax revenue. The decomposition clarifies when each force dominates and how the welfare ranking depends on key elasticities.

Preview of results¶

Our baseline calibration using US CPS data finds that income-based fines generate higher utilitarian welfare than flat fines in 95% of Monte Carlo draws. While the double distortion mechanism is present---income-based fines raise effective marginal tax rates for regular speeders by 0.2--2 percentage points---its magnitude is small relative to the deterrence equity gain. The mean welfare difference favors income-based fines ($\Delta W = 0.83$, 95% CI: $[-0.02, 3.22]$), and income-based fines reduce consumption inequality (Gini 0.325 vs.\ 0.343).

The result is robust to the labor supply elasticity: our quadratic disutility specification implies a Frisch elasticity of 1.0, four times the meta-analytic central estimate of 0.25, yet income-based fines still dominate. At the empirically estimated elasticity, the welfare advantage would be even larger. The welfare ranking is reinforced under inequality-averse social welfare functions, with income-based fines dominating even more strongly under Rawlsian criteria.

Roadmap¶

Section 2 reviews the literatures on optimal taxation, crime deterrence, speed-safety relationships, Pigouvian taxation with labor interactions, day-fine systems, and equity in monetary sanctions. Section 3 presents the model: utility specification, fine structures, first-order conditions, and mean-field equilibrium. Section 4 describes calibration using US CPS microdata and PolicyEngine marginal tax rates. Section 5 reports results, including welfare comparisons, distributional analysis, decomposition, and convergence diagnostics. Section 6 discusses sensitivity, institutional considerations, and policy implications. Section 7 concludes.