Few policy questions divide economists more sharply than the optimal top marginal tax rate. Diamond and Saez argue for rates approaching 73 percent on top earners. Mankiw and others contend that rates above 50 percent destroy value creation and ultimately harm everyone. Both camps claim rigorous theoretical foundations and empirical support.

This disagreement isn't merely ideological theater. It stems from genuinely unresolved empirical questions and fundamentally different assumptions about what high incomes represent. When economists plug different parameter values into the same optimal taxation formulas, they arrive at policy prescriptions separated by 40 percentage points or more. The Mirrlees framework provides elegant theoretical machinery, but that machinery requires inputs we cannot measure with precision.

Understanding this debate matters beyond academic interest. Major economies currently set their top rates based partly on economic analysis, and the range of defensible rates under different assumptions spans from roughly 35 percent to over 80 percent. Policymakers must navigate this uncertainty while making concrete decisions. This analysis examines the three core disputes driving the controversy: how responsive high earners actually are to taxation, whether their incomes reflect genuine productivity or redistributed rents, and where revenue-maximizing rates actually lie under realistic assumptions.

The elasticity of taxable income—how much reported income changes when tax rates rise—sits at the center of optimal top rate calculations. In the standard formula, the optimal rate equals 1/(1 + a × e), where e represents this elasticity and a captures social preferences. Small differences in elasticity estimates produce enormous differences in recommended rates.

Feldstein's influential 1995 study using tax reform data suggested elasticities around 1.0 or higher for top earners, implying optimal rates well below 50 percent. Subsequent work by Gruber and Saez, using broader panels and refined methodologies, found substantially lower estimates around 0.4 for the general population and perhaps 0.5-0.6 for high earners. More recent studies exploiting quasi-experimental variation have produced estimates ranging from 0.1 to 0.8 depending on income level, time horizon, and country.

Methodological choices drive much of this variation. Short-run responses include income retiming that washes out over longer periods. Tax reforms often coincide with economic changes affecting high earners specifically, creating identification challenges. Different definitions of taxable income—whether to include capital gains, how to handle deductions—yield different measured responses. And responses to anticipated permanent changes may differ from reactions to temporary rate fluctuations.

International comparisons add complexity rather than resolution. Scandinavian countries maintain high top rates with apparently modest behavioral responses, but their tax bases differ substantially from American definitions of taxable income. The United Kingdom's experience with a temporary 50 percent rate showed significant income shifting among the very highest earners, though separating real responses from timing effects remains contested.

The honest assessment is that plausible elasticity ranges for top earners span 0.2 to 0.8, with the true value almost certainly varying across countries, time periods, and specific rate changes. This range alone generates optimal rate recommendations from 45 percent to over 75 percent, before considering any other disputed parameters.

Takeaway

When evaluating tax policy recommendations, always ask what taxable income elasticity the analysis assumes—a difference between 0.3 and 0.6 can swing the optimal rate by 20 percentage points while both values remain empirically defensible.

Standard optimal taxation assumes incomes reflect marginal products—that someone earning ten million dollars contributes ten million in value. But a substantial body of evidence suggests high incomes partially reflect bargaining power, market manipulation, or zero-sum redistribution from other stakeholders. If so, taxing these rents doesn't reduce productive activity and the optimal rate calculation changes dramatically.

Piketty, Saez, and Stantcheva formalized this intuition by decomposing the elasticity of taxable income into real responses (changing actual economic activity) and rent-seeking responses (changing how the pie is divided without affecting its size). When top executives negotiate higher compensation by extracting value from shareholders or workers rather than creating new value, taxing this income has no efficiency cost. The optimal rate rises substantially.

Evidence on rent-seeking is suggestive but contested. Executive compensation grew far faster than stock market returns or company productivity would justify. Financial sector incomes capture substantial portions of the social value their activities create, if they create any at all. Cross-country patterns show that when top tax rates fall, the income share of the top 1 percent rises—but this could reflect either reduced distortion or increased rent extraction.

The methodological challenge is distinguishing rents from genuine returns to scarce talent. A hedge fund manager might earn enormous returns through skill that no one else could replicate, or through market power and informational advantages that constitute zero-sum redistribution. Their compensation alone cannot tell us which interpretation is correct, and we lack reliable methods to decompose observed incomes into productive and redistributive components.

Incorporating even modest rent-seeking assumptions dramatically affects optimal rates. If 30 percent of behavioral responses represent income shifting or rent extraction rather than reduced real activity, an elasticity of 0.5 becomes an effective real elasticity of 0.35, pushing optimal rates from around 55 percent toward 65 percent. Larger rent-seeking adjustments push rates toward the revenue-maximizing level, since efficiency costs of taxation diminish when you're merely taxing redistributed value.

Takeaway

The rent-seeking question forces a philosophical choice: treating all market incomes as earned through value creation mechanically favors lower rates, while acknowledging that bargaining power and market failures shape compensation supports higher rates—and neither assumption is definitively proven.

The Laffer curve—the observation that both zero and 100 percent rates generate zero revenue—establishes that some rate maximizes government revenue. This peak represents an upper bound for sustainable taxation regardless of distributional preferences, since rates beyond it sacrifice both revenue and efficiency. Locating this peak requires combining elasticity estimates with income distribution data.

Under the standard formula, the revenue-maximizing top rate equals 1/(1 + a × e) where a reflects the Pareto parameter of the income distribution (roughly 1.5 for the U.S.) and e is the taxable income elasticity. With an elasticity of 0.5, this yields a revenue-maximizing rate around 57 percent. With an elasticity of 0.25, it rises to 73 percent. With 0.8, it falls to 45 percent.

Current top marginal rates in major economies cluster between 40 and 55 percent when including subnational taxes. The United States sits around 50 percent for high-income residents of states like California or New York. France, Germany, and the United Kingdom fall in the 45-55 percent range. Scandinavian countries exceed 55 percent in combined rates.

Whether these countries are above or below their revenue-maximizing rates depends entirely on which elasticity estimates one accepts. Those emphasizing Feldstein's higher estimates argue most developed economies already approach or exceed their Laffer peaks. Those accepting Saez's lower estimates suggest substantial room for revenue gains from higher rates, particularly in the United States and United Kingdom.

Dynamic considerations complicate this further. Very high rates sustained over decades might increase elasticities by encouraging emigration, career choice changes, or capital flight that shorter-term studies cannot capture. Conversely, high rates might reduce pre-tax inequality by discouraging rent-seeking, meaning that the same rate generates more revenue in steady state than transitional estimates suggest. Neither effect is precisely quantified, adding additional uncertainty to already wide confidence intervals.

Takeaway

Revenue-maximizing rates almost certainly fall somewhere between 45 and 75 percent for most developed economies—wide enough that countries face genuinely uncertain tradeoffs rather than clear violations of economic principle in either direction.

The fierce debate over top marginal rates reflects genuine scientific uncertainty, not merely political preferences dressed in economic language. Reasonable parameter combinations within defensible ranges generate optimal rate recommendations spanning more than 30 percentage points. Any economist claiming precise knowledge of the correct top rate is overstating what the evidence actually supports.

This uncertainty doesn't mean all positions are equally valid. Estimates assuming zero rent-seeking and high elasticities favor lower rates; those incorporating rent-seeking adjustments and accepting lower elasticity estimates favor higher rates. Both frameworks are internally consistent, but they embed different empirical and philosophical assumptions that current research cannot definitively resolve.

For policymakers, the practical implication is that setting top rates within the 45-65 percent range represents defensible application of optimal taxation principles under different plausible assumptions. Rates far below this range sacrifice revenue under most parameter combinations. Rates far above risk approaching genuine Laffer limits. The honest answer to 'what is the optimal top rate' is that we genuinely don't know within a substantial range—and policy should acknowledge rather than suppress this uncertainty.