The fundamental relationship between tax rates and government revenue has long been a central focus of economic debate and fiscal policy design. Not all tax hikes bring in more revenue. Setting the rate too high might not yield as much as a lower tax rate would, because tax rates influence how much people work and how much businesses invest—both critical economic activities that directly affect the income tax base.
Because of this delicate dynamic, economists and policymakers have long sought to determine the “Goldilocks rate”—the precise threshold where government revenue is maximized without triggering counterproductive economic contraction. A comprehensive new research paper sheds significant fresh light on this enduring fiscal question, finding that the income tax rate that maximizes revenue may actually be lower than previously estimated once modern behavioral responses and the complexities of the tax code are fully accounted for.
Is the Peak of the Laffer Curve Actually 73 Percent?
To understand the trajectory of modern tax theory, one must examine the foundational literature that shaped policy discussions over the past decade. An exceptionally influential policy paper in the “optimal tax theory” literature was published by prominent economists Peter Diamond and Emmanuel Saez in 2011, titled “The Case for a Progressive Tax.” That paper sought to identify the theoretical “peak” of the Laffer curve—the marginal tax rate at which government tax revenue is maximized.
Building directly upon earlier theoretical work developed by Saez, the 2011 paper estimated a top revenue-maximizing tax rate of 73 percent in the United States. However, the complex findings of that study were frequently misunderstood by policymakers, political commentators, and the media. Some analysts incorrectly concluded that the paper’s findings meant the top statutory income tax rate should be set at that elevated level, overlooking crucial methodological nuances embedded within the research.
To start, the top revenue-maximizing rate estimated in the paper was inclusive of all taxes levied at the federal, state, and local levels, a broad aggregate that explicitly included all payroll taxes and consumption-related sales taxes. Crucially, the paper did not estimate that government revenue would be maximized if the top individual income tax rate—which stood at 35 percent at the time of publication—was drastically increased to 73 percent.
A second, and critically overlooked, contributor to the widely cited 73 percent figure was the assumption of a tax base that was substantially broader than the actual U.S. tax system was at the time, and remains today. The authors of the 2011 study estimated that using the existing, narrower tax base, which is riddled with numerous deductions, exclusions, and tax credits, would yield a top revenue-maximizing rate of 54 percent. Furthermore, when isolating just the top federal income tax rate under those narrower conditions, the figure dropped to 48 percent. For historical and geographic context, that rate is not much higher than the combined federal and state income tax rates routinely faced by high-income taxpayers residing in high-tax states such as New York and California.
New Evidence Points to Flatter Laffer Curves
The influential Diamond and Saez 2011 paper was certainly not the final word on this complex economic topic. In the years following its publication, a multitude of academic papers have taken alternative modeling approaches, producing a wide range of total government revenue-maximizing tax rates from all sources, typically falling between 46 percent and 73 percent.
In a new research paper building directly upon this established body of work, economists Rachel Moore, Brandon Pecoraro, and David Splinter argue that many preceding studies utilized tax bases that were either unrealistically broad or overly narrow, while frequently missing critical behavioral adjustments made by taxpayers. Given the structural realities of our current tax system, they find that the true Laffer curves are significantly flatter than Saez and other authors previously estimated. This flatness implies that the actual revenue gains from increasing top statutory tax rates are substantially smaller than what those earlier, simpler models suggested.
Policymakers and fiscal analysts are naturally interested in how a change in the top ordinary income tax rate would directly impact government revenue, but a remarkably large share of income earned at the top of the distribution actually falls outside this standard ordinary income tax base. For instance, while the ordinary income tax base does capture some forms of capital income—such as short-term capital gains and the profits generated by pass-through businesses—other major types of capital income, including long-term capital gains and qualified dividends, are subject to distinctly lower tax rates. Moreover, the ordinary income tax base is eroded even further by targeted deductions and credits designed to incentivize various kinds of private investment.
The authors of the new study significantly improve upon previous academic methodologies by constructing a variety of distinct tax bases and utilizing highly detailed administrative tax data covering high-income earners to estimate specific behavioral responses along the Laffer curve. Their advanced economic model allows them to capture multiple, distinct margins of adjustment undertaken by individuals and corporations in direct response to shifting tax rates.
For example, corporations and business owners may strategically switch their organizational structures to noncorporate entities if tax rates on pass-through business income fall relative to corporate rates. Similarly, individual income taxpayers frequently increase tax-deductible activities, such as charitable giving or participation in tax-preferred investment vehicles, when their marginal tax rates rise.
When rigorously accounting for these additional margins of economic behavior, the authors find that the revenue losses resulting from federal rate increases roughly double compared to previous analytical estimates that relied on a single, uniform tax base. Using a single broad tax base yields a revenue-maximizing federal income tax rate of 55 percent. However, that revenue-maximizing rate falls sharply to 48 percent when utilizing multiple bases and fully accounting for all possible behavioral responses—what the researchers define as the “true” tax base. At that optimal 48 percent rate, the model yields an additional $43 billion in revenue, representing approximately 0.2 percent of gross domestic product (GDP).
The picture shifts even further when broadening the scope beyond federal collections. Taking into consideration all taxes collected at the federal, state, and local levels, the top federal income tax rate that maximizes total government revenue drops even lower. The top federal individual income tax rate that would maximize total government revenue from all combined sources is calculated to be 39 percent.
An increase to that 39 percent rate would yield a modest $1 billion in additional total tax revenue. At that specific threshold, however, the consequential losses incurred by state and local government revenues almost fully offset the marginal federal revenue gains. Notably, a rate of 39 percent is also very close to the current top federal individual income tax rate of 37 percent.
Overall, this new empirical research demonstrates that raising top income tax rates on the existing, complex tax base may not generate nearly as much revenue as previously estimated by policymakers, largely due to the pervasive opportunities taxpayers have to shift income across different tax bases within the broader tax code. Policymakers would be wise to keep these dynamics in mind when evaluating proposals to increase tax rates on high-income earners without enacting foundational structural reforms to the tax code itself.
