:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/31855869/478648941.0.jpg)
At the core of Washington's economic-policy debate is a premise shared by both Democrats and Republicans: raising taxes on the rich will hurt the economy by discouraging super-talented, super-productive rich people from working as hard.
But in a recent paper "Taxation and the Allocation of Talent," Benjamin Lockwood, Charles Nathanson, and Glen Weyl challenge that assumption. Higher tax rates, they argue, could push talented individuals to eschew lucrative-but-socially-useless jobs in favor of more broadly beneficial careers in teaching and research.
The idea: low taxes lead people to waste their lives on money-making
When choosing a career, people — and especially talented people — have the opportunity to trade off financial and non-financial rewards. A high school English teacher can probably take longer vacations than a partner at a corporate law firms, but the lawyer gets paid more and can take fancier vacations. A research scientist may attract more praise from the community than the designer of high-frequency trading algorithms, but the trader can buy a gorgeous loft in Manhattan. Raising taxes on the rich reduces the relative size of those financial rewards.
The career choices talented people make matter not just for themselves, but for the rest of society. Jobs differ in the extent to which success helps others. Major scientific breakthroughs help a scientist advance her career but are also broadly beneficial to society. A great teacher may impact a smaller circle of people, but is still helping many people beside herself. By contrast, lawyers and traders seem to largely compete with each other in zero-sum games. If high taxes push talented people into careers where their work helps others that could raise the growth rate and increase human welfare completely apart from revenues.
The evidence for
Like many economics papers, the work primarily consists of exploring a variety of mathematical models. The empirical data used is largely second-hand (the authors use a study of Harvard graduates career choices as well as IRS data about the occupations of the top 0.1 percent of the income distribution, and they cite a wide array of preexisting literature about the benefits of different professions) and not necessarily intended to produce precise results.
The authors show that under a variety of plausible assumptions the socially optimal top marginal income tax rate is very high — in the 70 to 90 percent range — largely because high tax rates would deter talent entry into finance and encourage talent entry into research/academia and teaching. The authors also find that this sort of high tax regime is a distinctly second-best policy alternative and that the vast majority of the benefits could be captured with fewer unintended consequences through hypothetical more targeted policy measures aimed at specific occupations.
Reasons for doubt:
The authors methods do not attribute any positive externalities to occupations in the arts and entertainment. To the extent that one believes "starving artists" are making contributions to society not captured by their monetary incomes, the true optimal tax rates will be even higher.
Conversely, the theory that skill in pursuit of financial careers adds little social value, though widely believed, has hardly been proven in an airtight manner. Even if some trading occupations are largely zero-sum, the financial sector arguably contributes broadly to the effective management of private firms (through the reality or threat of takeover) and to the commercialization of innovation (through the venture capital sector). Furthermore, it is at least not obviously correct to believe that the people who make excellent hedge fund managers would also be excellent elementary school teachers were they to choose another occupation.
Ultimately the paper is an extremely provocative theoretical contribution that suggests a potentially fruitful line of empirical inquiry. Given that the case for higher taxes tends to rest on equality, this research offers an interesting additional consideration. But ultimately the result hinges heavily on the estimation of hard-to-estimate parameters, so much so that the authors have created a handy web app into which you can insert your own favored values.
