Michael Doran’s scholarship on the taxation of executive pay is consistently insightful, and his latest article on section 162(m) is no exception. That provision, enacted in 1993, disallows a deduction for amounts paid to CEOs and other top officers of publicly traded corporations in excess of $1 million unless such compensation is “performance-based.” Doran convincingly argues that section 162(m) has done little to stem the rise of CEO compensation and that it has not caused corporations to tie pay to performance in a meaningful way. And while I am less sure than Doran that section 162(m) should be scrapped, Doran certainly makes a strong case for getting rid of the provision.
According to Doran, section 162(m) has proven to be a failure for at least three reasons. First, firms can skirt the $1 million limit by paying CEOs in stock options or by tying compensation to performance goals that are easily met. Second, the provision applies only to current employees—and so does not impose any constraints on nonqualified deferred compensation paid out to former CEOs. And third, corporations can ignore the cap, pay their CEOs more than $1 million outright, and simply forgo the deduction for amounts over $1 million paid. Indeed, more than a quarter of publicly traded corporations in 2013 blew through the cap and gave up the deduction.
The one thing section 162(m) has accomplished is something that its proponents never intended: raise revenue. The U.S. Treasury, by its own estimate, collected an additional $1.1 billion in corporate income tax revenue in 2013 on account of firms exceeding the cap and forfeiting the deduction. Doran acknowledges this unexpected revenue windfall, but says that the revenue boost cannot make up for (what he sees as) the unattractive distributional consequences of section 162(m).
It is on this last point that I am not entirely persuaded. Doran’s distributional analysis begins from the (correct) premise that “[a] company that pays an executive non-deductible compensation increases its own corporate tax liability” (p. 21). He goes on to say (and again I agree) that “the corporate tax burden is borne by some combination of labor and capital” (p. 22). From this he concludes that “the economic burden of disallowed tax deductions falls far more heavily on some combination of workers and investors than on the executives whose compensation was supposed to be in the policy crosshairs” (p. 23). But the conclusion does not necessarily follow from the premises.
While it is true that the corporation, not the CEO, is legally liable for the increased tax when it pays the CEO more than $1 million, legal incidence and economic incidence are not the same thing. In economic terms, denying a deduction to the corporation under section 162(m) is equivalent to imposing an additional tax on the CEO at a 35% rate on compensation over $1 million, while requiring the corporation to withhold and remit the tax. The standard assumption in the incidence literature is that the employee, not the employer, bears the entire burden of a labor income tax (though, as discussed below, that assumption might not hold true in this particular context). Doran, for his part, assumes that the employer (the corporation) bears the entire burden of a tax economically equivalent to a labor income tax, and that the burden is then allocated among the firm’s other workers and shareholders just as a corporate income tax might be.
To illustrate: Say that Goldman Sachs pays its CEO, Lloyd Blankfein, a base salary of $2 million (as it did in 2014). Only the first $1 million is deductible under section 162(m). The cost to Goldman of paying $2 million to Blankfein and deducting the first $1 million only is equal to the cost to Goldman of paying $2.54 million to Blankfein and deducting it all. (Arithmetically, $2.54 million minus a deduction worth 35% x $2.54 million equals $1.65 million; $2 million minus a deduction worth 35% x $1 million also equals $1.65 million.) Doran implicitly assumes that in the absence of section 162(m), Goldman would have paid $2 million to Blankfein and distributed the remaining $540,000 to shareholders and workers. But maybe instead the extra $540,000 would have gone to Blankfein. Or maybe some would have gone to Blankfein and some to shareholders and workers.
To be sure, we might have reasons for thinking that the supply of public company CEOs is more elastic than the demand for their services, in which case the economic incidence of what is effectively a tax on CEO pay would fall primarily on the corporation (and thus on the corporation’s shareholders or workers). Executives might move seamlessly between public and private firms; Blankfein might be able to command a $2 million base salary from a privately held investment bank; and so in order to retain him, Goldman might have to pay Blankfein a $2 million base salary regardless of section 162(m). Or perhaps the supply of public company CEOs is inelastic: there might be a limited number of people who can perform these jobs effectively, and there might be limited movement of executives between publicly traded and privately held firms. In that case, a larger share of the tax would fall on suppliers (i.e., CEOs). The point is: Doran’s objection to section 162(m) on distributional grounds is itself grounded on an assumption about the provision’s economic incidence that is far from obviously true.
This is, admittedly, no resounding defense of section 162(m). It amounts to saying: “Maybe the distributional consequences of section 162(m) aren’t so unattractive; we don’t really know enough to say.” What I can say is that regardless of whether one agrees with Doran’s distributional analysis, his article makes a valuable contribution to the section 162(m) debate. Those who read it will be—forgive the pun—well compensated.