Wei Cui’s new paper, Taxation Without Information: The Institutional Foundations of Modern Tax Collection, challenges the now-conventional wisdom that effective tax collection depends upon third party reporting. Cui suggests that effective tax collection in fact depends upon the existence of business firms for whom compliance with the law—tax as well as non-tax—is the norm. Cui argues that this insight should lead us to rethink our assumptions not only about modern tax collection, but also about modern business firms: “we should stop thinking of business firms as ‘fiscal intermediaries,’” Cui writes, and instead “conceive of firms as sites of social cooperation under the rule of law” (p. 3).
Cui’s paper is ambitious, important, and—I think—largely right. He has persuaded me that third party reporting is not nearly as integral to tax collection as I previously believed. If there is a weak point in his argument, it is this: the evidence he produces in support of his “social cooperation” theory is equally consistent with the claim that business firms facilitate legal compliance precisely because they fail to engender close cooperation among their members.
Cui’s carefully constructed argument begins by taking aim at a recent and influential article by economists Henrik Kleven (London School of Economics), Claus Thustrup Kreiner (University of Copenhagen), and Emmanuel Saez (UC Berkeley). Kleven and coauthors state that “to a first approximation, tax enforcement is successful if and only if third-party reporting covers a large fraction of taxable income.” Cui argues that this approximation is nowhere near accurate. More than a third of total government revenue in the United States (federal plus state plus local) comes from sales, property, and corporate income taxes that do not rely on third party reporting. And the value added taxes (VATs) that constitute a large source of revenue in most countries other than the United States rely primarily upon first party reporting. This last point surprised me: I had assumed that credit invoice method VATs required firms to report information about individual transactions with counterparties, against whose returns those reports could be checked. Cui notes that in fact, firms generally do not submit information to the government regarding specific transactions “but instead aggregate transactions into lines on simple tax returns” (p. 5).
While third party reporting plays a more central role in the taxation of labor and investment income in the United States, Cui contends that it need not be this way. We could instead impose excise taxes on employers paying wages, banks paying interest, and corporations paying dividends, much like the final withholding systems in place in many European countries today. As Cui notes, the employer portion of the U.S. federal payroll tax already works like this—and works well. There is no third party reporting involved: the only party reporting information to the government (the employer) is also the one obligated to pay the tax.
Not only do we have lots of examples of taxes that succeed without third party reporting, we don’t have a terribly good account for why third party reporting succeeds where it does. A standard explanation for the success of third party reporting in the labor income context is that firms and employees have “opposing incentives”: firms don’t want to underreport because they desire a deduction for wages paid. But as Cui notes, there are many instances in which the effective tax rate on an employee’s income far exceeds the effective tax rate on the employer. (E.g.: My tax rate is a lot higher than my employer’s, because the University of Chicago is tax exempt. And yet the University of Chicago files W-2s for its professors nonetheless.) One might add to this that the “opposing incentives” story presupposes the enforcement of the corporate income tax: if firms can evade that, they won’t care about the deduction for wages paid. In other words, the “opposing incentives” explanation for the effectiveness of third party reporting relies on the effectiveness of a tax involving almost no third party reporting.
Another theory as to why employers generally comply with reporting requirements is that firms face a high risk of audit, but Cui dismisses this explanation as well. He notes that “the audit rate for parties required to perform information reporting is not known to be higher than in other areas of tax administration” (p. 9, n.31). IRS statistics bear this out. In fiscal year 2014, the audit rate for individuals was 0.9%; the audit rate for C corporations with assets under $10 million was barely higher than that (1.0%); and the rate for partnerships and S corporations was even lower (0.4%). Corporations with assets over $10 million faced a higher audit rate 12.2%), but the fact that a corporation is audited does not mean that every single transaction will be examined. We do not know the per-transaction probability of IRS scrutiny, but we can guess that it is vanishingly small. The threat of detection seems like an incomplete explanation for widespread compliance.
Cui proposes an alternative account for why firms comply with their third party reporting obligations: most large firms in modern economies generally comply with the law across the board. There are, of course, exceptions to this general rule (Cui cites Volkswagen’s cheating on emissions tests), but the exceptions are just that. And, Cui continues, “[i]f a firm is generally compliant with the law, then compliance with the tax law should simply be expected” (p. 16). This is so even with respect to taxes such as the VAT, the sales tax, the corporate income tax, and the employer portion of the payroll tax that do not depend upon third party reporting.
The obvious next question is: Why do large firms in modern economies generally comply with the law? Cui offers two explanations. First, he suggests that firms are formed to earn economic rents, not to profit from illegal activities. “Therefore,” Cui writes, “if a firm is formed not in order to profit from illegal activities, but with the purpose of earning rents from other identified opportunities, then it would not be surprising if the firm does not maximally exploit opportunities arising from illegal behavior—that is simply not its purpose” (p. 17). Second, and relatedly, Cui posits that “participants in [modern firms] are more interested in the orderly division of profits than the disorder implied by cheating (on taxes and other regulatory matters)” (p. 18). Neither explanation strikes me as fully satisfying. Both explanations leave us to ask: Regardless of why firms are formed, and regardless of why individuals participate in them, might not they still want to evade taxes so that there are even more profits to divide?
A counter-narrative to Cui’s “social cooperation” theory might go as follows: Any criminal enterprise involving two or more people depends upon trust. The employer will collude with the employee to evade taxes only if the employer trusts the employee not to rat her out. Trust is required not only for the maintenance of collusion but for its initiation as well: the employer will propose collusion to the employee only if the employer trusts that the employee won’t go straight to the police. Situations in which we see high rates of tax noncompliance (e.g., evasion of the “nanny tax,” underreporting of gifts between family members, evasion of tax by small cash-based businesses) are situations in which the participants in collusion share strong personal and sometimes familial ties—i.e., situations in which there is trust. Collusion to evade the law (tax and non-tax) is less likely within the large modern firm because the large modern firm fails to facilitate trust among its various participants.
Put differently, tax evasion in the modern business firm requires cooperation. This is true even when the tax involves no third party reporting. The CEO of a large firm cannot evade the corporate income tax on her own; she will need the cooperation of the CFO and the accountants to make it work. Her problem is that she doesn’t trust the CFO and the accountants to play along. On this view, modern business firms comply with the law not because they are “sites of social cooperation,” as Cui suggests, but precisely because they are not.
How might one adjudicate between Cui’s social cooperation theory and this non-cooperative theory of corporate legal compliance? I admit that I haven’t figured out a way. Each might apply in some settings but not in others. What I can say with confidence is that Cui’s paper will cause readers to rethink their views about the importance of third party reporting. And, impressively, Cui manages to accomplish that goal in just over 23 pages—a model of concision to which we can all aspire.