A key term in Margaret Levi’s Of Rule and Revenue (1988) is “quasi-voluntary compliance.” As far as I can tell, Levi uses this term to refer to a situation in which citizens comply with the state’s demands (e.g. to pay taxes) out of a combination of strategic and normative considerations. The strategic considerations involve the calculation of the probability of being caught and the punishment that would be exacted; the normative considerations include the sense that the bargain between the citizen and state is “fair” — that the citizen believes that the state is providing sufficient public goods in return for his tax payments, and that the burden for the state’s activities falls on the citizenry in an equitable way. Thus Levi wants to emphasize (e.g. on page 53-54) that citizens will base their compliance decisions on their understanding of the government’s enforcement procedures not just because those enforcement procedures affect the citizen’s own probability of being caught, but also because they affect the citizen’s view of whether others will be paying their fair share. (And, citizens care about whether others are paying their fair share not just because this affects the probability that they will be caught but because they place a normative value on fairness and will refuse to pay taxes if they perceive the system to be unfair.)
If this is indeed what Levi means by “quasi-voluntary compliance,” then her explication the concept is disappointingly convoluted. At points Levi presents the idea of quasi-voluntary compliance (QVC) as if it were a fundamentally different conception of compliance from what would come out of a Becker-style costs-and-benefits analysis. Unless I’m missing something, it’s not: it merely adds normative concerns to the citizens’ calculations. This does contribute something important to the analysis of rulers’ strategies: if citizens care about fairness or legitimacy, then rulers must take care in choosing tax rates and enforcement systems that they do not violate citizens’ sense of fairness and thus undermine their revenue goals; rulers trying to maximize revenue should also think about the public good they are providing (and citizens’ perceptions of those goods) because, even though these goods are non-excludable and thus would not be part of a purely materially-motivated citizen’s compliance decision, the government’s output matters to fairness-minded taxpayers. But this contribution would have been a lot clearer to me if it were introduced as an extension and modification of a common costs-and-benefits calculus rather than a wholly different concept of compliance.
Given that the definition of quasi-voluntary compliance (QVC) has to do with the utility function of citizens, in practice it should be hard to tell the difference between QVC and mere compliance (the situation in which people pay their taxes because of coercion pg. 64). The main point Levi makes is that citizens will comply less when they perceive that others are not complying and the state is not delivering promised goods (e.g. pg. 68). This could be because of fairness, but it could also be mere compliance: if others are not complying, I may be able to get away with not complying too because others’ non-compliance tells me that the enforcement system is not working well (and that my own non-compliance may be overlooked because so many others are also not complying). (Also, if the state is not providing public goods, maybe it’s because no one else is paying.) The question is how one would know how much (if at all) these normative considerations really matter. At some points QVC seems to be in effect when coercion is not observed, but of course the most successful coercion is the kind that is implicit. In the introductory chapters Levi does offer one phenomenon that would indeed be predicted by “quasi-voluntary compliance” but not in a non-normative account: “rulers invest in deterrence that constituents perceive as being directed toward others” (pg. 67). The challenge would be to distinguish a) deterrence that is carried out in order to boost tax revenues from those being targeted from b) deterrence that is carried out in order to boost tax revenues from others (and not because those others think “I could be next!”).
On to cases. How does QVC enter into Levi’s analysis of states’ revenue strategies? In discussing the Roman Republic, Levi refers to QVC in her explanation for the use of “tax farming” in remote provinces; her argument seems to be that without the social rewards of high income and citizenship, and without detailed information on citizens, collecting taxes in the usual way would have been too expensive. “Without quasi-voluntary compliance, the agency costs of the census rose significantly” (pg. 80). In early parliaments in France and England, QVC helps us understand the role of parliaments in providing a venue in which the ruler could justify his taxes (“parliamentary consent shrouds a ruler’s policies in legitimacy”, pg. 118). In the introduction of the income tax in 18th-century Britain, Levi sees QVC at work in the attention that was paid to the justifications for the tax and the equitability of its administration. In 20th-century Australia, Levi sees QVC as helping to explain individual taxpayers’ objections to their assessments, as well as tax revolts and labor union campaigns against tax avoidance. I find these claims to be more convincing as we move ahead in time, in part because in the latter cases Levi brings to bear more evidence of debates about justification and enforcement (which would be hard to explain in another way), whereas in e.g. the Roman case one could easily explain the use of tax farming in the provinces without recourse to QVC.
I think the best evidence that citizens’ sense of fairness is an important constraint on taxation is that public debates surrounding taxation focus on both government waste and evasion/progressivity/fairness; if citizens were merely being held up by a predatory state and paying just because they don’t want to be arrested, then they should not care about what the government does with their tax revenues or whether anyone else is similarly being held up. On the other hand, for both kinds of debates one can argue that pure economic explanations are sufficient. When citizens complain about government waste or otherwise complain that they are paying too much for the government services they receive, it could simply be that the citizens legitimately want lower taxes for all of society (see e.g. Meltzer-Richard) or simply for themselves (the standard free-rider prediction). Likewise, when there are debates about how the tax burden is to be shared, it need not be that citizens actually care about fairness: it could be instead that simply want to pass more of the burden on to others. I am intrigued by the idea of building on Levi’s QVC idea to investigate patterns in parliamentary debate over time, but this indeterminacy gives me pause: what speech would persuade a skeptic that the speaker is concerned about fairness and not simply about reducing his own tax bill?
In conclusion, I come away from this book intrigued by the contention that fairness helps to explain the history of state revenue extraction but disappointed in the somewhat muddled way in which the ideas are presented and the lack of evidence for the importance of fairness. I think (and I suspect Levi might agree) that this is a book that would have benefited from an explicit application of game theory, or at least a bit more rigor in specifying e.g. the citizens’ outside option, what non-compliance would mean, etc. The idea that normative considerations matter to citizens in determining whether they will pay their taxes is intriguing and believable, but I don’t see a way to convincingly demonstrate to a skeptic that considerations of fairness are fundamental and not epiphenomenal.