Henry Farrell, for example, compared paying taxes to shopping at Wal-Mart. Far be it from me to criticize anyone who sends me free books, but this does not really work.
Umm, no. I sent her Tom Slee’s book, which uses the analogy of shopping at Walmart to demonstrate that vulgar revealed preference arguments do a very bad job of capturing situations of interdependent choice. This is something that is quite clearly laid out in the extended Alex Tabarrok description of Slee’s argument which I quoted in my original post. What’s at stake here isn’t shopping; it’s interdependence. When choices are genuinely interdependent, behaviour doesn’t necessarily tell us anything about the ‘true’ preferences of the actors in question. What it does tell us about, (if we think that actors are behaving rationally) is what actors think the best reply to other actors’ strategies is in a given strategic situation. I’d like it very much if Megan – and others who use similarly poorly-thought-through arguments – would read about and absorb this basic lesson of game theory. It complicates the analysis of social situations in some very useful and fruitful ways.
Leave aside my questions about whether people really prefer downtowns to Wal-Marts, which is hard to agree upon empirically—I say I care deeply about poverty in Africa, but if that’s true, how come I bought a new iPod instead of sending the money to Chad? Collective action problems generally apply to situations where the outcome is binary: either you have a Wal-Mart nearby, or you don’t. Tax revenue is not binary—it’s an upward sloping line.
This claim is both bizarre and wrong. Olson’s original analysis of collective goods starts from the assumption that collective goods are continuous, and that each marginal contribution to the good results in a marginal increase in supply. Either Megan hasn’t read Olson (who is the key figure in this literature), or she hasn’t understood his claims and assumptions. Later analyses (e.g. Russell Hardin’s 1982 book, Collective Action) bring in the idea of lumpy ‘step’ goods as a supplement to Olson’s arguments. Neither Hardin nor any other collective action theorist that I am aware of posits that collective goods need to be binary or even lumpy, for the sound reason that we have no practical or theoretical warrants to believe that they need be binary (or lumpy).
Some of the things the government spends the money on are binary—but given the existing level of tax revenues, this is simply not a reasonable objection to sending the government additional money. People who say they want higher taxes on themselves generally think the government does not have enough money to do the things it is already doing; as long as you think the government has a better (in some moral sense) use for the money than you do, then you have a moral obligation to send it in.
Given the basic wrongness of Megan’s argument above, this is all quite irrelevant.
(As an aside, I am afraid that Henry made a common mistake in referring to me as an economist. I am but a lowly MBA, and have never claimed otherwise, but for some reason a lot of my readers are confused.)
This wasn’t a mistake on my part, it was snarkiness that in retrospect I shouldn’t have engaged in. The “noted economist” description linked to one of Megan’s Atlantic colleagues who had indeed erroneously described her as an economist. I find myself very frequently annoyed by the apparent mismatch between the confidence of Megan’s judgments as to what economics does and doesn’t tell us, and her grasp of the debates among economists and other social scientists. Still, I should certainly acknowledge that even if Megan often seems to imply a deeper command of the literature than she in fact possesses, she has never to my knowledge described herself as an economist.
On the broader question of why people do, or don’t pay their taxes, the most directly relevant research that I am aware of is by John T. Scholz and Mark Lubell (this AJPS article is most germane; but they have other work dealing with these issues too). Scholz and Lubell provide good reason to believe that taxpaying is indeed a large scale collective action problem; although they don’t use this term, they clearly think of it as being in part an assurance game (that is one in which people’s willingness to contribute is conditioned on their belief that others will contribute too). They find that individuals’ willingness to pay taxes that they could probably fudge seems to covary with the degree to which they trust both government to do what is right and behave honestly and other citizens not to underpay their taxes. Not only that, but Scholz and Lubell find that these factors seem more important to people’s decisions than the self-interested fear of getting caught cheating. This suggests, contrary to Megan, both that tax paying is in part a collective action problem and that people’s cooperation in this game is partially conditioned on their trust that other people are paying their taxes too.