I probably most appreciate James Buchanan for his work on constitutional economics. I think there's a lot of value in public choice theory too, but I can't help feeling that a lot of people in twentieth century economics have been concerned about these issues (hell, a lot of people in every century of economics have been concerned about these issues), and public choice to a large extent just put a new gloss on old economic ideas. That's fine and revisiting ideas and sprucing them up can make a valuable contribution. But what excited me about Buchanan's constitutional economics was that it seemed to me to offer a fundamentally new (economic) way of thinking about an issue I cared a great deal about.
I first got acquainted with Buchanan learning about public choice as an undergraduate at William and Mary, but I did not see his work on constitutional economics until my public policy classes at George Washington University, with Prof. Rycroft. I'm frustrated that I can't put my hands on exactly what we read, but they were articles - not Calculus of Consent - and if I recall they were in political science journals, not economics journals.
The basic idea of Buchanan's constitutional economics was that public decision really comes in two stages, not one: the constitutional stage, and the political stage. People generate constitutions that create an institutional environment constrained in ways that they perceive to be beneficial. This has implications for how we think about the subsequent political stage. It rejects that Lysander Spooner take on things that says that unanimous consent is required for just policy decisions, because people will consent to a constitutional arrangement where legislation passes with less than unanimous consent because they think that the whole package of policy that such an institutional environment produces is preferable to policy produced in a unanimous consent environment. You can think about it as a sort of nested optimization.
Not only is unanimous consent not necessary at the political stage, but majoritarianism isn't necessary. You could have a supramajority or you could have minority vote requirement. Or you could have more complicated bicameral or representational schemes. Anyone who knows anything about the literature on social welfare functions knows that there is no satisfying single answer to these voting scheme type questions. My take on what Buchanan said (and I'm not sure if Buchanan would put it this way) is that we can talk about selection of institutional frameworks, not for their optimization of some elusive social welfare function but because the sub-optimal policy selections made in that institutional framework are expected to be preferable to the sub-optimal policy selections made in competing institutional frameworks. I definitely am not sure if Buchanan would endorse this, but I would characterize this as a blending of Kenneth Arrow and John Rawls.
It's the Arrow element that I'm not sure Buchanan would agree with me on. He would certainly agree with the tie to Rawls. And this brings me to something else I find valuable in Buchanan's work on constitutions. Rawls gives us a nice thought experiment but it's very hard for me to know what to do with his veil of ignorance, because the fact is there is no veil of ignorance as Rawls conceived of it. It's a nice ideal (and perhaps I should just say "we should implement that ideal with as much fidelity as possible" and be done with it), but since I'm interested in these questions from a scientific as much as from a normative angle, I find Buchanan's approach a lot more useful. The conception of the constitution making process is similar to Rawls except there's no presupposition that we have any veil of ignorance except ignorance of what policymaking the future will bring. We don't know what we'll be voting on (although we may have an idea) or how the votes will turn out. But we do know our economic interests and our social position, etc. Given those preferences we generate institutional environments with constitutions. It is the veil of ignorance adjusted for what we actually have to work with in the real world. You can keep Rawls to think about some ideal, and ideals are nice to have in these discussions. But as a question of positive social science Buchanan's alternative approach has major advantages (and of course major similarities to Rawls).
Practically speaking, my appreciation of the discussion of constitutions comes from three sources. First chronologically is my long-standing interest in the American founding, the principles of the founding fathers, and constitutionalism and federalism as the critical contribution of American thinkers to the liberal tradition. Regular readers know the value I place on all this and that's long pre-dated my study of economics. Getting an economic analysis of these ideas has obviously held a strong appeal for me.
The second reason I like Buchanan's constitutional economics is that we can really generalize this to the economics of the generation of institutional frameworks. A lot of these principles apply to non-constitutional institution formation, after all. Buchanan has in mind a process that is perhaps more relevant for more deliberately generated institutional frameworks, and of course we can think of institutional frameworks that are less deliberately generated that are important as well - but if you're a person that thinks the institutional context is relevant to answering economic questions, then you're going to take interest in what Buchanan has to say about constitutions.
The third reason is that at the time I first read Buchanan on constitutions I was really getting into researching my great-grandad's own experience as the president of the Maryland Constitutional Convention of 1967-68, which I've talked about on here before. That effort failed, and I have a continuing interest in what my great-grandad's goals were (his vision really dominated the choices of the convention - he didn't micromanage it, but he definitely put his stamp on it and was widely supported by the delegates), what he did right, what he did wrong, and why it failed. At some point I would really like to have the opportunity to write something up formally on the convention (ideally for the fiftieth anniversary in a couple years), and I can't imagine I won't be citing Buchanan if I did.