Catherine Boone’s review is very magnanimous. Self-interest therefore dictates that I agree with almost all of it. Her suggestion that “it does seem like Bates’ model could accommodate an analysis” like mine does, however, give me an occasion to point to three counts on which I do not see such compatibility.

First, my book seeks to demonstrate the pitfalls of working with a parsimonious “model” of (political) action such as Bates’. I push against the notion that political actors’ goals can be adequately captured by the kind of simple preference function that rational choice accounts are typically premised upon. (Protestations that rational choice theory does not commit one to substantive assumptions regarding preferences notwithstanding, research practice almost invariably ends up making precisely such empirically unscrutinized assumptions that I label materialist-utilitarian. I take this bug to be an essential feature.) Likewise, the mode of political decision-making traced in my book often enough does not comport with a model of instrumental means-end calculations. And even where such calculations are central, the particular institutional, discursive, and historical situatedness of various actors creates such a complex mixture of contradictory logics and motivations that it makes little sense to think either in terms of the collective interests of putative entities such as “the political elite”—or of “outcomes” as the straightforward expression of the maximization of such interests. My book thus portrays state-led rural development that “failed” the Tanzanian peasantry as in many ways the unintended consequence of the complex interplay of significantly non-instrumental actions of a diverse set of diversely situated and motivated actors. In short, one thrust of the book’s argument is precisely a salvo against the parsimonious model of (political) action that is at the heart of rational choice analysis.

On a second, related count, I do not see that my book fits the “model” of empirical research practice that Bates’ account sets out. Boone approvingly notes my book’s critique of “the weaknesses of the empirical bases” for our arguments about (just?) African politics. I make the point that these weaknesses are no random accident. Whether wholly successfully or not, my attempt at an ethnography of the Tanzanian state’s development practices aims to provide a realistic, thick description of political processes as they unfold—and to grapple with questions of motivations and intentions by paying close attention to the standpoint of historically situated actors. This is different from the model of empirical research that is exemplified by Bates’ canonical text. It, like the typical rational choice account, “explains” suitably framed descriptions of outcomes by reasoning backward from them to perhaps—from the perspective offered by an armchair—plausible, but empirically unsubstantiated motivations and means-end calculations. Mike McGovern’s critique of the “casualness of the analytical moves from correlation to causation” when “much of the intellectual heavy lifting … is … done at the level of implication or commonsense guessing” among some exponents of “Popular Development Economics” is apropos here: it applies, too, to much rationalist analysis in political science (“Popular Development Economics—An Anthropologist among the Mandarins,” Perspectives on Politics 9 [2, 2011]: 348).

Lastly, to put this negatively, my book fails to deliver when measured against the “model” implicit in Bates’ account of what social-scientific research ought to aim for: general covering laws. Here, my choice not to try to measure up is grounded in an understanding that complex political phenomena such as “rural development policy” simply are not governed by such laws.