Torsten Persson and Guido Tabellini have produced the most ambitious study yet that attempts to identify and estimate the effects of constitutional design on economic outcomes. They draw on data from more than 80 countries and control for demographic, historical, regional, and economic characteristics. The purpose is to determine whether the shape of political institutions has measurable impact on economic policies (e.g., fiscal balance, social welfare spending), on government performance (e.g., corruption indices, protection of property rights), and on direct measures of economic performance (e.g., productivity of capital and labor). The political institutions that draw the most attention are constitutional regime type and the method of electing legislators.

The authors are sensitive to subtle distinctions among constitutions and electoral rules, yet they follow previous political science research in focusing much of their analysis on the basic differences between presidential and parliamentary systems, and between majoritarian and proportional electoral rules. From these distinctions, they draw a number of intuitions on which they base their statistical analysis. One is that majority elections turn more dramatically than do proportional ones on swings in support among small groups of voters—those in pivotal districts, for example. This could produce more acute responsiveness by elected officials in majority-election systems, and so less use of their office for private gain. It could also encourage politicians to produce public policies that yield more targeted, as opposed to universalistic, benefits. Another is that presidentialism might foster accountability by maximizing internal checks and balances that prevent rent-seeking by politicians, whereas parliamentarism may maximize responsiveness to the broadest set of citizen demands by requiring governments to hold together broad, and potentially heterogeneous, support coalitions in order to survive in office.

The empirical scope of the authors' analysis demands a formidable effort at data collection, as well as a host of decisions regarding measurement and coding. Space limitations preclude discussion of these here, except to say that Persson and Tabellini are consistently clear and explicit about their decisions, providing readers with the necessary information on which to evaluate their results. At any rate, issues of inference are even more challenging than those of data in this project, and here the authors provide a real service in Chapter 5, which is a primer on statistical and econometric methods to address fundamental problems that plague cross-national comparisons in general—reverse causation, selection bias, unobserved variables, interactive effects, and potential differences across observations in functional forms of relationships among the variables (what they call “comparing the uncomparable”).

Persson and Tabellini absolve from an obligation to read the chapter any readers willing to take on faith their methodological skills and their judgment, but for readers (like this one) who are anything short of econometrically proficient, the chapter is a worthwhile tutorial. The advantage over most pure methods texts is in already being immersed in the empirical data and the hypotheses regarding relationships among the variables at stake for the authors. With the thick theoretical and empirical context of the book at the front of one's mind, the issues of inference and methods are much easier to grasp, and the methods lessons in this chapter will serve any consumer or producer of cross-national comparisons well beyond this book.

The empirical results reported in the book are substantial, but a few key insights warrant particular mention. One is that with respect to accountability, the effect of presidentialism appears to interact with the quality of democracy, such that when presidentialism is good, it can be very good, although (unlike Mae West) when it is bad, it is not even better—in fact, far from it. That is, among regimes with high Freedom House scores, presidentialism reduces corruption and rent-seeking policies, whereas among lower-quality democracies, presidentialism contributes to policies that undermine property rights and reduce the productivity of labor. This prompts the natural question: Does presidentialism itself affect the quality of democracy? Empirically, Persson and Tabellini's presidential systems are lower-quality democracies than the parliamentary systems in their analysis, but fully untangling the relative import of various other factors (wealth, colonial history, regime age) besides regime type that may contribute to democratic quality is difficult.

Another critical result is the ratchet effect the authors detect with respect to spending policy under one particular institutional format, but not others. Specifically, parliamentary regimes with proportional electoral rules respond most dramatically to negative economic shocks by increasing government spending, but then do not subsequently scale back when times get better. These regimes are good Keynesians in bad times, but not so good ones in good times. Persson and Tabellini speculate that this is because the coalitions that generally govern in parliamentary and proportional systems have trouble agreeing on a distribution of the fiscal pain that budget cuts would imply.

Presidential regimes, by contrast, are lousy Keynesians, period, actually exhibiting mildly pro-cyclical spending patterns. This could be because presidential regimes tend to be found in poorer countries, which face more severe restrictions on financing government debt, and so for whom countercyclical spending is less available as a policy option to begin with. Moreover, whereas parliamentary/proportional regimes tend to have difficulty cutting spending under any circumstances, presidential systems postpone spending cuts until the year after elections, suggesting a brand of manipulation of the political business cycle. Governments in proportional electoral systems, parliamentary and presidential alike, by contrast, raise spending in election years.

Readers interested in institutional design should take these reported results as the tip of the iceberg and read the whole book. The Economic Effects of Constitutions advances our understanding of political economy and is certain to become a standard reference in the comparative study of institutions.