福山写的《Why Nations Fail》书评

Acemoglu and Robinson on Why Nations Fail
Francis Fukuyama

Daron Acemoglu and James Robinson have just published Why Nations Fail, a big book on development that will attract a lot of attention. The latest fad in development studies has been to conduct controlled randomized experiments on a host of micro-questions, such as whether co-payments for mosquito bed nets improves their uptake. Whether such studies will ever aggregate upwards into an understanding of development is highly questionable. By contrast, Acemoglu and Robinson have resolutely focused on only the largest of macro questions: how contemporary institutions were shaped by colonial ones, why it was that regions of the world that were the richest in the year 1500 were among the world’s poorest today, or how rich elites were ever persuaded to redistribute their wealth. In Why Nations Fail, Acemoglu and Robinson restate and enlarge upon earlier articles like “The Colonial Origin of Institutions” and “Reversal of Fortune,” but in contrast to their academic work, the new book has no regressions or game theory and is written in accessible English for general readers.

Acemoglu and Robinson (henceforth AR; Simon Johnson of the old AJR team dropped out of this volume) have two related insights: that institutions matter for economic growth, and that institutions are what they are because the political actors in any given society have an interest in keeping them that way. These may seem like obvious statements, but many people in the development business haven’t gotten the message. Among development specialists there is what AR term the “ignorance” hypothesis: failure to develop is the result of not knowing either what good policies are (this was the old Washington Consensus) or, now that the focus has shifted to institutions, what good institutions are or how to create them. Many development agencies act as if leaders in developing countries want to do the right thing, if only they knew how, and that development assistance should therefore consist of sending smart people from places like Washington out to teach them, perhaps accompanied by some structural adjustment arm-twisting.

By contrast AR argue that bad institutions are the product of political systems that create private gains for elites in developing countries, even if by doing so they impoverish the broader society. (Think Nigeria, which has many multimillionaires while 70 percent of the population lives below the poverty line.) Doing the “right thing” would take away the rents they receive, which is why no amount of hectoring or threats to withhold the next loan tranche has much effect on their behavior. They are making almost the identical point to the one made in the 2009 book Violence and Social Orders by Douglass North, John Wallis, and Barry Weingast (NWW), who argue that most underdeveloped societies are what they term “limited access orders” in which a rent-seeking coalition limits access to both the political and economic system. Indeed, I see no real difference between the “extractive/inclusive” distinction in AR and the “limited/open” access distinction in NWW.

This conclusion about the primacy of institutions and politics for development has important implications for policy as AR point out. If growth is a byproduct not just of good policies like trade liberalization, which can in theory be turned on like a light switch, but rather of basic institutions, then the prospects of foreign aid look dim. Bad governments can waste huge amounts of well-intentioned outside resources; indeed, the flow of aid dollars into poor countries can undermine governance by undercutting accountability, thereby leaving societies worse off than they would otherwise be. As the American nation-building efforts in Afghanistan and Iraq have indicated, moreover, foreign efforts to help construct basic institutions are an uphill struggle. Bad institutions exist because it is in the interests of powerful political forces within the poor country itself to keep things this way. Hamid Karzai understands perfectly well how clean government is supposed to work; it’s just that he has no interest in seeing that happen in Afghanistan. Unless the outsiders can figure out a way to change this political calculus, aid is largely useless.

So far, so good. AR have done a great deal over the years to focus the attention of both theorists and policymakers on institutions, and to shape the emerging consensus on the importance of politics to growth within the economics profession. It is, then, very disappointing that their more fully fleshed out book fails to go very much further than these broad conclusions, skirting critical issues of exactly what sort of institutions are necessary to promote growth, and failing to come to grips with some critical historical facts.

The first problem with their analysis is conceptual. They present a sharply bifurcated distinction between what they call good “inclusive” economic and political institutions, which are sometimes also labeled “pluralistic,” in contrast to what they call bad “extractive” or “absolutist” ones. Unfortunately, these terms are way too broad, so broad indeed that AR never provide a clear definition of everything they encompass, or how they map onto concepts already in use. “Inclusive” economic institutions, for example, seem to include formal property rights and court systems, but also have to do with social conditions that allow individuals access to the market such as education and local custom. “Inclusive” political institutions would seem to imply modern electoral democracy, but they also include an impersonal centralized state as well as access to legal institutions, and forms of political participation that fall well short of modern democracy. We find, for example, that England following the Glorious Revolution of 1688-89 was incipiently inclusive, despite the fact that well under ten percent of the population had a right to vote. When AR first used the term “extractive” in their early articles, it referred to truly extractive practices like the mines of Potosí or the sugar plantations of the Caribbean which extracted commodities out of the labor of slaves. In the current book extractive seems to mean any institution that denies any degree of participation to citizens, from tribal communities to ranchers in 19th century Argentina to the contemporary Chinese Communist Party.

Since each of these broad terms (inclusive/extractive, absolutist/pluralistic) encompasses so many possible meanings, it is very hard to come up with a clear metric of either. It also makes it hard to falsify any of their historical claims. Since more real-world societies are some combination of extractive and inclusive institutions, any given degree of growth (or its absence) can then be attributed either to inclusive or extractive qualities ex post.

The use of such broad categories and the failure to distinguish between the different components of political “inclusion” greatly diminish the book’s usefulness, because one wants to know how these components individually affect growth, and how they interact with one another. There is for example a large literature comparing the separate impacts of a modern state, rule of law, and democracy on growth, which tends to show that the first two of these factors have a far greater influence on outcomes than democracy. There is in fact a lot of reason to think that expansion of the franchise in a very poor country may actually hurt state performance because it opens the way to clientelism and various forms of corruption. The Indian political system is so inclusive that it can’t begin major infrastructure projects because of all the lawsuits and democratic protest, especially when compared to the extractive Chinese one. Furthermore, as Samuel Huntington pointed out many years ago, expanded political participation may destabilize societies (and thereby hurt growth) if there is a failure of political institutions to develop in tandem. All of the good things in the “inclusive” basket, in other words, don’t necessarily go together, and in some cases may be at odds. You never get much hint of this in Why Nations Fail, however, since the authors seem to argue the more inclusion the better, along any of its axes.

Like many other works making use of history but written by economists, the AR volume contains some pretty problematic facts and interpretations. It makes the case, for example, that Rome shifted away from an inclusive Republic towards absolutist Empire, and that this was then responsible for Rome’s subsequent economic decline. Leave aside the fact that Rome’s power and wealth continued to increase in the two centuries after Augustus, and that its eastern wing managed to hold on remarkably until the fifteenth century. It can be argued that the shift from the narrow oligarchy of the Republic to a monarchy with highly developed legal institutions actually increased access to the political system on the part of ordinary Roman citizens, while solving the acute problem of political instability that bedeviled the late Republic.

Similarly, following on a tradition begun by Douglass North and Barry Weingast, AR point to the Glorious Revolution of 1688/89 as a critical juncture marking both the establishment of secure property rights and an “inclusive” political system. The latter point is fair enough, but English property rights were rooted in a much older tradition of common law dating from the Norman invasion, and had created a strong commercial civilization well before 1689. The Glorious Revolution was much less important in establishing the credibility of property rights per se, than of the Crown as a borrower, which explains why English public debt exploded in the century following that event.

Given their overall framework, the hardest thing for AR to explain is contemporary China. China today according to them is more inclusive than Maoist China, but still far from the standard of inclusion set by the US and Europe, and yet has been the fastest growing large country over the past three decades. The Chinese restrict access to the market, engage in financial repression, fail to secure property rights, have no Western-style rule of law, and are ruled by a non-transparent oligarchy called the Communist Party. How to explain their economic success? Rather than see this as a threat to their model (i.e., more inclusion, more growth) AR pull a slight of hand by arguing that Chinese growth won’t last and that their system will eventually come crashing down (like Rome did, after about 200 years?). I actually agree that China will eventually crash. But even if that happens, a theory of development that can’t really explain the most remarkable growth story of our time is not, it seems to me, much of a theory.

The broad conclusions of Why Nations Fail are, thus, incontrovertible and of great importance to policy (which is why, incidentally, I gave it a positive blurb). One only wishes then that the authors had made better use of basic categories long in play in other parts of the social sciences (state, rule of law, patrimonialism, clientelism, democracy, and the like) instead of inventing neologisms that obscure more than they reveal.