I was sad to hear of the passing of the great Douglass North earlier this week (there are obituaries from the Economist, New York Times and Washington University in St. Louis, the last focusing on his teaching). While I cannot claim deep knowledge of North’s whole body of work, I loved his book Violence and Social Orders, co-authored with John Joseph Wallis and Barry R. Weingast. Amazingly enough, the book completely lives up to its subtitle: A Conceptual Framework for Interpreting Recorded Human History. North was a social scientist who, like Gellner and Levi-Strauss, aimed straight for the big questions about what modern life is; most economists do not even know how to ask those questions let alone answer them.

The book is difficult to summarize simply so I won’t try–but one of the lessons I learned from it is in fact pretty simple. The difference between good institutions and bad institutions in terms of economic growth is not that good institutions generate higher growth. Rather, countries with good institutions are flexible and better at responding to changing conditions, so they have fewer and shallower economic downturns. That results in a higher rate of trend growth over the long term. It’s a powerful and intuitive idea: the important thing is not finding the secret sauce for economic growth, but avoiding and recovering from mistakes. Here is the relevant passage:

An underappreciated feature of the different patterns of social orders relates to why poor countries stay poor. Economic growth, measured as increases in per capita income, occurs when countries sustain positive growth rates in per capita income over the long term. Over the long stretch of human history before 1800, the evidence suggests that the long-run rate of growth of per capita income was very close to zero. A long-term growth rate of zero does not mean, however, that societies never experienced higher standards of material well-being in the past. A zero growth rate implies that every period of increasing per capita income was matched by a corresponding period of decreasing income. Modern societies that made the transition to open access, and subsequently became wealthier than any other society in human history, did so because they greatly reduced the episodes of negative growth. …

Strikingly, the richest countries are not distinguished by higher positive growth rates when they do grow. In fact, the richest countries have the lowest average positive growth rates by a substantial amount. …When they grow, poor countries grow faster than rich countries. They are poor because they experience more frequent episodes of shrinking income and more negative growth during the episodes. Countries below $20,000 income do not exhibit a strong relationship between income and positive growth rates. The same is not true for the relationship between income and negative growth rates. …The poorest countries experience both more years of negative income growth and more rapid declines during those years…

All societies are subject to random and unpredictable changes in the world around and within them. Changes in external factors like climate, relative prices, and neighboring groups as well as changes in internal factors like the identity and character of leaders, internal feuds and disputes, and relative prices all contribute to persistent alterations in the circumstances with which societies must cope. The variations in the economic performance of limited and open access societies over time reflect the inherent ability of the two social orders to deal with change. …There is no teleology implied by the framework. Nonetheless, the framework illuminates why open access societies are better than natural states at dealing with change.

The latest historical economic research in fact seems to strongly support North’s thesis. The new issue of the Journal of Economic Perspectives has a nice review article on recent work compiling long-term GDP series for several European countries. There is only one chart but boy is it a doozy–just think of all the years of scholarly effort that went into creating these data series:

The chart seems to very much supports North’s thesis: historically there were indeed many episodes of growth, but they were usually followed by significant reversals. At least, until the Industrial Revolution in England came along and delivered much more consistent gains. Here are the authors, Roger Fouquet and Stephen Broadberry:

The new data shows trends in GDP per capita in the key European economies before the Industrial Revolution, identifying episodes of economic growth in specific countries, often lasting for decades. Ultimately, these periods of growth were not sustained, but they noticeably raised GDP per capita. It also shows that many of these economies experienced periods of substantial economic decline. Thus, rather than being stagnant, pre-nineteenth century European economies experienced a great deal of change. …

The paper tentatively finds that the likelihood of being in a phase of growth increased and the risk of being in a phase of decline decreased in the nineteenth and twentieth centuries. …Between the fifteenth and eighteenth century, there was an average of two economic downturns per country per century, while the nineteenth and twentieth centuries experienced less than one economic downturn per country per century. In the fifteenth and sixteenth centuries, economic downturns occurred about 8 percent of the time; in the seventeenth and eighteenth centuries, they were experienced 4–5 percent of years; and, in the nineteenth and twentieth centuries, downturns occurred 2–3 percent of the time. Thus, there appears to have been a modest reduction in the likelihood of experiencing downturns over the centuries from the fifteenth century. …Explaining the source of these differences could prove to be important for understanding how economies managed to generate sustained economic growth.

Looks like North was onto something.