In this post I write about the connections between Paul Romer’s work, which is essentially applied theory, and the empirical work on long-run economic growth done by economic historians. How was Romer influenced by the work of economic historians? has he influenced economic history? and have his theories been confirmed by the recent work of economic historians? (preview: I will argue that the answers are: yes; not much; and no).
Addendum shortly after publishing: my point above is not that Romer is wrong in general; in fact some of his ideas *about ideas* are fundamental for us to think about growth in the past. (Read on if this isn’t clear yet.)
Paul Romer’s was a well-deserved and long-anticipated prize. Many predicted he would eventually win, including myself in my very first academic article, written when I was a undergradute and published in 2008. (alternatively, click here for an ungated version). I now find it mildly amusing how assertive I was when I wrote: “Paul Romer is going to win the Nobel Prize in economics”. I continue to believe that this was a good choice.
Many have written about the nature of his main contributions, all of which, as I have said, have been on applied theory; see for instance, see the posts in Dietrich Vollrath’s blog, here and here, or in Paul Romer’s own blog.
Romer’s work had some influence on economic history, but not much. There is, for instance, a 1995 article by Nick Crafts which looks at the Industrial Revolution from a New Growth perspective, but it is fair to say that economic historians were perhaps not quick to pick up the New Growth theory train. Part of this was surely because its implications seemed to apply mostly to frontier economies and did not seem to apply to much of human history, a limitation which Unified Growth Theory would later attempt to overcome.
And yet, Romer himself has often spoken about economic history and relied on the data of economic historians. He now seems to have won mostly due to his 1990 article, but his earlier work on increasing returns (ungated version here) had a graph from Maddison, for instance (he discusses how Maddison influenced him here). And the process of growth itself was documented by economic historians using graphs such as the following; as my friend Max Roser tweeted:
(as a side note: Max, the source for this is not the Bank of England; it is Broadberry, Campbell, Klein, Overton, and van Leeuwen; empirical work is very demanding, so citations need to be fair with the people who did the hard work in putting these figures together).
One of the most empirical papers Romer has written is “Why indeed in America?”, which was the culmination of much of what he had done before. It was also one of the last papers he wrote before entering a writing hiatus. In this paper he explicitly argues for the complementarity of economic history and growth theory. He argues that the USA achieved economic supremacy after 1870 due to having the largest integrated market in the world. He writes:
“differences in saving and education do not explain why growth was so much faster in the United States than it was in Britain around the turn of this century. In 1870, per capita income in the United States was 75 percent of per capita income in Britain. By 1929, it had increased to 130 percent. In the intervening decades, years of education per worker increased by a factor of 2.2 in Britain and by a nearly identical factor of 2.3 in the United States. In 1929, this variable remained slightly lower in the United States. (Data are taken from Angus Maddison .”
Notice that there are three empirical statements here. Romer’s story builds on these facts, so if the facts change, the story must too. Theory depends on facts.
The first fact (according to Romer) is that the US only converged to British per capita GDP levels after 1870. Second, that this was not due to matters such as education or savings. Third, the reason was market size. As economic historians, we have made much progress in measuring each of these matters since 1995. Let me consider each in turn.
Timing of convergence of the USA to Britain
The important thing to keep in mind here is that it is by no means certain that the USA had not catched up earlier. The methodological issues are complicated and in fact today’s other (and equally deserving) Nobel prize, Nordhaus, wrote a fascinating paper about the problems involved in these types of measurements. (A popular description of this work can be read here.) As far as USA vs Britain is concerned, though, Marianne Ward and John Devereux summarize the debate as follows:
“Prados De la Escosura (2000) and Ward and Devereux (2003, 2004, 2005) argue for an early US income lead using current price estimates. Broadberry (2003) and Broadberry and Irwin (2006) defend the Maddison projections while Woltjer (2015) hews to a middle ground. The literature has recently taken an unexpected turn as Peter Lindert and Jeffrey Williamson, Lindert and Williamson (2016), find a larger US lead before 1870 and one that stretches further back in time than claimed by either Prados De la Escosura (2000) or Ward and Devereux (2005).”
Comparative levels of education
Recent evidence suggests that the average years of post-primary education actually declined in Britain after about 1700. (ungated version here). This was not the case at all in the USA, where it is well-known that the state invested in high schools, so it seems unlikely that average human capital grew at similar levels in the latter part of the 19th century, as Maddison/Romer claimed.
Addendum: minutes ago when I first posted, I initially wrote “post-secondary” where I meant to write “post-primary”
I used to believe this part of Romer’s story. That was until I read this brilliant paper by Leslie Hannah: “Logistics, Market Size, and Giant Plants in the Early Twentieth Century: A Global View”. (Ungated version here). Notice that Hannah does not refer to Romer’s argument or even cite him. What he does instead is he destroys the commonly held idea that USA’s market size was larger that Europe’s already before the Great War (aka World War I). It is true that the USA had more railroads, but it also had much longer distances. In Northwestern Europe, transportation by a mix of ships, trains and horses was cheaper, especially once we consider the much denser (and highly urbanized) population. It is important to remember that prior to WWI, Europe was living the “first age of globalization”, with high levels of integration and relatively low tariffs.
So, this part of Romer’s story cannot be right.
In conclusion, what does this all mean? will these new facts affect where growth theory will go? only time will tell, and growth theory itself is by no means moving much these days, as Paul Romer himself has addmited in recent interviews. What these facts suggest, though, is that other things must have mattered.
As I said in the beginning, I believe that Paul Romer’s applied theory work is important (as it is that of others that might have won, such as Aghion and Howitt). The natural complementaries between the work of economic historians and applied theorists suggest that we need to listen to each other in order for science to move forward. Hopefully, new generations of economists will do a bit of both, as have some people who now work on Unified Growth.
But in the future, it is fair that the Nobel committee gives more prizes to empirical work as well. Because theory can’t live without facts, but economics Nobels have been highly biased towards theorists (whether pure or applied).
Final addendum, about one hour later than the original post: Max Roser read this post and has now corrected the citation. This is not the first time I give him a slap on the wrist about this sort of thing, but I know that Max, who is a friend and a promoter extraordinaire of this sort of work, is well-intentioned. Yet it is crucial that we insist on this being done fairly everytime, because if even Max occasionally does this wrong, that shows that this is the sort of thing that easily happens. The root cause is indirect citations, i.e. citation of someone who cited the data instead of citing the original work. Doing this takes credit away from those who did the basic empirical work, even when that is unintentional. So we all need to be careful to cite those who produced the original data.