Last week, Charles Calomiris wrote an op-ed in the Wall Street Journal arguing that big banks are better for various reasons. Simon wrote last week saying that Calomiris underestimated the political dimension, and that his proposed solution — a cross-border resolution mechanism for large institutions — is the policy equivalent of assuming a can opener.
I wanted to look at Calomiris’s specific claims. I think I’ve already dealt with the myth that banks “need to be large to operate on a global scale—and they need to do so because their clients are large and operate globally.” Calomiris also argues that there are economies of scope (it’s better to be big because you can play in multiple businesses). Here’s his evidence:
“True, some empirical studies in the field of finance have failed to find big gains from mergers. But those studies measured gains to banks only, and measured only the performance improvements of recently consolidated institutions against other institutions, many of which had improved their performance due to previous consolidation.
“Yet even unconsolidated banks have improved their performance under the pressure of increased competition following the removal of branching restrictions, which permitted the consolidation wave in banking. And when an entire industry is involved in a protracted consolidation wave, the best indicator of the gains from consolidation is the performance of the industry as a whole. One study of bank productivity growth during the heart of the merger wave (1991-1997), by Kevin Stiroh, an economist at the New York Federal Reserve, found that it rose more than 0.4% per year.”
Note that Calomiris concedes that you can’t find benefits from mergers by looking at merged banks directly; this is why he falls back on an industry average.
First of all, there must be a joke to be made here about correlation and causality. Wait, here it is.
Second, 1991-97 was only the beginning of the merger wave; The Riegle-Neal Interstate Banking Act wasn’t passed until 1994. Let’s assume that mergers after 1995 wouldn’t show up in the 1991-97 data. That includes Nations-Boatmen’s, Nations-Barnett, Nations-Bank of America, B of A-FleetBoston, Chemical-Chase, Bank One-First Chicago, J.P. Morgan-Chase, Wells-First Interstate, Wells-Norwest, and Wachovia-First Union.
Third, I was interested by that statistic, because 0.4% annual productivity growth from 1991 to 1997 seems like nothing to write home about – labor productivity growth (the figure you usually read about) over that period was about 1.7% per year for the economy as a whole. So I tracked down the source: Kevin J. Stiroh, “How did bank holding companies prosper in the 1990s?,” Journal of Banking & Finance, Volume 24, Issue 11, November 2000, Pages 1703-1745. (I’m not sure if anyone can download the paper or I could only do it because I’m on the Yale VPN.)
I found out that Stiroh is measuring total factor productivity, not labor productivity, so 0.4% is better than average for the non-manufacturing sector. (The economy average was 0.3%, but the manufacturing sector was 1.9%, so by implication the non-manufacturing sector was less than 0.3%.) So far, so good.
But what does Stiroh say about this productivity growth? His main explanation for the productivity growth is not consolidation, but information technology: “The finding of steady productivity growth, in particular, is important since it is consistent with the idea that the massive investment in new technology is working to improve the performance of the banking industry.” This is not proven in this paper, but Stiroh went on to write a bunch of other papers on the link between information technology and productivity. For example, this paper (on the entire economy, not just banking) concludes:
“IT-producing and IT-using industries account for virtually all of the productivity revival that is attributable to the direct contributions from specific industries, while industries that are relatively isolated from the IT revolution essentially made no contribution to the U.S. productivity revival. Thus, the U.S. productivity revival seems to be fundamentally linked to IT.”
That second paper also finds (Table 2) that productivity acceleration — the difference between productivity growth in the 1995-99 period relative to the 1987-95 period — was lower in finance, insurance, and real estate than in the economy as a whole, including the services sector. I don’t know exactly what this means, but at first glance it doesn’t look good for banking consolidation.
Going back to the first paper (the one Calomiris cites), what does Stiroh say about bank size? “Despite the strong overall performance, roughly 10% of costs were due to inefficiency and 30-40% of potential profits were missed. Moreover, efficiency does not significantly increase with bank size as one might expect if economies of scale are an important determinant of success. Rather, there are efficient and profitable BHCs in every size class and increased size does not guarantee success.” Figures 6 and 7 show that there is no difference in cost efficiency across size classes and that the largest banks actually seem to have lower profit efficiency.
However, Stiroh also says that continuing consolidation seems to offer the possibility of reducing these inefficiencies, so on balance I would say he is slightly positive about size. Stiroh also has a paper on banks in Switzerland which I didn’t read, but whose abstract says:
“We find evidence of economies of scale for small and mid-size banks, but little evidence that significant scale economies remain for the very largest banks. Finally, evidence on scope economies is weak for the largest banks that are involved in a wide variety of activities. These results suggest few obvious benefits from the trend toward larger universal banks.”
I wish I had more time to read more of these papers. It seems to me that Stiroh has done a lot of serious empirical research on banking productivity and finds the evidence mixed — consolidation should be good, but it doesn’t really show up in the data. I haven’t read much of his work (or talked to him) and I certainly don’t want to imply that he is against big banks. I just think that citing a study he did of the 1991-97 period to back up a claim that banking mergers are good is a bit of a stretch.
By James Kwak