The too-big-to-fail problem is clearly at the heart of the recent financial crisis. Obviously, our policymakers believe that some financial institutions are so large and "systemically important" that they cannot be allowed to be fail, the large financial institutions in question understand this, and the result is a moral hazard problem. Through some failure of financial regulators to use their endowed powers, insufficient regulations, or due to the ability of financial institutions to evade regulatory scrutiny, large financial intermediaries took on more risk than was socially appropriate. This caused, or at least exacerbated, the financial crisis.
Leaving aside the issue of what should be done about large non-bank financial institutions (a key problem in itself), what should be done about the too-big-to-fail problem associated with large US banks? If we think too-big-to-fail is a problem, then it is clear that the problem is getting worse. As Ned Prescott (Richmond Fed) pointed out in a discussion at a recent conference at the Philadelphia Fed, the number of US banks has fallen dramatically recently, and concentration has increased - a larger fraction of market share is accounted for by the largest banks. There are potentially two reasons for this. First, with the deregulation of the US banking industry - the dismantling of various barriers to setting up additional branches and otherwise expanding individual banks - banks can reap the returns from economies of scale. Second, to the extent that there is too-big-to-fail, large banks are implicitly subsidized.
In my opinion, some economists tend to overemphasize the second effect, and want to lead us to believe that economies of scale are no big deal in banking activity. Certainly my old friend John Boyd (U. of Minnesota Finance Department) thinks this way. Others with this attitude are Simon Johnson and James Kwak, who have written here about it. Their proposed solution to too-big-to-fail is breaking up the large banks, much like we broke up ATT in 1984. For some econometric evidence on economies of scale in banking see this paper by David Wheelock and Paul Wilson.
There are still many small banks in the US, mainly due to historical accident and regulation. US banking started as a system where banks were chartered and regulated by state governments and typically restricted to operating within the state - sometimes restricted to one establishment (a unit bank). It was not until after the Civil War that we had a system that included National banks, regulated by the OCC (Office of the Comptroller of the Currency). A key problem with a unit banking system is that a local bank with loans collateralized primarily by local real estate is risky - it is poorly diversified and will fail with high probability. Diversification is where the economies of scale in banking come from. A large bank that can branch nationally is better diversified, and therefore more efficient.
How do we pool risk and get the diversification we need from a unit banking system that cannot provide it on its own? Securitization. For example, in the mortgage market, Fannie Mae and Freddie Mac did a tolerable job of creating a set of standards for conforming loans that mortgage lenders met, with the mortgages sold to Fannie Mae and Freddie Mac, bundled as tradeable securities, and then held and traded by various financial institutions. There is some loss in efficiency relative to a system with large banks, though, as local mortgage lenders have little discretion about who to lend to. They may have good information that a borrower is creditworthy, but if the borrower does not fit the Fannie Mae cookie cutter profile, he or she does not get a loan. Further, as became clear with the subprime crisis, securitization can be associated with some severe incentive problems, which we are painfully aware of now. In addition, Fannie Mae and Freddie Mac, due to political pressure, ineptitude, corruption, or some combination of the three, relaxed its standards, and had to be taken over by the federal government.
Large banks that branch nationally can clearly avoid the incentive problems associated with securitization. Loans are originated and held by a given bank, and loan officers who do a poor job of screening loans suffer termination, unlike the fly-by-night mortgage broker villains of the financial crisis. Does such a banking system exist? Of course, everyone knows that. Our neighbors (and my friends and relatives) in the Great White North have what appears to be one such first-rate banking system. The Canadian banking system is dominated by the five largest banks - the Royal Bank of Canada, the Toronto Dominion Bank, the Bank of Nova Scotia, the Bank of Montreal, and the Canadian Imperial Bank of Commerce. Indeed, where I grew up, a town of 10,000 people at the time (Cobourg, Ontario), there was one of each. These five banks have been around (or the banks that merged to yield these giants have been around) since the 19th century, and they are remarkably stable. There have been a total of three chartered bank failures in Canada since 1900 - one in the 1920s and two small regional bank failures in the 1980s. Canada has had deposit insurance since 1967, but apparently the CDIC (Canadian Deposit Insurance Corporation) does not have much to do. There were no bank failures in Canada during the Great Depression, and Canadian banks received no direct government support (outside of the essentially zero interest rate policy of the Bank of Canada) during the financial crisis. Further, the Canadian banking system is apparently not some slothful oligopoly. They compete successfully in international financial markets, and have always been well ahead of the average US bank in terms of electronic banking - for example the use of checks (or cheques in appropriate Canadian English) faded long ago.
Looks pretty good, right? Well, not according to Peter Boone and Simon Johnson, who write here. The gist of this piece is that (i) Canadian banks may look safe, but they are actually highly levered and risky; (ii) Canadian banks are subsidized by government-provided mortgage insurance; (iii) Even if we wanted a Canadian banking system in the US, we couldn't have it anyway.
Boone and Johnson state that "Canadian banks were actually significantly more leveraged — and therefore more risky — than well-run American commercial banks." First, given the distortions in the rest of the piece, I'm suspicious about Boone and Johnson's leverage numbers. However, suppose I think that the leverage comparisons they report are accurate. Any good Sloan School Professor should know that leverage and risk are not the same thing. For example, a bank that holds only treasury bills could be very highly levered and essentially riskless - riskiness of the bank is determined by capitalization and the riskiness of the asset portfolio. The financial crisis ran the stress test on Canadian banks - they just can't be characterized as risky. Notice as well in the Boone and Johnson piece that they pick and choose their banks. Obviously they're not making comparisons to Bank of America, Citigroup, or Wachovia (oh right, that one doesn't exist anymore).
Next, Boone and Johnson say: "If Canadian banks were more leveraged and less capitalized, did something else make their assets safer? The answer is yes: guarantees provided by the government of Canada." Any mortgage made by a Canadian bank with less than a 20% down payment must be insured, and the insurance is provided by a government agency, CMHC (the Canada Mortgage and Housing Corporation, which has a role something like the FHA). CMHC provides the mortgage insurance, and the bank that makes the mortgage loan pays the insurance premium. If the mortgage-holder defaults, CMHC pays off the bank.
To understand what is going on here, we have to go deeper into Canadian banking regulation. My hypothesis is that the Canadians have somehow solved the too-big-to-fail problem. The banks are too-big-to-fail, but they never fail, so problem solved. How do they do it? Canadian banks are regulated by the Office of the Superintendent of Financial Institutions (OSFI). Apparently, there is ONE banking regulator, rather than the highly inefficient alphabet soup of overlapping and competing regulators we have in the US. In Canada, the Bank of Canada collects banking data, and the Bank of Canada and the CDIC cooperate with OSFI, but it is OFSI that has the regulatory authority. Further, OSFI also regulates insurance, and guess what? OSFI regulates the mortgage insurance activity of CMHC. There may be some subsidy implicit in the CMHC insurance, since CMHC is in part a vehicle for direct government lending and some of that subsidy may spill over. However I can't see that it's the big deal that Boone and Johnson want to make it out to be. The mortgage insurance certainly looks redundant - the banks seem quite capable of diversifying the risk on their own - but it looks relatively innocuous.
Boone and Johnson want to conclude that embracing large banks in the US, and simply doing a better job of regulating them, is a recipe for disaster. Baloney. Why not figure out what OSFI does, try to replicate it, do away with Fannie and Freddie, and do what we can to discourage mortgage securitization? Economizing on the replication in regulation in the US would be nice, but you can't have it all.