Too efficient NOT to consolidate
Cross-posted at News N Economics blog, by Rebecca
Here’s yet another historical record broken in 2009:
“Only three insured institutions were chartered in the [third] quarter, the smallest quarterly total since World War II.”
This fact is from the FDIC’s latest Quarterly Banking Profile. There are probably non-economic reasons for this, i.e., the application process to qualify as a new charter institution (see the types of charters here) is likely much more stringent than in previous years; but nevertheless, this fact reiterates the trend in the number of banking institutions, most definitely down.
The FDIC is awash in problem institutions. The well reported number of bank failures jumped to 132 in 2009 (as of November 20, and you can find the data here). However, that’s just share of the much larger “problem”. According to the same quarterly profile, there are now 552 “Problem” institutions in the FDIC charter system holding $346 billion of assets on balance – that’s 2.4% of nominal GDP.
As such, it seems that consolidation is all but a foregone conclusion. But watch out, because the new 4-letter-style phrase, “too big to fail”, is heavy on the tongues of US policymakers. Senator Bernard Sanders (Vermont) introduced the “Too Big To Fail Is Too Big to Exist” bill last month, which defines such an institution as (see the bill here):
“any entity that has grown so large that its failure would have a catastrophic effect on the stability of either the financial system or the United States economy without substantial Government assistance.”
Ahem, so how big is that? Peter Boone and Simon Johnson at the Baseline Scenario define “too big to fail” as bank liabilities amounting to 2% of GDP (roughly):
“So to us, 2 percent of G.D.P. seems about right. This would mean every bank in our country would have no more than about $300 billion of liabilities.
A large American corporation would still be able to do all its transactions using several banks. They would even be better off — competition would ensure that margins are low and the banks give the corporates a good deal. This would help end the situation where banks take an ever-increasing share of profits from our successful nonfinancial corporations (as seen in the rising share of bank value added in G.D.P. in recent decades).”
But there are economic efficiencies, like scale economies, that need to be considered. David C. Wheelock and Paul W. Wilson at the St. Louis Fed find statistically significant increasing returns to scale (i.e., bigger banks, lower costs) in the US banking system. They use a non-parametric estimator to estimate a model of bank costs and find the following (link to paper, and bolded font by yours truly):
“The present paper adds to a growing body of evidence that banks face increasing returns over a large range of sizes. We use nonparametric local linear estimation to evaluate both ray-scale and expansion-path scale economies for a panel data set comprised of quarterly observations on all U.S. commercial banks during 1984-2006. Using either measure, we find that most U.S. banks operated under increasing returns to scale. The fact that most banks faced increasing returns as recently as 2006 suggests that the U.S. banking industry will continue to consolidate and the average size of U.S. banks is likely to continue to grow unless impeded by regulatory intervention. Our results thus indicate that while regulatory limits on the size of banks may be justified to ensure competitive markets or to limit the number of institutions deemed too-big-to-fail, preventing banks from attaining economies of scale is a potential cost of such intervention.”
Better put: the cost of consumer and firm loans will be higher in the long run if too much intervention prevents the banking system from capturing scale economies. Furthermore, they suggest that even the largest institutions experience increasing returns (i.e., these).
I should say that I have absolutely no experience in non-parametric estimation and cannot vouch for the econometrics. However, the results are timely; and furthermore, the Federal Reserve Bank of St. Louis’ economics research is well-regarded. Point: I trust it.
As a note, David Wheelock wrote a very interesting piece a while back about the inefficiencies of mortgage foreclosure moratoria during the Great Depression …interesting stuff (paper link here).
The fact that bank costs decline with size isn’t a huge boon. All it means is that larger banks get monopoly power.
What Rob says.
Being able to escape state regulation and to charge 33% credit card interest should not be confused with “economies of scale” no matter how fancy the econometrics.
There’s been plenty of evidence of diseconomies of scale in many operational nodes of banking. We are in a financial crisis, in large part because some fools thought that there were economies of scale in mortgage underwriting, and drew most mortgage underwriting into the hands of a few giant aggregators: Countrywide, IndyMac, World Savings>Golden West>Wachovia, Washington Mutual. That pursuit of economies of scale worked out so well, I certainly wouldn’t want someone to ignore the “econometric evidence” provided by reactionaries at the St. Louis Fed, to actually regulate banking and prevent another crisis. I want me, “economies of scale”!!!
It is interesting that you wrote on this topic. Seeking Alpha had the same article from Baseline as written by Simon Johnson this morning you have quoted and defining TBTF although the article did admit there is no real parameter that can completely capture what is TBTF. The number or percentage based upon what-is -elusive. Some C&P:
“1) As legislation on restructuring the banking industry moves forward, attention on Capitol Hill is increasingly drawn to the issue of bank size. Should our biggest banks be made smaller?
2) There is a strong precedent for capping the size of an individual bank: The United States already has a long-standing rule that no bank can have more than 10 percent of total national retail deposits. http://money.cnn.com/2007/01/16/news/companies/bankofamerica_limit/index.htm This limitation is not for antitrust reasons, as 10 percent is too low to have pricing power. Rather, its origins lie in early worries about what is now called “macroprudential regulation” or, more bluntly, “don’t put too many eggs in one basket.”
3) This cap was set at an arbitrary level — as part of the deal that relaxed most of the rules on interstate banking — and it worked well (until Bank of America received a waiver).
4) Probably the best way forward is to set a hard cap on bank liabilities as a percent of gross domestic product; this is the appropriate scale for thinking about potential bank failures and the cost they can impose on the economy. Of course, there are technical details to work out — including how the new risk-adjustment rules will be enacted and the precise way that derivatives positions will be regarded in terms of affecting size. But such a hard cap would the benchmark around which all the specifics can be worked out.” http://seekingalpha.com/article/176664-the-costs-of-not-fixing-a-broken-financial-system?source=email
This is a great topic and worthy of additional examination by the pols and powers to be. One question, am I to assume non-parametric is non-normal distribution of data? My LSS seeping through . . . Thanks for writing on this topic.
Thank you for your comments! However, I think that one must look past the short-term banking crisis and toward the longer-term objective. Of course, regulation fell flat – this seems quite obvious. 2004-2007, in my view, is partially (I say partially because one must consider the easy money of the Fed, see David Beckworth’s blog) a failure of regulation, and not the market itself. Like most markets, the banking system is subject to over production and marginal social costs that are not incorporated by Countrywide and the like. However, that has absolutely nothing to do with the long-run cost structure and availability of loans. If there is a point at which the banking system is stifled because of regulation, I think that it should be avoided; over-regulation may be condemning.
Hi Run 75441,
How are ya? I agree – this is a great topic, and worthy of much more discussion. But unfortunately (again, in my view), the regulation topic takes a very second place to others, like health care (which is likewise important, of course). I have listened to Simon Johnson a lot over the last 2 years; he has a lot to say about the handling of the banking crisis – much criticism. We will see where it goes from here.
By the way, you are correct – a non-parametric test makes minimal assumptions about the population from whence it came; hence, no normal assumption required.
The whole issue of banks has been a great subject of debate since Alexander Hamilton and Thomas Jefferson. We had Andrew Jackson who thought banks were evil incarnate, and should be quashed. (The local banks of the time were close). Jackson of course did the central bank in. It is at least a little historically ironic that people in the successor party to that of Hamilton now oppose the central bank, and the Jefferson/Jackson party favor it. Because of this the country for a long time had small banks and several states had unit banking rules. (One office only) So a lot of banks failed in the 1920-1935 period because they had no geographic diversity. Canada on the other hand has basically 5 large nationwide banks and for various reasons did not see a lot of failures during the 1920-1935 period.
***This is a great topic and worthy of additional examination by the pols and powers to be. One question, am I to assume non-parametric is non-normal distribution of data?***
I think “non-parameteric” means that a normal distribution is not assumed. i.e. that the math works whether the data is distributed normally or not. 2sugs would probably know. Perhaps he’ll drop by and clarify.
I do hope that “non-parametric” is not going to become the meaningless buzz-word d’jour.
Is it not kind of a self fulfilling prophecy, that as consolidation takes place, failures will increase do to the rise of monopoly power. Monopoly power can appear regionally, even locally before we see it nationally. It is the success mechanism of the big retail chains.
It is an error to assume the rise in banks in trouble is a result of market efficiency and not actually the result of small area monopoly power. I suspect small area monopoly power would be viewed as market saturation and thus lead to the argument that the strongest will survive?
But, the complete error in all of economic since the Chicago boy came to influence is the focus on the efficiency of money. We should focus on the efficiency of people’s life. An efficient life is a life of continually reducing risk. It cost money to do that. More money than what it costs to continually increase the efficiency of money.
It is a non-normal data distribution. I was just checking to make sure. You would go to non-parametric data curve determination if the Anderson Darling Test revealed a P Value < .05 in a Normal Test. Economists have a separate lanquage of their own and I was just checking to see if it was the same as my understanding. In Lean Six Sigma, I do statistical analysis also in business and plant environments. In MiniTab, there is a section for "Individual Distribution Identification" I use to identify the type of distribution the data represents. Beyond MiniTab, we use DataFit for more complex data distribution. And yes 2Slugs would know and probably Cantab would know also from what I have read of his. Thanks for your input though . .
Bank regulation in Canada is a different game than here in the US – much more of it – and why the economy did not experience a banking crisis. Of course some would argue that Canada has its own housing problems in the not-too-distant future (Garth Turner at the Greater Fool). Thanks for the comparison!
I suspect that no matter where in the world one looks one can develop problems particulally if you a part of the problem development industry (media is a large part of this industry, politicans another, and pudits a third) .
Housing has been a problem in society almost forever, or probably forever, and in every semi devloped or more society. Just like the poor will always be with us and perhaps proceeding from it the homeless will always be with us.
“Housing has been a problem in society almost forever, or probably forever, and in every semi devloped or more society. Just like the poor will always be with us and perhaps proceeding from it the homeless will always be with us.”
Slaves and serfs had places to live, such as they were. The rich no longer own people, but they do not have to feed and house them, either. Despite our prosperity, people are consigned to economic serfdom. Poverty is passed down from generation to generation. America prides itself on being the land of opportunity, but we do not provide much of a path out of poverty for the children of the poor. Surprisingly, Europe, the home of aristocracy, does better now than we do.