It could be that large banks can finance themselves more cheaply because they are more efficient, that is, that there are economies of scale in banking. For some time, the received wisdom was that there was no evidence of such economies beyond relatively modest-sized banks, with balance sheets of approximately $100 billion. More recently, several papers have found that economies of scale may continue beyond that level. For example, the title of a paper by Joseph Hughes and Loretta Mester, "Who Said Large Banks Don't Experience Scale Economies? Evidence from a Risk-Return Driven Cost Function" suggests that large institutions may be better able to manage risk more efficiently because of "technological advantages, such as diversification and the spreading of information...and other costs that do not increase proportionately with size." That said, these authors conclude that "[W]e do not know if the benefits of large size outweigh the potential costs in terms of systemic risk that large scale may impose on the financial system." They add that their results suggest that "strict size limits to control such costs will likely not be effective, since they work against market forces..."
The TBTF theory of why large banks are a problem has to contend with the history of the Canadian and Australian banking systems. Both these systems have several very large banks, but both systems have been very stable--in the Canadian case, for 150 years. Beck, Demirguc-Kunt, and Levine (2003) examined the impact of bank concentration, bank regulation, and national institutions on the likelihood of a country suffering a financial crisis and concluded that countries are less likely to suffer a financial crisis if they have (1) a more concentrated banking system, (2) fewer entry barriers and activity restrictions on bank activity, and (3) better-developed institutions that encourage competition throughout the economy.32 The combination of the first finding with the other two appears paradoxical, but the key barrier to competition that was absent in Canada was the prohibition of nationwide branch banking, a factor emphasized by Calomiris and Haber in their discussion of the Canadian case. In addition, I put serious weight on another explanation offered in private conversation by a veteran of the international central banking community, "Those Canadian banks aren't very adventurous," which I take to be a compliment.
...
Why is the TBTF phenomenon so central to the debate on reform of the financial system? It cannot be because financial institutions never fail. Some do, for example, Lehman Brothers and the Washington Mutual failed in the Great Recession
Almost certainly, TBTF is central to the debate about financial crises because financial crises are so destructive of the real economy. It is also because the amounts of money involved when the central bank or the government intervenes in a financial crisis are extremely large Another factor may be that the departing heads of some banks that failed or needed massive government assistance to survive nonetheless received very large retirement packages.
One can regard the entire regulatory reform program, which aims to strengthen the resilience of banks and the banking system to shocks, as dealing with the TBTF problem by reducing the probability that any bank will get into trouble. There are, however, some aspects of the financial reform program that deal specifically with large banks. The most important such measure is the work on resolution mechanisms for SIFIs, including the very difficult case of G-SIFIs. In the United States, the Dodd-Frank Act has provided the FDIC with the Orderly Liquidation Authority (OLA)--a regime to conduct an orderly resolution of a financial firm if the bankruptcy of the firm would threaten financial stability. And the FDIC's single-point-of-entry approach for effecting a resolution under the new regime is a sensible proposed implementation path for the OLA.
Closely associated with the work on resolution mechanisms is the living will exercise for SIFIs. In addition, there are the proposed G-SIB capital surcharges and macro stress tests applied to the largest BHCs ($50 billion or more). Countercyclical capital requirements are also likely to be applied primarily to large banks. Similarly the Volcker rule, or the Vickers rules in the United Kingdom or the Liikanen rules in the euro zone, which seek to limit the scope of a bank's activities, are directed at TBTF, and I believe appropriately so.
What about simply breaking up the largest financial institutions? Well, there is no "simply" in this area. At the analytical level, there is the question of what the optimal structure of the financial sector should be. Would a financial system that consisted of a large number of medium-sized and small firms be more stable and more efficient than one with a smaller number of very large firms? That depends on whether there are economies of scale in the financial sector and up to what size of firm they apply--that is to say it depends in part on why there is a financing premium for large firms. If it is economies of scale, the market premium for large firms may be sending the right signals with respect to size. If it is the existence of TBTF, that is not an optimal market incentive, but rather a distortion
Would breaking up the largest banks end the need for future bailouts? That is not clear, for Lehman Brothers, although a large financial institution, was not one of the giants--except that it was connected with a very large number of other banks and financial institutions. Similarly, the savings and loan crisis of the 1980s and 1990s was not a TBTF crisis but rather a failure involving many small firms that were behaving unwisely, and in some cases illegally. This case is consistent with the phrase, "too many to fail." Financial panics can be caused by herding and by contagion, as well as by big banks getting into trouble.
In short, actively breaking up the largest banks would be a very complex task, with uncertain payoff.