Some observers argue that government policy favors the biggest handful of banks at the expense of all the others, which leads many of them to call for a break up of those mega-banks. The Government Accountability Office (GAO), an arm of Congress, released a report today that is intended to address part of the question – whether the largest banks pay less to raise funds because markets assume taxpayers may bail those firms out. This advantage is often referred to as an “implicit subsidy”, because it effectively subsidizes the operations of the biggest banks, but is based on an implication of government support, not an official policy. (Please see my recent primer on implicit subsidies for a longer explanation of this complicated subject).
The GAO report is far from conclusive, but appears broadly consistent with my own view, based on initial analysis of the report. First, it highlights that it is very hard to define and measure implicit subsidies, which suggests to me that we should be cautious about taking policy actions based on such measurements. The report principally relies for its quantitative conclusions on an extensive econometric model created by the GAO. They emphasize the difficulty of accurately specifying the complex model and therefore they ran 42 variants using different definitions of key variables. Using 2013 data, the most recent available, they found that 8 variations of the model showed a positive implicit subsidy that was statistically significant, 18 showed a negative implicit subsidy that was statistically significant, and 16 could not find a subsidy in either direction that was statistically different than zero. To say the least, this is not a compelling argument that there are substantial positive subsidies today.
Second, the value of implicit subsidies varies over time. They were significant prior to the financial crisis and became much more important during the crisis itself, as government support became explicit and the value of that backing shot up as credit concerns grew about the banks. Since then, they have fallen sharply due to a combination of: explicit policy against future taxpayer rescues; legislation and regulation to make it much less likely rescues would ever be needed and, if they were, to shift their cost away from taxpayers; and the increase in safety margins of capital and liquidity held by the banks reducing the probability these banks would need a rescue. These subsidies have fallen so far that it is difficult to prove they still exist, as the GAO findings noted above show. My strong intuition is that they are still there, but on a quite modest scale. After all, most market participants think it is still at least a little bit more likely that a very large bank would be rescued by the government, than that a small one would be, and that not all of the costs of that rescue would be recovered from the bank’s funders.
The GAO does not try to quantify the other important piece of the question about government policy – to what extent does legislation and regulation now add greater costs for the largest banks? There is explicit government policy now that puts burdens on the very largest banks that do not exist for all other banks, including some that are still quite large by most standards. The most obvious is that the mega-banks are required to fund themselves more from shareholders and less from debt and deposits, which are cheaper. (There is some academic argument that this balances out on a societal basis, due to higher tax collections, but no argument that the individual bank avoids higher costs.) There are some other notable costs for the largest banks, described in my primer. The costs may total as much as one quarter of a percentage point on assets annually — larger than most current estimates of the borrowing cost advantage. It is therefore distinctly possible that there is now a net disadvantage from government policy for the largest banks.
Overall, there does not appear to be a sound reason to change current government policy based on worries about implicit subsidies. The net subsidy may actually be negative now. If it is positive, it is now small and hard to measure. Government policy is already focused on avoiding future taxpayer rescues and we should continue down the path of implementing sound approaches to achieve this, such as the FDIC’s work on Single-Point of Entry approaches to dealing with large banks that go under.