Are Large Banks Valued More Highly?

Alvaro Taboada is Assistant Professor of Finance at Mississippi State University, and René Stulz is Everett D. Reese Chair of Banking and Monetary Economics and the Director of the Dice Center for Research in Financial Economics at The Ohio State University Fisher College of Business. This post is based on a recent paper authored by Professor Taboada, Professor Stulz, and Bernadette A. Minton, Professor of Finance and Arthur E. Shepard Endowed Professor in Insurance at The Ohio State University Fisher College of Business.

In Are Large Banks Valued More Highly?, we investigate whether the value of large banks, defined as banks with assets in excess of the Dodd-Frank threshold for enhanced supervision ($50 billion in 2010 constant dollars), increases with the size of their assets, using Tobin’s q and market-to-book as our valuation measures. There is a widely-held view that banks gain from becoming large enough to obtain access to a stronger regulatory safety net in the form of “too-big-to-fail” (TBTF) subsidies. If this view holds, large banks should be valued more because they have an asset that other banks do not have, namely a claim on public resources, and the value of this asset grows as these banks become larger. Yet, this popular view ignores the possibility that large banks may bear larger costs than other banks because they are TBTF. For example, these costs may be in the form of greater regulatory scrutiny, political risk, or regulatory requirements that force them to pursue suboptimal policies. With these potential higher costs, the issue of whether TBTF banks are valued more highly is an empirical matter.

Using a variety of approaches, we find no evidence that large banks are valued more highly or that their valuations increase with size. We pay special attention to the sample period ending before the global financial crisis (GFC) since so many observers have argued that TBTF was an important contributing factor to the crisis. We find that the valuations of large banks are negatively related to their size from 1987 to 2006. The negative relation remains significant if we extend our sample to 2010, but when we add the post-Dodd-Frank years in the sample, the relation becomes weaker in the cross-section and does not hold when we look at within-bank increases in size.

In addition to the argument that the safety net gives banks incentives to become large, it is often argued that it gives them incentives to become riskier. If the riskiness of bank income increases with size for larger banks, so that banks become more likely to incur costs of financial distress as they grow larger, large banks would be valued less than small banks. There is no consistent evidence that bank risk increases with size for large banks. Both equity volatility and tail risk increase with bank size for small but not for large banks. As a result, large banks do not, on average, have higher equity volatility and tail risk than small banks.

Using a bank z-score as a proxy for a bank’s probability of distress, we find that distress risk increases with bank size for small banks but decreases for large banks. Yet, in contrast to these results, leverage is unrelated to size for small banks but increases with bank size for large banks. Further, the systematic risk of banks, as measured by a market model beta, increases with bank size for small banks as well as for large banks. In general, one would expect that if two banks are identical except that one has higher leverage, the bank with higher leverage has higher stock return volatility and distress risk. The fact that tail risk and equity volatility do not increase with size for large banks and that distress risk decreases with size for large banks is consistent with a negative relation between asset risk and bank size for these banks.

The results relating Tobin’s q and market-to-book to bank size are inconsistent with the hypothesis that large banks are valued like other banks except for a premium resulting from their enhanced access to the safety net because of TBTF. So, why do values of large banks fall as their size increases? We investigate several potential explanations.

If performance falls with size for large banks so that the present value of cash flows as a fraction of assets falls with size for these banks, larger banks would have lower valuation measures than small banks. This explanation does not have support in the cross-section of banks when performance is measured by return on assets (ROA) or return on equity (ROE). Tobin’s q also should be lower for large banks if their cash flows are expected to grow less than those of small banks. There is no evidence that ROA or ROE growth falls with bank size for large banks.

In valuation formulas, bank cash flows are discounted at the bank’s cost of capital. Hence, if the cost of capital were higher for large banks than for small banks, large banks would have lower valuations. To address this issue, we examine whether large banks have different stock returns from small banks by comparing stock returns and risk-adjusted stock returns. There is no evidence that large banks have stock returns or risk-adjusted stock returns significantly different from small banks.

Large banks engage in different activities from smaller banks. In particular, non-interest income increases with asset size. Taking into account non-interest income increases the negative relation between bank Tobin’s q and size for large banks. Specifically, activities associated with non-interest income appear to enhance bank value, but not as much for large banks.

One of the most important balance sheet differences between large and small banks is that large banks typically have a sizeable portfolio of trading assets. There is strong evidence that the value of banks is negatively related to trading assets. A bank that increases trading assets by 1% of assets and reduces its holding of securities by 1% (in other words, reduces securities held for investment and increases securities held for trading) is associated with a decrease in Tobin’s q of 0.19%. Importantly, trading assets help explain the relation between Tobin’s q and bank size before the GFC and hence this relation is not the product of the crisis. However, the ability of trading assets to explain the relation between Tobin’s q and size for large banks is sensitive to regression specification.

Following the GFC, regulations for banks changed substantially, mostly because of Basel III and Dodd-Frank. We do not find evidence that the relation between bank Tobin’s q and size is significantly different during that period. However, Tobin’s q is related to some bank characteristics differently during that period. Specifically, more capitalized banks are valued less and the market appears to discount non-interest income relative to 1987-2010.

An important caveat is that while we show that TBTF banks are not worth more than small banks because of TBTF, we do not demonstrate that TBTF does not add value to large banks, since we cannot exclude that large banks would be valued even less without TBTF. However, similarly, we cannot exclude that TBTF decreases bank value through regulatory and political costs.

The full paper is available for download here.

Both comments and trackbacks are currently closed.