Watching the de-leveraging process on Wall Street continue, most recently with the torment of Lehman Brothers (LEH), we are often reminded of the statement by author Martin Mayer that there are no economies of scale in banking. Mayer, who we feature in the next issue of The Institutional Risk Analyst, was talking mostly about commercial banking when he made that statement, but the investment banks fit into that shoe as well and seem to be doing an even worse job than commercial banks when it comes to achieving real profitability.
Go back and review the Sell Side research focused on financial names published over the past 18 months. Note the change in the tone of the opinions and, in particular, the volatility of earnings forecasts for the major universal banks, then and now. Almost seems like different industries. The growing visibility of the true risk profile of many financial institutions raises a fundamental question: Do banks and brokers really ever make money, especially in risk-adjusted terms? Or to put it another way, is the combination of technology and deregulation -- aka "innovation" -- in the financial markets creating or destroying value for investors in financial institutions of all descriptions?
Earlier this year, a client asked us to extract the risk-adjusted return on capital or "RAROC" metrics for the top 100 US banks from The IRA Bank Monitor going back some twenty years and compare the same with a range of financial statement and market indicators. One of the resulting charts from this effort suggests very strongly that the real, risk-adjusted profitability of the US banking industry has been declining since the mid-1990s, when securitization and other types of "innovation" really kicked into gear.
Notice that even during the period of "supra normal" asset and equity returns reported by US banks during 2004-2007, the average RAROC for the top-100 banks was trending lower. The Q1 2008 results only confirm this trend, especially for some of the largest US banks. The latest RAROCs and other Economic Capital components are available to users of The IRA Bank Monitor and the Compendium of Bank Risk.
Perhaps more puzzling, during this period the standard deviation among the 100 largest, financially sophisticated banks was also falling, suggesting that diversity among the group grew less and thus the risk profiles were likely converging. Notice the huge reduction in the SD for the group over the past decade.
At first we worried that the sharp decline in RAROCs displayed in the early 1990s was due to statistical anomalies, but The IRA asked Joseph Mason, Herman Moyse, Jr./Louisiana Bankers Association Professor of Finance, Louisiana State University, to review the results. His reply was that the falling RAROCs may be attributable to three factors -- deregulation, litigation and securitization.
The IRA: Joe, doesn't the sharp drop in RAROC in the early 1990s look like a data inconsistency? Dennis lined up the FDIC data labels perfectly, but still the change seems drastic. What do you think?
Mason: You are correct to be wary of the early vintage FDIC data, but that period also was one of great change for the banking industry. Your focus on the top 100 banks probably avoids some of the S&L crisis era capital distortions in the FDIC data that might affect RAROC, so on balance I think that the chart does tell that story.
The IRA: Do tell.
Mason: Starting in the early 1990s, you had a series of court decisions, legislative initiatives and an increase in OBS securitization activity, events which all contributed to changes in bank business models and risk profiles. From 1989 to 1994, we saw some baby steps in terms of bank securitization, senior subordinated structures and the like. By 1995 when I joined the OCC, the volumes were rising and we started to be concerned about all of the implicit recourse events. You see RAROCs fall until 1997 with the 120% LTV crisis. Then you get the turning point in 1999 when 120% LTV dies in the marketplace and you set the stage, as the chart suggests, for the run-up to the home equity loan crisis.
The IRA: And just by coincidence HELOCs are in the headlines again today. So you would agree with the chart that the real, risk-adjusted returns for the banking industry have been going down for the institutions most affected by "innovations" like securitization and OTC market structures? Is this a function of the economy or specific market developments like securitization?
Mason: Both. The economy is important, but I would say that the effects of securitization are probably going to drive your SD and RAROC.
The IRA: But despite the promise of "innovation," the results suggest that the industry is working for less and less real return, a trend which if true implies that banks are really poor investments, especially if regulatory demands for higher capital levels come to fruition.
Mason: The bottom line is competition. In the 1990s, we not only deregulated banks but also deregulated funds, for example, and allowed them to enter the low-risk areas of the banking business. Sometimes new players just entered the traditional banking market without need of deregulation like money market funds. So banks had increased competition and the result was lower profits.
The IRA: This reminds us of our discussion earlier this year with Bob Feinberg ("GSE Nation: Interview with Robert Feinberg," March 17, 2008), who argues that the banking industry model is broken. Indeed, you yourself have suggested that the banking industry never really recovered from the 1930s and was just carried along by the war mobilization. As Feinberg and Alex Pollock at AEI both would say, the banks all became GSEs after the Depression and WWII -- just competitive ones!
Mason: Correct. If we really do want a safe and sound banking sector, we must start to pay attention to issues like the effect of securitization on bank business model choices. The group you have chosen in this study emphasizes the securitization effect by excluding smaller banks, but it nevertheless tells the story.
The IRA: Thanks Joe.
Keep in mind that these RAROC calculations we illustrate in detail within the current version of The IRA Bank Monitor don't fully encompass the OBS risk which these banks have securitized over the past half decade or more, risk that is now likely to come back on-balance sheet during 2008 as the FASB finally does away with QSPEs. Add the full weight of existing OBS exposures back onto these bank Economic Capital profiles, a task IRA is currently completing as part of a larger credit conditions indicator project, and the overall RAROCs are likely be even lower than our preliminary work indicates.
What does this suggest for universal banking in the US as a viable business model?
To add further to the calculus, when you consider the belated calls by US regulators to increase capital held by this very same group of large, sophisticated banks, the prospective RAROC going forward falls off the proverbial table. If JPMorgan Chase (JPM) reported a single-digit RAROC in 2007, a number that moved significantly lower in Q1 2008, what would it be with say 2x current regulatory capital? Would investors even be willing to provide this much capital to Jamie Dimon and his CEO peers at other moneycenter banks?
Interesting, is it not, that the Federal Reserve and other US regulators now seem to agree with the fundamentals-based bank EC measures that IRA first published three years ago, measures which suggest that many of the largest US banks require substantially more capital to support their true economic risks (See "Rogue Traders and Economic Capital," Janury 28, 2008)? The regulatory community has not yet come to our view of, say, requiring 5x Tier One Risk-Based Capital for JPM, for example, but check back with us at the end of 2008.
The bottom line is that when investors look at financial stocks and particularly the largest US banks, the question must be asked: Are any of these business models sufficiently robust and stable to warrant a place in any portfolio allocation, especially a portfolio that requires investment-grade assets? Our answer is this: We test bank safety and soundness at 1,000bp of aggregate loss or the low end of a "BB" rating.
The Q1 2008 RAROC and EC results for the top 100 US banks generated by IRA suggest that many of these institutions do not meet that investment grade test, regardless of the nominal external default ratings they may have obtained from conflicted Sell Side ratings agencies. What does such an assertion suggest for the target internal ratings methodologies within Basel II, rules which global regulators claim will enhance bank safety and soundness?