Andrew P. Hofer
Made as Instructed
Are the "Stress Tests" really a Walk in the Park?
The Federal Reserve released the results of its Supervisory Capital Assessment Program (“SCAP” or, as it is more commonly known, the “Stress Test”) ten days ago. It is hard to imagine that policymakers view it as anything but a success. Bank stock and bond prices rallied through the announcement, and several institutions immediately took advantage of the comparatively frothy conditions to raise debt and equity capital. Ultimately, the definition of sufficient capital is a matter of market consensus, but there are reasons to be concerned that the additional capital required by the Stress Tests will not be sufficient to adequately support the nation's biggest banks through the next few years.
We are not envious of the constraints placed on policy makers as they devised the “Goldilocks” capital adequacy test ¾ neither too strong nor too weak. The public and Congress were in no mood to authorize additional TARP capital. The government needed to put market fears about the banks to rest while conserving as much of the estimated $110 billion in available TARP funds as possible. Furthermore, had the tests concluded that the Nation's banks were woefully undercapitalized, that would create renewed uncertainty and perhaps return markets to the dark days of last November. On the other hand, if the SCAP had been perceived as too easy, the market reaction could have been similarly negative. Predictably, the Federal Reserve shifted the capital goalposts around to insure the banks would have capital needs within an amount that could easily be raised. We outline some of the more forgiving aspects of the Stress Tests below:
Assuming a relatively modest downturn
The “adverse” economic scenario used to perform the tests was quite mild compared to the more bearish consensus forecasts. For instance, the “adverse” case suggests a lower unemployment rate than has already materialized over the last few months. Our “base” case is closer to the fed's adverse case, and there is precious little distinction between the Fed’s two cases. We certainly believe it is possible that the economy could follow an even weaker path, and that any “adverse” stress test ought to represent more extreme probabilities. The unemployment rate is critical to consumer loss estimates as it is highly correlated with credit losses on consumer loans. The Fed’s base case is constructed from the average of economist surveys, but the more adverse scenario is simply too close to that consensus, especially given developments in the months since these scenarios were initially conceived.
Giving substantial credit for future earnings
While adding another $600 billion in losses, the Stress Test gives immediate credit for future “pre-provision” earnings of over $350 billion, significantly lowering the external capital demand on these 19 institutions. This assumption seems generous in both amount and timing, since these profits have yet to materialize even as the associated loan losses would surely be realized immediately upon sale of any asset. It is also important to note that this level of profitability is contingent on the continuation of current market conditions, including a steep yield curve, high credit spreads and continued government purchases of mortgage-backed securities.
Mixing and matching capital ratios and risk-weighting
The Fed took the more conservative approach of focusing on “Common Capital”, exclusive of coupon-paying preferred shares. While this focus on loss-absorbing capital makes the test seem tougher, the Fed then set the target quite low. First, the test required only 4% capital, which gives allowable leverage of 25:1. This is actually quite high relative even to 1990s bank capitalization. Then, the Fed increased the available leverage further by using “Risk-Weighted Assets” as the denominator. Risk-Weighted Assets exclude part or all of government obligations, funds in other bank deposits, municipal bonds, and most distressingly (or de-stressingly) mortgages on 1–4 family homes (see Exhibit II). In the case of Citigroup, for instance, “Risk-Weighted Assets” are half of reported total assets. When the government requires 4% capitalization for the Stress Test, it is really requiring only 2% capitalization on Citibank's stated balance sheet. To give a sense of the range of possible outcomes here, consider Exhibit III, depicting institutional leverage under various definitions of tangible common equity (“TCE”) and Tier 1 Capital and comparing those to the SCAP.
As senior creditors, counterparties and depositors of these institutions, we do not see much value in moving money around between junior layers of capital. So we are puzzled and amused at the Fed's success in creating the impression that converting TARP preferred into Equity magically provides new resources. The Fed explanation argues from a true premise ¾ that common equity is more appropriate for loss absorption since it pays no coupon and has no fixed par value. Yet it simply does not logically follow that ‘systemic’ creditors are any more secure for the exercise. Furthermore, such an assertion calls into question both the original TARP investment and the old regulatory concept of Tier I capital, which included all of these layers of junior capital. This novel focus on common equity allowed the Fed to “double dip” on their TARP investments. There isn't a bank on the list that couldn't satisfy the SCAP shortfall simply by converting its existing TARP monies, which would have required no actual new investment. The Saturday Night Live parody is essentially correct: this was a “Pass/Pass” test.1
Large, but not unprecedented loss assumptions
Finally, and most seriously, we are concerned that the loan loss assumptions used by the Fed may not qualify as a true “Stress Test”. The first lien residential mortgage loss assumptions, which are fairly tightly grouped around 8%, seem low, especially with respect to institutions that offered (or purchased institutions that offered) so-called “affordability” products. While we would not expect losses in most bank portfolios to go as high as subprime mortgages, the Fed’s projections seem very low in light of system-wide prime mortgage delinquencies and foreclosures presented in Exhibit IV. The Fed’s assumptions for second lien losses have a median of 13%. With housing prices down as far as they are, very little second lien debt has equity underneath it, suggesting that performance in this category could be far worse.
We also have grave concerns about the amount of commercial real estate exposure on bank balance sheets. This is not only an issue for several of the first 19 banks, but for many smaller banks. The Fed’s loss assumptions for commercial real estate varied widely, from 2% to about 45%, depending on the institution, but the median was just over 10%. We foresee significant problems in this area, as several hundred billion dollars of commercial real estate loans on bank balance sheets will mature from 2010 to 2012, much of it underwritten on rent and price assumptions that no longer make economic sense. If these loans cannot be refinanced, commercial real estate values will decline even further in a glut of foreclosures. Lenders will count themselves lucky to experience losses under 10%. There is hope that Treasury Programs can be structured to stimulate purchase of real estate loans. Right now, however, we are not encouraged by PPIP's moribund state, and see real reluctance in the private sector to join in partnership with an administration that appears to be willing to flex its regulatory muscle in unpredictable ways.
We have similar concerns about the nearly one trillion dollars of leveraged corporate loans coming due in the next few years, although the restructuring of these loans is less likely to create the vicious price-delinquency cycle so familiar in real estate.
Our hats are off to the Fed. The Stress Tests managed to trace a fine line between being too transparently easy and being so tough that they raised new questions about banks. We are encouraged by the ease with which financial institutions are raising new capital from the market and how this enabled the Treasury to use some of its remaining TARP reserve to shore up several domestic life insurers. Last but not least, the rally in financial institution debt, up until the last few days, has been quite strong. However, we believe the “adverse” scenario mapped out in the Stress Test is actually quite likely, with unemployment trending higher than Fed projections as the commercial real estate and leveraged loan segments of this deleveraging crisis play out. We are concerned that banks may be less than amply capitalized when greater losses materialize. At that point, who will be asked to provide additional capital? What will be the public mood for further rescue of these institutions and those who finance them? Between the repayment or conversion of TARP investments and the various proposals for conservatorship of “systemically important institutions” (not just their regulated subsidiaries) on Capitol Hill, we see a government that is preparing to engage in wholesale restructuring of financial institution debt should the SCAP not be our last measure of capital adequacy for the Nation’s banks. This possibility introduces an element of uncertainty for senior creditors.
Throughout this year we have targeted underweight allocations to banking and finance credits, focusing instead on institutions that we believe can survive without government assistance. Given the rocky road ahead, and the necessary consolidation in the banking industry, we are maintaining this underweight. We find better value in the corporate and asset-backed sectors, where cash flows are more durable and the vagaries of government support and public opinion do not loom as large in our credit analysis.
This commentary also available in PDF format
