Jean-Paul NICOLAÏ, Directeur Général, OTC Conseil
Fanny LIONET, Chief Operating Officer, OTC Conseil Americas

Only the future will tell us whether the stress test exercise that the banking sector has undertaken will prove to be relevant. Some commentators have called into question the validity of the hypotheses and the way banks and authorities are said to have negotiated the results, even to the point of predicting the apocalypse. Taking a look at the hypotheses, the results of the tests, and at the way expectations make the world go around, maybe, at the end of the day, the European strategy – not revealing the results of the stress tests – might be a good one.

The US exercise
What if...

The recent past has demonstrated that many banks do not have sufficient financial backing to survive exceptional market conditions. But how do we know how strong the next shock will be? And supposing it happens tomorrow, then how will banks survive it?

Targets
In order to ensure that banks have adequate capital to face adverse conditions without running into insolvency and bankruptcy – and risk triggering another systemic crisis – the Federal Reserve ran a stress test analysis, based on the Supervisory Capital Assessment Program (SCAP).

The aim of the exercise was to measure risk, and to get banks to fund their potential capital needs, in a period when their nationalization was clearly a subject of controversy.

The SCAP is built to estimate Profit and Loss statements and reserve needs under adverse scenarios for eligible US bank holding companies (BHCs) with assets exceeding $100 billion dollars. These stress tests allow us to ascertain whether the banks under scrutiny are able to fulfill their vital role in the economy.

Treasury plan
Banks have six months from May 7, 2009 to raise private capital or to benefit from the Capital Assistance Program (CAP) of the US Treasury.
The capital granted through the CAP will be preferential securities convertible into ordinary shares with a 10% discount off of the February 9, 2009 reference price. These preferential securities will pay a yearly coupon of 9% and will be convertible at the initiative of the issuer, with the consent of the regulating authority.
After a period of seven years, these securities will be automatically converted into ordinary shares, if they have not already been reimbursed or converted. No upper limit was defined concerning the amount that the Treasury might grant.

Within the framework of the Capital Purchase Program (CPP) and with the agreement of the regulating authority, banks will be allowed to ask for capital in preferential securities, in addition to previously received capital. A bank could exchange securities from its CPP into new CAP securities. CAP beneficiary banks will be subjugated to restrictions in terms of executive remuneration, dividends and acquisitions. They will also have to specify the foreseeable usage of their granted capital and of their credit activity volume.

Hypotheses
The potential losses were calculated based on a bank’s assets, including those related to mortgages and contingent liabilities, over a two-year period, starting from 2009. Two economic stress scenarios were applied to the banks: a baseline scenario and a more pessimistic one (c.f. table 1).

Economic Scenarios
- The baseline scenario is based on the following hypothesis: a 2% contraction of the US economy in 2009, followed by growth of 2.1% in 2010. The considered unemployment rate would be 8.4% in 2009 and at 8.8% next year, and a drop in real estate prices by 14% in 2009 and a further 4% in 2010.
- The more adverse scenario anticipates an economic contraction of 3.3% in 2009, with only 0.5% growth in 2010, an unemployment rate of 8.9% this year and 10.3% in 2010, and a drop in real estate of 22% in 2009 and 7% the following year.

Source: The Supervisory Capital Assessment Program: Design and Implementation - Board of Governors of the Federal Reserve System - April 24, 2009


For both scenarios, a certain default rate is applied depending on loan categories, with higher estimates projected in the more adverse scenario (c.f. table 2).

Source: The Supervisory Capital Assessment Program: Design and Implementation - Board of Governors of the Federal Reserve System - May 7, 2009


Results
Focus on the worst-case scenario

The SCAP stress tests propose an estimated $600 billion loss for the 2009-2010 period under the more adverse scenario. The lion’s share of the loss was generated by accrual loan portfolios, for which the estimated loss is 9.1% of the assets of 19 banks over the two-year period. Trading and investment losses in the SCAP were estimated up to $135 billion. Added to those already recognized as of mid-2007, we would then arrive at nearly $950 billion in total losses for the 19 banks by the end of 2010.

Once the overall stress test calculation is applied, considering losses, revenues and reserve requirements, the 19 BHC will need to add $185 billion to capital buffers, in order be compliant with the SCAP target capital buffer requirements at the end of 2010.

This capital buffer estimate was calculated as of the end of 2008. However, since then a number of the 19 banks have already restructured their capital or assets that, quite often, have increased their Tier1 Capital.
In addition, the preprovision net revenues of many of the firms exceeded what was assumed in the more adverse scenario by billions, allowing them to build their capital bases. The effects of this rendered the additional capital necessary to establish the SCAP buffer equal to some $75 billion over a two-year period, as shown in the table below.

Source: The Supervisory Capital Assessment Program: Design and Implementation - Board of Governors of the Federal Reserve System - May 7, 2009

Criticisms of the Hypotheses
For his part, Federal Reserve Chairman Ben Bernanke has defended the relevance of the tests. He pointed out the high level of loss rates on loans – which are even higher now than during the Great Depression. Despite Bernanke’s support, and even if the man in charge of the “New New Deal,” Larry Summers, the Director of the White House's National Economic Council,” is a great figure in the world of economists, it is quite difficult to consider such hypotheses without any concern. The first major review sheds light on the hypothesis of economic growth – operating under the assumption of a growth rate of - 2% in 2009 seems quite strange when, for example, the OECD forecasts -4%. As an adverse scenario, -3.3% economic growth is a relatively benign hypothesis. The same arguments hold for the unemployment rate and for real estate price depreciation. In fact, the hypotheses underlying the stress test exercise have come under severe criticism: “All the ingredients they have so far are weak, and there are several missing ingredients,” Joseph Stiglitz said in a recent interview. The people who designed the plans are “either in the pocket of the banks or they're incompetent.” New York University Economics Professor Nouriel Roubini had similarly tough words about the process, as reported by Forbes Magazine: “Stress test results are not worth the paper they're written on.” While both of these world-renowned economists are also well-known mavericks, their criticisms are fair points.

Criticisms of the “Bargain”
According to The Wall Street Journal, the Fed has significantly scaled back the size of additional capital needs, following hard negotiations with the banks. This has led to speculation of a secret bargain between banks and the government, and calls into question the relevance of the exercise, one of the foremost ambitions of which was precisely to dispel some of the uncertainty that hangs over American banks. The amounts renegotiated are indeed significant. For instance, the additional capital needed for Bank of America, initially around $50 billion, is now around $35 billion in the final result.

However, these types of adjusted figures are par for the course. Dialogue between regulators and the regulated is a normal step in any audit and prudential assessment process. Supervisors have to allow the supervised the chance to respond before establishing the final terms. Intimating that there was a “secret bargain” in play might be a fallacy, provided that the arguments deployed by the banks in the process are acceptable and transparent. Largely, the problem boils down information asymmetries in a classic principal-agent problem context. For instance, cost-cutting or revenue forecasts are likely to be more accurately and forensically evaluated and projected by each individual bank than by the outside regulator.

Criticisms of the Perimeter
We are facing a thornier problem of transparency, driven by not insignificant oversight on the part of the regulator to make a careful assessment - especially regarding the toxic financial products banks have in their books. Clearly, as Paul Krugman pointed out, the stress test exercise was not a rigorous audit.

Another major liability regarding the perimeter is the lack of a systemic risk component within the stress test set. Since the crisis broke, all the regulation research has turned to the macro-prudential dimension, as a driver of systemic risk. In this context, what about the “non-major” US banks and their position in the stress test exercise? For example, as bankruptcies increase in number, how pertinent and realistic are scenarios from which the savings and loan associations and the regional banks are absent? In addition, the systemic component of the crisis is not limited to the domestic dimension. Also, local banks with significant risk exposure to commercial real estate market are, as counterparties to major banks, “another brick in the wall.”