Stress Tests in a graphic

One of the differences between the way the US and European authorities responded to the financial crisis was in the rigor of their respective stress tests. The American stress tests credibly signaled the health of financial institutions and thereby stemmed the spread of the panic, whereas that in Europe did little to allay market suspicions and fears.

This graphic nicely captures how these tests restored normalcy to the credit markets.

Once the test results were announced, the TED spread (indicates the perceived risks in lending to banks), CP spread (a similar indicator for business), and the Baa spread (indicating perceptions of corporate risk) all fell precipitously. 

Uncertainty and balance sheet in credit crises

In an NBER working paper, Arvind Krishnamurthy of Kellogg School draws attention to two amplification channels that operate during liquidity crises - balance sheets and uncertainty. He writes,

"The first mechanism works through asset prices and balance sheets. A negative shock to the balance sheets of asset-holders causes them to liquidate assets, lowering prices, further deteriorating balance sheets, culminating in a crisis. The second mechanism involves investors' Knightian uncertainty. Unusual shocks to untested financial innovations lead agents to become uncertain about their investments causing them to disengage from markets and increase their demand for liquidity. This behavior leads to a loss of liquidity and a crisis."

He writes that the balance sheet mechanism creates an endogenous source of liquidation shocks that depends on the leverage of agents. Also, uncertainty inhibits the process of price discovery, rendering market prices uninformative about fundamental value, and thereby triggering off a positive feedback loop on uncertainty. Further, as in the case of the sub-prime mortgage crisis, the downward spiral in asset prices started with uncertainty about counter party risks (due in part to the complexity of credit market instruments) which soon became evident on the balance sheets as a balance-sheet/asset-price feedback was set in motion. Accounting rules like Mark-to-Market (information revealed from which are difficult to interpret in an environment where the price discovery mechanism is impaired) is an important mechanism through which balance sheet amplifiers are triggered off.

About recovery from an uncertainty-driven crisis, he writes that it is

"only resolved over time as investors understand where they went wrong and formulate new models of the world; in short, as the uncertainty is resolved. Part of this process involves information revelation. What mistakes have investors made? Which investors have large exposures to the relevant assets, and how big are their losses? In an environment where the price discovery mechanism is impaired, information revealed from accounting statements is hard to interpret. Further, in an environment where balance sheets are weak, a financial institution may be reluctant to realize losses, impeding information revelation. These forces tend to perpetuate an uncertain environment, which may be one factor behind the duration of the subprime crisis. Indeed, from this perspective, a benefit of the stress-tests performed by the Treasury in the current crisis is that they force information revelation and reduce uncertainty."

Update 1
Mostly Economics has more on "how changes in asset prices are transmitted internationally through their effects on the balance sheets of highly leveraged financial institutions" and points to two papers.

First, Michael B. Devereux of the University of British Columbia, has a paper that develops a "model of the international transmission of shocks through de-leveraging across financial institutions". He writes, "In a macro-economic model in which highly levered investors hold interconnected portfolios across countries, we show that the presence of binding leverage constraints introduces a powerful financial transmission channel which results in a high correlation among macroeconomic aggregates during business cycle downturns, quite independent of the size of international trade linkages."

Second, Paul Krugman was one of the first to point to an international finance multiplier, different from the conventional foreign trade multiplier, that palys an increasingly important role in linking markets together and enabling the transmission of global economic shocks. He defines an international finance multiplier, as that "in which changes in asset prices are transmitted internationally through their effects on the balance sheets of highly leveraged financial institutions". The multiplier effect is transmitted through, what Kaminsky, Reinhart, and Vegh describes as the "leveraged common creditor". Prof Krugman therefore argues that since all economies now share leveraged common creditors, balance sheet contagion has become pervasive.

I have blogged about this in an early post here.

Stress tests results finally released

After subjecting 19 of America's biggest banks to the most public scrutiny in decades, in the form of stress tests to reveal their ability to withstand further economic crisis and losses, federal regulators ordered 10 of them to raise a total of $75 billion in extra capital and gave the rest a clean bill of health.

The results of the stress tests (pdf here) comes as a relief to the financial markets given that the banks that came up short will have to raise far less than some investors had feared. Among the biggger banks, Bank of America will have to raise $33.9 bn, Wells Fargo $13.7 bn, Citigroup $5.5 bn, and GMAC $11.5 bn. The details of the additional capital required, potential losses and the area of these losses, for each of the 19 banks is available here.





The stress tests (full paper here) were tailored to test the capital adequacy of these banks - would banks have enough capital if the recession proves deeper and loan losses larger than the current consensus expectation. The banks are being given two numbers, the total capital they need and how much has to be in the form of common stock, the strongest form of capital. The tests have been made under the assumption that unemployment touches 10.3%, home prices fall 22% this year, and the economy contracts 3.3% this year and does not grow next year. Under this worst case scenario, the results estimate that the losses by the 19 banks could total $600 billion this year and next, or 9.1% of the banks’ total loans.

The additional equity will have to be either raised privately (either by selling shares to the public or a big investor, or by selling some of their businesses, or by conversion of some of the privately held preferred stock to common stock) or from the government (by fresh equity injections or conversion of the existing preferred shares into common stock). The banks are eager to avoid having the government increase its stake drastically because that would dilute the holdings of the banks’ existing shareholders.

The stress tests have been criticised for making rosy assumptions and were effectively designed to make everyone pass. Further, regulators gave the banks a break by letting them bolster their capital with unusually strong first-quarter profits and also by letting them predict modest profits even if the economy again turns down. Nouriel Roubini and Mathew Richardson have argued that the stress tests are only posponing the inevitable nationalization. This debate has been covered here and here. The NYT and Slate debates the stress tests here and here. More on the tests here and here. FAQ on the stress tests available here.

As Simon Johnson points out, the stress tests are a smoke-screen to let the too-big-to-fail banks "earn their way back to solvency through exercising their greater market power (Lehman and Bear Stearns are gone), government-subsidized debt (courtesy of the Federal Deposit Insurance Corporation), and various forms of implicit subsidy (through “legacy” loan removal programs)".

Update 1
The Cleveland Fed analyses the baseline and adverse case scenarios of the stress tests. In the baseline case, real GDP falls by 2.0 percent in 2009 before rebounding to 2.1 percent in 2010; the unemployment rate averages 8.4 percent in 2009 and 8.8 percent in 2010. House prices decline 14.0 percent in 2009 and fall an additional 4.0 percent in 2010. The more adverse (but not necessarily “worst-case” scenario) assumes a sharp 3.3 percent real GDP contraction in 2009 followed by scant 0.5 percent growth in 2010; the unemployment rate averages 8.9 percent in 2009 and 10.3 percent in 2010. House prices drop 22.0 percent in 2009 and 7.0 percent in 2010.

Update 2 (23/7/2010)
In an effort to reassure the markets and identify the banks with credit problems, 91 European banks were subjected to stress tests, in which all but 7 passed. To pass the tests, a bank’s Tier 1 capital, a measure of reserves, could not fall below 6 percent of assets in the face of a new recession and a sovereign debt crisis. The tests were the most extensive conducted in Europe, covering 65 percent of the total banking market and 20 countries from Ireland to Poland.

Stress test results

The much awaited stress results, not publicly announced but shared privately with the 19 banks that underwent the tests, have turned out to a damp squib as expected. The Fed announced that though the banks had enough capital to offset a raft of new losses, it would still need a new cushion of financing on top of the current minimum levels as a buffer against higher losses if the economy worsened. This strengthens the widely held belief that the government would support the largest banks even if their financial health eroded and that the banks may be forced to dilute their common stock in exchange for further equity infusions from the government.

There are also indications that the events till date may have widened the gap between the stronger set of banks led by the major investment and custodial banks like Goldman, Morgan Stanley, JP Morgan Chase, US Bancorp etc and the embattled majors like Citigroup, Bank of America, and Wells Fargo and the regional banks. However, the financial sleight of hand that dressed up the better than expected last quarter results among some of the banking majors, means that even those perceived as being better positioned may only be illusory.

The Fed also released the details of the parameters used in the stress tests, available here (and here).

Stress test charade?

The NYT reports that after labouring through records of nineteen of the largest banks for the past eight weeks in the name of "stress tests" under the Geithner Plan, nearly 200 federal examiners appear to have come to an interesting conclusion - the banks are in a better shape than what everyone thinks, but despite all the bailout money pumped in, they will still need further bailing out! This raises the question - if all the banks have passed the stress tests, but will still need to be bailed out, then what was the purpose of these tests?

The tests, led by the Federal Reserve, rely on a series of computer-generated "what-if" projections in the event the economy deteriorates, so as to judge all the financial institutions against a common set of standards. Those include unemployment rising to 10.3% by next year, home prices falling an additional 22% this year, and the economy contracting by 3.3% this year and staying flat in 2010. Critics have questioned these standards as all being too optimistic. As part of the tests, the banks analyzed each category of loans they held and compared their results with the "high" and "low" range of government loss estimates. The banks were also asked to project their earnings over the next two years to give the regulators a better sense of how much capital they would have to absorb the coming losses.

Since the tests are being conducted outside public view with little or no disclosure, there are serious dounts being expressed about the rigour of these tests. The credibility of stress tests will be open to question if all those exposed to it pass the test. The Geithner Plan had hoped that the stress tests would provide a "seal of good book-keeping" to bridge the information asymmetry and encourage investors to lend to stronger banks, thereby enabling them to raise capital for their regular activities and in the process unfreeze the deadlocked credit markets.

Naked Capitalism has this scathing indictment of the stress tests, describing them as a "charade" and a "complete sham". In the absence of transparency about the stress test details and the controversy surrounding the use of tangible common equity as the measure of capital adequacy ratio, some others have called the TARP stress tests asinine.

The Geithner Treaury will argue that these tests give the government adequate information to assess the extent of problems and formulate specific bailout plans for each institution, instead throwing money into dark holes, as was hitherto being done.