Sections

Research

Stress Test Loan Losses and Profit Expectations: A Comparison

Douglas J. Elliott
Douglas J. Elliott Former Brookings Expert, Partner - Oliver Wyman

May 8, 2009

The stress test results are best understood by comparing them with other detailed analyses of the financial state and prospects of the banks. My earlier piece, “Interpreting the Bank Stress Test Results,”, provided a detailed analysis of the implications for the potential stress test results of estimates from two other credible sources: the International Monetary Fund (IMF) and NYU’s Nouriel Roubini, widely viewed as the most pessimistic of those who have provided thorough analyses of potential credit losses. This paper provides a partial update, comparing the actual stress test results for credit losses and expected profits with estimates derived from the IMF and Roubini analyses.

Table 1 shows the key comparisons for the 19 banks, in aggregate, relevant to the net reduction in capital expected by these banks through 2010.

Table 1: Expected loan and securities losses and expected profitability, ($ billions)

Regulators

IMF

Roubini

System-wide loan and securities losses for 2009 and 2010

NA

550

1250

Portion of system-wide losses absorbed by the 19 banks1

NA

70%

70%

19-bank loan and securities losses for 2009 and 2010

599

385

875

Portion already in capital figures (purchase acc’ting. adjust.)2

64

64

64

Net loan losses

535

321

811

System-wide bank earnings in 2009 and 2010, after dividends

NA

300

300

Portion of system-wide profits from 19 banks3

NA

70%

70%

19-bank earnings including loan provisioning for 2011 losses

363

210

210

Reduction in capital through Dec 2010

172

111

601

Planned capital actions and excess Q1/2009 earnings4

110

110

110

Net reduction in capital through Dec 2010

62

1

491

1. Author estimate, based on rough proportion of system-wide assets owned by the 19 banks
2. From reported stress test results. Reflects U.S. accounting rules for loan losses in bank acquisitions
3. Author estimate, based on rough proportion of system-wide assets owned by the 19 banks.
4. From reported stress test results. Represents higher Q1/2009 profits than implied by full year estimate built into 2009 and 2010 profit figures, plus capital raising already committed or executed.

The overall pattern is clear: the reported stress test results incorporate significantly more conservative loss assumptions than does the IMF, but most of this relative conservatism is offset by an assumption of stronger bank earnings. Not surprisingly, the stress tests are significantly less conservative on both loan losses and expected profits than the estimates of Dr. Roubini. Of course, the real stress test will be comfortably surviving 2009 and 2010. We will only know the true capital needs after the fact.