14 March 2014

VoxEU: Macro-prudential stress tests should not rely on regulatory risk weights

This column argues that macro stress tests would be more effective if capital requirements were measured differently from the current regulatory risk weight-based approach, and in particular, were based on total assets and on market risks.

Authors: Viral Acharya, Robert Engle, Diane Pierret

The capital ratio of a bank is usually defined as the ratio of a measure of its equity to a measure of its assets. Regulatory capital ratio usually employs book value of equity and risk-weighted assets, where individual asset holdings are multiplied by corresponding regulatory ‘risk weights’. Macro-prudential stress tests rely on models that translate an adverse macro-economic scenario into losses to assets on the balance sheet of banks. These losses are assumed to be first borne by equity. The resulting capital ratios determine which banks fail the test under the stress scenario, and what supervisory or recapitalisation actions are undertaken to address this failure.

Recent concerns on the denominator of capital ratios – the risk-weighted assets – have been expressed in multiple surveys that point out the inconsistency in the calibration of risk weights (BCBS 2013, Haldane 2012). This column argues that the inadequacy of risk-weighted assets is also responsible for producing an inadequate ranking of the required capitalisation of banks in stress tests. 

Overall, the authors' findings suggest that stress tests would be more effective if capital requirements were measured differently. A risk-based capital requirement is not sufficient because:

• The increase of risk over time is not captured adequately with static risk weights; and
• risk weights are flawed measures of bank risks cross-sectionally, as banks game their risk-weighted assets (cherry-pick on risky but low risk-weight assets) to meet regulatory capital requirements without necessarily reducing economic leverage.

The authors recommend that regulatory stress tests complement their assessment of bank and system risks by using the simple leverage ratio-based and market-based measures of risk. They welcome the new Basel III Tier 1 leverage ratio, but this has not yet been incorporated into the European stress test design. The mis-measurement of bank risks is likely to be present in future stress tests as long as the reliance on static regulatory risk weights prevails. In particular, the upcoming Asset Quality Review (AQR) to be conducted by the European Central Bank as the Single Supervisory Mechanism of the eurozone Banking Union will involve a stress test to assess capital shortfalls. The regulators conducting this exercise should be wary of relying only on the current scheme of static risk weights, and instead, also employ the simple leverage ratio and market-based capital shortfall measures, such as SRISK

Full article

© VoxEU.org