Further analysis of the stress tests in Financial Stability 2/11.

• Author: Johansen, Ronnaug Melle
• Position: Report

The purpose of Norges Bank's stress tests is to test the resilience of the Norwegian banking sector to negative events of low probability. Negative shocks that rarely occur at the same time may, for example, be combined. A stress test does not provide a set answer to the question of how the banking system will handle a severe crisis. Should banks risk falling below the minimum capital adequacy requirement, they can respond by raising capital or limiting lending. The stress test functions as a useful illustration of how important risk factors could influence bank earnings and loan losses.

The authorities conduct stress testing in two different ways. (2) In one approach, stress testing is conducted by financial institutions based on a macro scenario specified by the authorities. This method is known as the "bottom-up" approach, referring to the way the test focuses on how the macro scenario affects risk in each of a bank's exposures and then aggregates the overall impact on banks' profits and capital adequacy. Another approach, often used by central banks, is the "top-down" approach, also referred to as macro stress testing. This is the approach used by Norges Bank (3). The macro stress test in Financial Stability 2/11 (FS 2/11) is based on the results from the macro model in Norges Bank's suite of models for stress-testing. Projections of high-risk debt to the household and corporate sectors are used to calculate loan losses given each bank's aggregate lending to the different sectors. An overall assessment of the bank's profit and loss account and balance sheet is also conducted to provide a basis for assessing the effect of an adverse scenario on banks' capital adequacy. The macro stress test ensures a consistent assessment of credit risk in the banking system by applying the same method of loan loss calculation in all banks.

Norges Bank does not publish data on individual banks. The purpose of Norges Bank's stress tests is to test the resilience of the Norwegian banking sector to highly negative events. Our analyses are based on aggregated information on the composition of banks' balance sheets. For example, Norges bank cannot assess the risk related to individual banks' exposures. Furthermore, if individual banks' results were to be published, a closer dialogue with banks would have to be maintained throughout the process. The results must therefore not be interpreted as stress tests of individual banks.

The stress tests in Financial Stability 2/11

In FS 2/11, Norges Bank conducted a stress test based on a severe international downturn. Growth among Norway's trading partners falls sharply and is assumed to be most pronounced in Europe. This has a substantial impact on the Norwegian economy as about 70 per cent of Norwegian exports go to Europe. The oil price falls to below USD 50, resulting in lower investment in Norway. Because of widespread turbulence in the global economy, it is assumed that traditional "safe havens" are perceived as less safe than during previous turbulent periods. Only a moderate depreciation of the krone is therefore assumed, despite the steep fall in the oil price. This is a typical example of a stress test where shocks that rarely occur at the same time are combined. In addition, problems in the European banking sector are assumed to result in an increase in credit market premiums and higher bank funding costs. Credit growth declines both as a result of lower demand for credit and because banks restrict lending.

To illustrate the effect on both the economy and on banks' risk-weighted assets of the possibility that banks may have to tighten credit standards in a downturn, an alternative adverse scenario was constructed where the decline in credit growth is less pronounced. In this alternative scenario, growth in the economy picks up again more quickly than in the main adverse scenario. The alternative scenario builds on two important assumptions. First, it must be assumed that there is demand for credit, which will be uncertain in a situation involving a sharp decline in the global economy and a fall in oil prices. Second, banks maintain normal lending standards. A bank that is in danger of falling below the regulatory minimum capital adequacy requirement is not likely to maintain normal credit standards. How long it will take before a bank is forced to limit its lending in a period of contraction partly depends on how strongly capitalised it is in the first place. In the model calculations presented in FS 2/11, some banks fell below the minimum requirement in order to maintain lending volume. It is therefore likely that these banks would have responded by reducing lending. The result cannot therefore be interpreted to indicate that some banks failed the stress test. This illustrates that it is important from an economic point of view for banks' to be in a strong enough capital position to withstand a severe downturn without having to tighten lending.

The probability that a course of events such as in the adverse scenario would actually occur is low. Nonetheless, banks as a whole maintain capital levels above the regulatory minimum requirement in the main adverse scenario. The stress tests show that due to the increase in capital adequacy ratios since 2009 the Norwegian banking sector is better equipped to weather a severe international downturn.

Banks' Tier 1 capital ratios

A bank's Tier 1 capital ratio is an important measure of its solidity. The Tier 1 capital ratio mainly reflects the ratio of a bank's equity capital to risk-weighted assets. (4) The bank's total risk is reflected in the denominator, which is the sum of risk-weighted assets. The regulatory minimum requirement for the Tier 1 capital ratio is currently 4 per cent, with a proposed increase to 6 per cent under Basel III. In connection with measures to increase the resilience of the EU banking sector to future losses, the European Banking Authority (EBA) has proposed a temporary increase in the minimum Core Tier 1 capital ratio (5) to 9 per cent by the end of June 2012 for a number of European banks.

Banks' Tier 1 capital ratios--IRB models for calculating risk weights

Following the introduction of Basel II, banks may seek approval to use internal models to estimate loan portfolio credit risk. This is called the internal ratings-based (IRB) approach. The purpose of this approach is to align the sum of risk-weighted assets more closely with the actual risk profile of banks' loan portfolios. (6) In a macro stress test, which does not include full details of banks' individual exposures, it is uncertain to what extent the wider introduction of IRB models in the projection period will have an effect on banks' Tier 1 capital adequacy.

Implementing IRB models in one or more of a bank's portfolios usually results in a non-recurring effect on the bank's risk-weighted assets, reflecting the transition from standardised risk weights to risk weights calculated internally by the bank. In the event of a fall in risk-weighted assets, a bank's Tier 1 capital ratio will in isolation increase. To avoid an excessive reduction in banks' Tier 1 capital when IRB models are introduced, risk-weighted assets should not be reduced by more than 20 per cent relative to the Basel I level.

The banks (7) included in the stress test in FS 2/11 have been using IRB models approved by the supervisory authorities since 2007. The effect of the introduction of IRB models on banks' Tier 1 capital ratios has varied across banks. DNB Bank, for example, began to use IRB models in much of its corporate portfolio from the fourth quarter of 2010. Combined with increased equity capital, this led to a substantial rise in banks' Tier 1 capital adequacy ratios (see Chart 1). Overall, risk-weighted assets for the stress test banks fell by 9.9 per cent in the fourth quarter of 2010, even though the banks' total assets only fell by 0.5 per cent. As a result, Tier 1 capital adequacy ratios were 1.3 percentage points above the level previously projected in the baseline scenario in Financial Stability 2/10.

Several of the banks reported that they would be using IRB models more widely in their loan portfolios. Chart 2 shows the uncertainty related to the introduction of IRB models in the projection period. Under the assumption that risk-weighted assets fall to the same extent in the fourth quarter of 2011 as in the fourth quarter of 2010, Tier 1 capital ratios will increase considerably. If a similar fall occurs at end-2012, Tier 1 capital adequacy ratios in the adverse scenario will be...