To answer the above questions, it is important to carefully scan the technical note released by the EU on the macroeconomic scenario and reference risk parameters on July 23.
Methodology of the Stress Test:
For risk professionals brought up in tghe environ of simulation and probability, the stress test methodology would look rather simple. At the base of it there is a macro model that develops the relationship between key macro variables like output, inflation, unemployment, interest rates and exchange rates. Once there is a robust econometric model ( either structural regressions or time series variety, the stress tests moves into the second level ( mainly dealing with regressions to associate sector level historic PDs and LGDs with the macro parameters. At the third level, there are baseline ( trend) and adverse scenario ( shock) that needs to be tested on the bank's balance sheet. The stressed balance sheet will then be used to examine the adequacy of capital ratios.
For the recent EU tests, the Country level Macro models were provided by specialists in the ECB . This is not surprising, as such models were already in use in the central banks for monetary policy purposes. Such econometric models will now have an expanded use, in the area of financial stability.
Second, the usual suspects for providing country level PD and LGD numbers is there, the EDF of Moody's KMV model. This, when merged with market data from Bloomberg, we have scenarios developed at country level for the Stress Test. Thereafter, the Basel Pillar I takes over as each Bank balance sheets need to be run through the constructed world. We still have less clarilty on the last few steps, but this is broadly rthe method.
My view about the framework;
1. For emerging markets, the problem lies at the very inception even as the attempt to merge macroeconomic models with micro level information from Banks is indeed a step in the right direction. Here, the regulators of the emerging markets may face a hurdle due to absense of robust macro models. Unless substantial effort in macro modelling is placed, it is difficult to arrive at a robust stress test framework.
2. There is very little evidence to know the degree of confidence one can place in the model. In other words, the model needs to be transparent and subjected to model risk assessments. Unless confidence is developed, applying additional tests for all countries, the required comfort level will be missing.
3. Once again, we are relying on the same set of rating and default information that had cxome under criticism during the financial crisis. Additionally, in emeging market, using advanced country default information would be highly unreliable and certainly less representative. While there are very few alternatives, caution is needed here.
4. To any econometricians who have dabbled with structural macro models and time series models ( Cointegrations and Vector autoregressions), it is easy to know that assumptions on lag length, ordering of parameters and many other factors can create havoc with the results.
Therefore, while the EU initiative is a step in the right direction, the stress tests will need time to establish credibility. During this timeframe, appropriate tests needs to be devised to ensure rigorous model validation exercise. Notwithstanding the usefulness of stress testing as a useful communication tool of the Central Bank, it is too early to be judgemental. it is only the first step in the rocky road to regulatory reform.
For the emerging markets, there is an imminent need to work on the basic building blocks of the stress testing framework and it will not at all be useful to implant models from an entirely different economic and financial environment.
