One of the steps involved in validating a model is back testing. If a model cannot even explain the past, the prospects of it explaining the future are grim. The same argument can be extended to methodologies of financial analysis- if a methodology of analysis cant predict most financial events of the past, using it to predict the future is a folly of formidable proportion. In the case of credit analysis, a methodology in use today should be able to predict past credit disasters. This is on account of the fact that sensible analysts would retool their analytical framework to take into account the shortcomings of their previous methodologies after those shortcomings were exposed by market forces. Consider the following methodologies: 1) Sovereign Credit assessment methodologies of Credit Rating Agencies 2) Big Bank Credit assessment of Credit Rating Agencies 3) Quant techniques of pricing credit risk using so called hazard rates and Poisson distributions. (it is a different matter that defaults do not follow a Poisson process- in times of abundant liquidity, there are negligible defaults. When liquidity tightens, there is a plethora of defaults. The Poisson modeling of credit defaults has mostly been devised by academics with no real world experience in the world of Credit) Now look at the Asian sovereign credit crisis, European sovereign credit problems, Lehman bankruptcy and allied bank credit events. Can the methodologies 1 to 3 be used to explain the past? If not, why use them to gaze at the future? The sample space of the drivers of credit default process can expand as more drivers of default are identified. All methodologies of credit analysis must incorporate, at the very least, the drivers of default that have been identified in the past. Future defaults might reveal new drivers. Superior methodologies will capture future these future drivers. The Asian credit crisis and bungled credit ratings revealed that sovereign credit analysis is incomplete if your methodology does not incorporate private sector borrowings in foreign currency. The European debt crisis revealed that the analytical framework is incomplete if sovereign credit analysis focuses exclusively on public sector debt, ignoring private sector debt. In an environment of rising private sector debt, asset prices tend to go up, causing government revenues from capital gains to shoot up. In such an environment, unemployment tends to be low- so government revenue from income taxes shoot up, while the expenditure for unemployment benefits fall. Since asset prices are high at that point, pension schemes seem fully funded. But there is limit to which the private sector can keep borrowing. At some point, private sector (household and corporate sector) defaults will shoot up. If the credit mania went on for some time, bank bailouts would be required. Stimulus packages might be necessary to bail out the economy. All these expenditures will cause government debt to go up. Irelands debt to GDP ratio was 25% in 2007. According to the IMF, it will hit 110% of GDP this year. Clearly ignoring private sector debt for sovereign credit analysis was a folly. But the credit rating methodologies of the rating agencies have not been amended for this. Obviously, the agencies will not be able to capture future sovereign debt events driven by private sector leveraging. That is the point of this piece on back testing. It is my conjecture that sovereign credit events can be forecast if one relies on two ratios-1) The consolidated Debt to GDP of society (this incorporates private sector and public sector debt) and 2) Marginal Productivity of Societal Debt over a period of time as measured by the ratio of change in GDP to the increase in Societal Debt (Public and Private Sector Debt). When the Marginal productivity of Societal Debt falls below cost of Debt, trouble looms in the horizon. It is not my case that merely measuring these two ratios will suffice. It is a good starting point, after which you have to apply abundant common sense to assess how societal competitiveness will evolve in future.