The effectiveness of risk models in the face of increasing loan defaults

“Remember that all models are wrong; the practical question is how wrong do they have to be to not be useful.” George E. P Box

In recent years, the number of analytical techniques available for managing credit risk has increased rapidly. Information technology has made information and analytics readily available to anyone with a computer.Today’s financial engineers seem extremely sophisticated with ultra effective credit analysis tools supposedly designed to enable them manage credit risk with precision.

Financial institutions today parade smart, well-educated young people who are armed with an array of systematic techniques and tools that their predecessors could not have thought possible. And they have huge resources at their disposal.

However it’s still sad to note that despite the amazing sophistication financial institutions seem to posses in analysing risk, they still grapple with issues of loan default. This is the issue all around the world. One would have imagined that the reverse will be the case.

The Central Bank of Nigeria 2016 Financial Stability report showed that the banking sector witnessed a rise in non-performing loans due to weakening economic conditions during the review period. A recent case study is the default of N541 billion naira loan by Etisalat Nigeria. Thanks for the intervention of the Central Bank and NCC, for had the banks taking over the company like they intended the already slumbering economy would have faced some more serious setback.

Understandably, the purpose of credit risk management is not to eliminate risk but rather to deal with risk in the most effective and sensible manner while protecting the interest of the lending institution. But even so loan default is always a demonstration that there were issues with identification and analysis of credit risk.

A credit officer is like a detective; his aim is to see reasons why the loan will not be granted after all, except the analysis proves otherwise. So in a case where loan default is on the increase, it sends a red flag to financial institutions to re-examine their credit analysis tools.

If nothing else, this situation will help financial institutions understand that, contrary to what it seems, they do not own a credit assessment black box that will reliably spit out the right answer to each and every question fed into it. Financial professionals must now also take model risk more seriously.

Models are used to represent some object in the real world. We build models so that we can use them to infer things about that object in the real world. Model risk in finance is therefore defined as the risk of financial loss resulting from the use of financial models. Put simply, it is the risk of being wrong; but to be more specific it is the risk of being very wrong. Model risk arises from many sources.

First, a model may not have captured all the relevant risk factors correctly. For example, the 2008 global financial crises resulted from the failure of the subprime mortgage models to predict the increased chances of default because the housing price appreciation risk was not a factor in these models.

Secondly, all the participants use the same model to arrive at the same conclusion giving rise to the same wrong judgement across board.

Thirdly, models cannot anticipate changing behaviour. At the highest level, what this shows is that there are risks associated with using analytics to create new financial institutions and strategies to exploit the inefficiencies in a market. In the process of financial engineering, you change the environment so that it no longer matches the one modelled. In the words of Emanuel Derman, ‘the reliance on models to handle risk carries its own risk’

In a world of seemingly imperfect models, senior management in financial services firms must renew their attention to traditional management tools: checks, balances, and controls, and, behind them, a strong risk culture. The rate of loan default should sound an alarm to financial institutions concerning credit risk management.

In these circumstances, it is extremely important for the firm to have a clear understanding of what their risk tolerances are and how they plan to manage those risks. That means there has to be clearly defined risk culture. The institution’s general understanding, custom and belief about credit must be clearly defined. A clear risk culture will produce better results than any conceivable model and analytical tool.

One important key to a strong risk culture is to have management make the company’s risk appetite abundantly clear to every employee. Business strategies need to align with risk tolerances and reward systems. Management needs to model good practice and get involved in risk decisions.

There is plenty of evidence to support the concept that market savvy, informed, experienced, well-capitalized, and disciplined financial players can prosper irrespective of the Nigeria’s economic situation. It will not be so easy for those who are not.

 

Brian Reuben