It has become clear that many banks weren't prepared for the economic and financial crisis that struck in 2008. The good news is that these hard times spurred analytic innovation and produced useful data to strengthen risk management.
Not only were the analytic models that banks employed in 2008 ill-prepared for the depth and length of the recession, but bank executives were undermined by their inability to do more to manage through that period.
Our postcrisis research yielded three major lessons about credit risk management:
• Risk is dynamic, and all elements of risk management must be treated dynamically.
• Rapid and significant changes in economic forces and market forces can render traditional risk management approaches less reliable.
• Credit providers need better economic forecasting relative to risk management for loan origination and portfolio management.
While most scoring models are able to rank order consumer risk accurately during turbulent times, there is empirical evidence that economic upheaval can have a significant impact on default rates even when credit scores stay the same. In other words, immediate past default experience may be a poor indicator of future payment performance when economic conditions deteriorate rapidly.
For example, in 2005 and 2006 a 2% default rate was associated with a FICO score of 650 to 660. By 2007, a 2% default rate was associated with a score of about 710 as rapidly worsening economic conditions (and the impact of prior weak underwriting standards) affected loan performance.
Although most banks already incorporate some type of economic forecasting into their policies, our experience indicates a substantial portion of this input is static and subjective.
However, progress in data collection and predictive analytics — particularly over the last three years — allows forecasting based on a more objective and empirical foundation.
Going forward, risk estimates should incorporate the potential future impact of changes in key economic indicators (unemployment, gross domestic product, housing prices, per-capita income) on credit risk. Such economic forecasts can augment credit-risk prediction in two ways.
First, it can be used to improve predictions for payment performance associated with any given score. These improved predictions can be incorporated into individual lending decisions and be used at the aggregate level to predict portfolio performance.
Second, economic data can be used to predict the migration of assets between tranches of risk grades.
When used in conjunction with aggregate portfolio default probabilities, this can form the basis of forecasting risk-weighted assets for the purpose of Basel capital calculations.
The ultimate benefit of adding this type of forecasting capability into credit-risk management is that lenders will have an enhanced ability to:
• Limit losses by tightening credit policies sooner and targeting appropriate customer segments more precisely.
• Grow portfolios in a less risky and more sustainable manner.
• Prepare for the future with improvements in long-term strategy and stress testing.
• Achieve compliance with capital regulations more efficiently. (Improved accuracy in reserving will also reduce the cost of capital.)
We applied this methodology to the portfolio of a top-10 U.S. credit card issuer.
We compared the actual bad rate against predictions from the traditional historical odds approach as well as the economically calibrated methodology. We found that the latter reduced the issuer's error rate (the difference between the actual and predicted bad rates) by 73% over three years.
In a second example, the European lender Raiffeisen Bank International addressed a significant increase in delinquencies in its personal-loan portfolio by utilizing the economically calibrated methodology to adjust its origination strategy. The result was an increase in profit per loan of $11.50.
Raiffeisen went on to apply a similar approach to the issue of grade migration. It used economic forecasts as adjusters to risk models as well as inputs into Basel risk-weighted assets calculations. The incorporation of economic data in this way also enables Raiffeisen to build countercyclicality into its calculations. Failing to do so leads inevitably to procyclical forecasts as models predict falling risk in good times and rising risk in bad times.
As part of our research, another U.S. credit card issuer retroactively applied this economic-impact methodology to its credit-line-reduction and collections strategies. An analysis of its 2008 data (conducted with the new methodology) found that the predicted bad rate for its portfolio rose more than 250 basis points compared with predictions based on a more traditional approach. The new approach would have decreased the amount of credit extended to a larger portion of the portfolio (and not decreased credit to those less sensitive to the downturn). The lender would have realized yearly loss savings of millions of dollars.
In addition, the new methodology would have allowed the bank to target 41% of its portfolio accounts for more aggressive early-stage delinquency collections. Assuming only a 3% increase in curing delinquencies, the issuer would have saved millions more on top of the loss savings it would have achieved from improved credit-line management.
Lenders now have the analytic tools to enable safer, measured growth while simultaneously preparing for a lingering recession. The use of such forward-looking analytic tools is fast becoming a best practice in risk management. With risk predictions better aligned to current and future economic conditions, lenders can adjust more quickly to a dynamic market and steer their portfolios through uncertain times.
