Over the past several years, credit union leadership has been challenged by regulators to document processes that heretofore had been credited to experience. It is understandable that leaders with successful track records would be frustrated by such a demand, but one would have to agree that the economic crisis that left many of us still uncertain about our future has made us question our experience and intuition. One might also be surprised to learn that recent studies have proven that our intuition can actually be a detriment to our decision making. For this reason, increasingly more financial institutions are using fact-based analysis to guide decision making today. In fact, the majority of financial institutions, as reported by FICO, are making improvements to their risk and decision management technologies a top priority in 2014.
In the early days of my financial career, I was assigned to the collections department of a large regional bank. Everyone I spoke to over the course of two years was delinquent on their account for one reason or another. One can imagine, then, how difficult it was for me to transition to the lending area. When asked to assess the risk of default of an applicant (this was before credit scoring was widely used), intuitively, I could almost always come up with a probable scenario where even the most credit worthy borrower could default. This is what Daniel Kahneman, Nobel Prize winner and author, calls Availability Bias. Our “intuition” is biased by the availability of information we have on the topic. For me, the availability of information I had on defaults outweighed my information on good payers, and I was biased toward a more risk-adverse response.
In his book, Think, Fast and Slow, Mr. Kahneman introduces the theory that our intuition, while relatively helpful, is almost always flawed and therefore fact-based decision making is superior. This is what analytics brings to loan portfolio risk management. When we have a fact-based architecture for decision making, our risk of falling victim to human biases is reduced. Analytics informs those fact-based decision models.
Consider what many credit unions refer to as a “Post Mortem” review of charge-off loans. When conducting a review after the fact, it seems relatively easy to see the factors that contributed to the default that should have been picked up in the initial review of the application. However, the “Post Mortem” is conducted in isolation, not in the context of a dozen or more other applications reviewed that day by the same underwriter. If you throw in a couple of other variables, things might not be quite so clear. Let’s say on the day the application in question was approved, there were eleven other applications that were declined by the same underwriter, and this was the best of the breed. Let’s also consider that the underwriter started their morning with a meeting in which the necessity for loan growth was discussed at length. These other factors create a bias that could make an otherwise bad decision look better in the proper context.
Kahneman also coins an acronym in his book, WYSIATI, or What You See Is All There Is. For many credit unions, this has played out with specific products, such as RV lending for example. A credit union has experienced high losses in RV lending, so they cease all RV lending. In reality, however, the high losses experienced may have only been related to an isolated number of loans. By eliminating all lending in the sector, the credit union stifles loan growth and is unable to meet certain members’ needs. If the credit union was to analyze the performance of all loans in the portfolio, it could isolate the characteristics of the loans that went bad and create fact-based decision models that allow them to continue lending in that sector.
Increasingly, lenders are using data to make informed decisions, eliminating the bias that exists naturally in our intuitive thought process. This is the battle line where a competitive edge will be lost or gained in the coming months and years. Getting a good handle on loan performance data, and employing the analytics to create informed decision models, will determine which lenders win the battle and which lenders do not.