Preparing for the new credit loss standard
Model behavior is not enough
Models will be crucial to how banks implement the new current expected credit loss (CECL) standard, but models are not enough.
August 23, 2017
A blog post by Jonathan Prejean, managing director, Deloitte & Touche LLP*
It’s happened at least twice in the last month. When I sat down with a banking client and asked them what they were doing to prepare for the FASB’s current expected credit loss (CECL) standard, their immediate response was “we’re working on our model.” When I followed up with broader questions about the overall process and whether they had considered potential business impacts, I got a shrug and “those are not things we have time for right now.”
That position is confirmed by a recent survey Deloitte conducted to explore how banks are approaching CECL implementation, as well as the potential challenges. Over half of respondents identified either the development of statistical models or obtaining the necessary data as the single most challenging implementation task.
Certainly those models will be crucial to how banks incorporate forward-looking information into their credit loss estimates—and deciding whether they can simply tweak the models they have been using for stress-testing, whether they need to start from scratch, or something in between is an important decision. But the point I have been trying to make with my clients is that the model is not enough. Those things they say they don’t have time for could have an outsized impact if they wait too long to consider them.
Let’s just take a look at two areas that are receiving relatively scant attention.
The allowance process extends from the gathering of the data to the recording of expected losses on the financial statements. Loss estimation models are an important part of the process, but not the only part. For example, many banks expect to modify the models they currently use for stress testing to estimate expected losses. But financial reporting requirements are far more rigorous than those used for stress testing, and the expected loss estimation will need to be controlled in a manner that complies with financial reporting requirements.
The enhanced process will involve areas of the business (e.g., economists, regulatory stress testing) that are not currently involved today. These groups will need to be educated on both the enhanced expected loss estimation and the financial reporting requirements so that they are comfortable making assertions that the numbers recorded in the financial statements are accurate. Staff involved in development or alterations to the model need to understand what the required outputs are going to be. At the very least, these changes call for a comprehensive training and communications plan that banks should be thinking about now.
Starting to think about process issues early has an additional advantage: banks, ever on the lookout for ways to become more efficient, can begin evaluating emerging technologies, such as robotic or cognitive process automation that show promise in this area.
It is not necessary for banks to perfect their loss estimation models before considering process. In fact, if they wait until their model is complete, chances are they will end up cobbling together a less-than-optimal process at the last minute. A far better strategy is to take a step back and start evaluating how this number will be used - for example, is it just financial and regulatory reporting or will it also drive front-office incentives and underwriting levels?
At first blush, the new standard should not affect a bank’s business – and that is certainly the position most bankers have taken when I’ve asked them about the issue. The argument goes like this: We are not changing our credit risk profile by implementing the new standard, so why would it impact our pricing, our product mix, or anything else?
But consider this: As banks prepare to comply with the standard’s international equivalent, IFRS 9, which goes into effect in January 2018, they are discovering that changing the way they classify and measure their financial liabilities is actually going to have a far greater impact on their risk appetite and their portfolio strategy than they had originally surmised.
For example, because banks will now recognize full lifetime losses immediately, it will negatively impact their capital. If they have a target return on that capital, they need to find ways to maintain this margin. One strategy for doing this would be to increase pricing. Furthermore, because the standard requires banks to estimate losses over the contractual life of a loan, will banks want to consider reducing the life of their loans? Or, because they recognize losses earlier, will they start to become more conscious of the credit profile of borrowers? Suddenly, “no business impact” is sounding very different indeed.
This brings me back to my original point. The impending CECL standard may have more implications for banks than they realize, and doubling down on their loss estimation models to the exclusion of all else could well be a risky strategy. Taking a more holistic view of their entire process—and taking it early—is far more likely to both head off potential problems and uncover hidden opportunities. Those are benefits most bankers should be able to embrace wholeheartedly.
*Note: The views expressed in this blog are those of the blogger and not official statements by Deloitte Touche Tohmatsu Limited or any of its member firms, including Deloitte & Touche LLP.
This publication contains general information only and Deloitte is not, by means of this publication, rendering accounting, business, financial, investment, legal, tax, or other professional advice or services. This publication is not a substitute for such professional advice or services, nor should it be used as a basis for any decision or action that may affect your business. Before making any decision or taking any action that may affect your business, you should consult a qualified professional advisor. Deloitte shall not be responsible for any loss sustained by any person who relies on this publication.