A Goldilocks strategy for loan loss forecasting

Nancy Carter & Kim Steinberg | Data Solutions Group, FIS

March 08, 2021

Setting reserve levels just right while preparing for CECL

Lending is forecasted to pick up in Q2, making now a good time for lending officers to evaluate portfolio strategies. In 2020, many financial institutions overcompensated for the pandemic-induced recession by setting aside too much capital for reserves to cover future loan losses. These actions took valuable capital out of the lending stream and, consequently, have impacted lending revenue. Many lenders are now seeking third-party expertise to help re-establish the right reserve levels so they can free up the maximum amount of capital for lending.


At the same time, as private institutions look ahead, new CECL requirements are coming at the end of 2022. FIs will need time to understand the impacts of CECL on their portfolios and make changes to lending strategies in advance to avoid a crippling surprise.

These are two compelling reasons for FIs to adopt forecasting models right now. As we’ll cover, some solutions in the market are not sophisticated enough. On the other hand, some are just too much. If this is something that keeps you up at night, it’s with good reason. But the good news is that you can solve the need for critical insights to better manage your business and impending compliance at the same time with the same tool.

Rethink Your Approach

Making loan loss reserve decisions based on loans that are currently or have historically been in default is no longer good enough. Many FIs still use the Weighted Average Remaining Maturity (WARM) methodology to calculate their expected credit loss (ECL). This spreadsheet-based formula relies solely on the historical charge-off rate average for the entire portfolio, leaving an FI open to enormous discrepancies and risk, especially in today’s fluid market, in which it is impossible to reasonably propose extrapolating historical data into the future as the sole means of setting forward-looking guidance. Therefore, lenders need to incorporate comprehensive economic indicators into forward-looking, predictive modeling. In order to make adequate predictions about loan conditions into the future, a data science algorithm is required that joins economic event data to critical borrower information, and tracks both for changes over time.

It’s not as hard as you might think to find a forecasting model that’s just right for you. Predictive models that include macro- and micro- economic factors and extensive borrower data are available today, and can help you to settle in to the appropriate reserve ratio for your portfolios, positioning you to maximize revenue opportunities as lending picks up in the coming months. Notably, being prepared to service lending customers as demand increases has the happy side effect of satisfying CECL requirements well in advance.

It’s Time to Embrace Forecasting

CECL may not have been popular, but it may yet prove useful. The intention of CECL, as a response to the Great Recession, was to address major shortcomings of the incurred loss model and prevent a sudden dearth of capital from grinding new credit to a halt. Its new forecasting requirements, while initially a heavy lift for institutions, can actually help FIs immediately by providing the right insight to balance reserve levels – mitigating the risk of being caught short while also ensuring that you can maximize every lending opportunity. Forecasting, in general, allows you to insulate your access to capital against fluctuating economic conditions.

While CECL was designed to ultimately force forecasting onto your radar, the sooner you can take an informed position on converging realities of economic conditions and customer demand for new credit, the healthier your balance sheet will be.

Adapting to the New Norm

Institutions that have accurate forecasting will thrive as they avoid over or under compensating on reserves. At the same time, they will be in position for new CECL requirements that go into effect December 2022. Furthermore, they’ll have the necessary runway to adapt their lending strategies and pivot if portfolios need rebalancing.

The ideal solution allows lenders to identify the contribution of loan products to the overall Expected Credit Loss (ECL). Using this as a prioritization technique allows institutions to focus efforts to revisit risk appetite and underwriting thresholds to the products that have the most impact to prescribed reserves. This will also provide insight into repackaging and resale tactics that could help institutions to write down or otherwise offload the products and loans that don’t fit within post-CECL default risk strategies.

At the same time, institutions can source new insights into their decisioning logic for new loan issuance, such as alternative credit scoring models underpinned by machine learning, as well as streamlined access to tax records. Newly issued credit tends to have a more pronounced impact on ECL, versus older vintage loans. Therefore, it is critical to get a baseline quickly in order to navigate servicing customers through the coming months.

As you evaluate whether to build your own modeling tools, leverage a current partner or find a third-party vendor, it’s important to act now. FIS has the added the capabilities to make this turnkey for your institution by partnering with Oliver Wyman to create an exclusive solution that you can act on right away. Developing a strategic response is possible once your institution has a baseline of expected loan loss reserve changes across the portfolio, both to boost short term revenue opportunity, as well as satisfy the CECL auditors.

For a more detailed look at this topic see the FIS CECL Point of View eBook.