How to Rock Macroeconomic Forecasting?

New Call-to-action

If the term “reasonable and supportable forecasts” means something to you, you might want to keep reading. If the phrase doesn’t ring a bell, scan through the new accounting standards update (ASU) on current expected credit loss (CECL) and you’ll understand soon enough why reasonable and supportable forecasts are worth talking about.


In a nutshell, FASB will be requiring that future macroeconomic conditions be considered when estimating credit losses in CECL. So how do you approach these forecasts? While we know all too well that there isn’t a trivial answer to this question, there are tools out there to help us obtain “reasonable and supportable” estimates. I’d like to informally discuss two such tools in this post: national projections and ARIMA time series forecasts.



National Projections 

This is, perhaps, the simplest resource for obtaining macroeconomic forecasts, as the Federal Reserve does the hard part for us. As an example, suppose you’re interested in obtaining unemployment projections for California. The fed produces three years’ worth of national unemployment projections. So, using historical data to quantify the relationship between California’s unemployment and that of the nation, you can apply that relationship to the national forecasts to obtain reasonable forecasts for California. A similar approach can be taken for any forecasts produced by the government.  



National Unemploymentnational unemployment.gif

ARIMA Time Series Forecasts
ARIMA, standing for auto-regressive integrated moving averages, is a statistical approach to making projections. Given a time series on any variable (e.g. unemployment), an ARIMA model uses historical patterns in the series to produce forecasts. Short-term forecasts are naturally more precise than long-term, as uncertainty typically increases with time.


While projections can be extremely difficult to nail down exactly, these are some tools at your disposal that can help you satisfy the “reasonable and supportable forecasts” requirement for CECL.

Rachel Messick

Product Manager/Data Scientist at Visible Equity