Estimating Recession Risk After The Weak Q1 GDP Report

The US economy slowed to a standstill in the first quarter, raising fresh questions about the macro outlook. But if the 0.2% gain in GDP in the first three months of this year is a reason to worry, the Federal Reserve prefers to emphasize the positive. The central bank labels the soft Q1 data as a victim of “transitory factors”. In yesterday’s FOMC statement the Fed advised that “although growth in output and employment slowed during the first quarter, the Committee continues to expect that, with appropriate policy accommodation, economic activity will expand at a moderate pace.”

Some economists agree, including Jonathan Wright, a former Fed analyst who’s currently a dismal scientist at Johns Hopkins University. “A weak GDP number starts to raise the concern that the economy has hit a soft patch,” he tells The Washington Post. “But there are several reasons for discounting the quarter one data, weather being one.”

Perhaps, although Bloomberg reports that several prominent forecasting shops have cut their second-quarters forecasts for GDP in the wake of yesterday’s disappointing Q1 release. “Given the fact the various regional purchasing manager surveys have been weak in April, it appears that producers will allow inventory positions to run off,” Joseph LaVorgna, chief U.S. economist at Deutsche Bank Securities, wrote in a note to clients. “This tells us that current-quarter growth is likely to run around 2.5 percent, not the 4 percent snap-back we had previously been anticipating.”

Looking ahead comes with all the usual caveats, of course. Meantime, what does the rear-view mirror tell us about recession risk? Not surprisingly, the potential for trouble is a bit higher in statistical terms after yesterday’s GDP release. But even after factoring in the latest number the threat remains relatively low when looking through a GDP prism.

The somewhat battered but still upbeat profile is based on estimating recession probabilities with a probit model. The two inputs for this methodology: the quarterly GDP changes from 1947 through yesterday’s initial estimate for Q1:2015 and NBER’s recession dates. The model looks for a relationship between the data sets and estimates a probability that the dependent variable–NBER’s recession dates–signals that a new recession has started, based on analyzing the history in context with the independent variable (GDP in this case). Crunching the numbers in R suggests that recession risk for the US remains quite low, at roughly 28%. That’s up from 9% based on last year’s Q4 data, although the latest estimate is still well below levels reached in recent years–levels didn’t equate with a new recession. In short, it doesn’t appear likely that the NBER will mark Q1 as the start of another recession, or so the available numbers in hand imply.


We shouldn’t rely on GDP alone to assess recession risk, but for the moment the crowd’s obsessing over yesterday’s news and so it’s useful to look at the data through a degree of statistical objectivity. The real work lies ahead, of course, as the Q2 profile continues to roll in. For now, it’s fair to say that business cycle risk is higher but remains well below the danger zone. The critical question, and one that can only be answered in the weeks ahead; Is the recent deterioration in the trend a temporary affair?