US Recession Probability Reaches 67%

Via Political Calculations blog,

What month will the National Bureau of Economic Research someday get around to saying marked the beginning of the next recession in the U.S.?

The NBER is notoriously slow in identifying when the business cycle in the U.S. either peaks before going into recession or troughs when coming out of one, often lagging behind these events by many months. That’s because they take a number of data series into consideration and will wait until many go through revisions before determining if the national U.S. economy has truly changed direction from growth to contraction or vice versa according to their model of the economy.

Because they’re so slow, analysts have built models to try to predict the timing of when the country’s business cycle has changed when evidence is building that it has, long before the NBER makes its “official” determination. Some of these models are oriented toward recession forecasting. They have been built to use currently available data to try to anticipate the most likely timing of when the NBER will be likely to say the business cycle changed from boom to bust.

In a sense, they’re using models to predict when the NBER’s business cycle model will someday find the U.S. economy went into recession! That brings us to a recession forecasting model whose results we’ve been featuring since October 2022, when a leading recession indicator first flashed a red warning light.

That model was introduced by Jonathan Wright in a 2006 paper while he worked at the Federal Reserve Board. This model uses just three data series to generate the probability the NBER will identify a month sometime between a date of interest and one year into the future. One of those datapoints is the rolling one-quarter average of the Federal Funds Rate. The other two are the rolling one-quarter averages of the yields of two constant maturity U.S. Treasurys, one for the 10-Year UST note, the other for the 3-Month T-bill.

For this data, the red warning light starts flashing when the yield of the 3-Month Treasury drops below the yield of the 10-Year Treasury. This is a clear signal the U.S. Treasury yield curve has inverted, with short term yields paying higher yields than long-term yields. Historically, yield curve inversions have occurred before the U.S. economy entered into recession. Wright’s innovation was to also take the level of the Federal Funds Rate into account, recognizing that how the Federal Reserve sets it in accordance with its monetary policies affects the likelihood of recession.

As of the close of trading on 27 April 2023, Wright’s recession forecasting model anticipates a 67.0% probability the period between now and the end of April 2024 will contain the month the NBER will someday say marked the beginning of a national recession in the U.S.

The latest update to the Recession Probability Track shows how that probability has evolved since our previous update one month ago.

The chart shows the current probability of a recession being officially determined to have begun between 27 April 2023 and 27 April 2024 is 67.0%. Assuming the Fed follows through on hiking the Federal Funds Rate by another 0.25% next week, the probability of an “official” recession will continue rising. Doing some back-of-the-envelope math using our recession odds reckoning tool, with the 10-Year and 3-Month Treasuries as inverted as they are today, the odds of recession will breach the 70% threshold in about two weeks. Even if the Federal Reserve stops hiking the Federal Funds Rate after its Federal Open Market Committee meets to set its rate next week, the odds of recession will breach the 80% threshold in early July 2023.

Analyst’s Notes

Two members of the NBER’s Business Cycle Dating committee have a new working paper in which they indicate Americans should expect a recession in 2023 and 2024. The role Federal Reserve officials in changing from expansionary to contractionary monetary policies looms large in their assessment.