Weighing Recession Probabilities

Episode Summary

Executive Summary: The hosts debate rising recession risk amid the Iran war, higher oil prices, weaker markets, and softening labor data. Their recession probabilities cluster around 40%-45%, while Moody’s random forest model recently moved from just over 48% to about 40% after more data. They agree the labor market is the key trigger, but baseline forecasts still assume no recession unless job losses, unemployment, and broader hard data deteriorate further.

Main Topics: Recession probability rising amid war-driven shocks (Priority: 5/5): The conversation centers on whether the economy is moving toward recession as the Iran war pushes up oil prices, lifts rates, and weakens financial markets. Moody’s random forest recession model (Priority: 5/5): Shandor Witcher explains the machine-learning model that aggregates many indicators into a 12-month recession probability, with labor market data carrying the most weight. Labor market as the decisive signal (Priority: 5/5): The hosts repeatedly emphasize payroll declines, unemployment claims, and the SOM rule as the main triggers that would justify changing the baseline forecast to recession. Why economists are cautious about calling recession (Priority: 4/5): They discuss recency bias, reputational risk, and the possibility that economists are reluctant to repeat overly pessimistic calls after missing the 2022-2023 slowdown. Limits of traditional indicators (Priority: 4/5): The group questions the usefulness of initial claims and notes that GDP could remain positive even if the labor market is effectively in recession. Conference Board leading indicator and market stress (Priority: 4/5): Shandor highlights the Conference Board composite leading indicator’s large peak-to-trough decline as a historically strong recession signal, even though it failed to predict the earlier slowdown. Forecast baseline still not recession (Priority: 3/5): Despite elevated risk, the team’s baseline assumes the conflict eases, oil prices retreat somewhat, and growth slows but avoids a formal recession.

Key Arguments: Recession risk is materially higher because the war has created a sustained oil-price shock, tighter financial conditions, and weaker confidence. Fiscal stimulus from tax cuts, defense spending, and accelerated depreciation is still cushioning consumers and businesses, keeping recession odds below what they would otherwise be. The labor market is the most important leading signal; persistent job losses and rising unemployment would likely force a recession call. Initial claims may be less reliable than in the past because of stricter UI eligibility rules and more gig work after layoffs. A recession could be present even if GDP stays positive, if payrolls and unemployment deteriorate enough. The Conference Board leading indicator’s long decline is a strong warning sign, even though it previously gave false alarms. Economists may be anchoring on the recent experience of being wrong about the 2022-2023 recession call, making them more cautious now.

Data Points: Chris recession probability: 42% - His estimate for recession starting within the next 12 months. Marissa recession probability: 45% - Her updated estimate given the ongoing Iran war and weak data. Moody’s model probability after February jobs release: just over 48% - Shandor’s random forest model jumped after the February employment data. Moody’s model probability after more data: about 40% - The model was rerun with additional indicators and revised down. 10-year Treasury yield change: up about 0.5 percentage point - Market reaction since the war began. 30-year fixed mortgage rate: 6.5% - Current mortgage rate cited as rates rose. Mortgage rate cited by Marissa: 6.64% - Her morning check of mortgage rates. Gasoline price increase: up $1 per gallon - Regular unleaded gasoline since the war started. Stock market decline: down 7%-8% - Approximate drop since the war began. 10-year Treasury yield level: 4.4% - Marissa cited the current 10-year yield. 2-year Treasury yield: 3.98% - Used to note the yield curve remains positively sloped. Conference Board leading indicator peak-to-trough decline: 18.4% - Shandor’s chosen recession-warning statistic. Largest prior decline in that indicator: 25.6% - Historical comparison for the leading indicator. SOM rule threshold: 0.5 percentage point rise in unemployment rate - Three-month moving average increase from the prior year’s low. Probability threshold for changing forecast: two-thirds - Mark’s rule of thumb for making a major forecast change.

Pivotal Quotes: "I think we're very labor market oriented, so persistent job declines, right, for several months." — Chris: Explaining what would push the team to adopt recession as the baseline forecast. "The key variable I didn't hear anybody mention is initial claims." — Shandor: Noting a labor-market indicator he thinks deserves more attention. "You could have GDP expansion with labor market recession, right, in this weird new environment that we have." — Mark Zandi: Discussing the possibility of recession-like labor conditions without negative GDP.

Implications: Listeners should expect recession risk to stay elevated unless labor data improve and the Iran/oil shock fades. The team is not yet calling recession, but a few more weak jobs reports or higher unemployment could quickly change the baseline.