Last year, we introduced our GMM Macro Regime Classification (MRC) model to create a dynamic, multiple macro factor classification of the current economic backdrop. Popular recession models tend to either generate lots of false positives or have large variable lead times (10s-2s invert before every recession but with a 2qrt to 2yr lead). Popular regime models tend to be narrowly focused (ISM clocks for instance) and their output is VERY input specific.
Our approach provides a shorter-term signal relevant for understanding current market rotations/risks and uses a broad data set that measures the breadth of the economic backdrop. That last part is particularly relevant now when the macro backdrop is unusually disjointed (deeply inverted yield curves, very low urate, normal credit spreads, etc.).
Today, we expand our model by adding regime probabilities. Based on our model history, the current period is classified as Transition nearly 80% of the time. The other 20% of the time, today’s backdrop would map to a Normal rather than Recession period. That indicates macro factors, in aggregate, are moving away from a recession. Regime classification is dependent on unknowable data, but reduced current readings suggest reduced odds of a recession, which supports a more risk-on factor/industry stance after a strong month of risk-off returns.
Growth has clearly slowed, and inflation remains too high for the Fed to pivot, so investors are focused on recession risk in 2023. Based on December data, our MRC model still puts the economy in a Transition period. Transitions (slowdowns that end in an acceleration or recession) are even more rare than recessions, and the current period is the second longest Transition in the nearly 40-year history of our model. Only the post-Dotcom Bubble period was longer. The lack of a clear growth trajectory (recession, slowdown, reacceleration) helps explain the elevated macro influence over the market and the still high level of correlations (stock and factor).
Though recession signals remain mixed, earnings estimate has already dropped. In line with our discussion (HERE), NTM EPS median monthly changes were positive in most non-Recession periods. Consecutive declines over the past four months either suggest a mild recession is approaching, which is our base case. If a recession is avoided the trend should reverse soon.
Recession Odds Dropped in December: Slowing growth and recession risk remain a major focus for investors into 2023, particularly after the yield curve inverted in late-October. Our updated Macro Regime Classification (MRC) model remains in a Transition period based on year-end data. The current Transition period has been unusually long, lasting 10 months, the second longest Transition behind only 2002.
The prolonged period of economic uncertainty helps explain why macro influence has remained high and correlations (stock and factor) are still elevated.
To better measure the market classification and internal transitions between the classifications, we further expand the MRC by adding period probabilities, that reflect 1) the current dominant classification and 2) the probability of that classification. The probability that the current period should be called a Transition was over 90% in early 2022, but has dropped to 79% today. At the same time, the probability the economic is accelerating toward Normal increased to 21%. Macro factors are moving closer to historical Normal periods, suggesting a recession may still be avoided.
Currently most of the macro readings are back to their normal range except for unemployment and inflation. Inflation has dropped from an extremely high level but is still well outside its typical range, and unemployment is very low. A steady decline in inflation is the Fed’s goal and is being priced into futures markets. How unemployment behaves will determine if a recession starts or can be avoided. What the MRC model tells us is that the rest of the economy is NOT signaling that a recession is imminent.
Other recession measures, such as NY Fed probability of recession based on treasury spread and Cleveland Fed model based on yield curve, are elevated. Recession probability by the end of 2023 has elevated above 50% based on Cleveland Fed model. Those trends help explain the sharp rotation into risk-off factors and sectors over the past month. Today, it appear markets are discounting a recession that the rest of the economy is not signaling.
Earnings Focus: We are not saying there will not be a recession, just that signs of a deep recession have not emerged. Though recession signals remain mixed, earnings estimate have been moving lower. As we discussed earlier (HERE), a sharp earnings drawdown is unlikely unless there is a severe recession. During historical classified periods, median monthly NTM EPS change for the S&P were all positive except during Recessions (this is based on forward estimates, trailing numbers follow a different pattern). The consecutive decline over the past four months either suggests a mild recession is approaching, which is our base case, or a recession can be avoided and the trend will reverse soon.
Sector EPS monthly changes have diverged between Recession and Normal periods. Only Staples and Health Care NTM EPS average monthly changes staying positive under Transition, Normal, and Recession periods. Earnings changes for Deep Cyclicals, especially Energy, are the most negative during Recession relative to Normal, while Defensives earnings tend to be less impacted by Recessions. Put simply, EPS trends of Defensives are not likely to be a support unless there is a deep recession.
