When macro models fail to predict markets, use micro models of giant companies instead

Most equity strategists have a collection of macro models to help assess broad economic conditions and predict stock market performance. These macro models differ from the micro models that analysts typically use to value a company, which range from simple fair price-to-earnings (P/E) ratio estimates applied to earnings forecasts to more detailed discounted cash flow (DCF) models. Of course, micro models can be applied to all companies in an equity index and aggregated; but proper macro models don’t use corporate financial data or valuation models to predict markets, instead they generally use macroeconomic data and rely on historical relationships for prediction. Some strategists are devout adherents of certain macro models; we find macro models useful when employed as a collection of market signals (e.g. Leading Economic Index, or LEI), but are most useful as guides for the inputs into our aggregated micro or intrinsic valuation models for the S&P 500.

 

Popular macro models: Associating macro conditions with market performance

Popular macro models include: 1) “Macro Compass,” this is usually based on the rate of change of real economic growth and inflation to then categorize macro conditions at the highest level into four quadrants; as with all macro models recent trends or forecasts are used to predict market performance being similar to outcomes during past same macroeconomic conditions; 2) “Institute of Supply Management (ISM) Clock,” this is based on the manufacturing ISM; when above 50 and rising it’s early cycle and bullish, if above 50 and falling it’s late cycle and cautionary, if well below 50 it’s a recession, if newly above 50 it’s a recovery; 3) the “Credit Impulse,” this monitors changes in the growth rate of loans, which is important for credit driven activities, especially construction activity; 4) the “U.S. Federal Reserve (Fed) cycle,” hikes above neutral or an inverted curve suggest late-cycle with elevated recession risk; 5) “cycle age,” years since the last recession are used to indicate risk of the next one. These are probably the most classic macro models for the stock market, but other popular ones focus on employment, fiscal conditions, election cycles, commodity prices, foreign exchange (FX) rates, asset allocations, monetary aggregates and especially since 2009 the Fed’s balance sheet.

 

We don’t rely on any single macro model and we are leery of recession signals

We don’t rely on any single macro model to predict the market, particularly those that aim to predict imminent recessions, but we use many macro models to help forecast S&P earnings per share (EPS). Because there is a loose relationship between gross domestic product (GDP) and S&P EPS growth outside of recessions, we use many sub GDP macro models. Our profit indicator models include the mfg. ISM, industrial production, investment spending on structures and equipment & software from the GDP accounts, gross exports, credit standards, bank loan growth, commodity prices, etc. We don’t treat cycles like reliable clocks, but we respect long-cycle forces that exist for many commodities and real estate and extended periods of sticky norms typical for macro variables like inflation, interest rates, FX rates and even volatility and risk premiums. We’ve long argued that an inverted yield curve is an imperfect recession indicator (like steep corrections), so is the age of an expansion, election years, a climb in oil prices or a drop in housing starts. Recessions are usually caused by monetary policy tightening against inflation or shocks, the latter difficult to predict and the former typically eases at the brink. We prefer to evaluate strength and weakness within a cycle than to predict longevity. Areas of strength and weakness often persist through the cycle. Ultimately, macro is uncertain and it’s difficult to diversify away macro risk.

 

Many macro models didn’t help the past year, some did but became less relevant

In 2023, US GDP surprised to the upside with robust 3% growth on services strength and inflation slowed but stayed at least 3%. Goods weighed on growth and inflation, whereas services boosted both. This was a crucial distinction within the broad data. Because S&P EPS was flat in 2023 owing to a mild but drawn-out goods manufacturing recession (mfg. ISM under 50 for 2 years), sluggish loan growth (mostly commercial, not household loans), down oil prices, sluggish foreign economies and a stable but strong dollar. All of these sub-cycle GDP and inflation factors affect S&P EPS more than GDP outside the broad crush of recessions. While S&P EPS was flat, neither a recession nor bear market occurred (rather 25% gain) despite the most vaunted of recession signals flashing red: an inverted yield curve. The curve remains inverted and a recession might still occur in late 2024 or 2025, but like the mid 1960s, mid 1980s and the most noted 1995 occurrence, soft landing do happen with a upward curve returning without a recession. Or the curve might stay flattish for a few years like 2004-2007 or for decades without a recession like in the UK. We respected the risk signal of the inverted curve and Fed tightening and still do, but we noted other factors mitigating recession risks such as the strength in services, sustained fiscal support, cheaper oil, but perhaps most importantly and continuing the surging equity market. Wealth effects from stock market gains boost spending and confidence and make financial conditions less restrictive than the Fed Funds rate alone would suggest. This brings us to Tech stocks and the phenomenon of the Great Eight the years since the pandemic.

 

S&P performance is now digital stock dominated; digital is the macro for the S&P

The Great Eight digital stocks are now 1/3rd of S&P 500 market cap. While S&P EPS remains global with positive sensitivities to manufacturing, commodity prices, credit conditions and loan growth, the digital trends now dominate. The earnings and market cap share of digital and big price moves seen at such pushed the S&P 500 toward growth style and digital trend sensitivities. Outside of recessions, we expect digital and healthcare earnings and returns to be the main drivers of S&P performance the rest of this decade. So we contemplate what kind of macro models help with that? Strategists must respect the now giant micro of Tech and Health Care and the intangible asset nature of these sectors.

 

Keep 2024E S&P EPS at $242, 20+ PE depends on yields and digital “duration”

We raise our 2024E S&P EPS at Great Eight and trim elsewhere after 4Q23 results and guidance. Despite strong 2024 EPS growth outlooks, the Great Eight PE is threatened by rising real Treasury yields and the sustainability of competitive advantages to provide them many more years of continued rapid economic profit growth at what is now macroeconomic scale. We think the long-term potential of Health Care is being overlooked and expect some Health stocks to join the future Great 10, 12, etc.