Quantitative Indicators Signal US Expansion

Summary

• The National Bureau of Economic Research (NBER) can declare a recession even without two consecutive quarters of GDP contraction, a common definition.
• The economic indicators the NBER uses to identify recession clearly signal the economy is currently expanding.
• Leading indicators do not suggest a forthcoming recession though their track record is mixed.

Market Implications

• I still expect two more Federal Reserve (Fed) 2024 cuts, against the market pricing 3.1 cuts.

Recession Identification: Context Matters

In this note, I update my February discussion of recession risks, which I still find low.

In February, I assessed recession risks based on previous post-WWII recessions. Until the mid-1990s, inflation and associated Fed tightening had driven recessions. Starting in the 1990s, financial imbalances have driven recessions. I concluded neither type of recession was likely this time as disinflation was likely to continue and the debt built up this time had happened in the public rather than the private sector.

Below, I focus on quantitative indicators of recessions.

The most common definition of recession is two consecutive quarters of negative QoQ growth. The NBER officially dates recessions in the US. The NBER takes between four and 21 months after a recession begins to identify it.

The common and NBER recession definitions do not always overlap (Chart 1). For instance, in 2022 growth contracted QoQ in Q1 and Q2 but this did not meet the NBER definition of recession. This likely was because negative growth was a payback for the very strong inventory buildup the previous two quarters. The recovery was actually in full swing.

Also, the NBER identified 2001’s recession without two consecutive quarters of negative growth. In addition, while infrequent, negative growth can happen outside of recessions.

Current data shows we are much more likely to be in expansion than recession.

US Economy Still Expanding

Current GDP data does not signal recession (Chart 2). However, the GDP data is released quarterly with a month lag and revised multiple times.

This could be why the NBER definition is mainly based on monthly data, real personal income and consumption, NFP and household survey employment, real manufacturing and trade sales, industrial production and real GDP and GDI (the data is maintained by the St Louis Fed).

The NBER states ‘There is no fixed rule about what measures contribute information to the process or how they are weighted in our decisions.’ Yet these series clearly identify recessions (Chart 3).

In all recessions, the average z-score of the NBER series has been -0.9 or below, with one false positive in 2013 Q1 that reflected the impact of tax changes on household income. Most importantly, the NBER indicators do not suggest we are currently in recession.

The NBER does not rely on unemployment as a marker of the business cycle largely because, as it explains, unemployment is a trendless indicator while the chronology attempts to identify turning points in economic trends.

Yet, a steep increase in unemployment is a sure marker of recession (this is the essence of Sahm’s rule, Chart 4). This nonlinearity reflects the negative feedback loop between production and demand (less employment begets less consumption and vice versa) that occurs during downturns. From that perspective, the recent runup in unemployment appears too gradual to signal recession.

Lastly, PMIs can fall below 50 during periods of expansion, especially the manufacturing PMI (Chart 5).

Determining whether we are in recession is easier than predicting one.

Leading Indicators Do Not Signal Recession

In this section, I discuss five of the better-known leading indicators of US business cycles.

I start with two models available on FRED. The GDP-based recession indicator uses the statistical properties of GDP changes during recessions and expansions to compute the probability of recession (Chart 6). A value of the index above 67% has historically been a reliable indicator the economy has entered recession. Once this threshold has been passed, if it falls below 33% reliably signals the recession is over. The current value of the index is 4%, i.e., it strongly signals expansion, which I think is rather unsurprising given how strong GDP growth has been.

Another model is the smoothed recession probabilities, which predicts turning points in the US cycle based on the co-movements of four of the NBER series. Namely nonfarm payroll, IP, personal income and sales (specifically, this a dynamic-factor Markov-switching model). The model currently estimates a 2% recession probability.

However, these two models tend to show either a very high or very low probability of recession. That is, they are more coincident than leading indicators of recession.

Two truly leading indicators of recession with a long track record are owned by the Conference Board (CB) and the OECD. They are based on similar data but have different objectives (Table 1). The CB Leading Economic Index (LEI) signals recession whenever its six-month SAAR change is below 5%. The OECD’s Composite Economic Leading Indicator (CLI) tries to signal turning points in the business cycle six to nine months before they occur.

The two indices tend to move in sync, hardly surprising since they rely on roughly the same data. Both indices tend to fall before recessions but not all declines in the indices indicate recession. Not all growth slowdown results in recession.

The OECD index missed the GFC and the CB index started falling long before the GFC (four years), so it is unclear how much of a signal it really was.

The CB index gave a false positive recession signal in 2022-24. On the other hand, it cannot be fairly compared with the OECD index that is less ambitious (identifying turning points rather than announcing recessions!)

Last is the yield curve, which has recently dis-inverted (Chart 9). Until the pandemic, recessions followed inversions. This time a roaring recovery accompanied the inversion.

I think the inverted yield curve simply indicates the market does not believe short-term yields will prove sustainable. In the past, this signalled Fed easing in response to recession.

By contrast, this time it signalled expectations of disinflation and Fed easing that data has validated. Therefore, I do not see the recent curve steepening as a sign of forthcoming recession but rather of economic normalization.

Models Remain Helpful

None of the indicators of recessions discussed above have a great track record. This could reflect that recessions by nature tend to be unexpected. If they were a consensus forecast, economic agents would change their behaviours and recessions may not happen.

Also, there is randomness to recessions. An economy could have limited resiliency and yet avoid a recession because it is lucky. In the US, temporary layoffs have become increasingly correlated with workers not at work because of weather but still employed (Chart 10). This suggests extreme weather is starting to impact employment, with weaker firms laying off workers rather than drawing on cash reserves or credit lines to keep them on the payroll.

The inability of firms to manage shocks like adverse weather may not show up if the US is lucky and avoids such events. On the other hand, repeated and large negative shocks could trigger the negative feedback loop between demand and production mentioned above.

Overall, I am not arguing recession models are unhelpful but rather that they cannot be taken at face value and must be complemented by fundamental analysis.

Market Consequences

The Fed has shown greater sensitivity to employment than inflation risks. While inflation remains substantial (see this week’s webinar), the above analysis suggests employment risks could be limited. This is why I stick to my expectations of two more hikes in 2024, against the market pricing three.

Dominique Dwor-Frecaut is a macro strategist based in Southern California. She has worked on EM and DMs at hedge funds, on the sell side, the NY Fed , the IMF and the World Bank. She publishes the blog Macro Sis that discusses the drivers of macro returns.

(The commentary contained in the above article does not constitute an offer or a solicitation, or a recommendation to implement or liquidate an investment or to carry out any other transaction. It should not be used as a basis for any investment decision or other decision. Any investment decision should be based on appropriate professional advice specific to your needs.)