United States: Double recession

With the Federal Reserve Bank of Atlanta’s GDPNow-cast for Q3 at 2.4% (as of September 30), the US economy may be coming out of the technical recession that started in Q1. At the same time, the US Treasury yield curve is still inverted, indicating that there is more to come next year. This suggests that we may be facing two recessions caused by the same factors, one technical and one official – i.e. approved by the National Bureau of Economic Research (NBER) – interrupted by an episode of positive GDP growth that may have started in Q3. What are the common causes behind these two recessions, why do we expect a second recession, and why do we expect it next year?

Let’s start with the recession that already happened. Since a technical recession is defined as two consecutive quarters of negative GDP growth, the US economy was in a technical recession in the first half of this year. However, there is still some doubt about this ‘recession.’ The third GDP estimate for Q2, released on September 29, not only confirmed that GDP growth was negative in the first two quarters (-1.6% and -0.6%, respectively), but it also showed that GDI growth was positive in both quarters (0.8% and 0.1%). In theory, GDP and GDI should be equal, because they measure the same output. However, GDP measures the value added of goods and services produced while GDI measures the income earned in this production (such as wages, profits, interest income, and rental income). In practice, there can be considerable statistical discrepancies (Figure 1).

At present, GDI seems to be more consistent with the robust labor market data – which disqualify H1 for official approval as a recession by the NBER – than GDP. This means that the ‘technical recession’ of H1 could still be a measurement error. However, it is clear that both measures showed a severe deterioration after the final quarter of 2021. While GDP shows a technical recession, GDI at least indicates a technical slowdown in the first half of 2022.

What about the third quarter? While employment growth remains strong, the Atlanta Fed’s GDPNow-cast for Q3 stands at 2.4% (as of September 30). This suggests that GDP growth may have turned positive in the second half of the year.

Unfortunately, once the technical recession is in the rearview mirror, we are heading for an NBER-stamped recession, probably next year. The economy is still getting hit by supply shocks, high and persistent inflation, and tighter monetary policy. Underlying momentum in the economy is fading and it is only a matter of time before businesses slow down hiring. At the moment, they are still hiring at a high pace to keep up with excess demand for goods and services, but when this comes down we are going to see a slowdown in employment growth as well. Eventually, when employment growth turns negative and the unemployment rate continues to rise, the NBER will declare an official recession. So an NBER-approved recession is only a matter of time. In fact, we may need one to get inflation out of the system. Therefore, a recession seems inevitable: even if the US is able to absorb the exogenous shocks to the supply side of the economy, the response of the central bank to the wage-price spiral (Figure 2) will cause a recession from within.

The Federal Open Market Committee (FOMC) expects to get inflation down through below-trend growth and some softening in the labor market. What’s more, the FOMC intends to keep at it until inflation is back at the 2% target. Regarding the probability of a soft landing, at the post-meeting press conference on September 21, Federal Reserve Chair Jerome Powell said he doesn’t know what the odds are and that he wished there were a painless solution, but there isn’t. He did expect a modest rise in unemployment and he claimed that it is plausible that job openings could come down without a large rise in unemployment. Powell said that nobody knows if this leads to recession or how deep this recession would be. He said that the chances of a soft landing fall if the Fed has to raise rates higher and keep them high for longer. While the Fed does not know the probability of a recession, or does not want to admit that it is above 50%, we can use the yield curve to put a number on it, based on empirical evidence.

The Fed’s increasingly aggressive hiking cycle has led to a strong rise in US treasury yields this year. The ten-year yield has moved up from 1.63% in early January to 3.83 on September 30. However, the short end of the curve – which is more closely related to the Fed’s rate policy – has moved up even more than the longer end of the curve. The two-year yield has moved from 0.76% in early January to 4.22 on September 30. Consequently, the yield curve has inverted on the 2-10 segment. Usually, an inverted US treasury yield curve is seen as a harbinger of recession. This is because inversions are often followed by recessions (Figure 3). Does this mean that we are heading for a(nother) recession?

In an earlier special, we looked at the empirical evidence regarding inversions and recessions and estimated a logit model that mapped the relationship between the 2-10 yield spread and recessions as defined by the NBER. The model was estimated on data well before the Covid-19 pandemic, because the exact timing of the 2020 recession (March-April) was not related to the business cycle. The estimation results indicated that inversions have to be interpreted with care. An inversion does not necessarily mean a recession. In particular, for the 2-10 spread our logit model suggested that we have to allow for a threshold level of -16 bps. Only when that level has been breached does the probability of a recession exceed 50%.

The same model can also be used to calculate the recession probability for the 2-10 spread at each point in time. The evolution of this probability is shown in Figure 4. The chance of a recession climbed above 50% in July, as the threshold level of -16 bps was breached. At present, with a spread of -39 bps on September 30, we calculate a 70% probability.

So even if the end of the technical recession in the US could be in sight, the yield curve suggests that an NBER-approved recession is still in the cards. Therefore, it looks like we are entering an episode between two related recessions, one technical, the other official. They are related, because ultimately the causes are the same. The exogenous negative supply shocks, the resulting high inflation, and the Fed’s hiking cycle to get inflation back to 2% all contributed to the technical recession and are likely to cause an official recession next year as well.

The negative contributions of net exports to Q1 GDP growth and inventory accumulation to Q2 GDP growth (Table 1) can be traced backed to exogenous supply shocks. The large decline in residential investment that contributed to negative GDP growth in Q2 was caused by high mortgage rates reflecting the endogenous response of the Fed to rising inflation caused by the negative supply shocks. The high and persistent inflation caused by the supply shocks and the endogenous wage-spiral that was allowed to develop by the slow response from the Fed is dragging down personal consumption spending. Strong household balance sheets, bolstered by generous Covid relief programs, continue to support consumption, but inflation is slowly eroding purchasing power. Consequently, consumption growth dropped below 3% in the first two quarters of this year.

Meanwhile, the Fed’s hiking cycle has hammered residential investment (-17.8% in Q2). It will work its way through the economy and is slowing down business investment and consumption as well. It is only a matter of time before this causes another episode of negative GDP growth, this time accompanied by rising unemployment. After all, the Fed may be forced to pump up the unemployment rate in order to ease wage pressures and squeeze inflation out of the system.

Note that our view that the Fed is going to need to hike the economy into recession in order to squeeze inflation out of the economy underpins our forecast of an NBER-approved recession, i.e. a recession that includes a substantial rise in unemployment. The inversion of the yield curve not only underlines this, but also provides information on the timing of the recession, because of the usual time lag of about 12 to 18 months between inversions and recessions. The 2-10 inversion crossed the threshold of -16 bps in July 2022, which would suggest the start of the recession will occur somewhere between July 2023 and January 2024. Therefore, we have put our forecast for a recession in the second half of 2023. The increasing inversion of the yield curve suggests that this recession is also getting increasingly likely. Based on the historical relationship between yield curve inversions and recessions, we calculated that the probability of an official recession at the 12- to 18-month horizon, given a current 2-10 inversion of 39 bps, is now 70%.

While GDP and GDI are still contradicting each other on whether the US economy was in a technical recession in the first half of this year, the Fed’s Volcker-esque hiking cycle and the inverted yield curve point to an NBER-approved recession next year. Both recessions are linked to the negative supply shocks, elevated inflation, and the Fed’s policy response. The current inversion of the yield curve indicates that the probability of the second recession is 70%.