In a recent survey, the chief financial officers of US companies have expressed their concern about the outlook for the US economy. 20.8% of respondents expect GDP to contract over the next four quarters.[1] Financial markets participants also feel nervous and show a mindset of ‘good news may actually be bad news’. The latest labour market report was surprisingly strong, triggering a jump in bond yields, a strengthening of the dollar and a decline in equity market futures (chart 1). These reactions reflect an expectation that the Federal Reserve may have to tighten more than previously expected. In such a scenario, the reaction of equities reflects the impact of higher risk-free rates, but also unease about the implications of more aggressive rate hikes in terms of recession risk and the earnings growth outlook.

Perhaps the current strength of the US economy is being overestimated. The Atlanta Fed nowcast for second quarter GDP shows an annualized negative growth rate versus the previous quarter of -1.9%. During the quarter, the decline in the manufacturing ISM index has played an important role in explaining the drop in the nowcast (chart 2).[2] In assessing the signal quality of a nowcast, it is important to look at the forecast errors. Chart 3 shows that until the GDP number for the fourth quarter of 2019, the forecast errors had fluctuated in a rather narrow range. Unsurprisingly, this changed with the Covid-19 pandemic, which led to important swings in activity and to large forecast errors.[3] Given the latest nowcast, it would require a very large forecast error for growth to be positive. Considering that first quarter growth was negative (-1.6%), a similar result for the second quarter would imply that the condition from the ‘popular’ definition of a recession -two successive quarters of negative quarterly real GDP growth– would have been met.

However, the National Bureau of Economic Research (NBER) business cycle dating committee, which determines the peaks and troughs in economic activity, considers a range of variables that are available monthly, rather than exclusively looking at GDP, for which data are published on a quarterly basis. Moreover, it also pays attention to the depth of the decline in economic activity. To call a recession, the decline should be significant, so a relatively small decline in real GDP in two consecutive quarters would not warrant doing so.[4] To avoid revisions, the NBER committee prefers waiting long enough before determining the start or end of a recession. Consequently, by the time that a recession is officially called, the US economy will have been in this state for quite some time already. For this reason, it is recommended to follow the evolution of the six indicators that are closely monitored by the NBER committee (chart 4). Except for real manufacturing and trade industries sales, which have been declining for three months in a row, all indicators continue to be on a rising trend, which suggests there is no imminent risk of recession yet.

[1] For a detailed analysis, see: US : an uneasy feeling, Ecoweek, 4 July 2022, BNP Paribas.

[2] For an quantification of the impact of the various data releases on the GDP nowcast, see: https://www.atlantafed.org/-/media/documents/cqer/researchcq/gdpnow/RealGDPTrackingSlides.pdf

[3] The forecast error corresponds to the difference between the advance estimate of real GDP growth and the latest GDP nowcast available when the advance estimate was published. Forecasts are available starting with the nowcast for 2014 Q2. The standard deviation of the forecast error is 0.60% for the GDP data covering the period 2014 Q2-2019Q4. Extending the sample until 2022 Q1 gives a standard deviation of 1.33%.

[4] “The NBER’s traditional definition of a recession is that it is a significant decline in economic activity that is spread across the economy and that lasts more than a few months. The committee’s view is that while each of the three criteria—depth, diffusion, and duration—needs to be met individually to some degree, extreme conditions revealed by one criterion may partially offset weaker indications from another.” (source: NBER).