There is no need for a meeting of the NBER Committee and its definition of a recession as “a significant decline in economic activity” to conclude that the US economy has already entered into a recession. The second estimate published by the BEA today for the first quarter of 2020 is only slightly revised down at -5% from the first estimate published a month ago which showed that the US economy contracted by -4.8% in Q1 2020 over Q4 2019 on an annualised basis. Analysts are now focused on second quarter figures which are likely to show the full impact of the lock-down and the business shutdown imposed in April.
Almost one year before the outbreak of the coronavirus crisis the inversion of the yield curve, in May 2019, had already indicated that the US economy would enter into a recession within a few months. Here is what Chuck Jones, a senior contributor to Forbes wrote in December 2019.
Probably because the Fed has become more accommodative, investors seem to have come down with amnesia that there is a lag between the inversion of the yield curve and the start of a recession. If history is repeated a recession could start between January and November 2020.https://www.forbes.com/sites/chuckjones/2020/12/31/2019s-yield-curve-inversion-means-a-recession-could-hit-in-2020/#57a9543b4229
In order to quantify the outlook for the quarters to come, taking into account the prevailing level of high sanitary and economic uncertainty, we estimate over the 1990-2020 period an econometric combining quarterly GDP with the monthly unemployment rate and the monthly average for the VIX equity volatility index. There is a well-documented robust economic relationship between GDP and unemployment (cf. the Ökun law). In addition, the impact of financial volatility and stress as a predictor of current and near-term GDP growth has been well documented by different authors as of late. One of the most interesting recent papers on this subject has been published in April 2019 by Tobias Adrian et al. under the title “Vulnerable Growth”. This sparse econometric model is primarily intended for nowcasting GDP growth over the next quarter, taking into account higher frequency data for unemployment and financial volatility.
Our model suggests that US GDP might fall by an expected -29% in the 2Q 2020 – although it may fall by as much as -50% at the lower bound of our 95% confidence interval – and that the level of unemployment might take up to three years to return to its pre-crisis levels. There is no question that there will be a rebound in quarterly GDP in Q3 from its record low level in Q2 as the lock-down is being gradually lifted across all the United States. Arguably, this is not captured in the central tendency of our nowcasting model although the rebound is still accounted for at the upper end of the 95% confidence level range. As for unemployment, the central projection of our model is likely to prove more accurate than for GDP looking beyond Q2, despite there being an incredibly high level of uncertainty, as is illustrated by the very broad 95% confidence interval, with a significant risk of unemployment overshooting above the central projection of the model, ceteris paribus.
US GDP and Unemployment nowcasting results
Our results are consistent with the median forecasts for GDP and unemployment reflected in Philly Fed’s survey of professional forecasters released on May 15. According to these median forecasts, Real GDP growth could rebound by 10% in Q3, on a QoQ basis, before petering out in the following quarters. However, the annual GDP growth projection for 2020 might prove too optimistic.
These results cast additional shadows on the assumption of a V-shape recovery for the US economy, which has already been abandoned by the Staff and Board of Governors of the Federal Reserve System. This can be grasped from a careful reading of the minutes of the FOMC meeting held on April 28, which we have previously analysed, or in a more straightforward fashion from Fed Chair Powell’s recent interviews (cf. the transcript of the latest 60 minutes interview).