GS Forecasts 2 Quarters of Negative Global Growth and Offers Insights on COVID-19's Fatality Rate

Goldman Sachs economists Jan Hatzius and Daan Struyven have updated their global economic growth forecasts, and are now projecting two consecutive negative quarters of global GDP growth followed by a sharp rebound that leaves year-on-year growth at 2%. To quote:

Our analysis shows effects on quarter-on-quarter annualized global GDP growth of -5pp in Q1 and -2pp in Q2, followed by a rebound in the second half of 2020, leaving our full-year global growth forecast at about 2%. All else equal, this would imply a short-lived global contraction that stops short of an outright recession.

It is interesting that the current fatality rate in South Korea is only 0.48% (18 deaths over 3,736 infections) where we have aggressive testing of COVID-19 (arguably the best globally) and a much better sense of the true number of infections (ie, the denominator). Yes, this is not the steady-state fatality rate, and we need more time to control for the lead-lag effects. But it is well below the global ex China rate of around 1.5%, and closer to the China ex Hubei rate (0.84%). Also interesting that SK's average age of 42yrs is materially higher than China (37), the US (38) and Australia (37). If the assessed R0 gets reduced below 2.3 as awareness spikes and human behaviours change (washing hands, not kissing cheeks, wearing more masks), communities may become more comfortable with mitigation over containment.

Goldman Sachs also interviewed Dr. Barry Bloom, the Joan L. and Julius H. Jacobson Research Professor of Public Health at the Harvard T.H. Chan School of Public Health. He confirms what I have previously argued about the COVID-19 fatality rate, which is that the denominator is likely to be much larger than currently reported, pushing the fatality rate lower...

Allison Nathan: We keep hearing about a 2% case-fatality rate. Is that an accurate reflection of the virus’ virulence?

Dr. Barry Bloom: The case-fatality rate estimates, or the percent of infected people who die, doesn’t provide any information about how lethal this virus is at this point. That’s because those estimates basically reflect the percent of hospitalized people who die from the disease. In the case of SARS, this was around 10% and a relatively accurate reflection of the disease’s virulence because virtually everybody infected had severe symptoms. But with this virus, which can be asymptomatic, we don’t know the total number of people infected; so the case-fatality rate’s denominator is unknown. In these situations, the initial case-fatality figures are almost always frighteningly high, but then come down when the total number of infected people becomes known. For example, rates for H1N1 were terrifying in the beginning, at 4%, but ended up at around 0.07%.

Dr Bloom also had some interesting remarks on the seasonal influences on the virus...

Allison Nathan: Will the onset of summer in the Northern Hemisphere help thwart the virus? 

Dr. Barry Bloom: That has been the case for influenza, which moves seasonally. And SARS wound down over the summer. But it is not at all predictable whether this infection will be seasonally controlled. There’s good evidence that coronaviruses don’t do well in warm, humid weather, and that they thrive in cold, dry weather. But is that because of the climate, or because people are huddled closer together in the wintertime? It’s not clear. A couple of small studies on parts of China versus places like Hong Kong that are in a tropical belt show no evidence that weather differences between these regions have been a big determinant in the viral spread. But we just don’t know.

Allison Nathan: So what do you think will most likely end this outbreak? 

Dr. Barry Bloom: If the virus turns out to be seasonal, it will decline in the summertime. And if we’re lucky, as in the case of SARS, it will not come back again. But it could instead become a recurring seasonal infection, like influenza, that reappears each year; 61k people died from the flu in 2018. All that said, we think it’s unlikely that seasonality alone will stop this outbreak. So what we really need to do is reinforce whatever benefits the seasonality might bring us with public health preparedness, so that we can detect and isolate cases while the numbers are still very small, and hopefully avoid having to use massive mitigation strategies like closing schools and major public events and congregations.