|Written by Dan diBartolomeo of Northfield
Catching Up on the Pandemic
The coronavirus pandemic has evolved considerably since our last article came out a month ago. That article was well received by our clientele and received coverage in the financial press of multiple countries. This material is posted on our website www.northinfo.com, and a link to it also appeared in the April issue of our newsletter.
To summarize the developments in the interim period, we can simply say that the best-case scenarios have gotten much worse, but the worst-case scenarios have gotten much better. Global mortality to date exceeds 120,000 souls as of today, three times our “optimistic” scenario of mid-March. Put simply, the process of slowing the spread of infections in Western Europe and the United States has taken much longer than in the reported data on China and South Korea. While the human tragedy of this loss of life is incredible, we note that over the five months since mid-November when the first human Covid-19 case is believed to have manifested, global mortality for routine reasons unrelated to the pandemic has been approximately twenty-four million (old age, other illness, traffic accidents). Over this interval, the coronavirus appears responsible for about one in two hundred deaths globally, though the figure has approached one in two deaths in a few cities at the respective local peaks of the crisis.
On the positive side, almost all the countries where the pandemic has infected large numbers of people to date are now reporting declining numbers of new infections and stable or declining numbers of deaths. Since the typical period of the illness is a couple of weeks, it is to be expected that trends in mortality will lag infections (both known and unknown) by a similar interval. In some countries, sporadic spikes in new infections simply reflect increased rates of testing which results in a counterintuitive but positive sign. Finding persons who are infected implies there are fewer people “infected but unaware” (who are most likely to spread the infection). A more detailed discussion of testing rates across a range of countries appears later in this article.
Even stable numbers of new cases implies an end to an exponential growth process that could have created tens of millions of deaths if left unchecked. For example, a US CDC estimate of the “worst case” for US mortality was 1.7 million deaths as of late February. This horrific number was later increased to 2.2 million in a separate analysis that assumed that only 50% of the American public would abide by “social distancing” orders. Luckily, the rate of US public compliance appears to be well over 90% nationally.
The current upper confidence bound on US mortality from the well-regarded University of Washington models is at 170,000, a tenth of the CDC estimate of two months ago. The central tendency of the US mortality estimate from the UWA IHME model is now about 69,000 with minimal mortality after June 1st. This timeline is consistent with our first order model’s original estimate of May 23rd for a return to relative normalcy. While still massive in human terms, it is of the same order as the 61,000 US deaths arising from regular seasonal flu over the winter of 2017-2018, despite the existence of a vaccine.
Implications for Equity Investors
While there is always some non-zero chance that the spread of the virus could spin out of control in heavily populated countries like India and Brazil, we have not yet seen massive numbers of infections so far. Using data on mortality spikes from 1895 to 2010 associated with armed combat (diBartolomeo and Hoffman, 2015) we estimated last month that even a “worst case” sixty million deaths globally would only impact global equity market returns by around .11% per annum over a ten year forward investment horizon. As a percentage of global population, even that incredible number of deaths would be similar to a single year of one of the two World Wars at .8% of the world population. The combination of the last year of World War I and the 1918 epidemic of Spanish flu caused mortality of about 2% of the then global population. Given the apparently reduced likelihood of that kind of worst-case situation, we assert that long term investors should be minimally impacted by the pandemic if they are globally diversified. Of course, as in war, the impact of the pandemic on individual country markets may vary widely.
On a shorter term basis, we previously illustrated that our US Short Term Risk Model (one day horizon) had increased the forecast annualized volatility level of the S&P 500 from around 12% in November to over 60% in early March, and then declined to around 50% by mid-March. As of the close of trading on April 15th, this value was 39.8%.
Our estimate is different from the traded VIX contract because the estimate for the S&P 500 is built “bottom up” from the estimated volatility and weights of the individual securities comprising the index. The individual security volatility estimates are revised from day to day by changes in option implied volatility as described in diBartolomeo and Warrick (2002), https://www.northinfo.com/Documents/534.pdf which was subsequently published in 2005.
In the previous article, we used the ex-ante volatility estimate to form an inference about how long investors believed the pandemic would last. As of March 16th, that estimate stood at seven months, or an end date in mid-October. Repeating the same procedure described in our March 2020 webinar, https://www.northinfo.com/Documents/939.pdf we obtain that if investors believed this increase in annual volatility to roughly 40% was permanent equity market averages would have fallen 58% since November. The actual decline is roughly 9% through April 15th. This size of fall would be justified by an increase of the long-term risk expectation from 12% to 13.2%. Using the property that variances are additive we obtain the “investor implied” remaining length of the pandemic at five months, yielding an end date in mid-September. This result suggests that investors remain conservatively biased, as the University of Washington pandemic model does not project actual disease impact beyond August 4th.
A Return to Normalcy?
The severity of the pandemic appears to have peaked in many countries including the US and most of Western Europe. As such, there is extensive debate about how and when to reduce various forms of “lockdown” restrictions on their populations so as to minimize the already great damage to their respective economies. We will assume that no country would undertake to reduce restrictions unless they believed the worst was over and that the number of future infections would be manageable within their health care systems.
The question at hand is whether a particular nation can be confident in their views. To address the issue of “assumption dependence” we will consider two metrics of readiness for a range of nations. The first is the number of coronavirus tests performed to date as a percentage of the population. The second is the percentage of tests that have given a positive result for infection.
Our sample consists of twenty countries, including fourteen countries reporting more than 20,000 infections, plus six additional countries that were particularly notable for other reasons. China is excluded from the analysis as they have not published detailed data on numbers of tests performed. All raw data was taken from the Worldometer website during on the evening of April 15th, US Eastern time.
In terms of percentage of population tested, our fourteen nations with greater than 20,000 infections had tested 1.04% of the populations on average as of April 15th. The highest testing rate was 2.3% in Switzerland, followed by Germany at 2.06%. The lowest value was .03% for Brazil. The UK and US levels were .59% and .95% respectively. At the top of the entire list of countries was Iceland at over 11% followed by UAE at 7.75%. Japan joined Brazil at the bottom of the table at just .07%.
We believe an even more important metric of readiness to return to normalcy is the percentage of positives among tests given. Given that the supply of tests is limited, medical authorities would naturally allocate tests to people most likely to need testing (those showing relevant symptoms), where a high percentage of positives is to be expected. Testing of random samples of the general population is likely to show much lower percentages of positive outcomes.
On this measure, the worst percentage was France with more than 44% positives, followed by Brazil at 40.7%. It should be noted that the US state of New York was also over 40%. The lowest was again the UAE with just .7% of tests given producing positives. Of the fourteen nations with more than 20,000 infections, the average national rate was 20.75%, with the US and UK at 19.75% and 26.2% respectively. Many of the more rural US states have also reported lower percentages of positives (e.g. Vermont = 7.1%).
Given the high coefficient of variation across even neighboring countries and regions, the process of coordinating a gradual return to normalcy may be prolonged as a matter of the “lowest common denominator.” It would seem that an increase in the mass production of testing kits will be a key determinant of the date at which the world can declare the battle with coronavirus is at least a stalemate, if not a victory.
Corporate Credit Risk
An obvious area of investor concern during the current coronavirus pandemic is the ability of corporations to make timely payments on their bond debt and bank loans. We will highlight some metrics of credit risk that appear to be material at the current time and provide some analysis thereof.
We begin with a very familiar measure, the yield spread between investment grade corporate bonds denominated in US dollars, and US Treasury bonds of similar maturity as shown in Exhibit 1. This incremental yield is compensation to investors for both the risk of default and the lower liquidity of corporate bonds. The sample period is from the beginning of 1992 to near the current date.
The views represented in this commentary are those of its author and do not reflect the opinion of Traders Magazine, Markets Media Group or its staff. Traders Magazine welcomes reader feedback on this column and on all issues relevant to the institutional trading community.