The Mystery of the COVID-19 Infection Fatality Rate (IFR) Unraveled (#14) with Dr. Linus Wilson

Dr. Linus Wilson discusses and reads his new study:

SARS-CoV-2, COVID-19, Infection Fatality Rate (IFR) Implied by the Serology, Antibody, Testing in New York City

Wilson, Linus, SARS-CoV-2, COVID-19, Infection Fatality Rate (IFR) Implied by the Serology, Antibody, Testing in New York City (May 1, 2020). Available at SSRN:


 The SARS-CoV-2, COVID-19, infection fatality rate (IFR) has been hard to accurately estimate. It is a key parameter for disease modeling and policy decisions. Asymptomatic spread and limited testing have understated infections in hard to predict ways across jurisdictions. We survey serology, antibody, studies of the COVID-19 infection to find official cases are understated by an average of 25-to-1. Further, we analyze the deaths and infections in New York City to estimate an overall IFR for the United States of 0.863 percent.

…5. Conclusion


The COVID-19 pandemic has a lot of uncertainty about the ratio of deaths to total infections. That confounds the calculation of how deadly the novel coronavirus is. The serology sampling in New York City and elsewhere makes estimates of infections more reliable. We estimate that the infection fatality rate (IFR) from serology studies in nine different sampling locations in the United State and Europe is on average 0.38 percent. We analyze the data from New York City in-depth to estimate that the IFR for all ages and genders in New York City was 0.85 percent. New York City is a preferable location to estimate IFR because it has one of the highest infection rates in the world. Thus, random sampling is less prone to an upward bias in false positives. In addition, New York City’s official counts are less likely to understate deaths than in other locations in the United States. We find that the infection fatality rates from New York vary a great deal by age and gender. Females ages 0 to 17 can expect infection fatality rates of 0.001 percent while males of age 75 and over can expect infection fatality rates of 9.127 percent.”

Dr. Linus Wilson[1]

Associate Professor of Finance

Department of Economics & Finance

B.I. Moody III College of Business

University of Louisiana at Lafayette

Moody Hall, Room 253

P.O. Box 43709

Lafayette, LA 70504

(337) 482-6209

linus [dot] wilson {at} louisiana [dot] edu


COVID-19: Save Lives or the Economy? The Benefits of Social Distancing in the Pandemic and the Value of Statistical Life (#13)

Dr. Wilson reads and discusses his new paper:

Wilson, Linus, Estimating the Life Expectancy and Value of Statistical Life (VSL) Losses from COVID-19 Infections in the United States (April 19, 2020). Available at SSRN:

His the “Download this paper” button to get a free copy.

Estimating the Life Expectancy and Value of Statistical Life (VSL) Losses from COVID-19 Infections in the United States


Dr. Linus Wilson

Associate Professor of Finance

Department of Economics & Finance

B.I. Moody III College of Business

University of Louisiana at Lafayette

Theses are the views of the author alone.




Americans aged sixty or older stand to lose 153 to 222 days of life expectancy from contracting COVID-19. Over 90 percent of the U.S. population was under stay at home orders by April 2020. These social distancing measures to slow the spread of the SARS-CoV-2 or novel coronavirus have led to over 20 million new applications for unemployment benefits. Are these economic losses justified? We find the value of statistical lives lost (VSL) from an unconstrained spread of the virus which hypothetically infected 81 percent of the population would amount to $8 to $60 trillion.



Journal of Economic Literature Codes: G22, I1, I18, J31, J65, K32


Keywords: actuarial tables, mortality, death rates, CFR, COVID-19, IFR, life expectancy, SARS-CoV-2, school closures, social distancing, stay at home orders, VSL

1. Introduction

This paper attempts to open the discussion of how to model the benefits of social distancing measures in terms of the value of statistical lives (VSL) saved in the SARS-CoV-2 or COVID-19 pandemic. To do this we compare the infection fatality rates IFR’s of Ferguson et al. (2020) to the VSL from several studies. We find in figure 3 panels A and B that the costs of 50 percent of the U.S. population being infected with COVID-19 in lives lost and VSL are, respectively, between 0.659 million and 2.305 million lives lost and between $5 trillion and $37 trillion in VSL losses. The Gross Domestic Product (GDP) was only $21.7 trillion at the end of 2019 according to Mataloni and Aversa (2020).

We use U.S. Census data to control for the age and gender of the population and show how a COVID-19 infection affects an individual’s life expectancy and compares to a typical year’s mortality. Life expectancy losses of between 153 and 222 days can be expected for Americans over 60 with a novel coronavirus infection, according to figure 2, panel C. Americans younger than forty can expect to lose less than two weeks of life expectancy from contracting the virus. In figure 1, panel C, persons over fifty can expect COVID-19 to be about as deadly or up to 70 percent more deadly than a year’s mortality risks. Persons younger than forty-years-old can expect less than half a year’s mortality risk in a COVID-19 infection.

To argue against the social distancing measures, either the IFR of COVID-19 must be nearer to the low end of Ferguson et al. (2020)’s 95 percent confidence interval or the social distancing measure must be very ineffective in reducing the reproductive number, R0, of the SARS-CoV-2 virus. As to the former, a few studies suggest a much lower mean IFR than the 0.9 percent from Ferguson et al. (2020). Ioannidis (2020) argues that, after adjusting for the age of the infected on the Diamond Princess cruise ship, the IFR for the U.S. population should be 0.3 percent, which is below the lower bound of the 95 percent confidence interval calculated in Ferguson et al. (2020) and used here. Likewise, a population-weighted study, Bendavid et al. (2020), recruited people to be tested in Santa Clara County, California regardless of symptoms for COVID-19. It found infection rates were severely under-reported. They calculated an IFR between 0.12 and 0.2 percent.

On the other hand, social distancing may be effective in reducing the spread of COVID-19. R0 is the number of additional persons that an infected person goes on to infect on average. Anecdotal evidence indicates that social distancing in late March and early April 2020 has been effective. Governor Andrew Cuomo of New York State, which has had the highest number of deaths and confirmed cases of COVID-19 of the U.S. states on April 16, 2020, argued in his daily briefing in CNBC (2020) that his modeling teams believed that R0 fell from a median of 2.5 in Wuhan before social distancing and 2.2 on the Diamond Princess Cruise ship to 0.9 in New York State after the mitigation efforts. Governor Cuomo argued that his advisors’ projected hospitals in the state would be overwhelmed with COVID-19 patients if R0 was persistently above 1.2.[1]  Rocklöv et al. (2020) estimate that uncontrolled R0 for COVID-19 on the Diamond Princess cruise ship was 14.8 before social isolation and 1.8 afterward. Chowell et al. (2011) argued that school closures in Mexico reduced the R0 of the H1N1 outbreak by more than 30 percent.

This paper will not attempt to measure the costs of social distancing which, no doubt, number in the many trillions of dollars in the United States alone. By April 7, 2020, Secon and Woodward (2020) reported that 95 percent of the U.S. population was under a stay at home order that meant all but “essential” businesses were shuddered. Morath and Chaney (2020) report that by April 16, 2020, 13 percent of the U.S. workforce or 22 million workers had filed unemployment insurance claims. The COVID-19 multi-state stay at home orders, and associated non-essential business shutdowns, began with California, on March 19, 2020, according to Mervosh et al. (2020). Before the SARS-CoV-2 disruptions, the U.S. unemployment rate stood at a record low 3.5 percent in February 2020 according to the Bureau of Labor Statistics.

Eichenbaum et al. (2020) estimate containing COVID-19 “optimally” with social distancing will lead to consumption dropping by 22 percent versus 7 percent without containment of the virus. Since consumption is about 68.1 percent of GDP, according to the St. Louis Fed, and 2019 GDP was $21.7 trillion, they are arguing macroeconomic consumption losses are about (0.22 – .07)*$21.7 trillion = $3.26 trillion. The loss of freedom cannot just be measured in just macroeconomic statistics. Aggregate consumption does not measure the loss of consumer and producer surplus. There may also be long-term effects to school-age children or households facing bankruptcy that may not be captured fully in Eichenbaum et al. (2020). People losing their jobs also lose their health insurance and may be more likely to die as a result. Certainly, more empirical work can be done to estimate the actual costs of social distancing measures relative to their R0 benefits.

One press release, Yale News (2020), from the Yale Tobin Center for Economic Policy estimated the daily losses of shutdowns at $19 billion per day or about $7 trillion per year. Nevertheless, $3 to $7 trillion annually is less than this paper’s low-end estimates of the VSL losses for anything resembling the unconstrained spread of the virus infecting 81 percent of the population that Ferguson et. al (2020) project. Our low-end VSL losses are $8 trillion for an 81 percent infection rate. For the high-end estimate, the value of statistical lives lost is about $60 trillion.

Overall, the results of this paper point to sizable personal risks for individuals over sixty becoming infected with COVID-19 in terms of the increased chance of death and reduced life expectancy. We also find that the number of deaths and value of statistical life (VSL) losses are extremely high from high rates of COVID-19 infection. Thus, major economic disruptions from social distancing, stay at home orders, and school closures will be justified if the infection fatality rate estimates are reasonable. Measurement error of IFR, pharmaceutical treatments reducing the fatality rates, or a vaccine could make social distancing economic disruptions not worth the cost.


[1] See the discussion around the 24:45 minute mark of CNBC (2020).”

The 2020 Presidential Election: A Race Against Mortality (#12) by Linus Wilson

A year from the inauguration, four of the top five Democratic 2020 U.S. Presidential election candidates in the polls are in their seventies. Using actuarial data and the history of Presidential assassinations, the top two contenders, Former Vice President Joe Biden and Vermont Senator Bernie Sanders, have a 24 to 29 percent chance of not surviving to the end of a hypothetical first term. The 77 and 78-year-old men’s chances of dying before the end of a second term as POTUS are between 46 and 56 percent.

The 2020 Presidential Election: A Race Against Mortality

by Dr. Linus Wilson, University of Louisiana at Lafayette

 Dr. Wilson reads his paper on episode 12 of The Finance Professor Podcast



Download the full paper at or go to or 


Below is the introduction of the paper:

  1. Introduction


Generations of Americans have never lived to see an American President die in office. Indeed, the last President to die was John F. Kennedy who in 1963 died from an assassin’s bullet at the relatively young age of 46 years. That was over 56 years from the start of the first in the nation nominating contest in 2020, the February 3, Iowa Democratic Caucuses. You would have to go back to Franklin D. Roosevelt’s death in office in 1945 in his record 4th term to find a U.S. President who died of “natural causes”. To be alive for that event in 2020, someone would have to be 74 years old. To have voted for FDR, someone would have to be at least 96 years old. (The voting age was not lowered from 21 to 18 years old until March 21, 1971, with the ratification of the 26th Amendment to the U.S. Constitution.)

According to Panetta (2020), on January 20, 2017, Donald Trump became the oldest President to be sworn in to his first term at 70 years 222 days old. Joseph Biden, Bernie Sanders, and Michael Bloomberg, aged 77, 78, and 77 were all at least three years older than Donald Trump, who was 73, on January 15, 2020. Biden, Sanders, and Bloomberg were 1st, 2nd, and 5th in the Real Clear Politics Democratic Primary polling average on that day. The 3rd place democratic candidate, Elizabeth Warren on that date was 70 and would be older than Donald Trump was at his inauguration in 2017 if she won the election and took office on January 20, 2021.

The likely winner of the 2020 election based on betting markets will be pushing the bounds of life expectancy at birth at some time during his tenure in office. He or she will likely be the oldest person to take the oath of office of President of the United States. This paper seeks to quantify the chances that the leading candidates will die before their first or second term in office using actuarial life expectancy data.

If we project out a year out from the January 20, 2021, inauguration, Former Vice President Joseph Biden, Vermont Senator Bernie Sanders, and Billionaire Michael Bloomberg have 24 to 29 percent chance of dying before the end of their first term as President of the United States. The range depends on the age and gender of the candidate and how we account for the chances of assassination. Those same leading Democratic challengers have a 46 to 56 percent chance of not surviving two terms as President of the United States (POTUS).

Kenski and Jamieson (2010) found that perceptions of the septuagenarian Republican nominee John McCain changed through the course of the 2008 election. Voters increasing perceived him a “too old” as election day neared. Thus, the current preference for older candidates by voters may change as the 2020 election progresses. Senator McCain died in 2018 at the age of 81, according to Pitzl (2018).

If the field of viable candidates survives to inauguration day, January 20, 2021, this paper finds that there is a 16 to 21 percent chance that the man or woman sworn in will die before the end of his or her four-year Presidential term on January 20, 2025.

In section 2, we discuss our data collection and how we determined if a 2020 Presidential candidate was “viable.” In section 3, we discuss the death probability calculations in table 2. In section 4, the paper concludes.