The discounted value of human lives lost due to COVID-19 in France

Background: This study estimates the total discounted value of human lives lost (TDVHL) due to COVID-19 in France as of 14 September 2020. Methods: The human capital approach (HCA) model was used to estimate the TDVHL of the 30,916 human lives lost due to COVID-19 in France; i.e., assuming a discount rate of 3% and the national average life expectancy at birth of 83.13 years. To test the robustness of the estimated TDVHL, the model was rerun (a) using 5% and 10% discount rates, while holding the French average life expectancy constant; and (b) consecutively substituting national life expectancy with the world average life expectancy of 73.2 years and the world highest life expectancy of 88.17 years. Results: The human lives lost had a TDVHL of Int$10,492,290,194, and an average value of Int$339,381 per human life lost. Rerun of the HCA model with 5% and 10% discount rates decreased TDVHL by Int$1,304,764,602 (12.4%) and Int$3,506,938,312 (33%), respectively. Re-calculation of the model with the world average life expectancy decreased the TDVHL by Int$7,750,187,267 (73.87%). Contrastingly, re-estimation of the model with the world’s highest life expectancy augmented TDVHL by Int$3,744,263,463 (35.7%). Conclusions: The average discounted economic value per human life lost due to COVID-19 of Int$339,381 is 8-fold the France gross domestic product per person. Such evidence constitutes an additional argument for health policy makers when making a case for increased investment to optimise France’s International Health Regulation capacities and coverage of essential health services, and safely managed water and sanitation services.


Introduction
France is one of the seven major advanced economies (G7 countries). The country has an estimated population of 64.994 million; a total gross domestic product (GDP) of Int$3,161.335 billion; and GDP per capita of Int$41,637.729 in 2020 1 . In 2018, approximately 10,918,992 (16.8%) of the population lived below France's poverty threshold of €1,008 per month of disposable income 2 . France has an inequalityadjusted human development index of 0.808 and a Gini coefficient of 32.7 3 . By 14 September 2020, there were 29,182,605 coronavirus disease-19  cases in the world, including 928,281 deaths, 21,027,161 recovered cases, and 7,227,163 active cases 4 . Europe had a total of 4,080,753 COVID-19 cases, including 212,545 deaths, 2,245,583 recovered cases, and 1,622,625 active cases. On the same date, France had conducted a total of 10 million COVID-19 tests that revealed a total of 381,094 COVID-19 cases, which included 30,916 deaths, 89,059 recovered cases, and 261,119 active cases 4 . France bore 9.3% of total cases and 14.55% of total COVID-19 deaths in Europe. France's densities of 5,836 COVID-19 cases and 473 deaths per million population were higher than Germany's densities of 3,117 cases and 112 deaths per million population.
Four factors might explain the relatively large number of COVID-19 deaths sustained by France. First, there was more than two months' delay in country-wide implementation of public health interventions that could have prevented (or slowed) transmission and spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). There is evidence that COVID-19 was already spreading in France by late December 2019 5 . However, the government only banned in mid-March 2020 gatherings of more than 100 people; the opening of nonessential public establishments; anchoring in inland and territorial waters of ships carrying more than 100 passengers; opening public establishments; opening schools and institutes of higher education; all religious gatherings; and embalming of dead bodies 6 .
Second, the average of 13 International Health Regulations (IHR) core capacity score for France was 82 (on a scale of 0 to the target of 100) in 2019 7,8 , denoting an overall IHR capacity gap of 18. As shown in Table 1, the country had IHR capacity gaps of 33 in legislation and financing, 20 in zoonotic events and the human-animal interface, 20 in food safety, 27 in laboratory, 20 in human resources, 27 in national health emergency framework, 7 in health service provision, 20 in risk communication, and 60 in points of entry 9 . The latter gap denotes suboptimal capacity at ports/airports/ground crossings for coordination and communication of pandemic surveillance; and appropriate medical diagnosis, isolation and care of ill travellers. The French points of entry capacity score of 40 were lower than the average score for the World Health Organization (WHO) European Region (EUR) of 61 by 52.2%.
Second, as shown in Table 2, generally the health system indicators for France are better than the EUR averages.  The type of economic evidence reported in this paper could be an essential input when health policy-makers make a case for increased investment in optimizing the IHR capacities, coverage of essential health services, and coverage of safely managed water and sanitation services to more effectively prevent and manage the current COVID-19 pandemic and future public health emergencies [13][14][15][16][17][18][19][20][21] .
A few macroeconomic studies have estimated the impact of the COVID-19 pandemic on business conditions in France 22 . However, there is a shortage of information on the value of human lives lost due to the pandemic. This study estimates the total discounted value of human lives lost (TDVHL) due to COVID-19 in France as of 14 September 2020.

Ethical statement
The study relied totally upon the analysis of secondary data contained in the International Monetary Fund (IMF), WHO, Worldometer, and Santé Publique France databases that are freely available to the public. Therefore, ethical clearance was not required.  25 , loving, religious practice, and performing expected societal roles ends upon death. It is also true that death halts individuals' spending on the consumption of goods and services, investments, government services (including payment of fees and taxes), and imports permanently. In other words, death terminates an individual's potential contribution to the creation of national output or GDP. Following the death of a human being at any stage of life, society losses not only the statistical person's contribution to GDP but also other intangible contributions, e.g. child's joy to parents, love to family and friends, companionship, fellowship, comradeship, sharing of knowledge (written or tacit) and social values.

Study location
Weisbrod 26 suggested measuring lost human capital as a result of premature death from any cause in terms of the deceased person's discounted future earnings net of their consumption 26 .
The TDVHL in France (TDVHL FRANCE ) due to COVID-19 is the sum of DVHL among persons aged 0-14 years, 15--44 years, 45--64 years, 65--74 years, and 75 years and over [13][14][15][16][17][18][19][20][21] . Formally 13-21 : Where: DVHL i is the discounted value of human lives lost due to COVID-19 in i th age group; i=1 is age 0--14 years, i=2 is age 15--44 years, i=3 is age 45--64 years, i=4 is age 65--74 years, and i=5 is age 75 years and over; The DVHL i in each of the five age groups was calculated using the following formula 13-21 : Where: to the last years of life lost (t=n); Y 1 is the GDP per capita of France; Y 2 is the current health expenditure per person in France; Y 3 is the average life expectancy at birth in France; Y 4 is the average life expectancy at the onset of death in the i th age group; Y 5 is the total number of COVID-19 deaths in France as of 14 September 2020; Y 6 is the proportion of total COVID-19 deaths borne by those in the i th age group. The baseline for the analysis is 2020.

Data and data sources
The data analysed in this paper and the sources are contained in Table 3.

Data analysis
The human capital model was analysed using Excel 2016 software (Microsoft, New York). The study reported in this paper replicates steps that were developed and applied in our recent valuation of human life studies related to COVID-19 [13][14][15][17][18][19][20][21] .
Step 2: Estimation of the undiscounted years of life lost (UYLL) from COVID-19 in France between December 2019 and 14 September 2020.
(a) Calculation of average ages of onset of death (AAOD) from COVID-19 for each of the five age groups. This entailed taking simple averages for each age group, e.g. for AAOD for 0-14 age group = (0+14)/2 = 7 years.
(b) Calculation of YLL by one person who died of COVID-19 in the age group as the difference between national average life expectancy for France and the AAOD for the specific age group. For example, YLL by a person dead in the age group 0-14 years = national average life expectancy for France (83.13 years) minus AAOD for the group (7 years) = 76.13 years. Thus, the YLL for one person who died in each of the five age groups was obtained similarly (see Table 4).
(b) Calculation of the discount factors applying the discount factor formula: ( ) . The discount factor for first YLL = (e) Total DYLL in each age group = DYLL per deceased person in age group multiplied by number of persons who died in age group. Therefore: UYLL in 0-14 age group = 29.80759833 × 4.597402597 = 137.04 discounted YLL (see Table 5).
Step 6: Calculation of the share of TDVHL accruing to the 13 regions and five territories of France 29 through multiplication of the TDVHL by proportion of COVID-19 deaths sustained by specific region and territory.
Step 7: A one-way sensitivity analysis was performed to evaluate the effect of changes in discount rate and the average life expectancy on the estimated TDVHL. This entailed recalculating the HCA model with (a) 5% and 10% discount rates 13 Figure 1 depicts the share of TDVHL across the 13 regions and five territories of France.

Distribution of the total discounted value of human life by regions and territories.
About 80.3% of the TDVHL accrued to only five regions of France, i.e. Auvergne-Rhône-Alpes, Bourgogne-Franche-Comté, Grand Est, Hauts-de-France, and Ile-de-France. Grand Est, Hauts-de-France, and Ile-de-France regions alone accounted for 66.18% of the TDVHL. The territories combined made up less than 1% of TDVHL. Findings from the HCA model: assuming 5% and 10% discount rates holding national life expectancy and other parameters constant Table 8 presents the effects of the application of 5% and 10% discount rates on the TDVHL due to COVID-19 in France.
Practical implications of the study findings Evidence on the economic value of human lives losses associated with COVID-19 may be useful to the Ministry of Public Health when advocating within the Government of France for sustaining or increasing investments into the national health system, disease surveillance and response system (including IHR core capacities), and other systems (e.g. water and sanitation) that tackle social determinants of health in the pursuit of the United Nations Sustainable Development Goal 3 to "Ensure healthy lives and promote well-being for all at all ages" and Goal 6 to "Ensure availability and sustainable management of water and sanitation for all" (p.14) 34 . Of course, the economic evidence reported in this paper is meant to complement the International Bill of Human Rights obliging the Government of France to assure every citizen's realization of the right to life (Article 3) and to "..a standard of living adequate for the health and well-being of himself and of his family, including food, clothing, housing and medical care and necessary social services… (Article 25)" (p.76) 35 .

Suggestions for further economic studies
• Comprehensive studies on the multidimensional impact of COVID-19 on household's wellbeing.
• Wide-ranging studies on the multi-sectoral impact of COVID-19 pandemic once the pandemic is eradicated.
• Consumer choice behaviour analysis in respect to uptake of various COVID-19 prevention interventions, e.g. handwashing with soap, use of safely managed drinking water and sanitation, use of face masks, and patronage of alcohol bars during COVID-19.
• Economic evaluations of cost and consequences of preventive interventions (including personal hygiene, physical distancing, safely managed human waste disposal, contact tracing, vaccines), diagnostics, and potential treatments for COVID-19. Where feasible, cost-effectiveness, cost-utility, and costbenefit analyses ought to be designed and conducted alongside ongoing and envisaged clinical, and effectiveness randomized trials 30,36 .
Limitations of the study First, HCA has been criticized for valuing non-market contributions to society at zero dollars 37 . For instance, traditional HCA values YLL among children below 14 years 38 , retired (62 years and above) 38 , homemakers (not employed outside the home), unemployed, and severely handicapped. In order to avoid discrimination against these vulnerable groups, which goes against the 1948 United Nations Universal Declaration of Human Rights 35 and the Constitution of the World Health Organization 39 , YLL at all the age groups were valued at equal net GDP per capita.
Second, the current study did not compare the costs and benefits of a raft of alternative preventive community-level interventions implemented by the French Government and citizens to limit transmission of COVID-19, e.g. bans on gatherings of more than 100 people, all religious gatherings, all travel, ships carrying more than 100 passengers, and embalming; closure of most public establishments, all schools and institutions of higher learning; and mandatory mask-wearing in public places 40 . It was also outside the scope of the current study to appraise the costs and benefits of various options for diagnosis of COVID-19, contact tracing, quarantine, and management of persons who test positive for COVID-19 .
Finally, Stiglitz, Sen and Fitoussi 41 have criticized GDP for not measuring economic wellbeing (or quality of life), ignoring income inequalities, and disregarding environmental damage caused by production processes.

Conclusion
The discounted value per human life loss of Int$339,381 is 8-fold the GDP per person of France. Such evidence constitutes an additional argument for health policy makers when making a case for increased investment to optimize IHR capacities, and coverage of essential health services, and safely managed water and sanitation services. The other rationales for increased investments in the development of resilient healthrelated systems include the fact that a pandemic, such as COVID-19, can trigger health systems and socioeconomic crises of significant magnitudes 42 ; and also the fact that every human being has the right to life 35 .

Data availability
All data underlying the results are available as part of the article and no additional source data are required. Introduction Paragraph 3 line 2 First, there was more than two months' delay in country-wide implementation of public health interventions that could have prevented (or slowed) Comment: This should be revised to read "First, there was more than two months' delay in country-wide implementation of public health interventions that could have prevented or slowed " (since prevention is not the same as slowed)

Methods
Step 3: Discounting of the years of life lost (DYLL). Comment:The authors indicate here that the discount of 3% is chosen based on their previous work but they cite other articles as references as well. Thus, the article should rather indicate that "(a) A discount rate of 3% was chosen because it has been used in previous COVID-19 related economic studies 13-21 , " Step 6: Calculation of the share of TDVHL accruing to the 13 regions and five territories of France 29 through multiplication of the TDVHL by proportion of COVID-19 deaths sustained by specific region and territory. Comment:In the introduction of the study it was not indicated that 13 regions and 5 territories of France will be used to show the share of these to the TDVHL. It thus pops up here with a previous mention of that in the introduction or even in the methods where information on the study site could be described. The characteristics of these territories may affect the outcomes of the study.
Step 7: A one-way sensitivity analysis was performed to evaluate the effect of changes in discount rate and the average life expectancy on the estimated TDVHL. Comment: The authors chose to use the world average life expectancy and the world highest average life expectancy of 88.13 for women from Hong Kong for the sensitivity. Given that the authors began comparing France with the rest of EU, I was expecting that the authors will first conduct the sensitivity analysis using average life expectancy for the Europe before comparing with the rest of the world.

Discussions
Contrasting of study findings with those from other countries: Comment:The authors refer only to their previous studies in comparing and contrasting their findings but I think the paper will benefit greatly by comparing and contrasting the findings with other studies in this same field as well. Second, the current study did not compare the costs and benefits of a raft of alternative preventive community-level interventions implemented by the French Government and citizens to limit transmission of COVID-19.

Recap of findings
Comment: This sentence above should read "Second, the current study did not compare the costs and benefits of a raft of alternative community-level interventions implemented by the French Government and citizens to limit transmission of COVID-19.
General Comment:This paper highlights issues that might explain the years of lives lost due to Covid-19 in France that may appeal to policy makers to make great investments in interventions needed to reduce deaths lost and increase GDP. More elaborately in this paper is the clear and robust description of the analytical methods for calculating the lives lost. This step-by-step methodological approach could easily be replicated in many countries and sites to produce more evidence that will guide policy makers in the prioritizing investments in Covid-19 to curb the pandemic.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate? Yes