Fight Aging!

Bloodless, heartless calculations of the value of your life are constantly taking place behind the curtains that society politely draws over some of the uglier realities of the human condition. Interactions with insurance companies might be the most visible signs of these calculations, but this is the tip of the iceberg. Humans assign value instinctively; to value is to be human. We don't just value objects, we value our lives, we value the lives of others. Based on an analysis of our actions, i.e. revealed preferences, one can estimate monetary equivalents to those life valuations and how they shift with time and circumstances. These estimates are produced constantly, and widely used in policy and industry circles, whether or not we might agree with them.

For all that it makes many people uncomfortable, this is an interesting topic, and one that can help in understanding why it is that the powers that be behave as they do in circumstances involving aging, medical research, medical regulation, centralized control of medical services, entitlements, and so forth. It is, I suspect, considerably easier to harness technological progress in order to reduce the cost of intervening to save a life or improve a life than it is to change human nature such that life and quality of life is valued more highly. We do not live in a perfect world, but we can at least work to make it better!

Valuing life over the life cycle

The COVID-19 pandemic has been associated with considerable economic and personal tolls. Two of the motivations often invoked to justify these interventions have been (i) the collective duty to protect society's most vulnerable members, and (ii) the consequences of pandemic-driven excess demand for medical care. The allocation of scarce medical resources in situations of excess demand for life support raised the specter of uncomfortable medical triage decisions between saving one person against another.

These considerations highlight the fundamental questions of (i) how to value longevity in general and how to adjust this value to account for (ii) the personal characteristics such as age, health, labor market and financial statuses, as well as (iii) the characteristics of the changes in death risk (e.g. magnitude, beneficial vs detrimental, permanent vs temporary, longevity mean vs variance). Indeed, the substantial costs to society of COVID-19 measures should be contrasted with the presumably large economic value of those lives saved by intervention. Moreover, the reallocation of such consequential financial and medical resources to the pandemic raises the issue of the long-term arbitrage of addressing a single illness at the potential expense of others. Put more bluntly, the delicate question of which lives should be prioritized - contemporary COVID-19 infected vs other current or future illnesses, young vs old, healthy vs unhealthy, rich vs poor - was brutally unearthed by the pandemic.

Addressing the first question of life value measurement involves proxying the (non-marketed) value of longevity through a theoretical (shadow) price. A natural candidate is the marginal rate of substitution (MRS) between additional life/mortality and wealth which, at the optimum, will capture the relative price of longevity. A second related alternative is the maximal willingness to pay (WTP) or the minimal willingness to accept compensation (WTA) for changes in life expectancy. The Value of a Statistical Life (VSL) is an infra-marginal approximation to the MRS that sums the willingness across agents to calculate an aggregate WTP or WTA to save someone, i.e. an unidentified (statistical) member of the community. Personalized life values can be assessed from the market value of an agent's foregone net revenues such as in the Human Capital (HK) value. Despite its usefulness in wrongful death litigation, the HK value is arguably less relevant for non-working (e.g. retired or disabled) agents, and therefore imperfectly applicable for society's more vulnerable members. Identified values can alternatively be recovered from the agent-specific MRS, WTP and WTA. An extreme example, potentially useful in both litigation and terminal care decisions, is a person's two Gunpoint (GPV) values: her willingness to pay to prevent and to receive compensation to accept imminent and certain death which gauges a specific person's willingness to save or lose her own life.

Secondly, adjusting identified life values for personal characteristics involves charting how ageing processes (e.g. the life cycles of wages, morbidity, and mortality risks, and finite biological longevity bounds), quality of life (e.g. health status, mix between market activities such as consumption and non-market ones, such as leisure) and disposable resources (financial wealth, labor income) affect an agent's shadow price of longevity. Third, since life values are to be inferred from changes in death risk exposure, the distributional characteristics of these changes are relevant. Indeed, whether the changes correspond to small or large, temporary or permanent increases or decreases in mortality risk and whether those changes affect the mean and/or the variance of longevity will alter the individual and societal willingness measures, and therefore the degree of substitution between personalized lives. For example, how do we compare the possibly large contemporary beneficial gains of intervention on the survival outcomes of currently infected persons versus the possibly small, but long-term detrimental increases in the risk of dying of agents whose interventions have been postponed is certainly relevant to both groups and to society as a whole.

In the model of Revealed Preferences presented in this paper, ageing is associated with (i) lower WTP/WTA per given change in death intensity, but (ii) higher willingness per given change in expected longevity. Indeed, the combined influence of falling wages, increased morbidity and mortality risks exposures and eroding remaining horizon imply falling net total wealth. Moreover, increasing mortality risks induces lower marginal (and therefore continuation) utility, although the mortality effects are dampened by age. Finally, the longevity returns of changes in survival fall in age, i.e. elders require much larger mortality changes to attain a given change in expected longevity. The combination of the three factors induces a lower willingness for changes in survival risk, but a higher willingness for expected longevity changes for older agents. The WTP to avoid certain imminent death falls from 1.75 M$ at 25 to 1.15 M$ at 65, whereas the WTA to accept death is unsurprisingly higher and falls from 4.13 M$ at 25 to 1.92 M$ at 65. Conversely, the WTP/WTA associated with changes in expected longevity increase in age, although the effects of ageing are weaker. The WTP per additional life-year through one-shot changes thus increases from 211 K$ at age 25 to 220 K$ at age 65.