Medical Progress and Mortality Derivatives

(Originally published on LinkedIn Aug 15 2014)

Introduction

In 2011, Swiss Re reported that $17 trillion of the total $23 trillion in pension-fund assets worldwide were exposed to longevity risks. This illustrates how pension funds worldwide are heavily exposed to the liability of paying out more because of people’s longevity — generally defined as how long they live beyond 65.

The Journal of Alternative Investments (http://www.iijournals.com/toc/jai/current) has devoted almost all of its Summer 2014 issue to this concern. It examines the risks that pension funds face if their beneficiaries outlast the money. This LinkedIn post reviews an important paper within this issue of the journal by Nakada et al (hereafter simply “Nakada”, singular) pithily entitled: “The Fundamentals of Longevity Risk.”

The objective of Nakada’s paper, and of the Summer 2014 issue as a whole, is to review the opportunities for trading and investing in longevity risk and longevity (or mortality or “death”) derivatives. In simple terms, opportunities to buy, sell, swap, speculate and hedge for or against the risk that pension liabilities will grow if people’s life expectancy grows. In the next sections, we first summarize Nakada’s analysis and then proceed with our own subjective and qualitative assessment.

Getting to the Roots of Longevity Risk

With one eye on historical records and the other on future drivers of mortality improvement, Nakada builds a model of longevity (i.e. how long you live beyond 65) risk.

There was hardly any increase in longevity from 1850 to 1900 and only very slow improvement until World War 2. The pace began to accelerate from that point, owed to “a golden age of medical advancement”. Public health measures, vaccines and antibiotics largely conquered infectious disease. From around the 1970s, mortality from cardiovascular disease began to slow as well, through better medical understanding and control of hypertension and atherosclerosis.

Drawing from these historical experiences, Nakada develops a model based on three variables — Lifestyle Changes, Medical Interventions and Health Environment — to investigate at what pace future longevity improvements may take place. Longevity improvements may derive either from further gains from known developments or from completely new future developments — particularly in the area of cancer which plays an important role in current mortality. Nakada finds the results from such a modeling exercise do not differ much from a simple extrapolation of historical trend, at least over a 5 year horizon.

Nakada however points out that, over a 5–20 year horizon, the evolution of current trends as well as completely new treatment paradigms could upend simple extrapolations. He identifies regenerative medicine and anti-aging research as two nascent technologies that could eventually increase longevity, His probabilistic view, developed from the spread of expert viewpoints, is that such an outcome lies considerably in the future. However, he emphasizes that his model retains flexibility to dynamically incorporate future breakthroughs in either field.

Moreover, Nakada understands that longevity changes could also be driven by extreme developments that today seem unlikely i.e. tail risks. He runs his model under 3 different extreme scenarios: (i) 80% of the population becoming obese by 2050 (ii) a reduction in cancer mortality to one-eighth its current level through lifestyle changes and breakthroughs in immunotherapy and DNA editing (iii) rapid progress in anti-aging technologies that soon enable people to age less in biological-years than calendar years. As it turns out in his modeling, the actual impact on longevity is muted under all these scenarios.

With the above analysis, Nakada concludes that longevity risk can be modeled by the sophisticated investor through a study of health data, behavioral trends and medical technologies. He argues that this longevity risk is not positively correlated with stocks and bonds and even has a well-behaved tail. Accordingly, as pension funds try and off-load the longevity risk they carry, an investor versed in the fundamentals of longevity risk could earn a new form of alpha.

Underestimating Tail Risk

Without having directly reviewed Nakada’s model or the kind of research and expert opinion that he has surveyed, we make some qualitative (admittedly subjective) observations.

First, Nakada may be heavily underestimating the pace of each of public health, lifestyle and in particular medical advances. As regards public health, developed country governments are getting increasingly sophisticated in their public health interventions. With growing realization everywhere that preventive care is eventually a much smaller burden on the exchequer than curative care, governments are increasingly and proactively reorienting their spending towards the latter. Going beyond simple hygiene and basic restrictions on smoking and high fat foods, governments are building national health and genomic databases and responding aggressively and flexibly to new infections. Simultaneously, lifestyle and behavioral changes are also spreading rapidly. The electronic and social media create powerful demonstration effects in favor of healthier living. More than that, the profit motive will also drive the wellness and fitness industries to promote healthier behaviors. As individuals and governments learn from each other and best practices become the norm, the cumulative effect on longevity can well become very powerful.

But most of all, Nakada may be truly underestimating the accelerating progress of medical science in particular over the last 10–15 years. As the Director of the NIH Dr. Francis Collins puts it, the arrow of progress we are riding in biomedicine took flight only some 40 years ago and is now traveling faster and further than ever before. In the 1970s people felt that genetics had something to offer to human medicine but only 100 years later. Not only the pace of progress but the accuracy of medical science has also vastly increased — for the first time medical science is dealing with a clear list of targets that are much more smart-bomb opportunities than carpet-bomb opportunities. In the view of many experts, the rate of medical progress over the next decade or so will be determined not so much by science but on how much funding will be available and how much flexibility the regulators show in shortening the approval cycle. With equity markets having generously funded biotech companies over the last couple of years and with regulators more willing to relax constraints (e.g. most recently with ebola), there are reasons to believe that a further of acceleration of medical science — in areas such as cancer, anti-aging and infections — may already be underway.

Put together, changes in public health, lifestyles and medical technologies could individually or collectively cause sudden and sharp increases in longevity. This could undermine the comfortable sense that longevity risk has a well-behaved tail.

Correlation with Macroeconomic Cycles

Nakada suggests that longevity risk has little correlation with stocks and bonds, thereby making it a great asset class for diversification. This argument may have some surface appeal.

However, detailed research of the relationship between macroeconomic cycles and shows a very complex picture and one that is becoming even more messy and confusing.

The subject has been studied by no less eminent an authority than Bob Fogel, who proceeded to deliver his Nobel Lecture on the subject. Several past studies of data from across developed countries showed that mortality rates increased strongly during upward cycles in the economy, and decreased during downward cycles. The causes for this strange but previously robust relationship are not well understood. One hypothesis is that when employment rates and the amounts of workload increase for the working population, there is less time for informal care-giving by younger relatives and friends for the chronically ill or oldest old.

Whatever the reason, more recent analysis has seemed to suggest that mortality is shifting from strongly procyclical to being only weakly related to the economic cycle — importantly, though, only in aggregate. When disaggregated, deaths due to cardiovascular disease and transport accidents still appear to be procyclical but strong countercyclical patterns of cancer fatalities and some external sources of death (particularly those due to accidental poisoning) have emerged over time. It is suggested that this may partially reflect the increasing influence of financial resources which can be used to obtain sophisticated and expensive treatments that have recently become available.

While the exact causal mechanisms will need to be unscrambled and fleshed out at a granular level, what seems clear is that Nakada’s view that stocks and bonds are not correlated with longevity risk is not well established. Macroeconomic cycles appear to impact mortality (and therefore longevity) as much as they impact stocks and bonds. Accordingly, longevity risk cannot escape being correlated with stocks and bonds. Therefore, any investment into longevity risk must reckon with how its underlying factors — in aggregate as well as disaggregated aspects — relate to stocks and bonds. At this point of time, with confusing data and evolving expert opinion, this must be classified as a big unknown.

Longevity Derivatives and Investor Interest

The classic hedge to be sold to pension funds would be that if pensioners live longer than expected, investors compensate the pension fund for the additional costs it faces. However, if pensioners die sooner than expected, investors profit.

However, a number of questions remain. To start with, as mentioned above, longevity risks and their correlation with macroeconomic cycles and other asset classes are as yet still poorly understood. Of course, Nakada has made attempts to address this through his paper and presumably the modeling and research work that backs it up. Other models known are the Lee-Carter model that has been by the US administration for federal budgets and for social security computations. More recently, Cairns-Blake-Dowd have also developed another model.

However, even if such risks could be modeled and priced, do we know what type of investors could take such risks? What is known is that many insurers were saddled with a lot of the risk of the subprime and asset-backed derivatives risk when the global financial crisis broke: they will be understandably wary of new asset classes. As for hedge funds with experience in say catastrophe bonds, many will find pension funds among their own LPs, so will have no capacity to absorb longevity risks without alienating such LPs. Some endowments, charities and family offices with longer horizons and tolerance for illiquid assets may conceivably find appetite for longevity risk but of course they too will want to be careful about experimenting with an unproven asset class.

Summary and Conclusions

Nakada and other papers in the Summer 2014 issue of the Journal of Alternative Investments make a valuable contribution to an important debate and nascent market phenomenon: probing the roots of longevity risk. Nakada frames the issues very well by breaking longevity risk into component factors: public health interventions, lifestyle changes and medical technology advances. Given the comprehensiveness that Nakada brings to the subject, it is fair to expect that the financial model he develops will be adding depth and richness to the Lee-Carter and Cairns-Blake-Dowd models already in existence.

That said, it would be interesting to engage further with Nakada’s work to evaluate if he fully incorporates the speed and enormity of advances in public health interventions, wellness-oriented lifestyles and in particular the ongoing acceleration of medical science. A first impression is that these have been underestimated, as have the tail risk that derives from there. Separately, one is not as sanguine as Nakada that macroeconomic risks have limited correlation with longevity risks. The economic literature suggests a far more complex and nuanced picture: while mortality from cardiovascular diseases increases in good times, for other diseases, e.g. cancer, longevity and cyclical upswings vary together. In light of the above, as of writing, a vibrant market in longevity/mortality derivatives seems distant because investors while need time to fully comprehend such risks and make bold to invest in them.

Nevertheless, the Journal of Alternative Investments needs to be commended for highlighting the subject. Nakada’s work is illuminating and very useful to those of us who work at the crossroads of healthcare and finance. One can expect that, driven by the need to address a genuine hedging need as well as in the search for alpha, many investors will be motivated to take a deep look at this emerging asset class.

References

Biffis and Blake (2009) “Mortality-Linked Securities and Derivatives”

Bloomberg (2011) “Death Derivatives Emerge from Pension Risks of Living Too Long”, May 17

Cairns, Blake and Dowd (2006) “A Two-Factor Model for Stochastic Mortality with Parameter Uncertainty: Theory and Calibration”

Collins (2013) “Politics on the Frontiers of Science: Major Breakthroughs are possible in neuroscience, cancer, AIDS and Parkinson’s — if Congress learns to set priorities”: Saturday Interview with Rago, The Wall Street Journal, November 8

Fogel (1993) “Economic Growth, Population Theory and Physiology” Nobel Lecture, December 9

Girosi and King (2007) “Understanding the Lee-Carter Mortality Forecasting Method”, September 14

Nakada, Breaux, Honarkhah, Hornsby, Tolla and Vessenes (2014) “The Fundamentals of Longevity Risk”, The Journal of Alternative Investments, Summer 2014 Issue; see also other papers on Longevity Risk in the same issue

Read (2014) “Beyond the Boomers: Facing the Global Challenges of Aging” August 12 www.linkedin.com

Rolden, van Bodegom, van den Hout and Westendorp (2014) “Old age mortality and macroeconomic cycles” J Epidemiol Community Health 68:44–50

Ruhm (2013) “Recessions, Healthy No More?”, (Revised) August