In 2009 my colleague and I were contacted by a hedge fund manager in New York City that was interested in investing in life settlements. He is well known, but I’ll protect his identity to save him the embarrassment about to be revealed. We were flown out to NYC and visited his lavish board room where there were a half dozen people in the room accompanying the hedge fund manager.
They were considering an investment in a large portfolio of lives offered by a bank and wanted to know if we had insights on how long these people were likely to live. The investment was an over/under “bet” – they needed to decide whether there would be fewer or more deaths per month than predicted by the bank offering the investment.
A 3-page brief describing the specifics of the investment was placed before us, including most importantly the assumptions used by the bank to generate the monthly survival estimates.
Normally it might have taken some time to draw a conclusion about such a request, but in this case, my colleague and I looked up after just a few minutes reading the brief and we both said simultaneously – “bet the under”.
Shocked at how quickly we arrived at our conclusion, the hedge fund manager said, “how do you know this?” Simple, I said, “the folks at the bank used life tables that were out-of-date – this is a slam dunk”.
The tables were not just mildly out of date, they were based on data that was more than a decade old. Since we knew that death rates for the U.S. population declined rapidly in that decade, that meant the bank was overestimating mortality and underestimating survival by a lot. More people would survive each month relative to the prediction made by the bank. The speed and confidence we exuded in our answer unnerved the folks in the room more than just a little.
The skeptical hedge fund manager blurted out “what if some of these folks live forever?” They won’t, we said, they’ll all die. “What if some live to 120 or 130?” “It’s highly unlikely”, we said, “the timing of death in humans is highly predictable and only one person in history lived past 120”. The hedge fund manager then went through a litany of medical breakthroughs that he believed are forthcoming, suggesting that there is great uncertainty in survival. It’s worth emphasizing that survival analysis was not his area of expertise; it is ours.
Our response was “sure, these advances might happen, but even if they do, unless we find a way to modify the biological aging of our bodies, all the people in this investment pool will die out in a highly predictable way”. The folks in the room obviously had not read any of our published papers on the biological forces that govern human longevity and survival.
I then asked the hedge fund manager two questions that really threw him and the other people in the room for a loop.
The first was whether he could begin and end the investment whenever they wanted in a calendar year? The answer was yes, but he wanted to know why. I said, “If you bet the over, start the investment in November and end it in the month of February of some future year; if you bet the under, start the investment in March or April and end it in a future October. “Why” he asked. “Because of the seasonality of influenza deaths – which kill anywhere from 30,000 to 70,000 people every year in the U.S. (notwithstanding COVID-19). In this way you can juice up the return on your investment by using nothing more than timing.” They had no clue!
The second question was whether the bank reported on the chronological age of the individuals, or did they have dates of birth. “Why” he asked. “If you bet the over, cherry pick the cohort for people born in the same year, to include people born earlier in the year like January through March; if you bet the under, cherry pick the cohort to include people born later in the year like October through December.” The reason should be obvious – people born earlier in the year are older than people born later in the same calendar year.
At older ages, this can make a notable difference in survival. For an investment cohort of this size, using this information had the potential to juice returns even further. Jaws were dropping right and left at this suggestion. Again, they had no clue!
What happened after we left?
The hedge fund manager didn’t believe us, so they didn’t make the investment. We were able to track the survival trajectory of the investment cohort for the next year, and sure enough, there were many fewer deaths each month than predicted by the bank. This was a classic case of “adverse selection” where we had information on survival that the bank apparently either didn’t have or didn’t pay attention to.
The hedge fund manager and his associates were extremely cautious, and we got that, but we handed them a golden goose of adverse selection on a silver platter and their arrogance and misinformation led them astray. They could have made a killing on that investment if you’ll excuse the play on words. My colleague and I would have bet our retirement funds on the under.
S. Jay Olshansky is Chief Scientist and Co-Founder at Lapetus Solutions