Life settlement investors don’t buy a life insurance policy without the data and analytical support of a life expectancy analytics firm. Life Risk News spoke to Chris Conway, Chief Development Officer at ISC Services, to learn more about how firms like his are adapting and changing to better model longevity risk for their clients.
LRN: Chris, let’s start with the obvious. Modelling life expectancy is a bit like economists making forecasts – it’s almost impossible to predict. So, what certainties – if any – can life expectancy analysts offer to life settlement investors?
CC: The most important thing to understand about life expectancy underwriting is that there are huge differences between “macro-longevity” and “micro-longevity.” Life expectancy underwriting in the life settlement sector is first and foremost an exercise in evaluating “micro,” as in individual, longevity risk. Not only are we trying to forecast the lifespan of a single individual, but we are generally looking at individuals who are approaching or already beyond the age of insurability. This is important because the life insurance industry is not issuing new life insurance to the members of the population we are evaluating; therefore, the data life insurance companies have in very large volumes, that spans many decades, does not include the information our sector uses to estimate life expectancy.
So, to provide investors with some degree of certainty, which we would define as consistency, we use a methodology that is very similar to that used by the life insurance industry. For example, we have a formal proprietary underwriting manual that provides research-driven guidance to our underwriting staff; we have policies and procedures that are consistently applied to the escalation of cases to our secondary and clinical review processes, and we are continually educating and training our underwriting team to apply our methodology uniformly and consistently. We do this by ensuring that our system enables our underwriting team to observe the entirety of the information we have about each subject insured over time. Lastly, and perhaps most importantly, we are focused on the scalability of the platform. In other words, we are focused on building a track record for the long term based on a methodology and platform that is not tied to an individual underwriter or even a small group of expert clinicians. We can’t offer consistency to our clients if what we do is tied to personalities as opposed to process, technology, and consistency.
LRN: No two people are the same, of course, and each insured individual will have specific health-related idiosyncrasies. So, where does accurate-as-possible life expectancy modelling start? Is it at the individual level or the population level, and why?
CC: The concept of accuracy in our sector has been misconstrued throughout the history of life expectancy underwriting. It’s understandable, even desirable, and it is frequently touted by some using a single metric, “A-to-E” [actual to expected], but again, in our view, consistency and scalability are far more important in the long term. Each individual case is indeed unique and involves an array of variables that make the work we do even more challenging. For example, we get hundreds or even thousands of pages of medical records in PDF format, for each insured. This “raw data” must be digitized, organized and summarized – not just indexed or sorted, but truly distilled into a meaningful summary of the information that matters for underwriting purposes.. Through that process, the data becomes information, and with the information derived from the data, we can seek to “know” what is and is not meaningful with respect to the insured, their lifestyle and capabilities, and their medical health. So, the process starts with the input we receive from clients. If the data is incomplete, redundant, indicative but not decisive, etc. then the work we do becomes more forensic, deeply investigatory, and deductive in nature. The data available about the population of which our insureds are members is relevant for purposes of comparison, but the unique characteristics of the individual are far more important to our work and our clients, since that is how they deploy capital and take risk.
LRN: Tell us about some of the improvements in the life expectancy analysis arena and drivers of these improvements, and why this should provide confidence to investors in life settlement funds.
CC: The improvements made in our sector are less about the models used to evaluate individual lives and more about the degree to which advances in medical science, healthcare and technology are changing either the guidelines applicable to evaluating older aged insureds, or the speed with which data can be transformed into useful information for the underwriting exercise. In addition, the more lives we underwrite and the more outcomes we experience over time, the more complete and clear the body of information and knowledge we have to refine and focus our underwriting processes. In the last decade or so, the impact of these changes has been that mortality tables focused on the population we serve have been developed and continue to be refined, technology has enabled parts of the process to be accelerated and standardized without a decrease in specificity, utility or a loss of context, and again, the concept of consistency can now be pursued and validated against a much more significant body of work to further refine the risk assessment process. As a result, investors can take more comfort in the viability of the asset class itself.
LRN: Are there any developments, technological or otherwise, that makes you think that life expectancy forecasting can improve further? Indeed, just how accurate can this really get?
CC: I think the work we do will develop much further over time in the sense that the outcomes we have predicted and continue to predict will eventually become known, and that will tell us whether we were “right,”. However, I’m not talking about actual-to-expected results so much as the degree to which the drivers of our predictions, the debits and credits we apply to each life through the underwriting process were or were not correct relative to the impairments involved. The more underwriting we do and the more outcomes we experience, the more we will be able to fine tune the considerations we apply to the evaluation of unique lives. However, no matter how finely tuned the instruments we build are, as they are “played,” in other words as time passes and medical science progresses, these tools will always be going “out of tune.” I don’t think we’ll see absolute certainty anytime soon, but I do think that there is room for improvement.
LRN: Lastly, Chris: Accurate valuation of policies, for which life expectancy analysis is a most critical component, is the number one driver of performance for life settlement fund managers. To what extent are firms like yours being included in the due diligence process undertaken by end investors and what are your thoughts here?
CC: Disappointingly, few have. And that’s something I find odd, because by talking to us to try and understand what we do would help them to ask the fund managers better questions during their due diligence process.
I think that one of the reasons for the lack of engagement by the end investors could be that not many of them know how to go about conducting due diligence on companies that do things they don’t know a lot about. But in life settlements, I’d argue that doing this is very important, because there are quite a few underwriting companies in our industry, and each has its own way of doing things, its own methodology and practice, its own “view” as to how this work should be done because this is a field that has not yet clearly defined itself or its own standards.
One thing I think is very important to highlight that investors should know is that different managers use life expectancy reports differently, and how they use the information we provide is critical for investors to understand. Investors often think the underwriting work is driven solely by actuarial factors, that the “calculator is the key,” but it isn’t. Calculators are driven by input; they don’t drive input. An incorrect risk assessment applied to the “best” table will produce an invalid result. Also, mortality tables are modified relatively infrequently, and changes must be predicated on empirical analysis conducted on work done over long periods of time, not on a whim, inference, or based on undocumented perceptions of “experience.” The truth is that the risk assessment part of what we do determines the input and the calculators themselves all generally do the same thing in the same way. However, all underwriters do not assess risk the same way and investors need to understand the bases for a given underwriters approach.
Evaluating a life expectancy underwriting company takes time, experience, and a fairly strong background in underwriting, not math. The tables matter, of course, but the underwriting risk assessment process is the critical first step. Underwriting companies are a significant and influential part of the life settlements infrastructure, and I’d urge investors to engage with life expectancy providers as part of their manager due diligence process more frequently.
Chris Conway is Chief Development Officer at ISC Services
