Captive.com logo

Captive Insurance News

Captive-Trends 2018

Captive Insurance Issues and Trends 2018

A FREE 23-page special report courtesy of Captive.com

Dig deep into important issues and trends in captive insurance. Download this FREE special report featuring practical knowledge and insights from eight respected captive insurance thought leaders!

Download FREE Report Now

5 Questions Concerning Your Captive's Asset Allocation Model

Question marks 600x300
April 04, 2018

How complacent have you become with your captive's investment portfolio? This is a particularly timely question given the recent rise in market volatility. It has been 10 years since the financial crisis, and for most of this past decade, market volatility has been fairly muted. However, if market volatility is indeed back, is your captive insurer prepared? That will depend, to a great degree, on the asset allocation model used by your investment manager to construct your investment portfolio.

Our interest in this question was piqued by a recent press release from Strategic Asset Alliance (SAAI) from their 2018 Insurer Investment Forum. One presentation titled "How Your Most Important Investment Decision Is Changing" caught our eye. The presentation outline reads as follows—"We all know that the strategic asset allocation decision determines 90+% of the long-run return of your investment portfolio. And, it seems virtually every investment manager and consultant is there to trot out his or her model and tell you why it's the best guide to making the correct decision. But how do you know? What questions should you be asking to better understand those asset allocation recommendations? How can you effectively compare one model to another to determine which is best for your company's portfolio?"

We have previously explored the question of modeling in an article titled "Is Your Predictive Analytical Modeling Wrong?" As a former investment consultant for a large insurance company, Captive.com editor John Foehl often visited the company's analytical modeling group. Displayed above the entrance of their office was the quote "All models are wrong, some models are useful." The original quote, "Remember that all models are wrong; the practical question is how wrong do they have to be to not be useful," is attributed to George E.P. Box, a renowned British statistician.

Given how pervasive predictive analytical modeling has become in all corners of insurance, the quote serves as an important reminder that models are mere approximations of the real world. While they certainly can be beneficial, models should not be used as a substitute for required critical thinking as supplied by management and the board of directors.

This brings us back to the presentation by Dan Smereck and John Mohr from SAAI. The presentation began with a discussion on the fundamentals of investing and the basic risks embedded within any captive insurer's portfolio—investment risk and inflation risk. The presenters subtly warned that the "basics" risk results in the permanent loss of the captive's capital. While this is frequently expressed as a number, it has significant consequences for the financial stability of the captive. Each time a captive insurer's investment manager talks about the proposed asset allocation for the captive's portfolio, the captive board should ask, "If this model is wrong, how bad could our loss of capital be?"

By way of illustration, we offer the following slide drawn from the SAAI presentation.

Slide

You get the idea. Some of the brightest Wall Street minds thought their models were right. If they could be wrong, how likely is it that other investment managers' models could be as well?

The presenters also discussed five financial model shortcomings, included below. While most of a captive's board members may not remember the verbiage, what is important is grasping the issues. This can assist board members in asking relevant questions the next time an investment manager comments as to why its asset allocation model is supremely suited to the captive insurer.

Financial Model Shortcomings

  • Computational irreducibility—The system being studied is so vast and complex that it is difficult to reduce down within a mathematical model with any precision. This is true for weather forecasting and financial markets.
  • Emergent phenomena—While individual rules work well and make sense, as these individual rules are aggregated within the model they can lead to unexpected results and chaos. Board members should be wary when an investment manager cannot reasonably explain the results from its asset allocation model when input variables are changed.
  • Radical uncertainty—Since models are based on human experience, how can events that have never been seen before be modeled? Therefore, the probability distributions produced by a model only capture "known" events. As a further caveat, while the probability of an event occurring may be infinitesimally small, it is still not zero and therefore should not be ignored. The failure of Long-Term Capital Management L.P. provides an excellent example of this phenomenon.
  • Ergodicity—Models assume that the probability distribution of events remains the same over time. A model's complexity will rise exponentially when it allows probability distributions to change over time.
  • Heuristics—This is any approach to problem-solving, learning, or modeling that employs a practical method not guaranteed to be optimal or perfect but sufficient for the immediate goals. Run times for asset allocation models using hundreds of thousands of iterations may be impractical and, therefore, the results, which may be labeled as optimal, may not be.

The next time a captive's board of directors sits down with its investment manager, here are five questions for the board to ask.

  • How are the expected risk/return correlations estimated for the various asset classes and over what time period?
    • What is unique about your organization's methodology, and how does it differ from your competitors?
  • What types of return probability distributions are utilized for each asset class?
    • How does your model sample the distribution during simulation?
  • What's the current worst-case scenario contemplated by your model?
    • Do you have multiple worst-case scenarios?
    • How do these translate into a potential loss of capital for our captive?
  • What is the goal the model is seeking to achieve, and how should we interpret these results?
  • Is your asset model correlated in any way with our underwriting/reserving model?
    • If yes, how so and to what degree?
All of the responses to these questions should be explained in a way that is easily understood by captive board members. If the investment manager cannot provide simple responses, perhaps the board needs to consider a different manager and asset allocation model.
Follow Captive.com on Twitter

Twitter Feed