Model Risk for Risk-of-Ruin Models

Risk of ruin is commonly used to refer to the probability of running out of money in retirement, assuming you do not adjust your spending.  It is a mathematical construct that is intended to help a person identify how much money they can reasonably expect to take as income from their retirement assets without needing to significantly adjust their expenses in later years.

Huy Lam, CFS, CFA®, Director of Financial Analysis and Planning at Schultz Collins, Inc., shared with attendees of RIIA’s 2016 Summer Conference, some of the many ways to measure risk of ruin and showed that depending on the methodology used, the probability of running out of money in retirement varies widely.  One huge source of variation is in the way asset returns are calculated. Another variation that makes a big difference is whether you are simulating ruin over a fixed time period (say 30 years) or over the expected remaining lifetime of your client (which, for a 65 year old, might only be 19 years).  Still another set of variables that can cause large differences in your ruin calculation are taxes, investment fees and transaction costs.

Using five different approaches to simulating asset returns, varying the planning horizon, and adding in taxes, fees and transaction costs, Huy discovered that the risk of ruin was as low as 4% or as high as 49%, demonstrating that precision and accuracy are two very different concepts.  It is possible to get a very precise number from any of these models – just pick a set of assumptions.  But the probability that your number is accurate is highly unlikely!

Dirk Cotton made the point that risk-of-ruin models are assuming that actual retirees will just keep spending the same amount every year even if they face clear prospects of ruin…and the probability THAT will happen is zero.  Instead, the retiree would reduce their spending and end up with a diminished standard of living.  This probability – a reduced standard of living – is harder to quantify but appreciably more relevant to the retiree.

Huy’s presentation offered an excellent review of stochastic modeling techniques, but he and Dirk shared similar concerns. Difficult (or impossible) to measure accurately, risk of ruin is a tool that may mislead investors if the underlying assumptions of the model are misunderstood.

Post a Comment