One good outcome of the stock market plunge of the last few years is that it has exposed ‘relative’ performance as a misleading indicator of financial health. The consultant who says to his client, “good news, your portfolio did better than the market, now please contribute umpty ump million pounds to your pension plan because it is hugely under-funded”, does not have a happy client.
Strangely, the consultant is unlikely to be terminated by the client. Why? Wasn’t it the consultant’s advice that got the pension plan in this predicament? He won’t be terminated because all consultants measure performance relative to the market. Isn’t it possible to measure performance in absolute terms? Yes! The Pension Research Institute has made the software for measuring absolute performance available for free to all, with no royalties required (see /free software). Then why don’t they do it? Because consultants could no longer do one performance study and sell it to every client they have. This ‘one size fits all’ approach is very profitable.
Consultants would have to provide a personalised performance service based on the absolute return that each client must earn at minimum in order to accomplish their goal. For the defined benefit (DB) plan the goal is to fund the plan within cost constraints. For individual retirement plans the goal is to retire at a specified time. In each case there is an absolute return that must be earned at minimum in order to achieve the client’s goal. I call this the minimal acceptable return, or MAR. Performance of all investment managers would then be compared to the client’s MAR. One would then be tracking performance toward their goal instead of measuring where they are relative to the market.
Would you use a doctor who measured your health relative to all of his other patients but had no ability to measure your health in absolute terms? Well, what if all doctors did it that way? I suspect there would be a huge public outcry for physical examinations that told patients the state of their health in absolute terms.
There should be a public outcry that performance of retirement schemes of all types be compared with the MAR that is absolutely necessary in order to accomplish the goal of each pension scheme. Risk should be measured in terms of falling below that MAR, and reward should be measured in terms of being above the MAR.

Portfolio versus manager
Some scholars who agree with me that performance of the client’s portfolio should be calculated against an MAR, do not agree that each manager should be measured against the same MAR. They believe it would be unfair to hold every manager to the same standard. What if the manager is in a style that is doing poorly, like large cap growth managers are currently experiencing? Is it fair to compare their performance with small cap value managers whose style is in vogue?
To begin with, almost all managers are a blend of several passive indices, eg, large growth plus large value, with a dash of cash equivalents. So, it would be best to calculate the upside potential and downside risk of each manager’s style benchmark and then find those who can beat their style benchmark.
Managers should only be compared with managers in their broad style category; but everyone in that category should be compared to the client’s MAR. What about performance of equity managers in different countries? The same principles apply. What about managers in different asset categories? Now we are talking about asset allocation, not performance measurement. Can asset allocation be done in a downside risk framework? Yes, but that software is not free.
The crux of the problem
The example in Graph 1 will illustrate the problem and the solution.
Suppose managers A, B, and C had the historic returns over time depicted in Graph 1. Equity manager A thought the end of the world was coming so he stayed in short term governments and got a constant rate of return of 5%. He says: “Well I didn’t get the best return, but I didn’t take any risk.” Really? 1, his standard deviation is zero. But, if manager A constantly earns a rate of return less than the 8% the client must earn at minimum in order to accomplish her goal, didn’t that incur risk of not accomplishing her goal? Yes! What is the upside potential for this strategy? ZERO, none, nada! In fact, manager A pursued the most risky strategy with the least upside potential and anybody who claims otherwise is not measuring risk and reward relative to the clients goal.
What about manager C. All consultants using traditional performance measures would claim he took the most risk. In fact, manager C had less downside risk than manager A. What about manager B compared with manager C? Hard to say which one took the most downside risk without using the bootstrap to calculate downside risk. What is clear is that manager C has a lot more upside potential to downside risk than manager B. So the upside potential ratio would clearly rank C above B.
Did manager C accomplish this through skill or luck? Nothing I have said will tell you that. But you can find the answer in a software program called ‘Stocktrib’ at
Can you blame someone for trying?
It is one thing to have the best concept of risk. It is another to get reliable estimates. Some firms are now calculating downside risk, but most of them are calculating it incorrectly. The most common error is using only the historic returns that fell below the MAR to calculate downside risk. That is like going to a doctor who only looks at the outside of your body to determine if there is something wrong with your lungs.
The correct medical procedure is to take an X-ray. The correct statistical procedure for calculating downside risk is the bootstrap procedure developed by Bradley Efron at Stanford University. There are a number of other statistical procedures that are necessary for reliable calculations of downside risk and upside potential. Suffice it to say, unless the calculations were made in accordance with the Forsey-Sortino software (which is free) I wouldn’t trust it.
What is the solution?
The market has a way of getting around barriers. Just as consulting firms replaced actuaries in importance because they wouldn’t or couldn’t measure performance, the internet will make those consulting firms that won’t adapt, obsolete. Will absolute performance measures guarantee success? No! Is it an improvement? Absolutely!
Frank A Sortino is director of the Pension Research Institute in San Francisco