Dirty Secrets Of Risk Management
By: Wendy Brodkin
Is risk management all it is supposed to be? Wendy Brodkin, of Goodman Institutional Investments, challenges some of the beliefs about the value risk management adds to management of money.
Risk management is the new flavour of the month in the pension fund industry, replacing the obsession with value-added over market benchmarks of only a few short years ago. Investment programs are now focused on managing risks recognizing the difficulty, if not the impossibility, of managing returns. Expertise in risk management and tools such as value at risk (VaR) is now a perquisite in many senior job descriptions. Increasingly, industry forums are addressing the subject.
There is nothing wrong with risk management per se. In fact, the underfunding crisis of this decade could have been avoided with a little more attention to risk exposures. Nevertheless, there is a problem that is slowly being recognized with the definition and implementation of risk management.
Risk management today is built on worn-out conventional wisdom that duped the industry throughout the bull market in bonds that started in 1982, right up to the dramatic global equity correction of 2000. Borrowing a phrase from Jeremy Grantham to explain, it is flawed with “dirty secrets.”
In 1990, Grantham wrote an article for AIMR’s first Investing Worldwide publication called ‘The Dirty Secrets of International Investing.’ He explained that these secrets were basically “unpleasant truths” that the industry was reluctant to address. For example, he pointed out that diversification doesn’t necessarily work when you need it and, therefore, is a seriously overrated benefit of global investing. It’s too bad we didn’t realize that 14 years ago!
The dirty secrets of risk management are similarly counter-intuitive and at odds with the way we manage investment programs. Consider, for a moment, the fundamental tools that are used to manage risk. Indicator Of Diversifying Power Correlation co-efficients are a widely used indicator of diversifying power in asset mix studies to allocate money amongst asset classes and in manager selection to combine different styles of money management in order to control risk. However, we know that correlations aren’t stable over time and, when markets are stressed, correlations converge. We also know that low/high correlations don’t necessarily mean different/similar returns. Our experience throughout the ’90s with the spectacular outperformance of U.S. stocks relative to Canadian stocks proved that. In the Winter 2003 issue of The Journal of Investing, Ernest Ankrim analyzed historic correlation statistics and returns in various markets and market segments. The results showed the opposite of what we believe. Periods of high correlations had large return differences between markets and periods of low correlations produced similar, not diversifying, returns!
Another structural flaw in risk management is the use of historic performance results. The industry is seemingly besotted with the past and disguises the numbers in a host of ways: standard deviations, information ratios, Sharpe Ratios, Sortino Ratios, etc. If that is not enough, the numbers are shown over varying time periods, often with the same end date sensitivity. Yet we know there is no embedded future forecast in these numbers and we know they don’t say anything about prevailing portfolio risks.
While returns-based analysis is fast, cheap, and easy to understand, only an analysis of the underlying holdings provides information about future risk and current exposures through such data as concentration, liquidity, and pricing levels.
A number of more sophisticated tools have gained some recent popularity including value at risk, sensitivity analysis, and stress testing. The problem is that they lack standardization and the quality of the output and interpretation varies widely. Some analyses are no more insightful than any of the old tools. They also rely on normally distributed results and historical events. More information is definitely a good thing, but it’s not a panacea for risk management. The industry still must consider the implications of the unexpected and events that can’t be easily modelled.
In any case, VaR techniques are not widely used for decision-making, likely due to the cost and complexity of the systems. The decision-making process to bridge the output from the analysis to the investment solutions is only just evolving. There are still many plan sponsors who have yet to identify what risks should be measured and over what time horizon. As important, the investment solutions to market risk and active management risk are generally accepted industry practice in the form of traditional active and passive portfolios. However, solutions to liability risk and capital losses are only now emerging.
Generally accepted industry practices are likely the biggest threat to any meaningful change in the industry. The rules of thumb that guided trillions of dollars through markets in the last two decades are flawed. Even historically, they generated only limited situational success at best and led to the underfunding crisis of 2000- 2002. These practices include the ‘buy and hold’ and ‘stocks for the long run’ strategies and the assumed linkages between risk and return, bond yields and stocks, the U.S. dollar, and gold. No fancy risk management overlay will compensate for the dominant underlying errors in these assumptions.
In many cases, the biggest error for Defined Benefit plans continues to be the exclusion of the liability risks in decisionmaking despite the fact that it is almost 20 years since pension legislation introduced the requirement of basing investment policy on plan specific characteristics. Risk/return scatter grams are still widely used as a snapshot of fund performance in boardrooms across the country, likely because no one has yet developed a commercially appealing asset/liability snapshot.
Even when liabilities are taken into account in asset/liability studies, the analysis is often irrelevant because it spans time frames beyond the life of the decisionmaking committee and long past the intervals of actuarial and accounting reckoning. And, in contrast to the ubiquitous risk/return charts, these studies are typically limited to every few years. Traditional performance reporting does virtually nothing in terms of monitoring the fund’s financial position and risk exposures. In fact, its results are most often at odds with the true position of the fund. In 1994, pension funds reported barely break-even results due to an equity market decline and two serious bond market crashes. However, it was a very good year for the asset/liability position because interest rates rose causing the liabilities to decline.
Similarly, the industry is breathing a collective sigh of relief with the doubledigit returns of 2003. However, in most cases, this return is inadequate to compensate for the increase in liabilities as a result of the interest rate decline.
What’s the answer? It’s certainly not to get out of the DB business. I hardly think that it is appropriate to transfer the risk to individuals when the whole North American pension industry and their assorted advisors struggled with it. The answer is also not to introduce new ways to measure risks of limited relevance. As one cynical, industry long-timer quipped in response to a pension committee adopting ‘decisionmaking from risk-adjusted space:’ “WHAT’S THAT ... A PADDED CELL?”
I think the solution must begin with identifying the real threat to the pension fund’s financial health and the risk tolerance around it. (In most DB plans, the primary risk is sensitivity to falling interest rates over three-year periods). The risk must then be measured and monitored with a plan to hedge the risk when it reaches some tolerance threshold. Risk management is about controlling risks, ensuring any intentional risks will be reasonably compensated, and believing that the unexpected can happen. For many plans, the only way to control the risks is through non-traditional investment programs and alternative strategies.
In conclusion, the industry is a lot smarter than before the tech bubble burst and certainly a lot more sophisticated. I remember back in the early ’90s when it wasn’t even standard practice for U.S. corporate pension plans to have an investment policy. Today, it is a study unit in the CFA program.
No doubt, we’ll look back in another 10 years with the clarity of hindsight and wonder how we could have been so wrong about so much, as we did over the last few years.
The issue is whether or not the mistakes will be fixable at that point. For some funds, this time around, they were not. The industry’s future depends on how we start to manage risk today.
Wendy Brodkin is president of Goodman Institutional Investments Inc.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -