Pension funds' slow crawl up the learning curve
By Robert Hunter
The risk management revolution that is sending shock waves throughout the financial world seems to be having little effect in the cloistered world of pension plan sponsors. Although most other financial institutions are investing huge sums in monitoring the risks they undertake, indifference to risk management techniques is surprisingly common even among the largest pension funds in the United States.
Three years ago, a study noted that pension plans lacked the most basic forms of risk reporting that would allow sponsors to monitor how investment managers manage the billions given to them. In the intervening years, surprisingly little has changed.
Much of the indifference can be traced to pension funds' elongated investment horizons. Most plan sponsors rely on asset allocation strategies, which are based on modern portfolio theory, as their sole risk management technique. “We've never been a market-timing shop,” says Steve Kornrumpf, acting director of the State of New Jersey Division of Investment. “We have a long, long, long range of pension liabilities—40 years or so. We recognize the fact that there are risks being in any financial market, and we're willing to accept those risks. And hopefully we'll get paid for them.”
Many pension fund consultants and custodians agree. “Large pension funds are in the business of taking risk, not eliminating it,” says Stephen Nesbitt, head of consulting at Wilshire Associates. “They don't have daily reporting requirements. Their goal is to fund long-term liabilities. We hear a lot about value-at-risk and risk-adjusted return on capital, but those are really much more applicable to financial intermediaries than to pension plans.”
Indeed, monitoring risk levels at the plan sponsor level is a difficult task, particularly for larger funds with sprawling portfolios and dozens of outside investment managers specializing in everything from domestic equity and real estate to cash management and international fixed income. Many of these plans lay the onus of risk management responsibility on their outside managers. “Most of our funds hire active asset managers, give them clear mandates, and they go from there,” says Janet Becker-Wold, vice president at pension fund consultant Callan Associates. “They handle all of their risks from a portfolio construction and asset allocation standpoint for the most part.”
|“Sponsors may assume that big money management firms are in control of their risks, but that's not always the case. The portfolio manager might understand exactly what he's doing, but by the time that information gets to the chief investment officer of the fund, much is lost.”
Moreover, even the biggest U.S. pension plans are frequently understaffed at the sponsor level, resulting in sparse risk management “departments.” A 1997 survey of more than 300 pension funds conducted by Barra/Rogers Casey found that sponsors had an average of 1.5 people dedicated to risk management, and that number was often an amalgamation of several people's part-time efforts. Only 12 percent of the plans surveyed expected to increase their risk management staffing this year.
The main reason for such modest efforts is cost. In smaller pension funds, risk management technology and staffing can easily add up to more than $200,000 a year; bigger plans can spend 10 times that much. The Barra/RogersCasey survey found that risk management budgets are generally greater than $100,000 among plans with assets of more than $15 billion, while more than half of the funds surveyed with assets of between $1 billion and $5 billion spend more than $100,000 a year on risk management.
Despite all the impediments, progress is being made. “Until recently, the sponsor world had been primarily return-focused. That's what people spent their time on,” says Ethan Berman, managing director and head of the risk management products division at JP Morgan. “The strongest quantification had been the Morningstar one-through-five-star system. But that's changing. You're starting to see more and more sponsors focus on risk-adjusted return.”
Much of the credit for the gradual turnabout goes to New York-based Capital Market Risk Advisors, which in 1996 assembled 11 plan sponsors, known as the “working group,” and more than 50 outside “commentators” to devise the Risk Standards for Institutional Investment Managers and Institutional Investors. The group identified 20 elements of risk management as essential to plan sponsors. The crux of the Risk Standards is Number 13: “Risk-adjusted returns should be measured at the aggregate and individual portfolio level to gain a true measure of relative performance.”
A year later, CMRA surveyed the pension fund universe to see the impact the Risk Standards had made. It found that while before 1997 few plan sponsors used risk-adjusted returns in any meaningful way, by 1997 more than 47 percent were moving toward measuring risk-adjusted returns. But it also found that more than 25 percent of those surveyed found the Risk Standards to be useful but didn't have a clue how to implement them. “In 1994, the volume of telephone calls we received inquiring about proactive risk management was about one every two weeks,” says Tanya Styblo Beder, a principal at CMRA. “Two years later, it became one call a week, and now we get around a dozen inquiries a week.”
Of course, simply inquiring about risk management does not a risk management system make. CMRA found that before 1997, 33 percent of sponsors adjusted the returns on individual portfolios via a common risk-adjusted return measurement to reflect the risk that internal and external managers take. By 1997, that number had crept up to 40 percent, and 47 percent of the sponsors surveyed said they planned to use a risk-adjusted return measure in 1998.
There is clearly a big difference between simply using risk-adjusted techniques and actively managing risk. “Risk management describes people who adjust the return profile and the risk profile of a portfolio,” says Maarten Nederlof, the former director of investor risk management strategies at CMRA who is now a managing director at Deutsche Bank. “So if I'm running an Asian portfolio, I get to decide how much risk I take on Malaysia today. Risk measurement, by contrast, describes the people who oversee all these funds, create reports that summarize the exposures by country and so on. Management is more proactive; measurement is passive, since it merely involves monitoring.”
|"Most of our funds hire active asset managers, give them clear mandates and they go from there. They handle all of their risks from a portfolio construction and asset allocation standpoint for the most part."
In other words, generating reports is one thing; acting on them is something different altogether. Ontario Teachers' Retirement System has been one of the few pension funds to acknowledge publicly that it had bridged the gap between measurement and action. Last year, it began experimenting with using value-at-risk on a portion of its assets. By June it had put all of its $41 billion into a VAR framework. It uses its VAR numbers, among other things, to track and value surpluses and, most important, to allocate specific risk levels to its outside managers, who are expected to outperform their assigned risk level rather than simply a benchmark. The plan allocates each manager a certain amount of money, then subtracts from it a portfolio consisting of the manager's benchmark. Next, it calculates a VAR based on the deviation from the benchmark, which it calls management effect at risk, or MEAR. Each manager is expected to produce excess returns at a fixed percentage of the plan-wide MEAR level—reportedly 10 percent. Ontario Teachers takes its risk management seriously—this year its risk budget is 3 percent of its assets, or roughly $1.23 billion.
Ontario Teachers is the rare exception rather than the rule in fund-wide risk management. Most pension funds are remarkably apprehensive about venturing out the risk management learning curve. The Barra/Rogers Casey survey indicates that asset allocation continues to be the most widely used risk management tool among sponsors. Although many sponsors feel the need to increase their risk management budgets, money alone is not the answer. While 97 percent of pension funds have created written risk management policies, only 46 percent assign risk monitoring to specific individuals. And many individuals complain that risk policies are not communicated clearly, reviewed regularly or implemented consistently across the entire plan.
“The problem gets more difficult when you have external investment managers who are involved in complex activities that may not have tracking systems,” says Nederlof. “Sponsors may assume that big money management firms are in control of their risks, but that's not always the case. The portfolio manager might understand exactly what he's doing, but by the time that information gets to the chief investment officer of the fund, much is lost. CIOs get summary reports produced by systems that can't provide high-resolution detail on certain types of assets. So while the valuations may be correct, they aren't necessarily going to show, say, an adequate stress-test of market movements.”
Indeed, many funds still lag considerably in stress-testing capabilities. Risk Standard 14 states that “simulation or other stress tests should be performed to ascertain how the aggregate portfolio and the individual portfolios would behave under various conditions. These include changes in key risk factors, correlations or other key assumptions and unusual events such as large market moves.” As of late 1997, however, CMRA had found that only 40 percent of investment managers stress-test their valuation models weekly, monthly or even quarterly.
Systems to the rescue
Many fund sponsors complain that the lack of a single, accepted method of aggregating risk lies at the root of the industry's risk management woes. Several vendors hope to foist onto the industry just such a product. The granddaddy of all institutional risk management systems is Bankers Trust's RAROC 2020, which analyzes a portfolio by risk type (equity, interest rate, currency or commodity), security type, and by manager or account based on a one-standard-deviation move for a one-year holding period. First launched in 1994, the product offers various reporting frequencies to appeal to a wide swath of the institutional investor marketplace, including pension funds.
In the RAROC 2020 environment, Bankers Trust acts as a “service bureau,” collecting data mostly from pension fund custodians and furnishing reports to fund sponsors. Aside from a longer time horizon, RAROC 2020 allows for stress testing based on what-if scenarios via a series of variance-covariance matrices, and lets clients monitor performance based on specific benchmarks. It also allows plan sponsors to compare their aggregate portfolio to their liability stream. In August, Bankers Trust unveiled an Internet version of the product.
Another Internet-based system hit the ether in August as well. Earlier this year, JP Morgan and State Street Analytics, a pension fund custodian, teamed up to bring Morgan's much-vaunted RiskMetrics system to pension funds to calculate fund-wide VAR. In August, the two announced the launch of VaR Calculator II, available on the Internet. Whereas the RiskMetrics methodology used by investment banks makes assumptions about the mean or expected return of any particular asset class over the next day, the JP Morgan/State Street system looks at longer time horizons. It allows fund sponsors to conduct stress testing and scenario analyses—what-if and canned scenarios—with State Street custodial data updated monthly.
“VaR Calculator allows clients to take a report from State Street Analytics on their assets that calculates their actual VAR based on their actual holdings, across their managers,” says Morgan's Berman. “And now they can get a desktop tool that can sit at the client, into which they can enter their own portfolio positions and do the calculations. We've seen a huge amount of interest in the product thus far.”
Another Internet system, Alerts, produced by pension custodian Northern Trust, allows plan sponsors to define certain risk parameters within the portfolios of their investments, which are stored in NT's Chicago headquarters. When managers exceed a risk limit, an e-mail is sent to the sponsor informing him or her of the violation and instructing the sponsor to log onto the Alerts system for a detailed analysis of the problem.
Perhaps the most ambitious of the risk management systems available to plan sponsors is Barra's Total Plan Risk, which determines how much risk individual investment managers were exposed to in order to generate their returns. “Alternative approaches estimate risk by extrapolating historical return patterns,” says Todd Doersch, senior vice president at Barra. “By contrast, TPR takes into account the fundamentals of the securities that make up the fund's portfolio. It captures the interrelationships between the securities' characteristics, and crunches the numbers to estimate volatilities in a more robust way. Rebalancing decisions based on TPR are more enlightened and better informed.” The main problem with the Barra system seems to be its rather hefty price tag. To date, only five plan sponsors have signed on as clients.
Despite these efforts, it's doubtful whether any system can become the industry standard. “There is no single off-the-shelf solution that covers everything from the mortgage market—the second-largest capital market in the United States—to the U.S. Treasury market, to all the foreign equity and bond markets,” says CMRA's Tanya Styblo Beder. “People who are using Internet-based solutions to do risk-adjusted return are faced with a challenge: In, say, emerging market securities, where do they get one year of history for securities issued in the last two months? You have to pick proxies for how those trade, and there's a lot of data-filling and gapping issues. It quickly becomes a tedious exercise for people.”
Others register similar concerns. “I'm a little skeptical about how useful these products are for a pension fund,” says a portfolio manager at Ameritech. “I think they make a lot more sense for trading operations at banks. They're not forecasting tools. Some people have the mistaken idea that a risk-measurement product is going to help them avoid situations like the meltdown in Southeast Asia and Eastern Europe. I'm skeptical about the usefulness of the output on a big, total-plan level.” Deutsche Bank's Nederlof agrees. “Knowing your VAR is virtually useless. What are you going to do with it? Two portfolios can have the same VAR and be different in many ways. The exercise of calculating VAR, on the other hand, does force you to get disciplined about a lot of things, which is valuable, but in most cases is more valuable than the number itself.”
Many place the blame for the relatively limited use of derivatives at the sponsor level (see box, Page 40) on an underdeveloped software market. “I don't think funds should operate in areas in which they're blind,” says a risk management consultant. “So either the more exotic derivative instruments need to have some analytics produced along with them that will make it possible for them to be tracked, or, as a number of funds have said, ‘Let's back out now and wait until the systems are available.'”
|"Value-at-risk is only a tool—there's nothing ideal or perfect about it. Everything is based on hindsight, and we don't know when the next crisis is going to hit."
head, State Street Analytics
It is clear that sponsors are looking for solutions that have not yet been developed. “VAR is only a tool—there's nothing ideal or perfect about it,” says Laurette Bryan, senior vice president of State Street Corp. and head of State Street Analytics. “Everything is based on hindsight, and we don't know when the next crisis is going to hit.”
Derivatives? Not Here!
Derivatives dealers have made valiant efforts trying to pitch over-the-counter derivatives to pension funds. But after a decade or so of trying, they don't have much to show for their efforts.
The derivatives-related disasters of 1994 certainly didn't help. Pension funds—particularly public ones—are acutely sensitive to bad press. But even that doesn't explain why so few funds will admit to using derivatives in a significant way.
Many funds say that, because they're so massive, they can afford to deal only in cash instruments—and the tiny bit of outperformance they could gain by using derivatives isn't worth the headache. “Ninety-five percent of our return is based on the asset allocation,” says Pat Mitchell, chief investment officer at California State Teachers, “and the next 2 percent to 3 percent is based on the benchmark that is used. Tweaking or leverage may be able to add a tiny bit, but we want to focus on the 97 percent to 98 percent we can control.”
New York State Teachers is emblematic of many large funds in that it allows its investment managers to use derivatives but avoids them at the sponsor level. “We have no derivatives strategy per se,” says a New York State Teachers spokeswoman, “but we do have outside international managers who manage a portion of our portfolio, and they use forwards and futures for currency hedging and so on.”
Although derivatives have yet to move into the pension mainstream, they are used more or less frequently for a variety of functions. Derivatives, for example, are frequently called in to help funds make short-term adjustments to their asset mix. If a particular fund's target asset allocation is, say, 60 percent equities and 40 percent fixed income, the equity market run-up of the last few years might well have shifted that mix to 70–30. Rather than reallocating 10 percent to fixed income by firing fund managers who have performed well, sponsors can use futures to rebalance the portfolio.
“We do asset allocation here—that's our foremost strategy, whether it's domestic or global,” says Tom Loeb, chairman and chief executive officer at Mellon Capital. “We use futures in asset allocation to get exposure and cut costs, because trading futures is as little as one-fifth the cost of trading the physical securities.” California Teachers' Mitchell admits that he occasionally uses Standard & Poor's 500 futures to tweak his asset mix. General Motors' George Bodine, meanwhile, says he uses a variety of futures—both domestic and international equity and fixed-income products—as well as currency forwards for overlay and direct exposure.
Other funds use derivatives for broad-based equitization strategies. Sponsors from time to time amalgamate idle cash that individual fund managers keep for liquidity. If, for instance, a group of equity managers each holds 10 percent cash, a 60 percent equity asset allocation of the fund could drop to, say, 54 percent. Sponsors can cover that idle 6 percent with futures to get the face amount of equity exposure.
Derivatives also come in handy in transitional situations. “If a sponsor fires a manager to shift allocation to equities,” says CMRA's Tanya Styblo Beder, “he can do an equity swap in which he pays fixed income and receives a rate of return linked to the stock market. He can fire a manager and have a dealer do the transition on the wire overnight. That has been a major benefit to the buy side as a whole.” Moreover, a fund that issues a pension obligation bond to raise money can use derivatives to put that money to use right away, rather than waiting until it hires one or more new managers.
Many sophisticated pension funds, such as Amoco, use derivatives to reallocate their plan's alpha, or the amount of return above the market's return (beta) that results from manager skill. There is considerable debate as to the real composition of alpha—some think that greater alpha results merely from managers taking greater risk, and hence should be tempered rather than cultivated. But alpha-porting strategies can add up fast. An additional 1 percent return on $1 billion over a 10-year period can provide a fund with nearly $200 million in excess returns. “We primarily use a lot of listed futures when we port alpha,” says Marvin Damsma, director of trust investments at Amoco. “We try to separate a manager's (or a strategy's) returns into two streams—the asset-class return and the excess return—and we move those asset-class returns by going long or short futures contracts. For example, if we give the money to an equity manager but we don't want the equity exposure, we sell that off by shorting equity futures. And then we can buy back other futures (bonds, for instance) in another market class, while the excess return stays the same.”
The cutting edge
Perhaps the most novel pension fund derivatives deal to date took place earlier this year, when Ontario Teachers and the Dutch fund Stichting Pensioenfonds entered into a total return swap on equity indices. The deal, which was consummated only after two months of board approval at Ontario Teachers and another six months during which lawyers hammered out ISDA master agreements, featured two equity index swaps totaling $60 million. Each fund had been interested in gaining international exposure to the other fund's domestic equity market, but because Canadian law limits pension funds to 25 percent international equity exposure, Ontario Teachers could not use the cash markets. Derivatives, however, do not count in the Canadian statute, so Ontario Teachers approached Stichting with an idea to swap the total returns of the Toronto Stock Exchange 100 index and the AEX index, a Dutch large-cap index. The swap has a one-year duration, with settlement at maturity.
The parties claim the deal saved them 25 basis points by dealing directly with each other rather than using a financial intermediary, and they avoided the costs they would have incurred had they dealt directly in the cash markets. “That deal speaks to the evolution of markets,” says CMRA's Dale Kunkel. “Instead of going to a dealer intermediary, I can now go directly to another pension fund and enter into a total return swap. This offers cost savings and international diversification without going through settlement procedures.”
Another class of OTC structures involves guaranteed investment contracts that allow plan sponsors to get a stable return while gaining a varying amount of upside. The products, which are marketed by Bank of America and other dealers, are quite popular in defined contribution plans, and lately have begun to register on the defined benefit radar screen as well. The simplest of BofA's pending deals involves a contract that is 90 percent Libor-linked, with 10 percent tied to a managed futures account. Not all 10 percent is invested in managed futures—some is held aside for liquidity. The benefits: diversity and stable value. “Academic studies have shown that managed futures have little or even negative correlation with traditional markets such as equities and bonds,” says Sai Raman, a risk manager at BofA.
A typical BofA deal costs between 40 basis points and 60 basis points, although prices can range from as little as 10 basis points to as much as 150 basis points. Most structures have maturities of five years, but different maturities will emerge as the products become more common. “The floor aspect gives fund managers a blended risk package that they can live with,” says Robert Whiteford, vice president of global capital markets financial engineering and risk management at BofA. “It is for sponsors who need a risk-reducing component in their asset blend. It gives them the effect of a medium-duration bond with a little bit of a kicker, and it's not correlated.”
BofA claims to have a dozen or so deals in the pipeline. “The demand for these guaranteed investment contracts is gigantic,” says Whiteford. “Ultimately, this market will expand. Right now no two deals are alike, but eventually they'll have a cookie-cutter aspect.” —R.H.