Gary L. Perlin ranks as a prominent derivatives user who knows a whole lot about innovation. Last March he became treasurer of the World Bank, taking over from Jessica Einhorn. But this isn't his first stint on K Street. In fact he began his career at the Bank in 1975, then moved to Wall Street. In 1982 Perlin returned to Washington for an 11-year stint at the Federal National Mortgage Association (Fannie Mae) culminating as senior vice president for finance and treasurer, a job he held until he left in early 1993 to rejoin the World Bank, also an early and very conspicuous player in the derivatives market. Though Perlin wasn't on board for the Bank's historic first currency swap in the late 1970s, just a few blocks away he did at lot of pioneering work while at Fannie Mae. There he became such a respected figure that he was invited onto the board of the Futures Industry Association and also joined the G30 working group.

In over a decade at the helm of the one the largest agency borrowing programs in the U.S., he was in a unique position to witness the radical changes in the capital markets that occurred during those tempestuous years. This interview was conducted by contributing editor Simon Boughey.

Derivatives Strategy: In the late 1980s and early 1990s, Fannie Mae embarked on a program involving much wider use of derivatives. Will you sketch the problems the agency faced and how you thought that derivatives would help?

Gary L. Perlin: If we go back to 1982, when I first arrived, it was a time of exceptionally high interest rate volatility and an exceptionally high absolute level of rates. I indicate those as two separate issues because Fannie Mae as a company had two sets of problems. One was the process by which it, in effect, sold forward commitments to purchase mortgages, which had a very strong options component. Because of the volatility, the fees that were being charged were well below a value that compensated Fannie Mae for taking the short-term risk during the commitment period.

The second problem was the very high absolute level of rates, which occurred at a time that the company had a huge duration mismatch. The fact that its portfolio was short-funded at the same time as there were very high nominal rates meant there was a significant negative net worth. So in 1982 there was both a flow problem and a stock problem.

We talked a lot about the flow problem at the time because it was the easiest thing to understand and the easiest thing to fix. Among other things, we just stopped issuing optional commitments, because we did not have a good hedging mechanism, nor did we have a reliable pricing regime. All of the commitments were converted from being optional delivery commitments to mandatory or forward delivery commitments.

It was not long after that we began to address the stock problem, because that was what represented the greatest risk to the balance sheet. This was a much longer process. It meant the issuance of longer-term debt to redress the duration imbalance. And it meant integrating the use of derivatives capacities so that the balance sheet would become more matched.

So, as early as 1982, we were using T-bill and Eurodollar futures to synthetically extend the average duration of short-term debt outstanding. At the time we used futures because there were no other derivatives. We also began to time the pricing of long-term debt issues more tactically. And like many people what we realized in the mid-1980s was that we had to start thinking about risk management from a more comprehensive standpoint. This meant not only using derivatives, but also thinking about our regular business in a context of risk management. All of a sudden, derivatives forced you to recognize the inherent risks in your balance sheet.

That was the really big challenge up until the late 1980s, when the next stock problem became one of dealing with another optionality mismatch in the balance sheet. We had a convexity mismatch because our mortgage assets were callable and our debt was not, so it was then a matter of spending the next four to five years dealing with that sort of risk.

DS: Could you not solve that simply by issuing more callable debt?

GLP: Exactly. That is what we ended up doing, and it was something that we could only redress in the long run by issuing callable debt. But if you go back to 1988, by which time the overall duration mismatch had largely been managed away, we had $100 billion worth of assets on the balance sheet and we had no callable debt at all. So we were required to move very deliberately over time. I believe when I left at the beginning of 1993, well over half of Fannie Mae's outstanding debt had some form of call feature.

Here the use of derivatives was more limited. It was not as though we went and hedged the optionality mismatch in the balance sheet, because we could not have advised that kind of exposure. We just had to build up the amount of calls on the liability side, and the way to do that was to increase the range and volume of option-embedded structures that we could sell to the marketplace. Often that required that we link up an off-balance-sheet position in conjunction with the new issue in order to achieve cost-effective access to a huge amount of options.

DS: So Fannie Mae's use of callable debt was quite different from some of your colleagues in the agency market, like those at the Federal Home Loan Bank, who were selling the option components to achieve deep sub-LIBOR funding?

GLP: That's correct. We kept our calls. Fannie Mae's balance sheet was almost exclusively comprised of long-term fixed rate mortgages, so it did not make sense to offer a debt structure that allowed you to get long options, and then to turn around and sell that.

DS: So what kind of off-balance-sheet instruments were you combining with new issues of debt?

GLP: To a large extent they were currency structures that would be swapped. Many of the structured issues had complex formulas, with floors and caps and so forth. Occasionally we would need to match those off against some derivatives positions so that we would be left with vanilla corporate callable-type structures. We hedged away the additional features.

DS: How were Fannie Mae's first forays into the Eurobond market in the late 1980s received?

GLP: They had mixed receptions because the U.S. agencies had structural disadvantages related to the registration requirements. Also, there was not much demand for callable paper in the Eurobond market. We focused on internationalizing our debt sales in 1986­1988, but that was more a matter of trying to redress the duration mismatch on the balance sheet.

We needed to sell a whole lot of very long term debt. It was just at the time when investment flows, particularly from Asia, became very important. Japanese investors became major buyers of long-term Treasuries and they found a lot of value in buying long-term agencies as well. But our use of the international market was more a matter of redressing the duration mismatch rather than the options mismatch.

DS: Why did it take Fannie Mae so long to issue callable debt to match its assets?

GLP: That's a good question. I think that the issue was one of priorities. Through the course of the 1980s, Fannie Mae's bigger risk was its duration mismatch, and to reduce interest rate risk we needed to focus on issuing a lot of longer-term debt. I think to have tried to hedge away the options mismatch would have made that job more difficult and more costly. Moreover, once absolute rates came down substantially, the options risk-which was exceptionally high for fixed rate loans that originated in the early 1980s where the coupons were substantially into the double digits-rose significantly. We also understood that we had to develop the agency callable market in such a way that would allow us to issue very substantial amounts of debt over time.

DS: Were you successful in erasing the duration mismatch and the optionality mismatch?

GLP: Over the ten years from 1982 to 1992, Fannie Mae was successful in doing that. Of course there were a lot of different pieces to the puzzle. Not only were we active on the liability side, but also on the asset side: for example, floating rate mortgages appeared, which helped fix things. But you cannot look at it without recognizing the scope of the duration mismatch. It took a very substantial period of time to change simply because of the volume of business to be done.

DS: Did Fannie Mae make any assessment of how much the derivatives program of the 1980s saved the agency-and the U.S. taxpayer, of course?

GLP: I am not aware of any overall assessment which was made. Individual evaluations were done, and the one that is clearest in my mind was in 1982. We had synthetically extended the life of our short-term debt by shorting some T-bill and Eurodollar futures on which we had a net margin outflow as rates declined. But because the decline in rates had an even more positive impact on Fannie Mae's cost of debt, our spreads to Treasuries shrunk considerably. We achieved tens of millions of dollars of net positive value through exercise.

We tried extremely hard throughout the decade to make sure that any hedge strategy was fully integrated into the overall asset/liability management strategy and looked to the net performance of the portfolio. Eventually it was clear that not only had we reduced the risk inherent in the balance sheet, but we had facilitated profitable growth in the portfolio by being able to match at the margin as well as on the embedded stock.

DS: Tell me about Fannie Mae's relationship with the dealer community. Did those relationships change as the program progressed and the market developed?

GLP: I think the relationship with the dealer community was always central to the exercise. The range of relationships went from derivatives to very extensive cash market relationships as well. The role of different firms developed over time. Some firms became very adept at structuring deals that provided us with attractively priced options, particularly the bulge-bracket firms that had good underwriting and risk management capacities. Our relationship with regional firms matured. Many of them had extremely good distribution capacities for debt that had optionality embedded in it. We had growing relationships with international dealers who recognized the company as a large player.

DS: What were the most important innovations in Fannie Mae's changing relationships with the dealer community?

GLP: Collateralized derivatives contracts really opened the world to us. Perhaps the first collateralized swap would have been done around 1989, and until then our swap guidelines were so credit restrictive that it meant that we did not have the capacity to do swaps with the majority of dealers. That inability to use their structuring capacity meant we could not tap the market for large amounts of callable liabilities.

However, the drive towards collateralization was a bit difficult. There were some dealers in the market who resisted moving towards the idea of collateral because they were very highly rated and did not want to lose that competitive advantage. This was in the days prior to special purpose vehicles, and some of the lower-rated credits did not want to set precedents of offering collateral or doing one-way mark-to-market agreements. Yet we were such a prospectively large business partner that dealers decided the profit potential made it worthwhile to engage in such credit-enhancing activities. It was something that needed to be brought along over time; it was not something that they immediately saw as being in their own best interest.

DS: How do you think Fannie Mae's experience was different from some of the other agencies?

GLP: I think the experience was different because many of the other agencies were using the same markets to obtain very attractively priced floating rate money. They ended up using many of the techniques that Fannie Mae also used, but they didn't necessarily have the same long-term market development interests that we had. Once the markets developed, what we needed to manage our balance sheet was often very different than what other agencies needed to manage their balance sheet. What we shared was a credit standing and a market access. So we very often acted very similarly in the markets, but with different long-term management strategies.

DS: Did the Federal Home Loan Bank and other agencies always get value for money when they sold calls and options?

GLP: It's my impression that most agencies invested significantly in their ability to assess the value in the structures. I think-particularly if they were not looking to maintain a long position in calls on the liability side of the balance sheet-they only engaged in transactions that made good sense for them. However, whether or not they got full value would have been a function of their ability to reverse-engineer the structures and their ability to disaggregate the execution of the underwriting versus the execution of the related swap. Sometimes they could do this and sometimes they could not. Sometimes the structures were so complicated that it was difficult to disaggregate.

Over time it became a relatively competitive market where you had a variety of agencies able to issue and find a way to use the proceeds and the embedded options with a variety of underwriters and swap counterparties. For any one structure there were probably several prospective issuers and several prospective underwriters which led to a very competitive dynamic.

DS: Was the heyday of the structured note market in the early 1990s detrimental to the agencies? Although it provided very attractive funding, in the long term did it do a disservice to the agency market?

GLP: The perspective that we had at Fannie Mae was that the long-run health of the market was important to our balance sheet. Many of our callable structures were issued in large sizes with very transparent structures. It is also important to recall that the reason that market developed was that investors were looking to take on market risk, because of the low absolute level of rates. However, the leveraged buyout and related fever of the late 1980s had made them very concerned about event risk on the credit side. It was clear to most participants in the market that the reason investors were buying strong credits was as an effective means of creating a market exposure without taking on a lot of credit risk.

Did some investors delude themselves into thinking that low credit risk meant low market risk? Probably. Did some of these issues become fodder in the hands of poorly managed investment entities? Probably so. Did the agencies always do enough work to know who was buying their paper? Probably not. But the competitive pressures were very strong.

DS: Fannie Mae must have been subject to some fairly strong sales pressures from different Wall Street shops. Bearing that in mind, did the Proctor & Gamble/Bankers Trust scandal surprise you, or did you think that somewhere along the line something like that was bound to happen?

GLP: Fannie Mae was quite knowledgeable about why we were using derivatives. Derivatives were always being used to bring about exposures that we otherwise would gladly have created on our balance sheet with cash instruments had they been available, or had they been available at the same sort of cost. But the dealers did not typically inquire about our strategy, nor were we necessarily forthcoming about it.

It is not surprising to me that many end-users whose principal business is not the management of financial risks were convinced about the value of taking on financial exposures to achieve extraordinary gains. They were neither well-advised by outsiders nor particularly well-regulated by their own corporate management. I do not believe that the dealer community could be expected to take on that governance role. Whether the dealers became a problem as well as simply not being part of the solution is really a matter for others closer to the situation to judge.

DS: On to your job at the World Bank. What are your sub-LIBOR targets these days?

GLP: That changes slowly over time. I think that at the World Bank we definitely have very strong motivations to obtain sub-LIBOR funding. The level is not picked out of the air, but based on what it is we perceive we can achieve for the volume of funding we need to do. It actually varies from year to year in terms of several basis points, sometimes tens of basis points, because of what is available in the structures and markets in which we operate. We certainly have aggressive targets, but only aggressive in terms of what we think our credit, our primary and derivatives market capacities and our volume requirements would allow us to achieve. But we are also very scrupulous about not saying what that target is, because otherwise we would end up dictating the target and perhaps lose touch with the market.

DS: Has the World Bank set an amount to be raised in 1996­1997?

GLP: In the fiscal year just begun, 1997, we have estimated that our borrowing may be a little bit elevated from the last couple of years, at $13 billion to $14 billion. The difference is that the currency composition will become increasingly less predictable as we are offering our borrowers greater flexibility (see Derivatives Strategy, July/August 1996). In fact we have told people to expect a much less predictable program from the bank with more mid-course changes in direction.