GEORGE HALL, CLINTON GROUP-JULY 1998|
George Hall began his professional career as a nuclear engineer for the Tenneco Corporation. After receiving his MBA from Wharton in 1986, he joined Citicorp Investment Bank where he directed mortgage securities trading. From April 1989 to April 1991, he was a Vice President at Greenwich Capital Markets, a subsidiary of The Long Term Credit Bank of Japan, where he headed the mortgage arbitrage group. In August 1991, he founded the Clinton Group with backing from Itochu to specialize in trading fixed income securities using arbitrage strategies. As the client base evolved, various programs were set up using different degrees of leverage and other limitations. In July 1994, the firm started its first pooled vehicle for outside investors, Battery Fund Ltd. and, in September 1996, it started a second offshore hedge fund, Trinity Fund Ltd., a more aggressive version of the same investment program. Through its history, the Clinton Group has provided consistent net returns ranging from 15% to 45% annually (depending on the programs) without a single losing month. With $1.5 billion under management, the Clinton Group stands as one of a handful of firms that have successfully weathered the various storms that periodically affect mortgage-backed securities arbitrage firms. George Hall spoke with HFN publisher, Antoine Bernheim, in mid-July 1998.
Profile of George Hall
Born: September 10, 1960 in New York
Education: BS, U.S. Merchant Marine Academy; MBA,
Wharton School, University of Pennsylvania
Last vacation: Grand Cayman
Last book read: Master of the Game: Steve Ross and the Creation of Time Warner by Connie Bruck
Hobbies: Boat racing
Favorite quote: "Those who do not study history are doomed to repeat it"
How he best describes himself: "Prefers the road less traveled"
Q. Every now and then, mortgage traders get surprised by an unanticipated change in prepayment speed as happened in 1994 and again recently. How have you managed to side step these pitfalls?
A. The problem in 1994 and again recently is not so much what happened to prepayments as what happened to interest rates. In 1994, there was a dramatic rise in interest rates and a corresponding slowdown in prepayments which caused securities to lengthen and prices to go lower. More recently, prepayments have picked up because interest rates have fallen. However, nothing happened that could not have been anticipated. There may be the impression that the mortgage market performed poorly in 1994 or recently, but I don't think it performed poorly at either time. I think mortgage traders tend to make assumptions that rates won't fall or rise too dramatically and they leave a fair amount of risk at the end points. This is akin to being short out of the money options. That leads to pretty good performance in a benign market but whenever there is volatility, losses occur. When performance is good as in 1993 or 1996 and 1997, it does not necessarily mean that mortgage traders are doing a good job. If you take risks and the market cooperates, you may not lose any money but it does not mean you are hedged appropriately: a good track record may be for the wrong reasons. At Clinton Group, we thoroughly analyze our portfolio each day to try to be sure that we are hedged against the risks associated with significant rate changes. There is a lot of analysis and computing that goes on which requires sophisticated systems. But with the right tools, it is not really that difficult to avoid risk and you can achieve good returns without having exposure to these changes in the market.
Q. Could you give us a sense of the current state of the mortgage market in terms of CMO issuance, dealer participation and the buy side players?
A. Issuance has increased in the first six months and has been quite a bit stronger than in the last few years as a result of higher level of prepayments. When higher coupon mortgages prepay, homeowners get new mortgages with lower coupons and they get pooled into pass-throughs and cut up into CMOs. So, CMO issuance increases with increased prepayments. With regard to dealer participation, the market is pretty active. Actually, we have never really felt a lack of liquidity in the market, even in 1994. On the buy-side, there are mortgage hedge funds that have come about over the last few years, there are also insurance companies and money managers. Nearly all financial institutions are involved in the mortgage market.
Q. Could you describe your investment approach and how you choose the instruments you trade?
A. We try to value securities not based on how they will perform in a particular market but rather how they perform on an expected return basis ie: if you take the weighted average performance of the security over any interest rate path, you try to find a security that gives you a greater expected return than a security of comparable credit quality. We then try to figure out how to hedge, so that we can lock in this relative value. Virtually all the securities have positive duration and we try to figure out how to reduce the duration by selling Treasuries, swaps or using some type of option. We have to analyze a lot of securities to find ones that we think are underpriced. We avoid predicting supply and demand and only buy a security if there is a fundamental reason to do so.
Q. Could you describe how your organization functions?
A. We have 38 people altogether. All our research and systems development originates from our trading desk of 8 people. We have a staff of computer programmers and there is a collaborative effort between research and trading. We use our own proprietary systems for modeling mortgage securities, doing option analysis, analyzing hedges, analyzing the portfolio as a whole and monitoring exposure to rate changes. With any model there are many assumptions and the output has to be interpreted. You give yourself an advantage if you have created the systems and are familiar with the models and the assumptions. We have 10 analysts who look at available securities from Wall Street dealers' inventory. We run these securities through our models, mostly overnight, and when we come in the morning, we have various option pricing models, cash flow models and hedging models to review. We do that not only for securities that are available to buy, but also to compare them to the securities in our portfolio. We always try to buy new securities that complement the portfolio or that are cheaper than securities we can sell out of the portfolio. I have certain guidelines in terms of the types of securities and what thresholds of return we are looking for and I personally review every trade each day.
Q. Are there instruments in the mortgage market that you do not trade as a matter of policy?
A. Every instrument has value at a price, however there are many securities that do not have the relative value we are looking for. Over a period of time there are many sectors that we are not involved in, not as a matter of policy but as a matter of fundamental value.
Q. What is the mix between trading and investing?
A. Our portfolios generally turn over 100% to 150% a year. There are two sources of return, one is the spread and one is capital gain. We look to buy and sell securities every day and improve our position. There is only a certain number of counter-parties in the market and liquidity also puts a limitation on the amount of trading that can be done.
Q. How do you control risk and volatility and what is your approach to cutting losses?
A. First, we only use a small amount of leverage. Leverage multiplies losses and the main risk is being overextended and having to sell securities to meet margin calls. We always maintain the capability to buy more. The key to cutting losses is to try to anticipate the potential risk to a portfolio ahead of time. Once the market has made a move, all you can do is dynamically hedge against future losses.
Q. How do you price your portfolios?
A. Our pricing systems uses several sources of information to arrive at appropriate prices. First, we see where dealers are offering securities. This tells us the offered side of the market. Also, on any given day, we are aggressively trying to offer securities for sale so we know where the bid side is. We also keep very detailed records of our trades and trades we hear about. Every two weeks, we put out a certain cross-section of our portfolio and solicit actual bids. There are thousands of mortgage securities and all of them have their own nuances but if you get information on a particular security, you can extrapolate that information and pretty accurately price where a similar security would trade. In addition, since we do a fair amount of borrowing by using repurchase agreements, we get a price from the lender. We take all that information to arrive at our prices. We have been extremely accurate over the last six years.
Q. What is the principal source of risk in your strategy?
A. Of course, there is the risk that we make a mistake and our hedging strategies don't work properly. But we have been doing this for a long time and we are confident in our hedging strategies. There is also the risk of spreads widening and there is not much you can do about that. If there is a lot of selling and demand dries up for one reason or another, then the securities would underperform relative to the hedge instruments and we would have a loss. But I think that is a limited risk. Although we take advantage of small inefficiencies in he market, I think the mortgage market is efficient enough and there are enough people analyzing these securities that when spreads do widen, demand increases.u