ESSAYS ON ISSUES THE FEDERAL RESERVE APRIL 2008 OF CHICAGO NUMBER 249

Chicag o Fed Letter

Hedges in the warehouse: The get trimmed by Brian Gordon, senior technical expert, Supervision and Regulation, and Adrian D’Silva, vice president, Financial Markets Group

When banks “warehouse” , that is, hold them temporarily before selling them through , how do they protect themselves against ? This article examines the effectiveness of hedging strategies in the context of the recent heavy losses in financial markets.

To Wall Street bankers, market conditions The originate-to-distribute business in the spring of 2007 looked close to model perfect. The world savings glut meant The largest and most sophisticated that liquidity was pouring into the U.S. banks in the world, which reported the fi xed income markets; interest rates were majority of the losses, use a common relatively low compared with historical business model known as “originate-to- averages; the U.S. economy was growing distribute” (OTD). These institutions strongly; and increasingly complex asset- originate loans for the specifi c purpose backed securities (ABSs) that banks had of selling them to others, typically developed as a way to trade loans to other through securitizations. This model were being sold as fast as they was widely considered to be benefi cial Banks typically use hedging could be created. for the banks because it moves off techniques to protect the the banks’ balance sheets and distrib- Then suddenly the fi nancial world utes them through the fi nancial system value of their assets. However, changed. Banks that had been posting among investors who can match their hedging is a diffi cult art where record profi ts started reporting record risk tolerances with the return goals. losses. While there are a lot of well- best practices are still evolving. Subsequent events have revealed that the documented reasons for the abrupt banks actually retained more of the risks change in fortune, we focus on one than had originally been believed; for seemingly obscure, but in retrospect instance, the banks retained “hung deals,” important, part of the picture in this which are loans intended for securitiza- Chicago Fed Letter. One source of losses tion that cannot be securitized under (although certainly not the only source) current market conditions. They also for the banks is that the loans that they may not have fully hedged their risks. were intending to hold temporarily— or “warehouse” before selling them to The need for speed others—fell in value. Banks typically In late 2006 and early 2007, the banks’ use hedging techniques to protect the biggest challenge was not in selling the value of such assets against declines ABSs they had created; rather, their during the warehouse period. How- challenge was in creating enough ABSs ever, hedging is a diffi cult art where to meet the demand of investors. Thus, best practices are still evolving; and banks were more focused on upside risk because of the bull market conditions (not having enough inventory to sell) that prevailed when the banks bought than “spoilage risk” (inventory getting the loans, they may not have fully old and stale). The key to profi tability, hedged their positions. especially as the market became more aggressive, was to minimize the amount in inventory. The longer a asset stays during this period banks often found of time the assets were actually owned by in the warehouse awaiting sale, the great- that spreads had tightened during the the bank. Indeed, the warehouse periods er the risk that the loan will lose value warehouse period. This progressive tight- dropped from nine months to as for some reason. ening increased profi t margins by more as three months for some products. Grow- than had been anticipated. It then en- Spoilage risk in the warehouse can affect ing competition among banks for loans couraged greater risk-taking by banks single loans and groups of loans. For led to increasingly aggressive bidding for because the more assets they booked, example, an individual loan can have an new loans from originators, with pur- the more money they made. increase in default risk. A group of loans chasers sometimes doing little or no due can be adversely affected when the mar- The OTD business model depends on diligence before submitting bids. Another ket moves against them (whether because a liquid ABS market because securitiza- tactic was to vertically integrate by buy- of interest rates or a general increase in tion is the primary distribution channel ing out suppliers of loans to lock away ). Both risks potentially reduce to investors. Thus, the bank is exposed production from the competition. the value of the of loans. In- to risk if the credit market becomes more Credit market turmoil terest rate risk is fairly well understood conservative or if the ABS market be- and relatively straightforward to , comes less liquid. Both conditions arose At the beginning of 2007, banks were an- leaving credit risk as the major risk of in the second half of 2007. Banks that ticipating record sales volume for securi- warehoused loans to the bank. had originated loans in late 2006 or early tized assets, so they were ramping up the 2007 in an aggressive credit market found amount of assets in their warehouses. Credit risk to single loans and groups of that they suddenly could not sell the They were also taking more risk with loans comes in two forms: spread risk and assets for a profi t, if they could sell them those assets. As the market became default risk. Spread risk is the larger at all. They were forced to liquidate the frothier, their ability to sell assets at and more volatile of the two. Default loans at a steep discount to what the bank tighter and tighter spreads increased. risk is typically lower because most of believed was their “intrinsic” value, or they could hold the assets and hope for By mid-2007, banks that had not hedged or had hedged a market rebound. Still, doing the latter required them to recognize an account- insuffi ciently found their downside unprotected. ing loss because the market was now sim- ply unwilling to pay what it had been This encouraged some to originate ever- the warehoused loans are new and few willing to pay in the recent past. more-marginal assets. In addition, since loans default soon after closing (sub- Hedging the warehouse their focus was on production and sales, prime mortgages that are fi rst payment hedging these warehouse lines against defaults being the notable exception). Hedging helps banks avoid losses by buy- downside risk, which was diffi cult and ing protection against a fall in the value Spread risk is the risk that the market will expensive anyway, did not seem as of the assets. However, banks did not become more conservative in its attitude high a priority. fully hedge their warehouse credit risk. toward credit risk between the time the Hedging warehouse credit risk is diffi - When the loan is funded and the time the loan is cult and expensive, especially in an en- began to affect the broader fi nancial sold. For example, a loan booked at a vironment where the possibility of large markets in mid-2007, suddenly the strat- spread of 150 basis points (bps) over short-term profi ts created by the bull egies that had been richly profi table Treasuries could be sold into a securiti- market tended to reduce risk managers’ became losers. As the markets tumbled, zation at a lower spread, say, 125 bps, focus on downside risk. so too did the value of the warehoused because the structure typi- loans. Banks that had not hedged or cally provides the with benefi ts, Hedging warehouse lines is notoriously had hedged insuffi ciently found their such as risk matching and liquidity, that diffi cult because of , which low- downside unprotected. offset the tighter spread. The difference ers hedge effectiveness or effi ciency. in spreads produces a 25 bps profi t for Basis risk is the risk that there is a diver- Presecuritization warehouses the bank because prices move in- gence between the asset that is being Securitizations are sold as pools of assets. versely to yield (spread is a component hedged (protected from loss) and the Banks accumulate a large amount of of yield). Alternatively, if credit risk in- asset providing the hedge (typically a assets for relatively short periods as they creases and the spread widens to 175 bps, ). This divergence is prepare a pool of assets for securitization. then the bank loses 25 bps. It is spread the result of a mismatch between the During these periods, the assets are widening (increase) that banks are prin- price movement of the asset being stored in the bank’s “warehouse,” which cipally seeking to hedge. hedged and the derivative contract used is conceptually similar to a physical ware- to hedge it. Hedge effi ciency is measured The OTD market began to grow quickly house used to store physical goods. Like by the correlation between the asset after 2002, and in that time, market toler- vegetables stored in a physical ware- being hedged and the hedge itself. ance for risk steadily increased. In fact, house, loans have “spoilage risk” while When effi ciency is high (strong negative correlation), the move in the products to hedge warehouse positions. customized swaps that may also in- opposite directions with the same speed These include basis risk, crease counterparty credit risk and are and magnitude. When effi ciency is low, issues, counterparty risk, and liquidity expensive to use because they are so there is little relationship between the risk. The hedge has basis risk because customized. They may also have basis movements of the two contracts. ABSs in the underlying index do not risk if the index returns diverge from the underlying assets being hedged. Thus, many banks that use these only The longer a loan asset stays in the “warehouse” awaiting sale, hedge part of their position and retain the greater the risk that the loan will lose value for some reason. some risk. For the same reasons that hedging is diffi cult and ineffi cient, it is also expen- Index products and related problems have the same risks as a pool of whole sive. For example, hedging a warehouse loans; thus, there will be The challenge is fi nding a credit deriv- line with a TRS is intuitively appealing, between the two, which reduces hedge ative contract with the following but the hedging counterparty might be effi ciency. This lower hedge effi ciency characteristics: expected to take on prepayment risk, means that the bank has a harder time which is diffi cult to estimate, and would • Its market value will move in the op- achieving treatment want compensation for that risk. posite direction to that of the ware- under FAS 133, which is the accounting housed loans; and guidance that governs hedge treatment. Default risk Failure to achieve hedge treatment can • Its market value will move with the The same forces that cause spread risk result in unwanted earnings volatility from same speed and magnitude of change. to increase (a general decline in eco- an accounting perspective, even if the nomic conditions) also increase the To meet this challenge, a series of index hedge achieves its economic goals. Since default risk of individual assets. This products have been developed over the banks are less likely to get the accounting risk increases the longer an asset is held past several years, notably the CDX (for benefi ts of hedge treatment, they are on the warehouse line. Banks can con- corporate bonds), LCDX (for corporate probably less likely to pursue it. loans), ABX (for ABSs, including those trol this risk by buying credit protection backed by subprime loans), and CMBX In addition, the index products are trad- on individual exposures in the ware- (for commercial mortgage-backed secu- ed over the counter, as opposed to on house. However, that strategy is available rities, or CMBSs). Each index takes the exchanges. This means that the bank is only for a small number of loans (only form of a synthetic credit default exchanging credit risk on a pool of loans the largest borrowers have single name that references roughly 20 underlying for counterparty credit risk—the risk credit derivatives on offer), is expen- securities that have a uniform credit rat- that the other party to a trade will not sive, and is prone to many of the same ing at the time the series begins (e.g., meet its obligations. Generally, since most BBB as of January 1, 2008). The index trades are made with large dealers, this tranching (or slicing) helps because each risk is reduced, but it adds a complicating Charles L. Evans, President; Daniel G. Sullivan, Senior Vice President and Director of Research; Douglas loan in the OTD business model will ul- element to the transaction. Also, the in- Evanoff, Vice President, fi nancial studies; Jonas Fisher, timately be sliced into tranches (classes dex products can be prone to illiquidity Economic Advisor and Team Leader, macroeconomic and to an imbalance of buyers and sell- policy research; Richard Porter, Vice President, payment of bonds) in its fi nal securitized form. studies; Daniel Aaronson, Economic Advisor and ers, which can distort prices. So for all Team Leader, microeconomic policy research; William Since the warehousing bank is in a nat- of these reasons, warehouse risk man- Testa, Vice President, regional programs, and Economics ural “” credit position, meaning it Editor; Helen O’D. Koshy, Kathryn Moran, and agers often will not use the index prod- Han Y. Choi, Editors; Rita Molloy and Julia Baker, holds the assets, the bank needs an op- ucts for hedging purposes. Many say they Production Editors. posite (or “short”) position to protect are better suited for speculative trading Chicago Fed Letter is published monthly by the against a loss in value of the assets held— than they are for hedging. Research Department of the Federal Reserve so the short position needs to increase Bank of Chicago. The views expressed are the authors’ and are not necessarily those of the in value as the credit risk of the assets Total return swaps Federal Reserve Bank of Chicago or the Federal Reserve System. increases. Using one of the index prod- A total return swap (TRS) is another ucts seems logical for this task because © 2008 Federal Reserve Bank of Chicago derivative product that is used for hedg- Chicago Fed Letter articles may be reproduced in it is possible to get a short position in a ing. Under this contract, the bank pays whole or in part, provided the articles are not that is made up of assets that the total return on an index (including reproduced or distributed for commercial gain closely resemble the ones being hedged. and provided the source is appropriately credited. mark-to-market gains and losses) on a Prior written permission must be obtained for For example, a bank with a CMBS ware- loan or group of loans in exchange for any other reproduction, distribution, republica- house line could use a short CMBX posi- tion, or creation of derivative works of Chicago Fed a fi xed payment. This fi xed payment Letter articles. To request permission, please contact tion to hedge its position. locks in the bank’s return and provides Helen Koshy, senior editor, at 312-322-5830 or an effective hedge. The TRSs are not email [email protected]. Chicago Fed Unfortunately, there are a number of Letter and other Bank publications are available without problems, however, as they are on the Bank’s website at www.chicagofed.org. practical challenges to using the index ISSN 0895-0164 illiquidity problems that the other hedging can be viewed as an unnecessary Given that the OTD business model had hedging products have. that reduces profi ts. matured only since the last (in 2001), banks had not lost money on In the wake of the credit turmoil of 2007, Conclusion their warehouse exposures to this extent securitization volumes have substantially The credit turmoil of 2007 was unprec- in the past. However, the importance of declined. Banks now fi nd that they have edented in many ways. This turmoil is hedging to protect downside risk is not originated assets during times of aggres- not yet completely resolved, so it may be new. Arguably, banks should have been sive underwriting conditions that are too early to list the lessons learned from able to foresee a cooling of the markets now aging on their warehouse lines. the events of 2007. New risks emerged, and a need to hedge downside risk, even This combination puts further down- and they have not yet been completely without being able to anticipate the ward pressure on values. evaluated. But some old risks, which precise course of events. might have been controlled, played a Incentives Hedging credit risk assets is both diffi - part as well. Another factor that discourages banks cult and expensive. In the most recent from hedging is incentives. Investment Banks lost money in ways that did not bull market, protecting downside risk bankers generally get a large amount seem plausible given market conditions seems to have become a lower priority of their total compensation in the form as recently as the spring of 2007. While at some banks. However, when the mar- of a bonus. The amount of the bonus certainly not the only source or even the ket turned, the lack of downside protec- is typically tied to a bank’s revenue or major source of losses, warehoused assets tion exacerbated the banks’ OTD losses. profi tability in a given year. The incen- awaiting securitizations that could not Incentives for bank managers, which tend tive system tends to be asymmetric with be completed became a problem for to reward short-term profi ts over long- respect to risk because bonuses scale up some banks. When the market value of term risk-adjusted performance, may rapidly as short-term profi ts increase, the warehoused assets fell, the banks have played a role as well. whereas the downside risk (being were forced to recognize losses. fi red) is fi xed. In a strong bull market,