Continued rapid growth of derivatives contracts at the pace of the past several years would begin to raise the mania flag. Even in a global financial marketplace there must exist a finite limit to shiftable risk.
- Jerry Jordan, 1995
As a method of hedging against declines in portfolio value, exchange-traded and OTC options offer several advantages over synthetic portfolio insurance as implemented with dynamic hedging. First and foremost, these option instruments are binding contracts, rather than strategies implemented as circumstances demand and allow. The option holder is thus assured the purchased level of protection - as long as the option issuer remains solvent.
Second, the price of protection is known up front. Unlike synthetic portfolio insurance, where the cost will reflect the volatility experienced over the life of the insurance strategy, option prices are set, either by competition in the listed option markets or by issuers of OTC options, to reflect the volatility expected over the life of the option. Whereas portfolio insurers using dynamic hedging assume the risk that volatility may be greater than expected, purchasers of portfolio puts shift this risk to the option counterparty.
Third, to the extent that it requires less trading of assets and to the extent that the demand for and the cost of protection are more transparent, portfolio insurance undertaken with options may pose less of a threat to market stability than portfolio insurance undertaken via dynamic hedging. Whether these conditions hold remains a moot point. There are at least two major reasons for concern.
One concern is that information about privately negotiated OTC options is not as readily available as that on exchange-traded options. For exchange-listed options, trading volume and price are a matter of public record. But where can one turn to get information about private OTC contracts? For corporate reporting purposes, most derivatives contracts are considered off-balance-sheet, hence not included in financial statements, except perhaps in a footnote. (The Financial Accounting Standards Board says it is going to begin requiring companies to report the fair market value of derivatives contracts on corporate balance sheets, staring Dec. 15, 1998; gains or losses on derivatives used for hedging may be reported in earnings in the same period in which gains or losses on the underlying hedged positions are recognized. Any improvement in market participants' awareness of the demand for portfolio insurance brought about by the increased use of listed options must be balanced against the lack of information about portfolio insurance using OTC options.
A second concern is the nature of the hedging undertaken by a dealer selling OTC options. In particular, the issuer of such options may itself use dynamic strategies to hedge the risks it has assumed. As John O'Brien said in the July 1993 Wall Street & Technology: "that organization usually is doing what suppliers of portfolio insurance were doing before 1987."
Unlike the pea in the notorious shell game perpetrated on unsuspecting tourists in the Big City, the risk covered by portfolio insurance puts is not so easily made to disappear.
Risks to insurance buyers
For exchange-listed options, the exchange clearing corporation serves as guarantor of the contract. With risk of default diversified across all exchange members, there is little likelihood that a put purchaser will be unable to exercise the option because of counterparty insolvency. With OTC options, however, the counterparty is no longer an exchange, but the individual issuer (generally a securities dealer). While it may remain unlikely any one firm will default over the life of any given contract, there is obviously a greater possibility of default for an individual firm than for the amalgam of firms constituting an exchange.
For the investor that has only a single counterparty contract, there is an ever present danger of that counterparty going under. The credit quality of one's counterparties is thus of prime concern for insurers purchasing OTC options. Wall Street firms suffered a steady deterioration in credit quality in the 1980s, as measured by Moody's and Standard &
Poor's ratings. Declining credit ratings hampered firms' ability to
attract customers in the lucrative area of OTC derivatives (including equity options, but most especially the much larger markets in interest rate and currency swaps and options).
Some firms sought to ameliorate this problem by setting up "special purpose vehicles," independent subsidiaries designed to deal only in OTC derivatives. SPVs are run by their own managers and directors, separate from those of the parent company, and are subject to special operating and accounting safeguards and ongoing independent audits.
Most importantly, their levels of capitalization are high enough to justify superior credit ratings.
Goldman Sachs Financial Products International, Merrill Lynch Derivative Products and Salomon Swapco all have obtained triple-A credit ratings. Thus buyers of OTC options for insurance purposes can feel reasonably comfortable their options will be honored, provided they monitor the credit ratings of their counterparties and seek to purchase options only from entities with superior ratings.
OTC option issuers, of course, will charge a premium for shouldering the risks that put purchasers lay off. Barry Schachter notes that one issuer he examined sold OTC options at price levels 45% above their theoretical values. In other words, the customer was being charged almost 11/2 times what it would cost the issuer to assemble or replicate an option position affording the specified level of protection. And, unfortunately, a portfolio insurance purchaser may find its attempts at comparison shopping thwarted, for two reasons.
First, OTC options are essentially proprietary vehicles tailored to the needs of individual customers. Prices are not publicly quoted and, even if they were, noncomparability across different options would make price comparisons difficult. Seha M. Tinic, in a 1995 article in the Journal of Financial Services Research, says issuers themselves price the options "according to sophisticated theoretical financial models, which rely heavily on the ability to decompose complex contracts into such simpler components as options, futures, forwards, etc., for which the information necessary for valuation is widely available in the prices observed in the organized securities markets"
Potential OTC option purchasers must perforce seek recourse to the same methods in order to determine whether they are getting good value for their money.
In this endeavor, however, they (as well as the securities dealers selling the options) may find themselves stymied by the inadequacy of today's option pricing models. In a 1989 Journal of Applied Corporate Finance, no less an authority than Fischer Black has said:
"The Black-Scholes formula is still around, even though it depends on at least 10 unrealistic assumptions. Making the assumptions more realistic hasn't produced a formula that works better across a wide range of circumstances."
One of the more obvious disparities between modeled and actual option prices is the so-called volatility smile. That is, the Black-Scholes formula allows one to estimate the volatility of the underlying market by "plugging in" known market prices for options, their strike prices and their expiration dates. According to Black-Scholes, the volatility implied by this exercise in interpolation should be invariant to the precise relationship between option prices and their strike prices; that is, implied volatility should be the same whether an option is in, at, or out of the money. In the real world, however, implied volatility rises as the option's price moves away from the strike price - either into or out of the money - forming a smile pattern.
Mark Rubinstein, in the 1994 Journal of Finance, wrote the volatility smile is an outcome of the 1987 crash, a form of "crash-o-phobia" reflecting heightened awareness of the potential for outlier-type price changes and a consequent increased demand (and price) for protection. Jens Carsten Jackwerth and Mr. Rubinstein, using an optimization technique for estimating expiration-date risk-neutral probability distributions, find that the probability of a 3 (or a 4) standard deviation decline (in the S&P index) is 10 times more likely after the crash than before. Klaus Toft, in a 1994 paper, also sees the volatility smile as an outcome of the crash, but argues for an alternative explanation that pins the smile on an increase in overall financial leverage, as corporations were forced to issue more debt in the aftermath of the crash.
Attempts to develop option models that better reflect the actual behavior of option prices, including refinements of Black-Scholes, binomial tree and lattice approximations, and Monte Carlo simulations, constitute one of the hotter areas in finance today. It is not our purpose to go into these in detail, but rather to draw the reader's attention to their implication: option pricing is not a simple matter of plugging the right numbers into the right equations. The right equations (and the right numbers, for that matter) remain, to a non-trivial extent, matters of conjecture.
One substantial risk any purchaser of portfolio protection via OTC options must surely run is that of overpaying for a level of protection that could be attained at cheaper cost either from other OTC dealers or from exchange instruments.
Risks to dealers
OTC option issuers face their own versions of the risks confronted by the option purchaser. Furthermore, their risk-control task is complicated manyfold by the number and the variety of contracts in which they may be engaged. As we have noted, the value of OTC equity instruments is dwarfed by that of interest and exchange rate vehicles (including swaps, options, swaptions and forwards). The dealer selling OTC equity options designed to reduce the risk of one corporation's pension fund may also be intermediating an interest rate swap to reduce the cost of financing another's capital, and an exchange rate forward to reduce the risk of yet another's foreign operations or supplies.
While the volume and diversity of its counterparties necessarily make assessment of credit risk a more complex task for a dealer than for the typical end-user, they also afford the benefit of diversification. The typical dealer will be less susceptible to sustained damage from a default by any one of its multiple counterparties than will the end user with a limited number of counterparties. Furthermore, the dealer is more likely to be able to benefit by "netting" arrangements, whereby contracts with any given counterparty are pooled so that positively and negatively valued contracts offset each other. With netting, a counterparty cannot choose to default on those contracts under which it owes the dealer while continuing to collect on contracts under which the dealer owes.
Dealers of OTC contracts also generally apply a number of formal boundaries in order to control credit risk. These include limitations on exposures to individual counterparties and on aggregate exposures to given credit rating categories and to given countries of counterparty origin. Dealers, like end-users, will require that their counterparties have relatively high credit ratings. Vijay Bhasin finds the average credit quality of derivatives users was significantly better than that of all firms with senior debt ratings. He also finds, however, the gap between the credit quality of derivative users and the average rating of all firms has narrowed over time.
Mr. Bhasin does not look at contract specifications, which may contain requirements that mitigate the increases in credit risk. Dealers may require posting of collateral or periodically mark contracts to market in order to offset perceived increases in default risk because of counterparty credit quality or the nature of the contract. The International Monetary Fund's 1993 survey of banks issuing interest rate and currency derivatives found they were beginning to mark contracts to market and require periodic margin payments to offset the increased risk of their longer-term contracts. It also found securities dealers charged higher premiums, in the form of wider bid-asked spreads, when dealing with riskier counterparties.
Derivatives themselves may reduce the probability of a counterparty default. As Ludger Hentschel and Clifford Smith Jr. point out in the Journal of Financial Engineering in 1995, there are two necessary conditions for default: The counterparty owes a payment on its derivatives, and the counterparty is insolvent. They argue that, to the extent the derivatives are used to hedge (and not for speculation), the probability of counterparty insolvency will be reduced.
Credit exposure for dealers may also be ameliorated by the nature of the relationship between macroeconomic factors and the value of derivatives contracts (although this relationship may also serve to heighten exposure). As John Hull, in a 1989 Journal of Financial and Quantitative Analysis, argues in regard to interest and exchange rate swaps, if bankruptcy becomes more likely when interest rates rise, then the exposure of interest rate swap portfolios may be stabilizing insofar as the counterparty paying fixed and receiving floating generally has the higher credit risk. Gregory R. Duffee, however, in a 1994 Finance & Economics Discussion Series, finds that, historically, default is more likely in periods of falling, rather than rising rates, so credit risk is increased for the counterparty paying fixed.
The task of assessing exposure to credit risk is nevertheless a difficult one for dealers (as well as end-users), inasmuch as it involves not only estimating counterparty default probabilities over often multiyear horizons, but also estimating the potential behavior of the derivatives over changing economic environments. Shortcomings in available models as well as unavoidable errors in forecasting model inputs make valuation of derivatives far from an exact science. Standard & Poor's Credit Week of November 1992 noted: "In general, the models and the systems' capability for tracking credit exposure are in a catch-up mode, and have experienced difficulty keeping up with the growth of the business."
The chore is further complicated by the now seemingly ubiquitous use of derivatives: "Participation in derivatives markets can cause firms to become connected through complicated transactions in ways that are not easily understood, making the evaluation of counterparty risk extremely difficult," according to an International Monetary Fund paper.
Dealers must also manage the market exposures associated with their derivatives positions. In selling a put on an equity portfolio, for example, a dealer places itself at risk of a market decline that will force it to purchase the securities at above-market prices, just as a dealer that enters into a swap to pay a counterparty a fixed interest rate in exchange for receiving a floating rate is at risk of a decline in interest rates leaving it with a negative cash flow. Many firms estimate the market risk of their derivatives positions (derivatives' responses to changes in underlying markets) by estimating value at risk. This is the loss in value over a given horizon that may be exceeded with a small probability. VAR estimates rely on models of the probability distribution of returns and their volatility, and often take into explicit account the possibility of price jumps. Again, VAR estimates are only as solid as the models and variables used to derive them.
(contd. in Part 4)