کیفیت خدمات بهینه تحت انحراف از ارزشگذاری هزینه حمل: شواهد از شاخص معاملات آتی بازار سهام اسپانیایی

We provide an analytical discussion of the optimal hedge ratio under discrepancies between the futures market price and its theoretical valuation according to the cost-of-carry model. Assuming a geometric Brownian motion for spot prices, we model mispricing as a specific noise component in the dynamics of futures market prices. Empirical evidence on the model is provided for the Spanish stock index futures. Ex-ante simulations with actual data reveal that hedge ratios that take into account the estimated, time varying, correlation between the common and specific disturbances, lead to using a lower number of futures contracts than under a systematic unit ratio, without generally losing hedging effectiveness, while reducing transaction costs and capital requirements. Besides, the reduction in the number of contracts can be substantial over some periods. Finally, a mean–variance expected utility function suggests that the economic benefits from an optimal hedge can be substantial.

Since its launching in January 1992, the Ibex 35 futures contract quickly became the most actively traded derivative contract in Meff Renta Variable, the Spanish equity derivatives exchange. In fact, the futures market on the Spanish Ibex 35 stock index is also one of the most active futures stock index markets in the world. Acceptance of a market for a stock index futures contract is related to the hedging ability of this derivative instrument. Operating with futures, it is not only possible to guarantee a certain profit, but to also bound the losses obtained over a given time period. Hedging spot positions in the Spanish stock market became especially relevant in recent years, because the systematic decrease in interest rates as a consequence of the fiscal and monetary policies aimed to achieve the European Union, caused a dramatic reallocation of private savings from riskless assets to stock exchange positions. The relevant issue in a hedging operation is to determine the hedge ratio, which provides the number of futures contracts that must be sold to counteract the opposite evolution in spot prices, so that, the potential losses in one market can be offset by the gains obtained in the other. A biased estimation of the hedge ratio implies that the losses in one market will be higher or lower than the profits in the other one. This is troublesome for a hedging strategy, whose aim is to transform a position in the spot market into a riskless portfolio. According to the cost-of-carry valuation (the standard forward pricing model), which assumes perfect markets and non-stochastic interest rates and dividend yields, the theoretical price at time t (Ft,T*) of an index futures contract maturing at time T equals the opportunity cost of keeping a basket replicating the spot index between t and T: equation(1) View the MathML source where St is the index value and (r−d) is the net cost of carry associated to the underlying stocks in the index, i.e., the riskless rate of return minus the dividend yield of the stocks in the index. Alternatively, Eq. (1) can be written equation(2) View the MathML source where rs,t=ln(St/St−1) and rf*,t=ln(F*t,T/F*t−1,T), the spot and theoretical futures returns, respectively. Under the previous assumptions, the relationship in (2) implies that: (a) the variance of returns in the spot market equals the variance of returns in the futures market, (b) the contemporaneous rates of return of the underlying stock index and the futures contract are perfectly and positively correlated, and (c) the non-contemporaneous rates of return are uncorrelated and no lead–lag relationships between returns should appear. However, in the presence of market imperfections such as transactions costs, asymmetric information, capital requirements and short-selling restrictions, there could be discrepancies between the traded futures price and its theoretical valuation according to the cost-of-carry model (see Mackinlay and Ramaswamy, 1988; Lim, 1992; Miller et al., 1994; Yadav and Pope, 1990 and Yadav and Pope, 1994; Bühler and Kempf, 1995; among others). Market imperfections may also produce a lead–lag relationship between spot and futures market returns, as well as between their volatilities. Then, it may be possible to anticipate price movements and risk fluctuations in one market from past information in the other market, a relevant question when using the futures contract as a hedging instrument for risky stock portfolios. In fact, there is a wealth of studies showing empirical evidence for the main international stock index futures markets supporting the existence of such lead–lag relationships (see, for example, Stoll and Whaley, 1990; Wahab and Lasghari, 1993; Pizzi et al., 1998; Iihara et al., 1996; Koutmos and Tucker, 1996; Racine and Ackert, 1999; among others). We start by providing empirical evidence in favor of significant mispricing in the Spanish stock index futures market. Assuming that the evolution of the stock index and the futures market returns are driven by heteroscedastic, geometric Brownian motion processes, we include a market-specific noise in the dynamics of theoretical futures returns. The motivation for such a noise is that, by itself, it produces a spread between theoretical and market futures prices, although such a hypothesis would only make sense when volatility in the spot and futures markets could not be summarized by a single factor. We use Engle and Kozicki (1993) approach to test for a single common ARCH factor between the two markets, conclusively rejecting such hypothesis. Hence, the two markets do not share an identical source of volatility, against the cost-of-carry model. We also provide empirical evidence for this model using data from 20/12/93 to 20/12/96 from the Spanish stock index futures market. A bivariate error correction model with GARCH perturbations is used to estimate the conditional second moments of market returns. Our model has the following characteristics: (a) it incorporates the long-run equilibrium relationship between spot and futures prices, (b) it takes into account the cross-market interactions between returns and volatilities, (c) it does not impose a constant conditional correlation coefficient in the matrix of second moments for market returns, a significant difference with most previous analysis (Park and Switzer, 1995; Iihara et al., 1996; Koutmos and Tucker, 1996; Racine and Ackert, 1999; Lien and Tse, 1999; among others), and (d) it captures the presence of an intraday U-shaped seasonal pattern for both spot and futures market volatility. Our model specification and technique estimation allow us to capture stochastically this intraday seasonal pattern for market volatilities rather than through deterministic variables, as it is standard in the literature. We estimate the model with hourly returns, using the nearest to maturity contract, to then recover estimates for the parameters in the theoretical model. Our estimates imply a less than perfect correlation between spot and futures returns, leading to an optimal hedge ratio below one to hedge the spot index portfolio, without losing any hedging effectiveness in ex-ante simulations of hedging strategies using actual data. The rest of the paper is organized as follows. The optimal hedge ratio under departures from cost of carry valuation is analytically derived in Section 2, and its main properties are discussed. In Section 3 we describe the data used in our analysis. We start Section 4 with some preliminary empirical evidence on regularities in returns and volatilities in Spanish spot and futures stock markets. We then present the econometric approach followed to estimate dynamic relationships across markets in conditional first and second order moments for returns, discussing the main results. In Section 5 we recover estimates for the theoretical parameters of the model. In Section 6 we make ex-ante simulations to investigate if taking into account departures from the theoretical cost-of-carry valuation enhances the hedging effectiveness of the futures contract. Finally, Section 7 summarizes and presents concluding remarks.

We have derived a two period hedging model allowing for departures from the cost-of-carry valuation of a futures contract on a stock index. Assuming a geometric Brownian motion for the dynamics of the spot index, we have modeled mispricing by introducing a specific noise in the dynamics of the theoretical futures price, possibly correlated with the noise common to both markets. The optimal hedge ratio is shown to depend on two factors: the relative size of the specific and common noises, and their correlation. A detailed analysis of this theoretical model shows that it can capture many interesting features of practical hedging situations, specially when the stochastic behavior of returns in the spot and futures markets widely differs. We have provided empirical evidence on the model using data from the Spanish stock index futures market over the sample period from December 1993 to December 1996. A bivariate error correction model with GARCH innovations has been used to estimate the parameters of the theoretical model, from which we have computed estimates for the optimal hedge ratio during this sample period. The model allows for transmission of returns and volatilities between both markets, showing that the futures market has a stronger influence on the spot market aspects than the other way around. Furthermore, we have provided significant evidence on intraday seasonality in volatility in both markets. This model should be a useful tool to discuss different characteristics of the dynamic relationship between spot and futures markets, beyond the implications for hedging exploited in this paper. Empirical results support that spot and futures markets do not have a common ARCH feature. Our findings suggest that there is a specific noise in the Spanish futures market with a small, negative correlation with the noise common to the spot and futures markets. A negative correlation in the theoretical model does not preclude futures market returns to be more volatile than spot market returns and, in fact, we observe that to be the case in 75% of the available data. Ex-ante simulations with actual data reveal that hedge ratios that take into account the estimated, time varying, correlation between the common and specific dis- turbances, lead to using a lower number of futures contracts than under a systematic unit ratio, without losing hedging effectiveness. Using less futures contracts for hedging implies lower transaction costs and smaller capital requirements. Besides, the reduction in the number of contracts can be substantial over some periods. Considering an investor with a mean–variance expected utility function, we have also shown that the economic benefits from an optimal conditional hedging can be substantial.