Margin Requirements, Price Fluctuations, and Market Participation in Metal Futures
Authors: Gikas A. Hardouvelis, Dongcheol Kim
Journal: Journal of Money, Credit and Banking, August 1995, 27: 659-671
THE OCTOBER 1987 STOCK MARKET CRASH renewed interest among regulators and economists in using margin requirements as a tool of controlling excessive speculation in cash and futures markets.
For example, the 1988 Brady Report’s recommendations of consistent margin requirements across related markets led to a heated debate on the desirability of permanently higher margins in futures contracts.
Permanently higher margins would be desirable only if they enhance the efficiency of the price mechanism by reducing mispricing and excess volatility. However, economic theory does not make unambiguous predictions on the consequences of higher margins.
Economists who believe that financial markets are often dominated by irrational speculators would tend to favor a permanent margin increase, whereas economists who believe that financial markets are dominated by rational investors would view a permanent increase in margins as harmful.
Empirical analysis is needed. Empirical work in cash markets has not reached unanimous conclusions. Hardouvelis (1990) reports a negative association between margins and volatility, excess volatility, and deviations of stock prices from fundamentals in the United States over we period 1935-1987. Moreover, Hardouvelis and Peristiani 91992)show mat margins have a clear price-stabilizing influence in the post—World War II Japanese stock market.
On the other hand, others, including Salinger (1989) and Hsieh and Miller (1990), argue that the U.S. evidence is not strong enough to support either view. Empirical work on the effects of futures margins is voluminous. Examples are Nathan (1967), Tomek (1985), Breeden (1985), Hartzmark (1986), Fishe and Go!berg (1986), Fishe et al. (1990), or Ma, Kao, and Frohlich (1993). This litera-ture finds a negative correlation between margins and market participation (open interest), but an ambiguous correlation with volatility. However, in futures markets, unlike cash markets, it is extremely difficult to use volatility in order to discriminate between the two alternative hypotheses.
The difficulty originates from the decision rules that futures exchanges follow. Unlike cash markets, where volatility has not historically influenced the decision to change margin requirements, in futures mar-kets, the exchanges systematically raise (lower) margins in anticipation of higher (lower) future volatility. Thus, even if there is a negative causal effect from margins to volatility, it would be hard to uncover it because it would probably be swamped by the volatility trend that the exchange is (correctly) forecasting.
The sample selection difficulty does not deter us from taking another look at fu-tures margins in the present article. The article’s main innovation lies on the use of a benchmark set of contracts. For each target contract we examine, we use a benchmark sample of other related contracts that do not undergo a similar margin change and, by comparing the behavior of target and benchmark metals, we assess more precisely the presence or absence of true causality from the change in margin requirements to the target contract.
To our knowledge, this is the first study that uses a benchmark sample to explore the causal influence of futures margins. The ap-proach is partially successful. It does reveal, for example, that the observed nega-tive relation between margins and market participation is causal running from the former variable to the latter, something the previous literature was unable to show. For reasons we explain later (footnote 7), our approach is less successful at discrimi-nating between the rational and irrational investor stories.