تاثیر معامله هزینه در کشف قیمت: شواهدی از میان فهرست شده قرارداد شاخص معاملات آتی سهام

Fleming et al. [J. Futures Markets 16 (1996) 353] hypothesise that ‘price discovery will tend to occur first in the lowest-cost market, as information-based trades are executed where they produce the highest net profit’. This paper exploits the institutional differences between Nikkei 225 Stock Index futures trading on the Osaka Securities Exchange (OSE) and Singapore International Monetary Exchange (SIMEX) to provide new evidence on the ‘transaction cost hypothesis’. The institutional differences between the two markets result in higher brokerage commissions and margins on the OSE relative to SIMEX. The analysis reported in this paper finds that both SIMEX and OSE Nikkei futures returns lead Nikkei 225 Index returns. However, a direct comparison of the rates of price discovery between the two futures contracts demonstrates that SIMEX futures returns strongly lead OSE futures returns. This evidence is consistent with the transaction cost hypothesis, and corroborates findings of earlier studies.

A large volume of research has found that a temporal or lead–lag relation exists between returns on stock index futures contracts and their underlying equity indices.1Fleming et al. (1996) propose that this relation arises due to differences in the cost of trading between markets. They argue that the activities of informed traders are concentrated in the lowest cost market which yields the highest net profit on their information. Consequently, new private information is first traded on (and impounded) in the lowest cost market followed by related markets in ascending order of costs. This implies that price discovery occurs more rapidly in lower cost markets. While prior research has provided tests of the transaction cost hypothesis by examining the lead–lag relation between derivative markets and their underlying assets, a further test can be conducted by examining temporal relations between similar futures contracts.2 As implied by Kim et al. (1999), such an examination provides a more internally valid test of the impact of transaction costs as the securities tested are more homogenous. This paper employs both sets of tests and examines the cross-listed Nikkei 225 futures contract.3 The Nikkei 225 futures contract listed on the Nikkei Stock Average (also called the Nikkei 225 index) provides a natural experiment to test the transaction cost hypothesis because it is traded simultaneously in three markets around the world: the Singapore International Monetary Exchange (SIMEX), Osaka Securities Exchange (OSE) and Chicago Mercantile Exchange (CME). Of particular importance to this paper, at least two markets exhibit significantly different transaction costs arising from institutional differences. First, the brokerage markets in Japan and Singapore are regulated differently yielding significantly different brokerage charges. Second, margin levels also differ significantly between the exchanges. Prior research examining temporal return relations for Nikkei 225 futures is relatively scarce. Papers such as Tse (1999) and Iihara et al. (1996) demonstrate that OSE futures lead the underlying index using a sample of data prior to 1992. In contrast, Swinnerton et al. (1995) also examine OSE Nikkei futures and the underlying stocks in the Nikkei 225 index and find that the futures play little, if any, price discovery role for a sample of tick data drawn from 1990. Lim (1992) examines the lead–lag relation between SIMEX Nikkei 225 futures and the underlying index using data sampled between 1988 and 1989 and finds that neither the futures nor the index lead or lag each other. Consequently, there are a number of discrepancies in the results of prior literature. Furthermore, the sample periods examined in previous research are close to contract inceptions and may be influenced by market maturation effects.4 The only studies to examine lead–lag relation between the Nikkei 225 futures contract on SIMEX and the OSE are Booth et al. (1996) and Shyy and Shen (1997). While Booth et al. find no evidence of a lead–lag relation between the contracts traded on the OSE, SIMEX and the CME using a daily observation interval. Shyy and Shen (1997) find evidence of 1-minute bi-directional causality between OSE and SIMEX Nikkei 225 futures. Their data, however, is limited to 13 trading days in November 1993. In this paper, the lead–lag relation between the underlying Nikkei 225 stock index, OSE Nikkei 225 futures and SIMEX Nikkei 225 futures is estimated using a larger set of intraday data drawn from a more recent period. The remainder of this paper is organised as follows. The next section presents a discussion of the relevant institutional characteristics of the Osaka Securities Exchange and Singapore International Monetary Exchange. Section 3 provides a discussion of theory and development of hypotheses. Section 4 describes the data and method used to test the hypotheses, while Section 5 reports the results. The final section provides a summary and conclusion.

This study provides the results of a new test of Fleming et al. (1996) transaction cost hypothesis using cross-listed Nikkei 225 futures traded on the Osaka Securities Exchange 148 A. Frino, A. West / Pacific-Basin Finance Journal 11 (2003) 139–151 and the Singapore International Monetary Exchange. In their paper, Fleming et al. (1996) argue that where two or more markets trade similar products, informed traders will transact in the market with the lowest transaction costs in order to maximise profits generated from trading on their information. While this conjecture has been previously tested by examining the lead–lag relation between derivative securities and their underlying assets, this study examines feedback between two futures contracts based on the same underlying asset in addition to the previous approach. It is demonstrated that the costs of trading SIMEX Nikkei 225 futures are significantly lower than trading a similar nominal exposure using OSE Nikkei 225 futures. The cost differences are attributable to lower margin requirements on SIMEX as well as the existence of negotiated brokerage commissions versus the fixed rate regime operating in Japan. Following Fleming et al. (1996) therefore, the primary hypothesis of this paper is that returns on SIMEX Nikkei 225 futures will lead returns on OSE Nikkei 225 futures. By examining the temporal relations between the Nikkei 225 index and OSE Nikkei 225 futures; the Nikkei 225 index and SIMEX Nikkei 225 futures; and between OSE and SIMEX Nikkei 225 futures directly, evidence supporting the transaction cost hypothesis is provided. It is demonstrated that while both OSE and SIMEX Nikkei futures lead the underlying index, returns on SIMEX Nikkei futures lead returns on OSE Nikkei futures. It is concluded that these results are directly attributable to the higher cost of trading on the OSE discouraging informed traders. We conclude that it takes longer for a given information set to be impounded in more costly markets.