Price discovery between regular and mini index futures in the Taiwan Futures Exchange

Abstract

This study explores the dynamics of the price discovery process between the regular and mini index futures traded on the Taiwan Futures Exchange (TAIFEX). In contrast to the US futures market, the TAIFEX operates under an automated electronic trading system for both types of contracts. After controlling for the differences in the trading mechanisms, we demonstrate that mini index futures make a greater contribution than regular index futures to the price discovery process within the TAIFEX. Regression analyses show that both relative liquidity and relative changes in liquidity between regular and mini index futures contribute to the price discovery of the mini index futures.

Introduction

As a result of globalization and innovations in information technology, futures exchanges around the world are frequently found to be introducing new financial products aimed at meeting the changing needs of investors and improving their overall competitiveness. For example, in September 1997, the Chicago Mercantile Exchange (CME) introduced a mini version of the S&P 500 futures contracts traded on an electronic system (E-mini), with subsequent rapid growth being exhibited in trading volume for these E-mini futures within the US.³

Both electronic and open-outcry trading systems have been adopted for many futures contracts, such as those of the S&P 500 and NASDAQ 100; however, E-mini futures contracts often simply involve a de facto reduction in contract size, since the contracts are generally smaller than those of floor-traded futures, although they have exactly the same underlying index. In general, the benefits of E-mini futures include more rapid execution speed, more accurate processing of transactions, lower trading costs and anonymity in execution.

Many of the proponents of E-mini futures claim that an electronic trading system enhances the price discovery process (see Ates and Wang, 2005a, Ates and Wang, 2005b, Gilbert and Rijken, 2006); however, those who oppose such claims argue that electronic trading eliminates the informational advantages possessed by market makers when engaging in floor trading, essentially because floor traders are able to obtain information from the order flow in the pit.⁴

From their empirical investigations into the evolution of the overall price discovery process between regular floor-traded futures and electronic-traded mini futures within the US, several of the prior studies have reported that mini futures generally play a dominant role over regular futures.⁵ It is argued that this dominant role attributed to E-minis is likely to be explained by the smaller contract size (Karagozoglu and Martell, 1999, Tse and Xiang, 2005) or the inherent differences between the trading mechanisms of the regular and E-mini futures markets (Ates and Wang, 2005a, Tse, Bandyopadhyay and Shen, 2006a, Chung et al., 2010).

Many other studies have examined the impact of trading mechanisms on information transmission within the futures markets; however, the findings are rather mixed. For example, Pirrong (1996) carried out analyses of the Bund (ten-year German government bonds), as well as futures contracts traded on the open outcry and electronic trading platforms in both the London International Financial Futures and Options Exchange (LIFFE) and the Frankfurt Deutscher Turner-Bund (DTB). Their findings indicated that the contribution to price formation in each market was similar, with no clear leader/follower relationship being discernible between floor-traded and electronic-traded futures.

Shyy and Lee (1995) had earlier reported that the Bund futures contracts traded on the DTB tended to lead the LIFFE; however, Shyy and Shen (1997) subsequently found no conclusive evidence in support of this lead-lag relationship between the open outcry system of the Singapore Stock Exchange (SGX) and the computer auction trading system for Japanese government bonds and Nikkei 225 index futures in the Osaka Stock Exchange (OSE).

Choy and Zhang (2010), Tao and Song (2010) and Pavabutr and Chaihetphon (2010) are the few ones that explore the price discovery of regular futures and mini futures on a same electronic trading platform. Choy and Zhang (2010) find that the regular futures play a dominant role in price discovery while the mini futures play a minor role. Tao and Song (2010) explore the price discovery of small (one-contract-size) trade in the mini Hang Seng Index futures. Their results show that small trade of mini Hang Seng Index futures plays an important role in price discovery. Pavabutr and Chaihetphon (2010) find that mini contracts contribute to over 30% of price discovery in gold futures trade even though they account for only 2% of trading value on the Multi Commodity Exchange of India (MCX). An important message from the above studies is that the price discovery of regular futures and mini futures on a same electronic trading platform is still unclear.

Furthermore, it has been found that market volatility may affect the trading performance of floor- and electronic-traded systems. Martens (1998) and Franke and Hess (2000) both reported that electronic trading systems make a greater contribution to the information share during quiet periods as compared to periods of relatively higher volatility. In contrast, Ates and Wang (2005a) found that E-mini index futures made a greater contribution to the information share during periods of high volatility as compared to periods of low volatility. It was assumed that this phenomenon was attributable to the ability of local traders to simultaneously trade in both the E-mini futures market and in the pits, since these traders are able to observe the trading information from the floor as well as the rapid execution of trades on the E-mini index futures trading platform.

In addition to differences in product design and trading mechanism, trading activity may well represent another important factor contributing to the leading role of E-mini futures in the overall process of price discovery. It is interesting to note that trading volume in most E-mini contracts is much larger than that for floor-traded contracts; and indeed, trading in E-mini futures in the US accounts for more than 75% of all trading volume in US index futures.

Relatively few papers have focused on the relationship between trading activities and price discovery. In one particular study, Zebedee (2001) examined the intraday relationships between S&P 500 ETFs, S&P 500 floor-traded futures and E-mini futures contracts; the results indicated that price discovery in the S&P 500 index derivative markets was time varying, with the tendency of E-mini futures to lead being attributable to the relative liquidity of the markets.

From an investigation into price discovery in the floor- and electronic-traded exchanges of the German stock market, Theissen (2002) found that the contributions made to price discovery by the trading system were positively related to market shares; however, the relationship between the size of the relative bid-ask spread and the contribution to price discovery was found to be weak. Ates and Wang (2005a) noted that the contribution to price discovery of E-minis was significantly affected by operational efficiency and relative market liquidity.

In the present study, we examine price discovery for regular (TX) and mini index futures (MTX) in Taiwan, with our study differing from the prior studies from several aspects. Firstly, both regular and mini index futures in Taiwan are traded on the same electronic trading system; thus, in contrast to the prior studies⁶ on regular futures (floor-traded) and E-mini futures (electronic-traded), we can control for the differences in trading mechanisms and examine the price discovery process between these two types of futures. Although there has been strong regulatory interest within the futures exchanges with regard to whether product design itself can affect the overall price discovery process, this appears to have been largely ignored within the extant literature.

Secondly, within the prior studies on the price discovery of regular and mini futures during periods of low and high volatility, researchers have had to overcome the problem of the differences between the trading mechanisms, which can clearly interfere with the impacts of contract size. We investigate whether market volatility affects the contributions to price discovery made by regular and mini index futures after controlling for differences in the trading mechanisms.

The futures markets are generally regarded as providing investors with ideal sets of instruments for their hedging activities during periods of high volatility, with mini futures providing smaller investors not only with an affordable instrument, but also with the option to trade in fractions of regular contracts during periods of high volatility. Furthermore, smaller trade sizes provide skilled traders with flexible trading and more precise hedging during periods of high volatility. Thus, the contribution to price discovery made by mini futures is likely to be greater during periods of high volatility, as compared to quiet periods.

Finally, the relative liquidity of mini index futures and regular futures in Taiwan differs from that in the US market. Within the Taiwan Futures Exchange (TAIFEX), the trading volume for mini index futures is lower than that for regular futures; thus, our results provide additional evidence on the relative impacts of liquidity and contract design on price discovery between regular and mini index futures. This also allows us to identify how the time-series variations in liquidity, over time, may affect the contribution to price discovery made by mini index futures based upon the gradual increase in trading volume in these instruments.

We adopt the Hasbrouck (1995) ‘information share’ and Gonzalo and Granger (1995) ‘permanent-transitory’ (PT) models to measure the contributions to the price discovery process made by the TX and the MTX, and find that the contribution made by the MTX has increased over time, ultimately overtaking the contribution made by the TX after 2004. To the best of our knowledge, our study is the first of its kind to provide direct empirical evidence on the dominant role of mini futures over regular futures based upon identical trading mechanisms. This provides clear evidence to show that the smaller contract size for mini index futures attracts more trading, thereby making a greater contribution to price discovery.

The contribution to the price discovery process made by the MTX is also found to be higher during volatile periods than during quiet periods. We further find that large MTX trades (trade size 4 or more contracts) contribute more to price discovery during volatile periods. The results indicate that MTX is also popular among skilled investors who want more flexibility by being able to purchase more contracts. During high volatility periods, due to higher risk in the market, skilled investors will prefer to trade MTX because they will have the option to scale back their trade size, if necessary. This advantage is not as important when the market risk (volatility) is low.

We carry out regression analyses on the daily information share of the MTX, with the results suggesting that the relative liquidity variables (such as market share, the relative bid-ask spread ratio and the relative number of trades) are major factors affecting the contribution of the MTX to the overall price discovery process.

We also find that the relative changes in liquidity between the TX and the MTX have contributed to the price discovery of the latter, and therefore conclude that relative liquidity and relative changes in liquidity are important factors determining the information shares of the related futures contracts. These findings are consistent with those of Theissen (2002) and Ates and Wang (2005a), who noted that liquidity was likely to be a determining factor in the overall process of price discovery, as opposed to trading mechanisms. Most importantly, despite the fact that liquidity for MTX contracts is lower (in terms of absolute value) our results further show that the higher relative liquidity and relative changes in liquidity are capable of making a strong contribution to price discovery.

The remainder of this paper is organized as follows. Section 2 provides an introduction to the institutional details of the TAIFEX and our data. This is followed in Section 3 by the presentation of the empirical models and methodology, with the empirical results subsequently being presented in Section 4. Finally, the conclusions drawn from this study are presented in Section 5.