مدت زمان، حجم معاملات و تاثیر قیمت بر معاملات در بازار معاملات آتی در حال ظهور

This paper examines the price impact of trading intensity on the MexDer TIIE28 interest rate futures contract, one of the world's most actively traded contracts. A novel volume-augmented duration model of price discovery decomposes trading intensity into liquidity and information components. Duration between transactions exerts a positive influence on price changes, while increases in order flow and trade volume exert positive and negative influences, respectively. The liquidity component dominates the information measure, suggesting that liquidity considerations dictate trade timing. These findings are rationalized with reference to MexDer's organizational structure, specifically the affirmative obligations placed upon marketmakers to trade a minimum volume.

Theoretical microstructure models of financial markets are generally underpinned by information- or liquidity-based explanations of trading patterns. Understanding the relationship between asset prices, trading volume and duration (the time between consecutive trades) in the context of a particular trading architecture, is critical to analyze both the information and liquidity aspects of trading activity and the resulting value of a financial asset. This study examines the price impact of the various components of trading intensity utilizing a novel volume-augmented duration model of price discovery. A globally significant futures contract, the 28-day equilibrium interbank interest rate futures contract (Tasa de Interes Interbancaria de Equilibrio or TIIE28) trading on the Mexican Derivatives Exchange (MexDer) is selected for analysis. At the end of the data sample period (October 2006), the TIIE28 ranked as the world's 3rd most actively traded futures contract behind the CME's Eurodollar and Eurex's euro-bund contracts ( Holz, 2007). 1 The paper's focus is upon the relationship between key trading process variables, specifically; order flow, duration, and trading volume, and the subsequent changes in the TIIE28 futures price. Prior research generally analyzes the asset price impact of one or two specific trade process variables in isolation, rather than considering their combined effects. For example, Madhavan et al. (1997) concentrate solely on the price impact of order flow assuming trade size to be constant. They maintain that this allows for a more parsimonious model, and argue that if markets are anonymous, informed traders will split up their large trades. Their model implies that trade direction is usually more informative than trade size. In contrast, in DeJong et al. (1996) trade size constitutes one of the key determinants of the bid-ask spread, while several other theoretical market microstructure models (Easley and O'Hara, 1987, Engle, 2000 and Manganelli, 2005) maintain that patterns in trading volume are the manifestation of the presence or absence of market relevant information. The key contribution of this paper lies in its specification of a model which explicitly includes trading volume as a component of the trade direction indicator. The prior literature rationalizes this choice along several dimensions. Easley and O'Hara (1987) argue that considerations of both liquidity and optimal trading strategy dictate that informed traders will sometimes engage in high volume intensive trading, while at others will segment their order flow to generate a larger number of informationally based trades over a given period of time. Thus, trade volume conveys information relating to market conditions and may directly affect prices. Engle (2000) and Manganelli (2005) model volume as a stochastic process when estimating the trading impact of duration. Moreover, DeJong et al. (1996) provide evidence that transaction size (and by association trading volume) is positively related to measured pricing effects. In addition to the above theoretically motivated reasons for including a trading volume variable in the estimated model, the organizational structure governing trading in MexDer's TIIE28 contract, in particular the affirmative obligations placed upon marketmakers as established in the markets “Liquidity Terms and Conditions”, dictates that volume should be incorporated when modeling trading intensity and liquidity. To provide the requisite institutional context, MexDer commenced operations on 15th December 1998, trading Mexican Peso–US Dollar futures contracts.2 The market operated via open outcry until the introduction of electronic trading in May 2000, transforming MexDer into a fully automated electronic limit order book market. However, low trading volumes continued, until a decisive development in a series of procedures designed to enhance market liquidity occurred in May 2001, namely the introduction of exchanged-designated marketmakers. 3 Marketmakers are committed among other responsibilities, to (i) offer buy and sell quotes with a maximum spread to create liquid markets in a minimum number of TIIE28 contract maturity dates, and (ii) trade a minimum number of TIIE28 contracts each month (the minimum volume rule). They are also granted with certain exclusive trading rights. Specifically, rather than having to submit orders through the electronic order book, marketmakers are also allowed exclusive access to a voice trading-by-phone service by which they can request MexDer's Trading Floor personnel to enter, modify or cancel orders. Moreover, the MexDer “Manual of Policies and Procedures” stipulates that only marketmakers are able to initiate certain important categories of trades such as “depth”, “cama” and “ronda” trades. 4 Following the introduction of marketmakers, MexDer's organizational architecture is perhaps best characterized as a hybrid electronic limit order book market. The liquidity impact was immediate. In the first month that marketmakers were introduced, average daily financial trading volume in the TIIE28 contract rose very significantly to USD 13,548 million, an increase of approximately 800% from the previous month's (April 2001) recorded average of USD 1733 million (Fernandez, 2003). The number of marketmakers has subsequently increased from the original eleven to fourteen institutions by the end of our sample period, by which time average daily notional financial volumes in the contract exceeded USD 9.5 billion.5 Many capital markets employ a hybrid market architecture characterized by a variety of affirmative obligations attached to designated marketmakers to enhance liquidity. This highlights the importance of including variables reflecting specific exchange trading arrangements, such as the MexDer's minimum volume rule, in structural price impact models.6 The large volume individual trades in the TIIE28 contract predominantly reflect the trading activity of designated marketmakers, who will almost always be on one side, and in many cases both sides of the transaction. As such, incorporating trading volume in a structural model not only potentially provides an important signal of the presence of informed trading (as in prior studies), but may also be crucial in capturing the price impact of the liquidity enhancement activity of exchange marketmaking rules, which is still not well understood (Panayides, 2007). A further important contribution of this analysis is to provide insight into the microstructure of a key contract on a significant emerging financial futures market, whereas prior research in this area tends to focus upon highly liquid equity markets in developed economies, or foreign exchange markets in the major currencies.7,8 This aspect has several dimensions. In comparison with equity or foreign exchange markets, futures markets offer additional potential insights on the price impacts of trading intensity, and how public and private information and market processes contribute to the determination of market equilibrium prices. Many equity markets are characterized by institutional rigidities such as short-sale constraints, which by their very nature mediate the response of asset prices to informational shocks. In a futures market traders can take long or short positions with equal facility. Moreover, trading volume in a futures market is likely to be an important factor impacting price variability as it is well known that the time to contract maturity influences futures trading patterns with traders seeking to unwind or rollover their positions as maturity approaches. The fact futures trading is exchange-based also implies that trading intensity in the market is more easily measured as compared to the multi-dealer, decentralized foreign exchange market. Finally, knowledge of the price impact of trading is central to execution cost management, particularly in emerging markets. Indeed, reducing execution costs, along with liquidity enhancement, are among the primary motivations for MexDer introducing marketmakers in 2001.9 Dealers and brokerage trading desks require knowledge of how trading process variables potentially impact returns in order to achieve efficiency in order execution, provide competitive services to their clients and attract institutional investors. To pre-empt the findings, the empirical results suggest that both duration and trading volume play a significant role in the pricing behavior of MexDer's TIIE28 futures contracts. First, the analysis of duration between transactions indicates that the transient and persistent components of duration exert positive and negative influences on price changes, respectively. Overall, the net effect corroborates the evidence provided by Ben Sita and Westerholm (2006). In contrast to the results of Dufour and Engle (2000), longer duration is found to have a greater impact on price changes than short duration. Second, the analysis of trading volume indicates that price changes tend to decrease when there is an increase in trade volume, while increases in order flow exert a positive influence on price changes, albeit to a lesser extent than the influence of duration. These findings are discussed in the context of the hybrid market structure and the affirmative obligations placed upon marketmakers at MexDer to enhance liquidity. The structure of the remainder of the paper is easily summarized. The next section presents an overview of the existing literature. Section three discusses the econometric model and the methodology employed. Section four provides further details of the TIIE28 futures contract, describes the data in detail and presents the results of the empirical analysis and sensitivity tests, while section five briefly concludes.

This paper examines the pricing impact of duration and trading volume in the context of a structural microstructure model of price discovery using data from the emerging futures market of Mexico, MexDer. In their empirical study, Dufour and Engle (2000) document that long duration has a lower impact on price than short duration. Extending Dufour and Engle (2000) by decomposing the duration effect into liquidity and information effects, Ben Sita and Westerholm (2006) find evidence that both components exert a positive influence on price changes. This analysis examines six years of tick by tick data for one of the world's most liquid futures contracts, the 28-day equilibrium interbank interest rate futures contract, TIIE28, on MexDer. In 2006 this contract accounted for 93% of total MexDer trading volume and 98.9% of the exchange's open interest, and was the 3rd most actively traded futures contracts world-wide. The findings involving transaction duration show that the transient and persistent components of duration exert positive and negative influences on price changes, respectively. Overall, the net effect is similar to the evidence provided by Ben Sita and Westerholm (2006). In contrast to Dufour and Engle (2000), long duration is found to have a greater impact on price changes than short duration. The analysis of trading volume suggests that price changes tend to decrease when there is an increase in trade volume. Additionally, increases in order flow exert a positive influence on price changes, albeit to a lesser extent than the influence of duration. The presence of volume persistence and trading volume clustering documented in the results is consistent with Admati and Pfleiderer (1988) who maintain that, in equilibrium, discretionary liquidity traders choose to trade together generating a clustering of trading volume. The preferred interpretation of these features is institutionally based, and centers around the affirmative obligations of marketmakers on MexDer and the exclusive rights they hold to initiate certain important categories of discretionary trades. This empirical evidence is consistent with these exclusive trading rights enabling marketmakers to efficiently manage both the price risk associated with holding an inventory of positions and any adverse selection costs arising from trading with informed traders. Note that this interpretation of the price impact measures accords with existing empirical studies which provide little support for the view that marketmakers rebalance their inventory through revisions to quoted prices, but rather mitigate inventory holding costs through discretionary trading when their marketmaking obligations are not binding (Panayides, 2007). The importance of the supply of liquidity components suggests that marketmakers trading behavior dominates the price impact of trade process variables on MexDer. Further, the results of this paper indirectly suggest that market making obligations or constraints such as position limits, or short sales constraints, dictated by the specific institutional context in which trading occurs, may be of relevance in determining the pricing impact of trade process variables in empirical market microstructure studies.