Samarth Shah. and. B. Wade Brorsen*

Transcription

1 OM SOHAM Electronic vs. Open Outcry Trading in Agricultural Commodities Futures Markets Samarth Shah and B. Wade Brorsen* Farm Foundation / Commodity Futures Trading Commission / Economic Research Service Commodity Contract Design Workshop Washington, DC June 9-10, 2009 *Samarth Shah is a PhD student and B. Wade Brorsen is a regents professor and Jean & Patsy Neustadt Chair in the Department of Agricultural Economics at Oklahoma State University. Partial funding was provided by the Oklahoma Agricultural Experiment Station through Hatch project OKL02170.

2 Electronic vs. Open Outcry Trading in Agricultural Commodities Futures Markets Futures and options exchanges worldwide are increasingly shifting from conventional open outcry markets to electronic trading. Previous literature has compared liquidity costs in electronic and open-outcry markets in financial derivatives and non-agricultural commodities. However, work on the impact of the introduction of electronic trading in agricultural commodity markets is limited. This study estimates and compares liquidity costs in a traditional open-outcry market and a new electronic market. The study uses electronic and open outcry hard red winter wheat futures contracts traded on Kansas City Board of Trade in Liquidity costs were estimated using both Roll s measure and average absolute price changes. The estimates ranged from 1.53 to 2.57 cents per bushel for open outcry contracts and 0.53 to 1.42 cents per bushel for electronic contracts. Thus, liquidity costs are lower in the electronic market than in the open outcry market. Key words: Bid-ask spread, electronic trading, KCBT, liquidity costs Futures and options exchanges worldwide are increasingly shifting from conventional open outcry markets to electronic trading. Reasons for this shift include reduced transaction costs, reduced trading errors, and increased speed of execution. Electronic trading eliminates the costs of maintaining a physical site for floor trading. An open outcry trading system is comparable to a dealership market, in which traders can trade with competing market makers at their publicly announced bids and asks while an electronic trading mechanism is comparable to a continuous auction system with automatic order matching in which traders communicate only via computer screens (Martens, 1998). Electronic trading requires an open limit order book that enhances market transparency, dissemination of updated prices, and speeds order execution. Previous work studied the effects of the migration from open outcry to electronic trading on relative efficiency, execution costs, and informational efficiency and shows that there is an incentive for this migration. Examples of such work include, Martinez, Ning, and Tse (2008), Chung and Chiang (2006), Ates and Wang (2005), Bloomfield, O hara, and Saar (2005), Aitken, Frino, Hill, and 1

3 Jarnecic (2004), Tse and Zabotina (2001), Blennerhasset and Bowman (1998), Frino, McInish, and Toner (1998), Martens (1998) and Pirrong (1996). However, this work has largely dealt with financial futures markets rather than agricultural commodity futures markets. Kansas City Board of Trade (KCBT) introduced electronic trading on the CME Globex platform on Jan. 14, This shift sought to provide high-speed trade execution, direct market access, and central counterparty clearing, to achieve fairness, transparency and anonymity in trading. At KCBT both electronic and open-outcry markets co-exist which provides a favorable scenario to compare both markets. The present study estimates and compares liquidity costs in KCBT s traditional open-outcry market and the new electronic market. The study also identifies the impact of different factors such as daily volume and days to maturity on liquidity costs. The results of the study will help hedgers and speculators understand various aspects of both market environments and aid their business decisions. The results of the study will also help regulators and exchange management in increasing fairness and efficiency of the market. Theory To compare electronic and open-outcry trading, first it is required to understand the execution of the orders in both trading environments. At KCBT, open-outcry trading occurs on a trading floor where members (traders) trade continuously through open-outcry. Traders publicly announce bid and ask prices. If a trader finds a bid or ask attractive, the trader simply sells at the bid or buys at the ask price. The transaction price is then made public. Quotes are valid for only a very short period of time. A trader can also request a quote, and then may accept the best offer or refuse trading. When there are more traders with the same offer or ask, the buyer or seller can choose with whom to trade. As there is no official market maker on the KCBT floor, an official order 2

4 book does not exist. Names of traders are not published by the exchange. This information is immediately available only to the people on the floor. Electronic trading is a continuous auction system with automatic order matching in which traders communicate only via computer screens without revealing their names. If two orders can be matched, then the automatic auction mechanism chooses as matching orders those with the best prices. If there are a number of identical best bids or asks the trade is assigned to the order that has been in the system the longest time. Large orders exceeding the limit order of the quote will be split up over more quotes according to price and time of entering the system. Information on transaction prices and volumes is published instantaneously in the electronic system. A quote is valid until it is explicitly withdrawn from the system. One of the obvious differences between the two trading systems is the limit order book. In electronic trading, traders have access to an anonymous limit order book, while in open-outcry trading no official limit order book exists, but names and behavior of other traders can be observed on the floor. Another major difference is the execution of orders. In electronic trading a large order can be matched with several orders of the limit order book at different prices. On the other hand, in open-outcry an order is sometimes executed at one price only. Therefore, it is often argued that large investors get a better deal in open-outcry markets than in electronic markets. Though volume per trade is higher in openoutcry markets generally total volume and number of trades are higher in electronic markets due to the fact that it facilitates fast order matching and execution of trade. Due to this high activity in electronic environment and quicker dissemination of information, liquidity costs seem to be lower in electronic trading as compared to the open-outcry system. The present study also identifies factors affecting liquidity costs in both trading systems. Previous work on liquidity in futures markets found that liquidity costs and trading volume are 3

5 negatively correlated while liquidity costs and price variability are positively related (Brorsen 1989; Demsestz 1986; Hasbrouck and Schwartz 1988 and Thompson, et al. 1993). This is because in high volume markets traders trade with little price effect to their transactions. However, in thin markets, the transactions of individual traders may have significant price effects and may therefore result in higher liquidity costs. This is also one of the reasons why liquidity costs in electronic markets are expected to be lower than those of open-outcry markets. Another factor which is believed to have an effect on liquidity costs is days to maturity of the contract (Brorsen 1989 and Thompson, et al. 1993). The number of days to maturity of the contract may influence liquidity costs if there is some additional cost or risks to the traders in contracts having distant maturity dates. Data The intraday prices used in this study were the Time and Sales data for hard red winter wheat futures contracts traded at Kansas City Board of Trade (KCBT 2008). At KCBT wheat futures contracts are traded with five expiration months March, May, June, September and December. The database contains a record of each trade price of the five contracts traded by both open outcry and electronic methods in Regular trading hours for open outcry trading at KCBT are 9:30 a.m. to 1:15 p.m. Monday through Friday. The electronic market operates during regular trading hours as well as 6:00 p.m. to 6:00 a.m. Sunday through Friday, although there are few trades recorded outside of regular business hours. For the present study one trading day for electronic trading was considered from 6:00 p.m. on previous day through 1:15 p.m. of the day. The data set also contains daily volumes for each contract. 4

6 Procedures The bid-ask spread is an accepted measure of liquidity cost in security and futures markets. As bid-ask quotes for futures markets are not recorded by the exchange, two proxies of bid-ask spread were used to measure the liquidity costs in futures market: the Roll s measure and average absolute mean deviation. According to Roll (1983) if markets are informationally efficient, the covariance between price changes is negative and directly related to the bid-ask spread. Roll s measure was calculated using the following formula (1) 2 cov, where, is change in wheat price at time t. Roll s measure gives more efficient estimation of spread in more frequent observation which limits its use for estimating spreads in trading days with less frequent trades. The other accepted proxy for the bid-ask spread was proposed by Thompson and Waller (1988), who suggested the average absolute value of price changes as a direct measure of the average execution cost of trading in a contract. The average absolute price changes were calculated as (2) Average absolute price change T F. T The liquidity costs for the five contracts were estimated in both electronic and open outcry futures markets using Roll s measure and average absolute mean deviations. Each measure was estimated on a daily basis and then averaged for the life of contract using daily number of trades as weight. To test hypotheses about the factors influencing liquidity costs, the following regression equation was estimated by ordinary least squares: 5

7 (3) where, RM is Roll s measure of contract i for day t, is volume per trade of contract i for day t, TV is total volume of contract for day, T is time to maturity of contract i in days, i is March, May, June, September and December, e is assumed to be distributed normally with mean and variance. Misspecification tests for the above model were done to check normality of error term, presence of heteroskedasticity and autocorrelation and linear functional form of the model. Results: There were five wheat futures contracts considered for the study. Descriptive statistics are presented in Table 1. Average trades per day for electronic markets were immensely higher than open-outcry markets. However, average volume per trade for electronic markets were considerably lower than the those of open-outcry markets which can be explained by the fact that in electronic trading a large order can be matched with several orders of the limit order book at different prices while in open-outcry an order is executed at one price only so large traders prefer to trade in open-outcry system. The monthly volume for electronic and open outcry markets is shown in figure 1. The daily volume of the July 2008 contract for electronic and open-outcry contracts is presented in Figure 2. The figures show that daily volumes of electronic contracts were higher than those of open outcry contracts throughout the life of the contracts except for few occasions. Total volume traded in wheat electronic futures markets during 2008 at KCBT were 1,882,302 contracts compared to 1,033,741 contracts in open outcry markets (KCBT 2008). The results indicate that the electronic market had higher market share and it dominated open outcry markets. 6

8 The liquidity costs for the five contracts in both electronic and open outcry futures markets are presented in Table 2. The average Roll s measure for electronic markets ranged from 0.53 cents per bushels to 1.43 cent per bushel while for open outcry markets it ranged from 1.23 cents per bushel to 2.71 cents per bushel. This indicates that the electronic markets were cheaper in terms of liquidity costs. Shah and Brorsen (2009) estimated the same measures for June 2007 open outcry wheat futures contract. They reported Roll s measure of 0.45 cents per bushel and average absolute mean deviation of 0.49 cents per bushel. Thompson, Eales, and Seibold (1993) also estimated the same measures for selected 1985 KCBT wheat contracts. Their estimates of average absolute deviations are cents per bushel for highly traded contracts, but were about double these values for lightly traded contracts such as the March contract during March or the September contract in February. The estimates of Roll s measure and average absolute mean deviation for June 2008 open outcry contract of present study were 1.23 and 1.39 cents per bushels, respectively. One of the reasons behind higher liquidity costs in 2008 as compared to 2007 for the same contract might be the lower trading volumes in The total trading volumes for wheat futures markets in 2007 at KCBT were 4,318,007 contracts while they were only 3,778,266 contracts in 2008 (KCBT 2008). The average absolute deviation estimates were also considerably lower in electronic markets compared to open outcry markets. With the higher prices and higher price volatility in 2008, the risk associated with scalping clearly increased. Figure 3 shows the number of trades by time of day. The open outcry market opens at 9:30 and closes at 1:15. The inverted U shape for number of trades is at least partly due to the length of time periods not being equal. What is striking about figure 3 is how few electronic trades occur outside the hours of operation for open outcry. 7

9 Figure 4 shows that liquidity costs are larger in the open outcry market at both the open and the close. Ekman (1992) argues that information traders are more likely to trade at the open and close and thus that is when price movements occur. Such changes in equilibrium prices would increase the estimate of liquidity costs. Also, the Goldman-Sachs index fund traded substantial long positions during When the fund rolled its positions into the next contract month (known as the Goldman-Sachs roll), it could have also caused greater price movement, especially at the close. The electronic market shows greater liquidity costs outside the regular trading hours, which could explain the small volume Klumpp, Brorsen, and Anderson (2007) found that producers had a slight tendency toward round number pricing. Figure 5 shows that round number pricing is much more prevalent in the open outcry market than in the electronic market. Prices were highly volatile during this period and began moving in whole cents in the open outcry market. Ordinary least squares was used to determine the relationship between liquidity cost, time to expiration of the contract, average volume per trade and total daily volume of the contract. The results of the regression for open outcry and electronic markets are presented in Table 3 and Table 4. The results show negative effect of daily volume on the liquidity costs for all the contracts in both electronic and open outcry markets except for May 2008 open outcry contract. The average volume per trade showed both positive and negative significant impact on liquidity costs. Since liquidity costs decrease with volume, it does create the question, would liquidity costs be lower if only the electronic market were traded? 8

10 Summary and Conclusion: The objectives of this study were to determine and compare liquidity costs in the traditional open outcry futures markets and the new electronic futures markets and determine the factors affecting these liquidity costs. To meet the objectives, intraday prices of five hard red winter wheat futures contracts traded on Kansas City Board of Trade exchange during 2008 were used where both electronic and open outcry markets coexists. Roll s measure and average absolute price deviations were used to estimate liquidity costs in futures market. The average Roll s measure for electronic markets ranged from 0.53 cents per bushels to 1.43 cent per bushel while for open outcry markets it ranged from 1.23 cents per bushel to 2.71 cents per bushel. Both the measures of liquidity costs in electronic market were considerably lower than estimated liquidity costs in open outcry futures market. Higher trading volume in electronic markets is one of the explanations of lower liquidity costs in this market. The study concludes that it costs less to trade in electronic futures markets as compared to open outcry futures markets. The implication of this result to hedgers is that transacting in open outcry market, except possibly for large transactions, is more expensive than transacting in electronic futures market. Hedgers should consider this fact when choosing between these two markets. 9

11 Table 1. Descriptive Statistics of Wheat Futures Contracts Traded at KCBT in 2008 Open Outcry Electronic Contract N Average Trades per Day Average Volume per Trade N Average Trades per Day Average Volume per Trade March (92.21) (66.36) May (78.37) (220.93) June (13.13) (2.37) Sep (36.94) (5.79) Dec (21.67) (3.04) Note: Figures in parenthesis are standard deviation. Average volume is in number of 5000 bushel contracts. Table 2. Measures of Liquidity Costs (cents/bushel) in Wheat Futures Contracts Traded at KCBT in 2008 Open Outcry Electronic Contract Roll s Average Absolute Price Change Roll s Average Absolute Price Change March 1.41 ( ) 1.31 ( ) 0.46 ( ) 0.38 ( ) May 2.17 ( ) 2.14 (9.0270) 0.88 ( ) 0.70 (7.3169) June 1.18 ( ) 1.23 (9.9804) 0.63 ( ) 0.41 (9.2452) Sep 1.38 ( ) 1.35 ( ) 0.28 (7.0920) 0.26 (4.4480) Dec 1.56 ( ) 1.44 ( ) 0.29 ( ) 0.30 (9.0059) Note: Figures in parenthesis are standard deviation. 10

12 Table 3. Results of the Regressions with Roll s Measure of Open-Outcry Contracts as Dependent Variable contract N Intercept Days to Maturity Volume per Trade Total Volume March (0.0003) (0.0244) (0.2024) (0.3915) May (0.0040) (0.4111) (0.1474) (0.0098) June (<0.0001) (0.0021) (0.1128) (0.3479) Sep (0.7078) (0.0174) Dec (0.0997) (0.0838) Note: values in the parenthesis are p values. (0.1482) (0.4897) (0.2224) (0.0157) Table 4. Results of the Regressions with Roll s Measure of Electronic Contracts as Dependent Variable contract N Intercept Days to Maturity Volume per Trade Total Volume March (<0.0001) (0.0026) (0.0264) (0.0002) May (0.0169) (0.3875) (0.8645) (0.1217) June (0.1340) (0.0377) (0.0007) (0.0030) Sep (<0.0001) (<0.0001) Dec (0.0086) (0.8265) Note: values in the parenthesis are p values. (0.3212) (0.0697) (<0.0001) (0.0005)

13 Open outcry Electronic Number of Contracts Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 1. Monthly Volume of KCBT Wheat Futures Contract in Open outcry Electronic Volume Jan Feb Mar Apr Jun Month Figure 2. Daily volume of Electronic and Open-outcry July 2008 Wheat Futures Contracts at KCBT Exchange 12

14 Number of Trades Electronic Open outcry Time of the Day Figure 3. Number of Trades at Different Time of the Day at KCBT in Liquidity Cost (cents/bu.) Open outcry Electronic Time of the Day Figure 4. Liquidity Cost at Different Time of the Day at KCBT in

15 Electronic Open outcry 50 % Ending Value Figure 5. Ending Values of Trade Price in Electronic and Open Out cry Markets at KCBT in

16 References: Aitken, M. J., A. Frino,A. M. Hill, and E. Jarnecic The Impact of Electronic Trading on Bid- Ask Spreads: Evidence from Futures Markets in Hong Kong, London, and Sydney. Journal of Futures Markets 24: Ates, A., and G. H. K.Wang Information Transmission in Electronic vs. Open-Outcry Trading Systems: An Analysis of U.S. Equity Index Futures Markets. Journal of Futures Markets 14: Blennerhasset, M. and R. G. Bowman A Change in Market Microstructure: The Switch to Electronic Trading on the New Zealand Stock Exchange. Journal of International Financial Markets 8: Bloomfield, R.,M. O hara, and G.Saar Make or Take Decision in an Electronic Market: Evidence on the Evolution of Liquidity. Journal of Financial Economics 75: Brorsen, B. W "Liquidity Costs and Scalping Returns in the Corn Futures Market." Journal of Futures Markets 9: Chakrabarty, Bidisha, and Pankaj Jain Understanding the Microstructure of Indian Markets. NSE Final Paper 128. Chung, H., and S. Chiang Price Clustering in E-mini and Floor-Traded Index Futures. Journal of Futures Markets 24: Demsestz, H "The Cost of Transacting." Quarterly Journal of Economics 82: Ekman, P.D Intraday Patterns in the S&P 500 Index Futures Market. Journal of Futures Markets 12: Frino, A., T. McInish, and M. Toner The Liquidity of Automated Exchanges: New Evidence from German Bond Futures. Journal of International Financial Markets, Institutions and Money 54: Kansas City Board of Trade Available at Klumpp, J.M., B.W. Brorsen, and K.B. Anderson Producers Preferences for Round Number Prices. Agricultural Finance Review. 67: Kumar, S Determinants of Liquidity in Open Electronic Limit Order Book Market. Working Paper, Indira Gandhi Institute of Research and Development. Martens, M Price Discovery in High and Low Volatility Periods: Open-Outcry versus Electronic Trading. Journal of International Financial Markets, Institutions and Money 8:

17 Martinez, V., B. Karan, N. Zi, and T. Yiuman "Competition for Order Flow and Market Quality in the Gold and Silver Futures Markets." Working paper 0036, College of Business, University of Texas at San Antonio. Pirrong, C Market Liquidity and Depth on Computerized and Open-Outcry Trading System: A Comparison of DTB and LIFFE Bund Contracts. Journal of Futures Markets 16: Roll, R "A Simple Implicit Measure of the Effective Bid-Ask Spread in an Efficient Market." The Journal of Finance 39(4): Shah, S. and B. W. Brorsen Liquidity Costs in Futures Options Markets. Paper presented at NCCC-134 meeting on Applied Commodity Price Analysis, Forecasting and Market Risk Management, St. Louis, Missouri, April. Thompson, S., J. S. Eales, and D. Seibold "Comparison of Liquidity Costs between the Kansas City and Chicago Wheat Futures Contracts." Journal of Agricultural and Resource Economics 18: Tse, Y. and T. Zabotina Transaction Costs and Market Quality: Open-Outcry versus Electronic Trading. Journal of Futures Markets 21: