4. Empirical Results
4.2 Bid-Ask Spread Components
In this section we employ the spread decomposition model of Lin, Sanger, and Booth (1995) to separate the bid-ask spread into the order processing component and the adverse selection cost component for the TF, the TE, and the TF futures in the call auction market and the continuous auction market, respectively. Before we actually estimate the model, we would like to study the behavior of the effective spread, zt, which is important to the model estimation.
Table 6 summaries the distribution of effective spread. A transaction with zt>0 means that it is triggered by a buying order while a transaction with zt<0 indicates that it is triggered by a selling order. Panel A of Table 6 shows that in the original call auction market, 46 percent of the transactions for the TX futures are triggered by buying orders; 50 percent are trigged by selling orders; and only 4 percent are traded at the quote midpoint. After the market transforms to a continuous auction market, 45 percent of the transactions are triggered by buying orders; 49 percent are triggered by selling orders; and the transactions occurring at the quote midpoint slightly increase to 6 percent.
As shown in Panel B, the pattern of the transactions for the TE futures is a bit different from that for the TX futures. After the transformation of the trading mechanism, the transactions triggered by selling orders increase by 1 percent from 43
percent to 44 percent while the transactions occurring at the quote midpoint decrease by 1 percent to 5 percent. The figures in Panel C of Table 6 suggest that the transactions for the TF futures, which are triggered by buying and selling orders, both decrease by 1 percent. In contrast, the transactions occurring at the quote midpoint increase by 2 percent.
It is obvious that in both types of trading mechanism, the transactions are triggered more often by selling orders than buying orders. Since we cannot clearly identify whether those transactions with zt=0 are triggered by selling orders or buying orders, the estimation of the spread decomposition model would be biased if the percentage of this kind of transactions is very high. Fortunately, less than 6 percent of the transactions is matched at the quote midpoint. Hence, it is unlikely that the transactions with zt=0 have significant impact on the estimation of the spread components.
We present the estimates of the spread components for the TX, TE, and TF futures in Table 7. We specifically report the estimates for those three parameters in the model, the adverse information cost (λ), the order processing cost (γ), and the persistence of order arrival (θ), the t-statistics for the estimates and the R-square values for the estimated models.
From Panel A we can see that the adverse selection cost for the TX futures
decreases by almost 20% from 0.53 in the call auction market to 0.33 in the continuous auction market while the order processing cost increases by 26% from 0.26 to 0.52. The estimates of the order persistence parameter decrease slightly by 8% from 0.23 to 0.15. In other words, the transactions are less persistent in the continuous auction market. Furthermore, the adverse selection cost is the largest component of the effective spread for the TX futures in the call auction market whereas the order processing cost becomes the major component in the continuous auction market.
Panel B indicates that after switching to the continuous auction market, the adverse selection component for the TE futures decreases from 0.44 to 0.22 while the order processing cost increases slightly from 0.26 to 0.32. However, unlike the TX futures, the estimate of the order persistence for the TE futures increases from 0.3 to 0.46 and becomes the largest component of the effective spread. Therefore, the continuous auction market is also effective in resolving the problem of information asymmetry for the TE futures. As shown in Panel C, the installation of the continuous auction trading system results in lower adverse selection cost and order processing cost for the TF futures. Nevertheless, the transactions of the TF futures become more persistent because the order persistence estimate increases sharply from 0.3 to 0.59. It is apparent that the continuous auction market also offers lower
adverse selection cost to the traders involving in the transactions of the TF futures contract.
According to the results in Table 7, we know that the change of trading mechanism may have different effects on the component structure of the effective spread for different futures contracts in the TAIFEX. While the order processing cost for the TF futures declines, those for the TX and TE futures actually increase.
This increase of the order processing cost is consistent with the finding of Lin, Sanger, and Booth (1995) and Huang and Stoll (1997) that the order processing cost declines as the trade size increases. It is because the trade size of the TX and TE futures increases indeed in the continuous auction market. The most consistent result in this section is that the continuous auction trading mechanism can effectively reduce the adverse selection cost for all three types of futures contracts. However, this result is inconsistent with the predictions of the models of Madhavan (1992), Pagano and Röell (1992, 1996), and Schnitzlein (1996). It would be intriguing to investigate this result further in our future research.
5. Conclusions
On July 29th, 2002, the TAIFEX decided to switch its trading mechanism from the original call auction market to the continuous auction market. This decision provides us with a unique opportunity to study the effect of trading mechanism of the
market quality of futures market. In this paper, the market quality is evaluated based on twelve commonly used proxies including the quoted spread, the effective spread, the percentage quoted spread, the percentage effective spread, the dollar-volume-weighted average percentage quoted spread, the dollar-volume-weighted average percentage effective spread, the trading volume, the number of trades, the trade size, the liquidity ratio, the price volatility, and the volatility of the rate of return. Besides, the change in the adverse selection cost due to the switch of trading mechanism is also investigated based on the spread decomposition model of Lin, Sanger, and Booth (1995).
Our empirical results confirm the finding of Elyasiani, Hauser, and Lauterbach (2000) that market liquidity indeed has many facts and cannot be represented by the bid-ask spread alone. In particular, the empirical results show that the quote spread, the effective spread, the percentage effective spread, and the dollar-weighted percentage effective spread are lower in the continuous auction market. However, the other liquidity measures do not illustrate the same pattern. Overall, although not very consistent, these results based on the liquidity measures suggest that the market quality of the continuous auction market is better than the call auction market.
Besides, the continuous auction market is also more effective than its counterpart in resolving the information asymmetry problem in the TAIFEX.
Reference
Amihud, Y. and H. Mendelson, 1986, Asset Pricing and the Bid-Ask Spread, Journal of Financial Economics 17, 223-249.
Amihud, Y., H. Mendelson, and B. Lauterbach, 1997, Market Microstructure and Securities Values: Evidence from the Tel Aviv Stock Exchange, Journal of Financial Economics 45, 365-390.
Blennerhassett, M., and R.G. Bowman, 1998, A Change in Market Microstructure:
The Switch to Electronic Screen Trading on the New Zealand Stock Exchange, Journal of International Financial Markets, Institutions and Money 8, 261-276.
Brockman, P., and D.Y. Chung, 1999, Bid-Ask Spread Components in an Order-Driven Environment, Journal of Financial Research 22, 227-246.
Christie, W.G., and R.D. Huang, 1994, Market Structures and Liquidity: Transactions Data Study of Exchange Listings, Journal of Financial Intermediation 3, 300-326.
Clyde, P., P. Schultz, and M. Zaman, 1997, Trading Costs and Exchange Delisting:
The Case of Firms that Voluntarily Move from the American Stock to the Nasdaq, Journal of Finance 52, 2103-2112.
Degryse, H., 1999, The Total Cost of Trading Belgian Shares: Brussel versus London, Journal of Banking & Finance 23, 1331-1355.
De Jong, F., T. Nijman, and A. Röell, 1995, A Comparison of the Cost of Trading French Shares on the Paris Bourse and on SEAQ International, European Economic Review 39, 1277-1301.
Elyasiani, E., S. Hauser, and B. Lauterbach, 2000, Market Response to Liquidity Improvements: Evidence from Exchange Listings, The Financial Review 41, 1-14.
Frino, A., T.H. McInish, and M. Toner, 1998, The Liquidity of Automated Exchanges:
New Evidence from German Bund Futures, Journal of International Financial Markets, Institutions and Money 8, 225-241.
George, T.J., G. Kaul, and M. Nimalendran, 1991, Estimation of the Bid-Ask Spread and Its Components: A New Approach, Review of Financial Studies 4, 623-656.
Glosten, L.R., L.E. Harris, 1988, Estimating the Components of the Bid/Ask Spread, Journal of Financial Economics 21, 123-142.
Hasbrouck, J., 1993, Assessing the Quality of a Security Market: A New Approach to Transaction-cost Measurement, The Review of Financial Studies 6, 191-212.
Huang, R.D., and H.R. Stoll, 1996, Dealer versus Auction Markets: A Paired Comparison of Execution Costs on NASDAQ, Journal of Financial Economics 41, 313-357.
Huang, R.D., and H.R. Stoll, 1997, The Components of the Bid-Ask Spread: A General Approach, Review of Financial Studies 10, 995-1034.
Huang, R.D. and H.R. Stoll, 2001, Tick Size, Bid-Ask Spreads, and Market Structure, Journal of Financial and Quantitative Analysis 36, 503-522.
Kehr, C., J.P. Krahnen, and Erik Theissen, 2001, The Anatomy of a Call Market, Journal of Financial Intermediation 10, 249-270.
Lauterbach, B., 2001, A Note on Trading Mechanism and Securities’ Value: The Analysis of Rejects from Continuous Trade, Journal of Banking and Finance 25, 419-430.
Lin, J., G.C. Sanger, and G.G. Booth, 1995, Trade Size and Components of the Bid-Ask Spread, Review of Financial Studies 8, 1153-1183.
Madhavan, A., 1992, Trading Mechanisms in Securities Markets, Journal of Finance 47, 607-641.
Madhavan, A., M. Richardson, and M. Roomans, 1997, Why Do Security Prices Change? A Transaction-Level Analysis of NYSE Stocks, Review of Financial Studies 10, 1035-1064.
Majois, C., and R.D. Winne, 2003, A Comparison of Alternative Spread Decomposition Models on Euronext Brussels, working paper.
Mendelson, H., 1982, Market Behavior in a Clearing House, Econometrica 50, 1505-1524.
Mendelson, H., 1986, Random Competitive Exchange: Price Distributions and Gain from Trade, Journal of Economic Theory 37, 254-280.
Menyah, K. and K. Paudyal, 2000, The Components of Bid-Ask Spreads on the London Stock Exchange, Journal of Banking and Finance 24, 1767-1785.
O’Hara,M.,1995,
Market Microstructure Theory, Blackwell Publishers Ltd, Oxford,
UK.Pagano, M. and A. Röell, 1992, Auction and Dealership Markets: What is the difference?, European Economic Review 36, 613-623.
Pagano, M., and A. Roell, 1996, Transparency and Liquidity: A Comparison of Auction and Dealer Markets with Informed Trading, Journal of Finance 51, 579-611.
Pirrong, C., 1996, Market Liquidity and Depth on Computerized and Open Outcry Trading Systems: A Comparison of DTB and LIFFE Bund Contracts, Journal of Futures Markets 16, 519-543.
Roll, R., 1984, A Simple Implicit Measure of the Effective Bid-Ask Spread in an Efficient Market, Journal of Finance 39, 1127-1139.
Schnitzlein, C.R., 1996, Call and Continuous Trading Mechanisms Under Asymmetric Information: An Experimental Investigation 51, 1996, 613-636.
Stoll, H.R., 1989, Inferring the Components of the Bid-Ask Spread: Theory and Empirical Test, Journal of Finance 44, 115-134.
Tse, Y. and T.V. Zabotina, 2001, Transaction Costs and Market Quality: Open Outcry Versus Electronic Trading, Journal of Futures Markets 21, 713-735.
Theissen, E., 2002, Floor versus Screen Trading: Evidence from the German Stock Market, Journal of Institutional and Theoretical Economics 158, 32-54.
Van Ness, B.F., R.A. Van Ness, and R.S. Warr, 2001, How Well Do Adverse Selection Components Measure Adverse Selection?, Financial Management Autumn, 77-98.
Table 1. The volumes of products traded on the TAIFEX
Contract Annual Volume 2003 Annual Volume 2002 % Increase Futures
TAIEX Futures (TX) 6,514,691 4,132,040 57.66
Electronic Sector Index Futures (TE) 990,752 834,920 18.66 Finance/Insurance Sector Index
Futures (TF)
1,126,895 366,790 207.23
Mini-TAIEX Futures (MTX) 1,316,712 1,044,058 26
Taiwan 50 Futures (T5F) 4,068 N/A (Launch date:
June 30th, 2003)
N/A
Total (Futures) 9,953,118 6,377,808 N/A
Options
TAIEX Options (TXO) 21,720,083 1,566,446 1286.58
Equity Options (STO) 201,733 N/A (Launch date:
January 20th, 2003)
N/A
Total (Options) 21,921,816 1,566,446 N/A
Total (Futures & Options) 31,874,934 7,944,254 301.23
Source: the article “2003:Record VolumefortheFifth ConsecutiveYearforTAIFEX”
on the official website of the Taiwan Future Exchange.
Table 2. The new and old trading mechanism of TAIFEX
Time New mechanism Old mechanism
Opening 8:45 a.m.
Electronic call auction Electronic call auction
Operation between opening and closing
8:45 a.m.-1:40 p.m.
Electronic continuous auction Electronic call auction
(Orders matching per 10 seconds)
Closing
1:40 p.m.- 1:45p.m.
Electronic call auction Electronic call auction
Table 3. Contracts specification of the TX Futures, the TE Future, and the TF
Ticker Symbol TX TE TF
Delivery Months Spot month, the next calendar month, and the next three quarter months Last Trading Day The third Wednesday of the delivery month of each contract
Trading Hours
08:45AM-1: 45 PM Taiwan time Monday through Friday of the regular business days of the Taiwan Stock Exchange
Contract Size NT$200 * Index NT$4,000 * Index NT$1,000 * Index Minimum Price Daily Price Limit +/- 7% of previous day's settlement price
Margin
The initial and maintenance margin levels as well as the collecting measures prescribed by the FCM to its customers shall not be less than those required by the TAIFEX. The margin levels will be adjusted and announced by the TAIFEX in accordance with "the Criteria and Collecting Methods regarding the Clearing Margins".
Daily Settlement Price The last trading price of the closing session or otherwise determined by the TAIFEX according to the trading rules
Final Settlement Day The first business day following the last trading day. All of the open interests after the final settlement day shall be settled on the final settlement price.
Final Settlement Price
The final settlement price for each contract is computed from the first fifteen-minute volume-weighted average of each component stock's prices in the index on the final settlement day. For those component stocks that are not traded during the beginning fifteen- minute interval on the final settlement day, their last closing prices would be applied instead.
Settlement Cash settlement
3 Source from introduction of products on the official website of Taiwan Future Exchange
Table 4. A comparison of liquidity before and after the enforcement of electronic
Quote spread 2.4137 1.96976 (1, 249020) 5535.57 0 ***
Effective spread 3.26473 2.27889 (1, 249020) 10163.4 0 ***
Percentage quote spread 0.04053 0.04407 (1, 249020) 912.99 0 ***
Percentage effective spread 0.05496 0.05105 (1, 249020) 379.977 0 ***
Dollar-weighted percentage quote spread 0.0405 0.04485 (1, 122) 8.832 0.004 ***
Dollar-weighted percentage effective spread 0.06966 0.06371 (1, 122) 5.015 0.027 **
Volume 12856.4 4551.33 (1, 122) 307.888 0 ***
Trade size 7.82471 1.90786 (1, 122) 589.097 0 ***
Trade number 1643.08 2385.39 (1, 122) 835.106 0 ***
Liquidity ratio 32784.2 8459.71 (1, 122) 431.963 0 ***
Return volatility 0.0429 0.03827 (1, 122) 2.045 0.155
Price volatility 27.1498 19.8455 (1, 122) 9.536 0.002 ***
Quote spread 0.23113 0.16161 (1, 191550) 13874.9 0 ***
Effective spread 0.26445 0.16817 (1, 191550) 12928.9 0 ***
Percentage quote spread 0.07541 0.08172 (1, 191550) 588.075 0 ***
Percentage effective spread 0.08638 0.08505 (1, 191550) 14.758 0 ***
Dollar-weighted percentage quote spread 0.0766 0.0842 (1, 122) 5.481 0.021 **
Dollar-weighted percentage effective spread 0.10752 0.10085 (1, 122) 2.226 0.138
Volume 2316.11 3167.54 (1, 122) 79.92 0 ***
Trade size 2.25729 1.50512 (1, 122) 126.025 0 ***
Trade number 1028.03 2105.36 (1, 122) 822.666 0 ***
Liquidity ratio 6480.94 6017.98 (1, 122) 2.965 0.088 *
Return volatility 0.06694 0.06194 (1, 122) 0.93 0.337
Price volatility 1.68267 1.21704 (1, 122) 10.933 0.001 ***
Quote spread 0.78825 0.72162 (1, 143174) 539.333 0 ***
Effective spread 0.87138 0.6786 (1, 143174) 3146.48 0 ***
Percentage quote spread 0.10442 0.11146 (1, 143174) 267.909 0 ***
Percentage effective spread 0.11533 0.10472 (1, 143174) 438.288 0 ***
Dollar-weighted percentage quote spread 0.10855 0.11528 (1, 122) 1.586 0.21 Dollar-weighted percentage effective spread 0.13894 0.11583 (1, 122) 19.385 0 ***
Volume 1436.86 2097.97 (1, 122) 30.167 0 ***
Trade size 1.84687 1.35072 (1, 122) 89.079 0 ***
Trade number 725.026 1544.59 (1, 122) 315.478 0 ***
Liquidity ratio 4162.28 6134.01 (1, 122) 38.588 0 ***
Return volatility 0.08042 0.05582 (1, 122) 69.427 0 ***
Price volatility 4.27654 3.33455 (1, 122) 5.893 0.017 **
ANOVA F-statistic
Panel A: Taiwan Stock Exchange Capitalization Weighted Stock Index (TX)
Panel C: Taiwan Stock Exchange Banking and Insurance Sector Index (TF) Panel B: Taiwan Stock Exchange Electronic Sector Index (TE)
Table 5. A summary of distribution of transaction price with respect to the quote bid-ask price
Old New
% %
Panel A: Taiwan Stock Exchange Capitalization Weighted Stock Index (TX)
Above ask 14 9
At ask 29 32
Between ask and bid 10 14
At bid 31 35
Below bid 16 10
Panel B: Taiwan Stock Exchange Electronic Sector Index (TE)
Above ask 10 7
At ask 33 31
Between ask and bid 11 18
At bid 35 35
Below bid 11 9
Panel C: Taiwan Stock Exchange Banking and Insurance Sector Index (TF)
Above ask 8 6
At ask 36 31
Between ask and bid 8 22
At bid 39 34
Below bid 9 7
Table 6. A comparison of distribution of effective spread
Old New
％ Max Average Min ％ Max Average Min
Panel A: Taiwan Stock Exchange Capitalization Weighted Stock Index (TX)
Zt > 0 46 51 1.6608 0.5 45 16 1.2344 0.5
Zt = 0 4 - - - 6 - -
-Zt < 0 50 -0.5 -1.7088 -55 49 -0.5 -1.2297 -21
Panel B: Taiwan Stock Exchange Electronic Sector Index (TE)
Zt > 0 43 2.7 0.1219 0.025 44 0.95 0.0865 0.025
Zt = 0 6 - - - 5 - -
-Zt < 0 51 -0.025 -0.1413 -2.5 51 -0.025 -0.0895 -1.25
Panel C: Taiwan Stock Exchange Banking and Insurance Sector Index (TF)
Zt > 0 46 8.7 0.4319 0.1 45 5 0.351 0.1
Zt = 0 2 - - - 4 - -
-Zt < 0 52 -0.1 -0.4498 -10.5 51 -0.1 -0.3606 -4.5
1. The effective, ztis defined as:
t t
t P Q
z
where Ptis the transaction price, represents one-half the signed effective spread with zt< 0 for a sell order and zt> 0 for a buy orders.
Table 7. Empirical Estimates of the spread components of the Effective spread
Old New
Estimate t-statistic R square Estimate t-statistic R square Panel A: Taiwan Stock Exchange Capitalization Weighted Stock Index (TX)
λ 0.531 140.493 0.160 0.327 52.75 0.019
γ 0.256 54.484 0.028 0.521 78.84 0.042
θ 0.229 75.699 0.052 0.151 57.32 0.023
Panel B: Taiwan Stock Exchange Electronic Sector Index (TE)
λ 0.440 114.654 0.170 0.216 84.498 0.053
γ 0.260 50.997 0.039 0.323 98.678 0.071
θ 0.300 79.674 0.090 0.459 184.14 0.210
Panel C: Taiwan Stock Exchange Banking and Insurance Sector Index (TF)
λ 0.399 139.616 0.299 0.154 92.684 0.081
γ 0.301 61.987 0.078 0.258 99.193 0.092
θ 0.300 67.258 0.090 0.588 227 0.346
6. Each parameter is estimated for the nearest –month contract of each of the three products, TX, TE, and TF in each of the periods, the old trading system and the new trading system.
7. The adverse information is estimated:
1
1
Qt zt et (9)
where Qt1Qt1Qt, Qtis the quote midpoint at time t; zt Pt Qt, Ptis the trade price at time t; λ is the adverse information component of effective spread.
8. The order processing cost is estimated:
1
1 1 ; γis the order processing component of effective spread.
9. The order persistence is estimated:
1
1
t t
t
z
z
where θ is the order persistence measure.