Research Proposal. Order Imbalance around Corporate Information Events. Shiang Liu Michael Impson University of North Texas.

Transcription

1 Research Proposal Order Imbalance around Corporate Information Events Shiang Liu Michael Impson University of North Texas October 3, 2016

2 Order Imbalance around Corporate Information Events Abstract Models such as the PIN measure proposed by Easley, Kiefer, and O Hara and Kyle s Lambda developed in the 1980's have been extensively used to detect the presence of informed investors and information asymmetry in financial markets. However, the presence of informed investors has been difficult to detect, so prior studies have made assumptions about market participants who may be informed, like large institutional investors. More recent studies [Collin- Dufresne and Fos (2013), Petchey, Wee, and Yang (2015) ] that use samples in which the presence of informed investors is known (SEC 13D filings and convicted cases of insider trading) have shown low levels of adverse selection and enhanced liquidity during periods when informed traders are in the market. We propose using a simple measure of information asymmetry, trade imbalance, to investigate levels of asymmetry around corporate dividend and earnings announcements. 2

3 Introduction Liquidity as measured by bid ask spreads and price impact of trades is a topic of great interest in current research in financial markets. Models that attempt to measure the existence and levels of information asymmetry in financial markets and the concept of adverse selection have existed for a number of years [Benston and Hagerman, (1974), Copeland and Galai, (1983), Glosten and Milgrom, (1985), Kyle, (1985), Easley and O Hara, (1987), Admati and Pfeiderer, (1988)]. These models assume that adverse selection manifests itself in actions of market makers who widen spreads to offset losses from trading with informed traders. Adverse selection costs are extracted by models that decompose spreads into order processing costs, inventory holding costs, and adverse selection costs [Glosten and Harris, (1988), Stoll, (1989), Hausbrook, (1991)]. Recent empirical studies test how well liquidity and adverse selection models describe information asymmetry [Barclay and Hendershott, (2004), Vega, (2006), Duarte, et al, However, the information that informed traders act on is private and they trade to disguise their actions, so in most cases it is difficult to know when informed traders are present and what information that they possess. So assumptions are made about the presence of traders, like institutions, who may possess valuable private information. Another model of information asymmetry, the Probability of Informed Trading (PIN) model of Easley, Kiefer, and O Hara, 1997, 1998 has been widely used. The PIN estimates the probability that a trader possesses inside information as a basis for the trade. A potential limitation of PIN is that it is based on maximum likelihood estimation and requires a relatively long estimation window for accurate parameter estimation, so PIN s are not available on a daily or intra-day basis. Also, more recent studies have produced confounding results of PIN s 3

4 decreasing before information events like merger and acquisition and earnings announcements [Aktas, et al, 2007, Benos and Jochec, 2007]. A study by Petchey, Wee, and Yang (2015) that uses a sample based on convicted cases of insider trading so that there is no question about whether or not trades were based on inside information also finds that PIN s decline in the period before the M & A announcement. Similarly, a study by Collin-Dufresne and Fos, (2012) that uses trade data from SEC 13D records finds that liquidity improves rather than declines during periods of active trading by 13D filers. Further, measures of adverse selection like Kyle s (1985) Lambda, PIN, and Amihud s (2002) illiquidity measure indicate that adverse selection is lower during active trading by 13D filers. These results suggest that alternate measures of adverse selection should be considered. Data and Sample Preliminary results are based on a pilot sample of dividend omissions for 32 firms over the period 2004 until A larger dividend omission sample will be collected via a web search. Dividend increase and decrease announcements will be obtained from the Center for Research in Security Price (CRSP) database and earnings announcments will be obtained from the IBES database. Samples will be screened for contaminating events within five days before or after the announcement Market microstructure information is obtained from the monthly Trade and Quote (TAQ) database. Trade direction is determined using the Lee and Ready (1991) method. Method of Study As reviewed above, there have been numerous methods employed to assess liquidity and trading on private information. In this study, we focus on one measure used to detect information 4

5 asymmetry, order imbalance. In their review of liquidity measurement using TAQ data, Holden and Jacobsen (2014) point out that estimates of order imbalance are likely to be unbiased, in contrast to estimates of spread and price impact. They also point out two advantages of order imbalance compared to PIN measures: (1) It can be computed over relatively short periods of time and (2) It does not have the potential numerical overflow problems when computing the log-likelihood function required for PIN calculations. Moreover, in Holden and Jacobsen s tests, absolute order imbalance calculated with no adjustments compares very closely with the absolute order imbalance calculated using daily TAQ. Considering all these factors and the fact that calculation requires only a simple calculation, it is a worthy candidate for further investigation. Absolute order imbalance is calculated as: Buys - Sells Absolute Order Imbalance = Buys + Sells (1) We propose to use Absolute Order Imbalance (AIM) and Signed Order Imbalance (SIM) along with liquidity measures like the effective spread to to determine the extent of information asymmetry and liquidity levels around dividend announcements and earnings announcements. In addition, we propose to validate the AIM measure by using it to investigate levels of information asymmetry using a sample of 13D filings, similar to the study by Collin-Dufresne & Fos. We include some preliminary evidence on dividend omission announcements in this proposal. Significance tests are based on an empirical distribution generated over a 100 day period prior to the 21 day event period, days -11 to -110 relative to the event date. 5

6 Preliminary Results Preliminary results are shown in Table 1. The AIM measure showed no evidence of significant order imbalance on any day during the forecast period. SIM results show significant sell imbalance on days -1,-2, -8, and -9 relative to the event date. There is also significant sell imbalance on days +5 and + 8. The pre-event sell imbalance offers some evidence of the possibility of trading on private information during that period. There is also no evidence of significant widening of the spread during the forecast interval based on the effective spread measure. In fact, the spread is significantly smaller that normal on numerous days within the forecast interval. So the spread measure shows that liquidity was normal or better than normal during the forecast period. Turnover is significantly higher than normal on the event day and the day following, reacting to the public release of the dividend omission information. Turnover shows no evidence of significant trading, possibly by insiders, prior to the announcement. Similarly, trade size is significantly larger only on the day following the announcement date. The number of trades per day is significantly larger on five of the ten days following the dividend announcement. Preliminary Conclusions Based on the results of the pilot sample, there is little evidence of significant insider trading prior to dividend omission announcements. Only the SIM measure indicates significant sell imbalance early in the forecast period. The measures of trading activity, turnover, trade size, and number of trades provide evidence of improved or no change in liquidity during the forecast period. These results seem consistent with the more recent literature that indicate that adverse selection declines and liquidity improves during periods when there is trading based on private or 6

7 inside information. Investigating additional events with larger samples should produce more conclusive results. As a minimum, the testing of AIM on a sample of 13D filings should provide evidence of its validity as a measure of information asymmetry. 7

8 Table 1 Dividend Omission Announcements Day AIM SIM Effective Spread Turnover Trade Size Number of Trades , ** ,899 1, * ,764 1, ,812 1, , *** ,733 1, ,417 1, ,749 1, ** 0.041* ,817 1, * 0.040** ,495 1, *** 2,759 1, * 0.026*** 3,524** 1, * ,676 1,843*** ,855 1,932*** ** ,833 1,781*** * 0.035*** ,653 1, ** ,713 1, ,122 1, * 0.037*** ,798 1,659** * ,686 1,605** *** ,683 1,387 * Significant at the 10% level based on the empirical distribution 8

9 ** Significant at the 5% level based on the empirical distribution *** Significant at the 1% level based on the empirical distribution 9

10 List of References Admati, Anat, and Paul Pfeiderer, 1988, A theory of intraday patterns: Volume and price variability, Review of Financial Studies 1, Amihud,Y., 2002, Illiquidity and stock returns: cross-section and time series effects, Journal of Financial Markets 5, Atkas, N. E. De Bodt, F. Declerck, and H. Van Oppens, 2007, The PIN anomaly around M&A Announcements, Journal of Financial Markets 10 (2), Barklay, M. and T. Hendershott, Liquidity exterrnalities and adverse selection: evidence from trading after hours. Journal of Finance 59 (2), Benos, E. and M. Jochec, 2007, Testing the PIN Variable, Working Paper, University of Illinois. Benston, G. And R. Hagerman,(1974), Determinants of the bid-ask spread in the over-thecounter market, Journal of Financial Economics 1 (4), Chae, Joon, 2005, Trading Volume, Information Asymmetry, and Timing Information, Journal of Finance 60, Collin-Dufresne, P. and V. Fos, 2015, Do prices reveal the presence of insider trading?, Journal of Finance 70 (4), 1555, Copeland, Thomas and Dan Galai, 1983, Information effects on the bid-ask spread, Journal of Finance 38, Duarte, J., Han, X., Harford, J., Young, L., 2008, Information asymmetry, information dissemination and the effect of regulation fd on the cost of capital. Journal of Financial Economics 87 (1), Easley, D., and M. O Hara, 1987, Price, trade size and information in securities markets, Journal of Financial Economics 19, Easley, D., Kiefer, N.M., & O Hara, M. (1997). One day in the life of a very common stock. Review of Financial Studies 10, Glosten, L and L. Harris, 1988, Estimating the components of the bid/ask spread. Journal of Financial Economics 21 (1),

11 Glosten, Lawrence, and Paul Milgrom, 1985, Bid, ask and transaction prices in a specialist market with heterogeneously informed traders, Journal of Financial Economics 14, Holden, C. And S. Jacobsen, 2014, Liquidity measurement problems in fast, competitive markets: Expensive and cheap solutions. Journal of Finance 69 (4), 1,747-1,785. Kyle, A. (1985), Continuous auctions and insider trading. Econometrica, 53, Lee, C.M. C., and Ready, M.J. (1991) Infering trade direction from intraday data. Journal of Finance Petchey, J., M. Wee, and J. Yang, Does the probability of informed trading measure informed trading?, Working Paper, University of Western Australia. Stoll, H., 1989, Inferring the components of the bid-ask spread: Theory and empirical tests. Journal of Finance 44 (1), Vega, C., 2006, Stock price reaction to public and private information. Journal of Financial Economics 82 (1)