Informed Trading and Liquidity


The objective of this chapter is to develop the underlying research hypotheses for the second part of this study. Based on the existing theoretical, empirical, and experimental literature, testable hypotheses are developed to answer the questions (research objectives) as outlined in Chapter 1.2


Limit Order Market Maker Order Book Market Order Informed Trader 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 162.
    For a quote-driven market structure, theoretical models are provided by Copeland/ Galai (1983) and Glosten/Milgrom (1985). Copeland/Galai (1983) formally analyze the trade-off of profits when trading with uninformed traders and the losses due to trading with informed traders. Glosten/Milgrom (1985) develop a model where the market maker is the sole provider of liquidity and quotes a bid-ask spread upon which one agent in each round can either submit a market buy or sell order. The informed trader’s optimal order placement foresees that he enters a buy market order when his assumed value of the instrument is above the posted ask and a sell market order if his assumed value is below the posted bid. In any other case, the informed trader does not enter an order. Both Copeland/Galai (1983) and Glosten/Milgrom (1985) ignore the depth dimension of liquidity as they assume one unit size for all trades. Kyle (1985) provides a dynamic model with uninformed traders, one informed trader, and several market makers. As the informed trader does not face any competition, he gradually resolves his informational advantage. In contrast, Holden/Subrahmanyam (1992) introduce competitive informed traders, resulting in an aggressive trading strategy.Google Scholar
  2. 163.
    See Glosten (1994), p. 1139ff. (proposition 3).Google Scholar
  3. 164.
    See Glosten (1994), p. 1129.Google Scholar
  4. 165.
    Similar results are found by Chakravarty/ Holden (1995), p. 216 and Daníelsson/Payne (2001), p. 5.Google Scholar
  5. 166.
    See Seppi (1997), p. 103. Parlour/Seppi (2003) analyze competition between a hybrid market structure and a pure order-driven market structure. They implement the limit order model of Seppi (1997) and model the economics of liquidity supply and demand.Google Scholar
  6. 167.
    For an explanation of transitory and permanent volatility, see Footnote 68 in Chapter 3.1.Google Scholar
  7. 168.
    See Handa/ Schwartz (1996b), p. 1860. See also Handa/Schwartz/Tiwari (1998), who describe order-driven market structures as ecological systems and Handa/Schwartz/Tiwari (2003), who provide evidence in accordance with Handa/Schwartz (1996b) for the CAC40 instruments at Paris Bourse.Google Scholar
  8. 169.
    See Parlour (1998), p. 793.Google Scholar
  9. 170.
    See Foucault/ Kadan/ Kandel (2005), p. 1172.Google Scholar
  10. 171.
    See Foucault/ Kadan/ Kandel (2005), p. 1197f.Google Scholar
  11. 172.
    See Foucault (1999), p. 105 and p. 110f. Winner’s curse means that a buy (sell) limit order of an uninformed trader is executed because its limit price is too high (too low).Google Scholar
  12. 173.
    See Foucault (1999), p. 112 and p. 116 (corollary 2).Google Scholar
  13. 174.
    See Griese/ Kempf (2006), p. 403. For DAX instruments, they find that approx. half of the instruments reveal a convex function while the other half show a concave function. See p. 413.Google Scholar
  14. 175.
    See Ahn/ Bae/ Chan (2001), p. 782, Irvine/Benston/Kandel (2000), p. 22f. and Chordia/Roll/Subrahmanyam (2002), p. 117. See also Beltran/Giot/Grammig (2005), p. 15 and Daníelsson/Payne (2001), p. 4, who find that depth for the buy and sell side depth are uncorrelated, i.e. liquidity suppliers enter limit orders only on one side of the market.Google Scholar
  15. 176.
    See Gomber/ Schweickert/ Theissen (2005), p. 10.Google Scholar
  16. 177.
    The presented theoretical and empirical literature focused on firm-specific determinants of liquidity. A unique research arm on commonality in liquidity has evolved. Commonality means a common variation or co-variation of liquidity across instruments, i.e. a systematic component of liquidity. The intuition for a liquidity risk factor is that other firm-specific attributes (risk and return) are influenced by systematic factors. Chordia/ Roll/ Subrahmanyam (2000), Hasbrouck/Seppi (2001), and Huberman/Halka (2001) analyze commonality in a market with designated liquidity providers, while Brockman/Chung (2002) and Domowitz/Hansch/Wang (2005) analyze the issue in a pure order-driven environment. Generally, all authors find affirmative results for a systematic component in liquidity. Chordia/Sarkar/Subrahmanyam (2005a) even find liquidity co-movement across asset classes (equities and bonds) and (2005b) analyze commonality across small and large firms.Google Scholar
  17. 178.
    See Kempf/ Mayston (2005), p. 16. In this study, commonality will not be analyzed but the fact that Kempf/Mayston (2005) find in line with existing empirical research for other hybrid and order-driven market structure commonality in Xetra (DAX instruments) is relevant for the first research question, if Xetra is comparable to other markets.Google Scholar
  18. 179.
    See Stoll (2000), p. 1480f.Google Scholar
  19. 180.
    See Demsetz (1968), p. 33ff.Google Scholar
  20. 181.
    See Tinic (1972), p. 81ff. and Benston/Hagerman (1974), p. 353 ff.Google Scholar
  21. 182.
    See Copeland/ Galai (1983), p. 1464.Google Scholar
  22. 183.
    See Daníelsson/ Payne (2001), p. 21f.Google Scholar
  23. 184.
    See Ahn/ Bae/ Chan (2001), p. 768f. and Bae/Jang/Park (2003), p. 535.Google Scholar
  24. 185.
    See, among others, Demsetz (1968), Tinic (1972), Benston/Hagerman (1974), Stoll (1978), and McInish/Wood (1992).Google Scholar
  25. 186.
    Admati/ Pfleiderer (1988), p. 7ff. develop a basic model with one risk-neutral market maker and several informed and uninformed traders.Google Scholar
  26. 187.
    Admati/ Pfleiderer (1988), p. 5. The authors use the term “thick market” instead of liquid market. As they refer to a market where trading activity has “little effect on prices”, they refer to a liquid market, as defined in Chapter 3.1.Google Scholar
  27. 188.
    See Admati/ Pfleiderer (1988), p. 34.Google Scholar
  28. 189.
    See Biais/ Hillion/ Spatt (1995), p. 1671f. and Ahn/Bae/Chan (2001), p. 773.Google Scholar
  29. 190.
    See Beltran/ Durée/Giot (2004), p. 13 and Ranaldo (2004), p. 56.Google Scholar
  30. 191.
    See Lin/ Sanger/ Booth (1995), pp. 1172–1176 and Madhavan/Richardson/Roomans (1997), p. 1055, who document that the adverse selection component of the spreads is highest at the start of trading and continuously declines throughout the trading day.Google Scholar
  31. 192.
    See McInish/ Wood (1992), p. 759f.Google Scholar
  32. 193.
    McInish/ Van Ness (2002) replicate the initial McInish/Wood (1992) study and find similar results. They implement spread measures to identify the intraday results for the different components of the spread.Google Scholar
  33. 194.
    See Madhavan (1992), p. 618f.Google Scholar
  34. 195.
    See Foster/ Viswanathan (1994), p. 510f.Google Scholar
  35. 196.
    See Nyholm (2002), p. 497.Google Scholar
  36. 197.
    See Brockman/ Chung (1998), p. 296f.Google Scholar
  37. 198.
    See Giot/ Grammig (2006), p. 879ff.Google Scholar
  38. 199.
    See Heidle/ Huang (2002), p. 392.Google Scholar
  39. 200.
    See Glosten (1994), p. 1152f.Google Scholar
  40. 201.
    See Admati/ Pfleiderer (1991), p. 444 and Chapter 4.2 for a description of sunshine trading.Google Scholar
  41. 202.
    Heidle/ Huang (2002), p. 395 find that informed trading is more pronounced in the anonymous setting of NASDAQ compared to NYSE and AMEX. Barclay/Hendershott/McCormick (2003), p. 2653ff. demonstrate that informed trading is even higher in the anonymous trading environment of ECNs (electronic communication networks) than on NASDAQ. Grammig/Schiereck/Theissen (2001), pp. 388–401 analyze the German equity market. They compare floor and electronic trading system of FSE, concluding that informed trading is much higher in the electronic trading system IBIS. Theissen (2002), p. 48f. and (2003b), p. 24 also finds that adverse selection costs are higher in the anonymous electronic trading system compared to the non-anonymous floor trading system in Germany. Jain/Jiang/McInish/Taechapiroontong (2003), p. 32ff. find similar results for the LSE, where the anonymous SETS system and a non-anonymous dealer market are operated in parallel.Google Scholar
  42. 203.
    See Easley/ Kiefer/ O’Hara/ Paperman (1996), pp. 1421–1428.Google Scholar
  43. 204.
    See Easley/ Kiefer/ O’Hara/ Paperman (1996) for NYSE, p. 1422. See also Huang/Stoll (1997), p. 1010, Chakravarty (2001), p. 302, Chung/Li (2003), p. 264, Chung/Li/McInish (2004), p. 12 and, Nyholm (2002), p. 499f. For open limit order books, evidence is found by Brockman/Chung (2000), p. 137 and Frey/Grammig (2006), p. 1026. Grammig/Schiereck/Theissen (2000), p. 631 do not find supporting evidence and explain their findings due to the relatively homogenous sample.Google Scholar
  44. 205.
    See Lin/ Sanger/ Booth (1995), p. 1164ff., Huang/Stoll (1997), p. 1004 and Glosten/Harris (1988), p. 128.Google Scholar
  45. 206.
    See Easley/ O’Hara (1987), p. 81. Hasbrouck (1988), p. 250f. finds that order size conveys information and increases in size. Easley/Kiefer/O’Hara (1997b), p. 178 demonstrate that large orders convey more information than small orders.Google Scholar
  46. 207.
    See Barclay/ Warner (1993), pp. 282, 285, and 302f.Google Scholar
  47. 208.
    See Hasbrouck (1995), p. 1196f.Google Scholar
  48. 209.
    See Chakravarty (2001), p. 299ff.Google Scholar
  49. 210.
    See Chakravarty/ Holden (1995), p. 233. In their single period model, the market makers first post a bid-ask quote, then both uninformed and informed traders simultaneously submit their orders, and lastly the orders are executed.Google Scholar
  50. 211.
    See Kaniel/ Liu (2006), p. 1871f. and p. 1892f.Google Scholar
  51. 212.
    See Harris (1998), p. 3f. The major constraint to the results of his analysis is that it only applies to small orders that execute at the bid-ask spread and accordingly do not impact prices. See Harris (1998), p. 62.Google Scholar
  52. 213.
    See Rindi (2002), p. 6. She stresses that the outcome of different transparency or anonymity regimes strictly depends on the underlying market structure. In a specialist structure, the disclosure of identities would help the specialist to reduce his adverse selection risk and accordingly liquidity would increase. In a centralized open limit order book with no dedicated liquidity provider, this effect differs significantly. Foucault/Moinas/Theissen (2004), pp. 30–34 provide evidence for the positive relation between pre-trade anonymity and liquidity for Euronext. Comerton-Forde/Frino/Mollica (2005) p. 534ff. find supporting results for the effects of pre-trade anonymity for Euronext, Tokyo Stock Exchange (TSE), and Korea Stock Exchange (KSE). Hachmeister/Schiereck (2006), p. 11f. provide analogous results for the relation between post-trade anonymity and liquidity for the Xetra trading system.Google Scholar
  53. 214.
    The trading protocol foresees continuous trading, price time priority, and the choice between market and limit orders. While the complete order book is visible ensuring pre-trade transparency, the traders remain concealed, foreseeing pre-trade anonymity. Bloomfield/ O’Hara/ Saar (2005), pp. 171–177 provide the details on experimental design and appendix A (pp. 194–197) provides the instructions given to the trading subjects.Google Scholar
  54. 215.
    See Bloomfield/ O’Hara/ Saar (2005), p. 168.Google Scholar
  55. 216.
    See Bloomfield/ O’Hara/ Saar (2005), pp. 186–189.Google Scholar
  56. 217.
    Bernhardt/ Miao (2004) are an exception as they model how informed traders profit from their information when information is generated throughout the trading day.Google Scholar
  57. 219.
    The cumulative price impact is calculated based on Barclay/ Warner (1993), where the price change for the current trade is computed as the difference between the price of the current and the previous trade. These price changes are then cumulated for different classes of trades.Google Scholar
  58. 220.
    See Anand/ Chakravarty/ Martell (2005), p. 290f.Google Scholar
  59. 221.
    See Anand/ Chakravarty/ Martell (2005), p. 302ff.Google Scholar
  60. 222.
    See Anand/ Chakravarty/ Martell (2005), p. 295f.Google Scholar
  61. 225.
    See Cao/ Hansch/ Wang (2004), pp. 14–18.Google Scholar
  62. 226.
    See Cao/ Hansch/ Wang (2004), p. 22f.Google Scholar
  63. 227.
    See Harris/ Panchapagesan (2005), p. 46.Google Scholar
  64. 228.
    See Harris/ Panchapagesan (2005), p. 62.Google Scholar

Copyright information

© Deutscher Universitäts-Verlag | GWV Fachverlage GmbH, Wiesbaden 2007

Personalised recommendations