【正文】
hoped to win investment banking business from. Large sums of money helped to blur the ethical lines: between December 1999 and November 2000, the Internet group at Merrill produced $115 million in investment banking fees for the conflicts of interest may arise in firms that sponsor sellside research and make markets in securities. Analyst remendations create profit opportunities on an agency basis through brokerage missions and on a principal basis through proprietary trading. Weiss (2003)notes that for the 10 firms cited by Spitzer, their aggregate share of revenues from missions (9%) and trading (%) exceeded those from investment banking (%).The academic literature has found indirect evidence of these conflicts. Ellis, Michaely and O’Hara (2000) find that the lead underwriter typically bees the most active market maker in the stock and that 23% of profits e from inventory gains and trading. Aggarwal and Conroy (2000) highlight the important role underwriters play in price discovery in the aftermarket for IPOs. Irvine (2003) finds that trading activity in Toronto Stock Exchange issues increases at the analysts’ firm in the two weeks following earnings and remendation changes. Chung, McInish,Wood, and Wyhowski (1995) demonstrate that analysts are more likely to cover stocks on the NYSE and Nasdaq with wide bidask spreads. Schultz (2003) also notes that a Nasdaq firm is more likely to make markets in stocks in which they have analyst paper finds empirical evidence of market makeranalyst conflicts in the period leading up to a remendation change. Section I begins by describing principal and agent relationshipsthat influence liquidity. There is also a discussion of rules and ethical standards regulating these Nasdaq2 limit order book, described in Section II, allows me to observe liquidity from all market makers. Historical data from the order book is collected in Nasdaq’s Nastraq II also develops a variety of measures of bid or ask pressure. The first relies on the frequency with which a market maker provides the best available price in the marketplace. A second measure takes into account depth on these occasions. A final measure looks at trading cross reference, in Section III, the Thomson First Call database of analyst remendations with data from the order book. I have three plete years of overlap in both data sets, 1999 to 2001. In total, I examine nearly 1, 600 remendations and the corresponding market making all three liquidity measures, I find evidence of increased bid activity prior to upgrades in the years 1999 and 2000 in Section IV. For downgrades in all years, and for upgrades in 2001, this pattern is not in evidence. In Section V, I f