سازندگان بازار محلی، نقدینگی و کیفیت بازار

We examine the role of geographically proximate (local) market makers in providing liquidity and improving the quality of a dealer market. Firms with active participation of local dealers enjoy lower quoted and effective spreads, as well as more informative prices. The beneficial effects from local market makers are not confined to a few “top” local dealers and they cannot be attributed to their participation in the firm's IPO syndicate or industry specialization. Further, we find that days with aggressive bidding from local market makers relative to their non-local counterparts are associated with significant positive abnormal returns, consistent with local market makers possessing information advantages. In summary, our results suggest that the information advantages of local market makers may be a contributing factor to the reduction in the cost of trading.

A unique feature of a dealer market such as the NASDAQ stock market is the presence of multiple market makers. With multiple dealers, features of market structure such as the total number of dealers, degree of dealer concentration, and the market share of the dominant dealer have been shown to play an important role in the determination of the bid-ask spread.1 The presence of multiple dealers also raises the possibility that dealers differ in their characteristics, with these differences impacting liquidity and cost of trading. In this paper, we examine the geography of market makers and study whether differences in location affect their market making activities, and the resulting trading costs for these stocks. NASDAQ dealers are geographically dispersed: approximately 61% of these market makers are not located in New York City and about 33% are not located in an urban area.2 Such geographic dispersion of dealers implies that some dealers are located close to the firms in which they make a market. We study whether proximity of dealers, and their active participation in market making, is a non-trivial feature of the market microstructure of stocks, and has an impact on liquidity and market quality. Why does the location of dealers matter for the firms in which they make a market? Recent literature on geography points to two potential effects that geographic proximity may have on a dealer's market making decisions. First, several studies find that retail investors bias their portfolio towards geographically proximate or local firms.3 A similar bias towards local stocks is also seen in the portfolios of institutional investors. This implies that a large fraction of the trading volume in a stock originates locally and is likely to flow to the local market maker. Indeed, Schultz (2003) documents that market makers in the same state as the firm account for a large fraction of the firm's trading. The local market makers' access to a potentially large and predictable order flow is likely to have a significant impact on their quoting and trading activities. Second, there is also increasing evidence that geographic proximity may lead to information advantages. These information advantages have been shown in many diverse situations like bank lending, accuracy of analyst forecasts, and acquisitions.4 If geographic proximity is associated with information advantages, then local market makers are likely to have superior information relative to their non-local counterparts. Such information advantages of local market makers could arise if they know their client base well, or are knowledgeable about whether or not certain orders are informed. In addition, through their social, civic, or business interactions with managers and employees of local firms, market makers may have better information about these firms. Their access to detailed coverage of regional business by local media as well as interactions with local financial intermediaries, like banks, may also give them information advantages over other market makers. If easy access to order flow reduces dealers' inventory risks, and information advantages associated with location alleviates their potential adverse selection losses to informed traders, it is natural to expect local dealers to be more aggressive than their non-local counterparts in making a market in nearby firms. Therefore, firms that attract active local dealers may enjoy the resulting benefits of more competitive quotes and informative prices than those without much local dealer participation. We examine this using a sample that consists of all NASDAQ stocks from January 1999 to December 2003, with local market makers defined as those located in the same Metropolitan Statistical Area (MSA) as the firm's headquarters. On average about 19.48% of a firm's dealers are classified as local. The importance of local dealers can also be assessed by the fact that they trade 11.49% of the share volume and improve inside quotes in 14.16% of the cases. This role of local dealers, whether captured as the fraction of all dealers, or of total volume or how often they improve inside quotes is significantly higher than what would be expected if market making business was randomly distributed. The finding that the stock's trading volume and quote revisions are disproportionately concentrated in local dealers is consistent with dealers' preference for making markets in local firms documented in Schultz (2003), as well as in a contemporaneous paper by Anand, Gatchev, Madueira, Pirinsky, and Underwood (2008), who find that local dealers account for a larger share of price discovery than their non-local counterparts. Whereas both Schultz (2003) and Anand, Gatchev, Madueira, Pirinsky, and Underwood (2008) study the impact of proximity on dealers' market making decisions, the implications of such findings on the cross-section of firms has not been examined so far. In particular, is a large presence of proximate dealers advantageous for the firm? In this paper, we fill in the gap by focusing on whether dealers' bias towards local market making activities impacts the liquidity and price informativeness of a firm's stock and therefore its cost of capital. To this effect, we find that active local market maker participation is associated with a significant reduction in both quoted and effective spreads as well as an improvement in market quality as formulated by Hasbrouck (1993). Specifically, an increase in the share of a firm's market making business by local dealers from the median to the top quartile is associated with a reduction of quoted spreads by 30% and in effective spreads by 12.87%. It is worth noting that this significant role of local market makers is after controlling for industry specialization of the dealer and his participation in the firm's IPO syndicate. Further, such local effect is not confined to a few “top” local dealers. We find that though the “top” local dealer significantly reduces spreads, the local effect is seen more broadly for other non-top local dealers as well. In summary, our findings suggest a significant role of local market makers in improving liquidity and market quality. The effect of local dealer participation can be illustrated by the following example. PentaStar Communications Inc. and ACT Teleconferencing Inc., two firms in our sample, operate in the same industry, are of comparable size, and are both located in the Denver-Aurora MSA. However, the percentage of times that local dealers are at the inside market is 39.47% for PentaStar and 5.31% for ACT. In line with our findings, PentaStar exhibits a lower spread, about 1.45% of price, compared to 2.71% for ACT. As there is increasing evidence that geographic proximity facilitates information transmission, we examine whether this effect of local market makers in improving liquidity and market quality can be at least partially attributed to their potential information advantages. If the effect of local market makers on spreads and market quality arises from their information advantages, such effects should be larger for firms that have greater information problems. Consistent with this, we find that the effect of local market makers appears to be larger for firms located in less populated areas, and those with high information-based trading (PIN) measures as proposed in Easley and O'Hara (2004). Further, if the local effect arises from information advantages, such effects should be more significant when local dealers have regional, as opposed to national, presence. We find some evidence that local regional market makers are associated with a greater reduction in spreads and improvement in market quality relative to non-regional local market makers. During our sample period, bid-ask spreads drop significantly following the decimalization implemented in March 2001. We study the pre- and post-decimalization periods to shed light on how this impacted the role of local market makers. We find similar effects of local dealers on improving market quality and reducing quoted and effective spreads across decimalization. Lastly, to directly examine the information content of local dealer quoting behavior, we test whether their active participation on one side of the inside market is associated with price movements in the corresponding direction.5 For example, when local dealers are active in improving the bid prices, there is likely to be good information about the firm, which should be reflected in a positive abnormal return. Indeed, we find that on days after the local dealers are disproportionately improving the bid prices relative to their non-local counterparts, the stock earns a significantly higher abnormal return than when they are not. Several studies have documented the prevalence of preferencing practices on the NASDAQ market. Internalization and payment for order flow discourages dealers from posting competitive quotes, resulting in a wider bid-ask spread.6 However, Blume and Goldstein (1997) find that a non-NYSE market does attract additional order flows when it improves the market best quotes, despite a large fraction of order flow being preferenced.7Bessembinder (2003) documents substantial (but imperfect) quote-based competition for order flows in NYSE-listed stocks. Further, it has been shown that competitive quotes increase market share (Barclay et al., 2003 and Chung and Chuwonganant, 2007), and the probability of executions (Goldstein, Shkilko, Van Ness, and Van Ness, 2008). This suggests that there exist incentives for local market makers to quickly update their quotes to reflect their new information, and indeed our results indicate that this is the case. The rest of the paper is organized as follows. Section 1 discusses the potential impact of predictable order flow and information advantages on spreads and market quality, while Section 2 describes the data. Section 3 documents results on the effect of local market making on a firm's spread and market quality and Section 4 examines the information content of the local effect. Lastly, Section 5 discusses robustness and Section 6 concludes.

Westudytheroleofgeographicallyproximatemarketmakersinprovidingliquidityand improvingmarketquality.Wefindstrongevidencethatactivequotingandtradingbylocal dealersisassociatedwithlowerquotedandeffectivespreads,aswellas,lowerdispersions in thepricingerror.Asthiseffectoflocalmarketmakersisstrongerinfirmswithgreater informationproblems,theresultssuggestthatlocalmarketmakersmayindeedhave informationadvantages.Thisisfurthersupportedbyourfindingthatdaysinwhichthe local marketmakersaredisproportionatelyactiveonthebidsideareassociatedwith significantlypositiveabnormalreturns. Thesefindingscarryimportantimplicationsformarketefficiency.Ashasbeen demonstratedbytheoreticalmicrostructuremodels,theposteriorbeliefsofmarketmakers will convergetothefull-informationvalueofanasset.Theactualconvergence,however,is not instantaneous.Whileitisdifficulttopreciselycharacterizetheactualadjustmentpaths, our empiricalworksuggeststhattheinformationadvantagesassociatedwithgeographic location mayindeedspeedupthepricediscoveryprocessaslocalmarketmakerstendto quote morecompetitivelytotakeadvantageoftheiruniqueinformationset. Further,priorworkby Schultz(2003) shows thatmarketmakerschoosetomake markets infirmsinwhichtheypossessinformationadvantages.Evidenceinthispaper indicatesthatfirms,especiallythosesufferingfromgreaterinformationproblems,gain significantlyfromattractingpotentiallyinformedlocalmarketmakers.Asliquidityand transactioncostsimpactanasset’sreturnandhenceafirm’scostofcapital,ourstudylends supporttotheargumentproposedby Easley andO’Hara(2004) that theinformation structuresurroundingafirm’sstockisanimportantdeterminantofthefirm’scostof capital.