Geographic dissemination of information

Abstract

Urban companies are located near millions more potential investors and sophisticated money managers than non-urban companies. More investors are familiar with urban companies and have access to informal information about them. The stock of urban companies is also more liquid than the stock of non-urban companies. We hypothesize that these factors lead information to be spread from urban companies to other companies. Urban stock returns lead rural/small city stock returns even controlling for size, industry, and analyst coverage. Closer examination of the lead–lag relation reveals that urgent trades, which are likely to reflect short-lived information, are much more common for urban firms. Information appears to be uncovered through informal means more easily available to people physically near a company. We discuss the corporate finance implications of our findings.

Introduction

Information plays a central role in all economic models in both asset pricing and corporate finance. A central question is how information is disseminated. Hayek (1945) notes that the price mechanism is an efficient provider of information. This is especially true for localized information of time and place. Smith (1991) remarks that local shareholders have a lower marginal cost of obtaining information. If localized information is more easily obtained by investors in close proximity to the main source of information, one might expect the dissemination of prices (i.e., information) to flow from more densely populated areas to more sparely populated areas of the country.

This paper explores the way physical proximity to investors affects the dissemination of information across stocks in the U.S. during 1973–2002. Many people and many potential investors are located near companies headquartered in urban areas, defined as the ten largest metropolitan areas of New York City, Los Angeles, Chicago, Washington, San Francisco, Philadelphia, Boston, Detroit, Dallas, or Houston. Money managers and major brokerage firms are sited predominantly in our urban metropolitan areas. Stocks of companies headquartered in rural or small city areas, defined as locations outside the top ten metropolitan areas, have fewer people and fewer investors nearby. We examine the lead–lag relation between urban and rural/small city stocks to test whether information spreads from stock of companies located in heavily populated areas to stock of companies located farther from potential investors.

We report two major findings. First, for a sample of over 19,000 different firms we show that the returns of urban company stock portfolios lead the returns of rural/small city company stock portfolios; rural/small city stocks do not lead urban stocks. This lead–lag relation is not explained by differences in size, industry, or analyst coverage. The information the financial markets appear to first incorporate into urban-based stocks is then absorbed into the stock prices of firms headquartered outside the largest ten U.S. metropolitan areas. This slow spread of information is consistent with the theoretical work of Hong and Stein (1999).

Second, there is direct evidence that more informed trading takes place in urban Nasdaq stocks than rural/small city Nasdaq stocks. We examine trades that take place for 1000 shares, the typical quoted depth for Nasdaq stocks. Several microstructure studies find that “urgent trades” of this type are more likely to be made on short-lived or perishable information than other trades. We demonstrate that urgent trades constitute a higher proportion of total trades for urban stocks than non-urban stocks. This finding is consistent with the role of serendipity in information collection (see Titman and Subrahmanyam, 1999). Urban-based companies have more potential investors who could, by chance, stumble across and trade on valuable information.

It is well established that both institutions and individuals invest disproportionately in nearby companies. Huberman (2001) speculates that one reason for the bias toward local companies appears to be familiarity. When news is revealed that is relevant for a number of companies, investors are likely to first trade the local stocks that are familiar — and for most investors these will be urban stocks. Hence, we expect to see information that is common to many firms spread from urban to rural/small city stocks.

Coval and Moskowitz (2001), Grinblatt and Keloharju (2001), and Ivkovic and Weisbenner (2005) point to a second reason for investors' bias toward local firms — superior access to information. Both individual investors and institutions earn significantly higher returns on their local investments than on their positions in stocks in farther away companies. Malloy (2005) finds that analysts do a better job of forecasting earnings in nearby companies even after adjusting for underwriting relationships.

Our key contribution is noting the informal ways information is discovered and diffused across stocks. Urban stocks are based in cities with many investors. Thus informal sources of information about them, such as conversations with employees or customers, are available to many potential traders and investors. Local media coverage of local companies also reaches large numbers of investors in urban areas. Our urban areas have large concentrations of institutional investors, brokers, and investment bankers, and we would expect these sophisticated investors to be the most adept at uncovering or interpreting information that is relevant for the valuation of numerous stocks in an industry or sector.

Understanding how new marketwide or industrywide information will affect the value of individual stocks requires background information on companies. The impact of new marketwide information will be more readily understood for local stocks. This suggests that more investors will more quickly grasp the implications of marketwide or industrywide news for urban stocks than for rural/small city stocks because it is easier/cheaper to obtain information on local stocks, this implies a dispersion of information from urban to other stocks.

There are two obvious corporate finance implications of the paper. First, the fact that rural firms lag the returns of urban firms has timing ramifications for the share repurchase decisions of corporations. If the stock market is having a strong return on a given day, the implication is for rural/small city managers to immediately purchase the shares instead of waiting a day. Since rural/small city firm stock returns lag the returns of urban companies, it would be cheaper for the firm to repurchase shares prior to the next day price increase. Recall that corporate managers have discretion in when the open market share repurchases are preformed.

The second corporate finance implication of our findings deals with the announcement of accelerated seasoned equity offerings. Bortolotti, Megginson, and Smart (2006) note that accelerated follow-on offerings are becoming much more prevalent. These accelerated equity offerings are popular due to the ability to get the deal done quickly. This exposes the issuing firm to less price risk during the underwriting period prior to the offering. Rural/small city mangers should delay having an accelerated bought deal after a market rise. These managers should wait until the following day to price the offering with the investment banker. This delay in the timing of the accelerated offering will enable the firm to get higher proceeds, all else being equal.

The structure of this paper is as follows. We discuss our hypotheses and provide a literature review in Section 2. In Section 3, we describe our data and methodology. The empirical results are in Section 4 and we conclude in Section 5.