asymmetric information, firm location, and equity issuance
TRANSCRIPT
Asymmetric Information, Firm Location, and Equity Issuance
Tim Loughran
Mendoza College of Business University of Notre Dame
Notre Dame, IN 46556-5646 574.631.8432 voice [email protected]
Paul Schultz
Mendoza College of Business University of Notre Dame
Notre Dame, IN 46556-5646 574.631.3338 voice [email protected]
March 27, 2006 ABSTRACT Information asymmetries between corporate insiders and outside investors may make external equity a prohibitively expensive source of financing. In this paper, we use location as a proxy for information asymmetries. Numerous studies show that investors are better able to obtain information on nearby companies. We posit that information asymmetries will be higher for rural firms, with few nearby investors, than for urban firms, with many nearby investors. As predicted, we find that rural firms wait longer to go public, are less likely to conduct seasoned equity offerings, and have more debt in their capital structure than otherwise similar urban firms.
*We would like to thank Robert Battalio, Jeffrey Bergstrand, Roger Huang, Jay Ritter, and seminar participants at the University of Notre Dame for helpful comments. We are grateful to Sebastian Rubano for research assistance.
Asymmetric Information, Firm Location, and Equity Issuance
I. Introduction
In an influential paper, Myers and Majluf (1984) observed that information
asymmetries between managers and outside investors could make it expensive to raise
funds through equity offerings and may lead some financially constrained firms to forgo
valuable projects rather than sell stock. Myers (1984) takes this observation further and
develops a pecking order theory of capital structure. In this theory, firms issue equity
only as a last resort, and capital structure is determined in large part by the firms’ ability
to finance internally.
There have been many studies of the pecking order theory (e.g., Fama and French
(2002)), and evidence on its validity is mixed. There have been far fewer tests though on
the fundamental insight underlying the pecking order theory, that information
asymmetries may affect equity issuance. Researchers have shown that firms time equity
offerings for periods when information asymmetries are small. They have also found that
proxies for information asymmetries such as trading volume, residual volatility, and bid-
ask spreads are related to the propensity to issue equity, and the market reaction to new
equity offerings.
The measure of information asymmetry that we use in this paper is the firm’s
location. This is an entirely different approach from other studies. We believe that
location is a variable that is particularly well-suited for seeing how information
asymmetry affects equity issuance. Several studies show that investors earn higher
returns on investments in local companies than on investments in more distant
companies. Put another way, being located far from a company puts an investor at an
1
information disadvantage that is clearly measurable in the bottom line. Other studies
show that security analysts who are located closer to a company produce more accurate
earnings forecasts than analysts who are located at a greater distance (see Malloy (2005)).
Again, greater distance implies a meaningful disadvantage in obtaining information.
We compare equity issuance by firms in urban areas, defined as the ten largest
metropolitan areas of the United States, with equity issuance by rural firms, defined as
firms located at least 100 miles from the center of any metropolitan area of 1,000,000 or
more people. The idea behind this measure is that, by definition, firms that are located in
rural areas have few nearby investors. The marginal buyer of shares in an equity offering
by a rural firm is therefore likely to be located some distance away. The information
asymmetries between this marginal investor and corporate insiders are likely to be large.
On the other hand, an urban firm is likely to have many potential investors nearby. The
marginal buyer of stock in an equity offering is more likely to be located nearby and
hence information asymmetries are likely to be smaller.
We particularly like rural location as a measure of information asymmetry
because it implies not only that there are few investors located nearby, but also that the
distance is economically as well as physically meaningful. Is a firm located in urban Los
Angeles, or one located in rural Bismarck, North Dakota farther from institutional
investors in New York City? Measured in miles, the company in Los Angeles is much
farther away. But, it is also much easier for the institutional investor to reach. There are
numerous direct flights from New York to Los Angeles every day. Getting to Bismarck is
difficult, and once there, the analyst is almost certainly stuck for the night.
2
Indirect evidence on the differences in information asymmetries between urban
and rural firms comes from differences in their underwriting syndicates. Broadly
speaking, reducing information asymmetries is the underwriter’s job when they help to
produce a prospectus and market an offering. If this is difficult or expensive to do, an
underwriter may forego participating in an offering. We find evidence that, all else equal,
fewer investment bankers compete to participate in IPO and SEO syndicates for rural
firms. Furthermore, underwriters who are used by rural firms tend to be less prestigious
as measured by lower Carter-Manaster rankings.
Our main results are consistent with the joint hypotheses that information
asymmetries between rural firms and investors are particularly large, and that firms avoid
issuing equity in the presence of these asymmetries. Rural firms are significantly older
when they finally decide to go public. Seasoned equity offerings are significantly less
common for rural firms, even after adjusting for differences in size, prior stock returns,
book-to-market ratios and other factors. As a result, all else equal, rural firms have
significantly less equity and more debt in their capital structure.
The remainder of the paper is organized as follows. Section II provides a literature
survey. The data used here is described in Section III. In Section IV we provide evidence
that rural firms issue less equity. Section V provides evidence on differences in
underwriting syndicates for urban and rural equity offerings. Section VI provides a
comparison of capital structures for urban and rural firms. We summarize our results and
conclude in the last section.
3
II. Literature Survey
A. Information Asymmetries and Equity Offerings
In their now classic article, Myers and Majluf (1984) considered what happens
when investment opportunities arise and management has information about assets-in-
place that is not available to outsiders. They show that if a firm is constrained to issuing
equity, and if the value of assets in place is higher than the market realizes, the firm may
avoid issuing equity to prevent harming current shareholders. This will in turn make
external equity an expensive form of financing. An announcement of an equity offering
will lead to lower stock prices as the market rationally assumes that the value of the
firm’s assets- in-place is lower than previously thought. This may lead firms with
correctly valued assets-in-place to avoid a project if it must be financed with external
equity.
Myers (1984) builds on these insights to revitalize an old-fashioned theory of
capital structure, the pecking order. In this view, the static costs and benefits of capital
structure composition are unimportant for most firms. What is important is the costs of
financing. Because of information asymmetries, these costs are far higher for equity.
Firms therefore prefer to fund projects internally, with external debt next, and with equity
only as a last resort. A firm’s capital structure is a byproduct of these cost-minimizing
financing decisions rather than an objective.
Korajczyk, Lucas, and McDonald (1991) observe that the degree of information
asymmetry between corporate insiders and outside investors will change over time for
individual firms. This suggests first that firms will time equity offerings for periods when
the asymmetries are small, and second, that the reaction to an equity offering will vary
4
depending on the degree of information asymmetry. They test these propositions on a
sample of 1,247 seasoned equity offerings that took place during 1978–1983. They posit
that information asymmetries will be high before earnings announcements and will
decline with them.
As predicted, Korajczyk, Lucas, and McDonald (1991) find that equity issuance is
far more common in the first half of a quarter after an earnings announcement than in the
second half. Equity offerings are especially scarce as a new earnings announcement
approaches. Furthermore, when compared to earnings announcements following equity
offerings, the earnings announcements before offerings are both more informative, and
more likely to convey good news. Finally, Korajczyk, Lucas, and McDonald find that the
stock price decline at the announcement of an equity offering is increasing in the time
since the last earnings announcement. As a whole, this study demonstrates that managers
are aware of information asymmetries and consider them sufficiently important to affect
the timing of stock sales.
Dierkens (1991) also finds that the reaction to equity offerings varies with the
degree of information asymmetry. She uses four measures of information asymmetry.
The first is the standard deviation of abnormal returns at quarterly earnings
announcements over the previous five years. The second is the variance of market-model
residual returns over the previous year. The third measure of asymmetric information is
based on a count of the number of announcements about the firm that appeared in the
Wall Street Journal over the prior year. Finally, the annual share turnover is used to
measure information asymmetry. Despite a small sample size of only 197 equity
offerings, Dierkens finds that abnormal returns around equity offering announcements are
5
significantly lower for stocks with high variances of return residuals or a small number of
Wall Street Journal announcements.
Our approach is similar to Bharath, Pasquariello, and Wu (2006). Like us, they
look at measures of asymmetric information and their effect on firms’ propensity to issue
equity. In their paper, they calculate four microstructure related measures of asymmetric
information: the effective bid-ask spread of Roll (1984), the inverse of daily turnover,
the relation between daily volume and first-order return autocorrelation suggested by
Llorente et al. (2002), and the PIN measure of informed trading from Easley et al. (1996).
They then standardize each measure and calculate a composite by averaging the four
standardized measures. The individual measures have some problems – for instance the
Roll estimate of the effective spread is negative almost as often as it is positive. And, as
Bharath, Pasquariello, and Wu observe, all of the measures are based on asymmetry
between informed and uninformed traders, not between managers and outside investors.
Nevertheless, tests using their composite information asymmetry measure indicate that
firms are particularly likely to avoid issuing equity when information asymmetries are
high. They conclude that the pecking order theory is a partial explanation for capital
structure decisions.
B. Location and Investors
There is strong, consistent evidence that investors tend to overweight their
portfolios in nearby companies. This implies that companies located in urban areas have
more potential shareholders than firms located in rural areas. It may therefore be easier
6
for urban firms to raise money by selling equity, and they may be more likely to return to
the equity market for additional capital.
Investors’ bias toward nearby companies is documented in several studies.
Huberman (2001) shows that customers of the regional Bell operating companies are
much more likely to buy shares of the telephone company providing their service than
another telephone company. Coval and Moskowitz (1999) examine the distance from
mutual funds’ headquarters to the companies the funds held in their portfolios in 1995.
On average, companies held in a fund’s portfolio were 10% closer to the fund’s
headquarters than the average distance of potential holdings. Individual investors are
even more biased than fund managers toward local companies. Ivkovic and Weisbenner
(2005) examine the stock investments of over 30,000 households in the continental
United States from 1991 to 1996. They find that the average household invests 31% of its
portfolio in stocks located within 250 miles. If investors had held the market portfolio
instead, only 13% of the average household’s investments would be this close.
A possible explanation for investor preference for local stocks is simply
familiarity.1 Barber and Odean (2005) observe that with more than 7,000 U.S. stocks,
investors cannot consider all securities in their investment decisions. They instead choose
among stocks that have captured their attention. Companies that are in the local news,
that employ an investor’s neighbor, or that an investor sees each day on the way to work
are more likely to capture his or her attention.
The other explanation for investing in local stocks is better access to information.
Much of the information that is useful for valuing stocks is informal, soft information. It
1 Massa and Simonov (2005) provide evidence that familiarity is not a behavioral bias, but is information driven. DeMarzo, Kaniel, and Kremers (2002) suggest that familiarity is a result of investors desire to hedge consumption relative to their neighbors.
7
comes from observing that a company is employing extra shifts, or from casual
conversations with the company’s employees or customers. Much of this information is
only available to investors who are physically close to the company. Evidence that
investors have better information on local stocks comes from their investment returns.
C. How Location Affects Investors’ Returns
Several papers suggest that investors earn higher returns on stocks of nearby
companies. Ivkovic and Weisbenner (2004) examine the returns of individual investors at
a large discount brokerage firm. These retail investors earn 3.7% more per year on local
stocks than on their other investments. When S&P 500 stocks are discarded the difference
between returns on local stocks and others is even higher, about 6% annually. Ivkovic
and Weisbenner find that the difference in returns between local stocks and others
appears for investors all over the U.S. and is robust to various risk adjustments.
Investors in other countries also earn higher returns on investments in local
companies. Bodnaruk (2003) examines Swedish investors’ stockholdings and location
every six months during 1995 to 2001. After controlling for various measures of risk, he
estimates that an investor who purchases shares of companies 100 kilometers away earns
between 1.8% and 3.8% less per year than an investor who buys stock in firms only 10
kilometers away.
Mutual funds also appear to earn significantly higher returns on their local-firm
holdings than on their distant-firm holdings. Coval and Moskowitz (2001) separate
mutual fund holdings into local and distant stocks, where local stocks are those with
headquarters within 100 kilometers of the mutual fund. Local stocks that are held by
8
funds earn annual returns that are about 3% higher on average than local stocks that are
not held by funds. Interestingly, all else equal, funds tend to turn over local stocks less
frequently than stocks of distant companies. Locally held firms tend to be small and
highly levered. Coval and Moskowitz suggest that these are the sorts of stocks in which
local investors may have an information advantage.
Other evidence that closeness to a company provides information advantages
comes from work on equity analysts by Malloy (2005). He finds analysts located nearer
a company’s headquarters provide more accurate earnings forecasts. This greater
accuracy is not explained by underwriting relationships. Enhanced accuracy of local
forecasts is particularly strong for firms located in remote areas, for small firms, and for
high book-to-market firms. Stock price responses to analyst rating changes are especially
strong for analysts located near a particular firm.
A major part of the job for underwriters of equity offerings is reducing
information asymmetries between issuers and investors. Corwin and Schultz (2005)
examine underwriting syndicates for 1,638 U.S. initial public offerings (IPOs) issued
between January 1997 and June 2002. They find that an underwriter is more likely to be
included in an IPO syndicate if it is based in the same state as the firm that is going
public. Underwriters who are located in an adjoining state (e.g., California for an Oregon
issue) are less likely to be included in the syndicate than underwriters based in the same
state, but more likely than underwriters based elsewhere. The comparative advantage of
investment bankers in underwriting local companies suggests that it is easier for them to
obtain information.
9
III. Data
We define a firm’s location as the site of their headquarters, which we obtain
from Compustat, Moody’s, and Nasdaq. This definition has obvious limitations. For
example, we define Wal-Mart as a rural firm (headquarters in Bentonville, Arkansas), but
their stores are in many locations. We still expect the site of the headquarters to be a good
measure of firm location, however. Many of the smaller firms in our sample have
facilities in only one place. In addition, even if a firm has numerous plants in various
states, the company is likely to be particularly familiar to investors living near the firm
headquarters. Finally, the location of a firm’s headquarters has proven to be a useful
approximation for the firm’s location in a number of papers, including Bodnaruk (2003),
Coval and Moskowitz (1999, 2001), Ivkovic and Weisbenner (2005), Loughran and
Schultz (2004, 2005), and Seasholes and Zhu (2005).
We classify a firm’s location as urban if its headquarters is in one of the ten
largest metropolitan areas of the United States according to the 2000 census. These
metropolitan areas include New York City, Los Angeles, Chicago, Washington-
Baltimore, San Francisco, Philadelphia, Boston, Detroit, Dallas, and Houston. Companies
are classified as rural if their headquarters is at least 100 miles from any of the 49
metropolitan areas with one million or more people. Rural firms are located in northern
New England and New York State, Appalachia, the Deep South, the plains states, west
Texas, Alaska, Hawaii, and eastern Washington and Oregon. In Figure 1, areas of the
United States that we define as rural are shaded.
Companies located within 100 miles of any of the metropolitan areas of one
million or more but not in one of the top ten cities are omitted. So, firms like Dell
10
Computer (located in Austin, Texas) or Microsoft (based in the metropolitan area of
Seattle, Washington) are omitted. We believe our definitions provide samples of clearly
urban and clearly rural firms. Only U.S. operating firms (as defined by the Center for
Research in Security Prices (CRSP)) are included in the sample.
Table 1 provides summary statistics for our sample of urban and rural stocks.
Following the Fama and French (1992) methodology, we create yearly portfolios for
1980-2002 at the end of June of year t to ensure that the prior year’s accounting
information from Compustat is publicly available. In all of our analysis, only firms with
a stock price more than $10, as of June of year t, are included. For each firm, yearly total
debt is obtained from Compustat and is the sum of long-term debt (data item 9) and short-
term debt (data item 34).
The equity value is the market value (shares outstanding multiplied by stock
price) as of the last trading day in June of year t and is obtained from CRSP. For the
book-to-market ratio, we use the prior fiscal year’s book value (defined as Compustat
book value of equity plus balance-sheet deferred taxes and investment credit minus the
book value of preferred stock) scaled by the end of December year t-1 CRSP market
value of equity.
The first row of Table 1 reveals that the proportion of urban firms issuing
seasoned equity in the subsequent year (July year t to June t+1) is higher than the
proportion of rural firms. On average, 7.5% of the urban firms issue equity in the
subsequent 12-months compared to 5.9% of the rural firms. The table also reports that
23.3% of urban firm’s total capitalization (book value of debt plus market value of
equity) is in debt as compared to 28.4% for rural firms, while rural firms have a
11
correspondingly lower proportion of equity in their capital structure. There are, however,
a number of other differences between urban and rural firms that need to be controlled
before concluding that location affects equity issuance.
One difference is that urban firms are larger. The mean capitalization of the urban
firms in our sample is just over $2.6 billion, while the mean market value of rural firms is
only $1.2 billion, about half as much. Urban firms also have lower book-to-market ratios
(i.e., more tilted towards growth) and are less likely to be listed on Nasdaq. The industry
composition of urban and rural firms is also different. We obtain SIC codes from CRSP
and use them to classify firms into the 48 industry groups defined in Fama and French
(1997).
In Table 1, we list the common industries present for our sample. Industry
composition differs significantly across urban and rural stocks. Only 3.9% of urban firms
are utilities as compared to 12.4% of rural stocks. Utilities, especially in the 1980s, had a
much higher equity issuance rate than non-utilities. Hence, given that rural firms are
much more likely to be regulated utilities, rural firms are still less likely, on average, to
issue seasoned equity. Business services is the industry of 11.6% of urban firms but only
3.0% of rural firms. Finally, 12.9% of rural stocks are banks, but only 6.1% of urban
stocks.
Like other papers (see Malloy (2005)), we find that rural firms are relatively
neglected by analysts. Analyst coverage data are obtained from Institutional Brokers
Estimate System (I/B/E/S). Analyst coverage is defined as the number of analysts
reporting current fiscal year annual earning estimates prior to June of year t. Firms
without any I/B/E/S coverage are assigned a value of zero analysts. The number of
12
analysts covering the average urban firm exceeds the number covering the average rural
firm by more than two. In the prior year (July of year t-1 to June of year t), urban firms
have higher buy-and-hold returns than their rural counterparts. Much of the high prior
stock performance of both samples is due to the $10 price screen at the June formation
date. Thus, firms with a substantial decline in price will be screened out while prior
winners with a stock price greater than $10 are included in the sample. Notice that our
analysis does not have a look-ahead bias. We have no requirement of subsequent stock
price or stock returns to remain in the sample. In the subsequent year, rural firms slightly
outperform urban firms (16.6% versus 16.0%). Counting each firm-year separately, there
are 21,499 urban observations but less than 5,600 rural firm-year observations.
IV. The Difficulties of Raising Equity Capital for Rural Firms
Casual observation suggests that rural firms are less likely than urban firms to go
public. To illustrate this, we counted the number of firms with headquarters in each state
that traded on the New York Stock Exchange (NYSE), American Stock Exchange
(Amex), or Nasdaq market at the end of 2000. We then standardized each count by
dividing by the number of firms with 500 or more employees in that state as reported in
the 2000 census, at www.census.gov/epcd/susb/2000/us/US/--.HTM#table2.
For Massachusetts, a state which is primarily urban, the ratio of public firms to
firms with 500 or more employees is 12.4%. Other states with primarily urban
populations have similar ratios: 13.5% for New York, 21.2% for California, and 8.9% for
New Jersey. For states that are primarily rural the ratios are much lower. The ratio of
public firms to firms with 500 or more employees is 1.7% for Maine, 1.4% for Nebraska,
13
1.1% for South Dakota, and 1.0% for West Virginia. Results are similar for other urban
and rural states. They are also similar when the number of firms with 200 or more
employees is used in the denominator.
The dearth of publicly traded firms from rural areas suggests that rural firms are
less likely to sell stock publicly than urban firms. Nevertheless, these results are far from
conclusive. The census data does not control for differences in characteristics of firms
across states. Also, the firms are required to employ 500 or more in the state, but are not
required to have their headquarters there. For a more rigorous examination of the effect
of location on equity issuance, we examine the characteristics of urban and rural IPOs.
A. Initial Public Offerings
Firms that are located in rural areas, far from underwriters, institutional investors,
and most retail investors are likely to wait longer before going public. We obtain the firm
age, defined as the number of years since the firm’s founding date, for IPOs that took
place over 1980-2002. Our data source for all IPO information is provided by Loughran
and Ritter (2004). Only IPOs with a first CRSP-listed stock price of more than $10 are
included in the sample. To limit the impact of outliers, age values over 80 years are set to
80. On average, rural firms were 20.95 years old at the time of their IPO while urban
firms were only 13.07 years old.
Of course, the longer operating histories of rural firms can reflect differences in
size or industry composition across urban and rural firms. To control for these
differences, we regress firm age at the time of the IPO on a dummy for Nasdaq listing,
dummies for the energy, business services, retail, banking, and utility industries, the log
14
of the firm’s market value, and dummies for rural location, whether the IPO had venture
capital (VC) backing and whether it used top-tier investment bankers. The top-tier
underwriter dummy takes a value of one if the lead underwriter has an updated Carter and
Manaster (1990) rank of 8 or more, and zero otherwise. Results are shown in Table 2.
For comparison, the first regression includes only an intercept term and a dummy
variable for a rural location. The coefficient on the rural dummy is 7.88, with a White
(1980) heteroskedasticity-adjusted t-statistic of 5.62.2 Rural firms are significantly older
when they go public.
The second regression includes the other explanatory variables. Firms which list
on Nasdaq after their IPO are 11.03 years younger than otherwise similar firms that
initially listed on the NYSE or Amex. The market value coefficient is insignificant when
the Nasdaq dummy is included. VC-backed IPOs go public almost 7 years earlier than
non-VC IPOs. Industry is also related to firm age. Retail companies are about seven years
older than average at the time of their IPO. Business services companies are 3.29 years
younger. For our purposes, the most interesting result is that the coefficient on the rural
dummy is 5.74, with a t-statistic of 4.12. Even after adjusting for all these other factors,
rural firms wait almost six years longer to go public.
We provide three different robustness checks. First, the sample time period is
divided in half. Column 3 reports the IPO regression results during 1980-1991 while
column 4 reports the results during the later time period, 1992-2002. In each subperiod,
the rural dummy is statistically significant with a coefficient of approximately 5.5.
2 The results appear to be robust to alternative specifications of the dependent variable. For example, if the dependent variable is the natural log of (age+1), the rural dummy coefficient remains positive and highly significant.
15
Second, we rerun the regression after discarding 552 IPOs of firms based in the
San Francisco Bay Area. This region spawned a large number of technology companies
in the late 1990s. Some of these urban companies were quite young at the time of their
IPO, and may have an inordinate amount of influence on the age results. The fifth
regression shows though that after discarding these San Francisco-based IPOs, rural firms
are still a highly significant 5.17 years older than similar urban firms when they go
public.
Finally, in the last column of Table 2, we report regression results when urban is
defined to include only stocks from the five largest metropolitan areas: New York, Los
Angeles, Chicago, Washington-Baltimore, and San Francisco. The definition of urban
that we use throughout the paper, the ten largest metropolitan areas in the U.S., is a
subjective one. We have tried alternative definitions though and they have had no
qualitative impact on the results. In this case, when we use the five largest metropolitan
areas as our definition of urban, we see that rural stocks are five years older than urban
stocks at the time of their IPO. Despite the decline in the number of observations, the
coefficient on the rural location remains highly significant with a t-statistic of 3.41.
B. Seasoned Equity Offerings
To see how location affects the likelihood of a seasoned equity offering (SEO),
(also called a follow-on offering) we perform logistic regressions with a dummy for
equity issuance during the year as the dependent variable. All firms that issued seasoned
equity in the 12 months after June of year t are assigned a value of one, else zero. The
explanatory variables are market value, book-to-market, analyst coverage, prior stock
16
performance, and dummies for rural location, Nasdaq listing, industry, and calendar year.
These regressions are reported in Table 3. The z-statistics are calculated conservatively
using standard errors clustered at the individual firms.
Not surprisingly, book-to-market, analyst coverage, and prior stock returns are all
strong predictors of whether or not the firm issues equity in the subsequent year. When
the stock price is high relative to the book value (e.g., growth companies), firms are more
likely to issue equity. If the stock price is low relative to the book value (e.g., value
firms), they are less likely to issue stock. It is not clear if the likelihood of an SEO
increases if the firm receives more attention from analysts, or if analysts are more likely
to cover firms that may issue equity in the future. Nevertheless, a larger number of
analysts are associated with a greater likelihood of an SEO. We also find that the
likelihood of an equity offering differs by industry. Utilities and energy firms are more
likely to issue equity than other firms. Banks are much less likely to conduct SEOs.
Finally, as Korajczyk, Lucas, and MacDonald (1990) show, returns over the prior
year is a highly significant determinant of the likelihood of equity issuance. This is also
consistent with the CFO survey results of Graham and Harvey (2001) which lists recent
stock price performance as the third most important factor in determining firms’ equity
issuance decisions. Firms are much more likely to sell stock to the public following a
runup in price than after a stock price decline.
Even after adjustment for these factors, rural firms are less likely to issue equity.
The coefficient on the rural dummy in the logistic regression is -0.27 with a
heteroskedasticity adjusted z-statistic of -3.27. Column 2 of Table 3 reports the marginal
effects for each of the variables in the logit regression. The coefficient of -0.015 on the
17
rural dummy implies that rural firms are 1.5% less likely to issue seasoned equity after
controlling for other factors. Across our entire sample, firms have a 7.2% chance issuing
equity in a given year. Hence a difference of 1.5% between urban and rural firms is a
relatively large difference. The 0.084 coefficient on the utility dummy means they are
over 8% more likely to have a subsequent year SEO.3
As a robustness check, column 3 reruns the regression after removing utilities.
When utilities are omitted (column 3), the coefficient on the rural dummy is -0.22 with a
statistically significant z-statistic. The fourth column reports the marginal effects of the
variables for the logit regression without utilities. All else equal, the probability that a
rural firm will issue an SEO in a given year is 1.2% less than the probability of an SEO
by an urban firm. The fifth column reports that when only utilities are included in the
logit regression, the only variable that is significant is the rural dummy. Rural utilities are
less likely to issue seasoned equity than urban utilities. This may be the cleanest test of
the influence of location on equity issuance. It is difficult to think of any significant way
in which the urban and rural utilities differ, but the rural utilities are less likely to seek
external equity financing.
The utilities-only logit regression is interesting for another reason as well. An
assumption underlying our work in this section is that firm location is exogenous. That is,
differences in the need to issue equity do not determine whether firms locate in urban or
rural areas. We argue that this is true generally. The proximity of resources, land, and
customers, and the costs of moving employees to a new location are the primary
determinants of firm location. It is particularly easy to make this case for utilities. A
3 These results are again robust to our definition of urban. For example, when urban is defined as located in the metropolitan areas of New York, Los Angeles, Chicago, Washington-Baltimore, or San Francisco, the coefficient on the rural dummy remains at -0.201 with z-statistic of -2.27.
18
utility serves a particular geographic area. It cannot economically locate its headquarters
away from its customer base and power generation. Hence it is implausible that results
for utilities are muddled by endogeneity. And, as Table 3 shows, rural utilities are less
likely to conduct SEOs than urban utilities.
V. Firm Location and Access to Investment Bankers
We have shown that rural firms wait longer to go public and are less likely to
conduct SEOs than urban firms. We feel the likely explanation for this is that information
asymmetries between managers and potential investors are greater for rural firms.
Investors who are located far from rural firms do not get access to the soft, informal
information that they obtain about nearby firms. In this section, we look at underwriter
participation in equity issues of urban and rural firms. Access to information is critical if
underwriters are to do their jobs. If it is difficult for distant underwriters to obtain
information on rural firms, we would expect them to be more reluctant to participate in
rural firm offerings. Consistent with this, Loughran and Schultz (2005) and Malloy
(2005) report that investment bankers provide less analyst coverage for rural firms than
similar urban companies.
A. Location and Investment Banker Quality
We gather information on syndicates for all IPOs and SEOs conducted by urban
or rural firms with offering prices over $10 during 1980–2002. We first examine whether
rural firms are constrained to use less prestigious underwriters. Identities of underwriting
syndicate members are obtained from SDC. We obtain Carter-Manaster rankings from
19
Jay Ritter’s website and use them to rank lead underwriters for urban and rural firm
equity offerings. Carter-Manaster rankings of underwriters are based on where
underwriters’ names appear on IPO prospectuses. The most prestigious underwriters, like
Morgan Stanley and Goldman Sachs, never appear below others on prospectuses and are
given a rank of nine. Underwriters ranked eight appear below the underwriters ranked
nine and ahead of others, and so forth. The lower the Carter-Manaster ranking, the less
prestigious is the underwriter.
The distributions of the ranks for urban and rural firms are reported in Table 4.
Panel A shows results for IPOs. Investment bankers with a Carter-Manaster rank of nine
underwrite 49.1% of urban firm IPOs but only 35.3% of rural IPOs. On the other hand,
39.2% of rural firm IPOs are underwritten by investment bankers with ratings of seven or
below while only 20.9% of urban firm IPOs are underwritten by these investment
bankers.
SEO results in Panel B are similar. A larger proportion of urban firm SEOs than
rural firm SEOs are underwritten by investment bankers with rankings of eight or nine.
And, while only 10.4% of urban firm SEO underwriters have rankings of six or less,
21.1% of rural firm SEOs have underwriters with these ranks.
Of course, there are significant differences between urban and rural firms in terms
of market values and industry composition. To control for these differences, we estimate
logistic regressions with a dummy variable for a Carter-Manaster rank greater than or
equal to eight for the lead underwriter as a dependent variable. If the offering is
underwritten by co-lead underwriters we use the highest ranking to determine the
dependent variable. This measure of underwriter prestige is regressed on a dummy
20
variable for a rural location, the log of the firm’s capitalization (in millions of dollars),
the log of the offering proceeds (in millions of dollars), and industry dummies. Results
are shown in Table 5.
Panel A reports results for IPOs. The first regression uses only the capitalization,
offering size, and rural location variables. Firm size is an important determinant of
underwriter prestige. In the first logistic regression, the coefficient on log of firm size is
0.46 with a heteroskedasticity-consistent z-statistic of 6.68. Likewise, the proceeds from
the offering is highly significant with a coefficient of 0.75 and a z-statistic of 8.67.
Finally, the coefficient on the rural dummy is -0.63 with a z-statistic of -4.20. Rural firms
are significantly less likely to retain a prestigious underwriter for their IPOs. The next
row in the table calculates marginal affects at the mean firm size and offering proceeds.
The marginal effect of -0.09 for the rural dummy indicates that rural firms of average size
that are conducting an average size offering are 9% less likely to have a prestigious
underwriter.
The second logistic regression reported in Table 5 includes five industry
dummies. Both firms in the service industry and retail firms are more likely to use a
prestigious underwriter for their IPO. Inclusion of these industry dummies has little effect
on the significance of the rural location dummy. A rural firm is still 9% less likely to
have a prestigious underwriter than an otherwise similar urban firm.
The next logistic regression includes dummies for the calendar year of the
offering. They are included because the investment banking industry consolidated over
the sample period, with fewer and fewer low ranking underwriters surviving. After
including the year dummies, a rural firm is 7% less likely to use a prestigious underwriter
21
than an urban firm all else equal. For the final logistic regressions, we restrict the sample
to IPOs by firms with capitalization between the 25th and 75th percentiles of all firms
conducting IPOs with prices of $10 or more to make sure that our results are not driven
by unusually large or small firms. Results remain strong. All else equal, a rural firm is
now 7% less likely to have a prestigious underwriter. The coefficient on the rural location
dummy remains significant at the 5% level even after omitting half of the observations.
Panel B repeats the analysis with SEOs. Some firms in our sample have more than
one SEO, hence we conservatively assume that errors are clustered by issuer. Results are
qualitatively similar to those found with IPOs, with the exception that offering proceeds
is no longer a significant determinant of the likelihood of a prestigious underwriter, while
the firm’s capitalization is even more significant. Rural firms are significantly less likely
to use a prestigious underwriter in seasoned equity offerings. When all offerings are
included and variables for industry and year are used, an otherwise average rural firm is
10% less likely to use a prestigious underwriter. When attention is restricted to SEOs of
firms with capitalizations in the 25th to 75th percentiles, rural firms are still 10% less
likely to use a prestigious underwriter.
It appears then, that rural firms cannot get the best underwriters to handle their
equity offerings. They are stuck with second tier investment bankers. This is consistent
with prestigious underwriters being reluctant to bear the extra costs of gathering
information on these firms, and being unwilling to jeopardize their reputations by
underwriting offerings when they have incomplete information on a company.
22
B. Location and Syndicate Size
The number of co-managers in a syndicate is a measure of the competition to
underwrite and offering. Corwin and Schultz (2005) observe that lead underwriters
almost never voluntarily add co-managers to an underwriting syndicate. Instead, issuers
choose co-managers from the investment bankers that competed to lead the offering.
Hence a large number of co-managers and joint lead underwriters implies that a number
of investment banks wanted to be the lead underwriter for the offering.
In Table 6 we provide the distribution of syndicate size for urban and rural equity
offerings. Syndicate size is defined as the number of lead or joint lead underwriters plus
the number of co-managers. Panel A of Table 6 reveals that a larger portion of urban firm
IPOs than rural IPOs have large syndicates. Syndicates with four or more members
underwrite 17.5% of urban firm IPOs but only 6.2% of rural firm IPOs. On the other
hand, 72.4% of rural firm IPO syndicates have only one or two investment bankers, but
only 52.3% of urban firm IPO syndicates are that small. Panel B reports syndicate sizes
for SEOs. Results are similar.
The differences in syndicate size could simply reflect differences in
characteristics of urban and rural firms. To control for these differences, we conduct
Poisson regressions of syndicate size on the natural log of the firm capitalization, the
natural log of the SEO proceeds, and industry dummies. Table 7 reports results.
Panel A presents Poisson regression estimates for IPOs. In the first regression, we
omit the industry dummies. Syndicate size increases with both firm size and proceeds of
the offering. Both variables are highly significant. The rural dummy is negative and
significant. The marginal effects are shown in the next row. A firm of average size that is
23
raising the average amount in their IPO will have 0.25 fewer syndicate members if it is
located in a rural area. In the next two Poisson regressions, industry dummies and year
dummies are included. Rural firms continue to have smaller syndicates, although the
heteroskedasticity-consistent z-statistic falls to -1.84 when yearly dummies are included.
When the industry and year dummies are in the Poisson regression, rural firms are found
to have 0.09 fewer syndicate members.
The last Poisson regression uses only IPOs by firms with capitalizations within
the interquartile range. The coefficient on the rural dummy remains negative, with a z-
statistic of -1.97. For firms with sizes in the interquartile range, a rural location implies
0.13 fewer syndicate members.
Panel B reports estimates of Poisson regressions with the syndicate size for SEOs
as the dependent variable. We again assume errors are clustered at the firm level. The
rural dummy is negative in all regressions, indicating that rural firms have smaller SEO
syndicates. The z-statistics range from -1.72 to -2.18 across the regressions. As a whole,
it appears that for SEOs, like IPOs, a rural location means a smaller syndicate.
To summarize, the Poisson regressions indicate that there is less competition to
underwrite rural firm equity offerings. While not every investment banker who competes
to be lead underwriter is included in a syndicate, almost every co-manager, co-lead, or
lead underwriter in a syndicate competed to be the lead. The fact that syndicate sizes are
smaller for rural firms suggests fewer underwriters were willing to bear the extra costs to
obtain the needed information on these rural firms.
24
VI. Differences in Capital Structure for Urban and Rural Firms
In the previous sections, we show that rural firms issue seasoned equity less
frequently than urban firms. Does a firm’s location affect its capital structure?
Information asymmetries that arise for rural firms may hinder their equity issuing ability
and may lead the companies to have more debt in their capital structure.
In Table 8, we report time-series average parameter values from annual cross-
sectional regressions of the proportion of capital structure represented by the book value
of debt on a dummy for a rural location and several other variables. The dependent
variable consists of observations for each firm each year over 1980-2002. All firms are
included in the first regression. The average coefficient on the rural dummy is 1.36 from
the 23 annual regressions, indicating that, all else equal, debt comprises an additional
1.36% of the total value of rural firms. The t-statistic is 3.66, allowing rejection of a null
hypothesis that the coefficient is zero at any conventional significance level. Nasdaq
listing is also important in determining capital structure. Debt is 6.0% less of the capital
structure for Nasdaq stocks. Not surprisingly, value firms (high book-to-market ratio)
have significantly more debt in their capital structure. Analyst coverage has a negative
effect on the debt proportion. As expected, the higher a firm’s prior year buy-and-hold
stock return, the lower is its debt proportion. Some of the industry dummies are also
highly significant. The proportion of the capital structure in debt is 6.6% less for Business
Services firms, 16.2% more for utilities, and 24.8% more for banks.
Because capital structure is so different for banks and utilities than for other firms,
we rerun the annual cross-sectional regressions after excluding them. Column 2 reports
the regression results after excluding all banks. The average R2 of the 23 annual
25
regressions declines slightly from 30.0% to 25.8%, but coefficients on variables are
qualitatively unaffected. In particular, the average parameter value on the rural dummy
increases to 2.35, with a highly significant t-statistic of 5.92. In column 3 of Table 8, we
exclude both banks and utilities with limited impact on the coefficients. In this column,
the rural dummy remains a highly significant coefficient of 2.35.
In the last column of the table, we include only utility firms in the annual
regressions. As was the case in the other regressions, the average coefficient on the rural
dummy for the utility only sample is positive and statistically significant. We interpret
the results of the table to mean that difficulties in raising capital through equity offerings
result in rural firms having more debt in their capital structure.
VII. Summary and Conclusions
In recent years, there has been a growing realization on the part of financial
economists that information asymmetries may play an important role in determining
capital structure. When outside investors are at a significant information disadvantage to
insiders, selling equity may be very difficult. As a result, firms with the largest
information asymmetries will have more debt and less equity in their capital structure.
In this paper, we examine equity issuance using location as a proxy for
information asymmetries. Recent papers show that both individual investors and
institutions overweight their portfolio with local stocks. Easier access to information is
thought to be one reason for the bias toward local companies. This is supported by the
findings of some papers that investors earn higher returns on investments in local
companies, and that analysts produce more accurate earnings forecasts for nearby firms.
26
Rural companies have few potential purchasers of stock located nearby, and
therefore the marginal investor for an equity offering by a rural firm is likely to be
located far away. This puts the equity investor in a rural offering at a more significant
information disadvantage to insiders than an equity investor in an urban company’s
offering. We expect then, that rural companies will be more reluctant to issue equity and
will have more debt in their capital structure.
That is what we find. Firms from rural areas wait longer to go public. They are
less likely to conduct a follow on offering, even after adjusting for firm size, prior stock
returns, book-to-market ratios, and other factors. When rural firms do issue equity, there
is less competition to underwrite the offering and the firm is often stuck with a less
prestigious underwriter. All else equal, rural firms have more debt in their capital
structure and less equity.
So, we conclude that geographic location is closely related to information
asymmetries and that information asymmetries between corporate insiders and outside
investors seem to be important determinants of equity issuance. It seems likely that if
anything, we have understated these effects. Our data do not permit us to determine if
rural firms are more likely to remain private than similar urban firms. Likewise, we
cannot determine if rural firms are more likely to forego projects because of the difficulty
of obtaining equity financing. These are interesting questions for future research.
27
References
Barber, B., Odean, T., 2005, “All the glitters: The effect of attention on the buying behavior of individual and institutional investors,” Working paper, UC-Davis. Bharath, Sreedhar, Paolo Pasquariello, and Guojun Wu, 2006, Does asymmetric information drive capital structure decisions?, Working paper, University of Michigan. Bodnaruk, A., 2003, “Proximity always matters: Evidence from Swedish data,” Unpublished working paper, Stockholm School of Economics. Carter, R.B., Manaster, S., 1990, “Initial public offerings and underwriter reputation,” Journal of Finance 45, 1045-1068. Corwin, S., Schultz, P., 2005, “The role of IPO underwriting syndicates: Pricing, information production, and underwriter competition,” Journal of Finance 60, 443-486. Coval, J., Moskowitz, T., 1999, “Home bias at home: Local equity preference in domestic portfolios,” Journal of Finance 54, 2045-2073. Coval, J., Moskowitz, T., 2001, “The geography of investment: Informed trading and asset prices,” Journal of Political Economy 109, 811-841. DeMarzo, P., Kaniel, R., Kremers, I., 2002, “Diversification as a public good: Community effects in portfolio choice,” Working paper, Stanford University. Dierkens, N., 1991, “Information asymmetry and equity issues,” Journal of Financial and Quantitative Analysis 26, 181-199. Easley, D., N. Kiefer, M. O’Hara, and J. Paperman, 1996, Liquidity, information, and infrequently traded stocks, Journal of Finance 51, 1405-1436. Fama, E., French, K., 1992, “The cross-section of expected stock returns,” Journal of Finance 47, 427-465. Fama, E., French, K., 1997, “Industry costs of equity,” Journal of Financial Economics 43, 153-193. Fama, E., French, K., 2002, “Testing trade-off and pecking order predictions about dividends and debt,” Review of Financial Studies 15, 1-33. Graham, J., Harvey, C., 2001, “The theory and practice of corporate finance: evidence from the field,” Journal of Financial Economics 60, 187-243. Huberman, G., 2001, “Familiarity breeds investment,” Review of Financial Studies 14, 659-680.
28
Ivkovic, Z., Weisbenner, S., 2005, “Local does as local is: Information content of the geography of individual investors’ common stock investments,” Journal of Finance 60, 267-306. Korajczyk, R., Lucas, D., McDonald, R., 1990, “Understanding stock price behavior around the time of equity issues,” in R. Glenn Hubbard, Ed.: Asymmetric Information, Corporate Finance, and Investment (University of Chicago Press, Chicago). Korajczyk, R., Lucas, D., McDonald, R., 1991, The effect of information releases on the pricing and timing of equity issues, Review of Financial Studies 4, 685-708. Llorente, G., R. Michaely, G, Saar, and J. Wang, 2002, Dynamic volume-return relation of individual stocks, Review of Financial Studies 15, 1005-1047. Loughran, T., Ritter, J., 2004, “Why has IPO underpricing changed over time?,” Financial Management 33, 5-37. Loughran, T., Schultz, P., 2004, “Weather, stock returns, and the impact of localized trading behavior,” Journal of Financial and Quantitative Analysis 39, 343-364. Loughran, T., Schultz, P., 2005, “Liquidity: Urban versus rural firms,” Journal of Financial Economics 78, 341-374. Malloy, C., 2005, “The geography of equity analysis,” Journal of Finance 60, 719-755. Massa, M., Simonov, A., 2005, “Hedging, familiarity and portfolio choice,” Review of Financial Studies, forthcoming. Myers, S., 1984, The capital structure puzzle, Journal of Finance 39, 575-592. Myers, S., N. Majluf, 1984, Corporate financing and investment decisions when firms have information that investors do not have, Journal of Financial Economics 13, 187-221. Roll, R., 1984, A simple implicit measure of the effective bid-ask spread in an efficient market, Journal of Finance 39, 1127-1139. Seasholes, M., Zhu, N., 2005, “Is there information in the local portfolio choices of individuals?,” Working paper, U.C. Berkeley. White, H., 1980, “A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity,” Econometrica 48, 817-838.
29
Table 1 Mean Summary Statistics for Urban and Rural Firms, 1980-2002
Urban
Firms (1)
Rural Firms
(2) % with subsequent year SEO 7.5% 5.9%
Debt Proportion 23.3% 28.4%
Equity Proportion 76.7% 71.6%
Market Value (in millions $) $2,632.1 $1,192.3
Book-to-Market Ratio 0.64 0.76
% on Nasdaq 41.4% 57.1%
% in Energy 5.0% 4.7%
% in Utility 3.9% 12.4%
% in Business Services 11.6% 3.0%
% in Retail 4.5% 5.1%
% in Banking 6.1% 12.9%
Number of Analysts 7.6 5.4
Prior Year Return, % 35.4% 28.0%
Post Year Return, % 16.0% 16.6%
Number of yearly firm observations 21,499 5,584
Each June of year t, urban and rural portfolios are formed. Only stocks with a stock price of more than $10 as of June of year t are included. A stock is located in an urban area if the company headquarters is in the metropolitan area of New York City, Los Angeles, Chicago, Washington, San Francisco, Philadelphia, Boston, Detroit, Dallas, or Houston. A stock is located in a rural area if it is not within 100 miles of the center of a metropolitan area of one million or more people as defined by the 2000 census. Market values are in millions of dollars as of June of year t. Debt holdings, as of the prior fiscal year, are from Compustat (items 9 (long-term debt) plus 34 (debt in current liabilities)). Total capital structure is book value of debt plus market value of equity. Industry classifications are defined in Fama-French (1997). Analyst coverage information is from I/B/E/S.
30
Table 2 Regression of Firm Age at the IPO on Various Explanatory Variables, 1980-2002
All IPOs (1)
All IPOs (2)
IPOs 1980-1991
(3)
IPOs 1992-2002
(4)
Excluding San
Francisco IPOs (5)
Only Five
Urban Cities
(6) Intercept 13.06
(33.38) 22.85
(10.39) 13.05 (2.34)
23.31 (8.65)
18.58 (6.60)
27.47 (10.15)
Rural Dummy 7.88 (5.62)
5.74 (4.12)
5.33 (2.13)
5.99 (3.55)
5.17 (3.66)
5.04 (3.41)
Nasdaq Dummy -11.03 (-7.68)
-7.99 (-2.78)
-11.80 (-7.03)
-9.76 (-6.46)
-12.50 (-6.76)
Log(Market Value)
0.31 (0.85)
2.21 (2.07)
0.25 (0.57)
1.06 (2.07)
-0.23 (-0.56)
VC Dummy -6.95 (-9.08)
-8.34 (-5.25)
-6.32 (-7.17)
-6.65 (-7.17)
-7.04 (-7.50)
Top-Tier Banker Dummy
2.39 (2.86)
3.00 (2.11)
1.96 (1.88)
2.71 (2.63)
2.66 (2.56)
Energy Dummy -5.36 (-2.23)
-2.95 (-0.73)
-7.61 (-2.66)
-6.08 (-2.45)
-6.99 (-1.92)
Business Services Dummy
-3.29 (-5.28)
-2.73 (-2.12)
-3.08 (-4.34)
-4.41 (-5.09)
-3.86 (-5.36)
Retail Dummy 7.27 (4.07)
4.95 (1.75)
8.35 (3.62)
7.74 (3.82)
7.40 (3.32)
Banking Dummy -2.03 (-0.55)
7.76 (0.51)
-2.50 (-0.68)
-0.78 (-0.19)
-2.28 (-0.49)
Utility Dummy -5.65 (-1.29)
-4.36 (-0.86)
-6.42 (-1.10)
-5.92 (-1.30)
-10.56 (-2.64)
R2 0.022 0.172 0.149 0.186 0.146 0.186 Observations 2,145 2,145 597 1,548 1,593 1,568 The dependent variable, IPO Age, is the number of calendar years since the firm’s founding date as of the IPO. Age values greater than 80 years are set to 80. Only stocks with a stock price of more than $10 are included. The rural dummy is set to one if the firm is not within 100 miles of the center of a metropolitan area of one million or more people as defined by the 2000 census. Nasdaq is a dummy variable equal to one if the firm is listed on Nasdaq, zero if the issuing firm is listed on NYSE or Amex. The market value (in millions of dollars) is the first CRSP-listed market value after the IPO. Industry dummies for Energy, Business Services, Retail, Banking, and Utilities are equal to one if the firm operates in the respective industry. White’s heteroskedasticity-adjusted t-statistics are in parentheses. The average first-day returns are 13.29% for rural IPOs compared to 31.66% for urban IPOs. The average rural IPO firm age is 20.95 years compared to 13.07 years for urban IPOs. Urban firms are based in one of the ten largest metropolitan areas of the United States. For regression (6), urban cities are defined as New York, Chicago, Los Angeles, Boston, and San Francisco.
IPO Ageij = a0j + a1Rural Dummy ij + a2jNasdaq Dummy ij + a3j Log(Market Value) ij + a4jVC Dummyij + a5jTop-Tier Bankerij + a6jEnergy Dummyij + a7jBusiness Services Dummy ij + a8Retail Dummy ij
+ a9jBank Dummy ij + a10jUtilities Dummy ij + eij.
31
Table 3 Logit Regression of the Probability of Issuing Seasoned Equity in the Subsequent Year
All Firms (1)
Marginal Effects
from Logit (2)
All firms excluding utilities
(3)
Marginal Effects
from Logit (4)
Only
Utilities (5)
Marginal Effects
from Logit (6)
Intercept -2.52 (-13.53)
-2.66 (-13.18)
-0.66 (-0.77)
Rural Dummy -0.27 (-3.27)
-0.015 (-3.49)
-0.22 (-2.54)
-0.012 (-2.69)
-0.50 (-2.07)
-0.049 (-2.06)
Nasdaq Dummy 0.10 (1.53)
0.006 (1.52)
0.12 (1.69)
0.007 (1.68)
-0.39 (-0.99)
-0.036 (-1.11)
Log(Market Value) 0.01 (0.43)
0.001 (0.43)
0.00 (0.02)
0.000 (0.02)
0.13 (0.99)
0.013 (0.97)
Book-to-Market -0.18 (-2.56)
-0.011 (-2.57)
-0.19 (-2.46)
-0.011 (-2.48)
-0.34 (-1.16)
-0.034 (-1.15)
Log(1+Analysts) 0.13 (2.86)
0.008 (2.87)
0.12 (2.67)
0.007 (2.68)
0.05 (0.26)
0.005 (0.26)
Prior Return 0.29 (8.90)
0.017 (8.77)
0.29 (8.74)
0.017 (8.61)
0.14 (0.58)
0.014 (0.57)
Energy Dummy 0.32 (2.46)
0.021 (2.18)
0.34 (2.68)
0.023 (2.36)
Business Services Dummy
0.01 (0.10)
0.001 (0.10)
-0.00 (-0.00)
-0.000 (-0.00)
Retail Dummy 0.20 (1.67)
0.013 (1.55)
0.20 (1.67)
0.012 (1.56)
Banking Dummy -0.46 (-2.80)
-0.023 (-3.36)
-0.47 (-2.86)
-0.023 (-3.47)
Utility Dummy 0.96 (7.91)
0.084 (5.80)
Yearly Dummies Yes Yes Yes Yes Yes Yes Observations 27,083 25,544 1,539
Each June of year t, urban and rural portfolios are formed. The dependent variable, Subsequent Year Equity Issuance Dummy, has a value of one if the firm issued equity in the subsequent year; it is zero otherwise. Only stocks with a stock price of more than $10 as of June of year t are included. The rural dummy is set to one if the firm is more than 100 miles from the center of any metropolitan area of one million or more people as defined by the 2000 census. Nasdaq is a dummy variable equal to one if the firm is listed on Nasdaq, zero if the issuing firm is listed on NYSE or Amex. Industry dummies for Energy, Business Services, Retail, Banking, and Utilities are equal to one if the firm operates in the respective industry. Yearly Dummies, except for 1980, are included in the regressions. White’s heteroskedasticity-adjusted z-statistics are in parentheses. Standard errors are clustered for individual firms. Subsequent Year Equity Issuance Dummy ij = a0j + a1Rural Dummy ij + a2jNasdaq Dummy ij + a3j Log(Market Value) ij + a4jBook-to-Marketij + a5jLog(1+Analysts)ij + a6jPrior Returnij + a7jEnergy Dummyij + a8jBusiness Services Dummy ij +
a9Retail Dummy ij + a10jBank Dummy ij + a11jUtilities Dummy ij + Yearly Dummies ij + eij.
32
Table 4 Carter-Manaster Ranks for Lead Underwriters of Urban and Rural Equity Offerings
Panel A: IPOs
Carter-Manaster
Rank
Number Urban Firm
IPOs
Number Rural Firm
IPOs
Percent of Urban Firm
IPOs
Percent of Rural Firm
IPOs 9 786 97 49.1% 35.3% 8 481 70 30.0% 25.5% 7 155 57 9.7% 20.7% 6 76 21 4.7% 7.6% 5 65 20 4.1% 7.3%
<5 38 10 2.4% 3.6% All 1,601 275 100.0% 100.0%
Panel B: SEOs
The sample includes all equity offerings of U.S. based firms reported in SDC for 1980-2002. Carter-Manaster ranks are obtained from Jay Ritter’s website and are based on the relative position of firms on IPO prospectuses. The highest possible Carter-Manaster rank is 9. Urban firms have headquarters in one of the ten largest metropolitan areas of the United States according to the 2000 census. Rural firms have headquarters at least 100 miles from any metropolitan area of 1,000,000 or more. Stocks with prices less than $10 are omitted.
Carter-Manaster
Rank
Number Urban Firm
SEOs
Number Rural Firm
SEOs
Percent of Urban Firm
SEOs
Percent of Rural Firm
SEOs 9 602 135 44.9% 33.1% 8 542 133 35.7% 32.6% 7 136 54 9.0% 13.2% 6 70 51 4.6% 12.5% 5 64 22 4.2% 5.4%
<5 24 13 1.6% 3.2% All 1,518 408 100.0% 100.0%
33
Table 5 Logistic Regressions of Top-Tier Investment Banker Quality Dummy on Firm Location
Panel A: IPOs
Included Observations
Intercept
Rural Dummy
Log (Size)
Log (IPO Proceeds)
Energy Dummy
Services Dummy
Retail Dummy
Banking Dummy
Utility Dummy
Year Dummies
Obs.
All -6.64 -0.63 (-9.56) (-4.20)
0.46 (6.68)
0.75 (8.67)
No 1,876
All (dy/x) -0.09 (-4.04)
0.07 (6.22)
0.11 (8.05)
No
All -6.32 -0.59 (-8.95) (-3.76)
0.42 (6.15)
0.75 (8.65)
0.13 (0.32)
0.37 (2.34)
0.51 (1.99)
-0.39 (-1.74)
0.51 (0.88)
No 1,876
All (dy/dx) -0.09 (-3.69)
0.06 (5.68)
0.11 (7.97)
0.02 (0.32)
0.05 (2.51)
0.07 (2.45)
-0.07 (-1.61)
0.07 (1.01)
No
All -8.84 -0.54 (-8.16) (-3.20)
0.49 (5.95)
0.85 (6.64)
0.41 (0.85)
0.39 (2.38)
0.38 (1.37)
-0.50 (-2.09)
0.27 (0.38)
Yes 1,876
All (dy/dx) -0.07 (-3.21)
0.07 (5.58)
0.12 (7.32)
0.05 (1.04)
0.05 (2.50)
0.05 (1.65)
-0.08 (-1.85)
0.03 (0.42)
Yes
Firm size in 25th to 75th Percentiles
-12.85 (-4.61)
-0.53 (-2.09)
0.97 (4.21)
0.53 (3.05)
0.22 (0.45)
0.49 (2.07)
-0.19 (-0.59)
-0.99 (-2.42)
dropped Yes 926
Firm size in 25th to 75th Percentiles (dy/dx)
-0.07(-2.07)
0.14 (4.33)
0.07 (3.13)
0.03 (0.48)
0.06 (2.23)
-0.03 (-0.56)
-0.18 (-1.97)
dropped Yes
34
35
Panel B: SEOs Included Observations
Intercept
Rural Dummy
Log (Size)
Log (SEO Proceeds)
Energy Dummy
Services Dummy
Retail Dummy
Banking Dummy
Utility Dummy
Year Dummies
Obs.
All -9.92 -0.55 (-10.94) (-3.04)
0.91 (11.97)
-0.01 (-0.20)
No 1,919
All (dy/dx) -0.09 (-3.73)
0.12 (17.84)
-0.00 (-0.33)
No
All -9.91 -0.65 (-10.69) (-3.55)
0.91 (11.64)
-0.01 (-0.10)
-0.12 (-0.34)
-0.46 (-1.93)
0.21 (0.66)
-0.33 (-1.07)
0.76 (1.56)
No 1,919
All (dy/dx) -0.11 (-4.23)
0.12 (17.69)
-0.00 (-0.22)
-0.01 (-0.34)
-0.07 (-2.05)
0.03 (0.90)
-0.04 (-1.04)
0.08 (3.14)
No
All -9.96 -0.70 (-8.09) (-3.78)
0.93 (10.19)
-0.02 (-0.35)
-0.13 (-0.36)
-0.54 (-2.20)
0.23 (0.71)
-0.43 (-1.36)
0.64 (1.29)
Yes 1,919
All (dy/dx) -0.10 (-3.21)
0.12 (15.77)
-0.00 (-0.10)
-0.02 (-0.33)
-0.07 (-1.74)
0.03 (0.71)
-0.05 (-0.98)
0.08 (2.12)
Yes
Firm size in 25th to 75th Percentiles
-12.99 (-5.14)
-0.72 (-2.77)
1.21 (6.55)
-0.18 (-2.21)
-0.20 (-0.42)
-0.65 (-1.88)
0.04 (0.11)
-0.50 (-0.86)
0.98 (1.74)
Yes 944
Firm size in 25th to 75th Percentiles (dy/dx)
-0.10(-2.42)
0.15 (6.83)
-0.02 (-2.18)
-0.03 (-0.40)
-0.10 (-1.59)
0.01 (0.11)
-0.07 (-0.74)
0.09 (2.49)
Yes
The sample includes all equity offerings of U.S. based firms reported in SDC for 1980-2002. Carter-Manaster ranks are obtained from Jay Ritter’s website and are based on the relative position of firms on IPO prospectuses. The highest possible Carter-Manaster rank is 9. Firms with a lead underwriter with an updated Carter-Manaster investment banker of 8 or more are assigned a value of one for the top-tier dummy. Urban firms have headquarters in one of the ten largest metropolitan areas of the United States according to the 2000 census. Rural firms have headquarters at least 100 miles from any metropolitan area of 1,000,000 or more. Size is the market capitalization of the issuing firm (in millions $) on the day of the offering. Stocks with prices less than or equal to $10 are omitted. Robust z-statistics are in parentheses. Marginal effects (dy/dx) of variable changes assume a shift from zero to one for discrete variables and mean values are assigned for other variables. Errors are assumed to be clustered at the firm level for SEOs.
36
Table 6 Syndicate Size for Urban and Rural Equity Offerings
Panel A: IPOs
Number in IPO
Syndicate
Number Urban Firm
IPOs
Number Rural Firm
IPOs
Percent of Urban Firm
IPOs
Percent of Rural Firm
IPOs > 7 11 0 0.7% 0.0% 7 8 1 0.5% 0.4% 6 10 0 0.6% 0.0% 5 57 2 3.6% 0.7% 4 194 14 12.1% 5.1% 3 484 59 30.2% 21.5% 2 603 122 37.7% 44.4% 1 234 77 14.6% 28.0%
All 1,601 275 100.0% 100.0% Panel B: SEOs
Number in SEO
Syndicate
Number Urban Firm
SEOs
Number Rural Firm
SEOs
Percent of Urban Firm
SEOs
Percent of Rural Firm
SEOs > 7 13 2 0.8% 0.5% 7 9 1 0.6% 0.2% 6 27 3 1.7% 0.7% 5 72 9 4.7% 2.2% 4 150 31 9.7% 7.4% 3 366 75 23.0% 18.0% 2 483 146 31.2% 35.0% 1 438 150 28.3% 36.0%
All 1,548 417 100.0% 100.0% The sample includes all equity offerings of U.S. based firms reported in SDC for 1980-2002. The syndicate size is the number of book managers, co-managers and joint book managers in the syndicate. Urban firms have headquarters in one of the ten largest metropolitan areas of the United States according to the 2000 census. Rural firms have headquarters at least 100 miles from any metropolitan area of 1,000,000 or more. Stocks with prices less than $10 are omitted.
Table 7 Poisson Regressions of Syndicate Size on Firm Location
Panel A: IPOsIncluded Observations
Intercept
Rural Dummy
Log (Size)
Log (IPO Proceeds)
Energy Dummy
Services Dummy
Retail Dummy
Banking Dummy
Utility Dummy
Year Dummies
Obs.
All -1.06 -0.10 (-10.28) (-4.46)
0.11 (10.16)
0.17 (9.17)
No 1,876
All (dy/dx) -0.25 (-4.45)
0.27 (10.11)
0.41 (9.20)
No
All -0.99 -0.08 (-9.68) (-3.48)
0.10 (9.31)
0.18 (9.49)
-0.05 (-0.74)
0.07 (4.01)
-0.07 (-2.02)
-0.12 (-3.23)
-0.04 (-0.55)
No 1,876
All (dy/dx) -0.20 (-3.48)
0.25 (9.24)
0.43 (9.55)
-0.11 (-0.76)
0.18 (3.95)
-0.15 (-2.07)
-0.28 (-3.41)
-0.09 (-0.56)
No
All -0.03 -0.04 (-0.25) (-1.84)
0.04 (4.14)
0.18 (9.86)
0.00 (0.00)
-0.02 (-0.91)
-0.02 (-0.59)
-0.10 (-2.57)
0.00 (0.04)
Yes 1,876
All (dy/dx) -0.09 (-1.84)
0.10 (4.13)
0.43 (9.82)
0.00 (0.00)
-0.04 (-0.92)
-0.04 (-0.60)
-0.22 (-2.68)
0.01 (0.04)
Yes
Size in 25th to 75th Percentiles
0.02 (0.07)
-0.06 (-1.97)
0.03 (1.14)
0.16 (8.55)
0.03 (0.64)
0.01 (0.41)
0.01 (0.29)
-0.10 (-1.42)
-0.14 (-1.36)
Yes 939
Size in 25th to 75th Percentiles (dy/dx)
-0.13(-1.97)
0.06 (1.14)
0.38 (8.47)
0.08 (0.63)
0.02 (0.41)
0.02 (0.29)
-0.21 (-1.50)
-0.30 (-1.46)
Yes
37
38
Panel B: SEOsIncluded Observations
Intercept
Rural Dummy
Log (Size)
Log (SEO Proceeds)
Energy Dummy
Services Dummy
Retail Dummy
Banking Dummy
Utility Dummy
Year Dummies
Obs.
All -0.99 -0.07 (-6.53) (-1.74)
0.13 (10.46)
0.06 (6.31)
No 1,918
All (dy/dx) -0.16 (-1.78)
0.29 (9.96)
0.15 (6.38)
No
All -0.96 -0.07 (-6.33) (-1.84)
0.12 (10.14)
0.06 (6.50)
0.20 (2.76)
0.06 (1.41)
-0.07 (-1.08)
0.03 (0.46)
0.10 (1.51)
No 1,918
All (dy/dx) -0.16 (-1.88)
0.28 (9.72)
0.15 (6.57)
0.50 (2.53)
0.14 (1.38)
-0.16 (-1.12)
0.07 (0.45)
0.25 (1.44)
No
All -0.32 -0.08 (-1.96) (-2.18)
0.08 (7.48)
0.00 (0.16)
0.13 (2.07)
0.02 (0.43)
-0.06 (-1.27)
0.09 (1.46)
0.14 (2.02)
Yes 1,918
All (dy/dx) -0.17 (-2.23)
0.19 (7.29)
0.00 (0.16)
0.30 (1.96)
0.04 (0.43)
-0.14 (-1.30)
0.22 (1.40)
0.34 (1.90)
Yes
Size in 25th to 75th Percentiles
-0.83 (-2.40)
-0.06 (-1.72)
0.12 (4.66)
-0.01 (-0.91)
0.12 (2.49)
-0.02 (-0.53)
-0.07 (-1.45)
0.03 (0.28)
0.03 (0.37)
Yes 960
Size in 25th to 75th Percentiles (dy/dx)
-0.14(-1.76)
0.28 (4.66)
-0.02 (-0.91)
0.30 (2.37)
-0.05 (-0.53)
-0.17 (-1.49)
0.06 (0.27)
0.07 (-0.02)
Yes
The sample includes all equity offerings of U.S. based firms reported in SDC for 1980-2003. Urban firms have headquarters in one of the ten largest metropolitan areas of the United States according to the 2000 census. Rural firms have headquarters at least 100 miles from any metropolitan area of 1,000,000 or more. Stocks with prices less than or equal to $10 are omitted. The syndicate size is the number of book managers, co-managers and joint book managers in the syndicate. Size is the market capitalization of the issuing firm (in thousand $) on the day of the offering. Robust z-statistics are in parentheses. Marginal effects (dy/dx) of variable changes assume a shift from zero to one for discrete variables and mean values are assigned for other variables. Errors are assumed to be clustered at the firm level for SEOs.
Table 8 Average Parameter Values from Annual Cross-Sectional Regressions of Debt
Proportion, 1980-2002
Debt Proportion
(1)
Debt
Proportion (excluding
banks) (2)
Debt Proportion (excluding banks and utilities)
(3)
Debt Proportion (including
only utilities)
(4) Intercept 13.23
(11.49) 14.57
(13.29) 14.92
(13.42) 18.61 (4.71)
Rural Dummy 1.36 (3.66)
2.35 (5.92)
2.35 (5.44)
1.67 (2.39)
Nasdaq Dummy -6.01 (-9.49)
-6.39 (-9.17)
-6.42 (-9.06)
-2.90 (-2.68)
Log(Market Value) 0.83 (4.89)
0.79 (5.70)
0.78 (5.76)
1.43 (2.77)
Book-to-Market 12.77 (13.43)
12.60 (12.72)
12.32 (12.31)
22.51 (15.42)
Log(1+Analysts) -0.52 (-3.70)
-1.16 (-9.33)
-1.22 (-8.70)
-0.95 (-1.91)
Prior Return -2.74 (-4.71)
-2.79 (-4.85)
-2.74 (-4.58)
-11.56 (-5.36)
Energy Dummy -0.49 (-1.33)
-0.39 (-1.06)
-0.31 (-0.81)
Business Services Dummy -6.61 (-13.78)
-6.51 (-14.18)
-6.57 (-14.39)
Retail Dummy -2.54 (-6.68)
-2.45 (-6.48)
-2.48 (-6.51)
Utility Dummy 16.18 (21.05)
16.09 (21.10)
Banking Dummy 24.82 (16.35)
Average R2 0.300 0.258 0.202 0.479 Observations 23 23 23 23
Each June of year t, urban and rural portfolios are formed. The dependent variable, Debt Proportion, is the ratio of the book value of debt (long and short term) to the sum of the book value of debt and the market value of equity. The parameter values are the average of the 23 cross-sectional regressions. The rural dummy is set to one if the firm is not within 100 miles of the center of a metropolitan area of one million or more people as defined by the 2000 census. Nasdaq is a dummy variable equal to one if the firm is listed on Nasdaq, zero if the issuing firm is listed on NYSE or Amex. Industry dummies for Energy, Business Services, Retail, Banking, and Utilities are equal to one if the firm operates in the respective industry. The t-statistics are in parentheses.
Debt Proportionij = a0j + a1Rural Dummy ij + a2jNasdaq Dummy ij + a3j Log(Market Value) ij + a4jBook-to-Marketij + a5jLog(1+Analysts)ij + a6jPrior Returnij + a7jEnergy Dummyij + a8jBusiness
Services Dummy ij + a9Retail Dummy ij + a10jBank Dummy ij + a11jUtilities Dummy ij + eij.
39
Figure1. Rural Areas of the United States (shaded).
40