Predators or Watchdogs?
Bankers on Corporate Boards in the Age of Finance Capitalism
Carola Frydman∗
MIT-Sloan and NBER
Eric Hilt†
Wellesley College and NBER
December 2010
Abstract: Using newly collected data on investment banks and NYSE-traded railroads, this paper
analyzes how bank-firm relationships affected corporate governance and firm outcomes in the early
twentieth century. The paper provides a new view into this question by exploiting a regulatory
intervention that attempted to curtail bankers’ presence on boards of directors. Following the Pujo
Committee investigation of the “money trust” in 1912, the Clayton Antitrust Act of 1914 included
provisions intended to weaken the influence of bankers in nonfinancial corporations. In particular,
the Act prohibited investment bankers from sitting on the boards of railroads for which they
underwrote securities. The paper exploits this exogenous source of variation in bankers’ presence
on corporate boards to determine the causal effect of the influence of bankers on early twentieth
century railroads. Our evidence suggests that a relationship with bankers helped the firms ease
financial constraints and increased firm values.
∗
†
Email: frydman@mit.edu.
Email: ehilt@wellesley.edu.
1
Introduction
The financial meltdown of 2008 has put concerns about the role of the financial sector in the economy
at the forefront of academic and public debate. Although the current scale and complexity of
financial institutions is historically unprecedented (Philippon and Reshef, 2009), the debate about
the appropriate role of the financial sector is not. Arguments about the unchecked power and
influence of Wall Street today are reminiscent of the backlash against financiers that followed the
panic of 1907. An examination of the financial system in that era, and the regulatory response
generated by the anti-Wall Street backlash, may produce important insights into the consequences
of financial regulations, the role of financiers in the corporate sector, and the influence that these
interactions have on the organization of American finance and industry.
The power and influence of bankers, and their prominence in industry and in the financial
markets, probably reached its apex during the first two decades of the twentieth century, which
have been termed the era of “finance capitalism.” Compared to other advanced economies at
the time, the financial sector of the United States was organized in a peculiar way, with private
banking partnerships holding a position of preeminence (Carosso and Sylla, 1991). Mostly organized
as partnerships, these investment banks were virtually unregulated and subject to few (if any)
disclosure requirements.1 Their partners often held directorships with the firms they financed, and
participated in their management. They also frequently held seats on the boards of other financial
companies, such as commercial banks, trust companies, and insurance corporations.
Whether or not these banker-directors’ presence was beneficial to the shareholders of the firms
they helped manage, or to the economy as a whole, was a contentious question at the time. Progressive critics of Wall Street argued that bankers utilized their client firms’ resources for their own
benefit, issuing excessive amounts of securities in exchange for exorbitant fees, and raising firms’
indebtedness to unsustainable levels (Brandeis, 1913). But the effects of banker representation on
a firm’s board include many potential benefits, as well as costs. On the one hand, a close bankfirm relationship may help solve agency conflicts, if the bankers act like independent directors,
and facilitate access to capital. The presence of a banker on a firm’s board might also “certify”
1
Some trust companies also engaged in investment banking, and some national banks entered the industry as well,
through securities affiliates (White, 1986; Kroszner and Rajan, 1997). Most investment banks, however, remained
organized as private partnerships until the 1970s, when the NYSE permitted publicly traded corporations to join (see
Morrison and Wilhelm, 2008).
1
that firm to the financial markets. Alternatively, bank control may be detrimental if bankers have
monopolistic power over the firms’ access to capital, or engage in extensive self-dealing. Moreover,
easy access to capital through close relationships with bankers might enable executives to finance
value-destroying projects. This question is difficult to resolve empirically, since bank-firm relationships are endogenously determined, and there is clear evidence in modern data that bankers tend
to be represented on the boards of firms that are in many ways atypical (Kroszner and Strahan,
2001).
This paper analyzes the effect of bankers on corporate boards in the era of finance capitalism.
To establish a causal link between bankers’ presence and firm outcomes, an ideal strategy would
consist of randomly allocating board seats to bankers. We propose to use the Clayton Antitrust
Act of 1914 as a natural experiment that attempted to sever bank-firm relationships. Influenced
by the findings of the Pujo Committee investigation of the “money trust” of 1912, the Clayton
Act restricted bankers’ role on firm boards, and thus created an exogenous source of variation in
the incidence of banker-directors. Because it exploits this exogenous severing of the ties between
bankers and nonfinancial firms, this strategy represents a substantial step forward in establishing
causal implications of the effect of banker-directors on firm outcomes.
Better known for its terms restricting anticompetitive practices, the Clayton Act imposed very
specific regulations on corporate directors in particular industries. For railroad corporations, it
effectively forbade self-dealing, in which a banker-director would underwrite a railroad’s securities:
either the bankers on a road’s board had to resign, or they could remain on the board but cease to
act as the firm’s bankers.2 Implemented in 1921, the terms of the Clayton Act relative to railroads
create a quasi-experiment with which to estimate the effects of bankers on firm boards. The focus
of this paper is therefore on publicly traded railroads, which were among the most important issuers
of securities in the early twentieth century (see Edwards, 1938).
Analyzing the effect of the Clayton Act requires having information on the characteristics of
firms, their boards, and their ties to banks. We use a newly hand-collected dataset on NYSE-listed
railroads from 1902 to 1929. Railroads were amongst the largest corporations at that time, and
2
For industrial corporations, the provisions of the Act were somewhat different. The Act prohibited anyone to
serve on the boards of competing firms. As many powerful bankers sat on the boards of large numbers of firms that
could be considered to be competitors in the language of the Act, these directors would have needed to resign from
many of those firms’ boards but had discretion on which board to step down from.
2
the presence of bankers on their boards was even more prevalent than in other types of firms.3 For
each railroad, we collect detailed accounting data from their income statements, balance sheets,
and capital stock descriptions, as well as a complete description of their board members. The
information on board composition allows us to measure the ties of railroads to banks by identifying
the railroad directors that were members of investment and commercial banks. Thus, an important
contribution of our paper is to present the first-ever comprehensive characterization of the incidence
of banker directors among publicly traded firms, and to construct a consistent dataset of the finances
of large corporations over the first three decades of the twentieth century.
To determine whether the presence of bankers on boards was detrimental or beneficial for
corporations, our strategy is to compare the outcomes for railroads that had bankers on their
boards in 1920, right before the implementation of the Clayton Act, with those that did not,
before and after the enactment of the Act, regardless of whether bankers actually stepped down
from the boards after the implementation of the Act. Alternative specifications focus on changes
that occurred after the Act was passed in 1914, even though its restrictions on railroads were not
implemented until 1921. It is possible that bankers made changes in their relationships with their
client firms in this interim period; the partners from some investment banks resigned from railroad
boards as early as 1914.
Our empirical tests of the effects of banker directors focus on the firm outcomes that were at
the center of the historical debates, as well as other outcomes of more modern theoretical interest.
With respect to the former, we analyze the effect of bankers on railroads’ interest costs and leverage,
in order to test whether bankers exploited their client firms by forcing them to over-issue costly
debt securities. We also test the effect of banker-directors on railroads’ mark-ups, to investigate the
progressives’ claims that bankers facilitated anti-competitive arrangements.4 With regard to the
latter, we test the effect of banker directors on railroads’ investment-to-cash flow sensitivities, to
investigate whether close ties to banks helped relax railroads’ financial constraints. Finally, we also
examine the effect of banker-directors on railroads’ valuation, the clearest indication of whether
these relationships were beneficial overall to the owners of these enterprises.
3
In the future, we will expand our analysis to include industrial corporations as well.
Any such anti-competitive arrangements would benefit the railroads, rather than harm them. Thus, these arrangements would not reflect a conflict of interest between banker-directors and the firms’ shareholders. We discuss
this issue in more depth below.
4
3
Our findings indicate that the participation of investment bankers in the management of railroads generally benefitted the railroads. Comparing the performance of roads with bankers on
their boards to those without a banker-director, we find that the market valuation (as measured
by Tobin’s Q) of roads with investment-banker directors fell relative to others in the post-Clayton
Act period. The results also provide some insight into the channels through which these effects
occurred. The rate of interest that railroads with investment bankers on their boards paid on their
debt rose after the implementation of the Clayton Act, relative to the rates paid by other roads.
We also find that the presence of bankers lowered the sensitivity of investment to the cash flows
of the firm. Thus, bankers appear to have eased financial constraints, although this effect was
experienced only by railroads with the partners of J.P. Morgan & Company, or a few of the other
most prominent investment banks, on their boards. Taken together, our results provide evidence
against the notion that banker-directors harmed their client railroads.
As with any difference-in-difference estimation, it is important to be able to rule out other
factors that may have altered the value of firm-bank relationships at the same time that the quasiexperimental variation is imposed. Other changes during this period that may have influenced the
value of firms’ financial ties include the government operation of railroads during World War I,
the rise of a broader investor class after the Liberty Bond drives during the war, the increase in
the availability of financial information as more firms listed on the NYSE, the introduction of new
securities issuance regulations (‘blue sky laws’), and the founding of the Federal Reserve. Even
if these events contributed to a decline in bank-firm relationships, it is not clear that they would
have led to a differential effect in outcomes across our treatment and control firms. Moreover, these
potentially confounding factors mostly affected firms slowly over time, whereas the timing of the
Clayton Act is more discrete. The Act was passed in 1914 but its implementation for railroads was
postponed until 1921. Thus, the effect on railroads should be centered around this later date.5
The data and results of this paper contribute to several areas of research. An important question
in modern corporate finance is whether the presence of bankers as directors is beneficial for the
shareholders of a firm. An large empirical literature has studied this question using modern data
in countries in which bankers have an extensive presence on boards, such as Germany (Gorton and
5
While our analysis focuses on railroads, the differential in the timing for industrials and railroads will allow us
to differentiate the Clayton Act from other explanations further when the accounting data for industrial firms is
available.
4
Schmid, 2000; Agarwal and Elston, 2001; Fohlin, 1998) and Japan (Weinstein and Yafeh, 1998;
Mørck and Nakamura 1999), as well as in those with a modest presence of bankers, like the US
(Booth and Deli, 1999; Kroszner and Strahan, 2001; Güner, Malmendier and Tate, 2007; Minton,
Taillard and Williamson, 2010). An important limitation of these studies is that they are generally
not able to address the endogeneity of board composition (Hermalin and Weisbach, 1988; 1998) or
bank affiliations. This paper improves upon those earlier works by exploiting a quasi-experiment
that generates exogenous variation in bankers’ presence on boards.
Among economic historians, the role of financial firms in the U.S. economy in the early twentieth century has been the focus of a longstanding and important literature (see, for example,
Edwards 1938; Redlich, 1951; Carosso and Sylla, 1991; and Calomiris, 1995). Most of the empirical
contributions to this literature have focused almost exclusively on the activities of one firm, J.P.
Morgan and Company (Delong, 1991; Ramirez, 1995; and Simon 1998).6 It is unclear whether
the results of those studies generalize to other, less prominent firms. Moreover, like the modern
corporate finance literature, this work generally can not address the problem of the endogeneity of
relationships between bankers and firms: perhaps the positive effects found for the presence of a
Morgan partner on a firm’s board were simply the result of the fact that J.P. Morgan and Company
formed relationships with superior firms. The results of this paper build on that earlier work by
extending it to all banks and firms, and by improving on their identification strategies.
2
Bankers and Railroads in the Era of Finance Capitalism
Railroads were the first true examples of “big business” in the United States, and from the later
decades of the nineteenth century through the early twentieth century were the largest corporate
issuers of securities. Large railroad systems were in some cases created through consolidations, and
in other cases reorganized, by private investment banks which gradually achieved the capabilities
necessary to raise the enormous amounts of capital required by these firms. The investment banks
that participated in major underwritings often held large ownership stakes and board seats with
commercial banks and trust companies, which provided credit to finance underwritings, and also
6
All of these contributions find substantial benefits of affiliation with Morgan; Hannah (2007) is skeptical of the
prevailing interpretations of these results.
5
with insurance corporations, which were large purchasers of railroad debt.7 Investment banking
partnerships cultivated close relationships with railroads and, beginning in the 1880s, put one or
more of their partners on the boards of directors of their client firms. The relationships investment
bankers formed with railroads were often quite durable, with the railroad often relying exclusively
on the investment bank represented on its board for securities underwritings. In some cases,
investment bankers were initially granted absolute control over their client railroad’s management
through a voting trust arrangement, which gave a small number of trustees (usually the bankers
themselves) sole voting authority over the firm’s stock (see Redlich, 1951).
In the early decades of the twentieth century, industrial companies became increasingly important as issuers of securities, and many new entrants into the investment banking industry focused on
those firms. Nonetheless, railroads’ securities offerings remained substantial, and several individual
railroads were among the largest issuers of securities in the United States well into the 1920s. The
real value of new bond issues by railroads remained roughly constant from 1910-14 to 1920-24, at
around $1.6 billion for each five-year period. However, issues by other sectors grew rapidly, so that
the relative share of railroads among all corporate bond issuance fell by more than half, from 49%
to 21%.8
Close affiliations between railroads and investment banks were often established during periods
of expansion by a railroad, when the firm would make significant acquisitions and/or issue large
amounts of securities, or during episodes of financial distress. In the wake of the panic of 1893,
investment bankers became actively involved in the management of many important railroads,
reorganizing their financial structures, consolidating their operations, and putting a halt to the
riskier strategies pursued by less prudent managers (see Carosso, 1970).9 In cases where railroads
faced bankruptcy, the participation of bankers in the reorganization was critical, as they were able
to design a reorganization plan that could be seen as protecting the interests of the bondholders,
7
Some trust companies, and, in the years after 1910, national banks entered the investment banking business,
becoming important bond underwriters in the 1920s (see, for example, White, 1986, and Kroszner and Rajan, 1997).
8
Values presented in Dollars of 1910. Data obtained from Financial Review, 1914 and from Commercial and
Financial Chronicle, various issues.
9
For example, the Chicago Great Western railroad, which was financed with a large amount of short-term debt,
went into receivership in 1908 as it was unable to roll over its debt. The stockholders and bondholders of the firm
formed committees that invited JP Morgan & Company to reorganize the firm’s capital structure. A Morgan partner
was placed on the railroad’s board, but the firm was also granted absolute control over the railroad through a voting
trust. Morgan underwrote the issuance of new securities that replaced some of the preferred stock and all of the
short-term debt with long-term debt, which matured in 1959.
6
and thus reduced their incentives to foreclose. Other railroads resisted entering into a relationship
where bankers would participate in their management, often because they were controlled by an
investor who did not want to surrender their control over management to bankers.10 These historical
accounts suggest that bankers generally became involved in the management of railroads when they
were needed. Railroads without investment bankers may have been simply more financially sound.
It is therefore likely that that the railroads with and without bankers differed in important ways,
which were not actually caused by the bankers, but were instead related to the conditions that led
to the bankers coming onto the board.
Banker directors were important in one other respect as well. In the late-nineteenth and early
twentieth centuries, many railroads with competing lines fell into vicious rate wars with one another,
with disastrous consequences for their profitability. Some banker-directors, particularly those affiliated with large numbers of railroads such as J.P. Morgan, publicly called for cooperation among
competitors, and attempted to facilitate collusive “gentlemen’s agreements.” Although the more
aggressive enforcement of antitrust laws in the years after the Hepburn Act of 1906 limited bankers’
ability to pursue these strategies, progressive critics of banker participation in railroad management
argued that the many interlocks created among the boards of competitors by bankers continued to
facilitate collusion in the industry.
2.1
Panic and Backlash
Prior to the panic of 1907, popular suspicion of financial institutions and concern over the anticompetitive practices they were believed to facilitate produced legislative efforts to reign in the
power of these institutions, for example with proposals for federal incorporation of firms engaged in
interstate commerce (see, for example, Mitchell, 2007). But the crisis provoked shock and outrage,
and the calls for strict regulations became more forceful. The panic of 1907 began with the collapse
of several trust companies and ended with a series of rescues organized almost single-handedly by
J.P. Morgan (see Bruner and Carr, 2007). Bankers were accused not only of profiting from the
crisis—an accusation made plausible by J.P. Morgan’s arrangement for U.S. Steel to acquire a controlling stake in an important competitor held by a failing brokerage house during the crisis—but
10
A concise history of banker involvement in railroad management, and the activities of other investors who
competed with bankers for railroads, is presented in Faulkner (1951).
7
also of actually causing the crisis (Carosso, 1987).
After various investigations and legislative efforts at both the state and federal level, the U.S.
House of Representatives authorized an investigation of the so-called “money trust,” the financiers
believed to secretively and self-servingly control all access to capital, in 1912.11 Led by Representative Arsène P. Pujo, the House Committee on Banking and Currency and its influential council,
Samuel Untermyer, conducted extensive hearings, calling many important bankers to testify. The
committee also collected copious information on banking and the financial system. Although the
committee was never able to fully complete the work Untermyer hoped to undertake, it issued a
very thorough report (Pujo Committee, 1913b). This report compiled important information on
the ties between financiers and corporations, and included a long series of recommendations for the
regulation of financial institutions, the protection of investors, and the restraint of anticompetitive
practices in finance and industry.
One of the most important reasons for the lasting influence of the Pujo Committee’s activities
was a series of essays written by the progressive lawyer Louis Brandeis, and published as the book
Other Peoples’ Money and How the Bankers Use It in 1913. Brandeis wove the data and findings
of the committee into a powerfully articulate narrative that attacked what Woodrow Wilson called
the “money monopoly” on many fronts. Brandeis showed that the most powerful bankers in the
country, the partners of J.P. Morgan & Co., the directors of First National Bank, National City
Bank, and the various other bank and trust companies they controlled, altogether held directorships
in hundreds of corporations, whose total assets amounted to more than $20 billion, a staggering
sum at the time. Brandeis claimed that the presence of even one of these bankers on the board
of a corporation gave them control of the firm’s resources, which they could utilize for the benefit
of the other companies they controlled at the expense of the company’s shareholders. Brandeis
stated that “when once a banker has entered the Board his grip proves tenacious and his influence
usually supreme; for he controls the supply of new money” (1913, p. 11). In his view, bankers
got their clients to issue excessive amounts of securities at exorbitant fees; forced their firms to do
business with other firms they controlled; and coordinated anticompetitive arrangements among
11
The term “trust” was used at the time to denote all the great combinations that dominated industries, since the
first such enterprises were legally organized as trusts. The term was used synonymously with “monopoly”—hence the
term, “antitrust.” These “trusts” are not to be confused with trust companies, which were financial intermediaries
similar to banks but subject to fewer regulations.
8
their companies. Brandeis argued that, in the extreme, the bankers’ actions could even cause the
failure of corporations.
2.2
The Clayton Act of 1914
Although the recommendations from the Pujo Committee were not passed as legislation, the committee’s findings were nonetheless enormously influential. The Clayton Antitrust Act, passed in
1914, was intended to sharpen the enforcement of the Sherman Antitrust Act, and prohibit the
anticompetitive arrangements the “money trust” was believed to facilitate. Among the new provisions of the Act was a prohibition against interlocking directorates among competing industrial
firms (Section 8), something which the Pujo Committee argued was widespread. Together with the
Federal Trade Commission, also created in 1914, the Clayton Act marked a significant enhancement
of antitrust enforcement in the United States.
The long history of extensive banker participation in railroad management, highlighted by the
Pujo Committee and Brandeis (1913), led the authors of the Clayton Act to impose particularly
strict terms specific to that industry, intended to curtail the influence of banker-directors. As the
progressive critique of bankers held that they engaged in tunneling, Section 10 of the Act explicitly
forbade any transactions between railroads and the firms of banker-directors. In particular, it states
that:
No common carrier engaged in commerce shall have any dealings in securities...to the
amount of $50,000, in the aggregate, in any year, with another firm, partnership, or
association, when the said common carrier shall have upon its board of directors or
as its president, manager or as its purchasing or selling officer...any person who is at
the same time a director, manager, or purchasing or selling officer of, or who has any
substantial interest in, such other corporation, firm, partnership or association...12
As underwriting constitutes “dealings in securities,” whereas lending may not, this provision clearly
applies to investment bankers, rather than commercial bankers.13 Investment bankers who sat on
the boards of their client firms could choose to remain on those boards and cease to act as their
12
Oct. 15, 1914, ch. 323, 38 Stat. 730. This term was repealed in 1988.
One possible exception would be a collateralized loan, where the collateral offered was securities owned by a
railroad. Such lending did indeed occur, but in general bank loans were a very small part of railroads’ capital stock.
Our evidence shows that railroads were mostly financed through corporate bonds and retained earnings.
13
9
underwriters, or resign. A relatively small number of commercial banks and trust companies
participated in securities underwritings, and therefore faced the same choice.
The implementation of Section 10 of the Clayton Act—the terms relative to railroads—did not
occur immediately, however. An two-year delay in its implementation was enacted, so that the
Interstate Commerce Commission could develop the capacity to enforce it (see House Judiciary
Committee, 1917).14 Further delays were then implemented in 1916 and 1918, in part because
railroads were extensively involved in the U.S. war effort in World War I. In 1917, the federal
government assumed control over much of the operations of U.S. railroads as part of that effort,
effectively guaranteeing them minimum rates of return on their operations. But control was restored
to the railroads themselves in 1920. Although Congress passed another delay of the implementation
of Section 10 in 1920, which would have postponed it until 1922, President Wilson vetoed this
extension, and its provisions went into effect on January 1, 1921. Contemporary press coverage
indicates that this was largely unexpected.
Would this restraint on banker-directors have had any impact on railroads? Since the publication of Brandeis’ book in 1913, some financial historians have expressed skepticism of his negative
view of investment banks’ influence on their client firms. Morrison and Wilhelm (2007) argue that
investment banker activism in firms’ management improved corporate governance and helped resolve problems arising from informational asymmetries, and De Long (1991), Ramirez (1995), and
Cantillo (1998) all find positive effects of the presence of J.P. Morgan partners on a firm’s board. In
a view that is not entirely compatible with these findings, other historians have expressed skepticism
about the extent of influence bankers obtained from the one or two board seats they typically held
on their client firms. Carosso (1970, p. 152), for example, claimed that investment bankers were
“unable to impose their will upon the other directors, who were always more numerous than the
representatives of Wall Street.”15 Redlich (1951) argues that investment bankers mainly selected
14
The act contains a provision enabling railroads to offer their securities business to an open bidding process, which
would be supervised by the Interstate Commerce Commission. If a director’s firm offered the lowest bid, then the
railroad would be permitted to accept that firm’s bid.
15
This objection is particularly salient for our study, since we are only able to identify the connection between
financial firms and railroads through the presence of bankers on the board of directors. There were, of course, other
mechanisms in which bankers may have exerted influence on firms, and we address some of them to the extent that
our data allows in our empirical analysis. Unfortunately, we are unable to observe whether a particular bank provided
loans to a firm or the ownership structure of corporations. However, these two concerns are probably less relevant for
railroads, since most of their borrowing took the form of corporate bonds and since the large size of their enterprise
precluded most individuals or financial institutions from having a large ownership stake for most railroads.
10
the top executives of firms, and then delegated management to them.
3
Data
No comprehensive dataset of accounting data or board composition exists for the early twentieth
century. We have collected such a dataset for this paper. Our main source is Moody’s Manuals of
Railroads and Corporation Securities, which presents firm-level data for large numbers of corporations, obtained from those corporations’ annual reports to shareholders. Prior to the Securities and
Exchange Acts, publicly traded firms were not required to disclose financial data. However, beginning around 1900, the NYSE began to require listed corporations to disclose basic income statement
and balance sheet information. Therefore, our sample is restricted to NYSE-traded firms.
Our current sample includes all railroads that traded at the NYSE in 1907, 1915, and 1920
(a total of 84 roads).16 For each railroad, we collect financial information for each year from
1902 to 1929 using the statements presented in Moody’s Manuals.17 This allows us to construct an
unbalanced panel dataset containing information on revenues, costs, leverage, and various measures
of firm performance (such as return on assets and Tobin’s Q, for example). We also collect stock
price data at an annual frequency from the New York Times.
Due to the lack of standardization in the disclosure of financial information, an important
challenge is to interpret companies’ accounting statements in a consistent manner to construct a
uniform set of variables.18 When possible, we use overlaps across manuals to ensure the consistency
and accuracy of the data within firms. Moreover, we use alternate variable definitions in our analysis
to take these idiosyncrasies into account.
While the accounting data is constructed at an annual frequency, we obtain board composition
and bank membership data at two- to five-year intervals. This restriction is imposed both due to the
16
We are in the process of constructing similar data for all industrial firms that traded at the NYSE in 1907 and
1915.
17
Because the disclosure of financial information for NYSE-traded firms improved at the turn of the century, many
firms do not report much accounting data until the 1905 manual. Thus, our data effectively starts in 1902. For
most firms, the manuals from 1905 to 1915 provide two years of financial data, so we can collect annual financial
information using manuals from every other year. Later manuals contain five to six years of data, allowing us to use
fewer volumes while still collecting information at an annual frequency.
18
Lack of consistency in accounting data is particularly problematic for industrial firms but less severe for railroads
because their accounts became much more standardized after the Hepburn Act of 1906. Although there remained
significant variation across companies in the accounts that were presented on their financial statements, the variation
across firms diminished over time.
11
large magnitude of the data collection effort and the difficulty in obtaining some of the rare primary
sources we work with. However, because directors of nonfinancial and financial firms change slowly
over time, we are still able to capture important variation in the composition of boards over time.
We collect the names of all directors and managers as listed in Moody’s Manuals.19 To identify
which directors are bankers, we use various data sources containing information on members of
financial firms. For investment banking partnerships, we collect lists of all partners from stock
exchange directories and partnership directories for various years in our sample period.20 We also
obtain lists of directors of commercial banks and trust companies from the Rand McNally Bankers’
Directory.21 In order to restrict our dataset to the individuals most likely to serve on the boards
of major corporations, we focus on investment banking partnerships that were registered members
of the NYSE, and of commercial bankers (that is, the president and directors of commercial banks
and trust companies) in New York, Chicago and Boston.
Cross-referencing the lists of investment and commercial bankers with those of corporate directors enables us to create a detailed panel dataset on the presence of bankers on boards. After
carefully correcting the name entries by comparing those for the same firm over time, we match
the names of corporate directors to those of bankers based on the full last name, the initials of
first and second name, and, when used, a suffix. A concern that arises in this procedure is that
matching on names may lead to erroneous matches. For example, our matching procedure will lead
us to assume that investment banker J. A. Smith is also the director listed as J. A. Smith, when
these may in fact be two different individuals. We have performed several checks to address this
important concern. First, at least one of our sources often lists the full name of the individual. In
that case, we can match initials and full names by using the name of the firm the individual belonged to.22 Also, we match bankers based on names and cities when the geographical information
is available. As an added robustness test, we verified our matching results against the descriptions
of interlocks between financial institutions and railroads’ boards in the Pujo Committee reports.
19
For railroads, we have this information for 1905, 07, 09, 11, 13, 15, 20, 23 and 25.
We observe investment banking partnerships in 1906, 10, 13, 15, 17, 20, and 1923. When matching these data to
those of directors, we link 1906 information to board data in both 1905 and 1907, the 1910 data to both 1909 and
1911, and the 1923 data to both 1923 and 1925.
21
We have collected this information for 1905, 07, 09, 11, 13, 15, 20, 23 and 25.
22
Our statistics on the presence of bankers on nonfinancial boards are also robust to alternative matches for cases
in which the same initials can be attributed to more than one full name, and to the handful of cases in which a suffix
(such as Senior or Junior) is added or dropped for a given name over time.
20
12
For a total of 179 bankers holding a total of 155 seats, we failed to identify bankers on only 4
railroad boards because these individuals were not listed in the Moody’s Manuals. Moreover, there
was only one banker in our sample that was not listed in the Pujo report. In all other instances, we
obtained a perfect match.23 It is possible that our matching procedure still exaggerates the degree
of interlocking across directorates and financial institutions. However, the presence of bankers on
boards that we document in this paper is so high that our results are surprising even taking mismeasurement into account. Moreover, there is no reason to believe that these matching concerns
are systematically changing over time. Thus, measurement error in the matching procedure cannot
explain the significant changes in the presence of bankers on boards that we document over time.
4
Bankers on Railroad Boards and the Clayton Act
Table 1 presents summary statistics for the boards of the sample railroads, calculated over 1905-25,
by matching the names of railroad directors to the names of partners in investment banks, and to
the names of directors of commercial banks, for the same year. These statistics were calculated by
matching lists of 7,769 railroad director name-years, 36,512 commercial bank director name-years,
and 17,418 investment bank partner name-years.
The matching results, presented in table 1, contain some genuine surprises. Whereas today
relatively few publicly traded firms in the U.S. have bankers on their boards (about 30% among
large firms; see Kroszner and Strahan, 2001), in the early twentieth century, 93% of railroads had
a commercial banker on their board, and 57% had a partner from an investment bank on their
board.24 These statistics, as well as historical accounts, suggest that bankers held an extraordinary
degree of influence in the economy during the period of finance capitalism compared to recent
decades, and participated in the management of a far greater share of large corporations.
Table 1 reveals important differences in the relationship between financial and nonfinancial firms
across types of financial institutions. Of the twelve total directorships for an average railroad, five
of those were typically held by commercial bank directors. Moreover, those five banker-directors on
average held directorships with nine different commercial banks, creating a large web of connections
23
We also plan to verify accuracy of our matches by consulting published Directories of Directors for board members
residing in New York City or Boston.
24
The percent of U.S. publicly-traded firms with at least one investment banker on their board for recent years is
much lower: 16% (Güner, Malmendier and Tate, 2007).
13
Table 1:
Bank representation on railroad boards, 1905-1923
Mean
SD
Min
Max
Board characteristics
Board Size
12.43
3.18
4
26
Commercial bank representation on board
At least one commercial banker on board
Seats held by commercial bankers
Number of different commercial banks represented
0.93
4.88
9.39
–
3.32
7.92
0
0
0
1
14
34
Investment bank representation on board
At least one investment banker on board
Seats held by investment bankers
Number of different investment banks represented
0.57
1.11
0.98
–
1.29
1.14
0
0
0
1
6
6
Note: Statistics calculated from 625 firm-year observations taken at intervals of two to five years
over 1905-25.
with the commercial banking community.25 In contrast, the representation of investment banking
partnerships on railroads suggests that relationships between these two types of firms were often
exclusive. On average, only one seat was held by an investment banker, who represented only one
investment banking firm. While some investment bankers were also directors of commercial banks,
investment bankers were partners of only one investment banking house.
In order to gauge the impact of the Clayton Act, which was passed in 1914 and implemented
in 1921 for railroads, we next examine the historical trends in railroad-banking relationships over
time. As discussed earlier, Section 10 of the Clayton Act limited banks’ ability to serve as the
underwriters of the railroad. Thus, we expect the Act to disproportionately affect the presence of
investment bankers on boards, as investment banks conducted most of the underwriting activity.
Moreover, we do not expect investment bankers to completely disappear from railroads’ boards.
Indeed, these bankers could have remained on the board as long as the bank no longer underwrote
securities for the firm.
Figure 1 plots the trend in banker representation on railroad boards by year. Panel (a) shows
the proportion of firms with at least one banker on their board for 1905-25. For commercial
bankers, this rate displays a slight decline over time, but it never falls below 89%.26 The rate for
25
Interlocks among the directorates of commercial banks were common; the commercial bank directors who sat on
our sample railroads’ boards held directorships with 2.2 commercial banks on average.
26
On average, there are a total of 288 commercial banks in a given year in our sample. Of these, about 31% had
representation on the board of railroads.
14
(a) Proportion of firms with bankers on their boards
1
Commercial Bankers
0.8
0.6
Investment Bankers
0.4
0.2
Clayton Act
Passed
0
1905
1910
1915
Clayton Act
Implemented
1920
1925
(b) Median proportion of banker−directors
1
Clayton Act
Passed
Clayton Act
Implemented
0.8
0.6
Commercial Bankers
0.4
0.2
0
1905
Investment Bankers
1910
1915
1920
1925
Figure 1: Evolution of banker representation on railroad boards
15
Student Version of MATLAB
investment bankers is somewhat more volatile, and exhibits many of the same fluctuations as the
line for commercial bank representation, only magnified.27 The most important difference between
the two lines is the change that occurs between 1920, before the Clayton Act was implemented for
railroads, and 1923, after its implementation. Over this interval, the fraction of railroads with at
least one investment banker on their boards fell from 59% to 41%, whereas the share of railroads
with commercial bankers on their boards actually increased somewhat.28 The pattern of investment
banker resignations thus conforms to our expectations about the effect of the Clayton Act.
The lower panel of figure 1 displays the median proportion of bankers among railroads’ directors
over time. For commercial bankers, this rate gradually decreases, falling from 55% (about 7 bankerdirectors) in 1905 to 23% (about 3) in 1923, although it increases slightly in 1925. The pattern
for investment bankers is quite different. The median fraction of investment bank partners among
railroad directors is remarkably stable up until 1920, at around 7% (one investment banker). In
1923, after the Clayton Act, this falls to 0—the median railroad no longer had an investment banker
on their board. And it remained at zero in 1925. Many investment bankers of course remained
on railroads’ boards, and some even had large numbers of investment bankers. But the mean
proportion of investment-banker seats on railroads boards fell by 29%, from 9% to 6.3%, over that
period.
Finally, deeper insights into changes in bankers’ representations on railroads’ boards can be
obtained by calculating measures of turnover among banker-directors. We define ‘net turnover’ in
a given year as the number of new banker-directors—that is, those who were not on the board in
the previous period—minus the number of banker-directors departing from the board (i.e., those
that were on the board in the previous period but not in the current period). This is a ‘net’ concept
of turnover since it explicitly incorporates replacements for directors who may have resigned.
Table 2 displays this data for investment bank partners on firms’ boards. Turnover among
investment bankers becomes quite negative between 1920 and 1923, which is consistent with a
large effect of the Clayton Act’s provisions in that year. Indeed, the average firm lost a quarter of
27
On average, there are a total of 570 investment banking partnerships in a given year in our sample. Of these,
only about 5.5% had presence on railroads’ boards.
28
Another sizable decline in the fraction of investment bankers on boards occurred from 1911 to 1915, suggesting
that some bankers left boards in response to the backlash during the Pujo investigation and the enactment of the
Clayton Act, even if the Act did not immediately applied to railroads. Partners of J.P. Morgan & Co., for example,
famously stepped down from many railroad boards in January 1914 (Simon 1998).
16
Table 2:
Turnover among investment banker directors, 1907-1925
Investment bankers
On railroad boards:
Net turnover
-0.24
-0.11
0.17
-0.06
-0.07
0.09
-0.25
0.01
Year
1907
1909
1911
1913
1915
1920
1923
1925
an investment banker in the 1920-23 period, a more than 20% decline relative to a mean of 1.16 seats
in each board held by partners of investment banks in 1920. The increase in net turnover among
investment-banker directors is consistent with the implementation of Section 10 of the Clayton Act
in 1921 having a significant effect on board composition.
What effect did the Clayton Act’s provisions on railroads have on their performance? The
press certainly noted the sudden changes made in the composition of railroad boards in response
to the Act, and reported concerns among railroad officials that the resignations would “work to
the great disadvantage” of the affected roads.29 On the other hand, progressives lauded the Act’s
provisions to “forbid...acts of piracy” by forbidding self-dealing among railroad directors.30 Some
labor activists argued that bankers’ control over both steel firms and equipment producers, on the
one hand, and railroads, on the other hand, enabled them to maintain high prices for railroad
supplies, thereby extracting resources from the railroads—and “unconscionable misuse” of their
power.31 In the end, whether close relationships with bankers helped or hurt firms during the
period of development of financial markets is an empirical question. We exploit the variation in
these relationships introduced by the Clayton Act to analyze this question in the next section.
29
“Many Changes Soon in Railroad Boards,” New York Times, 23 January 1921. The resignations of prominent
banker-directors attracted considerable attention in the press; see, for example, “Schiff and Kahn Quit Union Pacific,”“Mellon Leaves the P. R. R.,” “Two New Erie Directors,” New York Times, 2 December, 13 January, and 24
July 1921, respectively.
30
Machinists’ Monthly Journal, January 1921, p. 61.
31
“Rail Unions Charge Plot by Financiers to Force Idleness,” New York Times, 20 April 1921.
17
5
Railroads, Investment Banks and Performance
5.1
Accounting data
The board composition data analyzed in the previous section were matched to the financial data
collected from Moody’s to create a dataset that can be used to analyze the effect of the Clayton
Act on the performance of railroads. Thus far, we have completed coding the data for 82 railroads,
almost all of the 84 railroads that had common and/or preferred stock listed on the NYSE in 1907,
1915 and 1920.32 For each firm, we collect financial information in every year from 1902 to 1929
as long as a balance sheet and income statement are provided in several editions of the Moody’s
manuals from 1905 to 1930. For each year in the sample, which covers 1902-29, we observe on
average 66 railroads (ranging from a minimum of 22 in 1902 to a maximum of 74 in several years).
Overall, our dataset contains 1,859 firm-year observations with non-missing information on total
assets.
Definitions and summary statistics for the accounting data are presented in table 3.33 As
the first few rows of the table make clear, the NYSE-traded railroads were extraordinarily large
enterprises, certainly some of the largest in the United States. The average value of the total assets
of our firms is about $200 million (of year 1910 dollars), and the bulk of that, or 76%, was in the
form of “Property, Plant & Equipment (PP&E),” the long-term physical assets such as the road
itself, and locomotives and rail cars.34 Of those assets, about 42% were financed by long-term debt,
an indication that securities underwriting was quite important for these firms. Using the balance
sheet and income statement data, we calculate an “interest rate” variable, a rough proxy for the
rate paid on the debt, as their total interest expense divided by all long-term interest-bearing
liabilities.35 On average, this was around 4.6%.
32
These counts excluded the Canada Southern Railroad which, although trading in the NYSE in 1920, provided
financial information for only one year in our entire sample period.
33
To prevent our estimates from being influenced by outliers, we trim all variables at the top and bottom 1 percent
of the distribution for all years. Our regression results are usually stronger if we don’t treat outliers in this manner.
34
We obtain a fairly similar statistic (74.6%) when we exclude the 242 observations that report PP&E lumped
together with other non-physical assets, such as investment in securities.
35
Because of poor accounting reporting during the First World War years, short-term borrowing is difficult to
measure separately from trade credit and other current liabilities. Thus, we focus on long-term debt, which account
for most of the roads’ borrowing. For those years in which we can observe short-term borrowing independently,
this source of financing accounts on average for less that 1% of total assets. Moreover, the interest expense is often
lumped with other fixed charges. These accounting problems are usually common to all firms in a given year, and
are therefore taken into account by year fixed effects in our empirical analysis, but our results are also robust to
excluding problematic observations.
18
Table 3:
Definitions of accounting variables
Variable
Definition
Mean
SD
Min
Max
EBIT
Operating revenues - operating expenses (millions)
9.374
11.577
-16.07
105.13
PP&E
Property, plant and equipment (millions)
140.28
140.48
0
901.30
Assets
Total assets (millions)
196.50
203.69
1.4561
1,152.72
Age
Years since earliest incorporation date
38.736
25.235
0
124
Mileage
Miles of track operated
2,942.36
3,044.12
0
14,563
Book leverage ratio
Long−term debt
Assets
Interest expense
Interest−bearing liabilities
Operating revenues
Operating expenses
∆PPE
LaggedPP&E
Net income
LaggedPP&E
Net income
Assets
Market value of assets
Book value of assets
0.416
0.154
0
0.716
0.046
0.012
0.022
0.106
1.369
0.171
0.935
1.9
0.014
0.166
-0.68
2.255
Interest rate
Markup
Investment
Cash flows
Return on assets
Tobin’s Q
0.03
0.041
-0.076
0.628
0.019
0.017
-0.023
0.094
0.868
0.176
0.444
1.939
Note: Data calculated for all available firm-years for each variable. Number of observations ranges from 1,640 to 1,790 for
most variables. For Tobin’s Q, which requires stock price data, N=1,214, as many railroads’ stocks were illiquid and traded
infrequently. All variables in 1910 dollars based on the BLS-based consumer price index. All variables have been trimmed at
the top and bottom 1 percent.
In terms of the performance of these enterprises, their return on assets was around 1.9%, and
their return on equity 5.0%, with the difference generated by the high degree of leverage of the
roads. On average, the railroads generated cash flows equivalent to about 3% of their level of capital
at the beginning of the year (measured by their lagged level of PP&E), and paid out about 31% of
their net income in the form of dividends on common stock. For the firm-years for which we are
able to obtain information on stock prices, the average value of Tobin’s Q for the roads was about
.87.36
5.2
Railroads with and without bankers
How did the roads with at least one investment banker on their board in 1920 (the year before
the implementation of the Clayton Act) compare to those that did not have a relationship with
a bank in that year, over our entire sample period?37 A simple comparison of means between
36
We obtain the market price of common shares on the last day of trading in the year from the New York Times.
If the shares did not trade on that date, we use the trading price on the closest day in the two week period beginning
one week prior to the end of the year.
37
We focus our analysis on investment bankers because the Act was designed to affect the relationship between
railroads and those financial institutions that underwrote their securities. It is likely that a few commercial banks
19
Table 4:
Characteristics of railroads with and without bankers,
All firm-years, 1902-1929
Means:
No investment
Investment banker
Banker on board
On board
(1)
(2)
t
(3)
Physical characteristics
Total mileage
Log assets
Age
2,912
18.341
39.422
2,965
18.57
38.228
-0.344
-3.805
0.983
Debt, liquidity and investment
Book leverage ratio
Interest rate
Cash flow
Investment
0.406
0.047
0.030
0.014
0.423
0.046
0.029
0.014
-2.385
2.736
0.312
0.043
Performance
Tobin’s Q
Return on assets
Markup
0.893
0.022
1.376
0.853
0.016
1.365
3.799
6.652
1.341
Note: Column (1) presents means for all firm-years for railroads that did not have an investment
banker on their board in 1920, column (2) presents means for all firm-years for railroads that had
an investment banker on their board in 1920, and column (3) presents the t statistic for a two-sided
test of differences in means.
the two groups of firms over the entire sample provides some insight into the nature of railroads’
selection into these relationships. Although prior work has shown that J.P. Morgan & Company
was affiliated with firms that were superior relative to their peers (DeLong, 1991), it is not obvious
that we should expect selection to be positive in general: distressed firms, or firms in need of
restructuring or improved access to credit might seek out these relationships. Indeed it is precisely
under these circumstances that many firms took on investment bankers as directors (Carosso, 1970;
Ramirez, 1995).
Comparisons of means between railroads with and without investment-banker directors are
presented in table 4. Although the roads with bankers on their boards were somewhat larger, as
measured by total assets, they were similar in terms of mileage, age, the level of investment, and
cash flows.38 By contrast, measures of their performance indicate they were inferior: they had
in our sample also engaged in the underwriting of securities at that time. Exploring this channel in detail requires
further information on the underwriting of securities by financial institution.
38
We measure age using the year of incorporation of the railroad. This date often changes for a specific railroad
within our sample period, as firms restructure and reincorporate following financial distress, M&A activity, or government antitrust rulings. To avoid these problems, we calculate the age of roads using the earliest incorporation
year available for the firm.
20
lower rates of return, lower market valuations, and were able to generate somewhat lower revenues
for the same level of operating expenses. They did, however, pay less interest relative to their
interest-bearing liabilities, and were somewhat more leveraged.
These differences across the two types of firms are of course not causal. For example, the
railroads with investment bank affiliations may have performed better than they otherwise would
have without the participation of their bankers in their management. To obtain evidence of the effect
of the bankers’ role in railroads management, we perform regressions exploiting the implementation
of the Clayton Act.
5.3
Empirical Specifications
In 1920, 42 different investment banks had at least one of their partners on the board of a railroad.
The partners of these firms likely provided financial expertise, and also represented their firms
in underwriting transactions. The Clayton Act severely restricted these relationships, effectively
forbidding investment-banker directors from acting as a road’s investment banker.
Our analysis of the effects of bank-railroad relationships focuses on three main hypotheses,
drawn from the historical critiques of the role of bankers in the economy, and from the modern
literature on the value of ties between banks and nonfinancial corporation. These hypotheses, and
the corresponding empirical specifications, are as follows.
Hypothesis 1: Investment banker-directors exploited their client firms by forcing them to issue
excessive levels of costly debt securities. This is the argument of Brandeis (1913), and has two clear
empirical implications. First, if bankers were able to exert monopoly power over their client firms,
they could have forced them to issue debt at higher interest rates (thus benefiting the bank itself
if it received compensation for underwriting in the form of the debt securities and held corporate
debt for a long period of time, as was often the case—see Carosso, 1970).39 This has been found to
be the case, for example, for Japanese firms with strong ties to banks (Weinstein and Yafeh, 1998).
If a bank affiliation allowed bankers to exploit firms in this way, we would expect the interest rate
on the debt of railroads with bankers on their boards to fall after the imposition of the Clayton Act
relative to that of other firms, since the Act presumably would have reduced the monopoly power
of bankers over their client firms.
39
Ideally, we would also like to directly observe underwriting fees, but this information is not generally available.
21
A second implication would be that bankers may have forced their client firms to issue higher
levels of debt, irrespective of the cost, to maximize the revenues obtained from underwriting (as
long as the extra leverage did not severely increase the cost of financial distress). This would
imply that the level of indebtedness of railroads with banker-directors should have fallen after the
imposition of the Clayton Act. It is important to note that a decline in the leverage ratio of firms
following the Act is not conclusive evidence of this hypothesis. Indeed, the presence of bankers may
have helped firms ease financial constraints and access external finance if these close ties solved
asymmetries of information between investors and firms, or if bankers’ presence served to certify
the quality of the firm. If bankers facilitated access to capital markets, we would also expect a
decline in firms’ indebtedness following the Act, albeit one accompanied by a decline in the value
of the firm.40 Nonetheless, a necessary consequence of this hypothesis is that indebtedness should
fall after the imposition of the Act.
These hypotheses will be tested in the following empirical framework:
rit = αi + γt + βXit + λBankeronboardi × PostClaytont + ²it ,
(1)
where rit is the interest rate on railroads’ debt; αi and γt are firm and time fixed effects to control
for unobserved firm characteristics that are invariant over time and any time-specific macro effects;
Xit is a vector of time-varying firm characteristics (such as firm size and the age of the firm),
and Bankeronboardi × PostClaytont is an indicator equal to one for all firms where at least one
investment banker was a director in 1920 (prior to the implementation of the Act), multiplied by an
indicator for the post-Clayton Act period, which takes a value of one for any year including or after
1921. In this framework, λ is the local average treatment effect of the Clayton Act’s prohibition
on self-dealing by banker-directors, and it is estimated using variation in the presence of bankers
across railroads before the Clayton Act was implemented. In alternative specifications, the level of
firm indebtedness, measured as the ratio of the book value of long-term debt to total assets, will
be used as the dependent variable. If the progressive critique of banker-directors was correct, then
we would expect to see a negative effect of the Clayton Act (λ < 0) in these regressions.
40
Güner, Malmendier and Tate (2007) show that an increase in borrowing that may come with close ties to banks
may not in fact be beneficial, if it simply enables the executives of unconstrained firms to finance value-destroying
projects.
22
To be clear about our identification strategy, this specification presumes that railroads with
bankers on their boards in 1920 had them throughout the pre-Clayton Act period, and therefore
likely assigns some partially “treated” firms into the “control” group, if some railroads had bankers
in earlier periods but not in 1920. Moreover, firms are “treated” as long as they had a banker on
their board in 1920, irrespective of whether these bankers stepped down from the boards in the
post-Clayton Act period or not. This definition of treatment and control groups should bias the
results against finding an effect.
It is also possible that firms may have been able to get around the Clayton Act regulations in
different ways. For example, influential bankers may have stepped down from boards to continue
underwriting securities but managed to appoint associates not directly affiliated with the bank to
the board of the railroad. Indeed, we found a handful of instances in which former partners of an
investment bank became directors of railroads in the years following the Clayton Act, while the
current partners resigned from the board. However, this behavior would also bias us against finding
an effect of the Act. Our results suggest that firm and bankers’ ability to minimize the impact of
the Act was limited, and provides further evidence of the important role that having a presence on
a board (either as a way to obtain inside information, to exert influence on the firm, or to provide
certification to outside investors) played during the era of finance capitalism.
Hypothesis 2: Investment banker-directors facilitated collusion among competing railroads and
other forms of anti-competitive behavior. This was a frequent concern among progressive critics of the money trust, and was specifically raised by the Pujo Committee. Collusion would not
necessarily have been detrimental to the shareholders of the railroad but, it was argued, would
have had negative consequences on overall welfare in the economy. If banker’s directorships were
an important mechanism to facilitate anticompetitive behavior, then the profitability of railroads’
operations—their revenues divided by their operating expenses—should have fallen after the imposition of the Clayton Act.41 Unfortunately any fall in profitability following the Act would be
observationally equivalent to bankers having improved the efficiency of railroads’ operations, and
thus raised profitability by lowering costs. We will try to distinguish these two effects by examining
the impact of the Clayton Act on railroads that had bankers on their boards whose investment
41
It is important to note that operating expenses do not include interest payments or most investment expenditures,
and therefore this hypothesis is distinct from the others in its empirical implications.
23
banks were also represented on the boards of many other railroads—the bankers who would have
been best able to facilitate collusion. To test this hypothesis, we will use a similar specification
to that of equation (1), only with railroads’ operating profit ratio (or markup) as the dependent
variable.
Hypothesis 3: Investment banker-directors relaxed railroads’ financial constraints. If affiliations
with banks helped firms to solve asymmetries of information and have better access to financial markets, the presence of banker-directors would have benefited railroads by enabling them to become
less constrained by their cash flows in their investment decisions. Standard models in corporate
finance hold that the sensitivity of a firm’s investments to their cash flows reflects the degree of
financial constraints faced by the firm. This hypothesis therefore implies that the Clayton Act
should have increased the investment-to-cash flow sensitivity for railroads with bankers on their
boards in 1920 compared to those that did not have banker-directors. A standard empirical model
of investment-to-cash flow sensitivity would be:
Iit = αi + γt + δ1 CFit + δ2 Qit−1 + δ3 CFit × Qit−1 + δ4 CFit × log(Assets)it−1 + βXit + ²it ,
where Iit is a firm’s investment rate, measured as the real change in its capital stock, divided by
the value of its real capital stock at the beginning of the period; CFit is a firm’s cash flows, again
divided by the capital stock at the beginning of the period; and Qit−1 is Tobin’s Q for the previous
period.
In order to estimate the effect of banker-directors, and the change caused by the Clayton Act,
we will modify this specification as follows:
Iit = αi + γt + δ1 CFit + δ2 CFit × Bankeronboardi + δ3 CFit × Bankeronboardi × PostClaytont +
+ δ4 Qit−1 + δ5 Qit−1 × Bankeronboardi + δ6 Qit × Bankeronboardi × PostClaytont +
+ δ7 CFit × Qit−1 + δ8 CFit × log(Assets)it−1 + βXit + ²it ,
where Bankeronboardi is an indicator variable equal to one for all years in the sample for those the
railroads with investment bankers on their boards in 1920. In this specification, the coefficients of
24
primary interest are δ2 and δ3 . The estimate of δ2 is the average difference in cashflow sensitivity
of investments between firms with and without bankers on their boards in the pre-Clayton Act
period. This difference does not have a causal interpretation, and most likely reflects the impact of
selection. The estimate of δ3 is the difference-in-differences for the post-Clayton Act period, and
presents the causal effect of the Act’s terms on railroads with bankers. If this estimate is positive,
it would indicate that the Clayton Act increased the sensitivity of investments to cash flows for
railroads with bankers on their boards relative to the others.
Finally, we will investigate the net impact of the Clayton Act on railroads with bankers on their
boards by estimating the impact on their valuations, as measured by Tobin’s Q. If bankers were
harmful to their client firms, then their valuations should have risen after the Clayton Act was
imposed. The empirical specification used to estimate this effect will be the same as (1), only with
Q as the dependent variable.
5.4
Estimation Results
We begin with results for regressions testing Hypothesis 1, reported in table 5.42 The table presents
results for regressions as specified in equation (1), with year fixed effects, which account for all timevarying influences that affected all railroads, such as macroeconomic conditions, developments in
labor relations, or changes in government regulations on railroad rate setting, and firm fixed effects.
The results in column (1), which present the simplest specification, indicate that in the post-Clayton
Act period, the interest rate on the debts of railroads with bankers on their boards in 1920 rose
relative to railroads without bankers on their boards. This finding rejects the hypothesis that
bankers increased the financing costs of their client firms, and presents evidence that the opposite
was true. Perhaps through their financial expertise and strong connections within financial markets
and with major creditors, investment bankers enabled their clients to issue debt at lower interest
rates.
Previous empirical research on the role of bankers in this era has focused exclusively on the
activities of the most important firm, J.P. Morgan & Company. In column (2), we investigate
whether J.P. Morgan’s clients enjoyed a differential benefit compared to those of other investment
42
In all our regressions, we use the entire sample from 1902 to 1929. However, results are generally robust to
narrowing the analysis to a shorter time period around the implementation of the Clayton Act in 1921.
25
26
1,548
Yes
Yes
0.000
(0.000)
0.001
(0.002)
0.023
(0.032)
0.005**
(0.002)
0.004**
(0.001)
(2)
1,548
Yes
Yes
0.000
(0.000)
0.001
(0.002)
0.023
(0.032)
0.004**
(0.001)
0.005**
(0.001)
(3)
0.48
0.48
0.48
denote significance at 1%, 5%, and 10%, respectively.
1,548
Observations
+
Yes
Firm fixed effects
R-squared
Note: Robust standard errors in parentheses. **, *, and
Yes
0.000
(0.000)
0.001
(0.002)
0.023
(0.032)
0.004**
(0.001)
Year fixed effects
Constant
Log(lagged assets)
Firm age
Investment Bank not underwriter on board × Post-Clayton Act
Investment bank underwriter on board × Post-Clayton Act
Investment bank w/ seats on 2 or fewer RRs × Post-Clayton Act
Investment bank w/ seats on 3 or more RRs × Post-Clayton Act
Investment bank w/ 3 or fewer seats × Post-Clayton Act
Investment bank w/ 4 or more RR board seats × Post-Clayton Act
Investment banker other than Morgan × Post-Clayton Act
JP Morgan partner on board × Post-Clayton Act
Investment banker on board × Post-Clayton Act
(1)
Table 5:
Regressions: Interest rate
0.48
1,548
Yes
Yes
0.000
(0.000)
0.001
(0.002)
0.021
(0.032)
0.004**
(0.001)
0.005**
(0.001)
(4)
0.48
1,548
Yes
Yes
0.000
(0.000)
0.001
(0.002)
0.025
(0.032)
0.004**
(0.001)
0.006**
(0.001)
(5)
banks, by including two interaction terms: one for railroads that had a J.P. Morgan partner on
their boards, and another for railroads that had only bankers from all other investment banks. The
results indicate that both J.P. Morgan’s clients and those of other investment banks both enjoyed
lower interest costs, which increased following the Clayton Act. Columns (3) and (4) present
specifications based on alternative measures of the connectedness or influence of investment banks,
namely whether their partners had more than 4 seats on railroads boards, or whether they held
board seats on 3 or more railroads, both of which would put the firms in the 75th percentile of the
distribution of investment banks with railroad board seats. In both cases, the results indicate that
railroads with directors from the best-connected investment banks benefitted about the same as
railroads with directors from less-connected investment banks. Finally, the regressions in column
(5) investigate whether the effect is purely coming through underwriting, with a specification that
includes an indicator variable for whether we could find evidence that the investment bank on
a railroad’s board actually underwrote railroad securities (from any railroad) during our sample
period.43 Some bankers may simply have provided expertise or acted on behalf of client shareholders
to protect their interests, rather than acting as underwriters for the firm. Interestingly, the results
suggest that even when the investment bank was not an active underwriter, the railroad enjoyed
lower interest rates.
Results for a second set of regressions related to Hypothesis 1 are presented in table 6. These
regressions test whether railroads with bankers on their boards saw their level of indebtedness
fall following the Clayton Act, which would be consistent with either bankers forcing their client
firms to issue more debt, or with bankers simply facilitating greater access to capital markets. In
general, these results are ambiguous. Column (1), which presents the most basic specification,
indicates that leverage did fall after the Clayton Act, but the estimated effect is quite small and
statistically insignificant. In Column (2), the specification with the J.P. Morgan & Company
indicator, shows that the Morgan firms saw their level of indebtedness fall substantially, but clients
of other investment banks did not, or even saw their indebtedness rise, although the effect is
43
This is measured with significant error. There is no systematic data on underwriting for the pre-1920s period.
However, some volumes of Moody’s Manuals in the late 1920s list the original underwriters for outstanding debt
securities, many of which were issued before 1920. This variable is equal to 1 for any investment bank that was found
to have underwritten debt for any railroad in 1920 or before. Thus it does not necessarily record whether the bank
underwrote for that particular railroad (this could not be ascertained for many railroads), but rather whether the
bank could have underwritten securities for that railroad.
27
28
1,643
Yes
Yes
-0.001*
(0.001)
0.040**
(0.014)
-0.147
(0.219)
-0.059**
(0.011)
0.009
(0.006)
(2)
1,643
Yes
Yes
-0.001*
(0.001)
0.039**
(0.014)
-0.135
(0.221)
-0.006
(0.007)
-0.005
(0.008)
(3)
0.85
0.85
0.85
denote significance at 1%, 5%, and 10%, respectively.
1,643
Observations
+
Yes
Firm fixed effects
R-squared
Note: Robust standard errors in parentheses. **, *, and
Yes
-0.001*
(0.001)
0.039**
(0.014)
-0.135
(0.221)
-0.005
(0.006)
Year fixed effects
Constant
Log(lagged assets)
Firm age
Investment Bank not underwriter on board × Post-Clayton Act
Investment bank underwriter on board × Post-Clayton Act
Investment bank w/ seats on 2 or fewer RRs × Post-Clayton Act
Investment bank w/ seats on 3 or more RRs × Post-Clayton Act
Investment bank w/ 3 or fewer seats × Post-Clayton Act
Investment bank w/ 4 or more RR board seats × Post-Clayton Act
Investment banker other than Morgan × Post-Clayton Act
JP Morgan partner on board × Post-Clayton Act
Investment banker on board × Post-Clayton Act
(1)
Table 6:
Regressions: Leverage
0.85
1,643
Yes
Yes
-0.001**
(0.001)
0.042**
(0.014)
-0.179
(0.223)
-0.011+
(0.007)
0.016+
(0.009)
(4)
0.85
1,643
Yes
Yes
-0.001*
(0.001)
0.040**
(0.014)
-0.15
(0.222)
-0.002
(0.007)
-0.010
(0.008)
(5)
imprecisely estimated. The results of most of the other columns in the table are similarly ambiguous,
mostly showing a negative effect of the Clayton Act on leverage, with some estimates going in the
opposite direction. In general, we do not find conclusive evidence of the effect of banker-directors
on the indebtedness of firms.
Hypothesis 2 held that banker-directors facilitated collusion among railroads, which would imply that they should have increased operating profitability. The results of regressions testing for
this effect are presented in table 7, where the dependent variable is the inverse of the railroads’
operating ratio. Once again, a fall in the ratio after the Clayton Act could be consistent with reduced anticompetitive behavior, but it could also be consistent with reduced efficiency of railroads’
operations, if the bankers brought managerial expertise to their client firms.
All of the estimates in the table are negative, indicating a fall in operating profitability after
the Clayton Act for railroads with bankers on their boards relative to those that did not, although
the magnitude and precision of the estimate varies somewhat across specifications. To provide
insight into whether this should be interpreted as a collusion effect or an efficiency effect, column
(4) includes an indicator variable for whether a railroad’s investment bank held seats on at least
3 different railroads in total. One would expect that bankers with representation on the boards of
large numbers of railroads would have the greatest opportunities to facilitate collusion. And yet
the result is the opposite: bankers with seats on only 2 or fewer railroads’ boards had a much larger
impact on their railroads’ profitability. This is at least suggestive evidence against the notion that
the effect is coming through collusion.
Hypothesis 3 posited that banker-directors reduced financial constraints on railroads, which
would have been reflected in lower investment to cash flow sensitivities. Results of regressions that
estimate this effect are presented in table 8.44 Although in general the results vary somewhat from
specification to specification in their magnitudes, the direction is generally the same. The estimated
selection effect—that is, the difference between firms with bankers on their boards compared others
in the pre-Clayton Act period—is always negative, indicating that prior to the Act, the investments
of firms with bankers on their boards were less sensitive to their cash flows compared to other firms.
But the sign of the coefficient on the interaction with the post-Clayton Act period is generally
44
The reporting of Property, Plant and Equipment, which we use to determine the level of investments in physical
capital, is often lumped with investments in other types of assets. We omit observations that suffer this problem.
Results are similar, albeit noisier, when we include them.
29
30
1,565
Yes
Yes
-0.001
(0.001)
0.009
(0.016)
1.355**
(0.257)
-0.037*
(0.018)
-0.032**
(0.011)
(2)
1,565
Yes
Yes
-0.001
(0.001)
0.009
(0.016)
1.359**
(0.257)
-0.029*
(0.013)
-0.039**
(0.013)
(3)
0.69
0.69
0.69
denote significance at 1%, 5%, and 10%, respectively.
1,565
Observations
+
Yes
Firm fixed effects
R-squared
Note: Robust standard errors in parentheses. **, *, and
Yes
-0.001
(0.001)
0.009
(0.016)
1.356**
(0.257)
-0.033**
(0.011)
Year fixed effects
Constant
Log(lagged assets)
Firm age
Investment Bank not underwriter on board × Post-Clayton Act
Investment bank underwriter on board × Post-Clayton Act
Investment bank w/ seats on 2 or fewer RRs × Post-Clayton Act
Investment bank w/ seats on 3 or more RRs × Post-Clayton Act
Investment bank w/ 3 or fewer seats × Post-Clayton Act
Investment bank w/ 4 or more RR board seats × Post-Clayton Act
Investment banker other than Morgan × Post-Clayton Act
JP Morgan partner on board × Post-Clayton Act
Investment banker on board × Post-Clayton Act
(1)
Table 7:
Regressions: Mark-ups
0.70
1,565
Yes
Yes
0.000
(0.001)
0.001
(0.016)
1.468**
(0.259)
-0.018
(0.012)
-0.087**
(0.015)
(4)
0.69
1,565
Yes
Yes
-0.001
(0.001)
0.005
(0.016)
1.403**
(0.258)
-0.044**
(0.013)
-0.018
(0.012)
(5)
positive, reflecting increased sensitivity of investments to cash flows. This is suggestive evidence
that investment-banker directors helped ease firms’ financial constraints when they were allowed
to be on boards and therefore act as the bankers of the firms at the same time.
Were all investment banks able to provide this benefit to their railroads? The results in columns
(2) - (4) suggest the magnitude of the effect varied significantly between better connected, and less
connected, investment banks. J.P. Morgan & Company (column (2)), firms with many board seats
(column (3)), and firms with representation on many railroads (column (4)), all provided a much
stronger benefit compared to other investment banks. In column (5), we test whether investment
banks that actually underwrote railroads’ securities provided a stronger benefit, and the results
indicate that they indeed did: the clients of banks that did not underwrite securities did not see
a rise in the sensitivity of their investments to their cash flows as the banking relationship with
financier-directors were severed.
Finally, we can assess the net effect of the banker-directors on their client railroads by estimating
the impact of the Clayton Act on market valuations. Results for regressions for Tobin’s Q are
presented in table 9. The estimates provide consistent evidence that banker-directors raised their
client firms’ valuations, since in the post-Clayton Act period Q fell for railroads with bankers on
their boards. Although the results in column (1) indicate that the average effect of the Act was
negative, results in subsequent columns indicate that the effect was driven by prominent, or better
connected firms, and underwriters of railroad securities, and was much smaller for railroads that
only had ties to bankers from less connected investment partnerships.
6
Alternative Specifications and Robustness of the Results
The results presented in the preceding tables provide some evidence against the hypothesis, motivated by progressive-era critiques of financiers, that banker-directors harmed their railroads. In this
section, we explore these results further by estimating alternative specifications of our regressions.
The Clayton Act was passed by Congress in 1914, at a time when the influence of bankers within
the economy was the subject of intense debate and criticism. Although the provisions relative to
railroads were not implemented until 1921, it is possible that the relationships between bankers
and their client firms were already affected by the Act in 1914, perhaps in anticipation of its
31
Table 8:
Regressions: Investment to cash flow sensitivity
(continued on next page)
(1)
(2)
(3)
(4)
(5)
Cash flows
-0.040
(0.326)
-0.035
(0.352)
-0.082
(0.341)
0.000
(0.344)
-0.060
(0.336)
Investment banker on board × Cash flows
-0.457
(0.341)
0.376+
(0.194)
Investment banker × Cash flows × Post-Clayton Act
-0.813+
(0.454)
0.850+
(0.463)
-0.104
(0.393)
0.114
(0.252)
JP Morgan partner on board × Cash flows
JP Morgan partner × Cash flows × Post-Clayton Act
Banker other than Morgan on board × Cash flows
Banker other than Morgan × Cash flows × Post-Clayton Act
Investment bank w/ 4+ RR board seats × Cash flows
-0.524
(0.370)
0.441+
(0.253)
-0.28
(0.409)
0.219
(0.284)
Investment bank w/ 4+ RR seats × Cash flows × Post-Clayton
Investment bank w/ 3 or fewer seats × Cash flows
Investment bank w/ 3 or fewer × Cash flows × Post-Clayton
Investment bank w/ seats on 3+ RRs × Cash flows
-0.442
(0.358)
0.414+
(0.237)
-0.519
(0.552)
0.255
(0.370)
Investment bank w/ 3+ RRs × Cash flows × Post-Clayton Act
Investment bank w/ seats on 2 or fewer RRs × Cash flows
Investment bank w/ 2 RRs × Cash flows × Post-Clayton Act
-0.668+
(0.364)
0.564*
(0.234)
0.181
(0.482)
-0.324
(0.384)
Investment bank underwriter on board × Cash flows
Investment bank underwriter × Cash flows × Post-Clayton Act
Investment bank not underwriter on board × Cash flows
Investment bank not underwriter × Cash flows × Post-Clayton
32
Table 8: Investment to cash flow sensitivity, continued
(1)
(2)
(3)
(4)
(5)
Qt−1
-0.018
(0.064)
-0.029
(0.065)
-0.020
(0.064)
-0.017
(0.064)
-0.014
(0.064)
Investment banker on board × Qt−1
0.015
(0.058)
0.005
(0.010)
Investment banker × Qt−1 × Post-Clayton Act
JP Morgan partner on board × Qt−1
0.045
(0.071)
0.016
(0.011)
-0.049
(0.061)
-0.045
(0.028)
JP Morgan partner × Qt−1 × Post-Clayton Act
Banker other than Morgan on board × Qt−1
Banker other than Morgan × Qt−1 × Post-Clayton Act
Investment bank w/ 4+ RR board seats × Qt−1
0.025
(0.062)
0.002
(0.013)
-0.027
(0.070)
0.012
(0.013)
Investment bank w/ 4+ RR seats × Qt−1 × Post-Clayton
Investment bank w/ 3 or fewer seats × Qt−1
Investment bank w/ 3 or fewer × Qt−1 × Post-Clayton
Investment bank w/ seats on 3+ RRs × Qt−1
0.015
(0.059)
0.005
(0.012)
0.019
(0.108)
0.003
(0.020)
Investment bank w/ 3+ RRs × Qt−1 × Post-Clayton Act
Investment bank w/ seats on 2 or fewer RRs × Qt−1
Investment bank w/ 2 RRs × Qt−1 × Post-Clayton Act
Investment bank underwriter on board × Qt−1
0.016
(0.058)
-0.002
(0.011)
0.005
(0.084)
0.033+
(0.017)
Investment bank underwriter × Qt−1 × Post-Clayton Act
Investment bank not underwriter on board × Qt−1
Investment bank not underwriter × Qt−1 × Post-Clayton
Cash flows × Qt−1
60.011
(65.391)
1.405
(4.806)
0.000
(0.001)
-0.006
(0.013)
0.107
(0.216)
87.084
(68.136)
-0.196
(5.247)
0.000
(0.001)
-0.001
(0.013)
0.013
(0.215)
57.082
(67.777)
1.902
(5.064)
0.000
(0.001)
-0.006
(0.014)
0.136
(0.229)
61.996
(66.868)
0.986
(5.020)
0.000
(0.001)
-0.008
(0.014)
0.132
(0.221)
60.067
(64.144)
1.532
(4.787)
0.000
(0.001)
-0.001
(0.013)
0.014
(0.220)
Year fixed effects
Yes
Yes
Yes
Yes
Yes
Firm fixed effects
Yes
Yes
Yes
Yes
Yes
Observations
929
929
929
929
929
Cash flows × Log(lagged assets)
Firm age
Log(Lagged assets)
Constant
R-squared
Note: Robust standard errors in parentheses. **, *, and
+
0.18
0.20
0.18
0.18
denote significance at 1%, 5%, and 10%, respectively.
33
0.19
34
1,144
Yes
Yes
0.007**
(0.001)
-0.165**
(0.026)
3.624**
(0.406)
-0.115**
(0.020)
-0.007
(0.012)
(2)
1,144
Yes
Yes
0.006**
(0.001)
-0.161**
(0.025)
3.564**
(0.401)
-0.046**
(0.014)
-0.013
(0.012)
(3)
0.76
0.77
0.76
denote significance at 1%, 5%, and 10%, respectively.
1,144
Observations
+
Yes
Firm fixed effects
R-squared
Note: Robust standard errors in parentheses. **, *, and
Yes
0.006**
(0.001)
-0.162**
(0.025)
3.587**
(0.404)
-0.035**
(0.012)
Year fixed effects
Constant
Log(lagged assets)
Firm age
Investment Bank not underwriter on board × Post-Clayton Act
Investment bank underwriter on board × Post-Clayton Act
Investment bank w/ seats on 2 or fewer RRs × Post-Clayton Act
Investment bank w/ seats on 3 RRs × Post-Clayton Act
Investment bank w/ 3 or fewer seats × Post-Clayton Act
Investment bank w/ 4 RR board seats × Post-Clayton Act
Investment banker other than Morgan × Post-Clayton Act
JP Morgan partner on board × Post-Clayton Act
Investment banker on board × Post-Clayton Act
(1)
Table 9:
Regressions: Tobin’s Q
0.76
1,144
Yes
Yes
0.006**
(0.001)
-0.157**
(0.026)
3.519**
(0.408)
-0.040**
(0.013)
-0.011
(0.015)
(4)
0.76
1,144
Yes
Yes
0.006**
(0.001)
-0.164**
(0.026)
3.607**
(0.406)
-0.044**
(0.013)
-0.021
(0.014)
(5)
inevitable implementation. Moreover, bankers may even have changed the way they did business
with railroads, and perhaps curbed any “abuses,” in an effort to forestall the implementation of
the act itself.
It is certainly clear that the firm that attracted the most scrutiny and attention, JP Morgan &
Company, responded to the Clayton Act well before it was implemented for railroads. In 1913, the
year before the act was implemented, the partners of JP Morgan held board seats on 12 railroads,
but they began to resign from some of those seats in 1914, and the bank also ended the two voting
trust agreements they had in place in that year.45 By 1920 they remained on only 8 of those
railroads, and they had left 4 more by 1923. Although other firms did not seem to follow this
pattern—the percentage of railroads with an investment banker on their board remained roughly
stable over these years—it is possible that the effects of the act may have been felt beginning in
1914, rather than 1921. To the extent that the period 1914-21 represented a weaker treatment,
one might expect the results of regressions estimated around that point to have somewhat weaker
results than those presented above.
Table 10 presents results of regressions similar to the specification of equation (1), only the
Bankeronboardi variable is redefined to be equal to one for all years for railroads that had a banker
on their board in 1913, rather than in 1920, and the interaction applied to this variable is for
the period after 1914, rather than 1921. This specification is then estimated for a sample ending
in 1920. Thus, it captures the effect of the passage of the Clayton Act itself, rather than the
implementation of the Act.
The results are much weaker than those for 1921. Compared to the results reported in table 5,
the overall effect on the interest rate is much smaller. However, column (2) indicates a significant
effect for Morgan-affiliated firms. This may be due to resignations of the Morgan partners from
some of their client firms, or any other changes made by JP Morgan in response to the political
criticisms of their firm’s power. The results for leverage, in columns (3) and (4), are essentially zero.
On the other hand, the effects on the mark-up (columns (5) and (6)) are statistically significant,
although the magnitude of the effect is much smaller than the one found for 1921. For Tobin’s
Q, column (7) reports essentially no effect overall. But column (8) shows a large, negative effect
45
These were for the Chicago Great Western, and the Southern Railroads, and had been in place since JP Morgan & Company had become affiliated with those railroads. In voting trust arrangements, the shareholders are
disenfranchised and trustees exercise sole voting power.
35
36
yes
916
Firm fixed effects
Observations
R-squared
0.56
Note: Robust standard errors in parentheses. **, *, and
yes
+
0.074
(0.054)
0.000
0.000
-0.002
(0.003)
0.001
(0.001)
Year fixed effects
Constant
Log(lagged assets)
Firm age
Investment banker other than Morgan × Post-1914
JP Morgan partner on board × Post-1914
Investment banker on board × Post-1914
979
yes
yes
0.104
(0.339)
0.000
(0.001)
0.022
(0.021)
-0.005
(0.007)
979
yes
yes
0.135
(0.346)
0.000
(0.001)
0.02
(0.021)
-0.017
(0.011)
0.001
(0.008)
938
yes
yes
2.033**
(0.423)
0.001
(0.001)
-0.036
(0.026)
-0.025+
(0.013)
0.73
938
yes
yes
2.034**
(0.425)
0.001
(0.001)
-0.036
(0.026)
-0.026
(0.019)
-0.025+
(0.014)
Dependent Variable:
Leverage
Mark-up
(3)
(4)
(5)
(6)
0.57
0.88
0.88
0.73
denote significance at 1%, 5%, and 10%, respectively.
916
yes
yes
0.067
(0.052)
0.000
0.000
-0.002
(0.003)
0.006**
(0.002)
-0.001
(0.001)
Interest rate
(1)
(2)
Table 10:
Regressions: Effect of Clayton Act after 1914
0.82
679
yes
yes
4.232**
(0.583)
0.00788
(0.001)
-0.204**
(0.036)
(8)
0.83
679
yes
yes
4.394**
(0.571)
0.008**
(0.001)
-0.217**
(0.035)
-0.040+
(0.022)
0.047**
(0.018)
Tobin’s Q
0.014
(0.018)
(7)
for firms affiliated with JP Morgan, and, somewhat puzzlingly, a positive effect on firms affiliated
with other bankers. Overall, these results indicate that the Clayton Act had a modest effect on
our variables of interest since it was enacted until it was implemented, but this impact was mostly
specific to firms affiliated with JP Morgan.
A further concern that one might raise with these results is the possibility that they may be
due to underlying differential trends across groups of railroads in our data. That is, by comparing
the outcomes before and after the Clayton Act for groups of railroads whose outcomes were already
changing over time, we may be ascribing to the Clayton Act effects that may instead be due
to differential trends among treatment and control groups attributable to a different underlying
process. Since bankers were typically affiliated with railroads with particular characteristics—table
4 shows that they were more levered, for example—it is possible that firms with those characteristics
performed differently over time. We employ two approaches to address these concerns.
First, we create a placebo “Clayton Act” for the year 1909, and examine whether the firms
we have designated as our treatment group (that is, those with bankers on their boards in 1920)
were affected by this placebo treatment. We end the period of the sample first in 1913, before
the Clayton Act was passed, and then, in an alternative specification, in 1920, right before it was
implemented. Although the latter specification may include some effect of the Act itself, the longer
time span presents a greater opportunity to pick up the effects of any underlying differential trends
between our treatment and control groups in the pre-Clayton Act period.
The results in Panel A, for the 1905-13 sample, show essentially no effect of the placebo treatment. In general, the estimates are much smaller than those reported above, and not statistically
significant. The one exception is the effect on interest rates for railroads affiliated with bankers
other than J.P. Morgan, which, while small, is indeed significant. Results in Panel B (using the
1905-20 sample) generally consistent, although the overall interest rate effect becomes significant,
and the effect on the mark-up becomes significant as well. Still, it should be stressed that these
were obtained from a period that potentially encompasses the effects of the Clayton Act itself. In
sum, although the results generally do not indicate that our findings are driven only by differential
trends between our treatment and control groups in the pre-Clayton Act period, there is at least
some possibility of this in a few of our outcome variables.
Finally, we investigate the role of any trends in our data using specifications that explicitly
37
38
yes
yes
502
0.67
-0.030
(0.084)
0.000
(0.000)
0.004
(0.005)
-0.001
(0.001)
0.003*
(0.001)
0.022
(0.054)
0.000
(0.000)
0.001
(0.003)
0.002+
(0.001)
0.022
(0.054)
0.000
(0.000)
0.001
(0.003)
0.002
(0.002)
0.002
(0.001)
Interest rate
(9)
(10)
yes
yes
502
0.66
-0.038
(0.085)
0.000
(0.000)
0.005
(0.005)
0.002
(0.001)
-0.001
(0.001)
0.026
(0.018)
0.028
(0.290)
-0.031+
(0.295)
-0.035*
(0.018)
0.009
(0.011)
-0.001
(0.001)
0.031
(0.018)
-0.003
(0.010)
2.348**
(0.399)
0.002+
(0.001)
-0.055*
(0.025)
-0.034*
(0.016)
yes
yes
958
0.74
2.342**
(0.399)
0.002+
(0.001)
-0.054*
(0.025)
-0.029
(0.026)
-0.035*
(0.017)
yes
yes
684
0.82
4.314**
(0.594)
0.007**
(0.001)
-0.209**
(0.037)
(8)
yes
yes
684
0.82
4.293**
(0.602)
0.007**
(0.001)
-0.207**
(0.037)
0.004
(0.025)
-0.004
(0.017)
(16)
yes
yes
351
0.95
2.984**
(0.951)
0.006**
(0.002)
-0.141*
(0.058)
0.020
(0.023)
-0.022
(0.014)
Tobin’s Q
-0.002
(0.017)
(15)
3.185**
(0.937)
0.006**
(0.002)
-0.154**
(0.057)
Panel B: Sample period 1905-1913
Dependent Variable:
Leverage
Mark-up
(11)
(12)
(13)
(44)
1.127+
(0.667)
1.117+
(0.666)
yes
yes
516
0.71
0.000
(0.002)
0.019
(0.041)
0.000
(0.002)
0.020
(0.041)
-0.016
(0.023)
-0.008
(0.017)
Tobin’s Q
-0.010
(0.014)
yes
yes
351
0.95
yes
yes
534
0.91
-0.381
(0.749)
-0.001
(0.002)
0.051
(0.046)
-0.012
(0.015)
0.009
(0.009)
-0.009
(0.016)
(7)
yes
yes
516
0.71
yes
yes
534
0.91
-0.420
(0.746)
-0.001
(0.002)
0.055
(0.046)
0.003
(0.008)
Year fixed effects
yes
yes
yes
yes
yes
Firm fixed effects
yes
yes
yes
yes
yes
Observations
936
936
1,002
1,002
958
R-squared
0.54
0.54
0.87
0.87
0.74
Note: Robust standard errors in parentheses. **, *, and + denote significance at 1%, 5%, and 10%, respectively.
Constant
Log(lagged assets)
Firm age
Investment banker other than Morgan × Post-1909
JP Morgan partner on board × Post-1909
Investment banker on board × Post-1909
Year fixed effects
Firm fixed effects
Observations
R-squared
Constant
Log(lagged assets)
Firm age
Investment banker other than Morgan × Post-1909
JP Morgan partner on board × Post-1909
Investment banker on board × Post-1909
Interest rate
(1)
(2)
Panel A: Sample period 1905-1920
Dependent Variable:
Leverage
Mark-up
(3)
(4)
(5)
(6)
Table 11:
Regressions: Effect of a “Placebo” Clayton Act imposed in 1909
estimate differential trends between railroads with and without bankers. The effect of the Act
itself was unlikely to create a sharp, discrete change in many of our outcome variables, but rather
would likely have been felt slowly, as firms gradually replaced maturing long-term debt, as bankerdirectors gradually resigned, and as the ties between banks and their client railroads weakened.
We therefore test for an effect of the Clayton Act in our data by investigating whether there is a a
break in any underlying trend around the Act.
Before proceeding with the estimations, we first illustrate these issues in the data. Figure 2
illustrates the differences-in-differences between firms with and without bankers on their boards
over time (relative to an excluded year, 1903).46 The line in the upper panel, which presents the
pattern for interest rates, exhibits little trend prior to 1914, but then seems to turn upward around
that year, and certainly continues to trend upward after 1921.47 The lower panel, on the other hand,
shows that the differences in the markup exhibits a downward trend for most years in the sample,
with no obvious breaks close to the years in which the Clayton Act was passed or implemented. It
is likely that the results presented above for the change in the markup may simply be a reflection
of this trend.
In order to test whether the Clayton Act actually changed any pre-existing trends in the differences between firms with and without bankers, we estimate the following regression:
rit = αi + γt + βXit + π1 Bankeronboardi × yeart + π2 Bankeronboardi × (year − Clayton)t + ²it ,
where again Bankeronboardi is equal to one for all years in the sample for firms that had a banker
on their board in 1920; yeart is simply the year, which captures any underlying trend in the data;
and (year − Clayton)t , which is zero for all years prior to the Clayton Act, simply counts years
after either 1914 or, in alternative specifications, 1921. In this framework, π2 captures any break
in the trend after the Clayton Act was implemented.
Results for these regressions are reported in table 12. Column (1) presents the results for the
interest rate variable. Perhaps not surprisingly from the figure, this variable exhibits no trend
prior to 1914, and then subsequently there is a positive change in trend after that year which is
46
The figures plot the annual
P difference-in-differences between firms with and without bankers from the following
regression: rit = αi + γt + 1920
t=1902 λt Bankeronboardi × yeart + ²it , which contains firm and year fixed effects.
47
Standard errors are not presented in the figure, but only the last six data points are statistically significantly
different than the excluded year.
39
(a) Annual difference−in−differences: Interest rate
0.01
Clayton Act
Passed
0.009
Clayton Act
Implemented
0.008
0.007
0.006
0.005
0.004
0.003
0.002
0.001
0
1905
1910
1915
1920
1925
1930
(b) Annual difference−in−differences: Mark−ups
0.1
Clayton Act
Passed
0.08
Clayton Act
Implemented
0.06
0.04
0.02
0
−0.02
−0.04
1905
1910
1915
1920
1925
1930
Figure 2: Differences in differences over time
Student Version of MATLAB
40
41
1559
1559
yes
yes
yes
8.3747
(5.1615)
-0.0010+
(0.006)
0.0134
(0.0167)
yes
0.76
1097
yes
yes
0.4797
(6.5225)
10.8762+
(5.9499)
yes
0.0052**
(0.0010)
-0.1362**
(0.0246)
(8)
0.77
1097
yes
yes
-5.3983
(6.5154)
0.0783
(0.0589)
-0.1678**
(0.0259)
0.0011
(0.0039)
-0.0095+
(0.0056)
0.0045
(0.0032)
-0.0043
(0.0042)
Tobin’s Q
0.0027
(0.0032)
-0.0059
(0.0043)
(7)
-0.0674
(0.0753)
0.0053
(0.0172)
0.69
1473
yes
yes
-7.0140
(4.5085)
-0.0688
(0.0591)
0.0182
(0.0147)
-0.0012
(0.0030)
-0.0023
(0.0042)
-0.0040
(0.0025)
0.0015
(0.0033)
R-squared
0.50
0.52
0.86
0.85
0.69
Note: Robust standard errors in parentheses. **, *, and + denote significance at 1%, 5%, and 10%, respectively.
1473
Observations
yes
yes
-2.3665
(3.1790)
-0.0004
(0.0005)
0.0194
(0.0143)
0.0014
(0.0024)
-0.0061*
(0.0004)
0.0012
(0.0018)
-0.0030
(0.0021)
-0.0031
(0.0023)
0.0003
(0.0031)
Dependent Variable:
Mark-up
(4)
(5)
(6)
1493
yes
Firm fixed effects
yes
2.2823
(0.5469)
-0.0241**
(0.0057)
-0.0001
(0.0019)
0.0006*
(0.0003)
-0.0003
(0.0004)
-0.0005*
(0.0002)
0.0012**
(0.0003)
0.0012
(0.0016)
-0.0039*
(0.0019)
(3)
Leverage
1493
yes
0.2796
(0.4180)
0.0001
(0.0001)
-0.0009
(0.0019)
-0.0001
(0.0002)
0.0007*
(0.0003)
Year fixed effects
Constant
Log(lagged assets)
Firm age
Investment banker other than Morgan × (year-1914)
Investment banker other than Morgan × year
JP Morgan partner on board × (year-1914)
JP Morgan partner on board × year
Investment banker on board × (year-1914)
Investment banker on board × year
Interest rate
(1)
(2)
Table 12:
Regressions: Break-in-trend after the Clayton Act
statistically significant. In column (2), separate trends are estimated for Morgan-affiliated firms
and firms affiliated with other investment banks, and these groups of firms actually do exhibit
trends in the pre-1914 period. A very large and positive change in trend is found for the postClayton Act period, but only for the firms affiliated with bankers other than Morgan. Columns
(3) and (4) present results for leverage. Overall there is a positive change in trend after 1914, but
this effect is reversed when firms affiliated with Morgan and other banks are estimated separately.
Columns (5) and (6) present results for the mark-up variable. Consistent with the apparent trend
exhibited in the figure, there is no break in trend after the Clayton Act. Thus, we find no evidence
to support the view that banker-directors enabled railroads to engage in anti-competitive practices.
Finally, for Tobin’s Q, there are negative changes in trend overall after the Clayton Act, but this
is imprecisely estimated. The results in column (8), however, show that for firms affiliated with
JP Morgan & Company, there is a negative shift in the trend after the Clayton Act. (Alternative
specifications based on a break in 1921 exhibit substantially similar results.)
Overall the main conclusions that follow from this exercise are as follows. Although the markup,
our indicator for collusion and monopoly profits, was lower after the Clayton Act for firms affiliated
with bankers, this was almost certainly due to pre-existing trends for those firms. Thus we find
no evidence to support the view that bankers facilitated collusion among railroads. Secondly, the
increase in interest rates experienced by railroads affiliated with bankers after the Clayton Act was
not due to some preexisting trend. Finally, it is likely that the fall in valuation (Q) for railroads
affiliated with bankers following the Clayton Act was not simply due to some preexisting trend,
but the data are simply to noisy to estimate that effect precisely.
7
Discussion and Conclusion
The role of financiers in the economy in the early twentieth century was extraordinarily contentious.
Unlike today, when banker participation in the management of nonfinancial companies is relatively
rare, financiers were commonly represented on the boards of directors of large corporations at that
time. Progressives and labor activists saw a vast conspiracy among New York financiers—a “money
trust”—that used board seats to control American business. But many officials in these firms, and
many financial historians since then, have argued that the influence of these bankers on their client
42
firms was actually beneficial.
This paper presents the first comprehensive evidence of the presence of bankers on the boards
of American railroads in the early twentieth century. Moreover, it expands on the current and
historical literature on the role of bankers on boards by using the Clayton Antitrust Act, implemented in 1921, to estimate the causal effects of bankers on their client railroads’ operations. The
results suggest that investment bankers benefited railroads’ operations when they participated in
railroad management, despite the obvious opportunities for “self-dealing” or tunneling created by
their positions. In the years following the implementation of the Act, the valuations and operating
profitability of railroads with investment bankers on their boards fell relative to other roads. These
railroads also experienced an increase in their borrowing costs and in the sensitivity of their investments to current cash flows, suggesting that a close relationship with banks may have facilitated
improved access to financial markets.
There are different ways in which the presence of bankers on boards may have helped ease
financial constraints. Bankers may have used their superior information about the firm to decide
the optimal financing policies. It is also possible that the mere presence of a banker (or a wellconnected banker) on a board served as a seal of approval for uninformed outside investors. Or
the bankers may have simply provided financial advice. While we cannot separately identify these
channels, the strong results we obtain for banks that underwrote securities are at least suggestive
of a direct impact of bankers on firms’ access to capital. Regardless of the mechanism, these results
indicate that the power of bankers in the era of finance capitalism may not always have been wielded
in ways that harmed the enterprises they managed.
Our findings raise many questions about the nature and evolution of relationships between
financial and nonfinancial firms, and the effects that the organization of financial markets had on
the real economy. The presence of bankers on railroads’ boards (and also on the boards of industrial
corporations) in the first three decades of the twentieth century was much more extensive than the
in modern corporate America. The U.S. market system of corporate governance, which is often
heralded as an advantageous way to protect investors’ rights and foster growth, was not in place
during the era of finance capitalism. A better understanding of how bankers first rose to prominence,
and why their roles declined over time may provide important insights to the process of economic
development in American history.
43
Bibliography
Agarwal, Rajshree, and Julie Ann Elston. 2001. “Bank-Firm Relationships, Financing and Firm
Performance in Germany,” Economic Letters, 72, 225-232.
Allen, Franklin, and Douglas Gale. 2000. Comparing Financial Systems. Cambridge: MIT Press.
Atack, Jeremy. 1985. “Industrial Structure and the Emergence of the Modern Industrial Corporation,” Explorations in Economic History, 22 (January), 29-52.
Balke, Nathan, and Robert Gordon. 1989. “The Estimation of Prewar Gross National Product:
Methodology and New Evidence,” Journal of Political Economy, 97, 38-92.
Baskin, Jonathan B. 1988. “The Development of Corporate Financial Markets in Britain and the
United States, 1600-1914: Overcoming Asymmetric Information,” Business History Review,
62: 199-237.
Baskin, Jonathan B., and Paul J. Miranti. 1997. A History of Corporate Finance. New York:
Cambridge University Press.
Becht, Marco, and J. Bradford DeLong. 2005. “Why Has there Been So Little Block Holding in
America?” in Mørck, ed., A History of Corporate Governance Around the World. Chicago:
University of Chicago.
Berle, Adolf, and Gardiner Means. 1932. The Modern Corporation and Private Property. New
York: McMillan.
Bernanke, Ben S. 1983. “Nonmonetary Effects of the Financial Crisis in the Propagation of the
Great Depression,” American Economic Review, 73(3), 257-76.
Bernstein, Assaf, Hughson, Eric, and Marc Weidenmier. 2010. “Identifying the Effects of a Lender
of Last Resort on Financial Markets: Lessons from the Founding of the Fed.” Forthcoming,
Journal of Financial Economics.
Benmelech, Efraim. 2008. “Asset Salability and Debt Maturity: Evidence from NineteenthCentury American Railroads,” Review of Financial Studies, 22(4), 1545-84.
Booth, James R., and Daniel N. Deli. 1999. “On Executives of Financial Institutions as Outside
Directors,” Journal of Corporate Finance, 5, 227-250.
Brandeis, Louis D. 1914. Other Peoples’ Money and How the Bankers Use It. New York: Frederick
A. Stokes.
Bruner, Robert F., and Sean D. Carr. 2007. The Panic of 1907: Lessons Learned from the
Market’s Perfect Storm. Hoboken, NJ: John Wiley & Sons.
Byrd, Daniel T., and Mark S. Mizruchi. 2005. “Bankers on the Board and the Debt Ratio of
Firms,” Journal of Corporate Finance, 11, 129-73.
Calomiris, Charles W. 1995. “The Costs of Rejecting Universal Banking: American Finance in the
German Mirror, 1870-1914.” In Lamoreaux and Raff, eds., Coordination and Information:
Historical Perspectives on the Organization of Enterprise. Chicago: University of Chicago
Press.
44
Calomiris, Charles W., and Gary Gorton. 2000. “The origins of banking panics: Models, facts and
bank regulation,” in Calomiris, Charles W., U.S. Bank Regulation in Historical Perspective.
New York: Cambridge University Press.
Calomiris, Charles W., and Joseph R. Mason. 2003. “Consequences of Bank Distress during the
Great Depression,” American Economic Review, 93(3), 937-47.
Cantillo Simon, Miguel. 1998. “The Rise and Fall of Bank Control in the United States: 18901939,” American Economic Review, 88 (5), 1077-93.
Carosso, Vincent P. 1970. Investment Banking in America: A History. Cambridge: Harvard
University Press.
——. 1987. The Morgans: Private International Bankers, 1854-1913. Cambridge: Harvard
University Press.
Carosso, Vincent P., and Richard Sylla. 1991. “U.S. Banks in International Finance,” in Cameron
and Bovykin, eds., International Banking, 1870-1914. New York: Oxford University Press.
Chandler, Alfred. 1977. The Visible Hand: The Managerial Revolution in American Business.
Cambridge: Harvard University Press.
——. 1990. Scale and Scope: The Dynamics of Industrial Capitalism.
vard/Belknap.
Cambridge: Har-
Chernow, Ron. 1990. The House of Morgan. New York: Grove Press.
Cole, Harold L., and Lee E. Ohanian. 2000. “Re-Examining the Contributions of Money and
Banking Shocks to the U.S. Great Depression,” in Bernanke and Rogoff, eds., NBER Macroeconomics Annual. Cambridge: MIT Press.
Davis, Lance E. 1966. “The Capital Markets and Industrial Concerns: The U.S. and U.K., a
Comparative Study,” Economic History Review, 191, 255-72.
DeLong, J. Bradford. 1991. “Did J.P. Morgan’s Men Add Value? An Economist’s Perspective on
Finance Capitalism,” in Temin, ed. Inside the Business Enterprise: Historical Perspectives
on the Use of Information. Chicago: University of Chicago Press.
Dewing, Arthur S. 1914. Corporate Promotions and Reorganizations. Cambridge: Harvard University Press.
Directory of Directors Company. 1905-. Directory of Directors in the City of New York. New
York: The Company.
Dooley, Peter. 1969. “The Interlocking Directorate,” American Economic Review, 16(1), 26-38.
Edwards, George W. 1938. The Evolution of Finance Capitalism. New York: Longmans, Green
and Company.
Ellis, Charles D. 2008. The Partnership: The Making of Goldman Sachs. New York: Penguin.
Faulkner, Harold U. 1951. The Decline of Laissez Faire, 1897-1917. New York: Rinehart.
Federal Trade Commission. 1915-. Annual Report of the Federal Trade Commission. Washington:
Government Printing Office.
45
Fohlin, Caroline. 1998. “Relationship Banking, Liquidity, and Investment in the German Industrialization,” Journal of Finance, 53, 1737-1758.
Friedman, Milton, and Anna J. Schwartz. 1963. A Moneteary History of the United States,
1867-1970. Princeton: Princeton University Press.
Flandreau, Marc, Gaillard, Norbert, and Ugo Panizza. 2009. “Conflicts of Interest, Reputation,
and the Interwar Debt Crisis: Banksters or Bad Luck?.” Draft, Graduate Institute, Geneva.
Garraty, John. 1960. Right-Hand Man: The Life of George W. Perkins. New York: Harper &
Bros.
Geisst, Charles R. 2001. The Last Partnerships: Inside the Great Wall Street Money Dynasties.
New York: McGraw-Hill.
Goldsmith, Raymond W. 1958. Financial Intermediaries in the American Economy since 1900.
Princeton: Princeton University Press.
Gorton, Gary, and Frank Schmid. 2000. “Universal Banking and the Performance of German
Firms,” Journal of Finance, 95(1-2), 29-80.
Güner, A. Burak, Ulrike Malmendier, and Geoffrey Tate. 2008. “Financial Expertise of Directors,”
Journal of Financial Economics, 88 (2), 323-354.
Hannah, Leslie. 2007. “What did Morgan’s Men Really Do?” Draft, University of Tokyo.
Hawkins, David F. 1972. Financial Reporting Practices of Corporations. Homewood, Illinois:
Dow Jones-Irwin.
——. 1986. Corporate Financial Disclosure, 1900-1933. New York: Garland.
Hermalin, Ben and Michael Weisbach. 1988. “The Determinants of Board Composition,” RAND
Journal of Economics, 19, 589-606.
Hermalin, Ben and Michael Weisbach. 1998. “Endogenously Chosen Boards of Directors and
Their Monitoring of the CEO,” American Economic Review, 88, 96-118.
Hilt, Eric. 2008. “When did Ownership Separate from Control? Corporate Governance in the
Early Nineteenth Century.” Journal of Economic History, 68 (3): 645-685.
House Judiciary Committee. 1917. “Hearings to Amend Clayton Antitrust Act.” Serial 45, Part
2. Washington: Government Printing Office.
Huertas, Thomas, and Harold Cleveland. 1987. Citibank. Cambridge: Harvard University Press.
Interstate Commerce Commission. 1900-30. Annual Report of the Statistics of Railways of the
United States. Washington: Government Printing Office.
Kaplan, Steven, and Luigi Zingales. 1997. “Do Investment-Cash Flow Sensitivities Provide Useful
Measures of Financing Constraints?” Quarterly Journal of Economics, 112, 169-215.
Keller, Morton. 1963. The Life Insurance Enterprise, 1885-1910: A Study in the Limits of
Corporate Power. Cambridge: Belknap.
46
Khwaja, Asim Ijaz, and Atif Mian. 2008. “Tracing the Impact of Bank Liquidity Shocks: Evidence
from an Emerging Market,” American Economic Review, 98(4), 1413-42.
Kindleberger, Charles P. 1996. Manias, Panics, and Crashes: A History of Financial Crises. New
York: Wiley.
Kotz, David M. 1978. Bank Control of Large Corporations in the United States. Berkeley: University of California Press.
Kroszner, Randall S., and Raghuram G. Rajan. 1997. “Organization Structure and Credibility:
Evidence from Commercial Bank Securities Activities Before the Glass-Steagall Act,” Journal
of Monetary Economics, 39 (4), 475-516.
Kroszner, Randall S., and Philip E. Strahan. 2001. “Bankers on Boards: Monitoring, Conflicts of
Interest, and Lender Liability,” Journal of Financial Economics, 62 (3), 415-52.
Lamoreaux, Naomi R. 1985. The Great Merger Movement in American Business, 1895-1904. New
York: Cambridge University Press.
Lamoreaux, Naomi R. and Jean-Laurent Rosenthal. 2005. “Legal Regime and Business’s Organizational Choice: A Comparison of France and the United States During the Era of
Industrialization,” American Law and Economics Review, 7, 28-61.
—— and ——. 2006. “Corporate Governance and the Plight of Minority Shareholders in the
United States Before the Great Depression,” in Goldin and Glaeser, eds., Corruption and
Reform: Lesson’s from America’s Economic History. Chicago: University of Chicago Press.
Landstreet, J. Collins. 1984. “Antitrust Law - The Clayton Act - The Regulation of Interlocking
Directorates Between Competing Banks and Insurance Companies.” Tennessee Law Review,
51, 599-627.
La Porta, Rafael, Lopez-de-Silanez, Florencio, and Andrei Shleifer. 1998. “Corporate Ownership
Around the World,” Journal of Finance, 54(2): 471-517.
Mahoney, Paul G. 2003. “The Origins of Blue-Sky Laws: A Test of Competing Hypotheses,”
Journal of Law and Economics, 46: 229-51.
Martin, Albro. 1978. Enterprise denied: origins of the decline of American railroads, 1897-1917.
New York: Columbia University Press.
McCraw, Thomas K. 1984. Prophets of Regulation. Cambridge: Belknap Press.
Minton, Bernadette A., Jerome P. A. Taillard, and Rohan Williamson. 2010. ”Do Independence
and Financial Expertise of the Board Matter for Risk Taking and Performance?,” Working
Paper, The Ohio State University.
Miranti, Paul J. 1989. “The Mind’s Eye of Reform: The ICC’s Bureau of Statistics and Accounts
and a Vision of Regulation, 1887-1940,” Business History Review, 63(3), 469-509.
Mitchell, Lawrence E. 2007. The Speculation Economy: How Finance Triumphed Over Industry.
San Francisco: Berrett-Koehler.
Moen, Jon, and Ellis W. Tallman. 1992. “The Bank Panic of 1907: The Role of Trust Companies,”
Journal of Economic History, 52(3), 611-30.
47
Moody, John. 1900-. Moody’s Manual of Railroad and Corporation Securities. New York: Moody
Manual Company.
——. 1904. The Truth About the Trusts: A Description and Analysis of the American Trust
Movement. New York: Moody Publishing Co.
——. 1909-. Moody’s Analyses of Investments. New York: Moody’s Investor Service.
——. 1920. The Masters of Capital: A Chronicle of Wall Street. New Haven: Yale University
Press.
Morrison, Alan D., and William J. Wilhelm. 2007. Investment Banking: Institutions, Politics
and Law. Oxford: Oxford University Press.
—— and ——. 2008. “The Demise of Investment Banking Partnerships: Theory and Evidence.
Journal of Finance, 63 (February), 311-350.
Mørck, Randall, and Masao Nakamura. 1999. “Banks and Corporate Control in Japan,” Journal
of Finance 54, 319-39.
Moulton, Harold G. 1933. The American Transportation Problem. Washington: the Brookings
Institution
Odell, Kerry, and Marc D. Weidenmier. 2004. “Real Shock, Monetary Aftershocks: The San
Francisco Earthquake and the Panic of 1907,” Journal of Economic History, 64, 1002-1027.
Peach, William N. 1975. The Security Affiliates of National Banks. New York: Arno Press.
Philippon, Thomas, and Ariell Reshef. 2009. “Wages and Human Capital in the U.S. Financial
Industry, 1909-2006.” Working paper, NYU.
Pujo Committee. [U.S. Congress, House Committee on Banking and Currency.] 1913a. Money
Trust Investigation: Hearings Before the Committee Appointed Pursuant to House Resolutions 429 and 504 to Investigate the Concentration of Control of Money and Credit. Washington DC: Government Printing Office.
——. 1913b. Report of the Committee Appointed Pursuant to House Resolutions 429 and 504 to
Investigate the Concentration of Control of Money and Credit. Washington DC: Government
Printing Office.
Ramirez, Carlos D. 1995. “Did J.P. Morgan’s Men Add Liquidity? Corporate Investment, Cash
Flow, and Financial Structure at the Turn of the Twentieth Century,” Journal of Finance,
50(2), 661-78.
Ramirez, Carlos D., and Christian Eigen-Zucchi. 2001. “Understanding the Clayton Act of 1914:
An Analysis of the Interest Group Hypothesis,” Public Choice, 106(1-2), 157-81.
Ramirez, Carlos D., and J. Bradford DeLong. 1996. “Banker Influence and Business Economic
Performance: Assessing the Impact of Depression-Era Financial Reforms,” in Bordo and
Sylla, eds., Anglo-American Financial Systems: Institutions and Markets in the Twentieth
Century. New York: Irwin.
48
—— and ——. 2001. “Understanding America’s Hesitant Steps Toward Financial Capitalism:
Politics, the Depression, and the Separation of Commercial and Investment Banking,” Public
Choice, 106(1-2), 93-116.
Rajan, Rghuram G., and Luigi Zingales. 1995. “What Do We Know about Capital Structure?
Some Evidence from International Data,” Journal of Finance, 50(5), 1421-60.
Redlich, Fritz. 1951. The Molding of American Banking: Men and Ideas. New York: Hafner
Publishing Co.
Ripley, William Z. 1912. Railroads: Rates and Regulation. New York: Longmans, Green and Co.
——.1915. Railroads: Finance and Organization. New York: Longmans, Green and Co.
——.1916. Trusts, Pools and Corporations (Revised ed.) Boston: Ginn and Co.
Rockoff, Hugh. 2000. “Banking and Finance, 1789-1914,” in Engerman and Gallman, eds. Cambridge Economic History of the United States. New York: Cambridge University Press.
Roe, Mark. 1994. Strong Managers, Weak Owners: The Political Roots of American Corporate
Finance. Princeton: Princeton University Press.
Romer, Christine. 1989. “The Prewar Business Cycle Reconsidered: New Estimates of Gross
National Product,” Journal of Political Economy, 97, 1-37.
Schnabl, Philipp. 2009. “Financial Globalization and the Transmission of Credit Supply Shocks:
Evidence from an Emerging Market.” Draft, NYU.
Seligman, Joel. 1982. The Transformation of Wall Street: A History of the Securities and Exchange Commission and Modern Corporate Finance. Boston: Houghton Mifflin.
Sivakumar, Kumar N., and Gregory Waymire. 2003. “Enforceable Accounting Rules and Income
Measurement by Early 20th Century Railroads,” Journal of Accounting Research, 41(2), 397432.
Smith, George D., and Richard Sylla. 1993. The Transformation of Financial Capitalism: An
Essay on The History of American Capital Markets. New York: Blackwell.
Sobel, Robert. 1965. The Big Board: A History of the U.S. Stock Market. New York: Free Press.
——. 1968. Panic on Wall Street: A History of America’s Financial Disasters. London: McMillan.
Stover, John F. 1961. American Railroads. Chicago: University of Chicago Press.
Sylla, Richard. 1975. The American Capital Market, 1846-1914. New York: Arno Press.
Travers, Arthur H. 1968. “Interlocks in Corporate Management and the Antitrust Laws,” Texas
Law Review, 48, 819-58.
Weinstein, David E., and Yishay Yafeh. 1998. “On the Costs of a Bank-Centered Financial
System: Evidence from the Changing Main Bank Relations in Japan,” Journal of Finance,
53(2), 635-72.
49
White, Eugene. 1983. The Regulation and Reform of the American Banking System, 1900-1929.
Princeton: Princeton University Press.
——. 1986. “Before the Glass Steagall Act: An Analysis of the Investment Banking Activities of
National Banks.” Explorations in Economic History, 23(1), 33-55.
Wicker, Elmus. 2000. Banking Panics of the Gilded Age. New York: Cambridge.
Willis, Parker and John Bogen. 1929. Investment Banking. New York: Harper and Row.
Winerman, Marc. 2003. “The Origins of the FTC: Concentration, Cooperation, Control and
Competition.” Antitrust Law Journal, 71, 1-97.
Woodman, Earnest S. 1905-. Directory of Directors in the City of Boston and Vicinity. Boston:
Bankers’ Service Company.
50