Information, institutions and agency: the crisis of railroad finance in the 1890s and the...

Introduction and background

The American railroad industry underwent a massive financial reorganisation after a devastating wave of financial bankruptcies in the 189Os. The institutional and informational responses to this crisis were far-reaching, varied and fundamental. This study examines these responses with a view to better understanding the evolution of capital markets, and the development of institutions for regulating industry. It considers the long-term agency dynamics that were involved as financial equilibrium was sought to be restored.

Research in economics and its allied fields of accounting and finance has long focused on the problem of the uneven distribution of information between business managers and investors. Adam Smith worried about this dilemma as early as 1776 in his classic The Wealth of Nations, which had criticised the great trading monopolies, like the English East India Company, believing that the informational asymmetries associated with such large enterprises contributed to inefficiency and corruption (Smith, 1776). A century and a half later, Berle and Means in The Modern Corporation and Private Property (1932) echoed this concern by contending that the risk associated with informational asymmetry was corrosive to the efficient operation of the broad, anonymous financial markets that had emerged in the US by the 1920s. In their view, financial agency could best be addressed through greater transparency in the form of accounting disclosure certified by public accountants (Berle & Means, 1932). And since the 1970s, the creative theoretical works of Alchian and Demsetz (1972), Jensen and Meckling (1976), Watts and Zimmermann (1983, 1986) and others have greatly increased scholarly understanding of how agency relationships influence managerial, financial and accounting decision processes.

Although the predominant, contemporary emphasis on logico-mathematical modes of analysis has successfully advanced modern agency theory, constraints inherent in the methodologies applied in these research programs have limited their effectiveness in evaluating the significance of long-term, dynamic socio-economic processes. First, this empirical approach has generally considered agency questions in the narrow context of direct, dual-dimensional interactions between investor principals and their corporate agents.1 second, due to the emphasis on internal validity and the need for computational efficiency, modern agency approaches rely on static equilibrium analyses that limit comparisons against real world circumstances. Third, such lines of inquiry embrace a neoclassical view of the economy where individual units are essentially price takers and exert little market power individually. While such an assumption may help to satisfy the population homogeneity requirement necessary for valid statistical analysis, it is at a variance with the predominantly oligopolistic structures that dominate modern business and finance. Finally, economic agency models abstract the conception of the firm as essentially a nexus of contracts capable of diminishing agency risk through the adaptation of institutional arrangements from a bounded range of alternatives such as the distribution of ownership rights, incentive compensation contracts and the monitoring by boards of directors and independent accountants (Robbins, 1932; Friedman, 1953; Papandreou, 1958; Jensen & Meckling, 1976; Blaug, 1985, 1992; Whitely, 1986).

These circumstances suggest the need for a broader perspective in comprehending the nature of agency and its relationship with institutional and informational dynamics. Adapting Baskin and Miranti's (1997, p.304) framework to this context will enable us to more fully understand the rich interplay between environmental and firm-specific factors. Drawing on Galambos (1983) and Berniger (1986), the dependence of capital markets agency relationships (A) on short-term and long-term factors may be depicted as:

A=f(S^sub k^;L^sub k^)

where the S^sub k^ represents institutional and informational arrangements made to address agency problems in the short run, and the L^sub k^ represents adaptive institutional and informational responses to long-term factors including technology, professionalisation, and political economy that are identified by Galambos (1983) in his theory of organisational synthesis. The typical level of analysis in traditional agency-based research keeps the values of L^sub k^ constant and considers the functional forms and changes in the dependent variable in relation to changes in one or more variables in S^sub k^. These studies are therefore concerned, for instance with how agency relationships in the short term are affected by managers' decisions regarding firm capital structure, or executive compensation, or accounting method choice given that external institutional and informational factors embodied in L^sub k^ are fixed. But in a period of economic transition, the role of institutions cannot be assumed to be passive, and the functional forms of the L^sub k^ variables themselves change. Partial equilibrium analysis is of little use in these contexts. Thus an explanation of the viability or profitability of railroads during this period will consider not only short term institutional and informational responses that are affected by firm specific factors like cost of capital, earnings growth, and management's risk preferences, but also environmental factors like regulatory structures, property rights, and communication modalities. In particular, this paper examines how a shock to the system causes environmental responses and adaptive firm specific responses that together explain changes in agency relationships.

The environmental factors that unify the analysis in this paper relate to the role of institutions and information in shaping agency relations in business management and governance. The unique modes of inquiry applied by institutional historians and evolutionary economists have surmounted some of the barriers imposed by the requirements of quantitative research methodologies in evaluating complex, dynamic processes. Galambos (1970, 1983), Chandler (1977, 1990, 1992), Nelson and Winter (1982), Berniger (1986), Galambos and Pratt (1988), North (1990, 1994), and Miller (1994) have explained how institutional and informational change responded to the discovery through path-dependent learning of a better understanding about the nature of the relationship between business organisations and the broader socio-economic environment. Although major discontinuities have shaped the course of business development - such as the emergence of corporations of great scale and scope beginning during the late nineteenth century - the perfection of the new business form followed a pattern that seemed more evolutionary rather than revolutionary. The advancement of the new giant enterprise paradigm responded to the gradual discovery of better ways to manage the new organisations and to reconcile their functioning with the goals of critical stakeholder groups including investors, regulators, consumers and professional service providers. Both business and governmental organisations learned from their assessment of past experience about those policies that proved most successful in achieving basic goals. The understanding that developed from such learning became embedded in the organisational memory through the establishment of new institutional arrangements that facilitated the replication of actions that had led to past success.

By employing an evolutionary conception of agency, this study will analyse how innovative regulatory institutions emerged to specify, gather, classify and disseminate new forms of useful information for improving public oversight over the giant railroad enterprise. The focus centres on the institutional and information responses that followed the financial reorganisation of many leading railroads that went into receivership during the recession of 1893. The analysis of these events will amplify our understanding of the evolution both of property rights and of institutional arrangements for enhancing corporate transparency and capital market efficiency. First, the railroads by annihilating time and distance and providing reliable, all-season transportation lowered transaction costs in business and commerce and made possible the rise of the nineteenth century urban-industrial economy (Chandler, 1977, chapter 3 passim). second, this industry pioneered the development of accounting information not only for guiding internal managerial processes but in communicating both with a growing, anonymous body of creditors and equity investors and consumers of rail services who were deeply concerned about the equity of rates (Baskin & Miranti, 1997, p. 132-46, Previts & Samson, 2000). The statistical and accounting requirements of the Act to Regulate Commerce (Interstate Commerce Act) in 1887 increased the transparency of corporate affairs and, thereby, helped to reduce the risks associated with the asymmetric distribution of information between investors, consumers and their railroad officials. Third, the episode is significant because of the scale of industry bankruptcy, with about 25 per cent of total mileage in receivership by 1897 and because of the scope of the informational and institutional reforms that these dislocations eventually engendered. Although the sample is small, it is stratified in such a way as to account for a very high proportion of total assets placed under receivership during this era as well as a high proportion of the total asset value of the entire industry (Daggett, 1908). Fourth, railroad securities, particularly bonds, represented the most important category of publicly traded investments during this era partly because of the size and importance of the industry. Investors generally preferred bonds to equity contracts because of their informational attributes and superior legal rights. Bondholders had reliable information about contract duration and periodic returns in the form of interest and principle payments. They also enjoyed a priority over equity holders in the event of corporate liquidation. Moreover, the fact that the railroads functioned as natural monopolies in many markets and the general accessibility to public disclosure about the level of debt leverage in the capital structure of individual firms helped investors to gauge risk (Baskin & Miranti, 1997, p. 146-51). Fifth, the learning that derived from the financial crisis had a powerful influence in shaping the subsequent direction and nature of future institutional reform in the industry. It also helped contemporaries to understand the limitations inherent in corporate governance institutions in providing protection for investors and consumers.

This study extends a long line of historical inquiries into the significance of the railroad bankruptcies of the 189Os and their aftermath. The pioneering studies of Daggett (1908), Ripley (1912, 1915), and Dewing (1920) analysed these events so us to develop a better understanding of the economics of giant enterprise and applied these insights to propose new regulatory mechanisms that would preserve the benefits of scale while protecting the public from the unbridled exercise of great market power. Campbell (1938) detailed the steps taken to revive the industry after World War I through financial re-capitalisation and system consolidation. More recently Berk (1997) has argued that the reorganisation experience provided evidence that challenges Chandler's (1977, 1990) contention that the development of business strategies and organisational structures of giant business enterprise in the late nineteenth century was largely devoid of government involvement. Berk essentially believes that the lowering of corporate fixed financing charges through judicial reorganisation proceedings evidences substantial government involvement. Tufano ( 1997), on the other hand, focused on debt contracts in these reorganisations and described how the new economic conditions gradually shifted investor perceptions away from collateral toward income as an indicator of risk and worth.

This study advances understanding of this transitional period by evaluating the institutional and information reforms introduced to reduce risk perceptions among investors and to bolster confidence among consumers of the equity of railroad rates. The following section assesses the shortcomings of contemporary legal and regulatory institutions in addressing the many economic complications that emerged from extensive railroad bankruptcy and analyses the initial adaptive response involving the modification of prevailing institutional arrangements, particularly in law, to accommodate the unprecedented levels of business failure. The third section analyses the longer-term innovative response that led to the creation of new institutional and informational arrangements that reflected a greater sensitivity to the growing complexity and interdependence of modern industrial society. The conclusion explains how the response to the problem of massive economic dislocation in the railroad industry amplifies understanding of the role of information and institutions in mitigating the problems of agency in business finance.

Information, contracts and agency prior to the Depression of 1893

Two major institutional arrangements that supported railroad finance prior to the bankruptcy crisis of the 189Os included both contracts and specialised sources of information. This section discusses the types of railroad contracts and how they were distributed among various stakeholder classes. It also analyses the development of important sources of information about railroad finance and management, which helped to diminish investor risk concerns and enhance market efficiency. This section also identifies several informational shortcomings from the perspective of the Interstate Commerce Commission (ICC), which had the primary responsibility on the national level for improving the transparency of corporate affairs to assist both rate regulators and investors.

During the railroad industry's formative years, debt contracts served as the primary source for attracting expansion capital from outside investors because of their utility in overcoming a serious problem of informational asymmetry. Public ownership of equity securities remained limited because information about the potentialities and risks of the new industry remained initially limited. This was especially the case among British investors who sought safety by investing primarily in bonds (Cleveland & Powell, 1909, pp. 195-6; Madden, 1985, p.337; Wilkins, 1989). Although some equity capital had been raised through public subscription as in the case of the Boston and Worcester Railroad (Salisbury, 1967, chapter 4), the more common pattern was that of the Baltimore and Ohio and the Pennsylvania Railroad where local and state governments interested in enhancing their local economies through improved transportation became major subscribers (Cleveland & Powell, 1909; Burgess & Kennedy, 1949, chapter 5; Previts & Samson, 2000; Flesher, Samson, & Previts, 2002). Similarly, the Union Pacific's financing greatly benefited from the land grants and bond subsidies received from the federal government to accelerate the formation of better communications between the Eastern and Western states (Trottman, 1923, chapter 1). Although one of the earliest private investors to participate on a significant scale in the early railroads was New York merchant, John Jacob Astor (Chandler, McCraw & Tedlow, 1995, pp.1-47), equity remained primarily the investment vehicle of corporate promoters, financiers and managers prior to 1872 (Ripley, 1915, pp.106-8). While annual reports and financial statements date back to the incipiency of the industry in the 182Os, a general lack of understanding of rail economics and the lack of standardisation of accounting information contributed to the relegation of equity investment to speculative status.

Bond financing contracts, on the other hand, represented a more effective means of attracting capital from risk-averse external investors for several reasons.2 First, debt obligations enjoyed stronger property rights such as guarantees for the payment of both interest and principal as compared to equity dividends whose value depended on profitable operations. second, besides legal preferences over equity with respect to cash payouts and in corporate reorganisations, debt contracts often seemed more secure when backed by pledges of valuable collateral. According to Ripley, early European investors in US railroads limited their purchases primarily to mortgage bonds (Ripley, 1915, p. 106). Third, debt contracts inherently embedded much more evaluative information than those for common equity. Bond contracts specified both the interest rate and term. Moreover, since rail enterprises operated as natural monopolies in many of their service regions, the problem of default risk often seemed remote. Common stock, on the other hand, remained difficult to evaluate because of the difficulties encountered in assessing revenue trends and cost structures that affected profitability and the ability to pay dividends.

New institutions gradually emerged during the course ol the nineteenth century to accumulate and disseminate information about railroad finance and economics, thus gradually helping to reduce investor risk perceptions. Current information and commentary about industry developments were transmitted broadly through specialised business periodicals such as Henry Varnum Poor's American Railroad Journal (Chandler, 1956). University scholars, public leaders and engineers in the US and overseas took note of railroad enterprise by establishing theoretical frameworks to assist in economic analysis and making informed policy proposals for improving oversight and management (Lardner, 1850; Fink, 1875; Hadley, 1885). Government, initially on the state level, also became more deeply involved through the activities of their rate regulatory commissions in reporting information about the industry to facilitate the achievement of policy objectives.3 By 1886 the state initiatives were imitated by the federal government through the formation of the Interstate Commerce Commission (ICC).

The ICC required the filing of reports employing uniform formats, which were published in The Annual Statistics of the Railways of the United States. The federal system, developed under the leadership of Henry Carter Adams of the University of Michigan, not only sought to inform rate assessment and financial evaluation processes, but also to promote automatic or "self-executory" compliance to regulation through the broad publicity of accounts (Miranti, 1989). In this latter case the regulators believed that the high degree of corporate transparency provided by their reporting regime would serve as an effective deterrent to agent malfeasance (Adams, 1893). The appearance of new professions such as public accountancy also bolstered investor confidence in the reliability of corporate financial disclosures. While the earliest British chartered accounting associations dated back to the 185Os, the formation of the American Association of Public Accountants in 1887 marked the beginning of an associational focus dedicated to promoting auditing services in the US (Previts & Merino, 1998, p.138). Although the ICC did not require the independent certification of financial statements, investment bankers like J.P. Morgan sought to enhance the marketability of new railroad securities, particularly in London, by including in their prospectuses audit reports issued by well-known public accounting firms (Wilkins, 1989, pp.536-45).

Yet, in spite of the growing array of information sources, railroad finance experienced a severe crisis during the depression of 1893. This raised serious questions about the adequacy of the institutions that provided oversight over corporate affairs. The large, heavily indebted interregional rail systems bore the brunt of the decline. The Atchison, Topeka and Santa Fe, the Baltimore and Ohio, the Erie, the Union Pacific, the Northern Pacific, the Reading and the Southern all become financial casualties (Daggett, 1908; Campbell, 1938). The institutional response to these dire circumstances was twofold. The first was adaptive and largely involved the modification of common law structures for allocating losses and securing fresh financial resources in bankruptcy. The second was innovative and involved the development of new laws, institutions and sources of information to protect stakeholder interest and to establish more effective public oversight of the railroad enterprise.

The adaptive institutional response: bankruptcy and reorganisation

Much of the adaptive change with respect to bankruptcy involved the modification of old legal institutions and practices to accommodate the emergent economic problems of giant enterprise. Although well defined in the annals of Anglo-American jurisprudence, the traditional approaches to bankruptcy now had to confront unprecedented levels of lost investment value and that adversely affected the operations of several of the most important components of the national railroad network.

Reorganisation procedures involved the establishment of a complex set of agency relationships. Court-appointed receivers, for example, had the responsibility for arranging interim financing and for facilitating the formulation of a reorganisation plan that reconciled competing stakeholder interests and placed the entity on a sound financial footing. The receiver, usually a banking firm, had the additional charge for floating new security issues to support corporate resuscitation. Investor protective associations represented by bank trustees for each class of investment security provided assurances to their constituents about the proper exercise of guarantees, preferences and other rights specified in their underlying contracts. Teams of attorneys, accountants, engineers, appraisers and other expert groups also served as adjuncts to the primary fiduciaries by bringing to bear their specialised professional knowledge in resolving arcane questions affecting bankruptcy administration. The judiciary represented the principal governmental agency in the resolution of these matters. The effectiveness of the outcomes depended on the knowledge and acumen of the court's cadre of judges who, although often expert in bankruptcy, usually lacked detailed knowledge of the complexities of railroad finance and management. The ICC, the more expert agency with respect to industry knowledge, virtually had no direct input although the judicial decisions affected its role as the arbiter of rate equity and assuror of corporate transparency.4 Re-capitalisation and asset revaluation, for example, affected the measurement of profits and also the investment bases of external investors that informed regulatory decisions about the reasonableness of rates. Consumer groups understandably remained concerned about such adjustments, which might justify requests for rate increases. Many investors remained wary, believing that any unwarranted upward valuations of fixed assets resulting from such proceeding created watered stock and undermined railroad earnings power (Dewing, 1920, Vol.5, chapter 2).

By the 189Os the judicial system recognised two methods for handling the problem of corporate bankruptcy: (1) outright liquidation and the winding up of the affairs of the insolvent entity; and (2) a court supervised reorganisation which sought to place an endangered firm on a sounder financial footing. The courts preferred reorganisation in the case of the railroads to avoid the hardships that liquidation would impose on the public because of service interruption.

The initial stage in a reorganisation involved the placement of the enterprise in the hands of a receiver who took control of corporate property and infused liquidity through the issuance of receivers' notes. Purchasers of these securities had superior rights over all creditors or equity investors, being subordinated only to employee wage claims. The receiver also provided bridge financing that enabled the line to continue its operations during the frequently lengthy negotiation of an acceptable plan of reorganisation to all stockholder groups.

Receivers had two options in restoring railroad finances. First, they could secure permanent financing by assessing existing equity owners for additional cash. In this case the receiver imposed a requirement that preferred and/or common shareholders subscribe to the purchase of additional shares. Such subscribers might benefit from a recovery in the market value of their original shares as well as from capital gains on their assessments. Although they might prove effective in restoring solvency, assessments per se could not eradicate capital deficits, reduce the level of interest charges or change the relative proportions of fixed and contingent financing contracts (Daggett, 1908, pp.348-58; Dewing, 1920, Vol.5, chapter 5).

The second option, re-capitalisation, promised not only to infuse cash but also to eliminate capital deficits and to reduce debt and lower interest costs.

The charts presented in Figures 1a and 1b show how the financial structure was changed in the reorganisation to dramatically reduce fixed charges and make the railroads more operationally efficient. While reorganisations generally involved an average one-third increase in total capitalisation (Figure 2a), the weighting of longterm capital structures shifted from bonds to greater reliance on preferred stock (Figure 2b).

The involvement of J.P. Morgan and a few bankers was ubiquitous in the reorganisation and recapitalisation process. As a consequence, they achieved new, unprecedented levels of power and control over railroad operations, and the process became termed "morganization" (Chernow, 1990).

Another important step in re-capitalisation often involved the booking of accounting adjustments to reduce or eliminate any deficits in retained earnings. Any capital surplus derived from the issuance of common shares could be written off against such deficits. Another technique involved the upward revaluation to estimated fair market value of any long-term assets, such as land, and the recording of unrealised holding gains, which also diminished deficits. The justification for such adjustments derived from an estimated increment in the value of adjoining real estate that resulted from the successful railroad development and improved access to efficient transportation services. Similar offsets against deficits also came about from the assignment of higher fair market values or the extension of the useful lives for rolling stock and stationary facilities.

Besides accounting adjustments, the second feature of re-capitalisation involved the reduction of fixed payment contracts through the substitution of those with contingent claims. Some reorganisation plans called for the exchange of some portion of junior debt for common shares. Other plans forced the conversion of debt contracts to either preferred stock or income bonds, whose payments in the event of operating losses could be deferred without prompting the initiation of bankruptcy proceedings. The issuance of junior or convertible bonds at deep discounts from par with their implied promise of a significant capital gain at maturity could also enhance the attractiveness of re-capitalisation exchanges. In this latter instance corporate equity increased by classifying the discount not as a contra liability to the bond issue but rather as an asset5 (Daggett, 1908; Dewing, 1920, Vol.5, chapter 6).

While such applications of accounting and financial knowledge may have been effective in advancing judicial procedures, it did not always engender a high degree of confidence among all stakeholder groups. A critical problem lay in the adequacy of accounting adjustments to accurately reflect underlying real economic values. A lack of standardisation of accounting and the dominance of selfinterested banking institutions in these processes became the focus of public criticism and doubt. One highly authoritative centre for such criticism was the ICC, particularly the leadership of its bureau of statistics and accounts. Because of the unique informational requirements necessary to control railroad enterprises, Henry Carter Adams and his proponents believed that all oversight should be transferred from the realm of the common law and the judiciary to that of the administrative law and the ICC. In Adams's view the key to successful oversight lay in the development of new types of information to answer critical policy questions relating to the industry and its finances.

Institutional innovation: the extension of federal regulatory accounting and oversight

The collapse of railroad finance in 1893 represented an epochal change that provided a strong incentive for the further extension of regulatory institutions and information on the federal level to better protect the public interest. The focus of the subsequent reform drive centred on five informational and managerial issues: (1) the standardisation of financial accounting methodologies; (2) joint costing and cost of service measures; (3) the regulatory boundaries of the firm and transfer pricing with affiliates; (4) the valuation of rail industry assets; and (5) normative standards for evaluating enterprise financial planning and performance.

Standardisation of accounting methods

Prior to federal regulation, the rail industry had employed many different methods for accounting for their operating and financing transactions. Although the ICC initially had the power to specify standard reporting formats, the Supreme Court ruled, in a case against the Lake Shore Railroad (Knapp v Lake Shore & Michigan Southern Rail Company, 1905), that the ICC lacked authority to prescribe accounting methodologies [Miranti, 1989, p.484]. Moreover, as Richard Briefs research persuasively has argued, the lack of comprehensive measurement of capital costs encouraged the misallocation of scarce economic resources by providing investors with misleading signals about enterprise profitability (Brief, 1965, 1966). Some firms followed the practice of fully costing their capital assets at time of acquisition, others recognised costs at time of replacement, some did not make any allowance for depreciation, some followed regular depreciation schedules that approximated the useful life of the underlying assets, and still others moved between these many alternatives.6 Such practices diminished the usefulness of accounting information as a means for assessing earnings power and for making informed comparisons between firms in the industry. The previously noted adjustments to estimates of fair market value and the treatment of bond discounts directly impinged on the effectiveness of the ICC's role in helping to assure financial probity and in monitoring the reasonableness of rates (Ripley, 1915, pp.259-67; Bonbright, 1920, pp. 169-85). The Supreme Court had in the 1898 Smyth ? Ames case ruled that rail rates had to be set at levels that assured a reasonable return on the fair value of the capital invested (Smyth v Ames, 1898). However, the court further complicated accounting-informed regulation by declining to decide which of many different bases of accounting should be applied in industry monitoring, leaving this question for Congress and the ICC to resolve (Sharfman, 1931, Vol.1, No. 16, pp.74-7).

The resolution of the methodology problem did not occur until the passage of the Hepburn Act 1906, which had the strong support of the administration of Theodore Roosevelt (Kolko, 1965, pp. 129-44; Hoogenboom & Hoogenboom, 1976, pp.48-59; Skowronek, 1982, pp.249-59). While the new legislation also increased significantly the ICC's power over rates and other matters, it enabled the agency to begin the process of accounting standardisation based on uniform method rather than adherence to generally accepted principles. Uniform methodologies remained critical in advancing the ICC's rate evaluation mission because it allowed for the establishment of useful inter-corporate performance comparisons. The ICC also secured its extended powers in industry reporting by issuing a new general chart of accounts, by mandating the use of a more informative operating statement and by organising an examination staff with powers to audit the filings of regulated railroads (ICC, 1906, pp.59-62; ICC 1907a, pp. 139-44; ICC, 1907b, pp.9-24 and pp. 139-150; ICC 1907c, pp.9-24; Adams, 1908).

Joint cost and the cost of service rating

A second dilemma of accounting-based regulation related to the problem of joint cost analysis in rate evaluations. Because of the difficulties inherent in separating the substantial joint costs of operation between individual lines of service, the initial practice in the industry involved the use of value of service information as a guide to establishing rates. Rates were set at levels that the market could bear. Moreover, the ICC had also embraced this standard because it facilitated the achievement of income reallocation goals that it favoured. Specifically, value of service rating implicitly assured that the tariff for the carriage of high value manufactures would subsidise the transportation of basic staples such as foods and fuels. In this way the profitable manufacturing sector alleviated some of the economic pressures felt by a growing urban labour population and also helped to bolster incomes in the poorer economic regions, particularly the South and West with their heavy reliance on farming and extractive industries (ICC, 189Oa; ICC, 189Ob; ICC, 1914; Friedlander, 1969, pp.8-16).

The cross-subsidies inherent in value of service rating, however, eventually came under attack during the Advance Rate case of 1911. Following the passage of the Mann-Elkins Act the prior year, the railroads petitioned for the first time en masse for a comprehensive readjustment of rates (Ripley, 1912, pp.594-600). During this case attorney Louis D. Brandeis supported by management consultants Harrington Emerson and Henry J. Gantt, argued that the equity of the proposed rate increases could not be properly determined without information about the costs of providing particular classes of service, a goal beyond the capabilities of contemporary railroad accounting systems (Haber, 1964, pp.51-8; McCraw, 1984, pp.91-4; Oakes & Miranti, 1996).

While the cost-of-service arguments delayed the granting of the requested rate increases, over the longer term it induced the ICC to consider the applicability of marginal cost analysis for setting rates so that they would be competitive with those offered by the unregulated and cheaper river barge industry. The agency recruited Professor Max Otto Lorenz from the University of Wisconsin to undertake such studies (Lorenz, 1907, 1916; Ely, 1924). Although some firms like the Southern Pacific had also begun to experiment with marginal analysis by the 1920s, it had little effect on federal regulation until the 1930s when the New Deal's Transportation Acts extended the scope of the ICC's authority to include both trucking and barge transportation (Thompson, 1989). Then regulators used marginal costing as a mechanism for allocating the declining traffic volume between its three constituent industries. It facilitated the federal agency's ability to establish competitive rates between trains and barges for low value, high bulk cargoes and between trains and trucks for smaller, high-value shipments (Friedlander, 1969, pp.20-7).

Defining the regulatory boundaries of the firm and the problem of transfer pricing

An early issue that concerned reformers about the Act to Regulate Commerce derived from the initial failure to extend the scope of the ICC's regulatory authority to encompass ancillary services such as port lighterage, express delivery, equipment leasing and warehousing. Consequently, many believed that these allied business activities could function as opaque conduits for disguising the transfer of excessive profits from rail operations. The Erie Railroad, for example, which served for many years as the speculative vehicle of Jay Gould, had controlled much of the barge service that provided the vital linkages connecting docks and railroad sidings around New York harbour. Some suspected that through such interlocking services the measurement of rail operations could be structured so as to siphon off income, understate the true profitability of the line and, thereby, provide seemingly objective data in support of what in reality were unjustified higher freight rates.

The problem of regulatory scope remained unresolved until 1908 when revenue reporting requirements extended to some miscellaneous activities, including among others boat lines, canals, grain elevators, stock yards, and coldstorage warehouses (ICC, 1907c, pp.11-13). The subsequent passage of the MannElkinsAct 1910, besides establishing the ill-fated Commerce Court for adjudicating rate controversies and enhancing the power of the ICC over rates, also extended federal oversight over interstate telephone and telegraph and pipeline enterprises. Although these actions did not encompass the full scope of allied rail activities, it did increase in the horizon of regulation and the transparency of railroad business affairs (Ripley, 1912, chapter 17; Sharfman, 1931, Vol.1, pp.52-70; Kolko, 1965, pp.188-95; Martin, 1971, pp.183-93; Skowronek, 1982, pp.261-6).

Determination of original and replacement value of rail assets

The ICC also sought to overcome inconsistencies in accounting measurement that had long plagued railroad reporting through its sponsorship of a national inventory and valuation of all of the industry's assets. Although the Supreme Court in the Smyth case recognised the importance of accounting standardisation, the process did not get underway until the 1906 passage of the Hepburn Act. Consequently, there was great uncertainty among investors about the reliability of accounting information. As discussed above, during much of the nineteenth century major lines frequently employed significantly different capital cost accounting methods. Concerns also emerged about the completeness of records because of the loss of original cost data after mergers or reorganisations or after the reconstruction of operating facilities and road beds. Some information also remained suspect if prepared by construction companies controlled by unscrupulous promoters with strong incentives to inflate asset values in the initial offerings of bonds and shares (ICC, 1902, pp.80-2, 1903, pp.22-32, 1904, pp.10-19, pp.19-23; Ripley, 1912, pp.515-16).

These and other problems motivated state railroad commissioners operating through their representative organisation, the National Association of Railway Commissioners (later the National Association of Railroad and Utility Commissioners) (Smykay, 1955), to call for a comprehensive valuation of all industry assets to determine how much "water" existed in rail capitalisation, particularly after the reorganisations of the 189Os. This policy seemed desirable because among other issues it would help both to provide more reliable information for local property tax assessments and to provide a check on the reasonableness of fixed asset valuation and depreciation that affected the rate base and the general level of rail tariffs (National Association of Railway Commissioners, 1915, pp.382-6; and National Association of Railway Commissioners, 1916, pp.215-46).

Although the original Hepburn bill had included a provision for a national valuation of rail assets, this provision did not actually become law until the passage of the Valuation Act 1913. The latter legislation authorised the ICC to conduct a national inventory and provide estimates of both original costs and costs of reproduction less depreciation for all assets in 1913 prices. The ICC intended to compare these values against both the par and market value of all the capital invested in the industry as a mechanism for judging the probity and efficiency of rail finance. In addition, the federal agency conducted a mapping survey of all rail property nationwide. Although the valuation consumed much of the agency's budgets in the years that followed, it was a lengthy and time consuming process. The ICC only completed its initial preliminary estimate of a single line in 1918 and its first final report for an individual line did not appear until 1923 (Sharfman, 1931-1937, Vol.1, pp.117-38; Vol.3A, pp.33-42, 95-319, Vol.3B, pp.33-97; Martin, 1971, pp.267-349; Hoogenboom et al., 1976, pp.70-9).

Normative standards for evaluating financial planning and performance

The 1893 collapse had reinforced the view of Henry Carter Adams and other Progressive reformers, particularly in the administration of President Theodore Roosevelt, of the need for greater federal intervention in the form of extended authority not only over reporting but also over financial practice. Part of the problem in their view related to the dearth of reliable information and oversight institutions that had made possible the speculative manipulations on the stock market recounted in Massachusetts railroad commissioner Charles Francis Adams's classic, A Chapter of Erie (Adams, 1870). In addition the proponents of greater regulation believed that extending control to the ICC over railroad finance could both reduce risk perceptions among investors and minimise the opportunities for unscrupulous managements to unfairly raise their rate bases by inflating corporate capitalisation. Although Congress did not support the requested extension of these powers through the Hepburn legislation, the ICC did receive this authority after the passage of the National Transportation Act 1920. Besides the power to review and approve of all major railroad financing plans, Congress authorised the ICC to supervise the reorganisation of the national rail transportation system into 11 regional lines. The gradual accretion of governmental power over reporting, valuation and finance represented a necessary preliminary step in the minds of regulators to assure a rational and equitable restructuring of the industry, (ICC, 1920, pp.27-33; Sharfman, 1931-1937, Vol.1, pp.177-244; Vol.3A, pp.51-5; Norris, 1946, chapter 2; Kerr, 1968, pp.211-12; Hoogenboom et ai, 1976, pp.947).

By the 1920s virtually all of the institutional and informational innovations favoured by Progressive reformers had gained permanence in US railroad oversight. The system of governance that had emerged perceived the central problem of railroad management largely in terms of the probity and competence of its leaders and its allied professional groups. From an economic perspective the regulators also favoured the income redistribution inherent in its crosssubsidisation of rates from the wealthy, industrial East to the poorer agrarian South and West. Yet the intricate structure of institutions and information built up over a half century that supported the agency in advancing its mission did not seem particularly effective in preventing the recurrence of extensive bankruptcy during the course of the Depression of the 1930s. Although close federal monitoring doubtless was effective in assuring greater financial probity and rate equity, it did little to cushion the effects of a deep and lengthy industrial collapse. Moreover, the railroads could not be insulated from the serious disruption of heavy industry, a primary customer sector. The baseline of what constituted "safe" finance shifted radically during these years. The financing plans approved by the ICC during the 1920s frequently could no longer accommodate to the requirement of a radically changed economic environment. This eventually led as in 1893 to a resurgence in bankruptcies. The new focus of policy shifted from micro concerns about the ethical failure of business leaders to macro concerns about the failure to maintain aggregate demand.

Conclusion

What then does the preceding analysis reveal about the evolution of institutions and financial agency? How does it inform theoretical perspectives about the nature of their relationships?

This paper evaluates agency issues in a manner that differs considerably from the logico-mathematical studies that have been prominent in the extension of this paradigm in finance and economics. As argued earlier, these latter approaches necessarily focus on short run firm-specific factors by holding the dynamic environmental factors constant. By contrast, the methodology employed in this study stresses the dynamic process of institutional change and its interplay with firm-specific variables over long time horizons. Focusing primarily on a crisis among dominant firms in a leading industry in the 189Os it explains why the diminishment of agency risk and the restoration of financial equilibrium depended not only on the adaptation of an existing array of legal conventions but also on the innovation of regulatory institutions to collect and to disseminate new bodies of information and to monitor corporate agents. Moreover, the study makes clear that institutional innovation did not come about immediately but often took considerable time. This study thus focuses on the dynamics involved in moving from one equilibrium position to another - an issue that is "black-boxed" in the logico-mathematical studies of agency relationships. Although the pathway for reforming the use of accounting information to achieve better governance over rail financing seemed clear in the 189Os, it took three decades for the proposed changes to be fully embraced in the corporate governance system. Innovation was slow because it affected the economic interests of many stakeholder classes and, consequently, required the formation of broad political coalitions that through a protracted process of negotiation and compromise reconciled the inevitable conflicts that arose.

The steady revision of institutional relationships responded to a continuous process of collective learning. Unlike traditional agency models that assume institutional structures are exogenous, this study indicates that the comprehension of their role in agency mediation remains incomplete without consideration of the effects of path dependency. Institutions are problem specific social constructs, being formed as a result of the search for effective responses through surveys of past experience. This was the case in the railroad industry during the 189Os. What history revealed to contemporaries were specific shortcomings in the type and quantity of information available about the railroad enterprise as well as the abusive practices and performance failures that could be curtailed through the definition of more responsive institutions.

Theory building about the nature of the relationship between information, institutions and agency, however, also must encompass the many factors operating over longer time periods than is practical in the case of logico-mathematical analysis. As this study has documented, institutional evolution in rail finance encompassed both firm-specific owner-manager relationships and environmentally specific issues. It also had to take cognisance of the roles of many groups including shippers, governmental officials and politicians and banking and legal professionals - groups that all had input to this process of defining the structure of financial governance in the railroad industry during an important formative period. An important goal, thus, of this latter line of inquiry is to place in broad perspective the varied elements that drive evolutionary processes in the past. While it makes no claim to predictive power, it nevertheless has utility in informing debates about future directions of change. This comes about through the identification of predominant issues that impinge on the relationship between information, institutions and agency and by analysing how they eventually become resolved through the interplay of various social forces. Moreover, many of these factors remain relevant in the contemporary era. While historical knowledge analyses the relevant choices and outcomes in the past, it leaves to contemporary decision makers the challenging task of assigning probabilities of agency risk with respect to planning the disposition of informational and institutional resources in meeting future challenges.

The system of governance that had gradually arisen after the bankruptcies of the 189Os became a model for regulating natural monopolies in the US economy. The system relied on prescribed uniform accounting methods to both inform investors as well as rate regulators. Moreover, it differed markedly with respect to purpose and the structuring of accounting information from the approaches followed by the securities and Exchange Commission (sec) beginning in the 1930s. Not charged with the responsibility of regulating market competition, the sec delegated part of the responsibility for advancing the financial reporting process to private professional groups that standardised accounting information on the basis of generally accepted principles rather than uniform methods. Nevertheless, the regulatory system of prescribed accounting methods lasted for nearly a century and was not abandoned until the passage of the Staggers Transportation Act 1978. By this later time, however, concerns about US global competitiveness and the diminishment of national economic efficiency because of market regulation induced political leaders to support a new system based more on the free competition between the railroads and rival transportation modalities including trucking and barge transport.

Finally, this study seeks to broaden scholarly discourse on agency, institutions and information. The application of evolutionary analysis in studying these relationships represents a form of path-dependent learning. Among other goals this scholarly approach seeks understanding of the past perception of these matters and how they shaped the actions taken to achieve a more efficacious social ordering. It is an intellectual process, which inevitably compels the reconsideration of the bases and outcomes of earlier institutional developments. It, thus, also informs decisions about the utility of past action and whether inherited institutional structures should be continued, modified or abandoned. Such a line of inquiry also establishes an intellectual continuum, which connects past circumstances to the current condition of informational institutions. And it is through the critical evaluation of this experience that the future direction of change is foreshadowed.