II. FROM PRESENT VALUE TO REPRODUCTION COST^{Footnote 6}

Writing for the Supreme Court in an 1893 takings case involving a navigation company (channels were at the time second only to railroads in triggering regulatory controversies), and called to determine the exact value of “just compensation” for condemned property, Justice David Brewer proclaimed the following criterion:

This was nothing but the modern idea that the value of a business is equal to the present value of its future net earnings. Value should not be determined by looking backward at production cost, as in classical economics, but by looking forward, in terms of the expected income flows a given property could generate.^{Footnote 8}

However, notes Siegel (1984, p. 221), present value was an ambiguous concept when applied to regulatory issues. If it meant the value of an unregulated business, which by definition included its ability to gain supra-competitive profits, the purpose of regulation would vanish. Neither could it mean its regulated value—i.e., the value resulting from employing regulated rates in the calculation of future earnings. In such a case the constitutional protection of property would be dodged by basing the permissible rates on the rates already set by legislators. Only one solution existed, and American courts did not take long to realize that.

The proper present value of a business could be no other than the value set by the competitive market. Rate regulation was confiscatory when it reduced the overall return of the regulated business below that obtainable from an investment of comparable size and risk made in a competitive environment. While present value itself was a non-classical notion, using competitive markets as a benchmark for “just compensation” rested upon the well-established foundations of classical political economy. Both conservative and progressive jurists could invoke the authority of Adam Smith and his system of natural liberty to defend the choice as morally, as well as economically, grounded.^{Footnote 9} It was common at the time to criticize regulatory case law as arbitrary and ad hoc. But how could anyone accuse a court of having abused its discretionary powers if its decision had been guided by the just light of classical competition? Courts did not miss the golden opportunity of an inflow of external authoritativeness in so controversial an issue.

Under the doctrine of competitive present value, the judicial task was to determine what the value of a railroad or public utility would have been if operated in a competitive environment. A difficult task again, were it not for the help coming from business and engineering valuation techniques. The notion of reproduction cost provided the answer. While the notion had been theoretically developed in Henry Carey’s Principles of Political Economy,^{Footnote 10} it is very likely that late nineteenth-century US courts borrowed the idea of reproduction, or replacement cost, from more mundane sources.

For example, the 1886 volume of the Transactions of the American Society of Mechanical Engineers included a paper by a New Jersey engineer named Oberlin Smith, “Inventory Evaluation of Machinery Plant.” Smith discusses the evaluation techniques a business should apply in cost accounting, the problem being to ascertain “the true value of property kept account of” (Smith Reference Smith1886, p. 433). The only correct answer, he declares, is reproduction cost: “The grand principle which lies at the root of correct evaluation, and which should govern the appraiser throughout all his work, is, that any article is worth not what it did cost, but what it would cost to replace it to-day” (ibid., p. 436; original emphasis). As a further example, consider an 1886 paper by Brooklyn economist Paul Davis that dealt with the big national controversy surrounding the outstanding debt to the United States of the Union Pacific Railway Co. One solution then touted was to foreclose the government lien and proceed with selling the railroad lines, but, to assess its convenience, one should first be able to determine the value of Union Pacific properties. Here came Davis’s contribution. He suggested making recourse to the cost of reproduction principle, invoking the authority of civil engineer and renowned expert in railway matters Richard Price Morgan, who had already adopted this evaluation technique for the same railway company a few years before (Davis Reference Davis1886, pp. 68–70).

In a sort of mutually enforcing feedback, the notion of replacement cost owes a lot to the history of regulation itself. For instance, the concept was employed as far back as the 1840s in controversies raised against the British Railway Commission about the evaluation of capital improvements.^{Footnote 11} Accounting historian Germain Boer confirms that the earliest arguments about replacement cost in the US had no theoretical foundation or origin, but stemmed from rate cases involving public utilities and heard before regulatory agencies: “It was in the hearings before these commissions that replacement cost first received prominence in America” (Boer Reference Boer1966, p. 92). In short, railroad regulators used reproduction cost because engineers and accountants did; engineers and accountants in turn used it because some regulators had. Yet, in order to come full circle, an authoritative judicial sanction was required. The Supreme Court was ready to oblige.

III. THE SMYTH RULE

By the last two decades of the nineteenth century, engineers, accountants, and even a few regulators had agreed that the present value of a business was equivalent to the funds a potential buyer of the whole business would have to expend to construct anew that specific enterprise—i.e., the business’s reproduction cost. This technique rested on the more or less implicit assumption that buyers in a free market always had the option of building, rather than buying, and should bear the full consequences of their own choice. The reference to free—that is, competitive—markets proved crucial in the technique’s success. As Siegel remarks (1984, p. 222), under the reproduction cost approach, a businessman in a regulated industry would reap the fruits, positive or negative, of his own ability. If he had built his enterprise at less than its current reproduction cost, even regulated rates would guarantee him an extra profit. By contrast, if he had spent too much, he would suffer from his bad choices. This was exactly what would happen in the unregulated market, given that the reproduction cost always identified the right value of his business.

Of course, calculating the reproduction cost of a railroad, a navigation channel, or a public utility was no easy task, but still no harder than determining the historic cost, as required by the backward-looking evaluation criteria. Moreover, it was a matter of factual inquiry rather than mere judicial opinion. This catered to the courts’ need to avoid excessive discretion. In 1898 a unanimous Supreme Court established the reproduction cost rule in the landmark Smyth v. Ames case.^{Footnote 12} Despite mounting criticism (especially after WWI), and a considerable degree of ambiguity in the Court’s own wording, the rule would remain valid for the next forty-six years,^{Footnote 13} helping judges at all levels to separate rate regulation from confiscation.

The case involved Nebraska’s 1893 railroad rate legislation. With an opinion written by Justice John Harlan, the Court found the regulation unconstitutional: when set too low to permit the railroads a “reasonable return,” regulated rates amounted to an unconstitutional taking of private property. Factual analysis would have sufficed to reach this conclusion in the specific case, because the rates established by Nebraska legislators brought insufficient revenues for the railroads under any possible definition of “reasonable return.” However, the Court hinted at a preferred approach to the notion. In the key passage of his opinion, Harlan wrote:

These words were subjected to careful scrutiny over the following four decades.^{Footnote 14} Taken literally, they merely offered a list of evaluation criteria that eschewed any commitment to a specific determination of the vague concepts of “fair return” and “fair value.” So, Harlan’s statement has been considered ambiguous at best and useless at worst. According to Siegel, though, this was not how most contemporary courts and commentators understood it. At the time, it was widely recognized that the Court had endorsed the reproduction cost approach to rate regulation.^{Footnote 15} Indeed, all subsequent decisions on rate regulations took it for granted that the “fair return” to which a regulated company was entitled was the competitive dividend, and that the “fair value” upon which that return had to be calculated was the present reproduction value of the company’s assets, as required by the reproduction cost method, and not their original cost.

One year later, in San Diego Land & Town, Harlan himself clarified the Court’s endorsement of reproduction cost, calling for a “fair return upon the reasonable value of the property at the time it is being used for the public” as the measure of just compensation, while rejecting the appellant’s request based on the original cost.^{Footnote 16} Even more significantly, the next rate regulation cases—where the Court consistently applied the new method—were all decided unanimously, with opinions written by justices as diverse as progressive Oliver Wendell Holmes and conservative Rufus Peckham.^{Footnote 17} In 1903 Holmes quoted Harlan’s words from San Diego Land & Town to proclaim that “[i]t no longer is open to dispute that under the constitution,” evaluation for regulatory reasons should be based on reproduction cost. “That is decided,” Holmes wrote, “and is decided as against the contention that you are to take the actual cost of the plant, annual depreciation, etc., and to allow a fair profit on that footing over the above expenses.”^{Footnote 18} By the time of Lochner, reproduction cost looked like settled law.

As we said, the logical underpinnings of the reproduction cost principle resided in the notion of competitive markets. Fairness (of values and returns) was expressly translated in terms of the outcomes of classical competition. Smyth thus constitutionalized competitive values and returns as the fair yardsticks for measuring property and for protecting it against unconstitutional takings.^{Footnote 19} Classical competition gave substantive content to all those vague ethical or political attributes—justice, reasonableness, fairness—a regulated rate should satisfy. Indeed, its outcomes were themselves just, reasonable, and fair.^{Footnote 20} Under competitive conditions, the public would pay no more for a product or service than what would grant a competitive return to a freely entering rival: viz., the equilibrium rate of profit. The Court recognized that this was also the amount the Constitution should guarantee to railroad and utility shareholders. Competitive market returns, and only such, represented the morally justified profits that even privileged businesses like railroads and utilities were entitled to gain. Judicial determination of reasonable rates thus suffered no arbitrariness: competition, not the judge’s idiosyncratic will, drew the constitutional boundary between regulation and confiscation. Courts should just establish by factual analysis what the competitive return on the present market value of a given enterprise would be and compare it with that implied by the regulated rates.

Remarkably, the kind of competition the unanimous Court had in mind in Smyth and later decisions was still the classical one. As classical economists envisaged, competition was the dynamic process leveling the profit rate and canceling monopolistic rents unsupported by legal privilege.^{Footnote 21} Competitive markets were simply those where this process worked unhampered. Proof of their effective functioning rested in the tendency itself to the equalization of returns—the analytical core of classical economics.^{Footnote 22} Conversely, the Court’s jurisprudence on rate regulation did not hinge upon the neoclassical notion of competition as a market structure made up of small enterprises devoid of market power. Much as it happened in antitrust case law, the structural view of competition was still far away from the justices’ mindset.^{Footnote 23} Yet, the notion of value underlying the Smyth doctrine was not the classical, wholly objective one, but a more modern conception that could pave the way to fully subjective assessments.

In sum, by 1898 the Court had unanimously endorsed the strong connection between competition and property that lay at the core of classical political economy. The former provided no less than the measure of the latter, because free markets were the one and the only place where the value of property was determined. This free market value was, according to the Court, also the value protected by the American Constitution. One could not separate competition and property without violating the Constitution—and classical economic principles, too.

Under this respect, the Smyth doctrine had momentous consequences that went beyond rate regulation. The constitutionalization of competitive markets affected the Lochner era as a whole. “The Lochner era began,” Siegel writes, “when American jurists decided that the constitutional notion of property included its free market value” (Siegel Reference Siegel1984, p. 260). First settled in the railroad regulation context, the latter idea was soon generalized. All citizens were entitled, like railroads, to the free market value of their property. In Siegel’s account, even the central doctrine of the Lochner era, liberty of contract, emerged as a corollary of the constitutional protection of the competitive value of property (ibid.). Free market value is nothing but exchange value; the latter, in turn, depends upon the possibility of trading property with the utmost liberty. Hence, liberty of contract may also be construed as a necessary implication of the Court’s interpretation of the “just compensation” principle of rate regulation in terms of the reproduction cost rule—that is to say, of the ironclad link between classical competition and the constitutional protection of property rights.

However, the apparent triumph of the classical worldview epitomized by Smyth also contained the seeds for its eventual demise. If value was not anymore an objective entity, but something entirely determined on the marketplace, as a function of a piece of property’s ability to generate a flow of earnings, and if the latter were nothing but the outcome of the property’s capacity to cater to the marketplace’s tastes and needs, critics of that very worldview could then argue for the replacement of a different set of needs that property might legitimately be called to satisfy. For example, it could be claimed that the flow of earnings property was entitled to were those stemming from its ability to meet a politically determined social need; in other words, that value was a matter of policy and that the choice set available to policymakers had more options than just the classical, free market one.

IV. SMYTH BESIEGED

In the years after Smyth, federal courts were busy with rate regulation cases. Under the new doctrine, railroad companies could ask for injunctions against the enforcement of state-fixed rates. Federal courts of equity granted several such injunctions, significantly restricting state rate-making authority. Unsurprisingly, those in favor of more extensive government regulation of business power subjected Smyth and its underlying economics to intense criticism.

The reproduction cost method was attacked on pragmatic and theoretical grounds. Scores of economists, engineers, and legal scholars joined the debate. The intense controversy, which would last until the demise of the Smyth doctrine in the 1940s, climaxed in the wake of the Hepburn Act.^{Footnote 24} The 1906 Act strengthened the power of the Interstate Commerce Commission, establishing, among other things, the commission’s authority to fix maximum railroad rates. By so late a date, the classical faith in the markets’ self-policing ability had almost vanished from the economics profession. Hence, the economists’ take on the railroad issue focused on the allocative function rate regulation should accomplish, downplaying the themes of justice and equality that were central in the classical approach. Everybody still agreed that free markets would determine the long-run equilibrium value of goods and services and that this would conform to production cost. At the same time, almost everybody^{Footnote 25} recognized that in the case of railroads competitive forces would fail to push prices down to production cost and that there lay the rationale for rate regulation. Robert Harvey Whitten, a city planner who mastered the law as well as the economics, explained this rationale most effectively:

The regulator’s task was therefore to determine a rate that corresponded exactly to the “normal price,” which, in turn, should equate to the “normal cost of production.” Whitten defined the latter as “the amount which in the long run it is necessary to pay to secure the utilities demanded by the public. It is the amount that will secure an equilibrium between demand and supply.” And he defined the former as “the amount which in the long run it is necessary to pay to secure the utilities demanded by the public. It is the amount which constitutes an adequate inducement for investment” (ibid., pp. 422, 424). The allocative viewpoint was apparent: rate regulation should aim at generating sufficient profits to attract the capital necessary for future investment in railways, exactly as competitive markets would have done. The question was whether the reproduction cost method devised by the Smyth court could actually perform this task, without triggering either under- or overinvestment.

Several economists gave a negative answer on purely practical grounds. Reproduction cost was unable to fulfill its allocative role because of huge administrative problems. The valuation of present reproduction cost was a time-consuming process, largely based upon conjectural estimates and, above all, incapable by definition of providing a stable basis for fixing rates. Compulsory reference to current prices of inputs in the production process of railway services made the results exceedingly sensitive to changes in the price level. In particular, the inflationary process of the early twentieth century turned the method into a boon for railroads, by generating reproduction values that far exceeded the original outlays to build the lines.

One of the fiercest critics of reproduction cost was Harvard economics professor (and renowned racial taxonomist) William Zebina Ripley. In his oft-quoted treatise Railroads Finance and Organization, Ripley borrowed from the practical wisdom of railways commissioners to dismiss the method. “The ‘reproduction theory’ contemplates an imaginary community in which an imaginary corporation makes imaginary estimates of the cost of an imaginary railroad,” he complained.

The passage reveals not only Ripley’s distaste for the conjectural character of reproduction cost, but also his favoring of a backward-oriented approach like the historical cost method. This preference—which curiously mixed up a neoclassical focus on allocative efficiency with a classical penchant for objective valuation—was shared by most economists involved in the post-Smyth debate.

Practical difficulties also occupied center stage in Whitten’s analysis. He was especially concerned with the intrinsic instability of the incentives stemming from regulated rates when the latter were grounded on reproduction cost. Stability, and the ensuing provision of correct incentives to invest, would perforce require a continuous adjustment of the rate of return applied to the outcome of the evaluation method, but this was clearly a cumbersome task (Whitten Reference Whitten1914, p. 425).^{Footnote 26}

Problems in administering the reproduction cost method could get even worse whenever one wished to apply the technique in a truly rigorous way, by following the so-called value of service rendered approach. The correct method, Whitten explained, should neglect the value of physical property and aim instead at that of the service provided: “[The method] involves a reproduction of the service rather than a reproduction of the plant. If the old plant were wiped out, what would it cost at present to construct and operate a plant capable of performing the service now performed by the old plant?” (Whitten Reference Whitten1914, p. 426). The difference could be huge in terms of depreciation allowances because the given level of service could be rendered even with an obsolete capital.

Both Whitten and Ripley acknowledged that focusing on the service rather than the plant was the proper way of reasoning from a theoretical viewpoint.^{Footnote 27} But they also observed that the approach would clash with existing accounting techniques (Whitten Reference Whitten1914, p. 429; Ripley Reference Ripley1915, p. 357) and, above all, entail insurmountable administrative complexity. In Whitten’s words: “As thus stated, the reproduction method has so many difficulties that it is practically never employed. The reproduction of the service involves not only the determination of the cost of the most efficient substitute plant, but the determination of the present cost of reproducing the business. … The engineering costs of such survey and estimates would be enormous” (Whitten Reference Whitten1914, p. 427). How could the evaluator ever know all the details necessary to reproduce a company’s whole business?

Inevitably, the reproduction cost method had to be applied less than rigorously. But intrinsic imperfection in actual applications undermined the main argument in favor of the method itself, namely, its being a more rigorous evaluation technique—that is to say, one less dependent on the regulator’s or the judge’s arbitrary assessment—than the rough estimates of historical cost. “Rather than providing the definite, clear, and simple rule called for by a functional analysis,” Siegel notes, “the Smyth v. Ames rule was an administrative nightmare and an inaccurate one at that” (Siegel Reference Siegel1984, p. 236). Many economists thus preferred the historical—also called actual at the time—cost approach. On the one side, it was administratively easier; on the other, by a clever use of cost multipliers it could be tailored to meet contingent necessities, such as the need to attract new capital in some parts of the railroad industry.

Actual cost had the further advantage that “it eliminates or minimizes, through compensation, the changes of valuation arising from fluctuations in the level of prices in general” (Ripley Reference Ripley1915, p. 349). This remark touched a sensitive chord with contemporary commentators. Opting for one method or the other was hardly neutral:

The numbers were indeed staggering. Between 1913 and 1920 the US price level rose almost 150%. Under inflationary conditions, the reproduction cost method effectively meant a green light for railroads to charge what the market would bear, and a practical constitutional ban on government regulatory activities. The situation was the opposite of the time of Smyth, which came at the end of a period of sharp deflation. Falling prices meant that the method subjected regulated businesses to the risk of losing on their capital outlays. Accordingly, late nineteenth-century enemies of regulation had shunned reproduction cost, while pro-regulation interpreters had embraced it. Wartime inflation changed everything: now, present costs were much higher than historic ones, so reproduction cost was very favorable to railroads and other public utilities.

Scholars could not fail to notice. Princeton economist John Bauer was straightforward: “Gross reproduction cost is the standard of valuation which corporations are demanding. … Under present high prices it would probably give corporations the largest valuation that could be obtained through any single basis” (Bauer Reference Bauer1916, p. 579). The shocking reversal raised serious equity concerns. “[T]his change in the rate basis was merely fortuitous,” Ripley complained, “totally disconnected with the matter in hand. It would certainly appear more equitable that the rights and obligations of the companies should rest upon the amount of the investment … actually created and placed in the public service” (Ripley Reference Ripley1915, pp. 349–350). Actual cost was clearly preferable for achieving both the allocative and the equity goals of regulation. Whitten could thus conclude: “If normal actual capital cost were adopted as the rule for the future, accounting methods and rate regulation would be much simplified and the relations between the utilities and the public placed on a much more equitable and dependable basis” (Whitten Reference Whitten1914, p. 436).

The superiority of actual cost was not merely practical, but, some argued, also theoretical. A parallel stream of criticisms focused on a major analytical flaw in the argument supporting the reproduction cost method. As we know, the approach found its rationale in the equivalence between the market value of a business and its reproduction cost—an equivalence that was guaranteed by the opportunity for every entrant in a free market to choose between starting a business anew and taking over an existing one. Underlying all this was the principle of perfect capital mobility among alternative uses, a pillar of classical economics. The problem—as the best railroad economist of the time, Arthur Twining Hadley, had explained as early as 1885—was that railroads were a business like no other in the past. Railways services required huge investments in fixed capital that could not be withdrawn. Hence, railroads could continue for a long time pricing their services at a level that barely covered operating expenses. Their rates thus bore no necessary relation to their capital investment: “the rate at which it pays [for capital] to come in is much higher than the rate at which it pays to go out,” Hadley (1886, p. 223) famously proclaimed. This meant that observable market conditions were no benchmark at all for assessing what it would cost to reproduce the infrastructure necessary for providing the service.

By the second decade of the twentieth century, Hadley’s analysis had become common wisdom. “The real cost of transportation,” Whitten observed, “can only be determined by recognizing the only process by which transportation service can be supplied, that is, by devoting capital permanently to the enterprise” (Whitten Reference Whitten1914, p. 424). Capital mobility—and the related profit equalization drive—simply did not hold in the railways industry. Against classical claims, no entry/exit mechanism thus existed to bring either the market price or the return on capital to their “normal” level. The equivalence between the market value of a business and its reproduction cost was consequently lost. Or, as Whitten put it: “Cost of production determined by the reproduction method is largely hypothetical. It is not based on the actual conditions that limit the production of the utility” (ibid., p. 423). The critique seemed definitive. Nevertheless, the Smyth doctrine survived.

V. SMYTH DEFENDED

Though embraced by a majority of scholars, neither the pragmatic nor the theoretical criticisms met with universal consensus. Some authors still defended the reproduction cost method and rejected actual cost. For example, New York lawyer Edward Cashman Bailly observed: “As engineering skill and experience increase, and the courts become more intimate with the valuation problem, the measure of present value afforded by the cost of present construction meets with more favor” (Bailly Reference Bailly1911, p. 545). That engineers were the group more consistently in favor of employing reproduction cost^{Footnote 28} lent credit to the approach and provided the best reply to the pragmatic protest. “As for its accuracy and expediency,” Bailly underlined,

His conclusion was orthogonal to Ripley’s and Whitten’s: “It can hardly be conceived to-day that a valuation would be undertaken without great weight being given to the cost of reproduction” (ibid., p. 547).

This declaration of faith in reproduction cost was supported by the continued allegiance to the method by federal judges. The best example is a 1913 decision by the Supreme Court that, notwithstanding the justices’ refusal to apply it in the specific case, reiterated the centrality of reproduction cost for regulatory matters. Writing for a unanimous court in Minnesota Rate Case, Justice Charles Hughes defined the outcomes of the reproduction cost method as “mere speculation” and “mere conjecture”; unsurprisingly, he declined to endorse them.^{Footnote 29} But the denial referred only to the peculiar application of the method railroad stockholders had invoked in the case.

Railroad owners had argued that, in assessing the reasonableness of the regulated rates, allowance should be made for the fact that the “right of way” (viz., the right to use land for a railway line) cost more than the land’s market value. This is because the purchaser of the right, the railroad company, would find itself locked in (in modern economic jargon) in its bargaining with landowners on account of its specific railways investment and massive sunk capital. The argument’s goal was of course to inflate the reproduction cost and, therefore, obtain higher rates from Minnesota regulators.

The Court dismissed the thesis on two grounds. First, railroads had been granted eminent domain power over landowners precisely to escape these lock-in situations (Minnesota Rate, at 451). Hence, a railroad enjoyed a privilege over the land that already compensated it for the potential loss of bargaining power. Inflating the land’s value by recognizing its higher “railway value” would unconstitutionally sanction a railroad’s right to enjoy privilege-based, supra-competitive profits. Second, Hughes found it impossible to evaluate what the value of “right of way” through a certain land would have been in the absence of the railroad already built upon it. “[I]t is manifest,” he averred, “that an attempt to estimate what would be the actual cost of acquiring the right of way if the railroad were not there is to indulge in mere speculation” (at 452). The existing railroad could not be “obliterated” in order to recalculate the “railway value” of the land. Hence, the reproduction cost method could not be applied in that specific circumstance.

Ripley eagerly jumped on Hughes’s statements and concluded that Minnesota Rate had buried reproduction cost for good (Ripley Reference Ripley1915, pp. 355–356). Yet, this conclusion was far-fetched. The Court had actually proclaimed: “The ‘cost of reproduction’ method is of service in ascertaining the present value of the plant, when it is reasonably applied and when the cost of reproducing the property may be ascertained with a proper degree of certainty” (Minnesota Rate, at 452). Clearly, the justices had no problem with the method itself, provided if based upon demonstrable facts. Their decision simply meant the method could not be applied when grounded upon mere conjectures or, worse, when its use would grant constitutional protection to the profits generated by railroad privileges.^{Footnote 30}

Minnesota Rate is just one among a score of cases where federal courts confirmed their allegiance to the Smyth doctrine. In the early 1920s, Justice Pierce Butler even came close to affirming something the Smyth court never explicitly held, that reproduction cost was the only factor to be considered in determining a railroad’s, or other public utility’s, present value.^{Footnote 31} Keeping into account the inconclusive results reached by the economics of valuation, it is thus unsurprising that the debate between friends and foes of the reproduction cost method continued through the decade.

Clearly, no definitive answer could come from purely theoretical or practical arguments. As Siegel (1984, p. 242) remarks, commitments to values other than the purely economic motivated the supporters of alternative approaches. A broader issue was at stake under the surface of a seemingly technical subject. Under the original Munn doctrine (see Introduction), government regulation had the potential to spread across several branches of economic activity. The “public interest” notion developed by the Munn court was so vague that it could be stretched up to the point of establishing some general forms of public control over the economy. Decades of later jurisprudence by increasingly pro-laissez-faire courts had significantly reduced the risk of undue government activism. Still, competing conceptions of private property versus public welfare returned to the fore under the guise of the apparently neutral choice between reproduction and actual cost. Rate regulation thus remained a key battleground in the war between conservative and progressive approaches to constitutional law—a war that ended only with the so-called constitutional revolution of 1937 (with a coda, for Smyth, until 1944).

The history of economic thought may cast new light on this broader controversy. The early twentieth-century debate over evaluation techniques for regulatory purposes brings to the fore the difference between economics and political economy—and their possible irreconcilability. This is what happened in the case of classical economics and classical political economy, when applied to the analysis of industries that, like railroads, were characterized by massive fixed costs and non-standard forms of competition. Both the difference and the irreconcilability, it is argued, offer an original perspective for assessing the economic foundations of the New Deal upheaval.

By “political economy” I mean here a discourse on the relationships between individuals and society, and between the market and the state, which draws upon fields as diverse as economics, political science, law, philosophy, and sociology—the whole bag, that is to say, of what today goes under the “social science” label. Political economy usually involves a discussion of the general principles pertaining to the economic side of human activity, an assessment of the normative desirability of the different behaviors, and a catalog of the policy actions furthering those behaviors that are desirable and deterring those that are not.^{Footnote 32}

Classical political economy was predicated on faith in competitive markets; classical economics substantiated this faith with, among other things, the profit equalization theorem. Yet in the case of heavily invested industries, the latter could no longer work in the absence of its necessary premise: perfect capital mobility. A new economics was needed to account for new competitive conditions. But while the brightest minds of the younger generation of American economists were building the new theory, the political economy inspiring American judges was still that developed by the previous cohort of classical writers. Nowhere was the chasm between the allegedly eternal principles of classical political economy and those actual market conditions invalidating the theorems of classical economics so large—and its practical consequences so significant—than in the case of the railroad industry.^{Footnote 33}

VI. VALUE IS NOT A FACT

Throughout the long debate over the Smyth doctrine, only one conclusion was unanimously accepted. Everyone agreed that a railroad’s market value could not be used to measure the validity of regulated rates. Market value was exchange value; exchange value was the equivalent to capitalized earnings; capitalized earnings depended on rates, including regulated ones. In other words, value could not be the basis of rates because rates were the basis of value. This circular reasoning, which was implied by railroads’ complaints that a given rate did not pay enough interest on the value of their capital, was denounced by both friends and foes of the reproduction cost doctrine. Whitten’s 1912 treatise may be taken as exemplar:

The issue struck at the heart of classical economics. As Siegel (1984, pp. 246–247) remarks, only a notion of intrinsic value could break the circular relationship between value and rates. This was precisely what the alternative evaluation rules—reproduction cost and actual/historical cost—tried to do: to break the circle in the absence of an objective notion of value. Choosing between the rules thus did not mean selecting the most proper measure of a railroad’s value, but, more correctly, the best substitute for it. As usual, Hadley provided the most lucid statement: “[I]t is quite certain that the attempt to apply the term valuation to the assessment of a rate based on duplication cost as a primary factor or to call the result of such assessment a value results in grave confusion of thought and many practical mistakes. … a cost assessment neither produces value nor measures it” (Hadley Reference Hadley1928, pp. 176, 180).

For classical economists, value was a physical phenomenon, an objective fact: typically, a certain amount of labor or the gravitational attractor of market forces (viz., natural value). This notion of intrinsic value was lost with the transition to neoclassical economics: “Value is not a physical phenomenon,” a leading neoclassical author like Hadley stated, “but a social one—not an inherent quality of objects, but a measure of their adaptation to social needs.” Markets determined the actual value of property, not regulators or courts. It followed that assessing value for regulatory purposes was “essentially a political” endeavor, “not a scientific one” (ibid., pp. 176, 179).^{Footnote 34}

Within the neoclassical framework, the choice of the best rule to make up for the inscrutable value of a railroad was driven by the functional goal of approximating the most efficient allocation of resources without exceeding the boundaries of administrative feasibility. Luckily enough, competition provided a benchmark that catered to both classical and neoclassical sensibilities. The natural value of classical economists and the optimal, first best price of neoclassical theorists converged in the value determined by competitive markets. As we know, here lay the main reason for choosing the reproduction cost rule, for that notion made express reference to—indeed, was defined with respect to—the outcome of competitive markets.^{Footnote 35} The “competitive” label was especially appreciated by American courts, which could find in reproduction cost a familiar reference to classical economic ideas, in particular to the equivalence between competitive and natural outcomes.^{Footnote 36} When in 1898 the Supreme Court embraced reproduction cost as the constitutionally compelled notion of value, it had behind it a whole worldview supporting the choice. It was the worldview of classical political economy, where natural—i.e., competitive—values enjoyed a sort of ethical superiority over all others.

Problems for the Smyth rule arose under both viewpoints, classical and neoclassical. The trouble with the former was that the necessary reference to the outcome of competitive markets was completely lost when the rule was applied to the regulation of industries such as railroads, where no classical competition could ever subsist. This destroyed the only anchor classical economics could provide to ground reproduction cost in the absence of any other objective notion of value.

Functional justifications of the rule, typical of the neoclassical approach, did not work either. American courts could not subscribe to a functional standard because deep constitutional reasons militated against doing so. Functional evaluations were the quintessential matter of legislative policy, upon which courts could not exercise judicial review without infringing the separation of powers. In other words, either the reproduction cost rule found itself a non-functional justification, in some “higher value” deserving constitutional protection—values akin to those of justice and equality promoted by classical political economy—or it would find itself bereft of any logical foundation. The distance separating classical economics from the reality of the railroad industry thus sealed the rule’s fate.

Devoid of its intellectual mainstays, the very notion of property evaluation, not just the specific technique of reproduction cost, began to be seen by more acute observers for what it really was. Value was not an objective reality, a fact to be discovered. It was, as Hadley put it, a social convention. No complex inquiry could establish the fact of a railroad’s value for the simple reason that there was no such fact to be established in the first place. Take Justice Harlan’s list of factors in Smyth. These factors could no longer be a list of types of evidence to be considered in valuation proceedings because, again, no fact existed for the discovery of which evidence was required.^{Footnote 37}

The consequences of recognizing that value was not a fact were momentous. If it wasn’t a question of fact, then value could well be a question of policy. This conclusion did not escape progressive scholars, who understood that the demise of valuation as a fact-finding procedure offered an invaluable opportunity to redirect values, and the policies determining them, towards goals different from traditional ones. A brand new kind of political economy, cleansed of any classical dogma, looked like a concrete possibility.

In the new political economy, value would have no natural connotation; it would be a plastic, goal-based concept policymakers could adapt to varying social needs. At the same time, property and competition, detached from the Smithian system of natural liberty to which they owed their ethical superiority, would lose their absolute status. They would become socio-economic goals like any other, the protection of which would always be contingent on utilitarian, cost-benefit tests. Indeed, if value was a matter of policy, property rights themselves would be stripped of their role as safeguards and would become nothing more than means to achieve socially desirable ends—means that legislators could mold at will.

Columbia economist Robert Lee Hale was among the earliest to recognize that the conclusions drawn from the public utility area could be generalized to all kinds of property. “[I]n regulating the rates of utilities,” he wrote, “the law is trying the experiment in one limited field of turning its back on the principles which it follows elsewhere.” Hale had a clear grasp of where the novelty lay: “We are experimenting with a legal curb on the power of property owners.” The law, he explained, usually “acts on the assumption that whatever income a property owner can get without fraud by virtue of his ownership is legitimately his.” But in the case of public utilities, “standards of what it is proper for an owner to get out of his ownership have to be worked out de novo.” The upshot was radical: “The revision of property rights worked out within the utility field may very well serve as a model, wherever applicable, for the revision of other property rights” (Hale Reference Hale1922, p. 213).^{Footnote 38}

In the field of regulation, the new political economy meant acknowledging that the relation between a regulated company and the regulator was open-ended, amenable to being directed by proper regulatory choices towards any kind of social goal, be it the maximization of investments, consumer protection, or other embodiments of the public interest.^{Footnote 39} It meant recognizing that only two kinds of regulated prices existed: those that were useful in pursuing the regulator’s policy goals, and those that were not. Utilitarian considerations—themselves subordinated to the specific aims pursued by the policymaker—not absolute moral categories, should drive regulation.

VII. “THE ILLUSION OF JURISTIC NECESSITY”

The progressive jurist who most clearly understood that the world of classical regulation rested upon thin air was Harvard lawyer Gerard Carl Henderson. In the final pages of a long 1920 essay, drawing conclusions from his scathing review of three decades of regulatory jurisprudence, he denied that fair value—the pillar of the Smyth doctrine—could ever be “scientifically ascertained by observation and induction.” He deemed it “merely a cloak which conceals a process of arbitrary decision based on considerations of policy” (Henderson Reference Henderson1920, p. 1055).

Henderson explained: “The relation between the public utility and the community cannot be expressed in terms of a simple, quantitatively ascertainable fact, for the relation involves numerous and complex factors which depend on compromise and practical adjustment rather than on deductive logic.” The Smyth doctrine thus rested upon “a gigantic illusion” (ibid., p. 1051): that “there is a fact which can be discovered, if we are only persistent enough in our search for it, and which, once it is found, will provide a mathematical solution of all rate-making problems” (ibid., p. 912). Alas, this fair value, “the fact which for twenty years the [Supreme Court] has been vainly trying to find does not exist.” The notion thus deserved to “be shelved among the great juristic myths of history, with the Law of Nature and the Social Contract. As a practical concept, from which practical conclusions can be drawn, it is valueless” (ibid.; emphasis added).

Recognizing the mythological status of fair value could not be without consequences for the judicial approach to regulation. Henderson dedicated amusing pages to illustrating what would happen in an ideal community requiring a railroad system and having to decide how best to deal with a privately owned provider of the service (ibid., pp. 914–924). There could be only one outcome of the inevitable lengthy negotiations, to which everyone—citizens, shareholders, engineers, and, of course, lawyers and economists—would contribute: the admission that “[t]here is no generally accepted norm to which a contract embodying [the relation between the owners of a public utility and the government] can be assumed to conform” (ibid., p. 925). Thus, “when both lawyer and economist have had their say, there would still remain a substantial variety of possible adjustments, each different from the other, and each nevertheless accomplishing the primary objects which the parties had in mind” (ibid., p. 928). The only indisputable truth in the matter of regulation was that “the decision must turn largely upon considerations of justice and policy,” which can never be exactly formulated, neither juristically nor scientifically (ibid., p. 1055). This, in Henderson’s view, should prompt the Supreme Court to adopt a well-defined jurisprudential attitude: to “widen largely the scope of discretion accorded to the non-judicial body which in the first instance conducts or directs the valuation.” The judiciary should step back, because valuation for regulatory purposes had to “be regarded as a task which is largely legislative and administrative, in which the judgment of Congress, or of an administrative tribunal which adheres to the forms of due process, must within wide limits of discretion be deemed conclusive” (ibid.).

Henderson gave specific directions as to what his appeal for judicial restraint should entail: “If Congress formulates general principles of rate-making or of compensation, the [Court] should not upset these principles merely because it is of opinion that a different formulation would be juster.” Congress’s principles should always stand judicial scrutiny, “unless so clearly unfair as to outrage the common sense of justice.” Ditto for the discretionary deliberations of any administrative commission “appointed to investigate the relevant facts, to negotiate with the railroad interests, and to fix a fair and workable measure of compensation,” whenever “the commission has examined the evidence honestly and impartially, and has given all parties a fair hearing.” Regulation was a matter of “justice and policy,” not of imaginary facts; hence, “conclusions based upon considerations of justice and policy should not be overruled unless utterly beyond the pale of fairness and common sense” (ibid.).

The “illusion of juristic necessity” that Henderson (ibid., p. 910) blamed on the Smyth doctrine went beyond regulatory issues. Treating a railroad’s value as “a pure question of fact” conformed to so-called categorical thinking—the cornerstone of traditional legal analysis against which progressive jurists like Henderson were waging their war at the time.^{Footnote 40} Ascertaining whether a regulated rate guaranteed a fair return on capital to the railroad was conceived of as a dichotomic, yes-or-no question; accordingly, the answer was amenable to a clear, bright-line classification—fair versus unfair rates—of the type categorical jurisprudence so much appreciated. That the decision could be reduced to a fact-based yes or no—never overstepping the bounds of judicial discretion—was precisely the illusion that, according to Henderson, had lured the Smyth court into the reproduction cost doctrine. Exposing the purely conventional character of all evaluation exercises and, from there, the emptiness of any sharp categorization based upon them, meant showing that discretion was unavoidable, that no simple yes-or-no questions existed, that every decision was always a matter of “justice and policy.” In short, it meant striking a decisive blow at the foundations of categorical thinking.

Long before Henderson, another outstanding jurist had invoked the abandonment of traditional categorical analysis along similar lines. In 1894, future Justice Holmes had argued that the complexity of modern society required a jurisprudential move from differences in kind—typical of categorical reasoning—to differences in degree. Conflicts among alternative policies did not lend themselves to sharp, categorical solutions; the very fact that alternatives existed denied the possibility. A judge’s decision as to which of two or more alternative policies was the “right” one could never rest on absolute rational grounds. Law was a social creation, Holmes had proclaimed. Policy outcomes had always to be balanced before lines could be drawn (Holmes Reference Holmes1894, pp. 6–9).

With its trademark cost-benefit analysis, neoclassical economics offered the ideal tool to administer differences in degree and balancing tests. The technique required that weights be assigned to individual elements of cost and benefit. While applying cost-benefit analysis was a technical task, requiring the specialized knowledge of administrative bodies, assigning weights was an eminently political activity, fully within the bounds of legislative powers. Courts should interfere with neither task, limiting their interventions to the few cases when, as Holmes famously averred in his Lochner dissent, “it can be said that a rational and fair man necessarily would admit that the statute proposed would infringe fundamental principles as they have been understood by the traditions of our people and our law” (Lochner, at 76). The legislature, with its appended administrative bodies, rather than the courtroom, was the proper place in which the competing interests affected by public policies were to be balanced. Judicial deference had to be the rule; activist courts, the rare exceptions.

While Holmes never expressly applied this philosophy to regulatory matters, his support for a more liberal substantive analysis of regulation—in which rates would be evaluated according to their “usefulness” with respect to policy goals—was implicit. More generally, underlying his approach was the belief that classical political economy should occupy no special position in judicial review. The Smithian system of natural liberty was only one among several other policy views legislators might entertain, and an outdated one at that. Alternatives existed, including those entailing a more invasive role of the state in economic affairs, accompanied by a parallel limitation of property owners’ freedom of contract. This was indeed the rationale of his epoch-making Lochner dissent.

Historically speaking, the Lochner era would come to a close when a majority of the Supreme Court endorsed the views of progressive jurists like Holmes and Henderson. In the case that officially buried a half-century of laissez-faire constitutionalism, Chief Justice Hughes declared: “The Constitution does not speak of freedom of contract. It speaks of liberty and prohibits the deprivation of liberty without due process of law. … Liberty under the Constitution is thus necessarily subject to the restraints of due process, and regulation which is reasonable in relation to its subject and is adopted in the interests of the community is due process.”^{Footnote 41} This last, rather categorical, statement marked the end of the Lochner era.

Another implication of the progressive view was the exponential rise of the role of economic experts.^{Footnote 42} In the absence of a method to determine values objectively, the value question had to be avoided altogether and alternative forms of legitimization for government intervention had to be found. Society could be an alternative source of legitimacy. Yet society had first to be described and understood. As law historian Morton Horwitz put it: “Social reality—the Is—became the source of the Ought. Description was privileged over prescription. Value was to be discovered from social fact” (Horwitz Reference Horwitz1992, p. 211). This was the social scientists’ task. The twentieth century’s increasing delegation to positivist social science of crucial legislative, or quasi-legislative, duties thus embodied the attempt to evade the value question by elevating expertise and professionalism to the status of main source of legitimization. Neoclassical economics was at the forefront of the task—or, at least, it should have been.

It was at this juncture that the analysis of leading economists like Hadley and Frank William Taussig (neither of whom supported reproduction cost) could have rescued the Supreme Court from progressive attacks, at least as far as regulatory jurisprudence was concerned. A curious inversion took place. Economic theory at its best did offer a lifeline to the conclusions of classical political economy. Even bereft of classical assumptions, like perfect capital mobility, and results, like profit equalization, early neoclassical economics could still provide a rationale for deeming rate regulation unnecessary, thereby defending the system of natural liberty of Smyth and Lochner.

VIII. THE NEOCLASSICAL LIFELINE

Despite the absence of free entry and the substantial immobility of capital, competition still drove the railroads’ rate-making practices, including the most despised of all—and the true reason for a large chunk of regulatory efforts—rate discrimination.^{Footnote 43} This was the main lesson that could be drawn from the pioneering work of scholars such as Hadley and Taussig.

Hadley was the first economist to link the necessary conditions for successful price discrimination —monopoly power and the ability to separate markets—with a simple elasticity specification.^{Footnote 44} In Appendix II of his Railroad Transportation, the relationship between traffic carried and the rate charged was given both a mathematical statement and a diagrammatic illustration by way of a traffic demand curve. The basic assumption of Hadley’s analysis was that railroads are profit-maximizers: “The practical railroad manager has one general principle in this matter. He lowers rates whenever he thinks it will increase net earnings—in other words, as long as it will increase gross earnings faster than it increases operating expenses” (Hadley Reference Hadley1885, p. 261). The principle, which in Appendix II took the form of a differential equation, led the manager to lower rates “until the differential of gross earnings on a particular line of traffic ceases to be greater than (i.e., becomes equal to) the differential of the operating expenses” (ibid.). Under the implicit assumption of elastic demand, Hadley applied the marginalist principle to conclude, both in words and in formulas, that railroads would seek new traffic as long as the increase in revenues brought about by additional traffic exceeded the increase in cost (ibid., p. 263).

Hadley’s was a competitive theory of rate discrimination, in that he explained the phenomenon as an essential part of the competitive process in the railroad industry—of the way, that is to say, competition deployed its beneficial effects even in a market characterized by largely imperfect resource mobility. The railroads’ financial health, their ability to expand service to new markets and lower the average price of transport (thanks to market expansion under decreasing costs), even the possibility itself of their supplying any service at all—everything depended upon the adoption of rational discriminatory practices.

Even more explicitly than Hadley, Taussig foreran what is modernly called the second-best approach to efficient pricing. Today we know that in the presence of a multi-product firm, whose overheads need be covered by total revenues but cannot be assigned to individual products, a proper use of differential pricing can raise total output, spreading the joint costs’ burden among more customers. The idea is that the firm would set each product’s price above its marginal cost, at a level sufficient to cover overheads and obtain a normal return on investment. An optimal set of markups exists that would generate the required total revenue with a minimum loss of output with respect to the first best. These efficient markups depend on the elasticities of demand and are known in modern jargon as Ramsey prices. Whenever total revenues exceed total costs, including overheads and a normal return, the firm is earning supra-competitive profits. Hence, the proper test for monopoly power looks at total revenues and total costs, not at the size of the markup on any specific product: even a very high markup is not by itself an indicator of monopolistic profit.^{Footnote 45}

As early as 1891, Taussig showed a clear understanding of the technique and its implications. In his terminology, the most efficient way to repay railroad investments meant charging “what the traffic will bear.” The principle aimed at maximizing railroad traffic on the basis of the customers’ willingness to pay: i.e., of the different elasticity of the various portions of demand. Like Hadley, Taussig concluded that price discrimination—whatever its basis: freight classification, geography, or the amount of competition along a particular route—was rational economic behavior, and an efficient one at that. Regulators should not interfere with it; competition should be let free to deploy its beneficial allocative effects even in the railroad industry (Taussig Reference Taussig1891, pp. 454–456).

The bottom line of both Hadley and Taussig was of course that competitive forces could work only within a system that guaranteed the most complete protection to property rights and contractual freedom. Defending railroads from arbitrary regulation was not, therefore, a way of upholding their privileges, as progressive critics complained, but, on the contrary, the only available path for having competition destroy those very privileges, albeit only in the long run. In short, explicitly non-classical (viz., neoclassical) economic theory could be employed to buttress the gist of classical political economy, the intangibility of individual rights on property and contract under the neutral policing of market forces.

Yet, the Supreme Court never endorsed the teachings of top-notch railways economics. Lack of proper analytical understanding, or maybe sheer ignorance, made the turn-of-the-century Court unable to catch hold of the lifeline. A hiatus remained, separating the justices’ awareness that many classical economic categories did not apply to the railroad industry and their continuing adherence to the philosophy of the classical system of natural liberty. The latter without the former could not survive, if an alternative basis could not be found elsewhere. The stubborn application of reproduction cost, even in the face of convincing critiques like those by Ripley, Whitten, or Henderson, testified to the Court’s inability to comprehend that, by sticking to a technique the foundations of which lay in discredited classical economics, it exposed its flank to a more devastating attack on Lochner’s deepest roots, those residing in classical political economy.