Working with the classical expressions of demand, supply and market adjustments, the French engineer Jules Dupuit developed a model of competitive market adjustments in different time dimensions that anticipated Marshall's 'period analysis' on all important points. Dupuit's attempt to handle the complexity of economic phenomena in a scientific fashion, by utilizing the ceteris paribus method, and the integration of this method with the 'facts' of production in a 'typical' industry, generated a huge payoff for subsequent generations of economists. The payoff was enhanced by Marshall's refinements, and realized by Marshall's teaching, which took root in so many able students and disciples.
Marshall was less than generous to all those whose contributions were closely related to his own ... Dupuit [for one] ... received but footnote recognition and this not in the right places.
INTRODUCTION
That most famous and influential of neoclassical economic treatises, Alfred Marshall's Principles of Economics, is divided into six books that probe, respectively: the nature and substance of economics; its fundamental notions, such as wealth, income, production, etc.; wants and their satisfactions; the agents of production; the theory of value; and the distribution of national income. It can be arguably maintained, however, as we do here, that the essence of Marshallian partial equilibrium economics resides in Book III (where Marshall establishes the law of demand, its connection to marginal utility, the doctrine of consumer surplus, the notion of elasticity of wants and the relation of price and utility) and Book V (where he advances the partial equilibrium theory of demand and supply for different operational time periods and different 'laws of return'). For it is in these parts that Marshall develops what has come to be known as the neoclassical paradigm, which consists of the following core elements. (1) Demand and supply are fundamental market forces that exist independently as functions (i.e. schedules or curves) but interact to determine equilibrium price and quantity. (2) Demand is governed by consumer tastes, incomes and other prices, and is functionally equivalent to the marginal utility evaluations of consumers. (3) Supply is governed principally by resource availability and technology. (4) Supply adjustments take time, which may be represented by supply functions that vary according to the time allowed for adjustment. (5) Markets may be examined one at a time (i.e. in partial equilibrium), holding irrelevant factors in abeyance by invoking ceteris paribus. This particular line of argument, as distinct from, say, the general equilibrium approach of Marshall's contemporary, Leon Walras, is so intimately connected with Marshall as to bear his name--thus, we speak today of Marshallian economics.
The purpose of this paper is to investigate pre-Marshallian contributions to this line of argument, and in particular to expose the formative ideas of Jules Dupuit (1804-1866), a French engineer whose reputation in the history of economic thought is more or less confined to the narrow topics of demand and consumer surplus. It is our contention, which we support here with textual evidence, that Dupuit anticipated Marshall on practically all points of the socalled Marshallian paradigm. The repository of many of Dupuit's ideas in this regard is a paean to free trade entitled La Liberte commerciale (1861a), a highly important work in interpreting Dupuit's contributions to economic theory, Which nevertheless has been almost totally neglected in previous analyses of his work.[1] Although we attempt to show that the affinities between Dupuit's analysis and Marshall's are numerous and significant, we sidestep here the thorny problem of subjective originality (i.e. influence), and focus instead on objective originality (i.e. priority).[2] The actual impact of Dupuit on Marshall's thought remains problematic, largely because Marshall credited Dupuit outwardly with very little influence. In terms of core principles, general microeconomic theory was extant long before the dawn of the British neoclassical period, which may be said to have begun with Jevons. We therefore seek to enlarge understanding of the state of competitive market analysis a decade before Jevons and a generation before Marshall.
Our exposition proceeds in three stages. First, we compare Dupuit and Marshall on the nature of economic inquiry, the behaviour of markets and the nature of competitive equilibrium, all core elements of the Marshallian paradigm. Then we examine the nature of the influence on Dupuit of his predecessors, most notably Smith and Say. Finally, we assess the relative contributions of Dupuit and Marshall in terms of their respective visions of the role of static theory vis-a-vis actual economic processes.
I. MARSHALL AND DUPUIT COMPARED
Although an engineer by training, Dupuit was an economist by thought and deed. His career as an economist roughly spans the period between the birth of Marshall and the dawn of Marshall's own career as an economist.[3] Thus, the two writers were a generation apart. The intellectual filiations between them remain obscure, but there were remarkable similarities in their respective achievements in economic theory.
The nature of economic inquiry
The raison d'etre of Marshall's economic method was that order must be wrought from chaos. The use of scientific principles in economics was an attempt to feature the regularities of human 'economic' behaviour. In the midst of flailing criticisms by heterodox critics of traditional classical and neoclassical economics (socialists, historicists and Veblenians), Marshall simply observed that men's minds are limited and nature's riddles complex. Complexity, however, should not force economics to non-rational, non-deterministic 'solutions' or analyses. Rather, it should encourage economists to adapt, as far as possible, the scientific method. Near the beginning of his book Marshall (1920, p. 38) argued that 'The laws of human action are not indeed as simple, as definite or as clearly ascertainable as the law of gravitation; but many of them may rank with the laws of those natural sciences which deal with complex subject matter. From this simple analogy, Marshall concluded that economics must emulate the primary method of science, integrating aspects peculiar to economic markets (time in particular) and imitating the ceteris paribus method that scientists used in generating understanding (and predictability) concerning the natural world.
This is precisely the rationale that Dupuit gave for his methodology a generation earlier. Dupuit, too, set out to banish chaos from economic analysis. He maintained that organizing principles--principles that would arrange and set forth thoughts about how real-world markets behaved--were a vital prerequisite to understanding economic phenomena. These organizing principles, like real-world markets, always had to take time into account, but time introduced changes that subsequently affected solutions. Thus, elements had to be added to the problem a few at a time with the others held, to use Marshall's felicitous phrase, 'in a pound of ceteris paribus'. Causes, said Dupuit, must be 'isolated' before any meaningful analysis could proceed. Even so, science has limits, and an understanding of these limits constitutes the scientist's true compass. Dupuit (1861a; p. 23) emphasized his point in the following analogy:[4]
There are weeks when the curious flock en masse to the seaside to see the greatest tides of the century. Science, which has discovered the causes of tides, tells us that on a certain day the sun and the moon will be aligned in such a way as to raise the water far above its normal level, but then it turns out that the tide does not rise as expected. Does this mean that the theory has been weakened, even in the slightest way? Can it be that the influence of the sun and the moon on the tides is suspended for a day? Of course not; this great disappointment arises because the height of the tides depends on certain causes that are amenable to calculation and on others that continue to elude science. The day on which the great tide was expected, the wind may have come from the land, so that the cause which could not be foreseen thwarted that which was predicted. And so it is with economics. (Dupuit 1861a; p. 23)[5]
Technological change, gold discoveries or any one of a 'thousand other circumstances' may occur during the analysis of an economic problem, but that is no reason to deny that a judicious theory can capture the essence of a problem. Since 'men's minds are limited',[6] however, primary 'causes' must be identified and admitted one or a few at a time. On this subject, i.e. the nature and possibilities of economic theory based on rational principles, Dupuit and Marshall were in accord.
Demand and supply: market allocations
In his mature statement of economic theory, La Liberte commerciale, and in his earlier essays, including his pioneer treatment of utility and demand, Dupuit (1861a; p. 1844) appealed to Smith's famous dictum that the division of labour is limited by the extent of the market. The original extension of this conception at Dupuit's hands was that changes in institutions (e.g. technology, inventions, etc.) or changes in property rights (private assignment rather than government ownership or regulation) could lower the transaction costs of exchange, furthering the division of labour and adding to welfare and growth. Given institutions and property rights, the process by which prices and values were formed depended on supply and demand.[7]
La Liberte commerciale in fact contains one of the earliest and clearest conceptions of demand and supply. Dupuit had forcefully established demand theory in 1844, including a modern, but incomplete, list of ceteris paribus requirements.[8] That utility and scarcity are the foundations of value was reaffirmed in La Liberte commerciale, with Dupuit again arguing that
the relationship between value, utility and scarcity is extremely complicated and varied from one object to another, but one can easily comprehend that the more abundant something is, the greater the quantity of it that is offered for sale and the lower its price will be; that the more it is desired, the more demanded, the higher its price. It is this relationship that is called in political economy the law of supply and demand. (Dupuit 1861a, p. 44)
Supply and demand are precise but common-sense concepts that apply to both poor and rich, and, as commonly understood, the quantity demanded of most products exceeds the quantity supplied.
Dupuit posited that prices formed by supply and demand provide information and signals to market participants. Unlocking the circularity in demand and supply arguments found in the writings of his predecessors, Dupuit argued:
When the price of a product falls, it necessarily causes a large increase in consumption, because the price drops below the point where many people wanted the product but could not afford it. Thus a determinate level of consumption always corresponds to a certain price, and by the same token a certain price corresponds to a certain production; for a price that is too high leaves a large part of the product in the hands of the producer or the merchant, who must lower the price in order to dispose of this unwanted surplus; but as this occurs, it is essential to note one thing. Every production requires labor and all labor must be paid. It follows that when the current market price is no longer sufficient to pay the cost of production, the manufacture of the product stops or slows down, and the current price is thereby raised by the diminution of supply; when, on the contrary, the current price yields a large profit to the producer, he seeks to increase his profit by producing even more, and the price thereby falls due to the increase in supply. The cost of production therefore acts as a regulator of the current price, limiting its variations in either direction; but these variations cannot be eliminated altogether, because, inasmuch as production naturally precedes consumption, it is not possible to constrain the forces that tend to maintain the current price at a level suitably above the cost of production. (Dupuit 1861a, pp. 45-6)
The benefits of exchange are achieved in an iterative process through which buyers and sellers continuously exchange information. As Dupuit argued,
In each and every transaction, the current price of a product depends on its supply and demand; on the bargaining that ensues between the merchant and the buyer, in which the one wants to sell as high, had the other to buy as cheaply, as possible; specifically, it is the need of the one to sell, the other to buy, where buyer and seller exploit the knowledge of each other's needs, of their financial difficulties even, as each tries to turn the situation to his profit; without regard, one or the other, to the cost of production. The laws of economics are ruthless. . . . Dupuit 1861a, pp. 46-7)[9]
Markets provide, moreover, a critical allocative function--one supporting the economic allocation of all products during periods of plenty and periods of scarcity. Artificial interferences in markets, such as legal minimum wage impositions, misallocate resources and, in labour markets, create unemployment and out-migration to other lines of work or other countries.[10] Without legal obstacles, competitive equilibrium in output and input markets allocates products, services and inputs to those most able and willing to obtain them in free and open exchange. Interference in these processes, Dupuit argued, always have untoward effects.
Competitive market adjustments
Dupuit's conception of markets underpinning the theoretical and practical applications of supply and demand--was based upon experience and observation. His theory held that tariffs, tolls, and all monopoly-induced prices above marginal cost restricted mutually beneficial exchange and reduced economic welfare. Inherent in this view is a concept of equilibrium--whether or not it was ever achieved (or achievable). Most certainly, Dupuit (1853a, p. 344; 1853, pp. 53 4) conceived of equilibrium as a theoretical, not necessarily a practical, construct. As we shall see below, Smith set forth a similar discussion of market v. natural price but without a clear understanding of the interrelationships between marginal utility, demand and scarcity. By contrast, Dupuit pushed the analysis of competitive market adjustments far beyond the classical writers and into the modem age. In particular, he was (along with other engineers) the first to recognize and understand the nature of production costs and the opportunity costs of using various resources.
Nineteenth-century French engineers had an intimate knowledge of railway cost functions. On the rails at this time as much as three-fourths of total costs were fixed costs. One consequence of these fixed costs was that the opportunity costs of using different resources varied over time. The distinction between short- and long-run analytical periods in economics had its origins in this empirical fact. Dupuit was thus made aware of the time element in actual markets and its role in abstract theorizing. When combined with the methodological principle of ceteris paribus, he was able to forge a conception of competitive adjustment that compares favourably to Marshall's famous discussion in his Principles three decades later.
Constant-cost industries. The finest example of market adjustment principles in Dupuit's writings is contained in his analysis of the effects of protection and prohibitive tariffs in La Liberte commerciale, where he took up the subject of a prohibitive French tariff on English iron. He posited that, owing to the tariff, the same quantity of iron cost 250 francs in France and 150 francs in England. What effects would the total removal of the tariff have in the two countries if England (ultimately) had the ability to supply all of France's demands? Dupuit gave an answer to this question that established the significance of time periods of production as a general proposition: 'In all economic questions, one must distinguish between two periods, a transitional state [la transition] and a final state [l'etat definitif]' (1861a, p. 134). In other words, both short-run and longrun analysis is called for in every change of regimes (in this case the elimination of the French tariff).
Dupuit explained how at first, in the short run, the increase in England's production is determined by the existing capacity of its furnaces. The consequence of this production limit in the short run was that the increase in the English output of iron would be limited.
Figure 1 illustrates our own construction of Dupuit's argument. (Neither Dupuit nor Marshall illustrated adjustment theory with appropriate graphics.) The short-run result would be a small reduction of price in France resulting from the new imports of iron; but that would correspond to a rise of iron prices in England (to point b in the figure). Why? Because 'The high price in England is due to the fact that goods which are exported are not replenished by new production; the low price in France is due to the quantity of iron imported there' (Dupuit 1861a, p. 134). The demand for iron in England shifts rightward as in the figure. Without a large change in English production, the price of iron rises in England and falls in France. In consequence, consumption rises in France and falls (somewhat) in England. Determination of the exact price changes would be difficult since it will 'depend on the extent of the wants or demands in the two markets', but it would (in the short run) move between 150 and 250 francs.
The highly original part of Dupuit's analysis involves the nature and causes of the adjustment process in England. The short-term rise in price had critically important effects in the iron market:
the rise in England would increase profits and thereby repay capital investment in the iron industry at a higher rate than investments in other industries. Consequently more resources would be attracted to it, including more workers, because the industry can afford to pay higher wages. New forges would be built and more minerals mined, thus allowing exports to be increased. But, in proportion as production is increased, the price of iron will fall in the two markets; that is to say, the fall will continue in France and the rise will be arrested in England. This, then, is how the transition occurs: continued price reductions in France; momentary increase in England, followed by a continual decrease. (Dupuit 1861a, p. 134-5)
What happens next, as Marshall was later to explain in a discussion of the hypothetical fish market, depends entirely on resource and cost conditions in England's iron producing sector. Entry will occur because of the newly profitable conditions within English iron firms (see Figure 1). Higher wages will draw in more iron workers and time-consuming capital construction, resulting in more intensive mineral exploration and production. The result of the augmented production will cause a further reduction of price in France and a fall from the higher initial price in England. How far would price fall in England? Dupuit answered:
Evidently [to] the former price of 150 francs in England; for as long as it is above this level, English labor and capital will continue to be engaged in this production, seeing that we assume that their present price provides sufficient remuneration. We also assume, and this is essential, that the new minerals and fuels brought into use are as abundant as those that are used today in similar circumstances and places. (Dupuit 1861a, p. 135)
Prices rise in the short run because of the inability to move resources instantly to the production of iron. After the long-run adjustment, prices may return to their initial level (in England) if the new resources--labour and capital--are as productive as the initial inputs.[11] Dupuit thought that these constant-cost conditions would be replicated or approximated for a wide variety of merchandise.[12]
Increasing-cost industries. Marshall's version of the foregoing discussion by Dupuit became the standard neoclassical explanation of competitive adjustments for constant-cost industries. But Dupuit brought the analysis even closer to Marshall by discussing the entire range of possible long-run supply functions, such as those depicted in Figure 2. What if, for example, the English tariff on French wines were lowered or eliminated? Exports of French wine would increase in the short run, raising the price of wine in France and lowering its price in England. The nature of the cost function would differ. As with iron, 'higher profits from wine making would induce the French to devote a greater part of their farmland to viticulture; after, a, certain number of years the wine harvest would increase at the expense of other less profitable agriculture, and the price of wine would finally fall, but where would it stop?' At the old price? Dupuit thought not, because 'the new long-run price [prix definitif] must pay the expense of vine cultivation on the last lands cultivated. But these lands are much less productive than those in cultivation today ... [and must] ... raise price above the initial one for wine' (1861a, p. 136).[13] Such a situation would conform to the long-run equilibrium price and output at point c in Figure 2 after all adjustments. Clearly, Dupuit was describing an increasing-cost industry (LRS[sub t] in Figure 2) in the manner that subsequently came to be associated with Marshall.
Dupuit recognized even further limits on supply response, should resource reallocation be difficult or impossible vis-a-vis a change in demand. He wrote: 'If, instead of wine in general, we consider a particular vineyard, where nature puts limits on output, the result would be entirely different. The demand for Chambertin's Clos-Vougeot rises and one is unable to produce appreciably more; for products of that nature the result of free trade would be the immediate rise in price [i.e. in the short run] which would continue into the long run with no subsequent decline' (1861a, p. 137). The long-run supply curve would, in such cases, approximate the vertical (LRS[sub 2] at point e in Figure 2).
When Dupuit's analysis in La Liberte commerciale is appreciated to its fullest, it becomes obvious that there were few if any elements of Marshall's presentation of competitive equilibrium that Dupuit did not entertain earlier. But the affinities do not end here. Dupuit's priority extended to the fundamental methodological principles underlying the economic theory of competition (and indeed, all of microeconomic theory).
Ceteris paribus and competitive equilibrium: theoretical and actual
Dupuit enlisted the scientific method against those 'economists who engaged in reckless predictions' (1861a, p. 138), using his theory of competitive equilibrium as a case in point. His argument is as follows. Observation of price and quantity changes in the short and long runs (as described above) may produce observed changes which do not square with the theoretical predictions of supply and demand.[14] If, for example, the introduction of free trade (elimination of the tariff) creates 'the perfection of metallurgical processes or the discovery of richer minerals' and price falls to 120 francs for iron, or, conversely, if the infusion of Californian or Australian gold into Europe creates a rise in iron price to 300 francs, 'there would be no contradiction of the principles of competitive equilibrium (Dupuit 1861a, pp. 138-9).[15] In an exact parallel to Marshall's defense of the ceteris paribus method, Dupuit argued: 'When an effect depends on many causes, it is impossible to calculate exactly in a way that takes full account of the interdependence of phenomena. But it is appropriate for science to isolate each cause and calculate its effects separately; it is, in fact, the only means of study and discovery that science has' (1861a, p. 139).
This analysis makes pristine Dupuit's views on the relation between observation of economic processes and the role of theory. Theory--in this case, competitive theory--is essential to economic science for the understanding of economic processes. But causes must be isolated and examined one at a time in order to correctly and more fully understand effects. Dupuit's integration of the ceteris paribus (scientific) method with time periods of production and competition completes the 'neoclassical' theory of competitive adjustment.
II. DUPUIT'S PREDECESSORS ON MARKET ADJUSTMENTS
Both Marshall and Dupuit acknowledged Smith's formative influence on their thinking about competitive market adjustments. But Say was also a major influence on Dupuit, who was obviously more in tune than Marshall with the French tradition of economic thought. We therefore turn our attention, however briefly, to the early developments of the theory, in order to place Dupuit's formulation in historical perspective. Before Dupuit, the high points of the discussion were provided by Cantillon, Smith and Say (with Cournot's technical concept actually in strong contrast to the other three).
Cantillon-Smith- Say and the price market adjustment mechanism
The earliest writer to the set the foundations of market adjustment theory in a Newtonian framework describing the actual process by which price/cost margins were narrowed in the real world was Richard Cantillon. It is not clear whether Dupuit knew Cantillon's writings first-hand, but it is evident that Cantillon influenced the Physiocrats, who supplied a major part of the French heritage in economics that was absorbed by J. B. Say. Cantillon conceived the economy as an amalgam of interconnected markets that operated to achieve a kind of equilibrium. Market institutions evolve over time in response to 'need' and 'necessity', which, in turn, bind the inhabitants of society into a reciprocal network of interacting agents. The equilibrating force of the market place is supplied by a class of entrepreneurs who conduct 'all the exchange and circulation of the State'. These entrepreneurs are similarly bound in reciprocity, as they 'become consumers and customers one in regard to the other'. The number of entrepreneurs is therefore regulated by the number of customers, or by the total demand for their services, which amounts to the same thing. These self-interested agents make decisions regarding market transactions under uncertainty. In particular, the entrepreneur buys at a certain price in order to resell 'at an uncertain price, because he cannot foresee the extent of the demand' (Cantillon 1931, pp. 47-57).
While Cantillon's conception of price formation was preoccupied with the notion of intrinsic value (the measure of the quantity of land and labour of given quality entering production), he was nevertheless keen to point out that: 'there is never a variation in intrinsic values, but the possibility of proportioning the production of commodities and merchandise in a State to their consumption causes a daily variation, and perpetual ebb and flow of market prices' (Cantillon 1931, p. 31). In an early 'Austrian' view of the matter, Cantillon perceived that information possessed by market participants and the degree of coordination of individual plans affected movements from and to the point of 'equilibrium', where market price equals 'intrinsic price'.
In addition to his perception of output markets, Cantillon understood how price changes in the output sector of the economy reverberate through the input sector. His analysis is rich in suggestions of self-interest as a motive force, relative prices as signals to adjust resource use, and opportunity cost as a foundation for economic decision-making. Indeed, in a meaningful prologue to Adam Smith's discourse on the market mechanism, Cantillon referred repeatedly to 'natural' forces; he considered the market relationships he outlined as relatively immune to intervention, and he generally eschewed ethical judgments of economic outcomes.
Adam Smith did little to improve on Cantillon's discussion, although his imprint on future discussions of market adjustments was indelible. Indeed, one has only to substitute 'natural' in place of 'intrinsic' to appreciate the close affinity between Cantilion's and Smith's discussions of market adjustments. One notable improvement made by Smith was to make time a more explicit part of the analysis. Smith did not go so far as to establish the Marshallian distinction between short run and long run, but he clearly paved the way for Marshall's later formulation. He made it clear, for example, that market price is a temporary phenomenon (in which the full range of supply and demand come into play) whereas natural price is a long-run (ideal) concept, in which cost considerations dominate. In addition, Smith, like Cantillon, extended the theory of market adjustments so as to envelop both product and factor markets.[16] Importantly, moreover, Smith recognized the abstract character of his analysis, fully understanding that the narrowing of price-cost margins in theory could be forestalled in practice by changes in exogenous forces, including government regulation.[17] indeed, a close reading of Smith suggests that the supply--demand apparatus was fairly advanced near the end of the eighteenth century.
Say's discussion of market adjustments did not advance substantially beyond Smith, despite some early and prominent concessions to utility (Say 1819, p. 103). Approving of the distinction between market price and natural price, Say broached the subject of market adjustments to disequilibrium on much the same grounds. In the case where market price (prix courant) exceeds cost of production, for example, Say (1826, p. 298) declared: 'as this difference in the relation between the cost of production and the current price of the product holds out a prospect of larger profit than ordinary in this particular channel, it naturally attracts a larger proportion of productive agency, the exertion of which, by enlarging the supply, reduces again the current price to a level with the bare cost of production' (1826, p. 298). As with Smith, however, the time element is not specified in any precise manner. We therefore proceed to Cournot, from whom Marshall claimed to have derived much of his own theory of demand and market adjustments.
Cournot and the mathematical theory of exchange
Further progress on the part of the early writers to specify an adequate theory of supply and demand was no doubt hampered in part by the lack of mathematical technique. The early sponsors of the supply-demand theory had difficulty sorting out and integrating its component parts. As Schumpeter noted, 'They talked of desires or desires backed by purchasing power, of "extent" of demand and "intensity" of demand, of quantities and prices, and did not quite know how to relate these things to one another. The concepts . . . of demand schedules . . . and supply schedules . . . proved unbelievably hard to discover and to distinguish. . . .' (1954, p. 602). All of this changed, of course, with Cournot.
Although it is generally appreciated that the theory of demand forms the backbone of Cournot's microeconomics, the specific relation of Cournot's demand theory to the rest of his analytical, apparatus is peculiar in several respects. In the first place, Cournot's demand curve is devoid of utility considerations, which means that it constitutes a sharp break with Say. It also must be distinguished from Marshall's demand curve, even though Marshall acknowledged Cournot's influence on his thinking. In the second place, Cournot's demand curve is essentially passive in the matter of market adjustments. Cournot explicated the nature of market adjustments progressively, beginning with the absence of competition (monopoly) and moving through the intermediate stage of limited competition (duopoly) onwards to unlimited competition. In the first two instances, demand is given, and producers adjust to the nature of demand and the extent of rivalry in the marketplace. The chief characteristic of the demand curve faced by an individual firm in unlimited competition is that its slope becomes increasingly flat as the number of firms competing increases, so that, in the limit of pure competition, no single firm has any appreciable impact on market price--or, as we say today, each firm is a price-taker rather than a price-maker.[18] Despite all its brilliance and mathematical elegance, Cournot's theory says little about the nature of market adjustments over time--the area in which Marshall subsequently established pre-eminence, and in which Dupuit's contribution heretofore has been overlooked.[19]
There are no mentions of Cournot in Dupuit's work. What the two pioneers shared is a vicarious commitment to the mathematical method and a desire to make economics more scientific. In terms of how markets actually behaved, however, Cournot and Dupuit each held entirely distinct visions, despite the fact that both had excellent technical backgrounds and were eager to use technique in the service of economics.[20] By contrast, Dupuit frequently referred to Smith and Say, whose works he knew well. It was on this 'classical' foundation that Dupuit erected his own theory.
III. CONCLUSION
Working with the classical expressions of demand, supply and market adjustments, Dupuit developed a model of competitive market adjustments in different time dimensions that resembles Marshall's on all important points. Marshall focused on hypothetical markets, whereas Dupuit concentrated on real-world applications. Both analysts bifurcated costs, Marshall into prime and supplemental, Dupuit into fixed and variable. This division of costs seems to have dominated their respective views of supply adjustments in different time dimensions. One need only juxtapose Marshall's famous discussion of a hypothetical fish market--one of the most telling of his entire book--to Dupuit's discussion of the effects of tariffs on the iron market (in England and France) and the wine market (in France) to appreciate the theoretical affinities. While encased within a practical example, Dupuit's insistence that ceteris paribus be relaxed a bit at a time--using possible real-world adjustments of labour, capital (and in the case of wine, land)--produced an illustration of his adaptation of the scientific method that compares favourably to Marshall's description of 'adjustments' in the fish market. In the final analysis, Marshall was no more precise than Dupuit concerning the items in his pound of ceteris paribus. The same holds true, in a close comparison of their respective models, for Marshall's analysis of ceteris paribus conditions over the short-run or the long-run adjustment.[21]
We therefore assert that Dupuit invented a theory of competitive market adjustments and an organized theory for analysing market processes in situations involving constant-cost and increasing-cost industries before Marshall. New ideas inevitably involve a wrenching away from the old, but the break is not always dramatic. Like the classical economists of his day, Dupuit considered manufacturing (in l'etat definitif) to belong mainly to the constant-cost category, even though he was aware that resource costs could rise as more labour and machinery were added to production. Likewise, following Ricardo's lead, he considered agriculture to be characterized by increasing costs (with vertical long-run supply a possibility in the case of 'unique' goods or services). Dupuit's attempt to handle the complexity of economic phenomena in a scientific fashion--by utilizing the ceteris paribus method--and the integration of this method with the 'facts' or production in a 'typical' industry generated a huge payoff for subsequent generations of economists.[22] But the payoff was enhanced by Marshall's refinements,[23] and was realized by Marshall's teaching, which took root in so many able students and disciples.
It has long been axiomatic among historians of economic theory that, while Marshall's book did not contain great originality, partly because it was published at the noontime of the early neoclassical era, it nevertheless improved virtually all received economic concepts that it touched. The same is true of Marshall's treatment of competitive equilibrium and adjustment analysis. In the latter instance--the veritable centerpiece of his book--Marshall's presentation may now be seen to have provided filigree for the basic but substantial discussion given earlier by Dupuit. Marshall provided, for example, a lengthy discussion of what 'normal' and 'normal price' mean in four rather than two hypothetical time periods. Thus, Dupuit's 'natural' price becomes Marshall's 'normal' price; and Marshall (1920; pp. 363-80) added the 'market period' and the 'secular' long run to short-run and long-run analysis. The former is most certainly an improvement over Dupuit's discussion (although it is easy enough to read this 'instantaneous' period into his writings), but the latter appeared to get Marshall into analytical difficulties, as did his suggestion that decreasing supply price could characterize competitive equilibrium in the long run.24 The same is true of Marshall's focus on the long run, and especially the secular long run.[25] Additionally, Marshall invented a concept of the 'representative firm'--one that got its 'fair share' of external and internal economies and diseconomies. But a generation later, economists were still trying to untangle the analytical problems that these seemingly innocuous inventions provided. In retrospect, the Marshallian theory of short-run 'quasi-rents', which is emulated to great extent by modern theories of the firm (Alchian and Woodward 1988), held far greater analytical promise than the fiction of the representative firm. Dupuit certainly did not get this far in his own analysis, although he understood the creation and dissipation of rents in the form of consumers' and producers' surplus.
Dupuit's achievement was to express objective originality in the matter of the central proposition of what later came to be known as the 'Marshallian' theory of competitive market adjustment. Like Marshall, he defended the development of the 'pure logic' of static equilibrium 'states' as necessary ground for analysing the 'real world'.[26] In doing so he produced an engine of analysis which made the dissection of real world markets more 'scientific'. Despite Marshall's well-known aversion to giving credit to his intellectual forebears, we do not question his veracity when he said that he knew nothing of Dupuit as late as 1873 (Groenewegen 1995, p. 162).[27] Our purpose has not been to tarnish Marshall's reputation regarding the establishment of modern microeconomics but to burnish Dupuit's, who has heretofore been known mainly as the inventor of particular microeconomic phenomena (e.g. marginal utility, demand theory, consumer/producer surplus, monopoly and price discrimination).[28] Dupuit's excursion into a theory of competition which included time, short-run and long-run input and output adjustments and a concept of ceteris paribus, complements and completes his other important technical achievements and broadens respect for him as an economist. His theory of market adjustments was an early, substantial and luminous central panel in the rich mural that comprises mainstream contemporary economic theory.
NOTES
1. When mentioned at all, the work is portrayed as simply a primer on free trade. Parts of the book were initially published in the Revue europeenne in 1860 (Dupuit 1860). The book-length treatment was most certainly an embellishment of these essays, which were possibly conceived of as part of a larger treatise. (Dupuit refers to such a treatise on political economy as early as 1844.) According to Dupuit's biographer, in 1860 the French Ministry of Commerce urged Dupuit to publish these essays in book form in order to justify changes in the French customs system (Fleury 1867, pp. 42-3). The Minister was expecting 'political studies' of particular industries (coal, wine, etc.) but got a full-blown and rigorous demonstration of generalized free trade instead. Dupuit (1861a; pp. 178, 221), however, was fully acquainted with the 'political side' of free trade and urged Peel-style tariff reforms in France.
2. The distinction between subjective and objective originality is Schumpeter's. Schumpeter (1954, p. 839) cites consumers' surplus and the diagrammatic method of presentation as indications of Dupuit's priority over Marshall, but in our view this judgment underestimates Dupuit's priority in the areas of utility-demand theory and monopoly pricing. Despite his nearexhaustive command of economic literature, Schumpeter was capable of occasional lapses, as when he claimed that 'nobody had anything useful to say on monopolistic pricing [after Cournot] until Marshall published his masterly version of Cournot's theory' (1954, p. 960).
3. Marshall was born in 1842, two years before the publication of Dupuit's pioneer work on demand and marginal utility. Dupuit died in 1866, the year before the start of Marshall's self-proclaimed 'economic apprenticeship' (cf. Groenewegen 1995).
4. All translations of Dupuit are the authors' except Dupuit (1844) and (1849).
5. This passage is reminiscent of Hausman's (1992, pp. 33-53) discussion of J. S. Mill's methodology, but Dupuit (1861b, pp. 111-17) did not share Mill's view that economics is an inexact science. A few years later, Dupuit (1863, p. 476) endorsed the Bentham/Mill view of property rights, but it is unclear if, and to what extent, Mill influenced his views on economic method.
6. That this was Dupuit's method from the beginning of his inquiries into economics is fairly obvious. The matter of demand determinants is illustrative. In an early critique of Dupuit's conception of demand theory, a fellow engineer, Charles Bordas, made the point that an expression of the maximum sacrifice that an individual would be willing to make for some product was too variable to ever become the subject of science. As Bordas put it, 'what theory can one establish on so variable a basis and which depends on the taste as well as the means of each consumer?' (1847, pp. 41-2). Dupuit's answer attached ceteris paribus conditions to demand curves. As he noted, 'not only does the [maximum demand] price depend on the income of that person, as Mr. Bordas observes, but on his taste for meat, on his hunger, on the price of other nourishing commodities, and on a thousand other circumstances impossible to enumerate in a complete manner; but all these circumstances do not mean that this price does not exist for each object, for each person and at each instant' (1849, p. 184). Marshall, in his defense of the assumption of a 'stationary state', likewise describes the necessity of assuming that 'the character of man himself is a constant quantity' (1920, p. 368).
7. Dupuit's reliance upon demand and supply extended to income and wealth distribution as well: 'Taking a look at the distribution of wealth in society, one recognizes quickly that it is accomplished by virtue of the economic law of supply and demand, which gives 100,000 francs of income to a dancer, and 3,000 francs to an upright judge and to an elightened professor' (quoted in Fleury 1867; p. 15).
8. Dupuit was the first economist to link demand explicitly with marginal utility and to fully analyse alternative demand shapes. He based individual demand curves on maximum utility estimates of individuals, and 'summed' demand curves across individuals to arrive at market demand curves with shapes similar to rectangular hyperbolas (Ekelund and Thornton 1991). Dupuit's list of independent variables included, in addition to own quantity, wealth of the consumer, prices of related goods and a 'taste' parameter. In 1861 the list was expanded to include 'the quantity of money that each consumer has at his disposal' (Dupuit 1861a; p. 44).
9. We note here that Dupuit's concept of market functioning did not include an assumption of 'perfect information' (or anything approaching it) in the working out of real-world markets and in the rough and tumble of human economic exchange (higgling). Dupuit, unlike a host of writers to follow, chief among them William T. Thornton, was able to understand that the theory of competition--including the theoretical forces of supply and demand that support it--was but a necessary, sophisticated and scientific method of organising thinking about what must always and ever be a process. According to Schumpeter (1954; p. 975), Alfred Marshall himself dichotomized competition as 'theory' and competition as a 'rivalrous process', especially in analysing the 'infinite variety of market patterns' between pure competition and pure monopoly, even if he did not always keep such distinctions clear.
10. Dupuit was specific about the effects of a legal minimum wage (which he assumed to be above the equilibrium wage level): 'Under an arbitrary regime in which the minimum is set without regard to the needs of the workers, if the fixed rate is above that which is consistent with the competitive wage, a certain number of workers will be forced into unemployment, because the employers cannot afford to hire all the workers available; if it is below, there will not be enough workers for the available work. When competition pushes the wage below their needs, poverty drives some workers into another profession or another country, if possible, and the wage rate rises in consequence of the diminution of supply' (Dupuit 1861a, pp. 47-8). This was also the basis for his opposition to labour unions.
11. In terms of Figure 1, a demand increase augments quantity and raises price from points a to b as variable inputs are applied to iron production. Entry brought on by higher profits and wages could push the long-run adjustment to point, c, creating a horizontal 'constant-cost' long-run supply curve.
12. According to Dupuit, 'What we have said about iron can also be said about many other goods, whose costs of production do not increase with increased output. For these goods there is a persistent tendency for higher prices to be reduced to lower levels' (1861a, p. 135).
13. Dupuit also notes that tastes may change in favour of French wine in England, which would increase demand.
14. Dupuit fully understood the classical dichotomy separating value theory, i.e. the theory of relative prices, from monetary theory, i.e. the factors explaining the price level. Indeed, the quantity theory of money and the critical distinctions between nominal and real prices are all discussed in La Liberte commerciale (Dupuit 1861a, pp. 48-9, 51-4, 61).
15. Indeed, such statements regarding new technology strongly suggest that changes in technology alter the basis under which supply curves can be drawn. Marshall (1920, p. 379n) noted the same point in reference to secular long-run periods when a reasonable ceteris paribus cannot be maintained. There are, in short, limits to partial equilibrium analysis.
16. Supporting passages regarding natural and market price from the Wealth of Nations are so well known that we do not repeat them here, but see Smith (1937, pp. 56-9). As only a sample, consider how Smith described input market adjustments when the quantity of a good offered for sale exceeds its demand: 'Some of the component parts of its price must be paid below their natural rate. If it is rent, the interest of the landlords will immediately prompt them to withdrawn a part of their land; and if it is wages or profit, the interest of the labourers in the one case, and of their employers in the other, will prompt them to withdraw a part of their labour or stock from their employment. The quantity brought to market will soon be nO more than sufficient to supply the effectual demand. All the different parts of its price will rise to their natural rate, and the whole price to its natural price.'
17. We find one of his observations most telling in this regard: 'though the market price of every particular commodity is in this manner continually gravitating, if one may say so, towards the natural price, yet sometimes particular accidents, sometimes natural causes, and sometimes particular regulations of police, may, in many commodities keep up the market price, for a long time together, a good deal above the natural price' (Smith 1937, p. 59).
18. 'The effects of competition have reached their limit, when each of the partial productions, D[sub k] is inappreciable, not only with reference to the total production D = F(p), but also with reference to the derivative F'(p), so that the partial production D[sub k] could be subtracted from D without any appreciable variation resulting in the price of the commodity' (Cournot 1838, p. 90).
19. One reason for the 'time-bound' nature of Cournot's discussion of competition is that Cournot failed completely to place the analysis of cost within the context of a 'long' and 'short' run. His discussion of cost and maximization (1838, pp. 58-61) take on a mechanical and highly abstract quality despite the fact that at one point he identifies a category of costs called 'general expenses' (presumably in contrast to other 'variable expenses'). He was, in short, unable to provide a theory of the factors influencing cost and supply without a firm grounding in opportunity cost and time periods of production.
20. By personal circumstances, occupational choice, and individual disposition, Dupuit was even more steeped in economic investigation than Cournot, a fact that stems from his position within the Corps des Ingenieurs des Ponts et Chaussees.
21. In his focus on the short run, Marshall (1920, pp. 370-1) gives 'full attention to such influences as the inducements which good fishing wages will offer to sailors to stay in their fishing homes for a year or two, instead of applying for work on a ship'. In addition, the use of 'old fishing boats' and other boats adaptable to fishing is a consideration in explaining the short-run 'normal' price of fish. These passages may be compared directly with those of Dupuit (discussed above). We find only insignificant differences in thrust and emphasis.
22. Any random sample of beginning textbooks provides the truth of this statement. That is, even in the modern world of the Coase-Alchian-Williamson firm and transactions cost (Alchian and Woodward 1988), students are typically introduced to competitive equilibrium and adjustments via the theory invented by Dupuit.
23. It may be recalled that Marshall made much of the non-uniqueness of the abilities and talents necessary to become a fisherman in the long run. Further, neglecting the implications of a possible 'common-pool' problem with fishing, Marshall even argues that 'economies of scale' may characterize fishing boat production (capital), creating a long-run decreasing supply price for fish.
24. Marshall indicated the possibility of decreasing cost industries, that is, a negatively sloped LRS curve. That possibility embroiled his analysis in a lively debate later over whether such a curve was compatible with competitive conditions. Reliance on falling input prices as the deus ex machina to explain declining LRS requires some explanation for decreasing costs in those markets. Use of internal or external economies as the principal explanation, as Blaug (1985, p. 381) has argued, provides little insight into the analysis.
25. Marshall (1920, p. 379n) toyed with the notion of the 'secular long run' where technology and other factors change, indicating the dangers of using the 'statical method' in these cases. As Schumpeter (1954; p. 995) noted, these and other cases--those involving industries of increasing return cum technological innovations--merely 'depict historical processes in a generalized form'.
26. Schumpeter says that 'Marshall, though offering plenty of material about the theory of shortrun processes, always emphasized primarily the properties of the long-run normal without, perhaps, making it sufficiently clear that what he really meant was the pure logic of the economic process rather than any state of things that will actually emerge at any future time' (1954, p. 1172).
27. Nor is there reason to question Whitaker's judgment that Dupuit had no influence on Marshall's theory of consumers' surplus. (However, see Stigler 1990, p. 8). But Whitaker is incorrect when he argues that Dupuit's work was 'completely unknown in England until Jevons came upon it in the later 1870s' (1975, Vol. 2, p. 281). That may have been so for mainstream British economists but it was not the case for all policy experts. In 1865 Edwin Chadwick presented what is likely the first synopsis of Dupuit's work (primarily on demand theory) in the former's Address on Railway Reform. Chipman is also incorrect when he notes that, whether the idea was original with Marshall or not, he 'was the first to incorporate it into economic theory as the basic tool of analysis of welfare problems. It was fundamental to his [Marshall's] theory of public finance and his theory of monopoly' (1990, p. 178). A reading of Dupuit's essays from the 1840s reveals that these were the primary applications that Dupuit gave to the concept of consumers (and producers') rent. Indeed, his 'taxation theorem' became the acknowledged basis for the taxation and welfare theories of F. Y. Edgeworth and Harold Hotelling. Marshall (1920; p. 101) does credit Dupuit with first 'formally describing ... small increments of price as measuring corresponding small increments of pleasure' and, after all, the concept of a 'surplus' is not very far away from that operation.
28. What Marshall did and what he did not know of earlier writers' ideas and mode of presentation is and will undoubtedly continue to be the subject of dispute. Perhaps the most interesting case, except (for us) that of Dupuit himself, is the nature of Jevons's impact on Marshallian economics, as is the extent and nature of two of Marshall's acknowledged antecedents, Cournot and von Thunen.
REFERENCES
ALCHIAN, A., and WOODWARD, S. (1988). The firm is dead; long live the firm. Journal of Economic Literature, 26, 65-79.
BLAUG, M. (1985). Economic Theory in Retrospect, 4th edn. London: Cambridge University Press.
BORDAS, C. (1847). Of the measure of the utility of public works: answer to the article of the chief engineer, Mr. Dupuit. Annales des Ponts et Chaussees, 2nd series.
CANTILLON, R. (1931). Essai sur la nature du commerce en general, ed. and trans. H. Higgs. New York: Augustus Kelley, 1964.
CHIPMAN, J. S. (1990). Marshall's consumer's surplus in modern perspective. In J. K. Whitaker (ed.), Centenary Essays on Alfred Marshall. New York: Cambridge University Press.
COURNOT, A. A. (1838). Researches into the Mathematical Principles of the Theory of Wealth, trans. N. T. Bacon. New York: Macmillan, 1927.
DUPUIT, J. (1844). On the measurement of the utility of public works, trans. R. H. Barback. International Economic Papers, 2 (1952), 83-110.
----- (1849). On tolls and transport charges, trans. E. Henderson. International Economic Papers,(1962), 11 (1962), 7 - 31.
----- (1853a). Phage. In Dictionnaire de l'economie politique, ed. C. Coquelin, Vol. 2, pp. 339-44. Paris: Guillaumin.
----- (1853b). Voles de communication. In Dictionnaire de l'economie politique, ed. C. Coquelin, Vol. 2, pp. 846-58. Paris: Guillaumin.
----- (1860). La Liberte commerciale: son principe et ses consequences. Revue Europeenne, 11, 34780; 592-623; 834-58.
----- (1861a). La Liberte commerciale: son principe et ses consequences. Paris: Guillaumin.
----- (1861b). Reponse a M. Dunoyer a propos de son rapport sur l'ouvrage intitule. La liberte commerciale. Journal des economistes, 2nd series, 31, 111-17.
----- (1863). Reponse a M. Baudrillart au sujet de l'article 'L'economie est-elle une science ou une etude?'. Journal des economistes, 2nd series, 37, 474-82.
EKELUND, R. B., Jr and HEBERT, R. F. (1976). Dupuit and marginal utility: context of the discovery. History of Political Economy, 8, 266-73.
----- and THORNTON, M. (1991). Geometric analogies and market demand estimation: Dupuit and the French contribution. History of Political Economy, 23, 397-418.
FLEURY, E. L. (1867). La Vie et les travaux de M. Dupuit. Journal des Economistes, 3rd series, 7, 1-50, (page references are to Ekelund-Hebert translation).
GIDE, C. (1891). Principles of Political Economy, trans, from the 3rd French edition by E. P. Jacobsen. Boston: D. C. Health.
GROENEWEGEN, P. (1995). A Soaring Eagle: Alfred Marshall 1842-1924. Aldershot: Edward Elgar.
HAUSMAN, D. M. (1992). Essays on Philosophy and Economic Methodology. London: Cambridge University Press.
MARSHALL, A. (1920). Principles of Economics, 8th edn. London: Macmillan.
----- (1975). The Early Economic Writings of Alfred Marshall, 1867-1890, ed. J. K. Whitaker, 2 vols. New York: Free Press.
SAY, J. B. (1819). Cours a l'Athenee. In Jean-Baptiste Say: Cours d'economie politique et autres essais, ed. Philippe Steiner. Paris: Flammarion, 1996.
----- (1826). A Treatise on Political Economy, 5th edn., trans. C. R. Prinsep. Philadelphia: Grigg & Elliot, 1832.
----- (1828). Cours complet d'economie politique pratique, 2 vols. Paris: Rapilly.
SCHUMPETER, J. A. (1954). A History of Economic Analysis, ed. E. B. Schumpeter. New York: Oxford University Press.
SMITH, A. (1937). An Inquiry into the Nature and Causes of the Wealth of Nations, ed. E. Caanan. New York: Random House.
STIGLER, G. J. (1990). The place of Marshall's Principles in the development of economics. In J. K. Whitaker (ed.), Centenary Essays on Alfred Marshall. New York: Cambridge University Press.
WHITAKER, J. K. (ed.) (1990). Centenary Essays on Alfred Marshall. New York: Cambridge University Press.
