Constraints are conceptualized as a border or margin. The location of the margin for any individual corresponds to his or her endowment, broadly conceived to include opportunities. This endowment is determined by many things including physical laws (which constrain how forms of energy and matter may be transformed), accidents of nature (which determine the presence of natural resources), and the outcomes of past decisions made both by others and by the individual himself or herself.
A value that holds true given particular constraints is a marginal value. A change that would be effected as or by a specific loosening or tightening of those constraints is a marginal change.
Neoclassical economics usually assumes that marginal changes are infinitesimals or limits. (Though this assumption makes the analysis less robust, it increases tractability.) One is therefore often told that “marginal” is synonymous with “very small”, though in more general analysis this may not be operationally true (and would not in any case be literally true).
Marginalism assumes, for any given agent, economically rationality and an ordering of possible states-of-the-world, such that, for any given set of constraints, there is an attainable state which is best in the eyes of that agent. Descriptive marginalism asserts that choice amongst the specific means by which various anticipated specific states-of-the-world (outcomes) might be effected is governed only by the distinctions amongst those specific outcomes; prescriptive marginalism asserts that such choice ought to be so governed.
On such assumptions, each increase would be put to the specific, feasible, previously unrealized use of greatest priority, and each decrease would result in abandonment of the use of lowest priority amongst the uses to which the good or service had been put.
In 20th century mainstream economics, the term “utility” has come to be formally defined as a quantification capturing preferences by assigning greater quantities to states, goods, services, or applications that are of higher priority. But marginalism and the concept of marginal utility predate the establishment of this convention within economics. The more general conception of utility is that of use or usefulness, and this conception is at the heart of marginalism; the term “marginal utility” arose from translation of the German “Grenznutzen”, which literally means border use, referring directly to the marginal use, and the more general formulations of marginal utility do not treat quantification as an essential feature. On the other hand, none of the early marginalists insisted that utility were not quantified, some indeed treated quantification as an essential feature, and those who did not still used an assumption of quantification for expository purposes. In this context, it is not surprising to find many presentations that fail to recognize a more general approach.
Under the special case in which usefulness can be quantified, the change in utility of moving from state to state is
Mainstream neoclassical economics will typically assume that
The “law” of diminishing marginal utility (also known as a “Gossen's First Law”) is that, ceteris paribus, as additional amounts of a good or service are added to available resources, their marginal utilities are decreasing. This “law” is sometimes treated as a tautology, sometimes as something proven by introspection, or sometimes as a mere instrumental assumption, adopted only for its perceived predictive efficacy. Actually, it is not quite any of these things, though it may have aspects of each. The “law” does not hold under all circumstances, so it is neither a tautology nor otherwise proveable; but it has a basis in prior observation.
An individual will typically be able to partially order the potential uses of a good or service. If there is scarcity, then a rational agent will satisfy wants of highest possible priority, so that no want is avoidably sacrificed to satisfy a want of lower priority. In the absence of complementarity across the uses, this will imply that the priority of use of any additional amount will be lower than the priority of the established uses, as in this famous example:
However, if there is a complementarity across uses, then an amount added can bring things past a desired tipping point, or an amount subtracted cause them to fall short. In such cases, the marginal utility of a good or service might actually be increasing.
Without the presumption that utility is quantified, the diminishing of utility should not be taken to be itself an arithmetic subtraction. It is the movement from use of higher to lower priority, and may be no more than a purely ordinal change.
When quantification of utility is assumed, diminishing marginal utility corresponds to a utility function whose slope is continually or continuously decreasing. In the latter case, if the function is also smooth, then the “law” may be expressed
When goods and services are discrete, the least favorable rate at which an agent would trade A for B will usually be different from that at which she would trade B for A:
If, for example, Lisa will not trade a goat for anything less than two sheep, then her
When indifference curves (which are essentially graphs of instantaneous rates of substitution) and the convexity of those curves are not taken as given, the “law” of diminishing marginal utility is invoked to explain diminishing marginal rates of substitution — a willingness to accept fewer units of good or service in substitution for as one's holdings of grow relative to those of . If an individual has a stock or flow of a good or service whose marginal utility is less than would be that of some other good or service for which he or she could trade, then it is in his or her interest to effect that trade. Of course, as one thing is traded-away and another is acquired, the respective marginal gains or losses from further trades are now changed. On the assumption that the marginal utility of one is diminishing, and the other is not increasing, all else being equal, an individual will demand an increasing ratio of that which is acquired to that which is sacrificed. If any trader can better his or her own marginal position by offering a trade more favorable to complementary traders, then he or she will do so.
A thorough-going marginalism sees marginal cost as increasing under the “law” of diminishing marginal utility, because applying resources to one application reduces their availability to other applications. Neoclassical economics tends to disregard this argument, but to see marginal costs as increasing in consequence of diminishing returns.
Marginalism and neoclassical economics generally explain price formation broadly through the interaction of curves or schedules of supply and demand.
Demand curves are explained by marginalism in terms of marginal rates of substitution.
At any given price, a prospective buyer has some marginal rate of substitution of money for the good or service in question. Given the “law” of diminishing marginal utility, or otherwise given convex indifference curves, the rates are such that the willingness to forgo money for the good or service decreases as the buyer would have ever more of the good or service and ever less money. Hence, any given buyer has a demand schedule that generally decreases in response to price (at least until quantity demanded reaches zero). The aggregate quantity demanded by all buyers is, at any given price, just the sum of the quantities demanded by individual buyers, so it too decreases as price increases.
Both neoclassical economics and thorough-going marginalism could be said to explain supply curves in terms of marginal cost; however, there are marked differences in conceptions of that cost.
Marginalists in the tradition of Marshall and neoclassical economists tend to represent the supply curve for any producer as a curve of marginal pecuniary costs objectively determined by physical processes, with an upward slope determined by diminishing returns.
A more thorough-going marginalism represents the supply curve as a complementary demand curve — where the demand is for money and the purchase is made with a good or service. The shape of that curve is then determined by marginal rates of substitution of money for that good or service.
By confining itself to limiting cases in which sellers or buyers are both “price takers” — so that demand functions ignore supply functions or vice versa — Marshallian marginalists and neoclassical economists produced tractable models of “pure” or “perfect” competition and of various forms of “imperfect” competition, which models are usually captured by relatively simple graphs. Other marginalists have sought to present more realistic explanations, but this work has been relatively uninfluential on the mainstream of economic thought.
In any case buyers are modelled as pursuing generally lower quantities, and sellers offering generally higher quantities, as price is increased, with each being willing to trade until the marginal value of what they would trade-away exceeds that of the thing for which they would trade.
That is not to say that the price of any good or service is simply a function of the marginal utility that it has for any one individual nor for some ostensibly typical individual. Rather, individuals are willing to trade based upon the respective marginal utilities of the goods that they have or desire (with these marginal utilities being distinct for each potential trader), and prices thus develop constrained by these marginal utilities.
The “law” is not about geology or cosmology, so does not tell us such things as why diamonds are naturally less abundant on the earth than is water, but helps us to understand how relative abundance affects the value imputed to a given diamond and the price of diamonds in a market.
Many critics of marginalism would reply that the reason that diamonds are more expensive than water is not because of their relative natural abundance but because of their cost of production. The reason water is available abundantly and diamonds in relatively smaller quantities is because one is inexpensive to produce and one very expensive. Critics claim that thus the reason water is cheaper than diamonds is simply because it costs less to produce. If diamonds could be produced cheaply from carbon, as modern technology may make possible in the short term, then the price of diamonds will fall, even though the demand for their use has not altered. Therefore, as these critics would claim, it is the cost of production which determines price, not the marginal utility.
Marginalists simply respond that if this were true then, rather than our seeing some goods and services not produced because their costs exceeded their prices, consumers would make a practice of seeking expensive wares without regard to their use. (As proto-marginalist Richard Whately put it, “It is not that pearls fetch a high price because men have dived for them; but on the contrary, men dive for them because they fetch a high price.”) Marginalists explain that costs of production may be what limit supply, but that these costs of production are themselves sacrificed marginal uses, and will not be borne when they are expected to exceed the marginal use of what is produced. In other words, the marginalist certainly does not explain price as a simple function of the marginal utility of a single good for one person or for some “average” person, but nonetheless insists that it results from the trade-offs that each participant would be willing to make for the various goods and services at stake, with those trade-offs being determined by marginal uses. The critics who believe that costs of production determine price, by assuming a demand that will bear the cost, have begged the essential question that the marginalists purport to answer.
A great variety of economists concluded that there was some sort of inter-relationship between utility and rarity that effected economic decisions, and in turn informed the determination of prices.
The first published statement of any sort of theory of marginal utility was by Daniel Bernoulli, in “Specimen theoriae novae de mensura sortis”. This paper appeared in 1738, but a draft had been written in 1731 or in 1732. In 1728, Gabriel Cramer produced fundamentally the same theory in a private letter. Each had sought to resolve the St. Petersburg paradox, and had concluded that the marginal desirability of money decreased as it was accumulated, more specifically such that the desirability of a sum were the natural logarithm (Bernoulli) or square root (Cramer) thereof. However, the more general implications of this hypothesis were not explicated, and the work fell into obscurity.
In “A Lecture on the Notion of Value”, delivered in 1833 and included in Lectures on Population, Value, Poor Laws and Rent (1837), William Forster Lloyd explicitly offered a general marginal utility theory, but did not offer its derivation nor elaborate its implications. The importance of his statement seems to have been lost on everyone (including Lloyd) until the early 20th century, by which time others had independently developed and popularized the same insight.
In An Outline of the Science of Political Economy (1836), Nassau William Senior asserted that marginal utilities were the ultimate determinant of demand, yet apparently did not pursue implications, though some interpret his work as indeed doing just that.
In “De la mesure de l’utilité des travaux publics” (1844), Jules Dupuit applied a conception of marginal utility to the problem of determining bridge tolls.
In 1854, Hermann Heinrich Gossen published Die Entwicklung der Gesetze des menschlichen Verkehrs und der daraus fließenden Regeln für menschliches Handeln, which presented a marginal utility theory and to a very large extent worked-out its implications for the behavior of a market economy. However, Gossen's work was not well received in the Germany of his time, most copies were destroyed unsold, and he was virtually forgotten until rediscovered after the so-called Marginal Revolution.
William Stanley Jevons first proposed the theory in “A General Mathematical Theory of Political Economy” (PDF), a little-noticed paper delivered in 1862 and published in 1863. He later presented the theory in The Theory of Political Economy (1871), which was fairly widely read but not much appreciated. Jevons' conception of utility was that in the hedonic tradition of Jeremy Bentham and of John Stuart Mill, and Jevons explained demand but not supply by reference to marginal utility.
Carl Menger presented the theory in Grundsätze der Volkswirtschaftslehre (translated as Principles of Economics) in 1871. Menger's presentation is peculiarly notable on two points. First, he took special pains to explain why individuals should be expected to rank possible uses and then to use marginal utility to decide amongst trade-offs. (For this reason, Menger and his followers are sometimes called “the Psychological School”, though they are more frequently known as “the Austrian School” or as “the Vienna School”.) Second, while his illustrative examples present utility as quantified, his essential assumptions do not. Menger's work found a significant and appreciative audience.
Marie-Esprit-Léon Walras introduced the theory in Éléments d'économie politique pure, the first part of which was published in 1874. Walras's work found relatively few readers.
(An American, John Bates Clark, is sometimes also mentioned in this context. But, while Clark independently arrived at a marginal utility theory, he did little to advance it until it was clear that the followers of Jevons, Menger, and Walras were revolutionizing economics. Nonetheless, his contributions thereafter were profound.)
Although the Marginal Revolution flowed from the work of Jevons, Menger, and Walras, their work might have failed to enter the mainstream were it not for a second generation of economists. In England, the second generation were exemplified by Philip Henry Wicksteed, by William Smart, and by Alfred Marshall; in Austria by Eugen von Böhm-Bawerk and by Friedrich von Wieser; in Switzerland by Vilfredo Pareto; and in America by Herbert Joseph Davenport and by Frank A. Fetter.
There were significant, distinguishing features amongst the approaches of Jevons, Menger, and Walras, but the second generation did not maintain distinctions along national or linguistic lines. The work of von Wieser was heavily influenced by that of Walras. Wicksteed was heavily influenced by Menger. Fetter referred to himself and Davenport as part of “the American Psychological School”, named in imitation of the Austrian “Psychological School”. (And Clark's work from this period onward similarly shows heavy influence by Menger.) William Smart began as a conveyor of Austrian School theory to English-language readers, though he fell increasingly under the influence of Marshall.
Böhm-Bawerk was perhaps the most able expositor of Menger's conception. He was further noted for producing a theory of interest and of profit in equilibrium based upon the interaction of diminishing marginal utility with diminishing marginal productivity of time and with time preference. (This theory was adopted in full and then further developed by Knut Wicksell and, with modifications including formal disregard for time-preference, by Wicksell's American rival Irving Fisher.)
Marshall was the second-generation marginalist whose work on marginal utility came most to inform the mainstream of neoclassical economics, especially by way of his Principles of Economics, the first volume of which was published in 1890. Marshall constructed the demand curve with the aid of assumptions that utility was quantified, and that the marginal utility of money was constant (or nearly so). Like Jevons, Marshall did not see an explanation for supply in the theory of marginal utility, so he synthesized an explanation of demand thus explained with supply explained in a more classical manner, determined by costs which were taken to be objectively determined. (Marshall later actively mischaracterized the criticism that these costs were themselves ultimately determined by marginal utilities.)
The doctrines of marginalism and the Marginal Revolution are often interpreted as somehow a response to Marxist economics. In fact, the first volume of Das Kapital was not published until 1867, after the works of Jevons, Menger, and Walras were written or well under way; and Marx was still a relatively obscure figure when these works were completed. (On the other hand, Hayek or Bartley has suggested that Marx may have come across the works of one or more of these figures, and that his inability to formulate a viable critique may account for his failure to complete any further volumes of Kapital.)
Nonetheless, it is not unreasonable to suggest that part of what contributed to the success of the generation who followed the preceptors of the Revolution was their ability to formulate straight-forward responses to Marxist economic theory. The most famous of these was that of Böhm-Bawerk, “Zum Abschluss des Marxschen Systems” (1896), but the first was Wicksteed's “The Marxian Theory of Value. Das Kapital: a criticism” (1884, followed by “The Jevonian criticism of Marx: a rejoinder” in 1885). The most famous early Marxist responses were Rudolf Hilferding's Böhm-Bawerks Marx-Kritik (1904) and Политической экономии рантье (The Economic Theory of the Leisure Class)(1914) by Никола́й Ива́нович Буха́рин (Nikolai Bukharin).
In his 1881 work Mathematical Psychics, Francis Ysidro Edgeworth presented the indifference curve, deriving its properties from marginalist theory which assumed utility to be a differentiable function of quantified goods and services. But it came to be seen that indifference curves could be considered as somehow given, without bothering with notions of utility.
In 1915, Евгений Евгениевич Слуцкий (Eugen Slutsky) derived a theory of consumer choice solely from properties of indifference curves. Because of the World War, the Bolshevik Revolution, and his own subsequent loss of interest, Slutsky's work drew almost no notice, but similar work in 1934 by John Richard Hicks and R. G. D. Allen derived much the same results and found a significant audience. (Allen subsequently drew attention to Slutksy's earlier accomplishment.)
Although some of the third generation of Austrian School economists had by 1911 rejected the quantification of utility while continuing to think in terms of marginal utility, most economists presumed that utility must be a sort of quantity. Indifference curve analysis seemed to represent a way of dispensing with presumptions of quantification, albeït that a seemingly arbitrary assumption (admitted by Hicks to be a “rabbit out of a hat”) about decreasing marginal rates of substitution would then have to be introduced to have convexity of indifference curves.
For those who accepted that superseded marginal utility analysis had been superseded by indifference curve analysis, the former became at best somewhat analogous to the Bohr model of the atom — perhaps pedagogically useful, but “old fashioned” and ultimately incorrect.
When Cramer and Bernoulli introduced the notion of diminishing marginal utility, it had been to address a paradox of gambling, rather than the paradox of value. The marginalists of the revolution, however, had been formally concerned with problems in which there was neither risk nor uncertainty. So too with the indifference curve analysis of Slutsky, Hicks, and Allen.
The expected utility hypothesis of Bernoulli et alii was revived by various 20th century thinkers, perhaps most notably Ramsey (1926), v. Neumann and Morgenstern (1944), and Savage (1954). Although this hypothesis remains controversial, it brings not merely utility but a quantified conception thereof back into the mainstream of economic thought, and would dispatch the Ockhamistic argument. (It should perhaps be noted that, in expected utility analysis, the “law” of diminishing marginal utility corresponds to what is called “risk aversion”.)
Meanwhile, the Austrian School continues to develop its ordinalist notions of marginal utility analysis, formally demonstrating that from them proceed the decreasing marginal rates of substitution of indifference curves.
Marginalism has been criticised for being extremely abstract, as “unobservable, unmeasurable and untestable”. Marginal utility is subjective, as the value of an additional unit of consumption is based on the individual's circumstances. However, margins (constraints) are often observable, as are patterns of choice; hence the general form of marginalism is in theory observable and testable. The special case of quantification of utility is more problematic, but the expected utility hypothesis represents a testable version of the theory with quantification. (Nonetheless, though confirmation of the expected utility hypothesis might have confirm quantification, the specific measure would not thus be found, as data that were fit by any proposed measure would be equally well fit by any affine transformation of that proposed measure.)
However, observed patterns of choice in test situations often seem not to correspond to an ordering, and the expected utility hypothesis has been falsified as description. (See the article on behavior economics, and perhaps especially that on the Ellsberg paradox or that on the Allais problem.) Many behavioral economists argue that people often follow simple rules of thumb instead of engaging in a mental process of maximizing some function. The reply from some marginalist and neoclassical economists is that these rules of thumb have been shaped by experience so that they give very nearly the same result as maximizing and that, moreover, use of rules of thumb is itself an act of optimization insofar as the decision-making process itself entails direct costs.
The theory is attacked for downplaying the rôle of cost of production in price determination in favor of a focus on individual's tastes and preferences. In its most extreme Austrian School version, marginalism denies that a purely objective, cost-based component exists at all. Rather, the Austrian School argues that costs of production pervasively involve individual preferences for labor vs. leisure and saving vs. consumption.
Karl Marx died before marginalism became the interpretation of economic value accepted by mainstream economics. His theory was based on the labor theory of value, which distinguishes between exchange value and use value. In his Capital he rejected the explanation of long-term market values by supply and demand:
In his early response to marginalism, Nikolai Bukharin argued that "the subjective evaluation from which price is to be derived really starts from this price", concluding:
Similarly a later Marxist critic, Ernest Mandel, argued that marginalism was "divorced from reality", ignored the rôle of production, and that:
Maurice Dobb argued that prices derived through marginalism depend on the distribution of income. The ability of consumers to express their preferences is dependent on their spending power. As the theory asserts that prices arise in the act of exchange, Dobb argues that it cannot explain how the distribution of income affects prices and consequently cannot explain prices.
Dobb also criticized the motives behind marginal utility theory. Jevons wrote, for example, "so far as is consistent with the inequality of wealth in every community, all commodities are distributed by exchange so as to produce the maximum social benefit." (See Fundamental theorems of welfare economics.) Dobb contended that this statement indicated that marginalism is intended to insulate market economics from criticism by making prices the natural result of the given income distribution.
Some economists strongly influenced by the Marxian tradition such as Oskar Lange, Włodzimierz Brus, and Michal Kalecki have attempted to integrate the insights of classical political economy, marginalism, and neoclassical economics. They believed that Marx lacked a sophisticated theory of prices, and neoclassical economics lacked a theory of the social frameworks of economic activity. Some other Marxists have also argued that on one level there is no conflict between marginalism and Marxism: one could employ a marginalist theory of supply and demand within the context of a “big picture” understanding of the Marxist notion that capitalists exploit labor.