Marginal utility: core, elaborations, revisions

Three core assertions define MU theory. First, consumers seek a generic ‘utility’, ‘pleasure’, ‘value’, or ‘benefit’ that motivates their preferences and purchasing choices. To maximise it, each consumer, second, applies the decision rule of selecting goods with equal MU per unit of cost. Jointly, the first two assertions predict (third) how price changes alter consumers’ behaviour. Each consumer attains an equilibrium that maximises the utility she obtains given her stable preferences and fixed budget. The first law of demand follows: when a commodity’s price rises, buyers buy less of it, and vice versa when its price falls. More formally, the sign of the substitution effect is usually negative.

More credible, nuanced versions of these core explanations have been formulated. Behavioural and neuro-economics preface the motivational assertion with psychophysical assumptions about (among other things) habits (Houthakker and Taylor, Reference Houthakker and Taylor1974), neurological evidence that simple choice is stochastic (Hare et al., Reference Hare, Camerer and Rangel2009), and how rats and pigeons respond to reinforcement and stimuli (Battalio et al., Reference Battalio, Dwyer and Kagel1987). This leads to a three-stage explanation of consumer demand. First, psychophysical mechanisms are described. Next, this description is translated into MU concepts. We read of ‘Nature’ assigning utilities to consumption bundles (Robson, Reference Robson2002), the activity of consuming as ‘shadow price’ (Stigler and Becker, Reference Stigler and Becker1977), ‘internal evaluation’ as a ‘cardinal utility function’, or sensory rewards and evolutionarily adaptive behaviour as the von Neumann–Morgenstern hedonic utility function (Robson, Reference Robson2002). The term ‘utility signal at consumption’ (Fehr and Rangel, Reference Fehr and Rangel2011: 15) assimilates neuronal signals to utility and so on. Third, the resulting utility functions are mathematically modelled to predict consumer behaviour.

Other elaborations redefine what consumers maximise. Specifically, the particular characteristics (e.g. colour) of goods are what provide utility (Lancaster, Reference Lancaster1966). Utility may include transaction utility from getting unexpectedly cheap or high-quality goods (Thaler, Reference Thaler1999), changes in wealth (Kahneman and Tversky, Reference Kahneman and Tversky1979), the pleasurable anticipation of risk (Allais, Reference Allais1953), and (negatively) information-search costs. Consumers attach more utility to keeping a good than to gaining it originally: the ‘endowment effect’ (Allais, Reference Allais1953; Kahneman and Tversky, Reference Kahneman and Tversky1979). In theory, consumers might reach a ‘bliss point’ where their wants are satiated (Gowdy and Mayumi, Reference Gowdy and Mayumi2001), but short of that, and usually, they behave as utility maximisers.

Further elaborations relax the second core assumption about how much knowledge and precision consumer decisions involve. Rather than consider every conceivable bundle of goods, consumers approximately maximise utility by choosing among similar goods relevant to their immediate purchasing aims (Lancaster, Reference Lancaster1966). They weight decisions by subjective rather than objective probabilities (Kahneman and Tversky, Reference Kahneman and Tversky1979), apply practical, ‘pseudo-maximising’ rules-of-thumb by, say, comparing available goods and prices with the usually expected ones, and use ‘mental accounting’ (Thaler, Reference Thaler1999). Consumers sometimes make inconsistent or misinformed choices (Bernheim and Rangel, Reference Bernheim and Rangel2009), but not to the extent of invalidating MU theory as a generalisation (Buchanan, Reference Buchanan1969).

Yet further revisions facilitate the mathematical modelling. Revealed preference theory requires only that utilities be ranked, not cardinally measurable (Houthakker, Reference Houthakker1950). Utilities can be weighed for risk and futurity (Friedman and Savage, Reference Friedman and Savage1948).

These elaborated ex ante forms of MU theory better predict observed consumer behaviour, for example, why consumers sometimes maintain their demand when prices rise, or respond asymmetrically to price reductions and increases. The grouping of goods explains advertising, bandwagon and snob effects, and how universally desired goods contrast with more instrumental consumer goods (Becker, Reference Becker1962; Erdem et al., Reference Erdem, Keane and Sun2008). Many such predictions concern consumer ‘irrationality’ in acting, under certain conditions, in ways that do not maximise their utility. Without endorsing them, McChesney (Reference McChesney2013) listed 20 such analyses and more have appeared since.

Deeper revisions deny that MU theory describes anything psychological at all (Buchanan, Reference Buchanan1969; Friedman and Savage, Reference Friedman and Savage1948; McChesney, Reference McChesney2013). Becker (Reference Becker1962) showed that if consumers chose mixtures of goods randomly, and their selections were uniformly distributed across individuals’ consumption possibilities, the mean (and mode) quantities of goods demanded would still change inversely with price. In these views, utility maximisation is not an ex ante motivation, the psychological consequence of a pre-defined utility function, but the ex post effect of consumer purchases upon satisfaction, well-being, or experienced utility ‘irrespective of the processes generating behaviour’ (Bernheim and Rangel, Reference Bernheim and Rangel2009: 53). Rather, market structures and their Darwinian dynamics are what maximise utility (Becker, Reference Becker1962; McChesney, Reference McChesney2013). On this view, MU theory does

Consequently, the first two defining assertions of MU theory are assumptions or axioms, not observable. The third is an observable behavioural prediction (Samuelson, Reference Samuelson1938; McChesney, Reference McChesney2013). One might object that consumers’ errors mean that their purchasing decisions do not necessarily reflect rationality or ‘true’ preferences (Fehr and Rangel, Reference Fehr and Rangel2011). However, one can also observe what mistakes consumers make, correct accordingly the preferences that might otherwise have been inferred from their behaviour, and so infer their true underlying preferences (Kőszegi and Rabin, Reference Kőszegi and Rabin2007).

Finally, MU theory has been reformulated as purely a logic of choice:

Without evidence that people do in fact behave ‘economically’, this claim is empirically empty. Further assumptions must be added to yield empirically testable predictions about consumers (Buchanan, Reference Buchanan1969). It also illustrates how MU theory uses terms, such as ‘efficient’, ‘preferable’, ‘rational’, and here ‘must’, ambiguously between a factual sense (equalising marginal input-to-cost ratios does maximise output) and normatively (economic agents ought to do so). These ambiguities build normative assumptions into MU theory (Robinson, Reference Robinson1962; Gowdy and Mayumi, Reference Gowdy and Mayumi2001).

Redundancy and replacement

The familiar criticisms of MU theory suggest what kind of replacement is warranted and why.

Ex ante MU theory attributes explain consumer behaviour from price changes given stable preferences and preferences by utility-seeking. Between evidentially attested psychophysical mechanisms and attested predictions, behavioural and neuro-economics tend to interpose superfluous, unattested assertions about utility functions. Furthermore, the anterior psychophysical explanations partly contradict the assumption that consumers equalise goods’ MU, even unconsciously. Humans’ neuronal encoding involves what MU theory defines as ‘irrelevant’ considerations (Fehr and Rangel, Reference Fehr and Rangel2011). The more mathematically complex the modelling of MU-based predictions becomes, the less credible a description of consumers’ decision rules it offers. ‘Fast and frugal’ probabilistic algorithms are simpler and more practicable. They necessitate neither compensating one outcome with another, nor outcome-commensurability, nor the accuracy-effort trade-offs which cost-of-information theories assume (Gigerenzer and Brighton, Reference Gigerenzer and Brighton2009). The act of deciding itself helps formulate and set preferences (Ariely et al., Reference Ariely, Loewenstein and Prelec2003). Changes in habit often alter them. Price changes influence affect as well as cognition, hence preferences besides behaviour (Peine et al., Reference Peine, Heitmann and Herrmann2009). Since the costs and benefits of further information-gathering are unknowable, consumers can’t know when their choice is optimal (Hodgson, Reference Hodgson1997). Behavioural economics and industrial safety research report systematic bias in human decision-making and how the framing of choices affects what individuals decide (Kahneman and Tversky, Reference Kahneman and Tversky1979). However, the bigger doubt is whether, biased or not, consumers use utility-maximising decision rules at all, even at a preconscious neurological level. These reasons suggest removing utility analysis from ex ante theories and explaining consumer behaviour directly from psychophysical antecedents.

The ‘as if’ and the logic-of-choice reformulations look stronger. However, asserting that the behavioural predictions, but not the motivation and decision-rule assumptions, are empirically testable tacitly reformulates the connections between them as logical, not causal, relationships. Yet true implications can logically follow from false premises without making the false premises true (Lemmon, Reference Lemmon1971). By themselves, neither logical deduction, mathematical modelling, measurement, nor criteria for theory selection are causal explanations (Buchanan, Reference Buchanan1969; Coase, Reference Coase1988). Adding the right supplementary axioms can make the core MU theory predict any economic behaviour, so that any behaviour can afterwards be described as what consumers preferred (Hodgson, Reference Hodgson2012), but that is predictively empty unless the predictions were first specified independently. Unrestricted retroduction from observed behaviour often accommodates many, even opposite, explanations. The key, arguably sole, prediction of MU theory is that the substitution effect has a negative sign (Lancaster, Reference Lancaster1966), but many theories that do not assume utility maximisation also predict it. Only imagination limits the range of ‘as if’ explanations. Ariely et al. (Reference Ariely, Loewenstein and Prelec2003) suggest five. Then, selecting which explanation is valid requires evidence about what causal mechanisms were operating. Evidence of consumers’ motivation is often available from their actual purchases (even if misjudged or if they would have preferred to be in different circumstances) and, with methodological precautions, by questioning. Many neuro-marketing studies describe neural responses to marketing mix or to marketing stimuli (cf. Hubert and Kenning, Reference Hubert and Kenning2008), rather than the neurological character of motivation. Nevertheless, the neural mechanisms that perform choices, decisions, and hedonic responses are becoming observable, although research is only beginning.

Whether by market-level analyses of consumer spending (e.g. Erdem et al., Reference Erdem, Keane and Sun2008), directly eliciting consumers’ intentions (e.g. Ramirez and Goldsmith, Reference Ramirez and Goldsmith2009), or experimentally, empirical studies on balance confirm the first law of demand and negative substitution effects, although these effects are not necessarily the largest determinants of consumer behaviour (Houthakker and Taylor, Reference Houthakker and Taylor1974). However, ex ante versions of MU theory give erroneous causal explanations of what links consumer motivation, decision-making, and behaviour. The ex post versions give none. If, as Becker (Reference Becker1962) says, we can dispense with some of the empirically dubious assumptions of MU theory and still derive (or explain) downward-sloping consumer demand curves and other consumer behaviour, is it then possible to replace all of them? A few economists have agreed, but with exceptions (e.g. Earl, Reference Earl2023; Earl et al., Reference Earl, Markey-Towler and Coutts2022), still fewer have contributed to such an explanation.

By studying the facts of consumer behaviour, what choices consumers face, and how they choose and then behave when prices change, this paper aims to help fill that gap.

Motivation

A precondition for a consumer’s purchase is that she wants the good at all. The evolutionary and genetic origin, and neural embodiment, of human motivation is well-established. Hunger, shelter-seeking, child-rearing, and other instinctual proclivities arise from yet anterior biological processes (e.g. for food, nutrients, gut-driven satiety signals, adiposity-related hormones, and other mechanisms). They stimulate evolutionarily formed behaviours: self-preservation, altruism towards kin, and so on. There appear to be heritable influences on individual pre-dispositions towards some, but not all, decision-making processes (Simonson and Sela, Reference Simonson and Sela2011; Wichary and Smolen, Reference Wichary and Smolen2016).

Consumers’ psychological reward systems then develop and supplement these foundational instinctual motives (Hubert and Kenning, Reference Hubert and Kenning2008) into a complex, hierarchical motivational structure. Among others, some early neo-classical economists (Gossen, Menger, Marshall, Engel [Chai and Moneta, Reference Chai and Moneta2010]) proposed that non-instinctual desires arise from instinctual ones, whether as means to obtain instinctually wanted things, by association, or interactively with the development of affect (emotion, mood, attitude) (Holbrook and Hirschman, Reference Holbrook and Hirschman1982), in an experiential, path-dependent (Hansen, Reference Hansen2005), and therefore individualised sequence. The resulting structure of motivations is partly instinctual, partly learned, and mutable (Mort and Rose, Reference Mort and Rose2004; Van Osselaer and Janiszewski, Reference Van Osselaer and Janiszewski2012). Taxonomies of the motives within it differ, but many studies report the desires for experiences besides tangible things, for sustaining the consumer’s self-esteem, self-concept, cultural identity (Weingarten and Goodman, Reference Weingarten and Goodman2021), and social status (many studies since Veblen, Reference Veblen2005). Much consumption is symbolically motivated (Douglas et al., Reference Douglas, Wilk and Isherwood2021). Consumers’ motivations correspondingly range from preconscious to deliberative (Chartrand et al., Reference Chartrand, Huber, Shiv and Tanner2008). Because the more immediately instinctual motives tend to be stronger (Lindenberg, Reference Lindenberg1996), consumers tend to buy satisfiers for them first, then for the most proximate unsatisfied wants, and so on (Ironmonger, Reference Ironmonger1972).

To arouse consumer demand, these often latent motives have to be activated. Activation can be excitatory or inhibitory, chronic or temporary (Van Osselaer and Janiszewski, Reference Van Osselaer and Janiszewski2012). Homeostatic regulation activates the instinctual motives (e.g. eating). A person’s consumption goals, affective states (e.g. boredom [Maccarrone-Eaglen and Schofield, Reference Maccarrone-Eaglen and Schofield2017; Yoon and Meyvis, Reference Yoon and Meyvis2024]), and attempts to regulate mood or the emotions (Kemp and Kopp, Reference Kemp and Kopp2011) may also activate motivation (Wichary and Smolen, Reference Wichary and Smolen2016). So do external cues (Hansen, Reference Hansen2005; Chartrand et al., Reference Chartrand, Huber, Shiv and Tanner2008), including sensory cues such as goods associated with previous purchases (Pfeiffer et al., Reference Pfeiffer, Deval and Kardes2022), social norms and pressures, and the opportunity to obtain a desired good (Yoon and Meyvis, Reference Yoon and Meyvis2024) or substitute. A perception of differences between one’s actual and desired self-image (Kim and Rucker, Reference Kim and Rucker2012), or of a loss of control, can cue compensatory consumption.

Information

As explained above, learning, including information acquisition, extends, modifies, and individualises consumers’ motivation structures. In selecting goods, consumers predict the physical, experiential, symbolic, identity, and status effects of consuming them.

Consumers predict these effects partly from memory (previous learning) (Yoon and Meyvis, Reference Yoon and Meyvis2024). Because situational cues may have activating effects, consumers frequently judge goods by appearance, texture, smell, or colour, especially for routine, relatively unimportant purchases, making shopping and trial consumption common methods of information search (Holbrook and Hirschman, Reference Holbrook and Hirschman1982). Even deliberative consumers often use ‘naïve theories’ about goods’ quality, social-symbolic, or other characteristics (Deval et al., Reference Deval, Mantel, Kardes and Posavac2013). For instance, they take price, brand, or appearance as indicators of quality (Völckner and Hofmann, Reference Völckner and Hofmann2007). For many routine purchases, memory, cues, and naïve theories give consumers enough information not to search further. The size of the choice sets they face influences how thorough a search strategy consumers use and their propensity to seek the best possible or just a satisfactory choice (Levav et al., Reference Levav, Reinholtz and Lin2012). A minority of consumers seek information from consumer organisations. Many consult social media. This gives shopping a collective character although, at times, it is based on misinformation and fear, for example, panic buying, herding, and bandwagon effects during the COVID-19 pandemic (Naeem and Ozuem, Reference Naeem and Ozuem2021). Consumers also acquire ‘incidental information’ without searching, above all unsolicited marketing persuasion (Janssen and Fennis, Reference Janssen and Fennis2017). Extensive research reports how different marketing methods influence consumers’ beliefs, attitudes, motivations, and buying behaviour. (Kotler and Armstrong, Reference Kotler and Armstrong2010, is a widely used overview.) Producers, retailers, and advertisers sometimes even bamboozle consumers with, say, bait-and-switch and ‘halo-effect’ selling (Talukdar et al., Reference Talukdar, Gauri and Grewal2010).

Learning more general institutional beliefs and norms also exercises ‘reconstitutive downwards causation’ upon individuals’ habits (Earl, Reference Earl2023) and motivation structures, socialising besides individualising them. Motives based on self-identity, self-presentation, status, social identity, and justifying one’s purchasing decisions (cf. Ajzen and Fishbein, Reference Ajzen and Fishbein2000) necessarily incorporate knowledge of social conventions, symbols, and norms. Consumers also engage in contagious imitation (Ironmonger, Reference Ironmonger1972), fads, and fashions (Veblen, Reference Veblen2005; Wakefield and Inman, Reference Wakefield and Inman2003). Sociologies of consumption (e.g. Douglas et al., Reference Douglas, Wilk and Isherwood2021; Ritzer, Reference Ritzer2009) explain this socialisation in terms of different cultures, micro-cultures, and sub-cultures of consumption, consumer habitus, or practice theory (Arsel and Bean, Reference Arsel and Bean2013; Shove and Warde, Reference Shove, Warde, Dunlap, Buttel, Disckens and Giswijt2002). Their ‘lived culture and social resources’, and marketing representations, frame consumers’ horizons of feeling, thought, action, sense-making, sensibilities, norms, consumption communities, identities, rituals, and decision rules in acquiring and using consumer goods (Arnould and Thompson, Reference Arnould and Thompson2005). ‘Marketplace institutions such as magazines, websites, and transmedia brands’ perpetuate the resulting ‘taste regime’ (Arsel and Bean, Reference Arsel and Bean2013: 889). These cultural differences influence the extent of consumer cosmopolitanism (Han and Won, Reference Han and Won2018), how consumers perform mental accounting (Banerjee et al., Reference Banerjee, Chatterjee, Mishra and Mishra2019), and how such ideologies as religion and political conservativism (Lisjak and Ordabayeva, Reference Lisjak and Ordabayeva2023), and marketing (Kotler and Armstrong, Reference Kotler and Armstrong2010), shape consumers’ desires. Through default choices, social influence, warnings, reminders, and simplifications, policy-makers and marketers also try to ‘nudge’ consumers’ decisions.

Evaluative decision

As modified by learning, consumers’ consumption motives are also their criteria for evaluating the goods they know about. A good’s attributes and price both influence what quantity a consumer demands, but the attributes are causally prior. Unless she evaluates a good as desirable to begin with, its price hardly matters to her. Often, her choice is simultaneously both under-determined (when several products would all suffice) and over-determined (when a good would satisfy several motives at once).

Consumer decision processes range from preconscious to highly deliberative. Many, even most, are low cognition (Chartrand et al., Reference Chartrand, Huber, Shiv and Tanner2008; Hubert and Kenning, Reference Hubert and Kenning2008), including choices made from memory or which are:

1. (1) compulsive (Maccarrone-Eaglen and Schofield, Reference Maccarrone-Eaglen and Schofield2017);

2. (2) impulsive (Chen and Wang, Reference Chen and Wang2016);

3. (3) panicked (Arafat et al., Reference Arafat, Kar, Marthoenis, Sharma, Apu and Kabir2020);

4. (4) habitual (Hodgson, Reference Hodgson1997);

5. (5) hurried (Nordfalt, Reference Nordfalt2009).

Heuristics are more deliberative. Gigerenzer and Brighton (Reference Gigerenzer and Brighton2009) describe eight ‘fast and frugal’ heuristics applicable to consumer choice, other studies at least 19 more. Reported heuristics include:

1. (1) affect (‘choose what feels best’) (Slovic et al., Reference Slovic, Finucane, Peters and MacGregor2007);

2. (2) recognition (e.g. of a favourite brand) (Hubert and Kenning, Reference Hubert and Kenning2008);

3. (3) satisficing (Levav et al., Reference Levav, Reinholtz and Lin2012);

4. (4) compromise (‘choose the middle’) (Fehr and Rangel, Reference Fehr and Rangel2011);

5. (5) external rules (Hodgson, Reference Hodgson1997; Payne et al., Reference Payne, Bettman and Johnson1993), including suppliers’ suggestions;

6. (6) consciously ranking one’s wants hierarchically and, opportunities allowing, satisfying them in that order (Earl et al., Reference Earl, Markey-Towler and Coutts2022; Ironmonger, Reference Ironmonger1972);

7. (7) lexicographic choice: deciding which goods are best in terms of the most desired attribute, then tie-breaking by considering the next most desired attribute, and so on (Wichary and Smolen, Reference Wichary and Smolen2016). Trivial differences may finally break the tie between otherwise similar goods (Levav et al., Reference Levav, Reinholtz and Lin2012);

8. (8) scoring goods by a few simple criteria (Degeratu et al., Reference Degeratu, Rangaswamy and Wu2000) or a weighted additive rule (WADD), the process closest to explicit utility maximisation calculations but not identical;

9. (9) two-stage decision-making (Kahneman and Tversky, Reference Kahneman and Tversky1979): choosing first among, say, product attributes and then among the goods having the favoured attribute;

10. (10) allocating money to mental budgets and then to particular goods within each budget (Skwara, Reference Skwara2023).

Consumers sometimes learn, combine, or devise decision rules ‘on the fly’ whilst making their decisions. By repetition, habitual rules-of-thumb form (Levav et al., Reference Levav, Reinholtz and Lin2012). Consumers apply different heuristics, or combinations of them, to different decision situations (Hodgson, Reference Hodgson1997). Payne et al. (Reference Payne, Bettman and Johnson1993) give an overview of how individuals trade off accuracy (e.g. not neglecting important information) against cognitive effort (e.g. difficulty) in information processing.

Often, consumers try to avoid trade-offs between goods (Gu et al., Reference Gu, Botti and Faro2013), for example, by using non-compensatory decision rules (high levels of another characteristic do not compensate for lacking a necessary characteristic) or using more general goals (e.g. environmentalism) to arbitrate between lower-order goals (Levav et al., Reference Levav, Reinholtz and Lin2012). Consumers often under-weigh or ignore opportunity costs (Frederick et al., Reference Frederick, Novemsky, Wang, Dhar and Nowlis2009), although less so when facing tight financial constraints (Dias et al., Reference Dias, Sharma and Fitzsimons2022) or difficult choices (e.g. between undesirable alternatives or conflicting motives) (Gu et al., Reference Gu, Botti and Faro2013). In practice, consumers seldom consider the opportunity cost of every good they might conceivably forgo but only a comparison set of perhaps 15, often below 5, similar goods (Huang et al., Reference Huang, Lurie and Mitra2009). When they do consider opportunity costs, a hierarchical motivation structure will tend to favour essential over discretionary purchases (Weingarten and Goodman, Reference Weingarten and Goodman2021). Engel’s ‘law’ states that households with lower incomes spend a greater proportion of income on goods, which directly satisfy their physical needs, as subsequent studies on balance confirm.

From individual to market demand

Two main mechanisms determine whether consumer demand becomes effective (Becker, Reference Becker1962). As a price rises, the inability to pay excludes some consumers from the market. Then, differences in willingness to pay make those remaining buy less.

Ability to pay

Markets are just one institution through which consumers obtain goods. Institutional preconditions for consumer demand in a market are that potential consumers:

1. (1) Confront property rights. Others own the desired goods but they do not;

2. (2) Can readily obtain these goods only by purchase, not by (say) making them, using up stores, borrowing, sharing, barter, gift, or theft. Research on business models, intellectual property, and certain aspects of marketing describes how firms create that precondition;

3. (3) Own enough money (income, wealth, savings, credit) to buy the desired good.

At the extreme, a consumer’s exclusion price is the one at which she would have to spend all her money on the focal good except for prior non-discretionary items (e.g. goods such as food whose consumption is hard to reduce below a physical minimum). Many spend all their income on consumption (Martin and Paul Hill, Reference Martin and Paul Hill2012). Even for essentials, low income may drive consumers towards ‘extreme value shopping’ (Carpenter, Reference Carpenter2009). The opportunity cost of a purchase, and of an increase in its price, is proportionately greater, the less money they own, and the greater their price sensitivity (Houthakker and Taylor, Reference Houthakker and Taylor1974). Their usually unequal ownership of money (‘budget constraint’, ‘opportunity set‘) excludes some consumers from a market before others when prices rise, not through ‘willingness’ but the ability to pay. Being unable to afford a good does not necessarily stop a consumer from wanting it and in that sense being willing to buy, but it does prevent her from revealing that willingness in the market.

This exclusionary amount is the highest point where the individual’s demand curve can meet the vertical axis (Figure 2).

Figure 2. From individual to market demand.

First law of demand

Market-level demand patterns emerge as multiple consumers interact with potential suppliers’ product designs, price-setting, and marketing.

Extensive market research describes how consumer ‘segments’ differ in buying behaviour. Within segments, individual consumers vary in motivation, learning, decision processes, and money ownership. The first law of demand emerges from these cumulatively uneven effects. Represented graphically, each individual’s demand curve (dotted ID1, ID2, … etc. in Figure 2) meets the vertical axis at her exclusion price, or her reservation price when lower, and the horizontal axis at her satiation quantity of consumption (see below). Because individual consumers’ behaviour and available money vary, their individual demand curves differ in position. They may cross, be stepped in either direction, multi-segmented, or kinked (e.g. when social norms stipulate what goods to buy for, say, social events such as weddings). Horizontally adding individual curves gives a whole-market demand curve (Figure 2).

The aggregate curve slopes downwards because individual demand curves do, and consumers are excluded cumulatively as price rises. The non-excluded consumers differ in motivation, learning, evaluation processes, and money owned. As prices rise, those with successively higher reservation prices become unwilling to buy and then stop buying, whilst others continue. All that is necessary for the first law of demand to operate is that for some consumers, price changes trigger at least one of the above mechanisms. Assuming that consumers who stop buying because of one price increase seldom resume following another, each mechanism, and a fortiori, the combination, has cumulative effects over successive price rises. The exclusionary institutional mechanisms are causally prior twice over. When they apply, they pre-empt the excluded consumers’ preferences or ‘willingness’ altogether, and consumers’ preferences don’t affect whether, when, and how they work. Together, these motivational and institutional factors imply a down-sloping, possibly stepped, gapped, or non-smooth demand curve.

To understand price elasticity then requires explaining how individual differences in ability and willingness to pay are distributed and why. If a consumer’s spending on other goods remains unchanged whilst a price change makes her buy less of the focal good, that decrease is the income effect upon her. If the money she has remains unchanged, she can buy less of other goods, or save less, as the price of the focal good rises. The proportion of her spending on the focal good that she moves to other goods is the substitution effect. The income effect becomes defined either in real terms or the money equivalent at post-change prices and the substitution effect as a proportion of spending. Neither definition involves indifference curves or utility, unlike Hicks’, Marshall’s, or Slutsky’s definitions.

Non-price changes also affect the quantity demanded. They include products (e.g. mobile phones [Yoo et al., Reference Yoo, Choi, Hwang and Mun2021]) acquiring additional uses. Goods change in what they symbolise. Their want-satisfying characteristics change (e.g. the sensory properties of food). So does how goods are packaged, promoted, and sold. Marketing studies describe how firms use these changes to stimulate demand and for non-price competition. Consumers’ characteristics also change, including life-cycle stage, household size and composition, education, other demographic factors, and their opinions. Purchasing activity itself sometimes changes consumers’ preferences, right up to the instant of purchase, as point-of-sale marketing exploits (Degeratu et al., Reference Degeratu, Rangaswamy and Wu2000). The above explanations need not assume that consumers’ wants, tastes, or preferences are stable or consistent whilst their demand for a good is forming.

Consumption

Purchase enables consumption, whose usual effects on consumer demand are satiation and learning.

Consumers satiate their desires not just by buying goods but by combining them with their own time, effort, and other resources to produce a concrete benefit (e.g. a meal). To an extent depending on the nature of the good, consumption is a household production activity involving effort, cognitive, and embodied knowledge: a ‘consumption technology’ (Lancaster, Reference Lancaster1966). Motivation to buy also depends on the consumer believing herself able to consume the good, most obviously with goods (e.g. motorcycles) that require skill or strength to use. In any event, consumption takes time (Stigler and Becker, Reference Stigler and Becker1977), which imposes an upper limit even on the hedonic consumption of free goods. Since consumers’ time is finite, so is their capacity to use goods, hence the amount they can actively consume. (Ownership, unlike consumption, has no such limit.) At this amount, the individual’s demand curve meets the horizontal axis (Figure 2).

Satiation deactivates the original consumption motive (Chartrand et al., Reference Chartrand, Huber, Shiv and Tanner2008; Pfeiffer et al., Reference Pfeiffer, Deval and Kardes2022; Van Osselaer and Janiszewski, Reference Van Osselaer and Janiszewski2012), especially for familiar ‘established’ goods (Ironmonger, Reference Ironmonger1972) and for the most salient characteristics of the consumed good. Certain goods (e.g. haircuts) have a physical limit on how much a person can consume. The satiation of most instinctual needs, some hedonic experiences (e.g. listening to music [Ariely et al., Reference Ariely, Loewenstein and Prelec2003]), and some other goals (e.g. consistency of choice [Silverman and Barasch, Reference Silverman and Barasch2023]) are temporary but repeatable. Satiation is also observed when prices are triflingly low, when the consumer pays a flat rate irrespective of how much she consumes, for distress goods of which she consumes the minimum (and would prefer, none), and derived demand, when she only buys in order to achieve some further purpose. For derived demand and distress goods, these patterns appear when non-market institutions allocate the goods free of charge. (Patients do not seek repeat surgery just because it is free.) At the market level, income elasticities that are low and decline as income rises suggest satiation in ‘basic need’ goods (food, housing, energy) (Baxter and Moosa, Reference Baxter and Moosa1996). As individual consumers cumulatively become satiated, so does market demand, even if slowly. Over 15 years can elapse between the first and the last consumers buying a new high-technology consumer good for the first time (Ironmonger, Reference Ironmonger1972).

Experience reinforces (Levav et al., Reference Levav, Reinholtz and Lin2012) or alters consumer’s expectations about consuming a particular good (Holbrook and Hirschman, Reference Holbrook and Hirschman1982). Repeated consumption of a good reduces predicted enjoyment, hence the consumer’s likelihood of purchasing it again (Pasdiora et al., Reference Pasdiora, Brei and Nicolao2020). By learning these predictive associations (Van Osselaer and Janiszewski, Reference Van Osselaer and Janiszewski2012), consumers acquire consumption and purchasing habits and skills and inertia in product choices.

Satiation and learning feed back into consumers’ memory and motivation to consume. Because of these feedback loops, the often temporary nature of satiation, and because the other factors mentioned above are continually changing, the formation of consumer demand is dynamic. Any equilibrium in a consumer’s demand and consumption is likely to be transient.