Demand and Supply Analysis

Depth

Level

Demand and supply analysis is the study of how the prices and quantities of goods are jointly determined by the competing decisions of buyers, who want to purchase more when prices fall, and sellers, who want to sell more when prices rise. It explains market outcomes as the equilibrium that reconciles these two opposing forces.

Origins and the Marginalist Foundation

The apparatus we now teach as "supply and demand" was not born whole. Adam Smith (Wealth of Nations, 1776) distinguished "natural" from "market" price and gestured at gravitational adjustment, but had no formal schedules. The decisive break was the marginal revolution of the 1870s—William Stanley Jevons (1871), Carl Menger (1871), and Léon Walras (1874)—which relocated value from embodied labor or cost of production (the Ricardian–classical view) to subjective marginal utility at the margin of consumption.

Alfred Marshall's Principles of Economics (1890) then supplied the synthesis still taught today: the crossing scissor blades. Marshall's famous dictum—that asking whether value is governed by utility or cost of production is like asking "whether it is the upper or the under blade of a pair of scissors that cuts a piece of paper"—was a deliberate mediation between the marginalists and the classical cost-of-production tradition. Both blades matter; demand reflects marginal utility, supply reflects marginal cost, and price emerges from their intersection.

A conceptual caution Marshall himself stressed: the two-dimensional price–quantity diagram is a partial equilibrium device. It isolates one market, holding constant income, other prices, tastes, and expectations (ceteris paribus). Its rigorous general-equilibrium counterpart—where all markets clear simultaneously—belongs to Walras, later given existence proofs by Arrow, Debreu, and McKenzie (1954).

The Consumer's Side: Deriving Demand

From preferences to choice

The demand curve is not a primitive. It is derived from constrained optimization. A consumer with a preference relation—assumed complete, transitive, continuous, and (locally) non-satiated—maximizes utility $u(x)$ subject to the budget constraint $p\cdot x \le m$ . The first-order conditions equate the ratio of marginal utilities to the price ratio:

\frac{\partial u/\partial x_i}{\partial u/\partial x_j} = \frac{p_i}{p_j}

Solving yields the Marshallian (uncompensated) demand $x_i(p, m)$ . Its dual, from expenditure minimization at a fixed utility level, is the Hicksian (compensated) demand $h_i(p, u)$ .

The Slutsky decomposition

The central analytical result—due to Eugen Slutsky (1915) and rediscovered by John Hicks and R.G.D. Allen (1934)—decomposes the response of demand to a price change into a substitution effect (always weakly negative for the own price, by the concavity of the expenditure function) and an income effect (ambiguous in sign):

\frac{\partial x_i}{\partial p_j} = \underbrace{\frac{\partial h_i}{\partial p_j}}_{\text{substitution}} - \underbrace{x_j\frac{\partial x_i}{\partial m}}_{\text{income}}

This is where the "law of demand" earns its edge cases. The law—that demand slopes downward—is a theorem only for the substitution effect. A Giffen good arises when a strong negative income effect on a heavily budgeted inferior good overwhelms substitution, producing an upward-sloping demand. Long treated as a textbook curiosity, Robert Jensen and Nolan Miller (2008) provided credible field evidence among rice and wheat staples in Hunan and Gansu, China—closing a debate open since Marshall attributed the idea to Robert Giffen.

Aggregation and its discontents

Market demand is the horizontal sum of individual demands—but this masks a deep problem. The Sonnenschein–Mantel–Debreu theorem (1972–74) shows that aggregate excess demand functions inherit almost none of the structure of individual rationality: continuity, Walras's Law, and homogeneity are all they must satisfy. Downward-sloping market demand is therefore not guaranteed by rational individuals, and uniqueness/stability of equilibrium cannot be assumed in general. This is among the discipline's most under-advertised results.

Behavioral economics (Kahneman and Tversky; Richard Thaler) further complicates the derivation: reference dependence, framing, and preference reversals violate the integrability conditions (Slutsky symmetry) that make demand the shadow of a well-behaved utility function.

The Firm's Side: Deriving Supply

Cost as the dual of technology

The supply curve is likewise derived. A price-taking firm with production function $q = f(z)$ minimizes cost to obtain $C(q, w)$ , then maximizes profit $pq - C(q,w)$ . The first-order condition is the canonical marginalist rule:

p = MC(q)

with the second-order condition requiring $MC$ to be rising. The competitive firm's short-run supply curve is thus its marginal cost curve above average variable cost (below which it shuts down, since it cannot cover avoidable costs). In the long run, supply is $MC$ above minimum average total cost, and free entry drives economic profit to zero.

The upward slope and its exceptions

Supply slopes upward because of diminishing marginal returns in the short run and because expanding output draws in progressively higher-cost producers in the long run. But this too has edge cases:

Constant-cost industries yield a flat long-run supply curve.
Decreasing-cost (increasing-returns) industries—driven by scale economies, network effects, or learning-by-doing—can generate downward-sloping long-run supply, undermining the neat scissors. Piero Sraffa's 1926 critique in the Economic Journal argued precisely that Marshall's partial-equilibrium supply curve is coherent only under constant returns; increasing returns spill over into other industries (violating ceteris paribus), while decreasing returns internal to a firm imply market power. This "cost controversy" of the 1920s–30s helped launch imperfect-competition theory (Edward Chamberlin and Joan Robinson, both 1933).

The perfect-competition idealization

The competitive supply curve presupposes price-taking—an assumption that is self-referentially awkward. If every agent takes price as given, who sets it? Walras invoked a fictional auctioneer conducting tâtonnement; the honest answer is that perfect competition is a limiting idealization, approximated (per Aumann, 1964) only in economies with a continuum of negligible agents. Under monopoly or oligopoly there is no supply curve at all: quantity is chosen where marginal revenue equals marginal cost, and the same price can correspond to different quantities depending on demand elasticity.

Equilibrium: Where the Blades Cross

Market equilibrium is the price $p^*$ at which quantity demanded equals quantity supplied, so no agent has an incentive to deviate. Around it:

Comparative statics trace how $p^*$ and $q^*$ shift when an exogenous variable (income, input prices, technology, a tax) moves. Paul Samuelson's Foundations of Economic Analysis (1947) formalized this: the correspondence principle links the stability of equilibrium to the sign of comparative-static effects.
Stability itself is contested. Marshallian adjustment (quantity responds to price gaps) and Walrasian adjustment (price responds to quantity gaps) can disagree about whether a given equilibrium is stable—a live issue for backward-bending or perverse-sloping curves. Scarf's (1960) examples of globally unstable tâtonnement are decisive: convergence is not guaranteed.

Elasticity: the quantitative bridge

Marshall's price elasticity of demand, $\varepsilon = (\partial q/\partial p)(p/q)$ , converts the qualitative slopes into empirical magnitudes and governs incidence questions. Tax incidence is the workhorse application: the burden of a per-unit tax falls disproportionately on the less elastic side of the market, regardless of statutory assignment—a result at the heart of debates over payroll taxes, tariffs, and cigarette excise.

Identification: The Econometric Problem

Perhaps the deepest practical issue: observed price–quantity pairs are equilibria, points where both curves happen to cross. A scatter of such points identifies neither curve. This identification problem was posed by Philip and Sewall Wright (the appendix to The Tariff on Animal and Vegetable Oils, 1928, and the origin of instrumental variables), formalized by the Cowles Commission (Haavelmo 1943; Koopmans), and remains central. To recover the demand curve one needs a supply shifter that moves supply while leaving demand fixed (e.g., weather shocks to harvests); to recover supply, a demand shifter. This is the theoretical justification for modern IV and natural-experiment designs (Angrist, Imbens, and the "credibility revolution").

Open Questions and Frontiers

Micro-foundations of aggregate demand: given SMD, when may we treat a market or economy as if it had a representative rational consumer? (Only under restrictive Gorman-form preferences.)
Behavioral demand: how to do welfare analysis when demand is not the trace of stable preferences (Bernheim and Rangel, 2009).
Increasing returns and platform markets: two-sided markets and network industries routinely violate the classical supply construction, reviving Sraffa's concerns in digital dress.
Non-tâtonnement dynamics and disequilibrium: how markets actually reach equilibrium, if they do, remains theoretically unsettled—an agenda from Clower and Leijonhufvud to agent-based computational economics.

The scissors, in short, remain the discipline's indispensable first cut—but the expert reads them knowing each blade is a derived, conditional object, and that their intersection is an idealization the theory itself teaches us to distrust.