The Cobweb Model: How Prices Spiral Toward or Away From Balance

The setup: why a time lag changes everything

Static supply and demand assume buyers and sellers react to the same price at the same instant, so the market settles where the curves cross. The cobweb model breaks that simultaneity by adding one realistic feature: a production lag. A farmer deciding how many acres of corn to plant, or how many sows to breed, commits now based on today's price — but the crop or the piglets only reach the market a full season or year later.

Formally, three assumptions drive the whole result:

• Supply responds to expected price with a lag. Quantity produced in period t depends on the price in period t−1: Qₛˢ = c + d · Pₜ₋₁.
• Demand responds to the current price, with no lag. Once the goods are on the market they must be sold at whatever price clears them: Qₜᵈ = a − b · Pₜ.
• Naive (adaptive) expectations. Producers assume the price they see today will still hold tomorrow. They never anticipate that everyone else is reacting the same way.

That last assumption is the engine of the whole model. Because every producer extrapolates today's price into the future, a high price makes all of them expand at once — guaranteeing a glut, and therefore a crash, one period later.

The core recursion and the stability condition

Setting quantity supplied equal to quantity demanded each period gives the price a simple first-order difference equation. Solving a − b·Pₜ = c + d·Pₜ₋₁ for the current price:

Pₜ = (a − c)/b − (d/b) · Pₜ₋₁

Everything hinges on the multiplier −d/b, where d is the slope of supply (how strongly output reacts to price) and b is the slope of demand. Each period the deviation of price from equilibrium is multiplied by (−d/b). The minus sign is why the price alternates above and below equilibrium — the source of the oscillation. The magnitude d/b decides the fate of the spiral:

• d/b < 1 (supply steeper than demand) → CONVERGES. Each swing is smaller than the last; the spiral winds inward to equilibrium.
• d/b > 1 (supply flatter than demand) → DIVERGES. Each swing is larger; the spiral winds outward, prices explode.
• d/b = 1 → PERPETUAL CYCLE. The price bounces between two values forever, tracing a closed rectangle.

An equivalent, more intuitive statement uses elasticities at equilibrium: the market is stable when supply is less elastic than demand. If quantity supplied reacts more sharply to price than quantity demanded does, a small price wobble triggers a large output response, which triggers an even larger price wobble — instability.

A worked example: watch the spiral converge

Take demand Qₜᵈ = 100 − 2·Pₜ (so b = 2) and supply Qₛˢ = 20 + 1·Pₜ₋₁ (so d = 1). Equilibrium is where 100 − 2P = 20 + P, giving P* = 26.67, Q* = 46.67. Since d/b = 1/2 < 1, we expect convergence. Start the market off-balance at P₀ = 40:

• Year 1 price 40 → farmers plant for supply Q = 20 + 40 = 60. That 60 must sell at demand price where 60 = 100 − 2P → P = 20.
• Year 2 price 20 → supply Q = 40 → market price 100 − 2P = 40 → P = 30.
• Year 3 price 30 → supply Q = 50 → P = 25.
• Year 4 price 25 → supply Q = 45 → P = 27.5.
• Year 5 → P = 26.25 → Year 6 → P = 26.875 … closing in on 26.67.

The price sequence 40, 20, 30, 25, 27.5, 26.25… oscillates but each overshoot is exactly half the previous one (the factor d/b = ½), homing in on equilibrium. Flip the numbers so supply is the flat curve — Qₛˢ = 20 + 3·P against Qₜᵈ = 100 − 2P — and d/b = 3/2 > 1: the same starting nudge would give 40, −20, 70, −65… ever-widening swings that fly apart. Same diagram, opposite destiny, decided purely by which curve is steeper.

Where you actually see cobwebs: hogs, corn, and careers

The model was born from real data. In the 1920s–30s, agricultural economists were puzzled by the regular hog cycle: U.S. hog prices swung on a roughly 3–4 year rhythm. Sows take about a year from breeding to market-ready pork, and farmers bred heavily when prices were high — flooding the market a year later. The German-language literature calls it the Schweinezyklus (pig cycle), documented since the 1920s. Corn showed a linked corn–hog cycle because feed and livestock prices interlock.

Cobweb dynamics recur wherever supply is lagged and biological or capital-bound:

• Coffee & cocoa: a coffee tree takes 3–5 years to bear fruit. Price booms (e.g., the 1975 Brazilian frost) trigger a planting wave that gluts the market half a decade later.
• Cattle: the U.S. "cattle cycle" of roughly 8–12 years reflects the multi-year lag in expanding a breeding herd.
• Labor markets for specialists. Richard Freeman's classic studies of engineers and physicists (1970s) and later law and PhD markets found cobweb-like enrollment gluts: a hot job market draws a wave of students who graduate 4–7 years later into a saturated field.
• Semiconductors & commercial real estate: multi-year fab or building construction lags produce boom-bust capacity cycles with the same signature overshoot.

The critique: would people really keep falling for it?

The cobweb model's Achilles' heel is its expectations assumption. Naive expectations require producers to be systematically, repeatedly fooled — to keep believing today's high price will last even after living through several boom-bust cycles. Critics from the rational-expectations school (Muth's 1961 paper on rational expectations was literally motivated by hog and cobweb markets) argued that anyone who noticed the cycle could profit by doing the opposite, which would damp the swings.

Several patches keep the model alive rather than killing it:

• Adaptive expectations: producers forecast a weighted average of past prices rather than just the last one, slowing but rarely eliminating the cycle.
• Storage and inventories: storable goods (grain) let arbitrageurs buy cheap and sell dear, flattening the cobweb — which is why the effect is strongest in perishable or biological-lag markets and weakest for storable, financialized commodities.
• Nonlinear cobwebs & chaos: with nonlinear curves, the model can generate genuinely chaotic, aperiodic prices — a bridge to modern complexity economics. Real cycles are noisier and less regular than the clean textbook spiral, but the tendency toward lagged over-reaction is robust and well-documented empirically.

A subtle point people get wrong

The most common misconception is that a divergent cobweb means prices literally shoot to infinity. They can't. In reality, physical and behavioral limits bite: output can't fall below zero, demand can't go negative, and after a couple of brutal cycles producers change their behavior — they hedge, sign forward contracts, or simply learn the pattern. So an unstable cobweb in practice shows up as large, persistent, but bounded fluctuations, not an explosion.

A second subtlety: convergence does not require supply to be steep and demand flat in absolute terms — it's the ratio of slopes at the equilibrium point that matters. A market with both curves flat can still be stable if supply is even steeper than demand. And note the direction of comparison flips people up: because supply's slope d sits in the numerator of d/b, the market is stabilized by inelastic (weakly responsive in quantity, steep on the standard price/quantity diagram) supply relative to demand — i.e., producers who don't over-react to price. The counterintuitive lesson is that a highly responsive, nimble supply side is exactly what makes a lagged market unstable. This is also why the cobweb is a cautionary tale for policy: an agricultural price support that props up prices can perversely lock in the over-production wave the model predicts.

Cobweb (naive expectations) vs. Rational-expectations pricing in a lagged-supply market

Feature	Cobweb Model	Rational Expectations
Expected next price	= last observed price (backward-looking)	= model-consistent forecast of actual next price
Price path	Oscillates; may converge, diverge, or cycle forever	Jumps to (or near) equilibrium immediately
Systematic errors	Yes — producers repeatedly fooled by the cycle	None on average; errors are pure random noise
Who moves markets	Slow-learning producers reacting to yesterday	Forward-looking agents using all information
Empirical fit	Good for slow, opaque, biological-lag markets	Better for liquid financial / storable markets