The capex and commodity glut cycle

The last two modules read a commodity move across sectors at a single moment in time. This one adds the missing dimension: time. Commodity prices do not wander randomly. They move in long, sweeping cycles — years of shortage and high prices, then years of surplus and low prices — and the two halves are not independent events. The boom causes the bust. The high price is the very thing that summons the supply that later destroys it.

That is a deeply counter-intuitive idea, so hold it plainly: for a commodity, the most dangerous moment is often the one that feels safest. At the peak, prices are high, producers are posting record profits, everyone is confident, and — precisely because of all that — new mines, new smelters, new plants are being funded that will pour supply into the market two or three years later. The comfort at the top is what builds the glut. This module is about reading that machine, and about why a commodity peak is the moment to be most careful, not least.

A commodity cycle is the long swing in a raw-material price between shortage, when prices are high, and glut, when prices are low. What makes it a cycle rather than a random walk is a feedback loop that runs through investment.

Start at a shortage. Demand outruns supply, so the price rises. High prices make producing the commodity wildly profitable, so producers rush to build more — they commit capex (capital expenditure: money spent building new plant, mines or capacity) to add capacity, how much they can produce when running flat out. Here is the catch that drives the whole cycle: a mine or a smelter or a chemical plant cannot be built overnight. There is a long lead time — the years between deciding to build and the capacity actually producing. So the supply response arrives late, all at once, and usually after the shortage that justified it has already eased. When that wave of new capacity finally switches on, supply overshoots demand, and the market tips from shortage into glut: more supply than buyers want, which pushes the price down. Low prices then make production unprofitable, capex stops, weak producers close, capacity slowly shrinks — until demand catches up again and the next shortage begins.

The engine, in one sentence: high prices invite the supply that later crashes the price, and the delay between the two is what stretches the whole thing into a years-long cycle nobody can time.

The four turns of the loop are simple. What makes them dangerous is the lag in the middle.

Turn one — shortage. Demand tops supply; the price climbs; producer margins fatten. Existing plants run flat out and mint cash.

Turn two — the capex response. The high price and fat margins make new capacity look irresistible. Boards approve mines, smelters, plants. Crucially, everyone does this at once — every producer sees the same high price and reaches the same conclusion — so the industry commits far more new capacity than the shortage actually needs.

Turn three — the wave lands. Years later, after the long lead time, the new capacity comes on stream. But the shortage that justified it has usually already eased, and now a flood of supply arrives on top. The market tips into glut. Prices fall — often hard and fast, because commodity demand is slow to change while the new supply is a step-jump.

Turn four — the shakeout. Low prices make production unprofitable. New capex freezes. High-cost producers bleed and eventually shut. Capacity slowly leaves the market. This clears the glut — and quietly sets up the next shortage, because when demand recovers there is now too little supply. The loop starts again.

This is why the cycle is genuinely hard to escape. The decision that dooms the price — approving new capacity — is made when the price is high and the decision looks most sensible. The mechanism is the calm at the top is what breeds the coming instability: a long stretch of high, stable prices convinces everyone the good times are safe, which funds the capex that ends them. Stability is not the opposite of the crash; it is the cause of it.

Watch the cycle write itself onto one company's numbers. illustrative

A composite metal producer sits three years into a boom. The metal price is near a multi-year high. Its accounts look magnificent: revenue at a record, operating margin at its widest ever, debt paid down, and — because profits are so fat — a reported P/E in the single digits. On a naive screen it looks cheap and strong at once. It is the most seductive combination a cyclical ever offers.

Now read it through the cycle instead of the snapshot. The record margin is a peak margin, earned at a peak price. The low P/E is low because the E is inflated by that peak price — divide a fixed share price by a peak profit and the multiple looks small. And across the industry, boards flush with this cash are approving new mines and smelters, each with a multi-year lead time. The very health of the accounts is the fuel for the supply wave that will, in two or three years, arrive and push the price down. When it does, the peak margin normalises, the peak profit shrinks, and the "cheap" P/E re-rates upward not because the stock rose but because the earnings fell. The single-digit multiple was the most expensive moment, disguised as the safest.

Understanding the loop is powerful and comes with a strict limit: it explains the shape, never the timing.

You cannot know where in the cycle you are, in real time. This is the honest, uncomfortable centre of it. The peak only becomes obvious as a peak in hindsight; while you are living through it, a peak looks exactly like a strong, healthy market that might run for years more — and sometimes it does. Booms can extend far longer than seems reasonable when demand keeps surprising or supply is slow to arrive. Reading the cycle tells you the machine exists and roughly which quarter of the loop the evidence favours; it does not stamp a date on the turn. Any read that claims to call the top is a guess about where we are, worn as a certainty — you can know the cycle is real without knowing where on it you stand.

It cannot tell you how far or how fast. Some gluts are shallow and brief; others are deep and last years. The depth depends on how much capacity was over-built and how demand behaves on the way down — neither knowable in advance.

And it cannot, of course, tell you what to own. Recognising that a commodity is near a likely peak is not a signal to sell anything or a permission to short. It is a reason to read the accounts with the cycle in mind — to discount peak margins, distrust a peak-earnings P/E, and size any commodity exposure knowing the price that made it look good is the price most likely to reverse.

The glut cycle is a museum of expensive mistakes.

1. Buying peak earnings as if they were normal. Extrapolating a record margin into the future is the cardinal cyclical error. The margin is high because the price is high, and the price is high because supply has not yet caught up — a condition designed to end.

2. Reading a low P/E at the peak as cheap. The multiple is low because the earnings are inflated. On a cyclical, a single-digit P/E at a price peak is a warning, not a bargain.

3. Believing the boom is structural, not cyclical. Every peak comes wrapped in a story about why this commodity is different — permanent demand, a new era, supply that can never catch up. Usually it is the four most expensive words dressed in fresh clothes: the same cycle, narrated as a new age. Sometimes there is a real structural shift; far more often the price simply summoned the supply, on schedule.

4. Confusing confidence with safety. Producer optimism, analyst upgrades and easy financing all peak together, at the top. The mood is warmest exactly when the risk is greatest, because that mood is what funds the capacity that ends the boom.