Cyclicals versus defensives

The last two modules built one idea: the economy moves in a cycle, and the cycle reaches different businesses with wildly different force. This module gives the two ends of that spectrum their proper names and reads them side by side, because the contrast is one of the cleanest and most useful in all of macro investing.

At one end sit the cyclicals — companies whose profits swing sharply with the economic cycle. Steel, cement, metals, autos, capital goods, real estate, airlines: when the economy runs hot they can mint extraordinary profits, and when it turns they can bleed. At the other end sit the defensives — companies whose profits barely notice the cycle, because their demand does not. Soap and biscuits, everyday medicines, electricity: people buy these in a boom and a bust alike, so the earnings stream is steady across the seasons.

The question this module answers is not "which is better" — that is the wrong question, and answering it confidently is the classic mistake. The question is: how does the very same cycle produce a peak profit that can be a trap in one and a barely-moving earnings stream that can be a comfort in the other — and how do you read each without being fooled by the number on the screen.

Two forces, both met earlier, decide where a company sits on the spectrum.

The first is the nature of the demand. A cyclical sells things people and businesses buy when they feel confident and flush — a new car, a new factory, a new flat, a flight, a tonne of steel for a project. These are postponable. The moment the future looks uncertain, they are the first purchases delayed, so cyclical volumes drop fast in a downturn and surge in a boom. A defensive sells things people buy regardless — food, soap, medicine, power. This demand is not postponable in the same way; you do not stop eating because growth slowed. So its volumes are steady almost by definition.

The second is operating leverage, the amplifier from the business-cycle module. Cyclicals also tend to carry heavy fixed costs — plants, furnaces, aircraft, debt — so a swing in volume becomes a bigger swing in profit. Defensives tend to have lighter fixed costs and costs that move with sales, so profit follows volume more gently. Put the two together and the spectrum is stark: cyclicals have swingy demand and an amplifier, so their earnings are violently cyclical; defensives have steady demand and little amplifier, so their earnings are calm.

Here is the mechanic that trips up more beginners than almost any other, and it lives entirely in the cyclical column.

A cyclical's peak earnings — the unusually high profit it makes at the top of its cycle — are, by their nature, temporary. But a common way to judge whether a stock is cheap, the price-to-earnings ratio (the share price divided by profit per share), uses recent earnings as its denominator. Feed it peak earnings and the ratio looks tiny — the stock appears wonderfully cheap — at the exact moment its earnings are most inflated and most likely to fall. When the cycle turns, profit drops, the denominator shrinks, and the "cheap" stock re-rates to an expensive-looking or loss-making one, often while the price is falling too. The low ratio was not a bargain; it was a warning misread as an invitation.

The defensive column has the opposite shape and its own, milder trap. Its earnings are close to normal in most years, so its P/E usually means roughly what it appears to. But because everyone values that steadiness, defensives can trade at high ratios for long stretches — and a steady earnings stream bought at too high a price is still a poor investment. The defensive trap is not a disguised peak; it is over-paying for calm.

The repair is not a trick; it is a change of denominator. Instead of the latest, cycle-distorted profit, ask what the company earns across a whole cycle — a normalised, mid-cycle figure that averages the peaks and troughs. Judge the price against that, and the peak's flattering ratio and the trough's frightening one both dissolve into something more honest. The ratio is only as good as the earnings you feed it — question the denominator before you trust the number.

Put one downturn in front of four composite companies and watch the same cycle land in four different places. illustrative The two on the right invert the naive "a downturn is bad for everyone, roughly equally" reading — their earnings barely move.

One economy, one downturn, four earnings paths — from halved to barely moved. Now sit with the inversion the whole part has been building toward: the steelmaker's violent swing is what makes its peak so alluring and so treacherous, while the pharma maker's flat line is dull in a boom and precious in a bust. The property that makes a cyclical exciting at the top is the same property that traps you there; the property that makes a defensive boring is the same property that protects you. Neither is "the safe one" or "the good one" in the abstract — each is a different bargain with the cycle, and the price you pay decides whether it was a good bargain.

The cyclical/defensive frame is powerful and, like every powerful frame, easy to over-trust.

A label is not a verdict on the investment. "Defensive" describes an earnings stream, not a guarantee of return. A defensive bought at a rich price can lose money; a cyclical bought cheap near a trough can do very well. Safety is business plus price, never the label alone.

It cannot tell you where in the cycle you are. Reading a cyclical well requires a view on the cycle, and that view is fogged in real time. The peak-earnings trap exists precisely because no one rings a bell at the top. You can judge whether behaviour looks late-cycle and euphoric, but you cannot pin the peak in advance.

It cannot promise defensives stay defensive. A "defensive" can be disrupted — a new competitor, a pricing shock, a regulatory change — and a "cyclical" can structurally improve. Assuming a label holds forever is its own trap; businesses move along the spectrum over time.

It cannot make a peak-P/E honest. The single most important defence in this module is to value a cyclical on normalised, mid-cycle earnings rather than the latest peak. The trailing ratio, fed a peak, will mislead you every cycle. Later modules on valuation across the cycle formalise this.

1. Buying the cheap-looking cyclical at the peak. The low trailing P/E on record earnings is the most reliable value trap in the market. The cheapness is the peak talking; the earnings are about to fall.

2. Extrapolating peak profit into the future. A cyclical's record quarter is a moment, not a run-rate. Building a forecast — or a valuation — on it is how tops are bought and dividends are chased just before they are cut.

3. Treating "defensive" as "safe at any price". Steady earnings reduce business risk, not valuation risk. Over-pay for calm and you can still lose; the price is always part of the safety.

4. Ignoring debt on a cyclical. Heavy borrowing on top of swingy earnings is what turns a cyclical downturn from a bad year into an existential one. The balance sheet decides who survives the trough.

5. Confusing a label with a phase call. Reading either type well needs a view on where the cycle is — which you cannot know precisely. The labels describe sensitivity; they do not tell you what the cycle will do next.