Diversification: the only free lunch

Two friends each run a small business. One sells umbrellas. The other sells sunscreen. In a wet summer the umbrella seller does well and the sunscreen seller does badly. In a dry summer it flips. Each business alone has a wild year-to-year ride. But own half of each, and your combined income is steady — because when one is down, the other is up.

You just stumbled onto the one thing in finance that looks almost too good to be true. Combine things that don’t move together and you cut your risk without cutting what you expect to earn. People call it the only free lunch in investing. Let’s see exactly why it works.

A reminder: risk is the spread

Back in item 3 we said risk is the spread of outcomes — how wide the good and bad years swing around the average. A single stock has a wide spread. Some years it soars, some years it sinks. The average might be a respectable 8% a year, but you live through the swings, not the average.

In item 6 we put many stocks into a basket — an index. Now we look at what that basket does to the spread. This is the heart of diversification: not “own more stuff” but “own stuff that doesn’t all move the same way.”

The magic word is correlation

Correlation is just plain togetherness. Two things are highly correlated when they tend to rise and fall together. Two umbrella shops in the same town move together — a wet summer lifts both, a dry one sinks both. Owning two of them barely helps. Their bad years line up.

Umbrellas and sunscreen move oppositely. When one zigs, the other zags. And many things move independently — a software firm and a fish farm have little to do with each other, so their good and bad years fall in no particular pattern.

Here is the rule that runs everything that follows. Things that move together don’t reduce risk. Things that move independently do. Things that move oppositely reduce it most. When you blend assets, their swings partly cancel. The peaks of one fill the troughs of another. The blend is calmer than either piece.

And — this is the free part — the average doesn’t drop. Blend two things that each average 8%, and the blend still averages 8%. You gave up nothing in expected return. You only gave up some of the wobble.

A worked example: zig and zag

Two assets. Each averages +8% a year. Both are bumpy. Here are three years for each.

• Asset A: +28%, then −12%, then +8%. Average = (28 − 12 + 8) ÷ 3 = +8%.
• Asset B: −12%, then +28%, then +8%. Average = (−12 + 28 + 8) ÷ 3 = +8%.

Notice B is A’s mirror in the first two years — when A soared, B sank, and the other way round. They move oppositely.

Now hold half of each — put £1 into A and £1 into B. Your blend each year is the simple average of the two:

• Year 1: (28 + −12) ÷ 2 = +8%
• Year 2: (−12 + 28) ÷ 2 = +8%
• Year 3: (8 + 8) ÷ 2 = +8%

Look at what happened. Each asset alone bounced between −12% and +28% — a 40-point swing. The blend sat at +8% every single year. Same average, almost no swing. The zig of one exactly filled the zag of the other.

Real assets are never perfect mirrors, so real blends still wobble a bit. But the direction is always the same: the less alike your holdings move, the smoother the blend, and the average stays put. That is the free lunch, in three lines of arithmetic.

The trap: fake diversification

Here’s where people fool themselves. Someone owns ten stocks and feels safe. But all ten are oil companies. That isn’t ten bets — it’s one bet, made ten times. Oil firms move together. When the oil price drops, all ten fall at once. Their bad years line up perfectly. The spread of the basket is barely narrower than the spread of one.

Diversification isn’t about the number of things you hold. It’s about whether they move together. Ten oil stocks are one bet. One oil stock, one bank, one supermarket, one drugmaker, one homebuilder — now the bad years don’t line up, and the blend genuinely steadies. You’ll feel this directly in the lab in a moment: drag two jagged lines from moving-together to moving-apart, and watch the blended line go flat.

The whole, and its limit

Diversification is the closest thing investing has to something for nothing — less risk, same expected return, purely from combining things that don’t march in step. It’s why an index of hundreds of companies is so much calmer than any one of them.

But it has a hard edge, and humility lives there. You can spread away the risk that’s specific to one company or one industry — the dry summer, the oil-price drop, the single firm’s bad luck. You cannot spread away the risk that hits everything at once. In a real panic, fear takes over and people sell whatever they can. Umbrellas, sunscreen, oil, banks — all fall together. The things that normally zig and zag suddenly all zag. Their togetherness jumps toward one, and the free lunch closes for the day.

That shared, can’t-escape risk is the thing no basket protects you from — and the next item names its biggest single source: the one number, set in one room, that tugs on every asset at once.