When I was at university, I was taught that a diversified portfolio reduces risk for a given level of desired return. As you know, the reason it was such a powerful and enduring result is correlation: if assets don’t move together, losses in one can be offset by gains in another.

Great so far, but there is a catch. The catch is that we rarely know those correlations in advance.

Let me give you an example. In the mid 2000s before the GFC, Banks packaged subprime mortgages into securities, rating agencies stamped their AAA-approval, and investors piled in believing that the risks were spread. Of course, borrowers came from different backgrounds, professions, ages and ethnicities. However, many had precarious jobs, had borrowed a lot, and were truly financially exposed. When the downturn arrived, defaults surged everywhere. What looked diversified was actually one big, correlated bet.

European real estate faces a similar risk today. Diversification is being pursued through sectors. Alternatives - data centres, life sciences, student housing, self-storage - are the hot ticket. At first glance, they look very different: new occupiers, new growth drivers, new demand stories.

But investment flows are what tie them together. As capital pours in, yields compress, pricing stretches, and the trade becomes less about income and more about momentum. Herd behaviour makes them move in sync — not because of fundamentals, but because the same investors are chasing them.

And that creates fragility (and a high correlation). A squeeze in debt markets, a downturn in private equity or venture funding, regulation (and many other things I can’t think of) could bite across multiple subsectors at once. None of these risks are identical, but the effect would be the same, a liquidity crunch.

That’s the paradox. The occupational fundamentals of a data centre and a student residence have little in common, but investor sentiment links them. If capital stops flowing, correlation shows up fast - and diversification stops working.

The lesson is not to avoid alternatives or cling to the old office-retail-industrial mix. It’s to remember that diversification only works when assets behave differently under stress, not just when they look different on a PowerPoint slide. Real protection comes from owning things that react differently when the tide turns, not from buying more of what’s fashionable. Knowing the “real correlation” is impossible. But if there is a history of trading over various cycles, then it is much easier to estimate.

But with new, fashionable things, diversification risks becoming what it was in subprime: not a hedge, but a trap.