The Allocation Does the Work

Ask an investor what is in their portfolio and you will get a list of names. Ask what percentage each name represents and the answer is usually slower, vaguer, and occasionally surprising to the person giving it. That asymmetry is the single most reliable tell of an unconstructed portfolio: it is a collection of decisions rather than a design.

The distinction matters because a portfolio's behaviour — how much it falls in a bad year, how it responds to a rate shock, whether it is genuinely diversified or four bets on the same underlying driver wearing different tickers — is a function of weights, not of names. You can hold ten excellent businesses and, through sizing alone, own something with the risk profile of a single-sector fund.

Consider the same twelve holdings arranged three ways. Nothing about the research changes. Nothing about the quality of the businesses changes. Only the weights change.

The first construction is not wrong. It is a coherent choice, and for an investor with real informational depth in a few names it can be the correct one. But it is a fundamentally different investment from the third, and it should be evaluated, monitored and stress-tested as such. The failure mode is not concentration; it is concentration that happened by accident — weights that drifted there because a position ran, and nobody ever decided whether it should be allowed to.

The most common misunderstanding in portfolio construction is that diversification is a count. Twenty holdings must be safer than ten. They frequently are not.

What matters is not how many things you own but how much they move together. Twenty companies that all depend on the same input cost, the same credit cycle, the same regulator or the same currency are, in risk terms, close to a single position that has been elaborately disguised. Ten genuinely different exposures can carry materially less risk than thirty similar ones.

This is not theoretical. Prospera's own published Portfolio Intelligence methodology records exactly this outcome in practice: across the four model portfolios, an eleven-holding portfolio measured as less internally correlated than one holding fifteen. The count went down and the diversification went up, because diversification was being measured rather than assumed.

The hidden concentrations

The exposures that cause real damage are usually the ones that do not appear on any sector label:

• —Factor concentration — every holding is, in effect, the same style bet (small-cap momentum, say) expressed through different names.
• —Regulatory concentration — several distinct businesses whose economics all sit under a single regulator's decision. Prospera's own High Growth model discloses precisely this: three holdings, three different businesses, one regulator.
• —Customer or supplier concentration — companies that look unrelated but share a single dominant counterparty.
• —Liquidity concentration — several positions that are individually modest but collectively impossible to exit in a stressed week.

The uncomfortable truth about risk management is that the tools which work best are the ones that require no predictive skill at all. A forecast helps only if it is right. A constraint helps whether or not you were right — which, given how often forecasts are wrong, makes it the more valuable instrument.

1. 1

   A maximum position size

   Caps the damage a single catastrophic holding can do, without requiring you to identify in advance which holding that will be. This is the whole point: you do not need to know.

2. 2

   A minimum position size

   Less obvious and equally important. A position too small to matter still consumes monitoring attention and still carries the risk that it will be held sloppily precisely because it does not matter.

3. 3

   A sector or theme ceiling

   Ensures that one industry being structurally wrong cannot take the whole book. It will occasionally stop you from owning more of the thing that turned out to be best. That is the fee for the protection, and it is worth paying.

4. 4

   A solvency veto that applies at every risk level

   Volatility and expensiveness are risks you can be compensated for. The risk of a business not existing is not one of them. A portfolio can rationally accept a violent price and a demanding valuation; it should never accept the possibility that the underlying company does not survive.

Rebalancing is routinely described as a way to improve returns. It is better understood as the only mechanism in a portfolio that enforces 'sell high, buy low' without requiring an opinion about what is high and what is low.

When an asset rises past its target weight, the rule says to trim it. Not because it is overvalued — you do not know that — but because it now represents more risk than the design allowed. When an asset falls below target, the rule says to add. Not because it is cheap, but because it now represents less risk than the design intended. The valuation judgement is entirely absent, which is exactly why it works.

The cost is real and should be stated plainly: rebalancing systematically reduces exposure to whatever is trending, so in a long, strong, single-direction market it will underperform doing nothing. Every investor who abandons rebalancing does so at the end of exactly such a period, which is the worst possible moment to abandon it.

None of this is exciting. There is no rebalancing anecdote worth telling at dinner. But over a decade, the portfolio that had a written maximum position size, a measured correlation structure, a solvency veto and a rebalancing band will have delivered something much closer to what its owner intended than the portfolio that had a better list of names and no design at all.