Portfolio Management: Diversification Is Not Enough
Klarman's pushback on treating broad diversification as a substitute for genuine analysis of each individual holding.
Klarman is skeptical of relying on wide diversification as the primary tool for managing risk, arguing it can become a substitute for the harder work of genuinely understanding each individual holding well enough to have real conviction in it. A portfolio of forty superficially-analyzed positions isn't obviously safer than a portfolio of fifteen deeply-understood ones — it just spreads the same underlying analytical shortcuts across more names.
His actual practice at Baupost favored meaningful concentration in high-conviction ideas, sized according to how much margin of safety each one offered, combined with genuine diversification across the types of risk different positions carried — not simply diversification by count of holdings, which he treats as a weaker, more superficial form of risk management.
| Diversification by count | Klarman's approach | |
|---|---|---|
| Goal | Spread capital across many positions to reduce single-stock risk | Concentrate in high-conviction ideas, sized by margin of safety, while varying the type of risk across positions |
| What determines position size | Often an equal or formulaic weighting across many names | How much margin of safety a specific idea offers, and how well it's understood |
| Risk being managed | Idiosyncratic risk in any single stock, diluted by count | The risk of being wrong about a position you didn't understand deeply enough in the first place |
The distinction Klarman draws isn't "concentration good, diversification bad" — it's between diversification that genuinely spreads different kinds of risk (a special-situation spin-off, a distressed bond, a conventionally-cheap stock, each with different return drivers and different things that could go wrong) versus diversification that just multiplies the number of tickers without actually varying the underlying risk exposure much at all.
A portfolio holding twenty different retail stocks is diversified by ticker count, but if a broad consumer-spending downturn hits, most of those twenty positions are likely to move together, since they share the same underlying risk factor. A portfolio holding five positions — a spin-off, a distressed bond, two conventionally undervalued stocks in different industries, and a risk-arbitrage position on a pending merger — is more genuinely diversified in Klarman's sense, even with far fewer names, because each position responds to a different set of risks.
- Diversification by number of holdings can become a substitute for genuine understanding of each position — Klarman treats that substitution as a weaker, more superficial form of risk management.
- Position sizing, in his practice, follows margin of safety and depth of understanding, not an equal or formulaic weighting across many names.
- Genuine diversification means varying the type of risk different positions carry, not simply increasing the count of tickers held — a portfolio can hold many similar stocks and still be poorly diversified in any meaningful sense.