Diversification for Unknown Unknowns
Diversify because the worst risks are the ones you can't name.
Diversification for Unknown Unknowns reframes diversification as humility, not laziness. Coffin distinguishes known unknowns — risks you can list — from unknown unknowns: forces that aren't even on your radar yet (consumer-taste shifts, technological surprises, regulatory whiplash). Even perfect company research can't immunise you against these.
The framework holds that diversification's primary job is not optimising expected return but absorbing the surprises that no amount of due diligence reveals. It also opens the door to asymmetric upside — Apple's iPhone moment was an unknown unknown that quietly built into one of the largest products in history.
It's a counterweight to the active community's concentration ethos: even if you can pick winners, the world can override you in ways you didn't model.
- There are known unknowns and unknown unknowns; only the second require true diversification.
- Even excellent research cannot price unknown unknowns.
- Diversification protects from idiosyncratic blow-ups and exposes you to surprise upside.
- Concentration earns its keep only when you can defend why this risk is worth missing other opportunities.
- Position sizing should reflect what you don't know, not just what you do.
- Catalogue your known unknownsFor each holding, list the risks you can name — competitive, regulatory, financial. This sets the floor of what your research already covers.
- Acknowledge unknown unknowns explicitlyWrite a line in every thesis admitting which categories of risk you can't enumerate (consumer taste shifts, geopolitical regime changes, new technologies). This forces humility into sizing.Pro tipUse prior cycles' surprises (smartphone, ZIRP, COVID) to remind yourself how invisible they were ex-ante.
- Set a diversification floorDefine a minimum number of names, sectors, or geographies. The floor enforces diversification before any specific thesis.WarningDon't let a great-looking thesis push you below the floor.
- Cap idiosyncratic exposureSet per-position and per-theme caps that limit damage if a single company is blindsided. Apply the same caps to themes (AI, weed, crypto) as to single names.
- Re-test after major surprisesWhen something genuinely unexpected happens, audit which holdings were exposed and whether the framework absorbed it as designed. Adjust caps accordingly.
When the iPhone launched it was mocked and compared unfavourably to cheaper Nokia handsets. Few investors modelled it becoming Apple's biggest product, yet diversified holders captured the windfall.
Coffin notes that even with perfect company research, an unforeseen shift in consumer sentiment can revolutionise how a company performs. Concentration leaves no buffer.
Coffin sits inside an active-management firm and engages directly with the concentration-vs-diversification debate. His position came from watching how often external shocks — consumer shifts, tech revolutions, geopolitical events — overturned otherwise solid theses, and how often unknown upside (the iPhone) emerged from products that were initially mocked.