The discipline of not knowing

Every January the industry publishes its outlooks, and they arrive with a reassuring weight to them: a target for the index, a house view on rates, six themes to watch. By March nobody mentions them again, and this is not because the people who wrote them are foolish, but because the documents were never really forecasts at all. They were reassurance, printed.

I spent years producing versions of them, and the whole exercise felt genuinely rigorous at the time. It had numbers in it.

There is an old distinction, Frank Knight's, between risk and uncertainty. Risk is measurable: you don't know which card comes next, but you know the deck. Uncertainty is not knowing the deck at all, or even whether the thing in your hands is a deck. Most of finance is built for risk, whilst innovation is uncertainty wearing risk's clothes, because uncertainty doesn't fit in a spreadsheet and risk does.

Consider what we actually know about how the future arrives in equity markets. Bessembinder's work, which I have written about before, shows that most listed companies, over their whole lives, fail to beat treasury bills, and that the wealth is created by a sliver. The future turns up through a handful of names, and we do not know in advance which ones. Not roughly. Not directionally. At all.

And yet the not-knowing has a shape. Take the wave in front of us. Artificial intelligence is uncertain in the way that matters, in that no one can tell you which companies will end up owning it, and yet parts of it are strangely regular. The cost of running these models falls along an experience curve most macro forecasting would envy, and adoption traces the same S-curve it traced for the smartphone, and for electricity before that. The economics of the wave are knowable; which company captures it is not. Investors usually try to invert this, growing precise about the companies and vague about the wave, because companies have ticker symbols and waves do not. Coherent not-knowing means holding both at once: conviction about the water, agnosticism about the swimmers.

This is an uncomfortable place for a human being to stand, so mostly we don't. We convert the not-knowing into narrative instead. We say that artificial intelligence will be transformative, which is a story, and then we say therefore this company at this multiple, which is a leap dressed up as a conclusion. The story is doing emotional work, not analytical work; a forecast is a comfort object, and it gives the discomfort somewhere to sit.

The interesting question is not whether anyone can know the future, because nobody really thinks they can, not if you ask them directly. The interesting question is what we do with the not-knowing, and here the structure of the investment matters rather more than the temperament of the investor.

Venture capital has solved the behavioural problem with plumbing. A ten-year lock is not really a legal term, or not only that; it is a device for not looking. The venture investor accepts a portfolio in which most of the positions will die, holds it through a decade of silence, and is then celebrated for her patience, except that the patience is not entirely hers. It has been outsourced to the fund structure, which quietly removes the option to flinch. There are no daily marks to argue with, and no Tuesday afternoon on which the position is down 40 per cent and a committee wants a memo about it. The lock is institutionalised permission not to know.

Public equity offers no such mercy. The same uncertainty, which of these companies carries the future, gets priced every single day, in public, in front of your clients, so that your not-knowing has a ticker. When an early position in something genuinely new falls by half, nothing about the underlying question has actually changed, and yet everything about how it feels has. The market has not given you information. It has given you an emotion, hourly.

This is the part the discomfort hides. Whatever can be priced with confidence has already been competed away, because a known deck is worth nothing to own when everyone sees the same cards and bids them up to fair value. Whatever return is left over sits in the part no one can price: the wave whose winners are not yet named. Risk gets arbitraged; uncertainty does not, because there is no agreed shape to arbitrage against. The premium was never in the spreadsheet. It was always in the fog, which is precisely why the fog is so lightly occupied.

The capital has noticed which side is the more comfortable. American endowments now hold more in private and alternative strategies than they do in public equities; the latest NACUBO-Commonfund study puts the split at roughly 54 per cent against 31. Institutions have, in effect, paid to have their not-knowing warehoused somewhere nobody prices it. But the businesses themselves have not moved with the capital. For most companies of any consequence the public market is still the destination, the place a business goes when it needs permanent capital, a currency, and shareholders it has never met. The private phase is longer than it used to be. It is still only a phase. The future is incubated privately and delivered, eventually, to the exchange, where the not-knowing finally gets a price.

There is a subtler pretence, and it belongs to the people who believe they have opted out altogether. The investor who declines to hold any view of the future does not actually escape forecasting; he simply holds the index instead, and the index is a forecast. It is a portfolio of the last regime's winners, weighted by how completely they won, a record of the past rebalanced to look like the present. Owning it is a large and very quiet bet that the future will go on resembling what has already happened. It is the biggest forecast in the room, made by the people who insist they don't make forecasts, and it feels safe for the exact reason it may not be, which is that everyone is standing in the same place.

So the liquid investor in innovation is asked to do the one thing the venture investor is spared: to hold uncertainty with the exit door standing open. Every day the market offers you the chance to convert your discomfort into cash and calls it prudence. Selling early is the tax we pay for needing to feel right, and the record on this is not flattering. The same institutions that will hold a failed venture position all the way to the end of the fund, because they must, will trim a public winner the moment after it doubles, because they can.

None of this argues that not-knowing is a virtue in itself. Plenty of people don't know things and lose money with great conviction. The discipline is narrower than that. It is sizing a position so that being wrong is survivable and being right is material; it is accepting that the winners, if they turn out to be real, will spend years looking like mistakes, and that the drawdowns will arrive long before the evidence does. Volatility, from that seat, is not the risk showing up. It is the price of admission for owning the wave before it has been priced as a certainty. And it is declining to manufacture certainty on demand, for a committee, for a client, for yourself at three in the morning, because the certainty would be a fiction and the fiction would in the end prove expensive.

I notice that I still reach for the story myself. Someone asks what I think the next decade holds, and I can feel the answer forming, fluent and structured and complete, before I have decided whether I actually believe it. The fluency is the warning. The honest answer, which is that the future will be made by a small number of companies I cannot yet name, and that my job is mostly to stay solvent and present while I find out which ones, satisfies nobody in the room.

It has the sole advantage of being true.