A trading house operates a long-standing, profitable industrial asset in an emerging market. The government announces a reform whose bearing on strategic assets is neither specified nor dated. Nothing has occurred. The leadership nevertheless faces a choice no method can dispose of. To continue the planned expansion is to accept an exposure whose variables do not belong to the house: the political trajectory of the host country, the narratives that will form around it, and the decisions of actors whose logic is not the house's own. To freeze is to renounce real growth and positions that will not return. Whatever the leadership decides, something that mattered will be lost.
Decisions of this kind are not settled by better analysis. They reach the top precisely because the choice cannot be delegated in full. This essay is about what remains once analysis has reached its limit.
Two kinds of decision
There are decisions where the right answer can be found: the options reduce to a common measure — cost, risk, efficiency — and the work is to bring it out. And there are decisions where the options cannot be compared, because they are not measured in the same unit. Keeping an ill parent at home, or placing him where he will be better cared for. Keeping a promise, or breaking it to spare a pain one had not foreseen. Choosing one side here is not discovering that the other was worth less, it is renouncing something that remained real. The regret one keeps is not a fault of reasoning. It is the trace of having chosen.
Bernard Williams called the second kind incommensurable. In the context of decisions taken within organisations, where the first kind has a name everyone uses — optimisation — let us give the second the name it already half-bears: judgment. The two ask different things. For optimisation, one looks for the right method, the right model, the right information. For judgment, no method decides in your place. One can be helped to see what will be lost whichever way the choice goes, but the choice itself remains yours. No one can carry it for you without taking it from you. This is the faculty Arendt called judging, as distinct from thinking: settling a particular case that no rule covers.
The residue analysis cannot dissolve
Management schools and most advisory practices naturally put forward what can be modelled. They are built to analyse, compare, and clarify; and because they often do this well, they encourage the hope that complex situations can be resolved by analysis alone. Often, the analytical work does frame the decision and clarify the landscape. But it cannot always settle the matter in full. What remains unsettled is what reaches the leader's desk, and what his judgment must carry.
The trading house above will be advised. Public affairs counsel will read the reform, geopolitical analysts will assess the political trajectory, lawyers will scope the regulatory exposure. The leadership may be told that staying carries a 65 percent chance of preserving growth and a 35 percent chance of severe loss if the reform turns toward nationalisation, while withdrawal carries a lower but certain loss. The calculation can be made, it can even discipline the discussion. But it does not settle the decision.
One could object that all this reduces, in principle, to a single magnitude: the firm's expected discounted profit. In practice, that magnitude is itself eroded from within. The probabilities that enter the calculation are not measurements in the actuarial sense. They are informed estimates whose underlying analysts could, in good faith, disagree by twenty points. The calculation also depends on second- and third-order effects: what other governments, partners and financiers will conclude about the firm's reliability as a long-term partner, how the decision will shape the host government's next move, and what will be lost by ceding the ground to a competitor. Each is plausible. None can be calculated with precision. This is the domain Knight called uncertainty, as distinct from risk.
Beyond uncertainty, the case meets the kind of incommensurability the essay began with. For instance: a firm that defines itself as a long-term operator in difficult markets, one that knows the ground, holds relationships across political cycles, and does not abandon strategic positions lightly once people, capital, and local credibility have been committed, does not simply lose value when it withdraws. It becomes a different firm. Or again: the firm does not arrive empty-handed to its decision. It carries the decisions it has already taken: positions defended, appointments made, capital committed. Whatever is decided now will be read against what was decided before, both inside the firm and outside it. Analysis clarifies the exposure, it does not abolish the need for judgment.
In organisations, such decisions reach the top by structure. Everything tractable — calculating, comparing, optimising — has been progressively entrusted to tools, analysts, procedures. What reaches the senior leader is what those tools could not resolve: by construction, decisions of judgment. They arrive at his desk because every method below has run its course.
Settling what remains to be decided, once everything else has been delegated, is not a residual task. It is what leading means.
References
Arendt, Hannah. The Life of the Mind. Harcourt Brace Jovanovich, 1978.
Knight, Frank H. Risk, Uncertainty and Profit. Houghton Mifflin, 1921.
Williams, Bernard. Moral Luck: Philosophical Papers 1973–1980. Cambridge University Press, 1981.
