← All publications

From the literature

The entrepreneur, or the art of judging without a manual

In Risk, Uncertainty, and Profit, the entrepreneur is not just the person who owns the business or coordinates tasks. He is the one who decides when the correct answer is not given by a procedure.

Knight starts from a contrast. In a world without true uncertainty, management could become a routine activity: the data would be known, the consequences calculable, the resources allocated according to predictable rules. The manager would then be one specialist among others, paid for his work like any other skilled employee. Running a business would become a technical trade, not a distinct economic function.

Two decisions, not one

Uncertainty changes the nature of the function, and Knight breaks this change down with useful precision. Deciding under uncertainty is not one act but two, nested within each other.

The first is the judgement about the world: what will happen, what is this technology worth, will this market exist? The second, more rarely mentioned, is the judgement about how much confidence to place in one's own judgement. How much to commit on this conviction? With what margin of error? At what point to conclude that one was wrong?

These two decisions are independent. One can have an excellent reading of a market and disastrous confidence in that reading, which leads to under-investment. One can have a mediocre reading and unshaken confidence, which leads to catastrophe with panache.

The firm as an architecture of delegation

The point becomes even more interesting in large organisations. No leader can personally know all the technical, commercial and human problems of a company of any size. He must therefore choose people who will judge in his place.

Knight derives from this a singular skill, which is perhaps the most modern idea in the book: the ability to judge other people's capacity for judgement. The leader becomes an evaluator of evaluators. His own competence is no longer technical; it is meta-technical.

An organisation is therefore not just a collection of skills; it is an architecture of delegation. At every level, someone decides whom to trust, with what degree of autonomy and under what responsibility. When information is incomplete, the quality of this chain of judgement becomes a central economic variable, as decisive as the cost of capital or the productivity of equipment.

Along the way, Knight resolves a difficulty that his predecessors avoided: that of the salaried executive. If the entrepreneur is the one who judges, what are we to make of the chief executive who decides without owning? The answer lies in shifting the problem: the person who has really exercised the entrepreneurial function is the one who judged that this post should be entrusted to this person, and who bears the consequences. The function does not disappear; it moves up a level.

A theory of self-confidence

Knight adds an observation with unexpected macroeconomic consequences. Two people may have the same technical competence but very different levels of confidence in their own abilities. One systematically underestimates what he can do; the other takes himself for Napoleon with an Excel spreadsheet.

Yet these differences do not remain individual. They determine the number of businesses created, the price that would-be entrepreneurs are willing to pay for productive resources, and therefore the average level of profits and losses in the economy. If overconfidence prevails, competition for resources drives up the guaranteed payments and squeezes the residual: entrepreneurs, collectively, pay too much for the right to try their luck. Knight in fact suggests that this is the most frequent case, which is an elegant way of explaining why so many entrepreneurial ventures reward their employees and suppliers better than their founders.

Entrepreneurship, in this framework, is therefore less a glorification of courage than a theory of responsibility under uncertainty. It is not enough to decide quickly. One must decide when no rule exists, accept that the outcome may contradict the initial judgement, and bear part of the consequences. The rest is management, which is useful but explains no profit.

Frank H. Knight, Risk, Uncertainty, and Profit (1921), chapters IX and X. Editorial adaptation, not a quotation.