The workshop pays its people before it knows what their work will fetch. Its orders may disappoint; the product may find a public nobody expected. The wage names an obligation in advance, while the owner accepts what remains after the undertaking has met its obligations. Between these two ways of receiving an income lies the entrepreneur’s wager. Someone commits resources to an opinion about the future, puts promises behind that opinion, and waits for the future to decide the amount of the remainder. The arrangement can give production a direction and the worker a steadier income. It cannot promise that the remainder will be a reward.
Frank Knight made this difference central to his account of enterprise. A measurable risk supplies a distribution on which calculation can work; uncertainty leaves the distribution itself open to judgment. The founder considering an unfamiliar product has evidence, experience, and perhaps an estimate of the odds, without necessarily possessing a dependable frequency from which to price every contingency. Judgment remains possible. Indeed, it becomes indispensable. What changes is the kind of warrant the judgment can claim, and the difficulty of finding someone willing to insure its error at a settled price. The entrepreneur acts before that warrant can be completed.
The promise is what makes this judgment an undertaking rather than an opinion. The person who predicts a market correctly over dinner need not have arranged a single delivery or put a single payment at hazard. The entrepreneur orders the material, employs the people, and accepts terms that will have to be honored even if customers prove less enthusiastic than expected. Knight’s account places the guarantee beside the direction: the person who undertakes to meet the agreed payments acquires a reason to decide which commitments should be made. Those who accept the payments obtain some shelter from the variability of the undertaking’s receipts. The firm gives the judgment a balance sheet.
The workshop’s cash remainder must be distinguished from profit above the owner’s opportunity costs. The workshop promises sixty in wages and spends forty on other inputs. Receipts of a hundred and forty leave forty; receipts of eighty leave a deficit of twenty. Paying everyone in the second case requires twenty from somewhere outside that period’s receipts. Now allow fifteen for the owner’s own labor and capital at their next-best uses. Economic profit becomes twenty-five or minus thirty-five, although the cash remainder remains forty or minus twenty. The positive cash remainder is consequently not all payment for bearing uncertainty: part can remunerate resources the owner supplied. Nor does the pair of possible outcomes establish what the venture was expected to earn. An account of what can happen is not yet a distribution of how often it will happen.
To see the separate work done by an expectation, temporarily give this example known odds. Let the higher receipts occur with probability three-fifths and the lower receipts with probability two-fifths. Expected receipts are then a hundred and sixteen, the expected cash remainder sixteen, and expected economic profit one after the fifteen-unit allowance. A realized cash gain of forty would greatly exceed the expected sixteen; a loss of twenty would not show that the initial expectation had been irrational. These probabilities are an added assumption for the illustration, precisely the information Knight’s harder case may lack. They show why an anticipated return, a realized return, and a judgment about the decision cannot be read from the same number.
The survivor therefore presents an evidentiary problem. The successful founder is visible, has money to speak with, and can tell the undertaking backward from its outcome. The failed founder may have acted on equally serious grounds and have little audience left. Yet success can contain information about judgment. Suppose, in another deliberately simple comparison, there are two equally likely kinds of decision-maker: one succeeds three times in four, the other once in four. Observing one success raises the probability of the first kind from one-half to three-quarters. It does not raise it to certainty. If both kinds instead succeed half the time, the same observation tells us nothing about which kind acted. What an outcome establishes depends on the competing explanations for it, including how differently those explanations predicted the outcome.
Knight treated a limited record of success as an uncertain guide to judgment. His entrepreneur exercises judgment about things and about the judgments of other people; engineers, managers, merchants, and financiers contribute assessments that someone must weigh. He also allowed luck an important place in the realized result, and treated inference from a small number of trials as precarious. A good outcome may strengthen a claim to foresight without measuring the contribution exactly. A bad outcome may expose an error without proving that a better decision was available beforehand. The residual disciplines the enterprise because it has to be lived with. It does not supply a complete biography of the knowledge that produced it.
There is a further difficulty in calling the entrepreneur the bearer of the uncertainty. A wage can be fixed in amount while remaining exposed to the employer’s ability to pay. In the workshop’s bad outcome, twenty of additional funds will keep the hundred in promises whole. If only ten is available, the promised payments are short by ten; naming someone the residual claimant does not create the missing resources. A supplier may receive an invoice instead of equity and still depend on the enterprise’s survival. A worker may receive the agreed wage this month and lose the relationship next month. The contract reallocates uncertainty, but the degree of shelter depends on the resources and commitments that support it.
Knight’s own concluding account of the enterprise makes room for this division. The manager who directs, the owner who guarantees, and the worker whose security depends on the undertaking need not occupy cleanly separated positions. Guarantees are incomplete; entrepreneurial functions can be distributed. This matters to the question of earnings. If one person receives a residual while others bear consequential exposures that their contracts have not removed, the residual alone cannot measure the whole burden carried in making production possible. The inquiry must follow the actual promises and actual losses. It must also distinguish compensation for a useful undertaking from the particular rule by which its eventual gain is divided.
Coase pressed a different limit on the explanation. A purchaser can promise a contractor a fixed sum for a specified delivery while leaving the contractor in charge of how the work is done. The purchaser still bears uncertainty about the later sale. A comparable payment to an employee can instead buy a bounded right to choose the work as circumstances unfold. Bearing the uncertain receipts consequently does not, by itself, explain why production must be directed inside a firm. Knight makes the relation between commitments and the remainder intelligible; Coase asks why this particular form of coordination was chosen to make the commitments effective. A theory of the venture and a theory of its boundary have work to do together.
The entrepreneur may thus earn something real through judgment and commitment while receiving an amount that cannot be identified with either contribution alone. To call every survivor deserving would make the outcome certify its own history. To call every survivor merely lucky would throw away information that outcomes sometimes contain. The harder task is to ask what was undertaken, what alternatives were available, what resources stood behind the promises, and what the observed result can discriminate among those explanations. These are questions the profit figure opens. They are not questions it settles merely by being positive.
The remainder records what the venture left, not everything its maker contributed.
The venture brings an opinion about the future into the present as wages, orders, equipment, and obligations. Its gain can sustain further experiments; its loss can end them. Both consequences matter without turning either into an infallible verdict on the person who acted. The next plate follows an undertaking that has survived and grown. Its success has gathered more operations beneath a single direction, and the question becomes whether the enlarged command makes production cheaper, or whether the power to enlarge itself has acquired reasons of its own. The argument remains open.