Risk, Uncertainty, and Profit
By Frank H. Knight
The text has been altered as little as possible from the original edition (
Risk, Uncertainty, and Profit, Frank H. Knight, Ph.D., Associate Professor of Economics in the State University of Iowa; Boston and New York, Houghton Mifflin Co., The Riverside Press, 1921).A few corrections of obvious typos were made for this website edition. However, because the original edition was so internally consistent and carefully proofread, we have erred on the side of caution, allowing some typos (such as for proper nouns and within references) to remain lest someone doing academic research wishes to follow up. We have changed small caps to full caps for ease of using search engines.Lauren Landsburg
Editor, Library of Economics and Liberty
First Pub. Date
Boston, MA: Hart, Schaffner & Marx; Houghton Mifflin Co.
1st edition. Based on award-winning dissertation essay.
The text of this edition is in the public domain. Picture of Frank H. Knight courtesy of Ethel V. Knight.
Structures and Methods for Meeting Uncertainty
Part III, Chapter VIII
To preserve the distinction which has been drawn in the last chapter between the measurable uncertainty and an unmeasurable one we may use the term “risk” to designate the former and the term “uncertainty” for the latter. The word “risk” is ordinarily used in a loose way to refer to any sort of uncertainty viewed from the standpoint of the unfavorable contingency, and the term “uncertainty” similarly with reference to the favorable outcome; we speak of the “risk” of a loss, the “uncertainty” of a gain. But if our reasoning so far is at all correct, there is a fatal ambiguity in these terms, which must be gotten rid of, and the use of the term “risk” in connection with the measurable uncertainties or probabilities of insurance gives some justification for specializing the terms as just indicated. We can also employ the terms “objective” and “subjective” probability to designate the risk and uncertainty respectively, as these expressions are already in general use with a signification akin to that proposed.
The practical difference between the two categories, risk and uncertainty, is that in the former the distribution of the outcome in a group of instances is known (either through calculation
a priori or from statistics of past experience), while in the case of uncertainty this is not true, the reason being in general that it is impossible to form a group of instances, because the situation dealt with is in a high degree unique. The best example of uncertainty is in connection with the exercise of judgment or the formation of those opinions as to the future course of events, which opinions (and not scientific knowledge) actually guide most of our conduct. Now if the distribution of the different possible outcomes in a group of instances is known, it is possible to get rid of any real uncertainty by the expedient of grouping or “consolidating” instances. But that it
is possible does not necessarily mean that it
will be done, and we must observe at the outset that when an individual instance only is at issue, there is no difference for conduct between a measurable risk and an unmeasurable uncertainty. The individual, as already observed, throws his estimate of the value of an opinion into the probability form of
“a successes in
(a/b being a proper fraction) and “feels” toward it as toward any other probability situation.
As so commonly in this subject fraught with logical difficulty and paradox, reservations must be made to the above statement. In the first place, it does not matter how unique the instance, if a real probability can be calculated, if we can know with certainty how many successes there
would be in (say) one hundred trials
if the one hundred trials could be made. If we know the odds against us it does not matter in the least whether we place all our wagers in one kind of game or in as many different games as there are wagers; the laws of probability hold in the second case just as well as in the first. But in business situations it so rarely happens that a probability can be computed for a single unique instance that this qualification has less weight than might be supposed. However, in so far as objective probability enters into a calculation, it is hard to imagine an intelligent individual considering any single case as absolutely isolated. The only exception would be a decision in which one’s whole fortune (or his life) were at stake. The importance of the contingency and probable frequency of recurrence in the individual lifetime of situations similar in the magnitude of the issues involved should make a difference in the attitude assumed toward any one case as well as the mathematical probability of success or failure.
A second reservation of more importance is connected with the possibility referred to in the preceding chapter, of forming classes of cases by grouping the decisions of a given person. That is, even though we do not get a quantitative probability by the process of grouping, still there is some tendency for fluctuations to cancel out and for the result to approach constancy in some degree. There appear to be in the making of judgments the same two kinds of elements that we find in probability situations proper; i.e., (a) determinate factors (the quality of the judging faculty, which is more or less stable) and (b) truly accidental factors varying from one decision to another according to a principle of indifference. The difference between the uncertainty of an opinion and a true probability is that we have no means of separating the two and evaluating them, either by calculation
a priori or by empirical sorting. But in the second case the difference is not absolute; the sorting method does apply to some extent, though within narrow limits. Life is mostly made up of uncertainties, and the conditions under which an error or loss in one case may be compensated by other cases are bafflingly complex. We can only say that “in so far as” one confronts a situation involving uncertainty and deals with it on its merits as an isolated case, it is a matter of practical indifference whether the uncertainty is measurable or not.
The problem of the human attitude toward uncertainty (not for the present purpose distinguishing kinds) is as beset with difficulties as that of uncertainty itself. Not merely is the human reaction to situations of this character apt to be erratic and extremely various from one individual to another, but the “normal” reaction is subject to well-recognized deviations from the conduct which sound logic would dictate. Thus it is a familiar fact, well discussed by Adam Smith, that men will readily risk a small amount in the hope of winning a large when the adverse probability (known or estimated) against winning is much in excess of the ratio of the two amounts, while they commonly will refuse to incur a small chance of losing a larger amount for a virtual certainty of winning a smaller, even though the actuarial value of the chance is in their favor. To this bias must be added an inveterate belief on the part of the typical individual in his own “luck,” especially strong when the basis of the uncertainty is the quality of his own judgment. The man in the street has little more sense of the real value of his opinions than he has knowledge of the “logic” (if such it may be called) on which they rest. In addition, we must consider the almost universal prevalence of superstitions. Any coincidence that strikes attention is likely to be elevated into a law of nature, giving rise to a belief in an unerring “sign.” Even a mere “hunch” or “something tells me,” with no real or imaginary basis in the mind of the person himself, may readily be accepted as valid ground for action and treated as an unquestionable verity.
Doubtless in the long run of history there is a tendency toward rationality even in men’s whims and impulses. And if for no other reason than the impossibility of intelligently dealing with conduct on any other hypothesis, we seem justified in limiting our discussion to rational grounds of action. We shall assume, then, that if a man is undergoing a sacrifice for the sake of a future benefit, the expected reward must be larger in order to evoke the sacrifice if it is viewed as contingent than if it is considered certain, and that it will have to be larger in at least some general proportion to the degree of felt uncertainty in the anticipation.
*15 It is clearly the subjective uncertainty which is decisive in such a case, what the man believes the chances to be, whether his degree of confidence is based upon an objective probability in the situation itself or in an estimate of his own powers of prediction. We hold also that both the objective and subjective types may be involved at the same time, though no doubt most men do not carry their deliberations so far; the man’s opinion or prediction may be an estimate of an objective probability, and the estimate itself be recognized as having a certain degree of validity, so that the degree of felt uncertainty is a product of two probability ratios. It is to be emphasized again that practically all decisions as to conduct in real life rest upon opinions, and doubtless the greater part rest upon opinions which on scrutiny easily resolve themselves into an opinion of a probability—though as noted this “scrutiny” may not in most cases be given to the judgment by the individual making it.
The normal economic situation is of this character: The adventurer has an opinion as to the outcome, within more or less narrow limits. If he is inclined to make the venture, this opinion is either an expectation of a certain definite gain or a belief in the real probability of a larger one. Outside the limits of the anticipation any other result becomes more and more improbable in his mind as the amount thought of diverges either way. Hence it is correct to treat all instances of economic uncertainty as cases of choice between a smaller reward more confidently and a larger one less confidently anticipated.
At the bottom of the uncertainty problem in economics is the forward-looking character of the economic process itself. Goods are produced to satisfy wants; the production of goods requires time, and two elements of uncertainty are introduced, corresponding to two different kinds of foresight which must be exercised. First, the end of productive operations must be estimated from the beginning. It is notoriously impossible to tell accurately when entering upon productive activity what will be its results in physical terms, what (a) quantities and (b) qualities of goods will result from the expenditure of given resources. Second, the wants which the goods are to satisfy are also, of course, in the future to the same extent, and their prediction involves uncertainty in the same way. The producer, then, must
estimate (1) the future demand which he is striving to satisfy and (2) the future results of his operations in attempting to satisfy that demand.
It goes without saying that rational conduct strives to reduce to a minimum the uncertainties involved in adapting means to ends. This does not mean, be it emphasized, that uncertainty as such is abhorrent to the human species, which probably is not true. We should not really prefer to live in a world where everything was “cut and dried,” which is merely to say that we should not want our activity to be all perfectly rational. But in attempting to act “intelligently” we
are attempting to secure adaptation, which means foresight, as perfect as possible. There is, as already noted, an element of paradox in conduct which is not to be ignored. We find ourselves compelled to strive after things which in a “calm, cool hour” we admit we do not want, at least not in fullness and perfection. Perhaps it is the manifest impossibility of reaching the end which makes it interesting to strive after it. In any case we
do strive to reduce uncertainty, even though we should not want it eliminated from our lives.
The possibility of reducing uncertainty depends again on two fundamental sets of conditions: First, uncertainties are less in groups of cases than in single instances. In the case of
a priori probability the uncertainty tends to disappear altogether, as the group increases in inclusiveness; with statistical probabilities the same tendency is manifest in a less degree, being limited by defectiveness of classification. And even the third type, true uncertainties, show some tendency toward regularity when grouped on the basis of nearly any similarity or common element. The second fact or set of facts making for a reduction of uncertainty is the differences among human individuals in regard to it. These differences are of many kinds and an enumeration of them will be undertaken presently. We may note here that they may be differences in the men themselves or differences in their position in relation to the problem. We may call the two fundamental methods of dealing with uncertainty, based respectively upon reduction by grouping and upon selection of men to “bear” it, “consolidation”
*16 and “specialization,” respectively. To these two methods we must add two others which are so obvious as hardly to call for discussion: (3) control of the future, and (4) increased power of prediction. These are closely interrelated, since the chief practical significance of knowledge is control, and both are closely identified with the general progress of civilization, the improvement of technology and the increase of knowledge. Possibly a fifth method should be named, the “diffusion” of the consequences of untoward contingencies. Other things equal, it is a gain to have an event cause a loss of a thousand dollars each to a hundred persons rather than a hundred thousand to one person; it is better for two men to lose one eye than for one to lose two, and a system of production which wounds a larger number of workers and kills a smaller number is to be regarded as an improvement. In practice this diffusion is perhaps always associated with consolidation, but there is a logical distinction between the two and they may be practically separable in some cases. We must observe also that consolidation and specialization are intimately connected, a fact which will call for repeated emphasis as we proceed. In addition to these methods of dealing with uncertainty there is (6) the possibility of directing industrial activity more or less along lines in which a minimal amount of uncertainty is involved and avoiding those involving a greater degree.
One of the most immediate and most important consequences of uncertainty in economics may be disposed of as a preliminary to a detailed technical discussion. The essence of organized economic activity is the production by certain persons of goods which will be used to satisfy the wants of other persons. The first question which arises then is, which of these groups in any particular case, producers or consumers, shall do the foreseeing as to the future wants to be satisfied. It is perhaps obvious that the function of prediction in the technological side of production itself inevitably devolves upon the producer. At first sight it would appear that the consumer should be in a better position to anticipate his own wants than the producer to anticipate them for him, but we notice at once that this is not what takes place. The primary phase of economic organization is the production of goods for a general market, not upon direct order of the consumer. With uncertainty absent it would be immaterial whether the exchange of goods preceded or followed actual production. With uncertainty (in the two fields, production and wants) present it is still conceivable that men might exchange productive services instead of products, but the fact of uncertainty operates to bring about a different result. To begin with, modern society is organized on the theory (whatever the facts, about which some doubt may be expressed) that men predict the future and adapt their conduct to it more effectively when the results accrue to themselves than when they accrue to others. The responsibilities of controlling production thus devolve upon the producer.
But the consumer does not even contract for his goods in advance, generally speaking. A part of the reason might be the consumer’s uncertainty as to his ability to pay at the end of the period, but this does not seem to be important in fact. The main reason is that he does not know what he will want, and how much, and how badly; consequently he leaves it to producers to create goods and hold them ready for his decision when the time comes. The clue to the apparent paradox is, of course, in the “law of large numbers,” the consolidation of risks (or uncertainties). The consumer is, to himself, only one; to the producer he is a mere multitude in which individuality is lost. It turns out that an outsider can foresee the wants of a multitude with more ease and accuracy than an individual can attain with respect to his own. This phenomenon gives us the most fundamental feature of the economic system,
production for a market, and hence also the general character of the environment in relation to which the effects of uncertainty are to be further investigated. Before continuing the inquiry into other phases and methods of the consolidation of risks, we shall turn briefly to consider the differences among individuals in their attitudes and reactions toward measurable or unmeasurable uncertainty.
We assume, as already observed, that although life is no doubt more interesting when conduct involves a certain amount of uncertainty,—the proper amount varying with individuals and circumstances,—yet that men do actually strive to anticipate the future accurately and adapt their conduct to it. In this respect we may distinguish at least five variable elements in individual attributes and capacities. (1) Men differ in their capacity by perception and inference to
form correct judgments as to the future course of events in the environment. This capacity, furthermore, is far from homogeneous, some persons excelling in foresight in one kind of problem situations, others in other kinds, in almost endless variety. Of especial importance is the variation in the power of reading human nature, of forecasting the conduct of other men, as contrasted with scientific judgment in regard to natural phenomena. (2) Another, though related, difference is found in men’s capacities to
judge means and discern and plan the steps and adjustments necessary to meet the anticipated future situation. (3) There is a similar variation in the power to
execute the plans and adjustments believed to be requisite and desirable. (4) In addition there is diversity in conduct in situations involving uncertainty due to differences in the amount of
confidence which individuals feel in their judgments when formed and in their powers of execution; this degree of confidence is in large measure independent of the “true value” of the judgments and powers themselves. (5) Distinct from confidence felt is the
conative attitude to a situation upon which judgment is passed with a given degree of confidence. It is a familiar fact that some individuals want to be sure and will hardly “take chances” at all, while others like to work on original hypotheses and seem to prefer rather than to shun uncertainty. It is common to see people act on assumptions in ways which their own opinions of the value of the assumption do not warrant; there is a disposition to “trust in one’s luck.”
The amount of uncertainty effective in a conduct situation is the degree of subjective confidence felt in the contemplated act as a correct adaptation to the future—number 4 above. It is clear that we may speak in some sense of the “true value” of judgment and of capacity to act, but it is the person’s own opinion of these values which controls his activities. Hence the five variables are, from the standpoint of the person concerned, reduced to two, the (subjective or felt) uncertainty and his conative feeling toward it. For completeness we should perhaps add a sixth uncertainty factor, in the shape of occurrences so revolutionary and unexpected by any one as hardly to be brought under the category of an error in judgment at all.
In addition to the above enumeration of five or six distinct elements in the uncertainty situation we must point out that the first three variables named are themselves not simple. Judgment or foresight and the capacity for planning and the ability to execute action are each the product of at least four distinguishable factors, in regard to which the faculties in question may vary independently. These are (a) accuracy, (b) promptness or speed, (c) time range, and (d) space range, of the capacity or action. The first two of these require no explanation; it is evident that accuracy and rapidity of judgment and execution are more or less independent endowments. The third refers to the length of time in the future to which conduct is or may be adjusted, and the fourth to the scope or magnitude of the situation envisaged and the operations planned. Familiar also is the difference between individuals who have a mind for detail and those who confine their attention to the larger outlines of a situation. Even this rather complex outline is extremely simplified as compared with the facts of life in that it compasses only a rigidly “static” view of the problem. Quite as important as differences obtaining at any moment among individuals in regard to the attributes mentioned are their differences in capacity for change or development along the various lines.
Knowledge is more a matter of learning than of the exercise of absolute judgment. Learning requires time, and in time the situation dealt with, as well as the learner, undergoes change.
We have classified the possible reactions to uncertainty under some half-dozen heads, each of which gives rise to special problems, though the social structures for dealing with these problems overlap a good deal. The most fundamental facts regarding uncertainty from our point of view are, first, the possibility of reducing it in amount by grouping instances; and, second, the differences in individuals in relation to uncertainty, giving rise to a tendency to specialize the function of meeting it in the hands of certain individuals and classes. The most fundamental effect of uncertainty on the social-economic organization—production for a general market on the producer’s responsibility—has already been taken up; it is primarily a case of reduction of uncertainty by consolidation or grouping of cases. In the mere fact of production for a market, there is little specialization of uncertainty-bearing, and what there is is on a basis of the producer’s position in relation to the problem, not his peculiar characteristics as a man. To isolate the phenomenon of production for a market from other considerations we must picture a pure “handicraft stage” of social organization. In such a system every individual would be an independent producer of some one finished commodity, and a consumer of a great variety of products. The late Middle Ages afford a picture of an approximation to such a state of affairs in a part of the industrial field.
The approximation is rather remote, however. A handicraft organization shows an irresistible tendency to pass over, even before well established, into a very different system, and this further development is also a consequence of the presence of uncertainty. The second system is that of “free enterprise” which we find dominant to-day. The difference between free enterprise and mere production for a market represents the addition of specialization of uncertainty-bearing to the grouping of uncertainties, and takes place under pressure of the same problem, the anticipation of wants and control of production with reference to the future. Under free enterprise the solution of this problem, already removed from the consumer himself, is further taken out of the hands of the great mass of producers as well and placed in charge of a limited class of “entrepreneurs” or “business men.” The bulk of the producing population cease to exercise responsible control over production and take up the subsidiary rôle of furnishing productive resources (labor, land, and capital) to the entrepreneur, placing them under his sole direction for a fixed contract price.
We shall take up this phenomenon of free enterprise for detailed discussion in the next chapter, though we may note here two further facts regarding it; first, the “specialization” of uncertainty-bearing in the hands of entrepreneurs involves also a further consolidation; and, second, it is closely connected with changes in technological methods which (a) increase the time length of the production process and correspondingly increase the uncertainty involved, and (b) form producers into large groups working together in a single establishment or productive enterprise and hence necessitates concentration of control. The remainder of the present chapter will be devoted to a survey of the social structures evolved for dealing with uncertainty. Some of the phenomena will thus be finally disposed of, so far as the present work is concerned, especially those which already have a literature of their own and whose general bearings and place in a systematic treatment of uncertainty alone call for notice here. Other problems will be merely sketched in outline and reserved for fuller treatment in subsequent chapters, as has just been done with the subject of entrepreneurship.
Following the order of the classification already given of methods of dealing with uncertainty, the first subject for discussion is the institutions or special phenomena arising from the tendency to deal with uncertainty by
consolidation. The most obvious and best known of these devices is, of course,
insurance, which has already been repeatedly used as an illustration of the principle of eliminating uncertainty by dealing with groups of cases instead of individual cases. In our discussion of the theory of uncertainty in the foregoing chapter and at other points in the study we have emphasized the radical difference between a measurable and an unmeasurable uncertainty. Now measurability depends on the possibility of assimilating a given situation to a group of similars and finding the proportions of the members of the group which may be expected to exhibit the various possible outcomes. This assimilation of cases into classes may be exceedingly accurate, and the proportions of the various outcomes
may be computable on
a priori grounds by the application of the theory of permutations and combinations to determine the possible groupings of
equally probable alternatives; but this rarely if ever happens in a practical business situation. The classification will be of all degrees of precision, but the ascertainment of proportions must be empirical. The application of the insurance principle, converting a larger contingent loss into a smaller fixed charge, depends upon the measurement of probability on the basis of a fairly accurate grouping into classes. It is in general not enough that the insurer who takes the “risk” of a large number of cases be able to predict his aggregate losses with sufficient accuracy to quote premiums which will keep his business solvent while at the same time imposing a burden on the insurer which is not too large a fraction of his contingent loss. In addition he must be able to present a fairly plausible contention that the particular insured is contributing to the total fund out of which losses are paid as they accrue in an amount corresponding reasonably well with his real probability of loss; i.e., that he is bearing his fair share of the burden.
The difficulty of a satisfactory logical discussion of the questions we are dealing with has repeatedly been emphasized, due to the fact that distinctions of the greatest importance tend to run together through intermediate degrees and become blurred. This is conspicuously the case with the measurability of uncertainty through classification of instances. We hardly find in practice really homogeneous classifications (in the sense in which mathematical probability implies, as in the case of successive throws of a perfect die) and at the other extreme it is hard to find cases which do not admit of some possibility of assimilation into groups and hence of measurement. Indeed, the very concept of contingency seems to preclude absolute uniqueness (as for that matter there is doubtless nothing absolutely unique in the universe). For to say that a certain event is contingent or “possible” or “may happen” appears to be equivalent to saying that “such things ” have been known to happen before, and the “such things” manifestly constitute a class of cases formed on some ground or other. The principal subject for investigation is thus the
degree of assimilability, or the amount of homogeneity of classes securable, or, stated inversely, the
degree of uniqueness of various kinds of business contingencies. Insurance deals with those which are “fairly” classifiable or show a relatively low degree of uniqueness, but the different branches of insurance show a wide range of variation in the accuracy of measurement of probability which they secure.
Before taking up various types of insurance we may note in passing a point which it is superfluous to elaborate in this connection, namely, that different forms of organization in the insurance field all operate on the same principle. It matters not at all whether the persons liable to a given contingency organize among themselves into a fraternal or mutual society or whether they separately contract with an outside party to bear their losses as they fall in. Under competitive conditions and assuming that the probabilities involved are accurately known, an outside insurer will make no clear profit and the premiums will under either system be equal to the administrative costs of carrying on the business.
The branch of insurance which is most highly developed, meaning that its contingencies are most accurately measured because its classifications are most perfect, and which is thus on the most nearly “mathematical” basis is, of course, what is called “life insurance.” (In so far as it is “insurance” at all, and not a mere investment proposition, it is clear that it is insurance against “premature” loss of earning power, and not against death.) It is possible, on the basis of medical examinations, and taking into account age, sex, place of residence, occupation, and habits of life, to select “risks” which closely approximate the ideal of mechanical probability. The chance of death of two healthy individuals similarly circumstanced in the above regards seems to be about as near an objective equality, the life or death of one rather than the other about as nearly really indeterminate, as, anything in nature. To be sure, when we pass outside the relatively narrow circle of “normal” individuals, difficulties are encountered, but the extension of life insurance outside this circle has also been restricted. Some development has taken place in the insurance of sub-standard lives at higher rates, but it is limited in amount and could be characterized as exceptional.
The very opposite situation from life insurance is found in insurance against sickness and accident. Here an objective description and classification of cases is impossible, the business is fraught with great difficulties and susceptible of only a limited development. It is notorious that such policies cost vastly more than they should; indeed, the companies find it profitable to adopt a generous attitude in the adjustment of claims, raising the premium rates accordingly, it is needless to say. Accident compensation for workingmen, under social control, is on a somewhat better footing, but only on condition that the payments are restricted to not too large a fraction of the actual economic loss to the individual, with nothing for discomfort, pain, or inconvenience. In the whole field of personal, physical contingencies, however, there is nothing that is strictly of the nature of a “business risk,” unless it be the now happily obsolescent phenomenon of commercial employers’ liability insurance.
The typical application of insurance to business hazards is in the protection against loss by fire, and the theory of fire insurance rates forms an interesting contrast with the actuarial mathematics of life insurance. The latter, as we have observed, is a fairly close approximation to objective probability; it is in fact so close to this ideal that life insurance problems are worked by the formulæ derived from the binomial law, in the same way as problems in mechanical probability. Fire insurance rating is a very different proposition; only in rather recent years has any approach been made to the formation of fairly homogeneous classes of risks and the measurement of real probability in a particular case. At best there is a large field for the exercise of “judgment” even after literally thousands of classes of risks have been more or less accurately defined.
*18 More important is the fact that, in consequence, insurance does not take care of the whole risk against loss by fire. On account of the “moral hazard” and practical difficulties, it is necessary to restrict the amount of insurance to the “direct loss or damage” or even to a part of that, while of course there are usually large indirect losses due to the interruption of business and dislocation of business plans which are entirely unprovided for. Thus there is a large margin of uncertainty both to insurer and insured, in consequence of the impossibility of objectively homogeneous groupings and accurate measurement of the chance of loss. Corresponding to this margin of uncertainty in the calculations there is a chance for a profit or loss to either party, in connection with the fire hazard. The probabilities in the case of fire are, of course, complicated by the fact that risks are not entirely independent. A fire once started is likely to spread and there is a tendency for losses to occur in groups. In so far, however, as fire losses in the aggregate are calculable in advance, they are or may be converted into fixed costs by every individual exposed to the possibility of loss, and in so far no profit, positive or negative, will be realized by any one on account of this uncertainty in his business.
The principle of insurance has also been utilized to provide against a great variety of business hazards other than fire—the loss of ships and cargoes at sea, destruction of crops by storms, theft and burglary, embezzlement by employees (indirectly through bonding, the employee doing the insuring), payment of damages to injured employees, excessive losses through credit extension, etc. The unusual forms of policies issued by some of the Lloyd’s underwriters have attained a certain amount of publicity as popular curiosities. These various types of contingencies offer widely divergent possibilities for “scientific” rate-making, from something like the statistical certainty of life insurance at one extreme to almost pure guesswork at the other, as when Lloyd’s insures the business interests concerned that a royal coronation will take place as scheduled, or guarantees the weather in some place having no records to base calculations upon. Even in these extreme cases, however, there is a certain vague grouping of cases on the basis of intuition or judgment; only in this way can we imagine any estimate of a probability being arrived at.
It is therefore seen that the insurance principle can be applied even in the almost complete absence of scientific data for the computation of rates. If the estimates are conservative and competent, it turns out that the premiums received for insuring the most unique contingencies cover the losses; that there is an offsetting of losses and gains from one venture to another, even when there is no discoverable kinship among the ventures themselves. The point seems to be, as already noticed, that the mere fact that judgment is being exercised in regard to the situations forms a fairly valid basis for assimilating them into groups. Various instances of the exercise of (fairly competent) judgment even in regard to the most heterogeneous problems, show a tendency to approach a constancy and predictability of result when aggregated into groups.
The fact which limits the application of the insurance principle to business risks generally is not therefore their inherent uniqueness alone, and the subject calls for further examination. This task will be undertaken in detail in the next chapter, which deals with entrepreneurship. At this point we may anticipate to the extent of making two observations: first, the typical uninsurable (because unmeasurable and this because unclassifiable) business risk relates to the exercise of judgment in the making of decisions by the business man; second, although such estimates do tend to fall into groups within which fluctuations cancel out and hence to approach constancy and measurability, this happens only
after the fact and, especially in view of the brevity of a man’s active life, can only to a limited extent be made the basis of prediction. Furthermore, the classification or grouping can only to a limited extent be carried out by any agency outside the person himself who makes the decisions, because of the peculiarly obstinate connection of a
moral hazard with this sort of risks. The decisive factors in the case are so largely on the inside of the person making the decisions that the “instances” are not amenable to objective description and external control.
Manifestly these difficulties, insuperable when the “consolidation” is to be carried out by an external agency such as an insurance company or association, fall away in so far as consolidation can be effected within the scale of operations of a single individual; and the same will be true of an organization if responsibility can be adequately centralized and unity of interest secured. The possibility of thus reducing uncertainty by transforming it into a measurable risk through grouping constitutes a strong incentive to extend the scale of operations of a business establishment. This fact must constitute one of the important causes of the phenomenal growth in the average size of industrial establishments which is a familiar characteristic of modern economic life. In so far as a single business man, by borrowing capital or otherwise, can extend the scope of his exercise of judgment over a greater number of decisions or estimates, there is a greater probability that bad guesses will be offset by good ones and that a degree of constancy and dependability in the total results will be achieved. In so far uncertainty is eliminated and the desideratum of rational activity realized.
Not less important is the incentive to substitute more effective and intimate forms of association for insurance, so as to eliminate or reduce the moral hazard and make possible the application of the insurance principle of consolidation to groups of ventures too broad in scope to be “swung” by a single enterpriser. Since it is capital which is especially at risk in operations based on opinions and estimates, the form of organization centers around the provisions relating to capital. It is undoubtedly true that the reduction of risk to borrowed capital is the principal desideratum leading to the displacement of individual enterprise by the partnership and the same fact with reference to both owned and borrowed capital explains the substitution of corporate organization for the partnership. The superiority of the higher form of organization over the lower from this point of view consists both in the extension of the scope of operations to include a larger number of individual decisions, ventures, or “instances,” and in the more effective unification of interest which reduces the moral hazard connected with the assumption by one person of the consequences of another person’s decisions.
The close connection between these two considerations is manifest. It is the special “risk” to which large amounts of capital loaned to a single enterpriser are subject which limits the scope of operations of this form of business unit by making it impossible to secure the necessary property resources. On the other hand, it is the inefficiency of organization, the failure to secure effective unity of interest, and the consequent large risk due to moral hazard when a partnership grows to considerable size, which in turn limit its extension to still larger magnitudes and bring about the substitution of the corporate form of organization. With the growth of large fortunes it becomes possible for a limited number of persons to carry on enterprises of greater and greater magnitude, and to-day we find many very large businesses organized as partnerships. Modifications of partnership law giving this form more of the flexibility of the corporation with reference to the distribution of rights of control, of participation in income, and of title to assets in case of dissolution have also contributed to this change.
With reference to the first of our two points above mentioned, the extension of the scope of operations, the corporation may be said to have solved the organization problem. There appears to be hardly any limit to the magnitude of enterprise which it is possible to organize in this form, so far as mere ability to get the public to buy the securities is concerned. On the second score, however, the effective unification of interests, though the corporation has accomplished much in comparison with other forms of organization, there is still much to be desired. Doubtless the task is impossible, in any absolute sense; nothing but a revolutionary transformation in human nature itself can apparently solve this problem finally, and such a change would, of course, obliterate all moral hazards at once, without organization. In the meanwhile the internal problems of the corporation, the protection of its various types of members and adherents against each other’s predatory propensities, are quite as vital as the external problem of safeguarding the public interests against exploitation by the corporation as a unit.
Another important aspect of the relations of corporate organization to risk involves what we have called “diffusion” as well as consolidation. The minute divisibility of ownership and ease of transfer of shares enables an investor to distribute his holdings over a large number of enterprises in addition to increasing the size of a single enterprise. The effect of this distribution on risk is evidently twofold. In the first place, there is to the investor a further offsetting through consolidation; the losses and gains in different corporations in which he owns stock must tend to cancel out in large measure and provide a higher degree of regularity and predictability in his total returns. And again, the chance of loss of a small fraction of his total resources is of less moment even proportionally than a chance of losing a larger part.
There are other aspects of the question which must be passed over in this summary view. Doubtless a significant fact is the greater publicity attendant upon the organization, resources, and operations of a corporation, due to its being a creature of the State and to legal safeguards. It must be emphasized that this type of organization actually reduces risks, and does not merely transfer them from one party to another, as might seem at first glance to be the case. Superficial discussions of limited liability tend to give the impression, or at least leave the way open to the conclusion, that this is the main advantage over the partnership. But it must be evident that the mere fact of limited liability only serves to transfer losses in excess of invested resources from the owners of the concern to its creditors; and if this were the only effect of incorporation, the loss in credit standing should offset the gain in security to the owners. The vital facts are the twofold consolidation of risks, together with greater publicity, and diffusion in a minor rôle, not really separable from the tact of consolidation.
It is particularly noteworthy that large-scale organization has shown a tendency to grow in fields where division of labor is absent and consolidation or grouping of uncertainties is the principal incentive. Occupations in which the work is of an occasional and intermittent character tend to run into partnerships and even corporations where there is no capital investment, or relatively little, and the members work independently at identical tasks. Examples are the syndicating of detectives, stenographers, and even lawyers and doctors.
The second of the two main principles for dealing with uncertainty is Specialization. The most important instrument in modern economic society for the specialization of uncertainty, after the institution of free enterprise itself, is
Speculation. This phenomenon also combines different principles, and the mere specialization of uncertainty-bearing in the hands of persons most willing to assume the function is probably among the lesser rather than the greater sources of gain. It seems best to postpone for the present a detailed theoretical analysis of the factors of specialization of uncertainty-bearing in the light of the many ways in which individuals differ in their relations to uncertainty; this discussion will be taken up in the next chapter, in connection with the treatment of enterprise and entrepreneurship. At this point we wish merely to emphasize the association in several ways between specialization and actual reduction of uncertainty.
Most fundamental among these effects in reducing uncertainty is its conversion into a measured risk or elimination by grouping which is implied in the very fact of specialization. The typical illustration to show the advantage of organized speculation to business at large is the use of the hedging contract. By this simple device the industrial producer is enabled to eliminate the chance of loss or gain due to changes in the value of materials used in his operations during the interval between the time he purchases them as raw materials and the time he disposes of them as finished product, “shifting” this risk to the professional speculator. It is manifest at once that even aside from any superior judgment or foresight or better information possessed by such a professional speculator, he gains an enormous advantage from the sheer magnitude or breadth of the scope of his operations. Where a single flour miller or cotton spinner would be in the market once, the speculator enters it hundreds or thousands of times, and his errors in judgment must show a correspondingly stronger tendency to cancel out and leave him a constant and predictable return on his operations.
The same reasoning holds good for any method of specializing uncertainty-bearing. Specialization implies concentration, and concentration involves consolidation; and no matter how heterogeneous the “cases” the gains and losses neutralize each other in the aggregate to an extent increasing as the number of cases thrown together is larger. Specialization itself is primarily an application of the insurance principle; but, like large-scale enterprise, it grows up to meet uncertainty situations where, on account of the impossibility of objective definition and external control of the individual ventures or uncertainties, a “moral hazard” prevents insurance by an external agency or a loose association of venturers for this single purpose.
Besides organized speculation as carried on in connection with produce and security exchanges, the principle of specialization is exemplified in the tendency for the highly uncertain or speculative aspects of industry to become separated from the stable and predictable aspects and be taken over by different establishments. This is, of course, what has really taken place in the ordinary form of speculation already noticed, namely, the separation of the
marketing function from the technological side of production, the former being much more speculative than the latter. A separation perhaps equally significant in modern economic life is that which so commonly takes place between the
founding of new enterprises and their
operation after they are set going. To be sure, by no means all the business of
promotion comes under this head, but still the tendency is manifest. A part of the investors in promoted concerns look to the future earnings from regular operations for their return, but a large part expect to sell out at a profit after the business is established, and to devote their capital to some new venture of the same sort. A considerable and increasing number of individual promoters and corporations give their exclusive attention to the launching of new enterprises, withdrawing entirely as soon as the prospects of the business become fairly determinate. The gain from arrangements of this sort arises largely from the consolidation of uncertainties, their conversion by grouping into measured risks which are for the group of cases not uncertainties at all. Such a promoter takes it as a matter of course that a certain proportion of his ventures will be failures and involve heavy losses, while a larger proportion will be relatively unprofitable, and counts on making his gains from the occasional conspicuous successes. That is—to face frankly that paradoxical element which is really involved in such calculations—he does not “expect” to have his “expectations” verified by the results in every case; the expectations on which he really counts are based on an average, on an “estimate” of the long-run value of his “estimates.” The specialization in the speculative phase of the business enables a single man or firm to deal with a larger number of ventures, and is clearly a mode of applying the same principle which underlies ordinary insurance.
Other illustrations of the same phenomenon will come to the reader’s mind. Industries which utilize land whose value is largely speculative are more likely to rent rather than own their sites where the nature of the utilization makes such a procedure practicable. Even expensive machines and articles of equipment of other sorts, ownership of which involves heavy risks to a small concern, may be rented instead of bought outright. The owner of leased land or equipment is presumably a specialist in that sort of business and his risks are reduced by the grouping of a larger number of ventures.
Other advantages of specialization of speculative functions in addition to the reduction of uncertainty through consolidation are manifest, and no intention of belittling or concealing them is implied in the separation of the latter aspect of the case in the foregoing discussion. It is apparent in particular that the specialist in any line of risk-taking naturally knows more about the problem with which he deals than would a venturer who dealt with them only occasionally. Hence, since most of these uncertainties relate chiefly to the exercise of judgment; the uncertainty itself is reduced by this fact also. There is in this respect a fundamental difference between the speculator or promoter and the insurer, which must be kept clearly in view. The insurer knows more about the risk in a particular case—say of a building burning—but the
real risk is no less because he assumes it in that particular case. His risk is less only because he assumes a large number. But the transfer of the “risk” of an error in judgment is a very different matter. The “insurer” (entrepreneur, speculator, or promoter) now substitutes his own judgment for the judgment of the man who is getting rid of the uncertainty through transferring it to the specialist. In so far as his knowledge and judgment are better, which they almost certainly will be from the mere fact that he is a specialist, the individual risk is less likely to become a loss, in addition to, the gain from grouping. There is better management, greater economy in the use of economic resources, as well as a mere transformation of uncertainty into certainty.
The problem of meeting uncertainty thus passes inevitably into the general problem of management, of economic control. The fundamental uncertainties of economic life are the errors in predicting the future and in making present adjustments to fit future conditions. In so far as ignorance of the future is due to practical indeterminateness in nature itself we can only appeal to the law of large numbers to distribute the losses, and make them calculable, not to reduce them in amount, and this is only possible in so far as the contingencies to be dealt with admit of assimilation into homogeneous groups; i.e., in so far as they repeat themselves. When our ignorance of the future is only partial ignorance, incomplete knowledge and imperfect inference, it becomes impossible to classify instances objectively, and any changes brought about in the conditions surrounding the formation of an opinion are nearly sure to affect the intrinsic value of the opinion itself. This is true even of the method of grouping by extending the scale of operations of a single entrepreneur, for the quality of his estimates will not be independent of the number he has to make and the mass of the data involved. But it is especially true of grouping by specialization, as we have seen. The inseparability of the uncertainty problem and the managerial problem will be especially important in the discussion (in the next chapter) of entrepreneurship, which is the characteristic phenomenon of modern economic organization and is essentially a device for specializing uncertainty-bearing or the improvement of economic control. The relation between management, which consists of making decisions, and taking the consequences of decisions, which is the most fundamental form of risk-taking in industry, will be found to be a very intricate as well as intimate one. When the sequence of control is followed through to the end, it will be found that from the standpoint of the ultimately responsible manager, the two functions are always inseparable.
We are thus brought naturally around to a discussion of the most thoroughgoing methods of dealing with uncertainty; i.e., by securing better knowledge of and control over the future. As previously observed, however, these methods represent merely the objective of all rational conduct from the outset, and they call for discussion in such a work as the present only in so far as they affect the general outline of the social economic structure. Thus it is fundamental to the entrepreneur system that it tends to promote better management in addition to consolidating risks and throwing them into the hands of those most disposed to assume them. The only further comment here called for is to point out the existence of highly specialized industrial structures performing the functions of furnishing knowledge and guidance.
One of the principal gains through organized speculation is the provision of information on business conditions, making possible more intelligent forecasting of market changes. Not merely do the market associations or exchanges and their members engage in this work on their own account. Its importance to society at large is so well recognized that vast sums of public money are annually expended in securing and disseminating information as to the output of various industries, crop conditions, and the like. Great investments of capital and elaborate organizations are also devoted to the work as a private enterprise, on a profit-seeking basis, and the importance of trade journals and statistical bureaus and services tends to increase, as does that of the activities of the Government in this field. The collection, digestion, and dissemination in usable form of economic information is one of the staggering problems connected with our modern large-scale social organization. It goes without saying that no very satisfactory solution of this problem has been achieved, and it is safe to predict that none will be found in the near future. But all these specialized agencies for the supply of information help to bridge the wide gap between what the individual business manager knows or can find out by the use of his own resources and what he would have to know to conduct his business in a perfectly intelligent fashion. Their output increases the value of the intuitive “judgments” on the basis of which his decisions are finally made after all, and greatly extends the scope of the environment in relation to which he can more or less intelligently react.
The foregoing relates chiefly to the production side of the problem of economic information. In the field of information for consumers, we have the still more staggering development of advertising. This complex phenomenon cannot be discussed in detail here, beyond pointing out its connection with the fact of ignorance and the necessity of knowledge to guide conduct. Only a part of advertising is in any proper sense of the term informative. A larger part is devoted to persuasion, which is a different thing from conviction, and perhaps the stimulation or creation of new wants is a function distinguishable from either. In addition to advertising, most of the social outlay for education is connected with informing the population about the means of satisfying wants, the education of taste. The outstanding fact is that the ubiquitous presence of uncertainty permeating every relation of life has brought it about that information is one of the principal commodities that the economic organization is engaged in supplying. From this point of view it is not material whether the “information” is false or true, or whether it is merely hypnotic suggestion. As in all other spheres of competitive economic activity, the consumer is the final judge. If people are willing to pay for “Sunny Jim” poetry and “It Floats” when they buy cereals and soap, then these wares are economic goods. If a certain name on a fountain pen or safety razor enables it to sell at a fifty per cent higher price than the same article would otherwise fetch, then the name represents one third of the economic utility in the article, and is economically no different from its color or design or the quality of the point or cutting edge, or any other quality which makes it useful or appealing. The morally fastidious (and naïve) may protest that there is a distinction between “real” and “nominal” utilities; but they will find it very dangerous to their optimism to attempt to follow the distinction very far. On scrutiny it will be found that most of the things we spend our incomes for and agonize over, and notably practically all the higher “spiritual” values, gravitate swiftly into the second class.
Somewhat different from the production and sale of information is the dealing in actual instructions for the guidance of conduct directly. Modern society is characterized by the rapid growth of this line of industry also. There have always been a few professions whose activities consisted essentially of the sale of guidance, notably medicine and the law, and more or less the preaching and teaching professions. Recent years, however, have witnessed a veritable swarming of experts and consultants in nearly every department of industrial life. The difference from dealing in information is that these people do not stop at diagnosis; in addition they prescribe. They are equally conspicuous in the fields of business organization, accounting, the treatment of labor, the lay-out of plants, and the processing of materials; they are the scientific managers of the managers of business; and though they by no means serve business or its managers for naught, and in spite of a large amount of quackery, they probably pay their way and more on the whole in increasing the efficiency of production. Certainly they do a useful work in forcing the intelligent, critical consideration of business problems instead of a blind following of tradition or the use of guesswork methods.
The last of the alternatives named for meeting uncertainty relates to the problem of a tendency to prefer relatively predictable lines of activity to more speculative operations. It is common to assume
*21 that society pays for the assumption of risk in the form of higher prices for commodities whose production involves uncertainty and a deficient supply of these in comparison with goods of an opposite character. This subject will come up again in connection with the closely related question of a tendency of profit to zero, and it seems best to postpone discussion of it for the present.
*22 We shall find reasons for being very skeptical as to the reality of any such abhorrence of uncertainty as to decrease productivity in any line below the level that an equivalent fixed cost would bring about.
“summum bonum” or any other objective
“bonum” is an
ignis fatuus. We are compelled to believe that in a great proportion of cases we take more interest in action whose fruition is only probable than we would if it were certain.
(The Nature of Capital and Income, p. 288). I should prefer simply “grouping” as both shorter and more descriptive.
Property Insurance, chaps. XVI, XVII.
(Business Organization and Combination, chap. XXIII) uses the terms “The Corporation Problem” and “The Trust Problem” to designate what I have called the “internal” and “external” problems respectively. He properly emphasizes the importance of the former in view of the tendency of the evils of monopoly, etc., to overshadow it in the popular mind and in much of the literature of the subject.
Journal of Political Economy, vol. 26, Nos. 1 and 2.
Economic Theory of Risk and Insurance, chap. III.
Part III, Chapter IX.