Review of Frank Knight's "Risk, Uncertainty and Profit"

What three things could excite the bourgeois mind more?

Editors Note: It’s not really fun to talk to people on Twitter anymore, because people have gotten wigged out by the whole Elon thing (understandably) and have started heading for the exits. In response, we are opening the comments section here to Everybody in the hopes that we can get some arguments going that are attached to this piece in information-space. Cheers!

---

Risk, Uncertainty And Profit! What three things could excite the bourgeois mind more? Were this a book for stallions, it might as well be called Salt, Apple And Mare. So far, we have mostly examined the idea of “uncertainty” through the light of JM Keynes, and a brilliant lamp he has been. Yet in the wider economic literature, “uncertainty” is often credited to two different economists working towards different ends, who discovered and formalized it independently at about the same time. Today, we are going to talk about the other of these economists: Frank H. Knight.

Introduction: Keynes and Knight, Newton and Leibniz

Why bother with Frank Knight? Like Keynes’s Treatise On Probability (TOP), Knight’s Risk, Uncertainty And Profit (RUP) is surely a “work of genius, “but there are plenty of those”. The value of TOP - at least to the goal of the linked piece - was its unique place in the history of the sciences of unknowledge, as it reintroduced the optimistic perspective from which there was something serious that could be done about the vast problem of unknowledge.

Both TOP and RUP present perspectives on how to conduct oneself both in and out of the face of unknowledge. But whereas Keynes attempts to build a universal logic capable of comprehending the whole of possible experiences of unknowledge even if this creates a very complex system, Knight tries to build up multiple very clear perspectives, like little δαιμόνῐᾰ warning us from evil. The hope is that clarity of the perspectives help guide the reader through long deductive chains. In other words, Keynes tries to build a convex mirror which tries to contain all the world’s complexity one-to-one but Knight tries to build a telescope to see from one perspective into the farthest space. In short, Keynes is a logician and Knight a modeler.

This metaphor is quite apt as Knight scholar Ross Emmett finds a lot of parallels between Knight’s methodology and Rorty’s rejection of The Mirror Of Nature. Emmett explains that Knight sees the economist's point of view - to name just one point of view among others - not as a “mirror of nature” (in Rorty’s phrase) but something like what the lens was for Newton. It will do the experimenter no good to pretend his lens is achromatic when it is not. Quite the opposite, it is the study of chromatic aberration which led Newton to his deep understanding of light itself. This approach of seeing theory as a perspective which can be switched in and out of seems to have been lost in the post-Samuelson tradition - though Samuelson himself was an ardent perspectivist w.r.t. theory.

One can think of the economist not as a man, but as a little angel on the shoulder trying to homogenize and unify the social vision created by the life of the reader. In Ross Emmett’s words, Knight’s goal is therapeutic, to free the reader of the stresses created by the attempts at distinctions and the weaknesses of ordinary language.

Take for instance, the explanation of “declining marginal returns of food” by talk “of boys eating successive oranges or other ‘dinner-table’ illustrations as is so commonly done.” (RUP, Part 2, Chapter 3 ‘Theory Of Choice And Exchange’). The claim is that these declining marginal returns explain how many oranges are consumed: after enough of them, some alternative - even the alternative of ‘nothing’ - starts to look good. The obvious question is that this doesn’t explain why buying bags of oranges at once has declining marginal utility, much less what factors determine the scale of the fresh orange wholesale market.

The problem highlighted by the above example - Knight would say “the fallacy discovered” - lies in the fact that ordinary language does not distinguish between a function and a fluent or between differential and a fluxion. The latter choices satisfy the Newtonian vision of science as the theory of objects moving through spacetime and the former the Leibnizian vision of science as the total theory of relations. As Knight said, that “water seeks its level” is a truth which has nothing to do with the body of water ever finding a level. He even goes so far to say that the exigencies of natural language make it a: 

“practical necessity to describe the action of any force by stating the final condition which it tends to bring about, the conditions under which it would cease to work. Any other description is partial and arbitrarily so.”

Frank Knight, ‘Costs Of Production In Long And Short Periods’. 

Returning to declining marginal utility, Knight explains the “serious error resulting from this [diachronic] method is that it gives the impression that there is a difference between the utilities of different portions of supply”, as opposed to the homogeneity required by the economist’s point of view. If the declining marginal utility is a consequence of the fact that one eventually gets full while eating oranges, then the declining marginal utility metastasizes to the oranges themselves, and they no longer work as a homogeneous kind.

Leaving RUP for a moment, we can look to an older Knight for a moment of clarity. The distinction between fluents and functions above comes to a head in the Knight/Hayek capital controversy. Hayek misinterprets Knight’s vision of capital as a permanent fund as a fluent rather than a function - as unchanging rather than timeless. Meanwhile, Knight shellacked Hayek’s fluxional theory of the interest rate for, among other things, being unable to relate its fictional time periods (such as the so-called “period of production”) to historical time.

Opening RUP again, we see will find this vision:

“As Carver has observed, a (human) ditch-digger is economically as closely akin to a steam shovel as he is to a bookkeeper. Indeed, the possibility of a competitive organization of society depends on the fact of varying proportions, that no particular agency is indispensable, but that within limits they may be substituted for each other and therefore each must compete with others of different kinds for its place. … It follows at once that … no classification or measurement of productive services on the basis of their contributions has any meaning for the distribution problem. According to such a standard they all form one vast homogeneous fund.”

Frank Knight, Risk, Uncertainty And Profit, Chapter 4 (“Joint Production And Capitalization”)

The key words in this passage are ‘economically’ and ‘homogeneous’. Homogeneity arises from the principle of continuity, but it is only homogeneity from the economic point of view.

So we have finally opened RUP to see that it emphasizes a perspective of continuity in order to create a system of economics. Our reading will proceed with a similar systematic approach. RUP has three labeled parts. Contemporary reviews noted the intuitiveness of this structure (actually due to a late revision!), and so this review will follow the three part structure as well. Those three parts are: 1) a retrospective introduction, 2) a review of so-called “perfect competition” and 3) a proposed theory of “imperfect competition through risk and uncertainty”. 

What Is Economics?

But before all that, it is our due to have a bit of explanation, an accounting of whys and wherefores. Knight says - one struggles to not say “brags” - that RUP is a work of pure theory, standing athwart the “practicalist” tendency in American thought. Yet no one, least of all Knight, would deny that the sole appeal of economic theory is to empower one to better organize society - or at least, to avoid disorganizing society. The apology for this book is that the axioms and structures of economic theory are designed to accommodate a simple insight which has immense implications, which must be tediously worked out.

That insight - if it is an insight - is this: “Costs are psychic, in the sense that they are the affective result of inactions one must take when taking action.”. The trick is then to, as Keynes phrased it, to extend the “the notion of the margin … beyond [the psychological] to the conditions of any economic factor which can be regarded as capable of small variations around a given value.” (Keynes, ‘Alfred Marshall, 1842-1924’, section VI). 

There is no individualism implied in this “insight”, though one may of course take individualism as an extra constraint. In fact, in one of Knight’s most praised articles, Knight specifically emphasizes the role of social cost apart from entrepreneurial cost. The point of view of society is as scientifically valid as any and more economically relevant than most.

Still, a couple examples are useful for exposition, being easier to empathize with. A restaurant manager who washes her own dishes is paying an immense cost, despite not actually handing any cash over to someone working as a dishwasher. The cost here is–hey wait, if they’re washing dishes…who’s managing the restaurant?

For a more extreme case, consider overtime pay in labor market monopsony. Comparing this market equilibrium to labor market monopsony without overtime pay, we see that the labor market with overtime pay will have higher profits than that without (see Joan Robinson’s Essays In The Theory Of Employment, Part 1 ‘Full Employment’, Section 3). Only from a strict accounting point of view does this example of the so-called “paradox of costs” seem contradictory, as the ‘insight’ tells us that how costs are borne have nothing to do with the direction of cash flows.

Thus we see how Knight’s ‘insight’ about costs directly divorces economics from accounting. Perhaps this is why economists always seem dazed when they discover the idea that accounting could matter to macroeconomics. 

From this insight, we can work out much of what went on to become economics: if costs are psychic, then income must also be psychic if their difference - profit - is meaningful. This is Irving Fisher. A person who is hired for a new job experiences a psychic change long before the first paycheck. Getting the paycheck also produces a psychic change, but that’s one kind of change among the infinite constantly occurring. Thus it doesn’t matter in principle what the time structure between payments and services rendered is, as long as it distributes risks and uncertainties among the parties.

All the above considerations divorce price theory economics from questions of microstructure and finance. In the hands of a master such as the young Joan Robinson above this manner of thinking can deliver startling insights. But in the extreme, this kind of economics can become divorced from any kind of actual market considerations at all.

So we’ve learned economics is not accounting and not finance. So what is economics? We know why to study economics: to better organize society in the aspect of want-satisfaction. We know that simple doctrines such as Econ 101ism, Bastard Keynesianism and Vulgar Marxism fail to do so because of - to just name one thing - their inability to deal with unknowledge. But in the attempt to improve on them the question of “Why” becomes secondary compared to “How”, as in “How does economics help us better organize society?”. Thus, We have come to the time to see how Knight attempts this in the first part of Risk, Uncertainty And Profit.

Part 1: The Background

In the doctoral thesis that eventually became RUP, the introduction section had but two citations: to a Hegelian philosopher and to Walter Bagehot. These cursory citations speak to Knight’s background in the German social science literature rather than what would become his own method. In principle, the book is organized according to a semi-Millian method of successive approximations, with ‘Perfect Foresight’ as a first approximation, ‘Risk’ as a second, etc.. In practice, RUP as a whole much more closely resembles Max Weber’s Ideal type approach, as noted by Knight in one of his retrospective introductions.

Coming back to part 1 itself, I believe the feeling of inadequacy about this introductory part is due to not getting at the true historical and logical sources of the economic theory to be expounded, to the point that it doesn’t make a coherent argument by itself despite its ober dicta. If I may adopt some pretense, then I will attempt to provide the argument which should be there, though in CVAR’s style and not Knight’s. 

Il y a deux labyrinthes fameux où notre raison s’égare bien souvent : l’un regarde la grande question du libre et du nécessaire, surtout dans la production et dans l’origine du mal ; l’autre consiste dans la discussion de la continuité et des indivisibles qui en paraissent les éléments, et où doit entrer la considération de l’infini.

Leibniz, Essais de Théodicée, Préface

The center of Knight’s economic theory is the old Leibnizian problem of the duel between freedom of the will and continuity. I think the issues will be worth tracing forward from Leibniz to Knight.

The neusis by which Leibniz trisected the angle between freedom and continuity was modal metaphysics. For Leibniz the pure will was a unity, a monad, which contained within it all (and only!) the monad’s possible relations with situations. There is an exact analogy with the infamous Economic Man, indeed this is his entry into the literature. Other wills are presented to this unitary actor only as situations - “Les Monades n’ont point de fenêtres, par lesquelles quelque chose y puisse entrer ou sortir.”, he says in La Monadologie. Frank Knight himself phrases the psychology of the monad perfectly:

“The Economic Man … treats other human beings as if they were slot machines.”

Frank Knight, Ethics And Economic Reform

Yet there are positive and also normative problems with Leibniz’s beautiful vision. The positive difficulty comes first: whither education? As Rousseau taught us, actual human beings become rather than be. The reconciliation of Rousseau and Leibniz became the question of the Enlightenment - and perhaps it still is. To this Kant added the obvious normative problem: the actor ought to treat the other players as people, not as props. Leibniz seems to reduce the humane life to a sort of self-regard (much like Plato and Jesus had before him).

Hegel for one was unsatisfied by Leibniz’s resolution. About the undivided substance of the monad Hegel said “Diese ungeteilte Substanz der absoluten Freiheit erhebt sich auf den Thron der Welt, ohne daß irgendeine Macht ihr Widerstand zu leisten vermöchte.”. But the result of this self-enthronement was not power, but alienation and stress: “der sich entfremdete Geist, auf die Spitze seines Gegensatzes getrieben, in welchem das reine Wollen und das rein Wollende noch unterschieden sind, setzt ihn zur durchsichtigen Form herab, und findet darin sich selbst.”. Hegel believed he had resolved this dilemma by means of a new logic which could at least state the reciprocal dependence of freedom and determination. This resolution has two planks: 1) Each unity is formed continuously in space-time & 2) “Freedom of choice” is nothing more than the assertion that the state of the will as a created object is a sufficient statistic for the choice. The question of “ultimate causes” thus recedes into a merely historical one. Perhaps in the early 19th century, this proposal really felt like the breakthrough. Today, one realizes that the notion of the cyclical dependence of freedom and formation is closer to a statement of the problem than a solution. This Hegelian circle cannot be squared even with the use of a neusis. If Minerva’s owl flies only at midnight, then we can’t wait for her for lunch. And so, while Leibniz & Hegel’s arguments for purely metaphysical or logical solutions are worth appreciating, the starting point for an effective social science is in the ability to mix different types of facts: logical, metaphysical, empirical etc..

Returning to economics will reveal the problem more sharply. Thorstein Veblen, Knight’s grandteacher, is the perfect case in point. Veblen’s account of the reciprocal causation between production and consumption is especially clear in his Theory Of Business Enterprise. Chapter 4 on “Business Principles" is a statement of the problem in its economic aspect which can hardly be improved (and certainly not by me). Chapter 5 on “The Use Of Loan Credit” illustrates a specific version of the problem piquantly: entrepreneurs take on loans to buy enough capital to get the “reasonable” rate of profit and the “reasonable” rate of profit is that which employs all currently available capital. Spinning out a non-circular logic in this loopy causal situation is quite a challenge. 

We have thus arrived on Frank Knight’s doorstep.  Now it comes time to step back into RUP. The trick Knight has for knocking out the Hegelian circle is a one-two combination. The first step is to find idealized special cases of the situation which, for all their difficulty, can be solved. In the case of business enterprise, the special case would be one in which the feedback loop is entirely stablizing. Each round, investors set the reasonable rate of profit by their loan decisions. This is followed by entrepreneurs taking on loans to pay out dividends. In the case of “pure competition”, any attempt by a phalanx of entrepreneurs to reduce the degree of indebtedness (or dually, an attempt by a phalanx of investors to raise the rate of profit) merely creates disappointment as rivals come in to fill the gap. 

This leaves a question: why do the agents participate in the process at all? “There does seem to be a certain Hegelian self-contradiction in the idea of theoretically perfect competition after all.”, as Knight says. Here we shift from logic to metaphysics. The case of perfect competition is not actual, but it is possible. As an economist, one analyzes this case and appeals to a principle of continuity to attempt to work one’s way from this artificial stability back to the real world.

“The heart of this subject, again, from the economic standpoint, is … in continuity and change. A little reflection on these lines should make it less easy on the one hand to teach ‘naive economism’…”

Frank Knight, Risk, Uncertainty And Profit, Preface To The Re-Issue (1933)

In the end, the only justification for this strange mix of logic and metaphysics is the only justification for the study of general economics: the “belief in the possibility of improving the quality of human life through changes in the form of organization of want-satisfying activity.”. But in order to do this we need to move past this introductory section to the actual analysis.

Part 2: Perfect Competition 

Rien ne se fait tout d’un coup, et c’est une de mes grandes maximes et des plus vérifiées, que la nature ne fait jamais des sauts: ce que j’appelais la loi de la continuité, lorsque j’en parlais dans les premières Nouvelles de la république des lettres ; et l’usage de cette loi est très considérable dans la physique.

Leibniz, Nouveaux Essais sur l'entendement humain, Préface (emphasis added)

Arriving at the second part we encounter the viewpoint of the economist. To oversimplify, Knight understands the economist as a theorist who must analyze the special case and discover its hidden continuities with other cases. Neither of these perspectives is “correct” per se, though of course most who take one or the other regard their opposites as fools.

A simple geometric example may help the reader understand Knight’s approach. Imagine an isosceles right triangle, where the right angle is a bright red point while the hypotenuse opposite it is colored with a hundred different blues of various shades. In between these two extremes, the colors blend smoothly from the single red, through an infinite number of purples, to the hundred shades of blue. 

Economic theory for Knight works the way a theory of “purplization” would here: the red corner represents perfect competition, while the blues and purples represent different, more “realistic,” economic structures. In order to extend his analysis out from this bright corner of perfect competition, Knight relies on the principle of continuity.  

The advantages of expanding around the red corner are metaphysical & practical rather than empirical: the red corner/perfect market is the most uniquely determined point in the entire figure. It satisfies the most algebraic relations, and so represents a common starting point for paths to ultimately different shades of blue. Yet emphasizing the single red point invites an immediate objection: almost none of the triangle is red and markets are almost never perfect! When the economist responds  that purplization is actually the general tendency, he rightly is regarded as a loon even though he said truly. 

But there is a more important aspect of this analogy, one which explains Knight’s theory nicely. The statement “red tends to purple” can be read in two ways: as a geometric statement about the value of the differential of the color-valued function or as a dynamic situation wherein one sees red fade into purple over time. It’s easy to turn a geometric relation into a dynamic one. All you have to do is move your eyes from one side of the triangle to the other. But by the same token, how one looks the triangle over has no special relationship with the differentials. The eye may “start at purple and tend toward the red” without contradicting the economist. From the extreme continuity assumptions of the pure economist, refusing to recognize any true heterogeneity, the economic problem becomes capable of treatment by geometry.

The first result of this geometry is the famous theory of Supply & Demand. This part of Knight is mostly notable for the careful attention paid to difficulties in the theory: peeling out assumptions in the strictly economic sense (no bargaining between buyer & seller), warding off misinterpretations (especially naive moral/political connections) and working out arcane deductions. The basic idea is simply that ‘competition’ is when buyers can substitute between sellers and ‘perfect competition’ is when they are indifferent to the seller entirely. The result is “a sort of Emersonian principle of Compensation applicable to every item; each is worth what it costs, but also costs what it is worth.” (RUP, Part 2 ‘Perfect Competition’, Chapter 3 ‘Theory Of Choice & Exchange’). This again is not due to any blessed quality of markets, but because – from the homogenizing perspective of the pure economist – there is nothing for cost and worth to be but homogeneous. Two simples.

The next result, then novel and now thoroughly absorbed, is the famous law of variable proportions, which is to say the examination of the effects of possible substitution on output. Again, the primary draw of this section is the extended treatment of difficulties. In passing over this quickly, I don’t mean to imply that there is nothing to disagree with here. Rather, the issues such as they are seem to be more in detail and fine point.

As noted above, Knight scholar Ross Emmett called these sections - and indeed the whole book - ‘therapeutic’ in the sense of Richard Rorty. Knight’s goal is not to make you “think like an economist”, but to make sure that, of the angels on your shoulder, at least one is an economist. He just wants to make sure someone could understand that there are relations between price and quantity without making it necessary to replace their brains with price-quantity axes and Marshall’s Scissors.

Coming back to specifics, the final chapter of this part concerns “change with uncertainty absent”. It begins with a literal list of some forces which may alter an economy and ends with Knight expounding the so-called ‘eclectic’ theory of interest (as opposed to the pure productivity or pure time preference theories). Despite the allowances for eclecticism, this theory explains basically nothing about interest rates. If you can imagine doing interest rates without expectations at all, that is basically the state of early 20th century macro. In the language of “purplization” we developed above,, interest rates are one (relative) price which are so far from the red angle that the theory of purplization explains nothing–it’s all swirling through different shades of blue.

And if we cannot start at the corner of perfect competition, then where can we start?

Part 3: Imperfect Competition Through Risk And Uncertainty 

Generaliter si diversos eventus utiles disjunctim habere possit negotium, spei aestimatio erit summa utilitarium possibilium ex omnibus eventibus collectarwn, divisa per numerum eventuum.

Leibniz, De Incerti Aestimatione

Before we move on, let’s review the theory of perfect competition. Put simply, the theory of perfect competition is when goods are perfect substitutes for themselves and reasonably good substitutes for one another. For our purposes, there are two issues with perfect competition. 

The first is what Knight called the “Hegelian contradiction” of perfect competition. If all goods are perfectly homogeneous, there’s not really a reason to have a system of property in the first place - or to zoom in, no reason for a class of arbitrageurs if there is no arbitrage - and yet the property system is where the forces holding perfect competition together are meant to originate.

The second is the absence of unknowledge. The raison d’etre of perfect competition is to bootstrap a starting point for a rigorous understanding of relations on differentials of market figures which will hold in the bulk of possible markets. But the homogeneity assumptions involved in constructing the models can’t help but imply something like perfect foresight, which sets the differentials of all expectations to zero. “Expectations never change” is less an insight into markets than an explanation of the limits of perfect competition models. 

So if the homogenizing view of the economist can’t tell us about these markets, who does Knight think can? The Entrepreneur. The name suggests “the one who undertakes”, but as Knight notes this is only a useful definition in the early stages of capitalism, when the social dynamic was dominated by the rise of owner/operators. No reader of Veblen - and certainly not Knight - can help but take into account the decline of the owner/operator. But then who is the entrepreneur?

For Knight, the entrepreneur and the economist are simply two points of view rather than two different kinds of people. The economist seeks explanatory continuities while it is the entrepreneur who discovers the opportunity to create glaring discontinuities that meaningfully change the economic situation. Knight’s goal in part 2 was not to make an economist of the reader, but rather to gift them an economist for use as an angel on the shoulder. Likewise, in part 3, his goal is not to make you an entrepreneur but to gift you an entrepreneurial mindset as a devil on your shoulder.

In order to get that lil’ devil, we must return to the basic problem of psychology in a market society anew, and so Knight does. This chapter is a review of early 20th century philosophy and psychology at their best, and is probably the single most fair review that Knight ever wrote. In sum, Knight declares himself a Jamesian: psychology is the study of a mind made of nerves, the truth of associationism is vitiated by the irrational nature of actual associations and the stream of consciousness is a trickle cutting through the vast continent of mind. There is little to disagree with in this summary, except for a few oversimplifications and one glaring lapse only clear in retrospect. Any issue here could be solved with a few sentences of clarification.

But what cannot be solved with a few sentences is the total absence of liquidity as a psychological factor. Even a decade later, Knight would not incorporate liquidity preference as a fundamental psychological factor. In “Professor Hayek And The Theory Of Investment”, Knight says: 

“In this connection it is, of course, fundamental that in the modern world ‘cash’ itself arises largely out of debts whose power to perform this function depends on their own liquidity in terms of ‘money’ in a narrower sense.”. 

To explain cash in terms of liquidity is correct, but to explain liquidity in terms of money creates a completely fatal circularity. I hate to engage in armchair psychology, but this statement could only be made by someone fettered by the gold standard: banknotes, bank money and bills of exchange were ‘explained’ by gold - “‘money’ in a narrower sense.”. The scarequote explanation of the value of a bill of exchange denominated in florins as a special chemical property of florins is a now-obvious - as Knight would phrase it - fallacy.

To come back to RUP specifically, after a great deal of very interesting philosophizing, Knight separates the problem of unknowledge into three categories: 1) A Priori Probability, 2) Statistics and 3) Estimation.

Of these, the first two fields of thought dominate theory because they can be exact. Their main assumption is - as it always is for the pure theorist - homogeneity. If a die were known to be perfectly cubical, and tossed by a method known to be chaotic, then throwing it a couple thousand times to see the frequencies converge would be as perverse - and as empirical - as adding one to itself a few thousand times to see if you keep getting two. 

The primary philosophical assumption Knight emphasizes in the case of statistics is the partial homogeneity of the population which allows for the concept of the sample. Though human beings, stags in captivity and 1kg balls of uranium all have “lifetimes”, one learns little by the means of a random sample of them. In both cases the Leibnizian continuity principles involved can be strong enough to allow treatment with mathematical exactness.

Despite the dominance of the other options in the realm of letters, estimation is the category of probability that is of most interest when thinking through the problem of unknowledge. This is for the simple reason that estimation is the sole method of approaching unknowledge which can be wrong. Facts about a priori probability and statistics are always about relations between quantities, rather than about an actual quantity present. The uniqueness of estimation arises from the fact that most practical estimations are made without any clear population to reference and the fundamental assumption that we are –at least partially – in error. 

There is a final and decisive difference rooted in psychology, especially from a Jamesian point of view: 

“the possibility of forming a class of similar instances on entirely different grounds. That is, instead of taking the decisions of other men in situations more or less similar objectively, we may take decisions of the same man in all sorts of situations. It is indisputable that this procedure is followed in fact to a very large extent and that an astounding number of decisions actually rest upon such a probability judgment, though it cannot be placed in the form of a definite statistical determination.”

RUP, Part 3, Chapter VII ‘The Meaning Of Risk And Uncertainty’. 

From here we move from the individualistic science of psychology to social science. Let’s start with an example, say, a coffee shop that happens to be next to a haberdashery. Nobody would deny that the marginal consumer of hats and coffee is indifferent, at the perfect competition price, between a hat at the coffee shop and a coffee at the hat shop. But when you flip the situation and say that the haberdashery and coffee shop are perfectly indifferent between who works at their shops, you have moved into absurdity. Even the more detailed claim that they are indifferent between the marginal hour of labor service at the prevailing wage doesn’t really hold up.

The main reason is simple: laborers that work at a coffee shop repeatedly have built up a wealth of similar-but-not-exactly-homogeneous experiences that allow them to make better decisions when providing labor services in a coffee shop. This is now called “learning by doing”. Both Keynes and Adam Smith include “the acquired and useful abilities of all the inhabitants or members of the society” … as a “capital fixed and realized” (Adam Smith, Wealth Of Nations, Book II ‘On the Nature, Accumulation, and Employment of Stock’, Chapter I ‘On The Division Of Stock’). Veblen calls this contribution the “joint-stock of knowledge” (Thorstein Veblen, The Engineers And The Price System, ‘The Industrial System And The Captains Of Industry’).

Perhaps this all seems a bit abstract. To see how ‘learning by doing’ creates uncertainties we can return to the case of overtime rates. It is a matter of common observation that we see managers run ragged to keep the stock of contracted laborers working less than overtime length so as to avoid hiring a new laborer to work one hour. You can ask them why and get the obvious answer: the same pay for an hour of work from an experienced laborer gets the firm a labor hour and some maintenance of the firm’s fixed capital stock.

So we see the fundamental relationship between capital and labor is rooted in the fact that laborers are human beings, “and Franklin therefore defines man as a tool-making animal.” (Karl Marx, Capital, Volume 1, Part 3 ‘The Production Of Absolute Surplus Value’, Chapter 7 ‘The Labour-Process and the Process of Producing Surplus-Value’, Section 1 ‘The Labour-Process Or The Production Of Use Values’). Or as Knight puts the issue, “It will be evident on reflection that even the coarsest and most mechanical labor involves in some sense meeting uncertainty dealing with contingencies which cannot be exactly foreseen. It seems to be the function of all conscious life to deal with ‘new situations’.” (RUP, Part 3, Chapter X continued, ‘Enterprise And Profit: The Salaried Manager’). It is this ability to make mental and physical tools to deal with uncertainty which distinguishes the human ditch digger and the steam shovel from an economic perspective.

This is only one example - an important one! - of how social institutions are shaped by unknowledge and in particular by the need to manage by expectation. Knight here famously distinguishes between two extreme cases of expectations

1. Risky expectations (pure risk)

   1. Dominated by the first two distinctions between expectation and statistics.

2. Uncertain expectations (pure uncertainty)

   1. Dominated by the third distinction between expectation and statistics.

According to Ross Emmett, in the original thesis from which RUP arose (and much of the later literature) the distinction between risk and uncertainty was between insurable and uninsurable unknowledge. Both the original and RUP definitions have advantages and disadvantages. The advantage of the original definition is that it focuses one on markets as institutions that deal with unknowledge both successfully and unsuccessfully. Eventually it led to Ken Arrow’s model of complete markets as those which are purely risky. Contrariwise, the advantage of the RUP definition is that it better mirrors the psychology of the entrepreneur as entrepreneur, making distinctions between labor hours from different laborers etc. which seem arbitrary from a market point of view. I think there are situations where each definition is appropriate.

With risk & uncertainty finally on the scene, we come to profit. The need for stable expectations creates a tendency towards centralization of control, which Knight calls ‘cephalization’. Keep in mind that this tendency is a differential in the space of possibilia and not a fluxion that actually happens as the change in situations over time. Anyway, this tendency creates space for two kinds of economic management: contractual management and direct management. Thus net income is paid out in two shares: rent for contractual management and the profit residual for direct management. In practice, rents and profits in this sense are not perfectly distinguishable. Bond companies largely manage by contract, but can also have their stooges monitor production directly. 

Despite this entanglement, there is still some use in looking at the pure entrepreneur’s point of view, that is, the point of view of pure profit income. Because profit is the residual here – what remains after all of the contracts have settled – the entrepreneur bears the brunt of unknowledge. More realistically, they manage uncertainty, rather than “bear it”: after all, contracted laborers are often uncertain as to how many labor hours they will work and suffer costs if they estimate the number of hours wrong. Further, if we take into account contract renegotiation, a landlord of commercial property who raises rents so high that all the businesses on his land close has certainly born a cost from a missed estimation of his client’s ability to pay. These considerations reinforce Knight’s methodological plank that entrepreneur is a point of view and not a class of people. As Knight explains, the difference between “employer and employee status depends normally on the possession of a minimum amount of capital.” . (RUP, Part 3, Chapter IX ‘The Meaning Of Risk And Uncertainty’).

Knight finally moves from these highly general considerations to those concerning modern society: in particular, the corporation and economic growth & social progress. The corporation is a social form that concentrates control while diffusing ownership. One immediately recalls Veblen’s already linked Engineers And The Price System or the work of Berle and Means decades later, but Knight’s is all his own. There is plenty to disagree with here - such as “the disposition of laboring people to gamble recklessly with life and limb as well as income.” - but also several nice insights.

Knight’s analysis of growth - which is intimately connected to interest rates - is subject to the same issues about the analytical absence of liquidity preference as a factor that we mentioned above. The main argument of these chapters concerns the Veblenian circle from earlier, the cyclical dependence of entrepreneurial & investor decisions. Knight shows uncertainty is a stabilizing factor. Knowledge of production techniques from the distant past fades out as superior techniques are introduced, which can only be adopted at risk. In contemporary society, this feedback process takes place because the size of profits are residual to the accuracy and adequacy of management.

Markets are one technique for structuring the maintenance of this feedback, and are not the sole process today, historically or in the future. The main advantage of markets is that they are the best understood of currently-known techniques - we have learned truck and trade by doing truck and trade - and this is a powerful advantage. But being the current state of the art has also been the argument for a thousand forgotten industries in the past.

Conclusion

Nearly a decade after RUP hit the presses, the world was plunged into a years-long depression which threw everything into question. In the Soviet Union, the collapse of international demand for Russian grain imploded the New Economic Policy, causing the mini-revolution called the Great Break. The economists of GOSPLAN were among the hardest hit by this change of the guard. In America, the Great Depression wreaked its bleak havoc on every existing institution. America’s economists were more embarrassed than hit by the political changes that came.

Knight with his eclecticism about methods of social control was less embarrassed than some economists. It is true, however, that Knight’s economics used Say’s “Law”: the “law” that because supply creates its own demand, there can be no general depression. His use was free and unconscious. Even in his review of Keynes’ General Theory, where he claims to deny Say’s Law, phrases the problem as “the blocking” of growth towards full employment. He writes there that blockage is the“fundamental mystery” of business cycle theory. If one actually denied Say’s Law, there would be no mystery: one needs no blockage in a direction where there is no force.

That Knight holds basically the same position almost two decades after RUP shows the degree to which Knight stayed more or less the same person. Perhaps it’s not so surprising when one realizes he was middle aged when RUP, his first book, came out. The main addition of time to Knight’s intellectual life was tonal rather than structural. Knight was an old liberal of the old type, a type often admired - though rarely agreed with - by Marx. In his own time, Knight couldn’t help but feel depressed that the era of Geheimrat Goethe had passed to the era of Minister Beria. Knight and Keynes never gave up what Keynes called the ‘religion’ of their early beliefs. The difference between them can be expressed by Keynes being more on the side of the varieties of liberal experience rather than Knight’s side of the sources of liberal insight.

Perhaps a story will help drive things home. A Soviet censor once chided Bulgakov for missing the “progressive elements” when translating Goethe. Bulgakov wryly, but privately, responded that the censor couldn’t tell him what those elements were. Both sides have a point here, though only Bulgakov consciously. If I may conclude by risking prophecy - and how else could a review of Risk, Uncertainty And Profit end! - then I would first venture that the era of Berias is over, but the age of Goethe’s dreams has not come. If we are to make real those mysterious but powerful progressive elements, then perhaps we need to listen to people like Maynard Keynes, Frank Knight and others.