Friday, February 25, 2011

Action Aversion

[As seen on LvMIC]

If I just straight out told you that there is a model within the conventional economic theory that implies most of us would prefer a world in which we would not be acting humans, you would probably tell me that I have a few screws loose. To avoid this potential label, I will develop this argument step by step. First, let us see what it means to be an acting individual.

Do we act?

Humans act. This statement is irrefutably true because the act of refutation itself implies an end – to refute the statement that humans act.[1]

But, suppose that what humans do is not action – employment of means to achieve ends. Then it must be something else. What else could it be? Even though we cannot say this with absolute certainty, we might say that non-living matter does not act.[2] Thus, we might say, if humans do not act, then they could be compared to non-living matter.

However, if we assume that humans don’t act, that simply means that humans do not make choices. They do not substitute less preferred state of nature for a more preferred one by using means available to them. In other words, they do not displace one possible future by acting to achieve another, more preferred future. In that case, humans just follow the inevitable path set out for them outside of their consciousness, just like the rocks and stones around us do (at least in our view).[3] [4]

Consequently, our attempt to refute the action axiom was not actually action – it was an illusion of action. While we can not exclude the possibility that all of us live in an illusion that we actually make choices, when we do not, we do not know whether we do in fact live in such an illusion. As Ludwig von Mises says:
We may or may not believe that the natural sciences will succeed one day in explaining the production of definite ideas, judgments of value, and actions in the same way in which they explain the production of a chemical compound as the necessary and unavoidable outcome of a certain combination of elements.[5]
Thus, it is possible that what we perceive as our choice may be just an inevitable outcome of a chemical reaction, which was in turn an inevitable outcome of some other physical process. However, accepting this as true would be pure speculation. What we do know is that our self-observation does indicate that we do indeed make choices – that we act.

The World of Action

Human action is impossible in the world of complete information. In this world everyone knows everything there is to know, including the future. Thus, one cannot act if he or she already knows his or her future. There is no space for choice in these circumstances.

It is hard to make definite statements about preferences of all human beings but we could safely state that many humans would be terrified to live in a world of complete information. In other words, many of us, if not all, would prefer a world in which we can make choices, a world in which we can use means to change our life according to our desires, a world in which we can act. Thus, it seems, at least given the general sentiment of human beings, that they prefer a world in which action is not an illusion, a world in which we are indeed different from stones that roll down a hill governed by nothing more but the law of gravity.

Now we come to an important point: If humans really prefer to be humans in the way we perceive ourselves – as acting individuals, then the probabilistic concept of risk-aversion is contrary to the nature of human preferences toward action.

The World of Probabilistic Risk-Aversion

The probabilistic concept of risk can be represented by a situation when an individual (let us call him Jim) faces different future paths that offer different payoff. It is assumed he knows the probability distributions (or probability density functions) of these future events.[6] What he does not know though is which path of all the possible paths the future would bring.

This situation is often illustrated using flip of a coin. Suppose Jim is offered a game in which a flip of a coin determines the amount of money he will receive. If the coin turns heads, Jim receives, say $100. If, however, the coin shows tails, the he receives only $50. Alternatively, he may be offered $75 dollars straight out without flipping any coins.

The next step is to qualify Jim’s risk preferences. If he chooses to flip the coin, he would be qualified as risk-loving. If, on the other hand, he chooses the certain $75 dollars, he would be qualified as risk-averse.

This risk-aversion qualification is associated with the fact that Jim refused the so-called “fair gamble” where the expected payoff of getting a $50 and a $100 with equal probabilities (1/2) is $75. In other words, the expected outcome of the gamble is $75.[7] Thus, Jim’s choice is interpreted as an aversion towards the risky nature of the gamble since that is the only difference between a certain $75 and a “risky” expected $75.

Whitin the probabilistic framework, this means that if Jim, who is risk-averse, knows the probability distribution of different future events – getting $100 and getting $50, he would prefer a world in which he would get the amount of money half way between $50 and $100 with absolute certainty. If he had an option to choose between these two worlds, he would choose the world of the-middle-of-the-road certainty instead of the world where he could possibly end up in a better situation than the-middle-of-the-road but he could also end up in a worse position. It is said that he would trade off the possibility of potential $25 gain over the average $75 for the absolute certainty of $75.

The bulk of economic literature states that most people would choose the option in which they receive $75 with absolute certainty. In other words, most people are qualified as risk-averse. This means that most people would prefer to get the expected, middle-of-the-road outcome with certainty rather than face an uncertain future where they could get much more but also much less than in the world of absolute certainty.

If we generalize further and assume that Jim knows the probability distributions (or probability density functions) of all future events in his life, using the logic of the probabilistic representation of risk, one could conclude that Jim would prefer a world in which he would live his expected, middle-of-the-road life path with absolute certainty over an uncertain world.

On the surface, this does not seem too strange. Many of us would probably like an “average” life, without too many frustrating perturbations. But, the above conclusion is not about preferences about life in an intrinsically uncertain world. It is a statement about preferences between an intrinsically uncertain world and an intrinsically certain world.

But, if one knows his or her future, where is the logical basis for purposeful human action? Is it really true that most of us would prefer not to act? I would say that this is not true. Rather, the probabilistic concept of risk-aversion does not have the robustness of a general theory. In other words, this concept seems sensible in isolated cases but when applied as a general principle, it contradicts the nature of human existence.

Averting Uncertainty through Action

On the other hand, members of what some call the Austrian school offer a more robust, qualitative[8] understanding of risk. For example, Murray Rothbard makes a clear distinction between gabling and entrepreneurial risk-taking:
Gambling on the throw of the dice and betting on horse races are examples of the deliberate creation by the bettor or gambler of new uncertainties which otherwise would not have existed. The entrepreneur, on the other hand, is not creating uncertainties for the fun of it. On the contrary, he tries to reduce them as much as possible. The uncertainties he confronts are already inherent in the market situation, indeed in the nature of human action; someone must deal with them, and he is the most skilled or willing candidate. In the same way, an operator of a gambling establishment or of a race track is not creating new risks; he is an entrepreneur trying to judge the situation on the market, and neither a gambler nor a bettor.[9]
Thus, an entrepreneur does not “consume” risk for the sake of it. He faces the unavoidable risks that are inherent to the market process in order to earn profit. All of us are, in one way or another, trying to reduce the uncertainty inherent in the world around us by using the means available to us – by acting. In this sense we are action-loving and this action-lovingness comes from our aversion to an uncertain future. We act in different ways to achieve different kinds of more certain future.

For example, this article was written for the purpose of making the future in which people understand the epistemological limitations of the probabilistic concept of risk-aversion more certain. But it would be a mistake to conclude that this means that I would prefer a world in which I would know with absolute certainty that I will write this article and the inevitable future events this would leed to. This is an error that results from extending the probabilistic concept of risk-aversion into the context of human action.

Friday, February 4, 2011

A Fundamental Economic Concept Lost in Translation

[As seen on LvMIC]

The principle of comparative advantage is one of the fundamental insights of economics. In everyday language, this principle teaches us that we can produce more goods in total if each of us specializes in the production of a product he or she is most productive in compared to other people. This productivity is expressed relative to the quantity of all those goods that each individual could have produced but are now foregone because he or she specialized only in one product. The beauty of this concept is that in our world of omnipresent interpersonal differences someone always has a comparative advantage in something. Consequently, division of labour is a natural consequence of interpersonal differences and the law of comparative advantage.

The thesis of the article is that the concept of comparative advantage has lost a fundamental aspect of its meaning in the works of economists who rely on stylized models of “aggregate” economies.  These models present the concept as a national or regional phenomenon rather than individual.

However, according to the logic of human action, only individuals can experience costs and benefits and consequently act upon those costs and benefits. All the production and consumption decisions are performed by individual human beings based on their knowledge of their own production possibilities and preferences.

As Friedrich Hayek noted, time- and place-specific knowledge owned by any individual in the society is not directly available to other human beings, including the economic analyst. Nevertheless, some of this individual knowledge is reflected in market prices through specialization and exchange. While some economists have maintained these principles attached tightly to the concept of comparative advantage, some have marginalized or even completely ignored them through the process of translation into a language that employs aggregate variables.

The individual-specific view of comparative advantage is generally accepted in the “mainstream” microeconomic literature. It is said that individuals may differ in terms of the quality of inputs they own or skills of combining those inputs into final outputs. Consequently, there is a potential for increased productivity leading to mutual benefits if some individuals specialize in some activities while others specialize in some other activities and then exchange the surplus production with each other. This is the well-known law of comparative advantage or, as Mises calls it – the law of association. This law simply states that in the world inhabited by people who differ from each other in a near-infinite number of ways, possibilities for mutual gains from association with other individuals are omnipresent.

However, the traditional macroeconomic literature is built upon the Ricardian climatic and the Heckscher-Ohlin factor-endowment aggregate models where comparative advantage is national and it is based on the difference in technologies or relative quantities of aggregate inputs between countries. For mathematical convenience, it is generally assumed that inputs are homogeneous within a country or a region and that they are assembled into outputs in an identical way throughout the entire country or region. These assumptions are seen as useful abstractions, and not as serious logical flaws.

Even when heterogeneity within an economy is introduced (at the firm level), it exists only as a special wrinkle in the generic Heckscher-Ohlin or the Ricardian model. These modifications are generally intended to explain two-way trade between countries. This is the case when the same product is both imported and exported from the same country. This situation cannot be explained using the traditional models and thus some economists called upon firm heterogeneity within countries as an explanation. However, heterogeneity within an economy is not seen here as a fundamental general requirement for the existence of a market. These models would still produce market prices within each “national” economy through its mathematical mechanics, even if heterogeneity is assumed away.

There is a logical conflict in this reasoning. Market prices are exchange ratios observed in an interpersonal exchange of goods and services. For an exchange to occur, there need to be at least two individuals in an economy. If there was only one individual in a country, he or she could not constitute a market. All the exchanges that this individual would make would be intra-personal exchanges. He might choose to devote more time to the production of clothing and thus give up some of the food he might have produced instead. However, in this situation there is no exchange ratio to be observed – there are no market prices to be compared with the market prices in some other (hypothetically isolated) country.

In the mathematical language, the intra-personal exchange ratios (i.e., the individual trade-offs when deciding between two actions) are the shadow prices or the slope of the individual Production Possibilities Frontier, not market prices. Some economists using the aggregate models of comparative advantage often ignore this important distinction between an individual and a country and treat them as equivalent economic concepts.

In order to introduce a possibility of a market, we need another individual in this country. But, if this individual was identical in all respects[1] to the already present individual, there would be no logical reason why any of them would specialize in the production of one good and exchange some of that product with the other individual. They both own equal quantities of homogeneous inputs and are equally skilful in combining these inputs into any output they may desire (and their desires are identical). In this situation there are no gains either in productivity or in the alignment of individual means and ends resulting from specialization and exchange. There is no particular reason for a market to emerge in an economy inhabited by identical clones who own identical inputs that are assembled into outputs in a precisely identical way.

Therefore, input[2] heterogeneity is not an additional wrinkle in a general market model – it is a necessary condition for the existence of any market. More precisely, input heterogeneity is an unavoidable attribute of any market. Even if all individuals were identical in all respects before specialization, if this specialization will not change them in any way and make them more productive in the selected line of production, there is no particular reason to specialize.[3] Consequently, there are no market prices to be compared between two hypothetically isolated economies.

In this case, if there were two hypothetically isolated “economies” composed of individuals identical within an economy but different across economies, the only thing we could observe is two groups of self-sufficient autistic clones. This is why individual heterogeneity is important. The same laws govern interpersonal exchange, regardless of whether the exchange occurs within a fence we call a country border or across the fence.


This article is a reflection on the nature of misunderstandings among economists, looking at the law of comparative advantage as an example. There are different sub-languages within economics, and the structure of the language we use can be more or less conducive to fully encompassing the meaning of the concepts we wish to explain. Sometimes, the alignment between different languages is quite thin or even nonexistent so that some messages get lost in translation. However, it is important to constantly clarify the subtle differences in the understanding of economic concepts in order to facilitate better communication, and ultimately, discover the messages that are too important to be lost.


  1. It should be noted that, unlike Block et al. (2007) critique of Mises and Rothbard, this article looks at homogeneity in the pure mathematical sense implied by the aggregate models of “national” comparative advantage. Inputs available to any individual within a country are identical in all possible respects. Given the properties of the production function, the basic unit of inputs is infinitesimally small. It is also assumed that specialization does not affect the individual level of skill in either line of production. 
  2. Since human action ultimately enters all production processes, the actor’s physical body with all its physical and mental capabilities is an input as well. 
  3. Preference in production of one good over another as a reason for specialization is also excluded since it is assumed that all individuals have identical preferences.  The only remaining reason for specialization and exchange would be that all individuals prefer exchange for its own sake. But, this would be a trivial explanation because exchange would be its own end.