Hülsmann on the Pioneers of Marginalism

In his biography of Mises, Hülsmann takes a quick detour to discuss the foundation of the Austrian school and some of its early members which influenced Mises’ development as an economist. I’m having trouble, however, following some of Hülsmann’s distinctions between Carl Menger, W.S. Jevons, and León Walras. Allow me to excerpt from the book at length,

Mises (Hülsmann)Gossen had indeed anticipated Jevons’s and Walras’s theories. The three men had developed general theories that were analogous to Menger’s general theory of value and prices, but differed from it in their psychological orientation and in the exact type of explanation they offered.

In Menger’s theory, the term “value” does not refer to a psychological feeling, but rather to the relative importance for an individual of the marginal unit of a good X — that is, to the importance of X in comparison to the marginal units of other goods Y and Z. The market price of a good results from the interplay of sellers and buyers, for whom the goods bought and sold have different relative importance. In contrast, in the theories of the other three authors, the price of a good results from the interplay of sellers and buyers whose feelings or well-being are differently affected by control of the good. While Menger explained the pricing process as resulting from the importance of a good relative to the importance of other goods, Gossen, Jevons, and Walras explained the pricing process as the impact of a marginal quantity of a good on the psychology of the actor — an impact they called want-satisfaction (Gossen), utility (Jevons), and satisfied needs (Walras). Jevons’s marginal utility thus played structurally the same role that marginal value played in Menger’s theory — it delivered an explanation of market prices — but where marginal utility explains the price of a good by how the good ranks in importance compared to other goods, according to the needs of the individuals involved in the pricing process.

In the psychological approach of Gossen, Jevons, and Walras, the human psyche was the great common denominator for the economic significance of all goods; in the theory of Menger there was no common denominator. In his approach, “value” cannot be independent of the specific circumstances of time and space; it is inseparable from these circumstances and means different things in different economic settings. According to Gossen, Jevons, and Walras, the amount of “utility” derived from a good could be different in different situations, but according to Menger, the entire basis of value is different as soon as the economic context changes — because the good would then be compared to different other goods.

— Jörg Guido Hülsmann, Mises: The Last Knight of Liberalism (Auburn: Ludwig von Mises Institute, 2007), pp. 131–132.

I don’t have a strong hold on the primary literature; the only relevant book I’ve read is Menger’s Principles of Economics, but I read it some time ago. But, Hülsmann’s argument above is confusing. I don’t remember Menger’s theory of value relying on a concept of “relative importance,” but on a number of prerequisites which boil down to the individual having established a subjective relationship with an object, where the latter can satisfy (or help satisfy) some end the individual has in mind. The “relative importance” of goods, after all, can only be established if each good has an independent value attached to it (and, actually, this idea of “relative importance” seems close to the theory of the marginal rate of substitution). What I thought the major differences between Menger and Walras were was the latter assumed a certain set of prices to exist — thus the introduction of the abstracting proxy of the auctioneer —, whereas the former developed a causal theory of price formation. (I’m not sure as to the differences between Menger and Walras, but if this facet of Mises’ thought has any relation to Menger’s, then another difference would be emphasis on the importance of the psychology that explains human preferences; Mises, and maybe Menger, instead held that we take preferences as given.)

Maybe Hülsmann’s distinctions are too subtle for me; if someone could clarify his argument I’d be much obliged.

9 thoughts on “Hülsmann on the Pioneers of Marginalism

  1. Dan(DD5)

    I think the difference is seemingly subtle but actually quite substantial in terms of its implications on theory. This is why I commented in the other post on how Austrians are subjective vs marginalists. (The literal meaning of the terms here implies nothing about the difference)

    It’s easier to see the difference when thinking about the ordinal vs cardinal debate.

    Menger’s value system is ordinal by definition. It does not allow for value the be cardinal. Value itself results from comparisons between goods, or in other words, value is meaningless absent choice. Choice implies comparisons, comparison implies different values attached to goods based on a grading scale – A is more urgent (valuable) then B, but less urgent then C, etc……

    Warlas’ value system is not ordinal by definition. It theoretically allows for value to be cardinal (not that it necessarily follows that it is, just that it can be consistent with a cardinal value theory). Goods are attached values independent of other goods (Yes, they are subjective based on utility and so on), but the value itself results from the good itself and not the comparison between goods. Once values are attached, man can make comparisons between the goods. Note that in such a system, value of goods can be cardinal.

    I would summarize it as follows:

    Menger: Choice gives rise to value
    Warlas: Value gives rise to choice

  2. Abhinandan Mallick

    Hi Jonathon, these questions have long interested me so forgive me for the long

    I think the point he’s trying to make regarding the concept of “relative importance”
    as important (no pun intended) and distinguishing to Menger’s approach, even if
    Menger does not use exactly the same wording (though I’m assuming we’re talking
    about the translatin of Menger’s Principles). It is perhaps unfortunate that Hulsmann
    is not more explicit in this passage, else he could have avoided confusion on
    the part of readers like yourself.

    I think the relevant point is that with a Mengerian/Misesian approach is you derive diminishing marginal utility as a consequence of preference necessarily demonstrated by action among ends/purposes with differing ordinal rankings of importance (hence the relativity of the concept, since ordinal rankings among a set of objects are unavoidably relative) when the means to achieve these ends are uniformly comparable. The last condition is vital and distinguishing, and properly understanding it allows the avoidance of confusions regarding “exceptions” to the diminishing marginal utility theorem. Mises himself gives a very good example of how this occurs with an example in Human Action (p.128 of the Scholar’s edition). The man more than happily trades his
    raincoat for 10 logs of wood when he has 90 since then he can achieve the most
    valued goal of making a hut with 100 while with only 10, he would not even a higher number to trade in his raincoat (This has interesting similarities to value and price imputation among complementary factors/goods combinable in rigid production ratios, e.g. the value of a missing shoe being conveying the same ordinal importance as the entire pair). He correctly lambasts the mathematical economists of his day for not properly understanding the above condition and thus mistakenly citing such examples as “contradictions” of the diminishing marginal utility theorem (trivially d^2U/dX^2 would no be uniformly < 0).

    I think this is the correct way to think about the distinguishing aspect of Menger's formulation of marginal value theory, and is really the core of what Hulsmann refers to somewhat vaguely in terms of economic context in the passage above. I don't think Rothbard takes the right approach since he seems to skirt examples like the above in MES, by claiming that you always have to use the "right units" in discussing marginal utility always, since analyzing the consequences of actions people make on "not right"
    units of goods is importantly fruitful for our understanding of reality. Indeed consistently carrying this line of reasoning and applying it to the context of commercial valuation and imputation allows one to yield many fresh insights, including deriving praxeologically the law of costs (which I'm gad you've pointed out the importance of on numerous occasions).

    It is these considerations that I think decisively distinguish the Austrian treatment of the subject of diminishing marginal utility. The argument for diminishing lacks any casuistic reference to psychological concerns (still frequently referenced in undergraduate texts). Furthermore, as I'm sure you're aware, diminishing marginal utility, at least as they express it, is not a necessary requirement of neoclassical treatments of price theory. You can have a diminishing marginal rate of substitution without diminishing marginal utility and have diminishing marginal utility without a smoothly diminishing marginal rate of substitution.

    This raises important questions regarding the diminishing marginal rate of substitution condition. You cannot derive this concept beginning with ordinal preferences among ends achievable uniformly comparable means, as the Austrian concept of diminishing marginal utility is formulated. The concept does have an important sense of relativity however as you have intuitively recognised, and is thereby necessary in neoclassical value and price theory in order to allow for relative comparisons among the valuations of goods depending on this MRS. The latter is not achievable with the neoclassical marginal utility since the expression dU/ dX is devoid of reference to relative values. The MRS is altogether not necessary for the Austrians however since their marginal utility concept is an ordinal ranking vis a vis other means (ultimately among ends) to
    begin with, which is already necessarily relative as noted above. Furthermore, the diminishing MRS condition also seems to be more of a mathematical convenience, required to achieve unique tangencies with a flat budget hyperplane (in a PC world with perfectly informed price takers), in the same sense diminishing returns to scale is mathematically convenient and necessary for unique production cost optimisation solutions under perfectly competitive conditions (also with a flat budget hyperplane). Even Hicks seems to have realized the problems regarding the economic interpretation of dMRS, referring in the second edition of Value and Capital to it as a "rabbit out of the the hat" assumption (p.23).

    Overall therefore, while the marginal utility concept of the Austrians is a relative one, or one denoting relative importance, it is quite distinct from neoclassical MRS even though the latter is also a relative concept of sorts, while neoclasssical MU is not. Furthermore I think distinguishes the 2 approaches to the subject as well.

  3. JCatalan

    Abhinandan: Thanks for the response, it’s very helpful. When was ordinal preference ranking (of goods bundles) introduced into microeconomics? I know you spend some time on it, but is there another way you could explain to me the point on comparable means and ends? I know we’re you’re coming from (the basic argument as to why wealth in general doesn’t suffer from diminishing marginal returns), and I think one of the distinctions can be made with the neoclassical concept of a “bundle of goods.” But, I’m having a little trouble meaningfully distinguishing it from “mainstream” micro. Sorry for the likely frustrating response; I’ll have to re-read your comment.

    Dan: Hülsmann does refer to how economists like Gossen thought that utility could, in principle, be measured. But, neoclassical price theory is built on the assumption of ordinal preference sets. As I understand it, even those who thought that one day we could provide exact measurement (in “utils,” for instance) used ordinance preference sets as a proxy, if you will. But, I see what you’re saying near the end of the comment: a good’s value may be contingent on other subjective relationships that an individual can make with related goods.

    1. Dan(DD5)

      I emphasized this point: That Walras marginal theory does not imply cardinal value, but that is also does not imply that it is not. I’m not talking about what Walras himself and other neoclassicists assumed. But this issue of “assumption of ordinal preference sets” is actually another important difference between the two:

      Menger does not assume an ordinal value system. He derives it from Human Action.

      Walras and neoclassical economists, at best, just assumes it.

  4. Beefcake the Mighty

    Perhaps you are right about Walras, although I don’t think so. However, there is not the slightest question that the neoclassicist concept of utility is ordinal, not cardinal. Look into the representation theorems on this point.

  5. Abhinandan Mallick

    I’m not completely sure when that kind of language was introduced into micreconomics (perhaps by Debreu’s time or earlier), but I think since you’ve pointed to the ordinal rankings of bundles, it is worth taking note here that again something different is being discussed. First rankings of sets of bundles are taken, often allowed to be the same as one another (except with counterexamples such as lexicographic preference sets). Again however, these rankings are simply taken with certain requirements imposed on the ordering of bundles by “rationality” axioms and disconnected from the ordering of ends relating an ordering of units means including the marginal unit, and are rather rankings of total combinations of goods. It is true one could try to reason this way, ordinally ranking total bundles of goods owned, but it is not the way that the Austrians derived the diminishing marginal principle (there are of course caveats, since Menger and Bohm Bawerk did treat utility somewhat cardinally, an error Mises cites Franz Cuhel as correcting). Secondly, once you have these rankings of bundles in place you have to convert them into cardinal functions to arrive at marginal utility unavoidably as a difference between adjacent bundles and thus a cardinal concept. This is why MRS is necessary to arrive at a marginal concept to express tradeoffs between ends. It is often described as an ordinal concept, since as is repeated ad nauseum in classrooms it is preserved under monotonic transformations. But this is trivial since the ratio of 2 partial derivatives of a multivariate function is invariant under all functional transformations monotonic or otherwise (aside of where there are singularities). So while monotonic transformations can preserve ordering I don’t see how the MRS is necessarily a monotonic or ordinal concept. I can accept it preserves tangencies along an equipotential surface regardless of how one then subsequently transforms that surface, but this seems moot.

    To clarify the point I made regarding uniform comparability of means among ends; this occurs when ceteris paribus we can say each unit of e.g. n means can be applied potentially among n ends with the value of any unit thus reflecting the relative value of the nth marginal end achievable. There can of course be scenarios where this does not necessarily apply, e.g. if instead of 3 eggs each being able to achieve ends of diminishing importance (e.g. one for yourself to eat, one for your partner and the final for your dog), you could have the 3rd egg allowing the possibility of achieving the attainment of a far more valuable end thus causing the value the actor demonstrates to hold over it to reflect the importance of this much higher valued end (making a spanish omelette perhaps). Rothbard tried to avoid these issues altogether saying that in such a case the relevant marginal unit is 3 eggs. This is all well and good, but I’ve never found that response completely satisfactory, since the same considerations play an analogous role when different complementary goods allow for more higher valued ends to be achieved, but we don’t redefine units of means to be all the complementary means capable of achieving such an end when we encounter these scenarios either. I think rather, while recognising Rothbard’s point, these considerations do not affect the validity of the diminshing marginal utility theorem, but rather affect how we apply it to understand and explain valuation in the world using ceteris paribus conditions.

  6. JCatalan

    Abhinandan, thank you again for taking the time to comment and straighten things out for me. I’ll have to digest your points over some time.

    For all those interested, here is another excerpt from Mises: The Last Knight of Liberalism which seems to somewhat contradict the above quoted excerpt,

    In Menger’s definition of value — which contrasted somewhat with his actual analysis of value — value was a characteristic feature of a single economic good. In contrast, Mises defined the value of one good in explicit context with the value of another good with which it was compared, and he stressed that this “comparison” was based on choice as it involved “acts of valuation.”

    — pp. 388–389.


Leave a Reply

Your email address will not be published. Required fields are marked *