avatar

Armen Alchian and Alfred Marshall

by Kurt Schuler March 10th, 2013 10:39 pm

 

This post has no connection to free banking, but it is a footnote to the history of economic thought that would otherwise find no outlet.

Armen Alchian recently died at the ripe age of 98. His work showed how much could be accomplished by thinking deeply using basic economic ideas that are taught to college freshman but that even few experienced economists have thoroughly in their bones. Years ago I read Alchian's influential textbook with William Allen, University Economics, and a volume of his collected writings, Economic Forces at Work; the latter has been superseded by a two-volume collected works, which I have not read.

Alchian is known, among other things, for the so-called Alchian-Allen theorem, which first appeared in his textbook. Here it is as Alchian and Allen stated it (1964 edition, page 75):

"Suppose that grapes are grown in California and that it costs 5 cents a pound to ship grapes to New York whether the grapes are "choice" or "standard" ... suppose further that in California the "choice" grapes sell for 10 cents a pound and the standard for 5 cents a pound... If grapes are shipped to New York, the shipping costs will raise the costs of "choice" grapes to 15 cents and of "standard" grapes to 10 cents. In New York the costs of "choice" grapes are lower relative to "standard" grapes (1.5 to 1) than in California (2 to 1)... New Yorkers faced with a lower price of "choice" grapes relative to "standard" will consume relatively more "choice" grapes than Californians will."

What I have never seen mentioned in any of the recent or older papers on the theorem is that there is an anticipation of it in Alfred Marshall's Money, Credit and Commerce, published about 40 years earlier (Book III, chapter 1, section 3):

"Again, if two sorts of the same commodity are of unequal usefulness, but cost the same for carriage, the better sort is likely to be chosen for long-distance transport. Where wood is plentiful, thick beams of quick-growing inferior wood may be preferred to thin beams of better wood, which perhaps cost twice as much: but if both have been imported from afar, the cost of carrying the thick beams may cause preference to be given to thin beams of superior wood....

"Again, the commoner sorts of wine are generally consumed near  home, while the better travel far:...."

P.S. Brief Internet research for this post led me to the useful Deardorffs' Glossary of International Economics, which I had not previously seen.

29 Responses to “Armen Alchian and Alfred Marshall”

  1. avatar Greg Ransom says:

    Note that the Alchian-Allen theorem applies also to 'transporting' goods across time, ie via longer or shorter production periods -- the costs of 'choice' grapes are lower relative to 'standard' grape the longer the period of production, ie the longer the aging process as in wine production, so that you will find the relative price of 'choice' grapes going down relative to 'standard' grapes as the aging process extends, predicting that 'choice' grapes will be used more relative to 'standard' grapes as you move across the time scale of alternative production/aging processes.

    I've named this the Ransom Effect, its the application of Alchian-Allen to Hayek-Bohm-Bawerk, and it has implication for Hayekian trade cycle theory and the Ricardo Effect.

  2. avatar Greg Ransom says:

    "If two sorts of the same commodity are of unequal usefulness, but cost the same for installation, the better sort is likely to be chosen for long-period construction projects. Where wood is plentiful, thick beams of quick-growing inferior wood may be preferred for use as firewood or boxes to thin beams of better wood, which perhaps cost twice as much: but if both could be used for long-period construction, the cost of waiting for the productive rewards from the thick beams may cause preference to be given to thin beams of superior wood."

  3. avatar Greg Ransom says:

    "If two sorts of the same commodity are of unequal usefulness, but cost the same for carrying across time, the better sort is likely to be chosen for long-period production processes. Where grapes are plentiful, easy-growing inferior grapes may be preferred for use making bottom self wines to harder to grow grapes of better varieties, which perhaps cost twice as much: but if both could be used for long-period wine making, the cost of waiting for the productive rewards from the easy-growing inferior grapes may cause preference to be given to the hard-to-grow grapes of superior varieties."

  4. avatar Greg Ransom says:

    "The commoner sorts of grapes are generally consumed immediately or with little time expended in aging, while the better travel far across time."

  5. avatar Greg Ransom says:

    The Ransom Effect is simply a parallel application of what Hayek did to Mises to create the IE construct, taking Mises' point about transportation and prices and transferring the point to the movement of goods across time, eg 'transporting' summer fruit consumption across time to winter time consumption, you can give goods differing geographic place/price stamps and you can give them differing place-in-time/price stamps.

    If you can apply Alchian-Allen to differential costs at different geographic points due to the differential cost of physical transport, you can do the same with different calendar points in time, ie you can map the differential costs at different time points due to the differential cost of waiting, ie the differnetial cost of 'transport' through time.

    • avatar Kurt Schuler says:

      Greg, you will be entitled to call it the Ransom effect after (a) you have written a book in which it is expounded, like the Ricardo Effect, the Hayek Effect, and the Alchian-Allen theorem, and (b) you have undertaken a pretty thorough search of economic writings to determine that you are in fact the first to propose it, and that this is not yet another instance of "Stigler's law of eponomy."

      • avatar Gonzalo R. Moya V. says:

        Plus, "Ransom effect" would be a misleading name given that the word "ransom" has already a meaning of its own (i.e. besides being the last name of its "author") that is quite different from what the "Ransom effect" is trying to explain.
        As for your two requirements for the "Ransom effect" to be called as such, I would say neither is necessary as (a) nowadays, when people quote links from the internet in their papers with the same formality as they do with standard books, and (b) stigler's law of eponomy does explicitly forbid people labeling ideas with their own names, it is just too cocky to do so. I would also say that it only needs to be referred by others as such with the sufficient relevance and/or frequency to be socially agreed to be called that way. Good luck with that, Greg!

        • avatar Gonzalo R. Moya V. says:

          I wish commentors could just edit previous posts to make corrections instead of making a new one just to do so. In (a), delete "when"; and in (b), add a "not" after "does".
          To restate my point, (a) academic papers nowadays list websites in their bibliography (under the title "Works Cited:") if they are widely regarded to be serious (i.e. reliable or rigorous) enough. This blog, in my opinion, would fit well under that category, making Greg´s idea perfectly quotable. (b) Greg is on the right track: searching to verify if an idea already has a label (obviously without having a label to use a key words is daunting), so posting his idea in a blog is the best route to get feedback on his originality. I personally deem it quite interesting and new, its just the name that does not stick with me, how about "Inter-temporal application of the Alchian-Allen Theorem" for the title of an article, and then let OTHERS call the idea "The Alchian-Allen-Ransom Theorem". Much more sticky, huh?

  6. avatar Greg Ransom says:

    Kurt -- I'm clearly crowd sourcing here and looking for productive conversational input.

    Got any?

  7. avatar Greg Ransom says:

    Kurt, I've read all sorts of stuff across the history of economics in this general area, but never seen it in that literature, but one person can't possible have seen all of the economic literature. I've you've got an example of a earlier case of the idea, I would be in your debt if you would send me an example.

    How does one conduct a "thorough search of the economic writings" on such a matter? The idea of a deliberate search for a random unknown is not plausible. Where does one look except everywhere and anywhere?

    Kurt writes,

    "b) you have undertaken a pretty thorough search of economic writings to determine that you are in fact the first to propose it,"

  8. avatar Greg Ransom says:

    Kurt, what is there to 'expound? The idea is right there before you. Alfred Wallace wasn't required to write a book to be credited jointly for the idea of natural selection, and Darwin was forced to *stop* working on his book and to write up a brief article to claim his own paternity to the idea.

    And not well, once interrupted by Wallace, Darwin *never* wrote his long book fully limning his idea -- he wrote a short, rushed version we know today as "Origin of Species".

    Kurt writes,

    "you will be entitled to call it the Ransom effect after (a) you have written a book in which it is expounded"

  9. avatar Greg Ransom says:

    What totems of 'science' and epistemic status would I need to put into an article on this to get it published in the AER?

    I have no idea what the gatekeepers of "real science" in economics would demand of a contribution to 'economics'.

    Anyone want to help co-author an article for the AER on this with me?

  10. avatar Kurt Schuler says:

    Clearly you have a modest humor deficit. My remark was made half facetiously. Godwin's Law came from a single Internet post, after all. But I was half serious. False originality is a real problem in economics. People reinvent the wheel chipped from a boulder because they are unaware that somebody else already developed the pneumatic tire. So, I would say start with survey articles or a history of thought relevant to the subject, and ask people who have written about the subject.

  11. avatar Greg Ransom says:

    I've read hundreds of history of thought books and articles -- I've never seen the idea.

    If anyone here has seen it, let me know.

    Humor can be tough on the internet -- especially dry humor with no non-verbal signal eg ;-)

    Person to person humor involves non-verbal humor signals -- everyone is in on the joke. Just a reminder.

  12. avatar Greg Ransom says:

    "ask people who have written about the subject."

    As far as I know, no one has written on the subject. Got any leads or suggestions?

    Nobody does structure of production capital theory -- British and American economists hated and condemned the very thought of it from their first introduction to it (eg see Hawtrey & Keynes on Hayek & Bohm-Bawerk).

    Roger Garrison and Peter Lewin and the only people who I can think know much about it, and neither have done much very sophisticated with it.

    Anyone else?

  13. avatar Greg Ransom says:

    I really think the way to get this kind of thing published in a top journal is do an econometric type study tied to the fundamental economic logic of choice -- the econometrics would be the "science" that give the fundamental logic of choice "scientific legitimacy".

    Too cynical? I don't think so.

    • avatar Gonzalo R. Moya V. says:

      Greg, you not always need to support an Economics thesis with econometrics in order to have it published. For your particular case, you could add rigurosity to your paper by quoting interviews you make with wine producers. They would support your thesis if they tell you that they use the finer grapes for wine production while selling the other grapes for direct consumption. You could go to Napa, California, for instance, to get several interviews from wine producers in a single town in order to reduce your costs. Once those interviews are included, you should not have trouble finding a non-econometric journal to publish your findings. Again, good luck with that.

  14. avatar Greg Ransom says:

    Anyone have a graduate student who'd like to do some econometrics on this?

    • avatar Bill Stepp says:

      Econometrics?

      Prove the fallacy of econometrics.

      1.) Econometrics is mathematical history.
      2.) Mathematics is a priori knowledge.
      .:. Econometrics is a priori history.
      3.) A priori history is a fallacy.
      .:. Econometrics is a fallacy.

      QED

  15. avatar Michael says:

    Greg and all, I have a couple comments about Greg’s theory. I would appreciate comments from Greg or anyone.

    The Alchian-Allen theorem is based on transportation costs being proportionate to the weight, not the price, of the goods. Thus, the goods having the higher price per weight will have lower transportation costs per price. For these goods, the proportionate increase in price due to transportation costs is smaller, so the price premium of the higher-priced goods (as a fraction of the price of the lower-priced goods in the same market) in the distant market is less than that in the local market. Thus, according to the Alchian-Allen theorem, folks in the distant market will buy proportionately more of the higher price goods than folks in the local market will. However, Greg’s theory involves not transportation costs but rather “carrying” costs over time, i.e., interest. But that is proportionate to price, not to weight. Someone buying, or not selling, grapes in order to make wine and then age it and sell it years later will have interest costs proportionate to the cost of the initial investment, not to its weight. Thus, the interest cost does not seem to be a factor that would tend to make the prices of higher- and lower-priced goods proportionately closer.

    Also, transportation (hopefully) does not change the quality or character of the goods, but aging and other processes that take time do, or the entrepreneur would not engage in them and bear their costs. I understand why some economists consider identical goods to be different goods if in different locations, but that point is irrelevant here. The Alchian-Allen theorem compares two different markets (distant from each other), and both the higher-priced goods and the lower-priced goods are qualitatively the same in one market as they are in the other. In Greg’s scenario, however, the goods are qualitatively different after the aging or processing, which means that their relative values in the eyes of consumers may be different than they were at the beginning. With that difference, it seems impossible to have a rule that applies in all possible situations. In some situations, consumers might buy proportionately more of the aged/processed goods produced from the higher-priced initial goods (compared to those produced from the lower-priced initial goods) than they buy of the higher-priced initial goods (compared to the lower-priced initial goods), and in other situations, less.

  16. avatar Michael says:

    I have a comment on the Alchian-Allen theorem itself, and welcome others’ comments.

    It seems to be valid when considering a situation such as that quoted from Marshall. There, two goods serve the same purpose, but it takes a greater amount (size or weight) of one to do the same job as a lesser amount of the other. One wood is stronger for a given amount. But the weaker wood has a lower price per amount, so that its price to get the same job done may be the same as or even lower than the stronger in the local market where they are both produced. Transportation costs will be higher per job to be done for the one requiring more to get the same job done. Thus, to the extent transportation is needed, it becomes relatively more expensive per job than the other, so the folks in the distant market will tend to buy less of it than folks in the local market in which it is produced.

    However, I don’t see why it is valid where there is a difference in quality that is not reducible to quantity (e.g., with choice and standard grapes). In such a case, the two goods are not interchangeable by adjusting only the quantity. Buyers in the distant market are unable to buy less of the higher-priced good and accomplish the same thing as with more of the lower-priced good. They get a better flavor, but that’s not more of anything. I don’t see why it is true a priori that adding the same cost per weight to each will cause greater buying of the higher-priced one. Let’s assume that in the local market, buyers buy 200 pounds of the standard and 50 pounds of the choice when the prices are 5 and 10 cents per pound, respectively. Let’s also assume that the buyers in the distant market have the same demand schedules and would buy the same at those prices. Now with prices at 10 and 15 cents per pound, buyers in the distant market might buy 100 pounds of the standard and 10 pounds of the choice. In fact, buyers in the distant market might decide to forego the better flavor entirely at the higher price, switching to the lower to economize. The fact that the ratio of the higher price to the lower is less in the distant market than in the local market tells us nothing here if both prices increase. It would be different if the same ratio were achieved by the higher price falling and the lower increasing.

    It might seem that the Alchian-Allen theorem is valid if the difference in quality is not reducible to quantity, but *both* markets produce both high-priced, high-quality, and low-priced, low-quality, versions of the good. If both versions from each market are roughly equal in size/weight, the transportation costs add a greater proportionate cost to the lower-priced versions than to the higher-priced ones, so it is more difficult for the lower-priced goods (compared to the higher-priced ones) to compete in the importing market against the same quality goods produced there. Both markets will import little, if any, of the low-priced goods. But the reason is because of the competition with the lower-priced goods made in the local market, not primarily the competition with the imported higher-priced goods (the analysis of *that* competition is the same as in the previous paragraph, where only one location produces the goods). It may well be that there are *more situations* in which folks import proportionately more of the high quality good than the folks in the producing market buy of the same good – in both cases relative to purchases of the low quality produced in the same locale – when the importing locale also produces high and low quality of the goods, compared with when they don’t. But it would still not be universal.

    • avatar Gonzalo R. Moya V. says:

      Michael, you seem to misunderstand the main point of the Alchien-Allen theorem, which is when applied to perfect-substitute goods with different quality that is not reducible through quantity, as is the case with grapes.
      It is derived from Microeconomics´ Consumer Theory, which consists of comparing the ratio of prices to the ratio of preferences, also known as the "Marginal Rate of Substitution" (MRS).
      For perfect-substitute goods, the MRS is constant, making the decision of which grape to consume a "corner solution", called this way because the optimal basket of consumption between these two types of grapes consists of all of one and nothing of the other (rather than some combination of the two that results in the case of non-perfect substitute goods).
      So, basically, given that the same fixed amount of dollars has been added both to the numerator and the denominator of the ratio of prices when transporting the grapes to a different location, this ratio has been reduced (becoming closer to one) so that more individuals will end up at the corner of consuming all prime grapes and no standard ones, thus resulting in a higher collective demand for prime grapes at this location.

      • avatar Michael says:

        Thanks for the reply, Gonzalo. It doesn't seem that goods of different quality (not reducible to quantity), e.g., standard and choice grapes, are perfect substitutes. That was my point. There is no constant ratio or rate of substitution between such goods. I agree that if the price of both kinds of grapes were the same, that people would buy only the choice grapes. As I said earlier, I would also agree that if the ratio of their prices were to approach 1 by the choice price decreasing and the standard price increasing, consumers would shift from standard to choice. However, that does not hold if the ratio of their prices were to approach 1 while both prices increase.

        This is a side point, but perfect substitutes don't give you a "corner solution" if you're considering just the Marginal Rate of Substitution between the goods, and not looking at their prices. One could have varying amounts of the two goods in combination and remain at the same level of utility. Only if one considers prices, and only if the ratio between the prices differs from the Marginal Rate of Substitution, would one buy only good (the relatively cheaper one, i.e., the one whose price is lower than the price of the other times the rate of substitution of the latter for the former).

        • avatar Gonzalo R. Moya V. says:

          Michael, if for some personal/unknown reason standard and choice grapes are not perfect-substitute goods to you, try at least to imagine them as so, as they are meant to be understood as such in the example.
          Saying that people would only buy the choice grapes if they cost the same as the standard ones is contradictory with claiming that there is no constant rate of substitution between them. Remember from your Microeconomics textbook that with two normal (i.e. not-perfect-subtitute) goods, people value less the one they have more relative to the one they have less, being willing to trade them until their ratio of preferences (which are subjective) is equal to the ratio of prices (which are objective), thus ending up with some combination of the two.
          Saying that a ratio of prices does not approach 1 when both prices increase is mathematically false, so there is no room for debate there: Starting from the fact that a/b>1 if a>b and a/b<1 if a<b, both of these ratios converge to 1 whenever a constant 'c' is added both to the top and bottom as 'c' grows large (i.e. the fixed costs of transportation in the Alchian-Allen theorem). Just grab your calculator or an excel spreadsheet and play with some numbers to see this.
          Your side point, on the other hand, is indeed correct, I was just being succint in my explanation of Consumer Theory so I did not mention the special case when the price ratio is equal to the MRS when dealing with perfect-substitute goods. Under such scenario, you are right, there is an infinite number of solutions to the linear programming problem of utility maximization subject to the budget constraint.

    • avatar Gonzalo R. Moya V. says:

      Michael, I also noticed that you misunderstood Marshall's example. If you read it correctly, you will find that it is quite analogous to the one of the grapes: The thick beams of inferior wood were preferred to the thin beams of superior wood at the place of origin only because the former ones were more abundant -and hence cheaper- than the latter ones, which were relatively scarce -and hence more expensive- (the key word here was "quick-growing"). So, the inferior wood is made as useful as the superior one when cut into thicker beams, but in doing so it becomes relatively more expensive to tranport (due to the higher size and weight, as you said). Thus, the thin beams of superior wood enjoy a higher demand at the place of destination for being relatively less expensive than the thick beams of inferior wood when compared to the place of origin.

  17. avatar John S says:

    Kurt,

    Ok, this is only tangentially related, but Alchian did write about including asset prices when measuring inflation.

    Do you think that money supply growth in a Free Banking system would lead to a better relationship between consumer price increases and asset price increases? In other words, would there be less need after a bust (e.g. in our current situation) for overpriced assets to decrease in value or for the Fed to "reinflate" consumer prices to achieve a correct balance?

    • avatar Kurt Schuler says:

      Briefly: yes, but I think there would still be substantial asset price booms and busts because herding behavior exists and the herd is sometimes very wrong. I do not see a way to eliminate such mistakes under any monetary or economic system; doing so would require changing our biology.

  18. avatar Michael says:

    (This comment is the first one I made, just before the one above that begins "I have a comment on the Alchian-Allen theorem itself..." However, it still shows "Your comment is awaiting moderation" so I assume the moderator just overlooked it and I can't find a way to contact the moderator.)

    Greg and all, I have a couple comments about Greg’s theory. I would appreciate comments from Greg or anyone.

    The Alchian-Allen theorem is based on transportation costs being proportionate to the weight, not the price, of the goods. Thus, the goods having the higher price per weight will have lower transportation costs per price. For these goods, the proportionate increase in price due to transportation costs is smaller, so the price premium of the higher-priced goods (as a fraction of the price of the lower-priced goods in the same market) in the distant market is less than that in the local market. Thus, according to the Alchian-Allen theorem, folks in the distant market will buy proportionately more of the higher price goods than folks in the local market will. However, Greg’s theory involves not transportation costs but rather “carrying” costs over time, i.e., interest. But that is proportionate to price, not to weight. Someone buying, or not selling, grapes in order to make wine and then age it and sell it years later will have interest costs proportionate to the cost of the initial investment, not to its weight. Thus, the interest cost does not seem to be a factor that would tend to make the prices of higher- and lower-priced goods proportionately closer.

    Also, transportation (hopefully) does not change the quality or character of the goods, but aging and other processes that take time do, or the entrepreneur would not engage in them and bear their costs. I understand why some economists consider identical goods to be different goods if in different locations, but that point is irrelevant here. The Alchian-Allen theorem compares two different markets (distant from each other), and both the higher-priced goods and the lower-priced goods are qualitatively the same in one market as they are in the other. In Greg’s scenario, however, the goods are qualitatively different after the aging or processing, which means that their relative values in the eyes of consumers may be different than they were at the beginning. With that difference, it seems impossible to have a rule that applies in all possible situations. In some situations, consumers might buy proportionately more of the aged/processed goods produced from the higher-priced initial goods (compared to those produced from the lower-priced initial goods) than they buy of the higher-priced initial goods (compared to the lower-priced initial goods), and in other situations, less.

    • avatar Gonzalo R. Moya V. says:

      There is a mistake in your first line of reasoning, which you drag from misunderstanding Marshall's example, as I pointed out above. The Alchien-Allen theorem is not "based on transportation costs being proportionate to the weight": although the transportation cost is indeed proportional to the size and weight of what is being transported, in their example of grapes the two types had the same of both characteristics (only their qualities were different), so the initial difference in prices was only due to their difference in quality and this spread got narrower once added the same fixed cost of carrying the commodity from one site to another.
      The "only" contribution of Mr. Ransom (quotation marks because it is not my intention to dismiss it but rather the opposite) is to point out that the word "site" in my previous sentence does not have to imply only a physical place, but also a temporal one.
      Indeed, the cost of carrying the grapes from one time to another (e.g. processing and storage) is the same regardless of the type, so the wine from choice grapes would have an added cost that is relatively not as high as the wine made out of the standard ones (it's easier to see it as if the wine producer buys the grapes from a local supplier rather than producing them as well). This would generate an extra incentive to produce wine from choice grapes if people deem it better than the wine made out of standard ones and thus are willing to pay more for it.
      Economists in general do not "consider identical goods to be different goods if in different locations", doing so would be just pointless and counter-intuitive, but you are right when saying that Greg's example is not straightforward, as the grape becomes another commodity when transformed into wine. As you suggested, if people deem the wine made out of standard grapes better, there is no incentive whatsoever to produce wine from more expensive grapes to sell it cheaper, as he would be cutting his profit spread from both sides (plus, the opportunity cost of direct consumption would make the price of the standard grapes to go up, approaching the price of the choice ones). Therefore, I would recommend putting canned or bottled (pickled) food as a better example.

Leave a Reply

You must be logged in to post a comment. Login now.