Here is an old blog post from the blog "Unlearning Economics" that I thought might get some non-macro discussion going (doesn't anyone else get bored of chatting up the gold standard and ABCT?).
http://unlearningeconomics.wordpress.com/2012/03/27/on-the-incoherence-of-marginalist-labour-economics/
Basically, UE argues that wages argues against the "MVP theory of wages". He doesn't spell out exactly what he means by this (and it actually meant different things to different people in the past), but I take him to mean that wages are not set by the interaction of supply and "demand" (defined as the marginal revenue value product of labor) for labor.
He gives two reasons:
These 2 reasons sound almost identitical to me, but I think they are slightly different. I have my own reasons for doubting their validity, but I was wondering....
Is there a uniquely Austrian take on UE's arguments? Do you agree, do you disagree?? Am I getting his argument wrong? I am curious to know what my micro-minded forum-mates think.
Ambition is a dream with a V8 engine - Elvis Presley
@Johnathan,
Since the topic of the thread has to do with a criticism of Neoclassical price theory, borrowed from Keen's Debunking Economics, it's only natural that I direct you to actually read the book.
Just to make it clear. The topic of this thread was actually stated pretty plainly. I was wanting to talk about a specific question about wage determination that was brought up IN A BLOG POST that DID NOT mention Keen's book. Specifcally, UE claims that the MVP theory of wages falls apart because in most production processes you cannot determine MVP. And my question was how Austrians would respond to UE's argument. THAT is in Keen's book? Very interesting.
I was trying to direct you to the literature on the topic — literature bound to be much more elucidating than any thing this forum (as wonderful as it may be) can offer.
First, if your only goal was to direct me to sources that answer my question, then you could have fooled me. You have pointed me to a number of books and articles that cover a variety of topics, not all of which relate to my specific questions. So I ask for page numbers or chapters you think would be helpful. Yet you never respond. So our exchange has not been all that helpful. :-/
Second, why exactly is my very specific question about how Austrians would respond to UE's argument is not good for an " elucidating" discussion on this forum? I mean, shit, UE's argument was simple enough for him to express in a short BLOG POST! But you think it is too heavy for open form discussion? Why? Wait. Never mind. Don't answer that. I think that question is really too deep for you to fully answer in this limited format.
In any case, if you're not interested in reading the literature out there — reading that would go to great lengths to answer this question all other questions — well, that's too bad for you.
I never said I was uninterested in heterodox lit. But that isn't the point of this thread.
@abskebabs,
What's the maximum you'd be willing to buy this stack from the salesman for, assuming you'd only buy if you could improve the situation of your bottomline from what it currently is?
If I understand you correctly, then I think you are thinking about the problem one of the ways way I am. UE's main argument seems to be that if you can't compute MVP, you can't determine factor demand. I think the way around this problem is to note that there are other ways to think about factor demand.
One way would be the way Milton Friedman described in his Knives/Blades/Handles thought-experiment from his Price Theory textbook. There the derived demand for one input (say handles) was the difference between the "demand" price for the output (knives) and the "supply" price of other inputs (blades) for any quantity. This sounds like what you are saying the sentence above, since it basically translates into the "maximum" the firm would be willing to pay for the input.
And I think that makes sense. But I think if I were UE I would respond that we changed the problem by already assuming the price of the other factors of production are already given. So my second thought was that we could use what are called "conditional factor demands" as the second-half of each factor market. Conditional factor demands are "the cost-miniziming level of an input required to produce a given level of output". In the taxi example, figuring out the "conditional factor demands is simple. You need 1 driver and 1 car to produce 1 unit of taxi services. IOW: for y taxi services you need y drivers and y cars. Now, if we wanted to know the eq'm price for each factor we would just need to simultaneously solve for the eq'm in the output market and the two factor markets (a computational process the market does on its own).
Of course, that is kind of over kill for this discussion. And I know some forum goers don't like math, but really I am just saying that there are two ways to think about the firms decision to hire labor (or any input). One way is profit maximization (the way UE is talking about it) and the second way is cost-minimization (the way I'm talking about it above). In the end, they wind up being the same thing.
Either way you slice it, I think UE makes a good point, but I think there are ways around the problem. And there are ways of dealing with that. At the very least it is something neoclassical economists have been thinking about.
I am still not exactly sure how Austrians would respond (though abskebabs has given me a better idea). But it looks conversation is slowing down and school will be starting back up soon so I just wanted to get my thoughts on UE's criticism.
Hmmm it's interesting you mention Friedman in this connection, though I think the difference between "demand" price and "supply" price is not the way to answer the question I posed above, but I can understand where you're coming from with it w.r.t the question you originally posed (from UE), so I'll try to deal with it first. The procedure sounds very similar to one of the cases Bohm Bawerk covers in his book on Value and Price in the chapter on complementary goods (pp.170-178 in the edition of Positive Theory of Capital on this site - I personally prefer the Hans Senholz translation). Bear in mind he calculates in value terms however.
I don't see how the solution you propose would allow for a determinate price to be worked out between the amount paid for the taxicab and car drivers' labour if these were factors specific to the above described production process, and hence could only be used to produce the taxi services worth whatever price they are. If either factor is even remotely converitble to other uses (as is definitely the case with labour overall), then the price offered for it would at least need to match what the marginal producer employing a different production process of a different value (though this could also be a taxi firm) can offer for the factor.
The example I gave is one that I think that better illustrates an interesting case where we have to think how MVP applies than the one UE gave. Due to the situation I described, you can think of it in terms of hypothetical scenarios to clearly get the answer:
Scenario 1: The stack of B wasn't set on fire and no salesman arrived:
Then given the special paper can be employed with either of the machines to produce either a stack of A or a stack of B, but you already have your stack of B, so if we rank ordinally the best decision made:
Stack of B today => Stack of B tomorrow => sells for $300
Stack of Special Paper => Stack of A tomorrow => sells for $100
Total Revenue=$400
Scenario 2: Stack of B was set on fire but no salesman arrived:
The stack of special paper is a convertible factor, and thus diverted to its more important use:
Stack of Special Paper => Stack of B tomorrow => sells for $300
Total Revenue=$300
Scenario 3: We can thus see the marginal change in revenue dependent on ownership of B to Dunder Mifflin is only $100, the value of the lesser valued stack of A, inspite of the fact that B itself sells for a much higher price. Thus it would not make sense to pay more than $100 for stack B, and this extends the application of DMVP/DMU where products are reporducible from a common, convertible factor. in fact, I think the theorem probably extends in applicabillity beyond the limits BB and Reisman apply to it arbitrarily to "immediately reporducible" factors, though the proper exploration of this would open up a whole new can of worms. In any case, if he buys the stack of paper for <$100, e.g. $50, the situation is improved:
Stack of B purchased <= -$50
Stack of special paper today => Stack of A tomorrow => sells for $100
Total revenue= $350
Note also since the maximum price that would have been paid if the salesman offered a stack of special paper would have also been $100. Hence superficially, if one didn't see or realise the lesser valued stack A, one could easily fall into believing a fallacious cost of production doctrine by observing these types of real life transactions. BB has much to say on this in his discussion of th law of Costs in Positive Theory of Capital and elsewhere.
"When the King is far the people are happy." Chinese proverb
For Alexander Zinoviev and the free market there is a shared delight:
"Where there are problems there is life."
Well, what I was relating from the Friedman text was *just* how to come up with the derived demand for a production process with fixed proportions ( like 2 blades + 1 handle = 1 knife). In that case the derived demand for knife handles (or the most a knife company would be willing to pay for 1 handle) will be the difference between the price of a finished knife and the price of 2 blades (ignoring other factors of production).
However, you're right that if we actually wanted to know the equilibrium price of handles we would have to take account of the supply function for handles (which should reflect the price that those handles would fetch if they were not used for making knives).
I will take a look at your new example later today. Gotta jet off to work right now.
I wanted to point out some information that might be useful in looking at this problem.
These comments come from Mises in Human Action Chapter 7 - Section 1 The Law of Marginal Utility
First for marginal utility to be relevant the units of a good must be subjectively homogenous:
Quantity and quality are categories of the external world. Only indirectly do they acquire importance and meaning for action. Because every thing can only produce a limited effect, some things are consider scarce and treated as means. Because the effects which things are able to produce are different, acting man distinguishes various classes of things. Because means of the same quantity and quality are apt always to produce the same quantity of an effect of the same quality, action does not differentiate between concrete definite quantities of homogeneous means.
So, the point is if they aren't in fact homogenous you aren't comparing the units with each other. Secondly, the point is that each unit is value differently. And that if one has ends that require the use of a unit of a good, then the value of that unit can only be derived from the subjective value of the end that it will serve as a means to produce. A second end to which it might be used, is necessarily less valued than the first end (or vice versa, one of the two ends is preferred), and the subjective value of a second unit of the same homogenous good would necessarily be derived from the less valued or second end.
The cab driver example is very obvious. The marginal utility of the second unit to that specific consumer of the service, is 0. Because the means has no lesser preference to which it could be put, this is a function of the effect of the cab service in reality. If it moves me elsewhere in space and time, then there will be no second end to which I could apply the second cab (at the same location). Note that the location in space and time do affect the cab. A cab across town, or a cab here later in the day has a different value to me, than the one in front of me right now.
Note, there may be a case where a single cab is not the unit. What if I'm moving my entire family (5 of us) to a location. Then two (3 passenger cabs) is the marginal unit for the satisfaction of his most preferred end, not one cab.
In the second example, "team production" is what would be referred to as a process. You rightly identify that the labor of a carpener and the labor of a plumber are not homogenous. They are not the same good. To the purchaser of these labor products, he must purchase all of them, or none of them. If he cannot purchase all of them, his goal, a finished house is not an attainable end.
So we see decline in subjective value for homogenous units on the demand side. We can expect that in the market we would see this decline as a finite price paid in the market unit for homogenous units of that good. We would see specific quantities purchased and these would not be homogenous, in that if one purchaser needed 8 hours of plumbing labor and the other needed 16 hours each would not value the "unit" purchased by the other. However, in purchasing the goods, because the market would act (through exchange) to compare the prices in terms of some homogenous unit, in this case a cost per hour.
On the supply side for plumbing labor, one assumes, since man does not work at all times and in all places, that there is a disutility to labor. A 24 hour day limits a man to producing 24 units of labor per day. However, man works less than 24 hours per day, there are other activities he prefers to do. Given the option man will provide less. Again in deciding how many units to offer in the market place, he does so at the margin. The man provides units of labor at the market price, until the quantity of money he can get for that next unit of labor is valued less than his ability to engage in other non-productive endeavors, leisure. Or in productive efforts that don't return money to him.
In both cases, the elusive and hypothetical "equilibrium" price is set at the margin. Not because people buy homogenous units, but because in competing for specific quantities of goods and services in the market, the different subjective use-values are quantified into prices of homogenous units.
This is the logic of human action. This is, as I understand it part of the way in which Austrian economics explains how a market arrives at "market prices." And why the margin insight by classical economics is correct, it was just formulated without respect to the proper categories present in human action in the real world.
Sorry, I should have been clearer, I completely understand what you're getting at with the above, which is why I said it's also one of the examples BB deals with. What i was puzzled by was what you seemed to be referring to regarding a solution to the problem when the prices of the other factors of production (e.g. the 2 blades) were not known and all factors were only specific to this particular production process. You alluded that Varian had a way of mathematically determining them in such a case even where you couldn't already resort to factor prices (and it would only make sense for the other factors to have definite, non-arbitrary prices if theyw ere bidded for by other producers for different processes in the factor market in any case).
Hey guys, some great discussion here. I've been away so only just noticed it in my referrals.
I agree that the plumbing example was bad. If you build a house without plumbing, you've got a house without plumbing, rather than no house. Within the confines of that example I'd say a carpenter would work better, but the two men carrying a box is the clearest. Moshe Adler, in his book Economics for the Rest of Us, uses the example of Mcdonalds: you need somebody cooking the burgers, somebody taking the orders, and extra grills and tills to go with these people. If you take a McDonald's team as a whole, it's pretty hard to nail down individual productivity of inputs.
it sounded like he was using examples where there is no ability to replace one factor for another. Look at his taxi cab example. There is simply no opportunity to replace taxi cabs with people (I can keep buying cars, but unless they drive themselves, I wont be able to expand the quantity of taxi services i produce).
The point is the ability to replace a factor with another of the same type: a labourer with a labourer; a taxi with a taxi.
I'm surprised to see a level of agreement here as this is basically an argument for a minimum wage, which I'd guess Austrians aren't all too keen on. Looking through, the only disagreements I've seen tend to rest on an implicit 'the price is right' premise, but this is pretty difficult to verify without circular reasoning as we generally have only prices to measure what is 'right.' And even if the price is justifiable by other means, it simply cannot reflect the marginal productivity of an input if the input has no marginal producivity alone.
And I don't know if the empirical work citied in Keen's book is better than what UE references on his blog, but so far the empirical papers UE has posted have not exactly blown me over.
Have you read this paper? For me it shows pretty convincing evidence that firms vary all inputs at once, which seems to corroborate with the argument I present.
There's been disagreement with your premise if you read more carefully, but more than that, you are now making completely unwarranted inferences. Even if labour were a specific factor of production in the way you claim it is (and it isn't), you leap from the alleged difficulty in determing DMVP in some instances to this being an argument for the minimum wage. Based on what? It would merely mean the purchaser of the services would be in a difficult position to accurately price them, not that being unable to do so would incine them to retain services that are now both more expensive and still difficult to evaluate.
Freedom of markets is positively correlated with the degree of evolution in any society...
It's not really about the difficulty of discerning an MVP, it's about the fact that an MVP can only be said to exist when the factors of production are combined. Hence, any minimum wage that is lower than the combined productivity would simply reduce the returns to capital/land, rather than cause unemployment.
Even if labour were a specific factor of production in the way you claim it is (and it isn't)
Could you clarify this point?
It would be a specific factor of production if it were only serviceable to producing one kind of good (the most extreme case.) In that case Mises argues that bargaining is necessary to establish the contribution of the specific factor of production. "Labour" (i.e. the price at which humans sell their services) is the least specific of all the factors of production.
Moreover, the entire bloody point of MVP is to disaggregate and determine the contribution of each individual factor in producing the final good. If the MW goes up, the employer will still simply hire more productive, experienced workers, cut down on production until it becomes profitable again, shift it elsewhere (since the factors can be more fruitfully employed in other sectors of the economy) or simply not produce. Capital and land can both be shifted to other uses. Why would "combined productivity" be the sole consideration in determining whether or not to employ a factor of production in producing a good? I mean who is going to commit their capital or land to production when it can be put to other, more profitable uses?
Okay, this is gonna get really good, really fast. I hope that this is what you wanted Student, because I think this will turn out to be a really interesting thread.
As for Jonathan and Abskebabs, can you try and elaborate on your respective arguments? I sort of see where you're going with your example Abs, but I could do with further elucidation to mull over it in my head. These are subtle nuances of Austrian capital theory I wasn't aware of.
@unlearningecon,
Glad you dropped in! I've been enjoying your series on Keen's book btw. :)
Moshe Adler, in his book Economics for the Rest of Us, uses the example of Mcdonalds: you need somebody cooking the burgers, somebody taking the orders, and extra grills and tills to go with these people. If you take a McDonald's team as a whole, it's pretty hard to nail down individual productivity of inputs.
I 100% agree. Team production can certainly make it difficult (if not impossible) to measure the productivity of individual inputs. Do you have any opinions on Alchian and Demsetz's 1972 treatment of this topic? As I understand their argument, they would say that where firms cannot measure individual productivity, they will instead measure behaviors correlated with productivity (like the number of smoke breaks you take) to set wages. And if the cost of monitoring those behaviors is high enough, firms will abandon trying to offer a fixed wage altogether. Instead, they will offer different sorts of compensation packages (like working on comission).
http://www.aeaweb.org/aer/top20/62.5.777-795.pdf
hmmm. Let me quote the exact passage from your blog post that I had in mind when I wrote the thing you quoted.
Production, by definition, requires that all the factors of production be brought together. As a result, labour is employed at the same time as capital (and land, but we can lump the two together for the purposes of this post) in every circumstance. A taxi driver without his taxi is worthless, and vice-versa. http://unlearningeconomics.wordpress.com/2012/03/27/on-the-incoherence-of-marginalist-labour-economics/
To me this paragraph is about the firm's ability to substitute 1 input for another. In the case of producing taxi services, you can never substitute a car for a driver. As a result, the marginal product of an additional driver is zero. Thus leaving the MVP theory of wages unapplicable to taxi drivers (or laborers in similar situations). That is what I thought was at the heart of what you were driving at. Was this not what you actually had in mind? If not, let me know what I missed.
For some reason I can't bring up the paper. I am at work so it could be my slow internet connection. Is that the Eiteman and Guthrie paper you posted? About the survey of firms on their cost functions that was done way back in the '50s? The web-address looks the same. If it is, one reason I didn't find that paper too convincing was because it is based on survey data and I wasn't sure how accurate a businessman's response would be to the types of questions they were asking on the survey. Obviously, data quality is always a concern when you're doing surveys. But this type of concern is mitigated these days by doing cognitive interviews and other types of pre-testing before fielding the survey to confirm that the survey instrument would be understood by respondents. Yet E&G don't mention doing any of this type of pre-testing in their paper (considering it was written 60 years ago it may have been before pre-testing was considered a best-practice).
But, ignoring that, even if I did buy into Eiteman and Guthrie's conclusions as presented I'm not sure why they should be a threat to the Neoclassical theory of the firm. I think the past 30-40 years of theoretical work on increasing returns to scale show that the neoclassical model is capable of handling declining average costs. It is true that increasing returns to scale and the perfect competition model don't mix, but (like i was saying earlie) the perfect competition model isn't the only tool in the neoclassical toolkit.
UnlearningEcon:It's not really about the difficulty of discerning an MVP, it's about the fact that an MVP can only be said to exist when the factors of production are combined. Hence, any minimum wage that is lower than the combined productivity would simply reduce the returns to capital/land, rather than cause unemployment.
I guess you will agree that the people can decide themselves if they want to be employers or employees. What do think motivates them to chose one over the other?
Don’t you think that a minimum wage will just distort incentives on the margin away from being an employer to rather be employed no matter the MVP? With all the logical effects that less people are self-employed, entrepreneurs offer less jobs, and too many want to be employed, causing unemployment and a decrease in productivity, making redistribution necessary to sustain the unemployed, causing overall a lower living standard?
Imagine as opposite example a scenario in which people inatially would only get a ridiculous low wage like 1USD/h (No minimum wage laws). Hence it was unattractive to be employee but very attractive to be employer. This would cause people to save since capital would yield huge returns, they would start business and hiring loads of people since they cost quite nothing. Hence pushing up their wages etc. until employers and employees are in a balance only market pricing can bring about.
Doesn’t any (meaningful) minimum wage regime sabotage that no matter if you can determine the MVP in any individual case or not?
This is familiar territory - often when I debate somebody who disagrees with me on this, we end up talking past each other. I say 'surely you can see that a taxi driver needs his taxi, so they only have an MVP together?' They respond 'but they calculate their factor input prices based on market signals, which reflect relative worth etc.' I haven't really managed to properly articulate the real pinpoint of the disagreement yet, but I'll give it another go here.
@Jon Irenicus
OK I understand that labour can have different uses but I don't see how this changes that it must be combined with capital in order to produce something?
Moreover, the entire bloody point of MVP is to disaggregate and determine the contribution of each individual factor in producing the final good.
Sure, but I'm saying that this is not possible.
Capital and land can both be shifted to other uses. Why would "combined productivity" be the sole consideration in determining whether or not to employ a factor of production in producing a good? I mean who is going to commit their capital or land to production when it can be put to other, more profitable uses?
This is true. If the MW goes up then it will possibly alter the structure of production. However, labour will still have to be employed too - perhaps more, perhaps less, perhaps roughly the same - the overall effect is ambiguous. That's when I'd defer to the evidence (probably not worth dicussing here).
If the MW goes up, the employer will still simply hire more productive, experienced workers, cut down on production until it becomes profitable again, shift it elsewhere (since the factors can be more fruitfully employed in other sectors of the economy) or simply not produce.
Why would they not have hired these workers already if they are more productive? There's a possibility that the MW will create unemployment, if a worker is unable to produce more than he can when combined with capital. But otherwise it will just eat into profits.
@student
Thanks!
To me this paragraph is about the firm's ability to substitute 1 input for another.
Yes, you are right - it is possible that capital will be able to substitute capital for labour, absolutely. But it is also possible to subtitute labour for labour. The relative ability of each to do this will determine bargaining power, no matter the nature of the potential replacement.
I'm not sure what better we can do than asking businessmen what costs they face - surely they will know what is going on inside the firm? You may well have read Piero Sraffa's paper on why this will be the case, if you'd like a more logical approach. The neoclassical model is capable of showing this, absolutely - all you need is a flat supply curve! But that is not what is taught.
@Skylien,
Your post was interesting but I have a feeling I will end up responding to too many people and don't want to get into the minimum wage specifically. However it is entirely possible, within your example, that existing entrepreneurs would offer more jobs by expanding with less competition.