Monday, March 31, 2008

Comity?

Mike Beggs writes:
"I've thought for quite a while that mainstream economics is supported by an exoskeleton of empiricism rather than a backbone of general equilibrium theory. At least the parts that matter. Attacks on a monolithic ‘neoclassical economics’ - conceived as an abstract, ridiculously unrealistic construct - miss the mark, because outside the academy - and even in much of that - this is not how economics is done. Mostly modern economics is about extrapolating data series from past trends, and drawing maps rather than diagrams."
Nicky Kaldor wrote:
"It is the hallmark of the neo-classical economist to believe that, however severe the abstractions from which he is forced to start, he will 'win through' by the end of the day - bit by bit, if he only carries the analysis far enough, the scaffolding can be removed, leaving the basic structure intact. In fact, these props are never removed; the removal of any one of a number of them - as for example, allowing for increasing returns or learning-by-doing - is sufficient to cause the whole structure to collapse like a pack of cards." -- Nicholas Kaldor (1966), "Marginal Productivity and the Macro-Economic Theories of Distribution", Review of Economic Studies, V. 33, N. 4: 309-319.

40 comments:

YouNotSneaky! said...

"the removal of any one of a number of them - as for example, allowing for increasing returns or learning-by-doing - is sufficient to cause the whole structure to collapse like a pack of cards."

Those props have been removed over and over. The structure is still standing. I think Nicky was getting ahead of himself and assuming that what couldn't be done in 1966, once it was done, would do what he wanted it to do. But it didn't

Gabriel M said...

Empirical studies of increasing returns, using the Cobb-Douglas or Translog production functions are a dime a dozen. (i.e. Yatchew)

You can find increasing returns, in one form or another, in basically all endogenous growth models, with some notable exceptions.

At least 75% (guesstimate) of contemporary business cycle theory research is done with imperfect competition and costs to adjustment.

Learning-by-doing? Arrow, Lucas, the never-ending literature on the Liberty Ship data?

If Kaldor's point is that the "private" equilibrium in these circumstances no longer supports the 1-st and 2-nd welfare theorems then yes, but the same could be said to happen, and I haven't read him pointing this out, when government is anything but a stylized angel, for example. -- There's more to life than welfare, anyway.

Anonymous said...

The point is, surely, how does getting rid of one of the props impact on the rest of the structure.

Most obviously, how does increasing returns impact on the neo-classical theory of distribution. Under increasing returns to scale, Euler's Theorem implies that if factors are paid their marginal products, then total factor payments will exceed output.

So, getting rid of one prop may impact elsewhere. This may necessitate the illogical use of the missing prop is, say, justify rising inequality. In which case, the prop is indeterminate -- missing to prove one thing, there to justify another...

YouNotSneaky! said...

"This may necessitate the illogical use of the missing prop is, say, justify rising inequality"

Huh?

Anonymous said...

And do not increasing returns to scale imply a downward-sloping supply curve? In which case, what about the rest of the neo-classical model?

Also, there is this quote by Piero Sraffa from 1926:

"Business men, who regard themselves as being subject to competitive conditions, would consider absurd the assertion that the limit to their production is to be found in the internal conditions of production in their firm, which do not permit of the production of a greater quantity without an increase in cost. The chief obstacle against which they have to contend when they want gradually to increase their production does not lie in the cost of production -- which, indeed, generally favours them in that direction -- but in the difficulty of selling the larger quantity of goods without reducing the price, or without having to face increased marketing expenses"

Nice to see that mainstream economics started to look into this issue a mere 60-odd years since Sraffa discussed it...

Anonymous said...

"This may necessitate the illogical use of the missing prop is, say, justify rising inequality"

Huh?

I was under the impression that inequality was rationalised and explained in the neo-classical paradigm by use of marginal productivity theory.

As this is based on Euler's theorem and since increasing returns produce a different outcome under this assumption, removing the "prop" of constant returns means that marginal productivity theory cannot be used to justify income inequalities.

But, of course, it has been. So what is it? Has the "prop" related to constant returns been "removed" or not? If it has, then someone should tell all those economists justifying inequality using it. Or is it selectively removed?

It seems to me that Kaldor may still be right. Remove one prop and another part of the model falls...

Another example seems appropriate:

"A few neoclassical theorists have broke part of the mould and are incorporating increasing returns into the analysis of international trade and growth theory . . . However, Romer, for example, makes it evident that he is not straying beyond the boundaries, by designing his analysis as, "…a well-specified competitive equilibrium model of growth. Despite the presence of increasing returns a competitive equilibrium with externalities will exist…and is capable of explaining historical growth in the absence of government intervention" . . . Nowhere does he emphasise that increasing returns also implies a downward sloping supply function and the potential of resulting disequilibrium."

Herb Thompson (1997) "Ignorance and Ideological Hegemony: A Critique of Neoclassical Economics", Journal of Interdisciplinary Economics, 8(4): 291-305.

How does having a downward sloping supply function affect the rest of the neo-classical model? Does the "structure" still stand?

Anonymous said...

I think Mike Beggs is *right* when he says "this is not how economics is done." Arrow-Debreu GT theory is still taught, but most fields within economics(including macroeconomics) use special case G.E. models that have became conventions within the field, often because they do *fit* certain stylized facts. The neoclassical theory of distribution, in its simplest form, is a G.E. model which does a reasonable job at "fitting" historical developments in U.S. labour-capital shares of national income (perhaps Radek can comment on the international success of the model). It is not a model that can be derived from the more general Arrow-Debreu model in a straightforward manner by "adding up." Trade theorists work with G.E. models of imperfect competition. These models provide an explanation for the predominance of intraindustry trade, which (seems to) run contrary to the predictions of the Heckscher-Ohlin model. While there has been a proliferation of "applied" models which fit various stylized facts, my impression is that G.E. theory of a more general nature has declined in importance as a research topic.

-h.e.

Gabriel M said...

Anonymous, even with increasing returns, if a factor is fixed in the short-run, you can still get "traditional" cost conditions.

The fact that businesses usually have spare capacity, to buffer against demand fluctuations and such, does not mean that economies of scale are pervasive. All equilibria will be on the demand curve too, so what's the issue with that quote?

Furthermore, it's a first class straw man to say that:

>> "I was under the impression that inequality was rationalised and explained in the neo-classical paradigm by use of marginal productivity theory."

I don't know and I don't want to know what you mean by "rationalised" (I suggest you tell it to the horde of leftist neoclassicists, some of them of Nobel caliber). As for "explained", a contemporary neoclassicist (and let's say we use the definition/3 criteria from Arnsperger and Varoufakis) will point to several factors re: the determination of income:

* gool ol' supply & demand (i.e. endowment/current wealth);
* market power;
* political power;
* etc. etc.

If you want to pick a fight with J.B. Clark, a time machine will be in order, because you won't find anyone like that around, maybe except from the 1st year principles texts.

Robert Vienneau said...

I think quotes I've previously given from Roncaglia are consistent with the views of Mike Beggs, H. E., and Kaldor. Do mainstream economics pretend that the empirical content is in their floating ad hoc hypotheses, rather than in unfounded textbook economics?

I'm not going to bother linking to my posts with quotes from Mankiw and DeLong in which they rationalize increasing inequality on the basis of marginal productivity. Gabriel is being amnesic or disingenuous. For that matter, where does he think the non-empirical idea of "skills-biased technical change" that Krugman (this year) and Rodrik are reacting against comes from? (By the way, here's more from Kaldor.)

It is no good citing endogenous growth theory. I have already noted that Romer's work is nonsense.

This blog doesn't treat mainstream economics as monolithic. For example, I've drawn parallels between the heterodox economist Paul Davidson and the more mainstream economists Brian Arthur and Paul David.

I haven't decided what I think of some diversity in mainstream economics, even when I know some come to political conclusions I can agree with.

Robert Vienneau said...

Should be:

Do mainstream economics pretend that the empirical content is in unfounded textbook economics, rather than in their floating ad hoc hypotheses?

Gabriel M said...

I remember Mankiw saying something in that direction on his blog. I don't think he would put it in print in professional journals.

Re: skills and technical change, I find nothing disturbing in the idea that the ability to use a computer will get you a higher wage in today's economy, all thing equal. I suspect there is more than one theoretical way to get this result.

There's more to endogenous growth than Romer and, in any case, what you point out is that his model employs some (unstated?) restrictions on the production of those goods, rather than it being "nonsense".

One potential approach is to consider mainstream economics "floating ad hoc hypotheses" and another is to understand that mixing strong results with approximations and limited cases is as good of a way as any to go forward, when knowledge is imperfect and new phenomena plenty.

I guess it's ironic that while mainstream economists use their dirty, "ad hoc" methods to study interesting phenomena, others are busy witting the n-thousandth indictment of neoclassical economics. If the principles textbook marginal product theory would hold, these authors would get no income.

Anonymous said...

I should clarify that in my original post I meant no disrespect to heterodox theoretical criticisms, which I'm obviously sympathetic to. I meant only that I think an awful lot of economics, inside and outside the academy, is fairly pragmatic with theory and essentially empiricist. Especially the kind of economics practised in government agencies and private banks.

Of course developments at a theoretical level have an impact on how the empirical is organised. And certainly it has some ideological impact, giving an illusion of depth and certainty to economics as a social science. But ultimately I think the Cambridge critique was not fatal to 'neoclassical economics' because what it undermined really was just a 'parable' as far as applied economics was concerned.

YouNotSneaky! said...

This is becoming somewhat of a pet peeve of mine so I'm gonna comment. There's a difference between believing in the marginal productivity theory of distribution which says that all factors get their marginal products and believing that wages are related to productivity, purely technological productivity. So maybe, as in some increasing returns model, labor doesn't get paid its marginal product (whatever that is). But that doesn't mean that wages are completely divorced from productivity. Productivity goes up, wages go up. In terms of growth rates, if the monopoly/monopsony/other externality mark up is constant then the growth rate of wages will still reflect growth rate of productivity. As seems to be the case. Or take the return on education. Seriously, to deny any kind of link between productivity and wages seems outright silly. (And this is part of the reason why the "neoclassical structure" doesn't crumble even in the presence of increasing returns and the Euler Theorem not holding)

I don't know about Mankiw, but I think at least Brad DeLong's views are more of this general sort.

Also, Robert does misuse the terms "illogical" and "nonsense" in some contexts where as what he really mean is "I don't like some particular assumption of a model whose conclusion I disagree with".

Also, again, you can have IRS, you can have wages not equal to marginal products and still have changes in the income distribution driven by changes in "skill biased technological change". And while trade may play a role in acerbating the effects of these changes, wages are still linked to productivity. Even for Krugman. Or take the changes in returns to education.

Anonymous said...

Obviously labour productivity sets a limit to wages (at an aggregate level - I'm sure you're not suggesting that individual workers' wages are determined by their marginal productivity), but it's going too far to suggest it determines them. There have been large shifts in labour and profit shares over the decades.

YouNotSneaky! said...

"There have been large shifts in labour and profit shares over the decades."

Not really. While there has been some change recently, overall in the post WW2 period labor's share in US has fluctuated between about .68 and .72
(http://research.stlouisfed.org/publications/net/20040801/cover.pdf)

A bit more precisely while the wage and salary share has fallen from about .59 to .56 over the past two decades,
(http://papers.ssrn.com/sol3/papers.cfm?abstract_id=658225)
once you include other forms of employee compensation it has pretty much stayed in the band mentioned above.

The same thing is roughly true for other developed countries and is of course one of the famous "Kaldor's Stylized Facts".

Also, as far as I can tell, there doesn't seem to be much pattern in the fluctuations within that narrow band or any obvious correlation with other potential determinants of wages (bargaining power, unions, trade etc.).

And yes, I do mean to suggest that individual worker's wages are RELATED to their productivity. Ceteris paribus and all that.

Anonymous said...

Well it depends what you consider 'large'. And in fact the US variation is significantly smaller than most developed countries: in fact it's the flattest curve in the OECD, so possibly not the best example to prove your point! Among industrial countries as a whole the average variation over the period since the 1960s (which I think understates the whole WWII period) is about twice what you suggest.

http://www.imf.org/external/pubs/ft/wp/2006/wp06294.pdf (See the appendices)

Certainly though I would agree that the variations are not easily explicable, especially short-run variations. Over a longer period, the low profit rates of the 1970s are surely part of the story, as well as the high unemployment and lower bargaining power of workers since.

As for individual productivity, how on earth do you measure it, with most labour processes being collective? It's hard to judge the individual productivity of an accountant, say, relative to an assembly line worker.

Luis Enrique said...

Mike,

Even without knowing how to measure individual productivity, a firm may believe, with good reason, that losing one production line worker will have much less of an impact on profits than losing its chief accountant. That explains a marginal productivity based wage differential, doesn't it?

(just to be clear, I'm not saying wages = marginal productivity, I'm just saying that even with numerous imperfections in the labour market, and the difficulty of knowing marginal productivity, it's still reasonable to suppose a pretty strong link between marginal productivity and wages)

Gabriel M said...

Not all circumstances are alike. It might be the case that in some situations the game theory of team production is the baseline while in others, the marginal (or at least average) contribution can be estimated based on the technological recipe.

That being said, there are models out there, on CEO pay for example, where a company must have {0, 1} CEOs and with no CEO, there's no activity, so things break down there, somewhat.

YouNotSneaky! said...

Mike,

(I'm getting a "Page Not Found" error message from your link. So instead I'm referencing data from here:
http://www.econ.upf.edu/docs/papers/downloads/374.pdf
and assuming that it's similar in nature. Unfortunately it covers a shorter time period.)

Yes. You're right that the changes in the labor's share for the US are the smoothest among the OECD. They are certainly smaller than the ones for Europe. For France and Germany for example the labor's share has fluctuated between .6 and .72.

But... I think this is more support for my position. Look at it this way. The US has had a much more deregulated labor market than Europe where unions are stronger, there are numerous labor regulations and wages are often set through some kind of a "cooperative" bargaining process between workers and owners. So it's not surprising that the labor's share is more volatile in places where other considerations come into play. You have a fairly unregulated labor market you get a fairly constant labor share, says the neoclassical growth theory. You have a regulated labor market then you get many deviations from that neoclassical growth theory.

Second, are these shifts in labor share "large" in some sense? Well, for US I would say no - it'd be an interesting excercise to see how much just basic measurement error in micro wage data one would need to produce these +/- 2% fluctuations but my guess is "not much". So for all we know that sucker for US is a flat line and the fluctuations are just noise in the data. However, I do agree that the shifts in European labor share's are probably too big to be accounted for by noise. But as I note above, that's what you'd expect.

Finally. Actually we could skip this whole marginal productivity beeswax and recast the whole aggregate Cobb Douglas production function framework as one where output is actually produced through whatever means you desire, and the division of this output is carried out through by bargaining. If our bargaining solution is Nash then we get the same answer where the bargaining power of labor and capital is the relative speed at which their productivity diminishes. Or to make a point in a straightforward manner, even if wages and profits are set through bargaining then surely the (yes, marginal) productivity of labor and capital still plays a role. A worker who cannot produce anything under any circumstances will have no bargaining power. (Here perhaps a better concept would be the Shapley Value rather than Nash. If I recall correctly John Roemer's used it in the past for this sort of thing).

Finally, finally, hey, they've calculated the value of the marginal product of Wayne Gretzky (in fact it was the owners of the LA King's in 1988 who did it, not fancy shamncy academic economists. Who says that people don't believe this stuff outside of academia?) so just because the modern business enterprise is complex and some workers are complementary to others (say, accountants to assembly line workers) it doesn't mean it can't be done. Just treat accountants and assembly line workers as different kinds of production factors.

Anonymous said...

"There's a difference between believing in the marginal productivity theory of distribution which says that all factors get their marginal products and believing that wages are related to productivity, purely technological productivity."

Obviously a difference lost on those neo-classical economists who have justified rising inequality with marginal productivity theory...

"So maybe, as in some increasing returns model, labor doesn't get paid its marginal product (whatever that is)."

So the neo-classical theory of distribution is "whatever that is"? Interesting...

"But that doesn't mean that wages are completely divorced from productivity. Productivity goes up, wages go up."

Not for the last 30 odd years in America. Productivity has been rising, but wages have been relatively flat. The difference is being reflected in rising inequality.

Anonymous said...

"If you want to pick a fight with J.B. Clark, a time machine will be in order, because you won't find anyone like that around, maybe except from the 1st year principles texts."

Damn those introductory neo-classical economics texts, telling people lies about neo-classical economics!

And damn those talking heads who repeat those same lies to justify rising inequality!

Anonymous said...

"Not really. While there has been some change recently, overall in the post WW2 period labor's share in US has fluctuated between about .68 and .72"

To state the obvious, but "labor's share" includes wages and salaries. A CEO, in this sense, is "labor" -- they are hired by capital. So if CEO wages are rising (as they are) then labor's share of income can stay around .7 but it still signifies a fall for the average worker.

"A bit more precisely while the wage and salary share has fallen from about .59 to .56 over the past two decades . . . once you include other forms of employee compensation it has pretty much stayed in the band mentioned above."

To use an obvious example, if health costs are rising (as they are), then overall compensation can stay approximately still but this does not mean that worker living standards are.

So a fall from .59 to .56 is the significant factor, given rises paying for "compensation."

"And yes, I do mean to suggest that individual worker's wages are RELATED to their productivity."

Which does not explain the divergence in productivity growth and real wages for over 2 decades. Or is it a case that productivity has also fallen 3% over that time?

Anonymous said...

"the changes in the labor's share for the US smoothest among the OECD."

Yes, smoothly downwards, suggesting a almost continual shift of income from labour to capital...

"They are certainly smaller than the ones for Europe."

In Germany, the share for labour was approximately the same at the end of the period as at the start, suggesting gains for labour which have been recouped.

I wonder what American workers would choose? After 30 years to be about the same, or to have less?

"The US has had a much more deregulated labor market than Europe."

I wonder how many regulations unions are subjected to in America compared to Europe? Thatcher "deregulated" the labour market in part by increasing regulation of unions to such a degree as to make industrial action harder to do.

"You have a fairly unregulated labor market you get a fairly constant labor share, says the neoclassical growth theory."

It is a fairly constantly decreasing labour share, unlike the European one were workers seem to have made gains and then lost them.

As would be expected from a theory of distribution based on class struggle rather than mythical marginal products...

I wonder if the data was extended over the last 15 years what we would see...

YouNotSneaky! said...

You haven't even bothered to actually look at the data and papers I linked too have you?

Anonymous said...

Finally. Actually we could skip this whole marginal productivity beeswax and recast the whole aggregate Cobb Douglas production function framework as one where output is actually produced through whatever means you desire, and the division of this output is carried out through by bargaining.

I never quite understand why you like going down this road. I don't dispute that the aggregate production function *fits* certain stylized facts. The issue is whether this constitutes evidence in favour of a marginal productivity model of income determination. The work of Fisher, Felipe and others suggests that the *success* of the Cobb-Douglas production function is little more than an accounting identity, since aggregate output is not a physical measure of output. At the same time, both the aggregation literature and the Cambridge Capital Controversy cast doubts about the *existence* of an aggregate production function (even Solow has indicated that it is only to be taken as a potentially useful parable).

I'm not opposed to the idea that the methodology may be useful for forecasting or heuristic purposes, but the not inconsiderable conceptual problems with the framework seem to be almost completely ignored by economists, both in textbooks and vast tracts of the modern academic literature. It's hard not to be suspicious of the quality of explanations of income distribution using these kinds of models, when (i) there is a seemingly willful ignorance of the problems of these models and (ii) the models may have inherent ideological purposes.

-h.e.

YouNotSneaky! said...

"I never quite understand why you like going down this road. I don't dispute that the aggregate production function *fits* certain stylized facts. The issue is whether this constitutes evidence in favour of a marginal productivity model of income determination...the models may have inherent ideological purposes."

Well, the reason why I wanna go down that road is because essentially I don't care about the ideological purposes, nor do I feel like or particularly want to, defend the marginal productivity determination of income.

For me the more important thing is what does the Solow model says about sources of growth. And it says that it ain't capital accumulation. And that result obtains whether or not workers are paid their marginal product or wages are set through bargaining or whatever.

It's worth remembering that the Solow model is in fact barely an "economic model" and is rather more of an "engineering model". The only economic behavior that happens is that people save a constant fraction of their income, and even that isn't necessary for it to work. Everything else, except the constant shares part, follows mechanically from the diminishing return to capital and Inada conditions. The wages and profits could be set by some central authority or through bargaining or written down on stone tablets on top of the mountain. There'd still be a steady state and all growth in that steady state would still be due to factors other than capital accumulation.

So no, I don't think the issue of whether marginal productivity theory of income distribution holds or not matters for the big questions of growth and for the neoclassical growth model.

YouNotSneaky! said...

"Obviously a difference lost on those neo-classical economists who have justified rising inequality with marginal productivity theory..."

Non-sequitur and completely wrong to boot. You can NOT believe in marginal productivity theory of income distributions and STILL think that rising inequality is driven by changes in relative productivity levels, as I've tried to explain, and as you obviously've failed to comprehend.

"So the neo-classical theory of distribution is "whatever that is"? Interesting..."

No, it means that you don't need a smooth aggregate production function to talk about marginal products... which is what the statement referred to so I don't know why you changed it to "neo-classical theory"

"Not for the last 30 odd years in America. Productivity has been rising, but wages have been relatively flat. The difference is being reflected in rising inequality."

Wages have been flat. Total worker compensation inclusive of non-wage renumeration hasn't. Also you seem to be confusing the median with the average. Median wages can be roughly flat, but average wages increase, keeping the share of labor in GDP constant and also resulting in increasing inequality. This isn't that hard, you know?
Also also. The rise in inequality has been due to the rising differential in compensation between skilled (college degree) and unskilled workers (w/o coll degree). Not a shift from labor to capital. As evidence in the almost constant share (in US) which you ignored.


"To state the obvious, but "labor's share" includes wages and salaries. A CEO, in this sense, is "labor" -- they are hired by capital. So if CEO wages are rising (as they are) then labor's share of income can stay around .7 but it still signifies a fall for the average worker."

Yes, except that it would be median worker and not the average. But we're not discussing here whether or not inequality has risen - which it has. We're talking about whether a particular implication of the neoclassical growth model - constant shares of labor and capital in GDP - holds up in real life. And the answer is "yup, at least for US where you'd expect it to".
Also, yeah, look you can say "a worker who makes too much money really isn't a worker" and then as average wages go up you'll get a falling labor share. But that's cheating.

"To use an obvious example, if health costs are rising (as they are), then overall compensation can stay approximately still but this does not mean that worker living standards are."

We're talking about SHARES here buddy! You know w*L/Y. Since w is in $ and Y is in $ that means the $ cancel out and it doesn't matter what the cost of health care is.
Ok, ok, lemme suppose that you were actually being nuanced and making a subtle point. The poor and the rich spend different proportions of their budgets on health care and perhaps that should be taken into account. But guess what, the rich spend a larger portion of their budget on health care so if anything this'd go other way (part of the reason why there's less inequality in consumption data than there is in income data).
So come on, think about this stuff a little before you say something.

"
Yes, smoothly downwards, suggesting a almost continual shift of income from labour to capital..."

See the freakin' graph. See the comment above which admits that the share has stayed smooth, i.e. constant.

Then some more blah blah blah non-sequiturs not worth responding too.

"It is a fairly constantly decreasing labour share, unlike the European one were workers seem to have made gains and then lost them.

As would be expected from a theory of distribution based on class struggle rather than mythical marginal products..."

Again, look at the freakin' picture. The labour share has stayed constant once you're looking at total labor renumeration.
And I'm sure the irony of using the words "mythical" and "class struggle" in the above sentence totally escapes you. What's next, the materialist dialectic? The impoverishment of workers? The collapse of that OTHER system, capitalism?

Anyway. To the extent there's class struggle going on here, it's more of one between unskilled workers (no college degree) and skilled workers (college degree) because that's where this rising inequality is coming from. Not from a shift from labor as a group to capital as a group.
Sorry if all them facts don't fit in with the 'scientific version of history'. But then they haven't for awhile now, have they?

YouNotSneaky! said...

To reiterate: I'm not saying (and not many are) that the marginal productivity theory of income distribution as laid out by JB Clark is true. All I'm saying is that people who insist that wages have NOTHING to do with productivity are crazy.

Gabriel M said...

People, keep calm!

On the other hand, there are circumstances under which something akin to the marginal product theory holds. There's a section in Mas-Colell about people getting out of the economy what they contribute, to paraphrase.

Anonymous said...

YNS: I've been out of action with the flu for a couple of days and the discussion has left me behind, but to return to your response above (i,e, where you report my link didn't work - sorry, must be behind uni subscription firewall).

Actually, I think the fact that the US movements are smoother proves my point, that institutional and historical circumstances make a difference. I think you would have difficulty showing that the US labour market is more perfect or less regulated than anywhere else. If nothing else, the fact that a labourer takes at least 16 years, nine months to produce suggests it's not going to be your ordinary market. As Anonymous says, 'deregulation' is better termed 're-regulation', since it's usually involved regulating to make collective bargaining more difficult, often including restrictions on union activity.

At any rate your approach of drawing a distinction between 'regulated' and 'deregulated', or even a spectrum of 'more or less regulated' is going to run into problems, because it's very important to look at what exactly was regulated and how, and also - and possibly most importantly - the macroeconomic context.

Personally, if I was to investigate why the US curve is relatively flat, I would start by looking at the unemployment experience. The US never enjoyed a period of very low unemployment as did many other developed countries in the 1950s and 1960s. The 1970s is where the labour-share trend reverses, and also when mass unemployment re-emerged in the developed world.

I wouldn't be averse to using some kind of production function as a framework for historical analysis. But not necessarily Cobb-Douglas since it's built on an assumption of constant labour and capital shares over time! A realistic approach would need to recognise that production techniques cannot be switched all that quickly, and certainly not without new investment and scrapping of old plant. This is important because it opens the possibility for profit squeeze, even if we accept the possibility of labour-saving innovation in the medium to long run.

As for the bargaining side, I really doubt real-world bargaining can be represented by game theory as if every hire and wage negotiation starts from scratch. Wage structures are historical artefacts that only change slowly over time.

Anonymous said...

Incidentally, if we are comparing US labour shares with others, I'm not sure the top line in that St Louis Fed chart is the one to use. Including social security contributions includes something that in most other countries is done through tax and welfare transfers.

Anonymous said...

(That last anonymous comment is me BTW, but none of the others are).

On individual marginal productivity, I'm really sceptical that it's possible to define. Is it true that losing a 'chief accountant' would cut output more than losing an assembly line worker?

It seems the concept is useful in justifying an existing income differential - that if the difference exists, it must be because of marginal productivity - than of explaining or predicting it. Are there examples of successful prediction? How was Wayne Gretsky's calculated?

I wonder if there is some confusion between productivity in value terms and productivity in 'purely technical' terms as YNS put it. Clearly 'purely technical' productivity advances often end up reducing the relative value of the good produced rather than increasing the reward to the immediate producers of it. These immediate producers in fact become less productive in value terms, but more productive in 'purely technical' terms.

I would say a better explanation for relative income differentials between workers involves looking at how scarce each kind of skilled worker is in the labour market. This has no necessary relationship to marginal productivity, unless you think that labour processes are infinitely variable in how you combine different skills, which is clearly rubbish.

Anyway sorry if these comments are a bit rantish, I'm still a bit woozy on medication.

Luis Enrique said...

Mike,

Is it really true that losing a chief accountant hits profits more than losing an assembly line worker?

Well I could certainly tell a story where it is so, and no doubt there's another story where the reverse is so, and I'm not sure how we'd decide which story is more realistic.

Isn't the point that presumably whoever is doing the hiring at the firm in question evidently thinks that losing the chief accountant will hurt profits more than losing one assembly line worker, which is why they are willing to pay the wages they do? [I appreciate I'm rather begging the question here]

I think you're quite right about needing to look at the scarcity of skills in the labour market, but surely wages are determined by how scare skills are and how productive those skills are - what they are worth to the employer? Whatever the supply of the skill is, employers aren't going to pay more to hire a worker than that hiring that worker contributes are they? So with some reasonable assumptions about competition in the labour market and supply of skills responding to wages, you still end up with wages still bearing a pretty tight relation to marginal product, don't you?

Although, as the start of this thread discusses, with increasing returns you can't pay everyone the marginal product, so perhaps wages being some function of the marginal product in equilibrium, is where we end up.

Sorry if this seems like either splitting hairs or trying to teach you to suck eggs - I don't mean to be doing either.

Luis

YouNotSneaky! said...

"Incidentally, if we are comparing US labour shares with others, I'm not sure the top line in that St Louis Fed chart is the one to use. Including social security contributions includes something that in most other countries is done through tax and welfare transfers."

But wouldn't that just mean that you should try and adjust the shares in the other countries?
And social security contributions = taxes, so it might be the same thing. In other words my understanding is that the author wants to look at pre-tax wages and to do that you have to put the social security taxes back in.

YouNotSneaky! said...

"Clearly 'purely technical' productivity advances often end up reducing the relative value of the good produced rather than increasing the reward to the immediate producers of it. These immediate producers in fact become less productive in value terms, but more productive in 'purely technical' terms."

Well, in the neoclassical story that would depend on the elasticity of demand for the product that the producer can now produce more off.

Oh yeah, have fun with your medication!

Anonymous said...

Hey YNS:

"Well, in the neoclassical story that would depend on the elasticity of demand for the product that the producer can now produce more off."

Well of course, but how does that change the problem?

As for whether to count social security - surely it doesn't matter which stats are changed as long as they are consistent across countries, since it's change we are interested in. Though it seems the US figure not including soc sec is more variable.

If (for other reasons) we were interested in comparing shares of current output, I would count social security and welfare payments, rather than contributions _to_ social security. (But subtracting tax and contributions of course.) On the other hand if we are interested in comparing labour costs, count contributions and taxes. As the figures are, I assume most countries' data are pre tax and transfer; that's certainly the case for Australian data I'm familiar with.

Anonymous said...

Hey Luis,

I guess the formula just looks to me like an after-the-fact rationalisation rather than an explanation. Since labour is so heterogeneous, and given that it is not so easy to relate any kind of physical productivity to value productivity, without seriously restricting and unrealistic assumptions, I don't see how value productivity can be reduced to the level of the individual.

Most commodities are produced and distributed with complex labour processes incorporating qualitatively different work. Most of the time there is no option of producing with just a little less of one kind of work. It all has to be there.

"Whatever the supply of the skill is, employers aren't going to pay more to hire a worker than that hiring that worker contributes are they?"

Yes but if the value of the product depends partly on the cost of hiring the skilled labour required to produce it, this is circular reasoning. 'What the worker contributes' in value terms is largely 'what the worker costs to hire'.

Anonymous said...

"To reiterate: I'm not saying (and not many are) that the marginal productivity theory of income distribution as laid out by JB Clark is true."

So a key aspect of neo-classical economics is not true? Wow. Yet this theory is still being used to justify inequality. Someone should tell people that the introductory books on neo-classical economics are just telling lies about it!

"All I'm saying is that people who insist that wages have NOTHING to do with productivity are crazy."

So, are you saying that the productivity of workers has fallen 3 full percentile points over the last few decades?

And I think that people are arguing that productivity gains are not automatically given to workers. They need to fight for them and that depends on organisation.

Now, if unions have been destroyed or weakened (as per the recommendations of neo-classical economics) then we would expect the gains of productivity to accumulate in the hands of capital. This would be reflected in rising inequality.

And guess what? That is what has happened since the 1970s. Now, clearly, wages are becoming unrelated to productivity as the market has become closer to neo-classical economics...

Oh, but do not worry, the neo-classical theory of distribution is not true...

ROTFL!

Iain
An Anarchist FAQ

Anonymous said...

"Non-sequitur and completely wrong to boot. You can NOT believe in marginal productivity theory of income distributions and STILL think that rising inequality is driven by changes in relative productivity levels, as I've tried to explain, and as you obviously've failed to comprehend."

Tell that to the talking heads who have been doing just that...

"Wages have been flat. Total worker compensation inclusive of non-wage renumeration hasn't."

If that rise in compensation is simply keeping up with rising medical costs, well, the fact that wages have been flat is the important factor...

"Also you seem to be confusing the median with the average. Median wages can be roughly flat, but average wages increase, keeping the share of labor in GDP constant and also resulting in increasing inequality. This isn't that hard, you know?"

I know, that is why I pointed to rising CEO wages as being a factor in how "labour" share can be flat and workers seeing their standard of living being squeezed...

"Also also. The rise in inequality has been due to the rising differential in compensation between skilled (college degree) and unskilled workers (w/o coll degree)."

James Galbraith has exploded that myth in one of his books...

"Not a shift from labor to capital. As evidence in the almost constant share (in US) which you ignored."

Except, I have not. I have indicated how an apparently constant share can equate to increased gains to capital, once you realise that CEO wages are better thought of as gains to capital. They are bosses, after all.

"We're talking about whether a particular implication of the neoclassical growth model - constant shares of labor and capital in GDP - holds up in real life. And the answer is 'yup, at least for US where you'd expect it to'."

Except, you have shown that income to labour has fallen steadily. Compensation may be steady, but if that simply means rising health costs are covered then that signifies a squeeze on labour. Equally, rising CEO income also means that labour is squeezed as their "salary" may be recorded as "labour" but it clearly should not be...

"Also, yeah, look you can say "a worker who makes too much money really isn't a worker" and then as average wages go up you'll get a falling labor share. But that's cheating."

No, I'm saying is someone is paid to be a boss then they are a boss, not a worker. Their role is what counts, not how much they get pay or whether they get paid at all...

"'Yes, smoothly downwards, suggesting a almost continual shift of income from labour to capital...'

"See the freakin' graph. See the comment above which admits that the share has stayed smooth, i.e. constant."


Smooth does not equate to constant. Smooth means "not jaggy" and the US curve is quite smooth and it is not constant. It clearly slopes downward. Hell, you even admit it is not "constant":

"You're right that the changes in the labor's share for the US are the smoothest among the OECD."

How can there be "changes" if it is "constant"? And the changes are, of course, clearly a curve which slopes downward... Feel free to deny the obvious, and the meaning of the word "smooth"....

"Again, look at the freakin' picture. The labour share has stayed constant once you're looking at total labor renumeration."

As I've noted, this fails for the obvious reason of exploding health costs and exploding inequality...

"And I'm sure the irony of using the words 'mythical' and 'class struggle' in the above sentence totally escapes you. What's next, the materialist dialectic? The impoverishment of workers? The collapse of that OTHER system, capitalism?"

Of, please! Class analysis is not limited to Marxists. Anarchists also have an analysis based on class struggle. Proudhon was discussing exploitation in 1840, long before Marx invented "materialist dialectic". Hell, even Ricardo was aware of class conflict...

And Marx distanced himself from the "impoverishment of the workers" in Capital Volume 1 -- although I do get the impression that many workers in America today are feeling pretty squeezed over the last few decades....

So it seems ironic that after 30 years of one-sided class war by the capitalists, which has been an attack on organised labour (under cover of "fighting inflation"), rising inequality, a steady decrease in labour's share by, what, 3 percentile points, and so on, that we get that "class struggle" is "mythical"!

Still, I suppose if we repeat "total compensation" enough times we can ignore all that...

And red-baiting is not a pleasant sight...

"Sorry if all them facts don't fit in with the 'scientific version of history'. But then they haven't for awhile now, have they?"

Oh, I must have ruffled some feathers... Listen, if may make life easy for you to assume that everyone who disagrees with you is a Marxist but it just is not true. Sorry, but I'm not one.

So maybe drop the red-baiting, eh?

Iain
An Anarchist FAQ

Luis Enrique said...

Thanks for the reply Mike.


As I said (begging the question), I do appreciate the circularity problem. Also, as you describe, that heterogeneous and qualitatively different jobs are complements, and the notion of individual productivity is hard if not impossible to pin down, even conceptually.

Still, it's not often the case that production involves such a high degree of complementarity that if one worker leaves, output falls to zero (I know you weren't saying that - I'm just reductio ad absurding) and I'm pretty sure most people running a business have some idea of how valuable individual workers are to them, that this idea is probably pretty grounded in reality (otherwise they'd make stupid hires and competitors could beat them), and it's got enough to do with the nature of the "production technology" to avoid the problem of circularity.