Keynes "hadn't got round to it"

...all the difficulties and rigidities which go into modern Keynesian income analysis have been shunted aside. It is not my contention that these problems don’t exist, nor that they are of no significance in the long run... Robert M. Solow, "A Contribution to the Theory of Economic Growth," 1956.

...the shunting aside opened up the opportunity for real-business-cycle theorists such as Finn Kydland and Prescott to use Solow’s steady-state model... for their explanation of short-run fluctuations. Harald Hagemann, "Solow's 1956 Contribution in the Context of the Harrod-Domar Model," 2009.

Far too much has been shunted aside.

The aim of Harrod's and Domar's models, according to Hageman, was "to extend Keynes's analysis into the long run by considering under what conditions a growing economy could realize full-capacity utilization and full employment." As for "the modern type of dynamic [that is, 'growth'] theory," according to Sir Roy Harrod, Keynes "hadn't got round to it":

Mr.  Nicholas Dimsdale:  Why is there so little in the General Theory on the direction of principal determinants of economic growth (which has, of course, been very much the concern of economic policy in the post-war period), and particularly did Keynes see this as a natural extension of handling the problems of unemployment? 

Sir Roy Harrod: I think the answer is no. I think of the timetable of it. Here was Keynes giving all his brains to the General Theory which is not, though it is what you call macroeconomics, dynamic. It is a general theory of how incomes and employment are determined at a given point of time. Then, poor man, he gets ill, the war breaks out, he writes this little booklet of which Austin Robinson has spoken just now, and he is entirely immersed in the war. You see, the use of growth as a regular economic concept had hardly come in before the war:  it has all blossomed among various writers since the war. I don’t see how Keynes can have been expected to have systematic ideas on growth; his systematic ideas related to full employment. The modern type of dynamic theory about what happens through time - he just hadn't got round to it. I am sure that he would have got round to and dealt with it very well; but the timetable of his life and death did not give him an opportunity.

In a 1965 New York Times retrospective on the Keynesian "revolution," John Kenneth Galbraith credited Leon Keyserling as a "tireless exponent of the [Keynesian] ideas. And he saw at an early stage the importance of enlarging them to embrace not only the prevention of depression but the maintenance of an adequate rate of economic expansion." In a subsequent interview, Keyserling was less bashful about the magnitude of his contribution.

Coming over to economic growth in particular, everybody talks about the influence of Keynesian economics. The Keynesian economics is really a static economics. It doesn't deal with economic growth at all. Furthermore, it was developed at a time of worldwide depression. Even Ken Galbraith, in an article in the New York Times a couple of years ago, when he was talking about the influence of Keynesian economics, mentioned me specifically as the one who had introduced the fundamental new factor of the dynamics of economic growth.

Keyserling was a member of President Truman's Council of Economic Advisers, became acting chairman in 1949.and chairman in 1950. His influence can be seen in the evolution of "economic growth" in successive volumes of the Economic Report of the President. In the January 1947 annual report, the term "growth" did not occur. In the midyear report, "economic growth" appeared once. The January 1948 volume contained 17 references directly to economic growth and featured a cheerleading section titled "Our Ability to Grow."

All of this alleged "extending" and "enlarging" of Keynes's analysis had little to do with Keynes's analysis -- other than shunting it aside -- and nothing to do with Keynes's own views about the long term problem, as outlined in his 1943 Treasury memorandum, "The Long-Term Problem of Full Employment" and re-iterated in a 1945 letter to T.S. Eliot:

Not long ago I was at a Conference where the Australians urged that all the Powers in the world should sign an international compact in which each undertook to maintain full employment in their own country. I objected on the ground that this was promising to be 'not only good but clever'. Civis, like the Australians, takes exactly the opposite line. He thinks that we can reach the goal by promising to be 'not so much clever as good '. 

It may turn out, I suppose, that vested interests and personal selfishness may stand in the way. But the main task is producing first the intellectual conviction and then intellectually to devise the means. Insufficiency of cleverness, not of goodness, is the main trouble. And even resistance to change as such may have many motives besides selfishness. 

That is the first, ought-to-be-obvious, not-very-fundamental point. Next the full employment policy by means of investment is only one particular application of an intellectual theorem. You can produce the result just as well by consuming more or working less. Personally I regard the investment policy as first aid . In U.S. it almost certainly will not do the trick. Less work is the ultimate solution (a 35 hour week in U.S. would do the trick now). How you mix up the three ingredients of a cure is a matter of taste and experience, i.e. of morals and knowledge.

It would be anachronistic to fault growth theorist for ignoring Keynes views on the long term problem of full employment. Presumably, Keynes's memorandum and his letter to T.S. Eliot were not available to early "dynamic" theorists. They were published in 1980 in volume 27 of Keynes's Collected Writings. One might wonder, though, why modern growth theorists have subsequently shown no interest whatsoever in re-evaluating their theories in light of those documents.

More remarkable is the complete lack of connection between growth theory and growth as a policy slogan. No, this is not the difference between map and territory. The theoretical models are maps of an abstract territory in which some kinds of features have the same names as the kinds of features on the actual territory -- "labor", "income", "output," "capacity utilization" -- but those labels remain undefined in any way that would correspond to the real features with the same name.

At the other end of the theory/sloganeering spectrum -- in the targeting of increases in GNP/GDP -- the features are also not defined in a way that would enable a consistent and durable system of measurement. Simon Kuznets's 1947 essay on "Measurement of Economic Growth" outlined many of the biases and obstacles in the way of defining and measuring growth." In a subsequent review of the Commerce Department\s National Income and Product Accounts, Kuznets argued that problems of duplication and of ignoring non-market production had not been overcome. In its reply to Kuznets, the department's economists pointed out it would not have been feasible to meet several of his standards. They had a point. But non-feasibility of the alternative does not in itself affirm the adequacy of the status quo.

In conclusion, Keynes's analysis was not "extended" by modern dynamic growth theory nor was it "enlarged" by Leon Keyserling's siphoning off and his sloganeering growthmanship. The analysis was -- in Solow's apt phrase -- shunted aside for a tautological model, just as the policy goal of full employment was shunted aside for the indistinct slogan of "growth."

Who needs a Keynesian policy goal of full employment anyway when the neoclassical model can simply assume full employment?

Napoleon Solow and the Phantom Mechanism

I would like to say why I think that the Doomsday Models are bad science and therefore bad guides to public policy. ... The basic assumption is that stocks of things like the world’s natural resources and the waste-disposal capacity of the environment are finite, that the world economy tends to consume the stock at an increasing rate (through the mining of minerals and the production of goods), and that there are no built-in mechanisms by which approaching exhaustion tends to turn off consumption gradually and in advance. You hardly need a giant computer to tell you that a system with those behavior rules is going to bounce off its ceiling and collapse to a low level. -- Robert M. Solow, 1973

Sandwichman is agnostic on the built-in mechanism fable.

On the one hand, Solow's "built-in mechanism" is a metaphor -- a depiction -- and of course there is no "mechanism" strictly speaking, just as God is not an old man with a long, white beard. There are instead more or less spontaneous reflexes of economic actors that in the aggregate have observable effects. Such reflexes, however, are multitude. The probability of all these reflexes co-ordinating themselves spontaneously and independently -- without help from a Maxwellian demon, Walrasian auctioneer or Invisible Hand -- to produce a conjectured effect far in the future is infinitesimal.

Suppose someone sits down where you are sitting right now and announces to me that he is Napoleon Bonaparte. The last thing I want to do with him is to get involved in a technical discussion on cavalry tactics at the Battle of Austerlitz. If I do that, I’m getting tacitly drawn into the game that he is Napoleon Bonaparte. -- Robert M. Solow

Aside from the multiplicity of so-called mechanisms, there is the slight inconvenience that a homeostatic regulator itself consumes energy to do its work (also known as 'transaction costs'). The more vast and complex the organism being regulated, the more energy the regulator will need to consume. When what is being regulated is the consumption of energy, a contradiction emerges: progressively more energy needs to be consumed to reduce the consumption of energy. There is thus a ceiling on vastness and complexity and a floor on reducing consumption.

It is not the finiteness of resource stocks, but the fragility of self organized natural cycles that we have to fear. Unfortunately, the services provided by these cycles are part of the global commons. They are priceless, yet ‘free’. Markets play no role in the allocation of these resources. There is no built-in mechanism to ensure that supply will grow to meet demand. Indeed, there is every chance that the supply of environmental services will dwindle in the coming decades as the demand, generated by population growth and economic growth, grows exponentially. -- Robert U. Ayres, 1998

Ayres highlighted another fly in the built-in mechanism ointment. To be fair, Solow did acknowledge the externality flaw in the price system. Fixing the flaw, he assured, would be simple and virtually painless:

The flaw can be corrected, either by the simple expedient of regulating the discharge of wastes to the environment by direct control or by the slightly more complicated device of charging special prices — user taxes — to those who dispose of wastes in air or water.  

What stands between us and a decent environment is not the curse of industrialization, not an unbearable burden of cost, but just the need to organize ourselves consciously to do some simple and knowable things.

Now that we know how that organizing ourselves consciously has turned out, couldn't we please have a built-in mechanism to do that part too? What if there is a built-in mechanism in the price and profit system that militates against the social capacity "to organize ourselves consciously to do some simple and knowable things"?

Finally, even if there was a built-in self-stabilizing market mechanism, it's interaction with natural systems may not lead to the gradual, advance adjustment that Solow conjectured:

We identify and prove that the interaction between a stable natural system and a self-stabilizing market mechanism can lead to cyclical or even chaotic behaviour. A built-in self-stabilizing market mechanism will not always serve the function of stabilization. Under certain conditions, it may increase the amplitude of fluctuations and have the effect of destabilization, as demonstrated in this paper. Therefore, by incorporating a self-stabilizing market mechanism, this model yields a result that contradicts Solow's (1973) conjecture that the market mechanism will have the effect of smoothing the time path of the world economy. -- Hans Gottinger, 1998

Mathiness, Growth and Increasing Returns

The following was originally posted on Ecological Headstand in October, 2012 under the title, Endogenous Growth Theory and Ecological Unequal Exchange: linkage, displacement and deflection of 'diminishing returns'. Paul Romer's rant on mathiness has provoked a response from Lars Syll regarding the issue of increasing returns to scale, which I discussed in this post.

What the late Stephen G. Bunker wrote bears repeating:

The crucial difference between production and extraction is that the dynamics of scale in extractive economies function inversely to the dynamics of scale in the productive economies to which world trade connects them.

Rather than repeat what Bunker wrote, though, I'm going to cite, later, a longer piece by Nicholas Kaldor from his 1985 Hicks Lecture that makes a somewhat similar point. But first, I want to present some background on an old debate and a 'new' theory.

In December 1926, The Economic Journal published an article by Piero Sraffa dealing with "that difficult branch of economic theory" -- the theory of increasing returns. Over the next five years it published responses from Cecil Pigou, G. F. Shove, Lionel Robbins and Allyn Young and, in March 1930, a symposium on the topic by D. H. Robertson, Sraffa and Shrove.

Almost exactly 60 years later, in October 1986, The Journal of Political Economy, published Paul D. Romer's "Increasing Returns and Long-Run Growth," an important contribution to so-called New Growth Theory. Romer took his cue explicitly from Young's 1928 paper, "Increasing Returns and Economic Progress" and although he mentioned the precedents of Adam Smith's pin factory and Alfred Marshall's distinction between internal and external economies, he skipped over the rest of the debate in which Young's contribution had appeared.

Critics have argued that Romer's usage of increasing returns and external economies is not faithful to Young's formulation, in that it "overlooked Young's emphasis on the reciprocal relations between the division of labor and the feed-back into aggregate demand as a requirement for growth," "neglected Young's categorical rejection of the usefulness of Walrasian general equilibrium models" and wrested "Marshall's microeconomic concepts of internal and external economies out of his theory of value and price to serve as a basis for amending constant return production functions to exhibit increasing returns for the macroeconomy" (Rima 2004, 181-182).

My concern here is with a more conspicuous omission in Romer's analysis -- the distinction between increasing returns as characteristic of manufacturing and diminishing returns as dominant in agriculture and extractive industries (Young 1928, 528-529). The words "agriculture," "land" and "rent" do not appear in Romer's 1986 article. When Romer mentions diminishing returns, it is only in the context of research activity or the limiting assumptions of classical conventional growth models. But diminishing returns is a specific limitation, not a generality that can be indiscriminately "offset" by increasing returns. In a lecture given at Harvard in 1974, "What is Wrong with Economic Theory," Kaldor explained that "it is the income of the agricultural sector, (given the "terms of trade") that really determines the level and the rate of growth of industrial production, according to the formula:"

Or, in prose, economic growth depends on either a relative reduction in the income of agriculture or increased demand from agriculture for industrial products. And, of course, increased demand from agriculture implies increased agricultural production, which at some point confronts the problem of diminishing returns. In his 1985 Hicks Lecture, Kaldor explained the inverse dynamics of scale between industrial and agricultural areas, parenthetically, in terms of the "differing manner of operation of perfect and imperfect competition":

The basic requirement of continued economic growth is that the various complementary sectors expand in due relationship with each other -- that is to say that general expansion is not held up by "bottlenecks" in key sectors. However, in the course of time, under the influence of technical progress, both of the natural-resource saving and labour-saving kind, the requirements of expansion may become considerably modified. In the manufacturing sector which becomes more important as real incomes rise, there are considerable economies of scale, as a result of which manufacturing activities are subject to a "polarization process" -- they are likely to develop in a few successful centres, and their success has an inhibiting effect on similar developments in other areas. The realisation of these economies of scale normally requires also that numerous processes of production which are related to each other are carried out in close geographical proximity.

As a result different regions experience unequal rates of growth of output and of population. The industrial areas experience a growing demand for labour which may involve immigration from other areas once their own surplus labour is exhausted. Technological development in primary production on the other hand, tends to be more labour-saving than land-saving, so that the growth of output may go hand in hand with a falling demand for labour; and though output per head may grow fast in real terms, the level of wages will tend to remain low (and may even be falling) as a result of a growing surplus population. Since labour cost per unit of output is the most important factor in determining selling prices (at any rate under competitive conditions) the low wages prevailing, in terms of industrial products, will mean that the terms of trade will move unfavourably to primary producers, which may be the main factor, along with the low coefficient of labour utilisation, for their state of "under-development" characterised by low standards of living. The important contrast -- which I regard as a major factor in the growing inequality of incomes between rich and poor countries -- resides in the fact that the benefit of labour saving technical progress in the primary sector tends to get passed on to the consumers in the secondary sector in lower prices, whereas in the industrial sector its benefits are retained within the sector through higher wages and profits. (The main reason for this difference lies in the differing manner of operation of perfect and imperfect competition.)

Kaldor's parenthetical explanation suggests more than it reveals. Sraffa's 1926 discussion is the key to unpacking why Kaldor specifies perfect competition as characteristic of primary production and associates imperfect competition with manufacturing industry. The key determinants, in that view, are the shapes of the firms' supply curves (increasing or diminishing returns) and the nature of external economies.

Externality and Ecological Overshoot

Marshall's notion of "external economies" has gone through a series of modifications to become today's "externalities." Pigou extended the concept beyond Marshall's industrial agglomerations and distinguished between “incidental uncharged disservices” and "incidental uncompensated services." The former became known as negative externalities and the latter as positive externalities, although typically it is the negative environmental externalities that are referred to simply as externalities. There is a seeming but misleading symmetry to the two terms and a similarly illusory quality of reciprocity within each of them. When a disservice is uncharged or a service is uncompensated there is a presumption that there might otherwise have been a "whom" to charge or to compensate and that the missing invoice could have been denominated in currency. In other words, the charging and compensating would appear to be a financial transaction between two parties, both of whom must be assumed to be legal persons. In reality, the services or disservices performed may (or may not!) be extremely indirect and the parties affected incredibly diffuse, both in space and time. Mundane examples of factory soot and laundry hanging out to dry may be more mystification than illumination.

In the case of the external economies of increasing returns and diminishing returns, respectively, although they function inversely to one another it is a double inversion that ultimately produces parallel incentives to firms in manufacturing and agricultural or extractive industries. In other words, while firms in the manufacturing center are routinely considered to be the beneficiaries of external economies that generate increasing returns in the sense that they receive uncompensated services, firms in the extractive periphery may also benefit from the externalization of diminishing returns in that they are able to avoid being charged for the environmental disservices they inflict. In effect, the cost of diminishing returns is first displaced to poor regions where it is then deflected onto society and the environment. Unequal exchange thus takes place, that is to say, in the global external economies, "behind the back", so to speak, of formal monetary transactions.