P8 Keynesian Complexity

This continues the sequence of posts on re-reading Keynes. The fundamental point about the labor market which is made in Chapter 2 is that the micro level negotiations on wages between firms and laborers do not determine the real wage in the macro-economy. Before explaining this point in detail, we want to show how it is just a special case of the general idea that the economy is a complex system which cannot be understood by looking at simple sub-systems.

The idea of complex systems is beautifully illustrated by the parable of the blind men and the elephant. Each one understood correctly and accurately one small part of the big picture. When we don’t understand the system as a whole, the descriptions of subsystems appear conflicting and contradictory. Once we have an understanding of the complex system, we can assemble the partial insights into a coherent whole. The main contribution of Keynes can be understood as an attempt to describe the economy as a complex system. Unfortunately, most of his followers were blind to the main insights of Keynes. Accordingly, there have been many different interpretations of Keynes; some followers saw the trunk, others the legs, and yet others the tail of the system that Keynes was describing. But no one appears to have understood the fundamental insights of Keynesian complexity: the system as whole does not act as a simple aggregate of the actions of the individual agents within the system. Pre-Keynesian macroeconomics was based centrally on the misunderstanding that the macroeconomy can be understood by scaling up the microeconomic behaviors of individual agents. While Keynes forcefully rejected this thesis, and created a complex system view of the macroeconomy, simple-minded followers failed to understand complexity, and went back to the pre-Keynesian views.

Blind Men Thesis:  20th Century Economists as a whole failed to understand the central insight of Keynes that the economy is a complex system. This means that what happens in the macroeconomy cannot be reduced to the microeconomic behavior of the agents.

Proof of the Blind Men Thesis: When we simplify an economy by creating a representative agent, we have thrown out the baby with the bath water in the process of simplification. A complex economy cannot be understood by reducing to a single agent, and then replicating this one agent. Similarly, aggregation of micro-behavior cannot be done to get macro-behavior. The very fact that this is widely done exhibits a failure to understand the central thesis and the fundamental contribution of Keynes.  We now list some commonly used approaches to understanding the macro-economy which illustrate the widespread failure to understand complexity:

1. The Sonnenschein–Mantel–Debreu theorem illustrates complexity of the demand function. The properties of the micro-agents are not replicated at the macro level when we aggregate the demand functions.  Frank Hahnregarded the theorem as the most dangerous critique against the micro-founded mainstream economics. Nonetheless, modern economics textbooks blithely ignore complexity and continue to assume that macro level aggregate demand satisfies microeconomic assumptions.
2. The widely used DSGE model with a representative agent reflects a failure to understand complexity. In his testimony to the Congress on the failure of economics, Solow understood this problem with the DSGE model and explained that “One important consequence of this “representative agent” assumption is that there are no conflicts of interest, no incompatible expectations, no deceptions.” Complex phenomena like unemployment and irrational exuberance cannot be captured in such models.
3. Solow himself fails to understand complexity. The Solow growth model which aggregates Labor and Capital, illustrates exactly the same problem. If the complex interactions between laborers and capitalists create macro-phenomena which are not reflected in the micro-economy, then aggregation will miss crucial elements of the system behavior and fail to capture the process of economic growth. In the same testimony cited above, Solow writes that aggregation is a good first approximation. The main point of complexity is that this is not true – in a complex system aggregation of micro behavior is a hopelessly bad approximation to system-wide behavior.
4. The Cambridge Capital Controversy is another illustration of complexity. Can we add up all the capital in the hands of all the capitalists and pretend that there is one aggregated quantity Capital, to create the Solow growth model? In a complex system, obviously not. Joan Robinson made the case that capital cannot be aggregated, but not on complex system grounds. Because she was right, Solow eventually conceded the point, and called aggregate capital a “Parable”. Modern textbooks blithely ignore this controversy, and continue to describe an economy using aggregates of capital and labor, directly in conflict with the insight that the economy is a complex system.
5. The widely used textbook examples of Supply and Demand Analysis in a single market is another failure to understand complexity. See Saglam and Zaman: The Conflict between General Equilibrium and the Marshallian Cross for an explanation of how the complex system captured by general equilibrium conflicts with insights obtained by looking at a simplified subsystem within one market.
6. The fundamental point about the labor market (which will be discussed in detail in a later post) being made by Keynes in Chapter 2 is a complex system point. We cannot understand the behavior of the labor market as a whole, by aggregating our understanding of negotiations between firms and laborers at the micro-level.In particular, supply and demand of labor interact, since as more labor is hired, aggregate income and aggregated demand increases. Thus the labor market cannot be studied in isolation from the goods market, and there are strong emergent effects.
7. Many other points made by Keynes depend essentially on complexity (for more examples, see John Foster (2006) Why Is Economics Not a Complex Systems Science?,Journal of Economic Issues, 40:4, 1069-1091). Among them, the central role of money in an economy emerges from complexity. If we can aggregate all the money, and similarly aggregate prices, neutrality may hold. However, if it is the interactions between agents which are of crucial importance, then exactly the same amount of aggregate money, distributed in different ways, will change system behavior. Modern macroeconomists fail to understand this, continuing to believe in the long run neutrality of money, even though Keynes explicitly denied it and provide clear (but complex) arguments for his position.

Ironically, failure to understand Keynes led to dismissal and contempt “Paul Samuelson felt he could say that “it is remarkable that so active a brain would have failed to make any contribution to economic theory . ..” (cited in John Foster 2006). Because Samuelson could not understand the complexity of Keynesian theory, he wrote that: “[The General Theory] is a badly written book, poorly organized; any layman who, beguiled by the author’s previous reputation, bought the book was cheated of his 5 shillings. It is not well suited for classroom use. It is arrogant, bad-tempered, polemical, and not overly generous in its acknowledgements. It abounds with mares’ nests and confusions: involuntary unemployment, wage units, the equality of savings and investment, the timing of the multiplier, interactions of marginal efficiency upon the rate of interest, forced savings, own rates of interest, and many others. In it the Keynesian system stands out indistinctly, as if the author were hardly aware of its existence or cognizant of its properties; and certainly he is at his worst when expounding its relations to its predecessors.”

Samuelson’s arrogance in believing that he understood the Keynesian system better than Keynes created the biggest barrier to understanding Keynes for 20th Century economists. Because of his stature, he became the authorized interpreter of Keynes, and very few went back to original writings to try to understand them. Those who did also failed to come to grips with complexity, and as a result, it is impossible to count the variety of interpretations of Keynes — see for example, Backhouse and Bateman. The Keynesian elephant has a huge number of parts, it seems.

The current search for micro-foundations for macro is another expression of ignorance about Keynesian complexity. Keynes macroeconomic already has micro-foundations, but the macro does not replicate the micro. Failure to understand complexity leads to the complaint that since the macroeconomic system cannot be reduced to the behaviors of representative agent, it is not micro-founded.

This psycho-history is important because, as Keynes wrote: “The difficulty lies not in the new ideas, but in escaping the old.”   To the extent that we believe we already know what Keynes said, through his bastardized interpretation at the hands of Samuelson, his contribution will escape our understanding. Many recognized the illegitimacy of Samuleson’s interpretation, but provided equally wrong alternatives, which failed to capture the central Keynesian insight of complexity. The over-confidence of Krugman that we don’t really need to understand Keynes since we can independently arrive at the truth does not seem empirically justified in light of the numerous failures on this front. Perhaps it is true, as some complexity theorists have said, that human minds are not built to understand complexity. However, today we have the computational tools necessary to model, analyze and understand complex systems. Thus major progress is possible, with the help of agent based models and complexity theory.