Recent Posts

Sunday, April 24, 2016

Anwar Shaikh On Micro And Macro Economics

As I noted in my discussion of Anwar Shaikh's new book Capitalism: Competition, Conflict, Crises, one of the most interesting parts of the work was his discussion of how microeconomic behaviour is reflected in aggregate performance. There are fundamental constraints upon individual behaviour that result in the same aggregate patterns at the macro level. Using the currently popular buzzword, macro behaviour is emergent,  and cannot be predicted from individual behaviour with reference to the aggregate system. In my view, this is useful for macro analysts, as it means we do not waste our time obsessing about the lessons of microeconomics. For mainstream ("neoclassical") economics, this line of thinking is disastrous, as it means that entire "microfoundations project" has all the usefulness of a foundation made of sand.

Why Aggregate Behaviour Matters

As discussed in "Finding The Solution In A Simple SFC Model," the macro accounting relations are not enough to pin down the solution of a Stock-Flow Consistent (SFC) model. We need to also model how the various sectors set spending flows. Standard SFC models use fairly basic behavioural functions to pin down the solution, whereas mainstream modellers insist upon extremely complex "micro-founded" behavioural rules (behaviour is the result of optimising decisions of households). (The complexity of those optimisations means that the actual solutions used by researchers tend to be approximations.)

Aggregates And Micro Behaviour

Chapter 3 of Capitalism is focussed on the discussion of micro and macro behaviour. The chapter summarises research published elsewhere, although he has some simulation code available (on request).

If we look at the example of demand curves, he is able to demonstrate that they are of a familiar downward sloping form as a result of two effects.
  1. Households face an income constraint.
  2. Engel's law: households will purchase a (socially-defined) minimum of "necessary" goods, and then increasing amounts of other (luxury) as their income increases. (This is a robust empirically observation.)
(To explain what is meant by "socially defined" in the last point, a machine for washing clothes would be considered a necessary purchase for a home owner in 2016 in North America, whereas it would have been a luxury in those societies in the 1930s.)

Shaikh describes four different behaviour patterns for individual households -- including neoclassical homogeneous agents with Cobb-Douglas utility functions -- and they all end up with an essentially identical aggregate demand function.

In other words, the micro behavioural rules did not matter for the aggregate results. So long as the SFC model approximation is reasonable, it would get the same results as a neoclassical model with highly complex behavioural rules (assuming the neoclassical researchers calculated the results properly).

(Production functions are another area where the relationship between micro and macro behaviour has been pinned down. This was the subject of a 1974 paper by Anwar Shaikh, and is discussed in Appendix 1.1 of Monetary Economics by Godley and Lavoie.)

A Physical Analogy

Although I am not a huge fan of analogies from physics, Anwar Shaikh usefully discusses the ideal gas law (PV = NRT, if that means anything to you). In summary, if you compress a gas, or heat it up, its pressure increases.

When initially proposed, the aggregate behaviour was believed to be the result of atoms bouncing around like little billiard balls. However, with the advent of quantum physics, we no longer think of atoms as being particles at a known position and velocity. That said, even the more complex quantum atomic behaviour ends up with the same aggregate result.

(At one point of my life, I took a course on quantum thermodynamics. This supposedly would have made it easier to understand the operation of silicon chips, but at the cost of being mystified as to how my heat pump works. If there are any electrical engineering faculty members reading this, please have your undergraduate students learn sensible thermodynamics as taught by mechanical engineers. Thanks.)

Characteristics Of Rigorous Aggregate Analysis

 Shaikh argues that there are five characteristics of "rigorous aggregate analysis." (Found in the section "Methodology For Economic Analysis" of Chapter 3; I have the ebook version and cannot offer a page reference.)
  1. "It should be rooted in some theory of the relevant factors at the micro level."
  2. "It should allow for the fact that only a few of these factors may be relevant at the macro level."
  3. The aggregate functional form may be quite different than the micro functional form (which implies that there is no such thing as a "representative agent").
  4. There are "many micro foundations that are consistent with any given aggregate pattern." 
  5. The tendency to return to equilibrium is a hypothesis, and "the existence, speed, and manner of operation must be explicitly addressed."
He argues that the "classical" economists have addressed this issue in a much more acceptable fashion than the neoclassicals. He cites Keynes, Kalecki, and Friedman as having discussing what affects individual behaviour, then transitioning to the aggregate patterns. The aggregate functional forms are different than the micro form.

Good News For Macro Analysts

For academic economists, the difficulty in the transition from micro to macro behaviour makes the creation of a Grand Unified Theory of economics difficult. It will always be possible to call into question the functional forms used in macro models.

For those who are more interested in the applications of macroeconomics, this may be somewhat of a blessing. Although it is impossible to prove that a particular set of behavioural laws is correct, it is also impossible to prove that they are wrong. This allows us to develop macro models without losing sleep about what is supposed to be happening at the micro level. Instead, we can just focus on the question of whether the aggregate behaviour resembles real world data sets.

Whither Neoclassical Economics?

Neoclassical economists justify the exclusion of other schools of thought from the top academic journals on the basis that they are not "microfounded." If these microfoundations are fraudulent, one realises that this exclusion was just a raw academic power grab. Unfortunately, it seems unlikely that the mainstream will admit that entire decades worth of research are essentially worthless.

(c) Brian Romanchuk 2016


  1. Brian,

    Is Shaikh's chapter 3 analysis based on any papers he references?


    1. The book is fairly heavily referenced, so there's a lot of citations in Chapter 3. He works out the consumption function example in the book, which May or may not have been published in that exact form. (He cites articles with similar work.)

      Since I do not have access to academic journals, I am not really in a position to track down his references.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.