Micro-Foundations for Keynesian Economics

  Lecture 8A of Advanced Macroeconomics   — Outline below covers the first 17m of the lecture linked below at bottom of post.

1. EXCESS Savings reduce Effective Demand, Normal Savings Do Not

It seems clear that shortfalls in aggregate demand can lead to recessions, but only in presence of fixed prices. Furthermore, normal levels of savings cannot create such shortfalls – an abnormally high level of savings is required. This is because of factors discussed in “ The Subtleties of Effective Demand ”. Basically, if a normal level of savings is reduced from Aggregate Demand, this money is saved and goes on to period T+1. Similarly, the savings of last period T-1, is going to come into the present period T. This will exactly offset the shortfall in Aggregate Demand created by the savings. However, this will not happen if for some reason there is EXCESS savings, over and above normal levels. This excess S(T) > S*  will be not be compensated fully by S(T-1)=S*, where S* is the normal level of savings.

What could lead to abnormally high savings? It appears that debt can force people to earn money to pay off debt, reducing aggregate demand. Thus it appears that the Keynesian mechanism for creating unemployment as an equilibrium phenomenon relies on debt – without explicit mention. Once the role of debt is highlighted as the source of shortfall in aggregate demand, we examine in detail Fisher’s theory of Debt-Deflation, which never received the prominence that Keynes did. In the recent times, this theory has been resurrected, and is solidly backed by empirical evidence. See: Fisher-Minsky-Koo theory of debt-deflation.

2. Empirical Evidence favors Keynes Conjectures

Examination of the empirical evidence around the Great Depression period supports Keynesian ideas about Unemployment. The paper “Real wages, working time, and the Great Depression” by Hart & Roberts, 2010 shows that as unemployment increased, wages and prices both declined while real wage remained fairly constant (see  Lecture 7  for details). This shows that wage rigidity is NOT the explanation for high unemployment during the Great Depression. Furthermore, it supports the Keynesian idea that cutting nominal wages will not reduce real wages. Also the data refutes the RBC models. If random shocks reduce supply of labor increasing unemployment then the Marginal Product of Labor should increase so that nominal wages should be counter-cyclical. In fact, observed nominal wages are pro-cyclical.

Even the the broad macro perspective of Keynes is supported, explaining his theory of effective demand requires articulation of the micro-structure of the model. Post-Keynesians have strongly resisted/rejected this neoclassical demand from micro-structure, because the neoclassicals accept only one kind of micro-structure – with optimizing agents, equilibrium outcomes, and no uncertainty. Since this type of micro-structure is patently absurd, nothing is gained by knowing whether or not we can construct maximization/equilibrium/known future kind of model which supports Keynesian conclusions. The New Keynesians attempt to do this by providing neoclassical foundations which lead to Keynesian conclusions. Although this can be done, it is irrelevant. As clarified in Models and Reality, from the critical realist methodological perspective, our models should be believable simplifications of reality. While we reject  Friedman’s Folly of accepting wildly unrealistic models , we do argue that we need to provide realistic micro-foundations, based on reasonable descriptions of human behavior, structure of markets, role of money, and genuine uncertainty. This is because, as we will see, the Keynesian macro-phenomena, are very sensitive to these details.

3.  Microstructure Matters for Effective Demand.

One of the crucial insights from the behavioral/experimental economics literature is that small and apparently irrelevant details of experimental setup can make major differences in outcomes. The same is true of Keynesian Macroeconomic conclusions. Depending on how we set up the micro-details, the aggregate outcome can vary greatly (see also,  Lecture 7: Micro-structure Matters ). Below we discuss four different micro-structures, each of which leads to different results at the aggregate, macroeconomic level. We will show that four different micro-structures lead to four different outcomes.

4.  Four Micro-Structures: Four Outcomes

We consider the same model we have been working with (see:  Simple Model Explains Complex Keynesian Concepts ). There are 10 Landlords (LL) each with 5 Acres of Land and simple fixed proportion production function 1 Acre plus 1 Laborer produces 10 units of corn. Suppose the Money wage is 100 PKR per laborer, and there are 40 laborers.

A: Several Micro-Structures leading to subsistence wages:

The situation is one of shortage of labor, so in principle laborers should have power. But suppose market structure is as follows. At the beginning of the period, wages are paid in NOMINAL terms. Then production takes place. All product belongs to the landlords. All 40 laborers will be hired, and will receive pay of PKR 4000. At this point the market for corn is not open so the price is not known. Now, the landlords have all the output, and they have two options – they can market the corn to earn money, or they can self-consume the corn. Suppose this is a one-period economy, and consumption of corn is subject to diminishing marginal utility, but no satiation. Then the landlords will just eat all the corn, and will not put any on the market. To get a more realistic outcome, suppose there is satiation. There is a maximum limit to consumption of landlords. Then Landlord will market the surplus which remains after their maximum consumption. Since this is a one period economy, and money is worthless unless used to purchase corn, however much the landlords place on the market, all of the quantity will be purchased by the laborers – all PKR 4000 will be spent, and price will adjust to ensure that all marketed surplus is sold. It is easily possible for Landlords to market just enough corn to keep laborers at subsistence level, while extracting all the money they have been paid in wages.

Other ways to get the same result involve preventing the landlords from eating corn – corn is coarse food for peasants while the landlords only eat imported food. The landlords can export food at the foreign price, and use revenue to purchase foreign food. They can retain some food to sell on domestic market. However little they retain, the laborers will spend all their money on purchasing this food, so they can extract the entire amount paid as wages. Again, laborers can end up with subsistence wages.

Note that in all these models, nominal wages are determined at the beginning of the period and real wages depend on landlord decisions about how much product to put on the market at the end of the period.  This is how a monetary economy works.

B: Landlords get Subsistence Profits

Suppose we eliminate money from the economy and determine wages in terms of units of corn directly. Suppose landlords compete with each other to hire labor, offering them some share of the output (10 units of corn per laborer) as the wage. Since there is scarcity of labor and surplus of land, competition will lead to landlords ending up with the smallest possible share of corn. For example, in an experimental setup (using oral double auction type bargaining), the landlords will end up paying 9 units of corn in wages, and keeping one unit of corn in profits. Any landlord offering less – 8 units of corn – can be outbid for the laborer by some landlord who offers a higher wage.

C: Cobb-Douglas Production Function

Instead of fixed proportions, assume a Cobb-Douglass production function. Then we can set up a market structure such that Labor and Land both receive their marginal products, as per neoclassical theory. For example, Oral Double Auction pricing, where Landlords compete for laborers and laborers compete with each other for work, will lead to this result. However, a Posted Prices setup, where Landlords announce wages, can lead to capture of surplus by landlords.

D: Social Norms of Fairness:

We can also imagine other types of societies with other types of cultural norms. For example in a society with strong norms of fairness and equality, we could have a sharecropping situation. Each landlord hires four laborers. One “fair” solution would be to split 50-50 so that each laborer gets 5 units of corn, while each landlord gets 20 – 5 each from the 4 laborers he hires. An alternative fairs solution would be that each landlord gets 20% of the produce from each laborers. In this case, all landlords and laborers would end up with equal amounts, 8 units of corn.

5.  Conclusions:

We can create a vast variety of different models with different structures, and get very diverse results. For example, if the landlord hires laborers and owns the produce, this results in different outcomes from situations where landlord rents the land to the laborer and the laborer owns the produce. Depending on how money is used, and the sequence in which payments are made, and production takes place, and how the market for corn is organized after production, many different outcomes can result.

From this we conclude that the demand for Micro-Foundations for Keynesian Economics is legitimate, because how we specify the micro-foundations makes a huge difference. Keynes does not specify Market Structure, Technology, Social Norms, Institutions, Social Classes, Sectors of Economy, and many other factors of importance. All of these factors make a difference to the outcome.  Depending on HOW we specify these, we can get to different understandings and explanations of Effective Demand. Whether or not unemployment can exist and persist will depend on these factors which Keynes did not specify.

End of First Part of Lecture. To be continued – Next part discusses how one can get to Keynesian type unemployment equilibria. Video below contains the full lecture, covering this first part, as well and many other topics. The above post covers the first 17 minutes of the lecture.




Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: