In their article, After Keynesian Macroeconomics, Thomas Sargent and Robert Lucas inquire into and critique the fundamental limitations of the Keynesian approach to Macroeconomics.  The approach to economics that arose from Keynes, was one that necessarily relied on historical data, upon which regression equations could be tested and specified.  Any Keynesian macroeconomic equation can be reduced to a regression equation; but to specify the parameters of the equation from observations of the outputs (Ps, Q, and such) requires a great deal of information about the entire system in the first place.  All estimates of these parameters (like rental rates, factor shares, or marginal products) can only be estimated from the historical data; which means that all Keynesian macro models are founded by ex ante information, and cannot adequately take into account the way that people have rational expectations about the future.  

Indeed, there are a number of related reasons why Keynesian structural macro models are not in fact structural, or theoretical, but rather empirically founded and therefore limited.  Keynes implied some very tight restrictions on macro relationships.  By identifying consumption expenditures as a function of income and current consumption only, or the demand for money as a function of interest rates only, Keynes made radical assumptions about that which money demand, and consumption expenditure was not contingent.  But saying what something is not, requires assumptions, in the same way we assume that some variables are truly exogenous, because they are not influenced by the system, but only influence the system.  This way of estimate ‘structural’ macro models must also exclude the individual from the equation.  When the model is based off of historical data and expectations of future values are based off of the previous values, then the individual acts more as a point of convergence of many historically determined macro variables, rather than an agent that makes rational, profit maximizing decisions.  This retrospective way of estimating models lead to the failure of the Keynesian economics in the face of high inflation and unemployment of the 1970’s.  Equations that always assumed a tradeoff between unemployment and inflation were systematically refuted, as both began to grow at faster and faster rates per year. 

But even more than a simple failure of Keynesian macro economics, the general approach to data and modeling was flawed.  Rather than simply adding or subtracting terms, aggregating or disaggregating, Lucas and Sargent argue that new-classical models are more sound in their assumptions, because by starting at the micro level, they can include the self-referential qualities of human agents—because agents are rational and can refer to the model from which they decide how to behave, the model itself should become part of the model, and change dynamically, rather than linearly through time.  Only at the micro level can we get to this lowest common denominator of all macro behavior; or so argues Lucas and Sargent.  

Lucas, Robert E and Sargent, Thomas J.  “After Keynesian Macroeconomics,” After the Phillips Curve: Persistence of High Inflation and High Unemployment, 1978.    

Billy Swanson

I didn’t go hunting for an article, because honestly, I was a little afraid of getting something bad/irrelevant off the web. So instead, I read “the new classical contribution to macroeconomics” by David Laidler, pp 334-358. I tried reading the Lucas-Sergant article entitled “After Keynsian Economics,” but couldn’t make too much sense of it. In this article, Laidler is arguing for the New-Keynsian/monetarist approach to macroeconomics. While recognizing that new classical had its contributions, it also had more than its share of false predictions and missteps. Therefore, Laidler argues that even though the Keynsian model failed to explain stagflation, there are other alternatives to the new-classical model; even if the contributions of new classical are inherently important to that alternative.

First, Laidler shows the reconcilability of monetarism and Keynsianism; a hybrid that is stronger than either model separately. The Keysnian “test” failed in the 1970’s when the gains in output and expenditure only corresponded to a short-term change in output, and Inflation rose at higher and higher rates, rather than shifting to a higher stable level. The monetarist contributions to the Phillips curve (such as the expectations augmentation) “pose no radical theoretical challenge to” Keynsian theories, but in fact resemble extensions of the ISLM model. From an analytic point of view, the new rendition of the Phillips curve–compliments of Friedman and Phellps—did not provide the deep explanation that Luca and the new-classical economists attempted.

There are a few major characteristics that distinguish new classical, from new-keynsian. In Keysnina macroeconomics, prices are inflexible (sticky). Therefore the supply of labor, inventories and capital must accommodate the sticky prices, causing aggregate demand shifts and more pressure on price changes. This system in one “in which prices are sticky, in which quantities change to absorb demand side shocks in the short run, and in which inflation expectations though mainly backward looking are endogenous.”(Pg. 338) In knew classical however, prices are assumed to be flexible and always the mechanism by which markets equilibrate. The individual is likely to misinterpret a money price change for a relative price change, therefore altering the quantity of goods s/he produces. Therefore, output and employment fluctuations occur because the individual is reacting to price changes. Prices have the effect on output; the opposite of the Keynsian approach. Furthermore, if prices are flexible and the market agents are not tied down by contracts and stick prices, in order for the rational maximizers to avoid unnecessary loses to income, s/he must form expectations consistent with the model. The “demand and supply schedules which determine the equilibrium structure of market prices” do not depend on “full and accurate information,” but rather the “agents perceptions of that structure.” (Pg. 339)

There are a few fundamental drawbacks to the approach of Neo-classical economics. First of all, Laidler points out that the “truthfulness” of a model should not be contingent upon weather or not it can explain the data on which it was designed; “this is just testing the logical ability of the economist.” Rather a model should be able to anticipate those events that the economist did not know about first, or define the model to address beforehand. The desperate attachment to maximization, deducibility, and the a priori non-empirical model, has hobbled the predictive abilities of neo-classical thought.

Empirically and historically, there are some demonstrations of these weaknesses. 1) Because the neo-classicals propose that the Phillips curve reflect the elasticity of the supply of labor with respect to real wages, we are able to test the hypothesis. Aggregate employment fluctuates too strongly relative to inflation, given that the changes in employment are only supposed to be a result of “is perceived nominal wage changes as reflecting real wage changes.” 2) Prices fluctuate and then output changes. The data show, however, that quantity changes seem to precede associate changes in price. 3) Because prices automatically adjust, the “economy should always be on its long-run demand for money function,” although empirically this is untenable. (Pg. 342)