Macroeconomic Theory

Spring 2011

Jeffrey Parker, Reed College

April 8 paper of the week

Assigned paper

Reading suggestions

• Romer discusses this paper briefly on pages 282-83.
• Don't be too worried about the simplicity of the AD curve. In logs, Lucas assumes that it has the form y = x - P, where x is nominal GDP and is assumed to be constant along the AD curve. This has become a pretty standard simplification in the literature. It assumes that the effects on AD from the IS/LM model can be captured by the single variable x and that the elasticity of aggregate demand is negative one. This is obviously a simplification, but may not do gross injustice to the empirical facts.
• The solution to his theoretical model on page 328-29 is broadly similar to Romer's imperfect information model of Section 6.2. Lucas's equation (13) is comparable to Romer's (6.29).
• The key column in Lucas's Table 1 is the last one, which measures each country's variance of aggregate demand (measured by nominal GDP variance). This is analogous to V_m in Romer's equation (6.29).

Questions for analysis

1. Based on Lucas's equation (13) and Romer's (6.29), what relationship would we expect between the elasticity of output (aggregate supply) with respect to AD fluctuations and the variances of aggregate demand? What is the intuition behind this result?
2. Comparing Tables 1 and 2, what relationship does Lucas observe? How strong is this relationship?
3. According to equations (13) and (6.29), the variance of "relative shocks" (V_z in Romer, τ² in Lucas) should also affect the elasticity of aggregate supply. What does Lucas assume about this variance in his test? Is this assumption reasonable? Is it ideal? If you had access to the data, how might you try to improve on it?
4. Looking at Lucas's Table 1, there is a substantial positive correlation between the average level of inflation (second column) and the variance of inflation. Highly variable inflation is strongly associated across countries with high inflation. The average level of aggregate demand growth Δx would be the sum of the first two columns of Table 1. Is the average level of AD growth highly correlated with the variance of AD growth? What, if any, relationship does Lucas's model predict between the average level of AD growth and the elasticity of the AS curve? What pattern shows up in his results? (This is the basis for an extended test that we will examine in a couple of weeks.)