Blog o’ Burns

Hoover, Kevin D. “Is Macroeconomics for Real?” (June 1999)

How sad would it be if the big punchline to this course was the fact that macroeconomics wasn’t real? I’d ask for a refund. Thankfully, that was not the conclusion of this article and hasn’t necessarily been the conclusion of most economists. In general, economists don’t doubt that stable and observable relationships exist at the aggregate level; however, can these relationships be explained in the context of an individual’s rational microeconomic behavior? I won’t pretend to completely (or even mostly… or even partly…) understand Hoover’s argument against the necessity of microfoundations, but I believe he is saying that macroeconomics is not just a construct of measurement or of theory– economic aggregates aren’t just derived from micro behavior; they also affect behavior in ways that alter economic “reality.”

The quest to base macro theory on microfoundations is founded in methodological individualism, which argues that explanations of social, political, and economic phenomena can only be adequate if they are based on the beliefs, attitudes, and decisions of individuals. Lucas and his fellow new classicals have tried to achieve this by using representative-agent models in which a single individual takes on national income as his/her budget constraint and makes microeconomic optimal choices which are supposed to represent the choices of the aggregate economy. However, few macro phenomena have been successfully reduced to their microfoundations.

Hoover distinguishes between two types of economic aggregates: natural aggregates are simple sums or averages, while synthetic aggregates are derived from their components in a way that makes them “dimensionally distinct” from those components. For example, changes in the general price level entail more than just looking at the price of one good or even just averaging the prices of many goods. Everyone has the sense that inflation exists and recognizes that a dollar doesn’t buy as much as it used to, but looking at prices alone won’t tell you much about changes in purchasing power. Economists have multiple ways of measuring inflation, but none will perfectly capture what the everyday person experiences– this just shows how aggregates can exist independent of how they are measured and how they affect the behavior of economic agents.

Though I’m typically the kind of person who likes tearing apart an argument, I have to say that I applaud Hoover’s questioning of the need for microfoundations. To me, new classical pretensions of finding economic truth by incorporating microfoundations seems like just a veneer covering up an over-simplified theory that assumes away any of the messiness that characterizes a real economy. Besides that, I never saw any evidence that these models were truly based on micro behavior. As far as I could tell, microfoundations were just used to explain why the economy is always operating at potential– since people/firms attempt to maximize utility/profits, they always react rationally to shocks in the economy, and thus output is always at its optimal level given people’s preferences for leisure and consumption. Such conclusions seem terribly convenient and don’t take into account possibilities for externalities, prisoner dilemma situations, and any number of occasions for market failure. Of course, all of these issues occur at the micro level, but I don’t feel like it’s necessary (or even possible) to model such problems at the aggregate level.

The point is that details matter. I consider myself a rational person, but my utility function probably changes on a daily basis based on my mood. If I’m having a bad day, I might buy some ice cream even though I wouldn’t consider it worthwhile on any other day. Let’s say that I’m a representative human being, and every person on earth increases his/her marginal propensity to consume ice cream on a bad day. Now imagine that there’s a shock to the economy, and suddenly everyone on earth is having an incredibly terrible day. A new classical would say: “And then all of the rational Stephanie’s went out to buy ice cream instead of doing homework, and ice cream prices rose due to higher demand, causing some of the Stephanie’s to drop out of line and equilibrating supply and demand.” However in the real world, it goes something more like this: “And then all of the Stephanie’s lined up at all of the Carl’s in the world and got way annoyed that the line was so long. The line was so long that the opportunity cost of getting ice cream (better grades from doing homework) exceeded the pleasure of eating it. However, if the line got shorter because other Stephanie’s dropped out, the net benefit of sticking around would make it worthwhile. As such, all of the Stephanie’s waited for each other to drop out of line. Meanwhile, the employees at Carl’s didn’t want to raise their prices, because they were afraid that all of their business would go to Coldstone instead. Besides, they knew that tomorrow would be a good day, so it didn’t make sense to change their prices for just one day. In the end, a suboptimal amount of homework got done and Carl’s didn’t make as much profit as it could have.” As you can see, all of these decisions were made at the micro level, but they were all based on perceptions of the larger economy. I hope that’s at least part of what Hoover is talking about.

Mankiw, Gregory N. “A Quick Refresher Course in Macroeconomics.” Journal of Economic Literature 28 (December 1990): 1645-1660.

In this article, Mankiw attempts to reconcile why developments in economic theory had not yet been translated into changes in applied macroeconomics. Until the 1970s, John Hicks’ interpretation of John Maynard Keynes’ vision of the economy through the IS-LM model was deemed the general consensus among economists at the time. Such theory became embodied in more practical, large-scale macroeconometric models such as the MIT-Penn-Social Science Research Council model. However, support for the IS-LM model dwindled in the early 1970s due to its inability to explain stagflation and its lack of adequate micro foundations. Friedman and Phelps’ prediction that the empirical relationship between inflation and unemployment would break down if exploited came true. Lucas also scored a blow on Keynesian thought with his critique arguing that many key macroeconomic variables depend on expectations. If a model cannot account for changes in expectations then it cannot evaluate alternative policies.

This breakdown in consensus caused economists to branch out in three directions of research to build macro models on firm micro foundations; these directions included: 1) modeling expectations (particularly rational expectations); 2) explaining macro phenomena using new classical models; and 3) resurrecting Keynesian theory with new Keynesian models.

New Keynesian Macroeconomics

While new classical models attempted to explain economic fluctuations within the Walrasian market-clearing framework, new Keynesians took another look at the possibility for market failure on a grand scale. They needed to explain why wages and prices failed to adjust instantly to equilibrate supply and demand. Lack of a theoretical justification for price and wages stickiness had helped bring down the original Keynesian model. Such an explanation would need firm mico reasoning as well.

Fixed Prices and General Disequilibrium

During the 1970s, Keynesians spent much time and energy trying to determine how markets interact when prices are fixed at nonmarket-clearing levels, primarily demonstrating economic policy’s influence on output and unemployment under such a regime. These models showed that the behavior of the economy depends greatly on which markets have excess supply and which have excess demand. The excess supply of labor (unemployment) can occur under two regimes: 1) classical unemployment in which real wages are too high all workers to be profitably employed; and 2) Keynesian unemployment in which unemployment occurs because firms cannot sell all they want to at the market price. In the classical case, unemployment occurs due to disequilbrium in the labor market, while the Keynesian case implies a disequilibrium in the goods market. Where is the problem? If imperfections exist in both markets, how do they interact? What keeps prices from adjusting to clear markets?

After the new classical revolution, American Keynesians switched focus to modeling the price adjustment process. Economists typically work under the assumption of perfect competition in which all firms are price-takers. However to model price adjustments, new Keynesians acknowledged that some economic actors have the power to be price-setters.

Labor Contracts and Sticky Wages

To explain sticky wages, a popular line of research was to model labor market failure as caused by labor contracts that specify nominal wages in advance. Such models were appealing because they reflected obseved institutions and implied that policy does have relevant effects on employment. Since nominal wages cannot adjust to economic disturbances, monetary policy can systematically determine real wages and thus employment.

Such models faced criticism on three grounds: 1) what micro foundation is there for firms and employees to enter into potentially harmful contracts?; 2) why should nominal wages play such an important role in determining employment?; and 3) shouldn’t real wages be countercyclical if this theory were true?

Monopolistic Competition and Sticky Prices

In the 1980s, Keynesians shifted focus from the labor market to the goods market. In particular, they looked at monopolistically competitive firms who face small “menu costs” when changing prices. Such costs include the time it takes to inform customers, the risk of making customers angry, and the effort to determine whether prices should change or not. If these costs exceed the benefit of changing prices, they will maintain socially sub-optimal prices.

“Menu costs” reflect nominal rigidities in the goods market. However, real rigidities may exist in the labor market in the form of efficiency wages. In such models, a cut in wage corresponds with a drop in productivity, since it causes the best workers to quit. Since firms monitor effort imperfectly, workers sometimes shirk their responsibilities and risk getting fired. Lower wages imply lower opportunity costs for losing one’s job, thus increasing the likelihood of shirking. If the cost of lost productivity outweighs the benefit of cutting wages, firms will not reduce wages in response to an excess supply of labor. Such real rigidities alone only cause classical unemployment; however when coupled with nominal rigidities, they provide a powerful explanation for Keynesian disequilibrium.