Macroeconomics at 100
Has macroeconomics progressed over the past 100 years, or are we merely treading water? There are good arguments for both sides. Before considering macroeconomics, I use an analogy in the field of urban planning. Then I’ll argue that macro looks a lot better if we view it as a series of “critiques”, not a series of models.
In the middle of the 20th century, city planners favored replacing messy old urban neighborhoods with modern high rises and expressways. Here’s Le Corbusier’s plan for central Paris:
Today, those models of urban planning seem almost dystopian. What endures are the critiques of the modernist project, such as the work of Jane Jacobs.
I believe that macroeconomics has followed a broadly similar path. We’ve developed lots of highly technical models that have not proved to be very useful, and a bunch of critiques that have proven quite useful.
Many people would choose 1936 as the beginning of modern macroeconomics, as this is when Keynes published his General Theory. I believe 1923 is a more appropriate date. This is partly because Keynes published his best book on macro in 1923 (the Tract on Monetary Reform), but mostly because this was the year that Irving Fisher published his model of the business cycle, which he called a “dance of the dollar.”
[BTW, Here’s how Brad DeLong described Keynes’s Tract on Monetary Reform: “This may well be Keynes’s best book. It is certainly the best monetarist economics book ever written.” Bob Hetzel reminded me that 1923 is also the year when the Fed’s annual report first recognized that monetary policy influences the business cycle, and they began trying to mitigate the problem. And it was the year that the German hyperinflation was ended with a currency reform.]
The following graph (from a second version of the paper in 1925) shows Fisher’s estimate of output relative to trend (T) and a distributed lag of monthly inflation rates (P). Many economists regard this as the first important Phillips Curve model.
Fisher argued that causation went from nominal shocks to real output, which is quite different from the “NAIRU” approach to the Phillips Curve more often used by modern macroeconomists—which sees a strong labor market causing inflation.
Today, people tend to underestimate the sophistication of pre-Keynesian macroeconomics, mostly because they used a very different theoretical framework, which makes it hard for modern economists to understand what they were doing. In fact, views on core issues have changed less than many people assume. During the 1920s, most elite macroeconomists assumed that business cycles occurred because nominal shocks impacted employment due to wage and price stickiness. Many prominent economists favored a policy of either price level stabilization (Fisher and Keynes) or nominal income stabilization (Hayek).
The subsequent Keynesian revolution led to a number of important changes in macroeconomics. In my view, four factors played a key role in the Keynesian revolution (which might also be termed the modernist revolution in macro):
1. Very high unemployment in the 1930s made the economy seem inherently unstable—in need of government stabilization policy.
2. Near zero interest rates during the 1930s made monetary policy seem ineffective.
3. Increases in the size of government made fiscal policy seem more powerful.
4. A move from the gold exchange standard to fiat money made the Phillips Curve seem to offer policy options—“tradeoffs”.
While I believe that the implications of these changes were misunderstood, they nonetheless had a profound impact on the direction of macroeconomics. There was a belief that we could construct models of the economy that would allow policymakers to tame the business cycle.
Most people are familiar with the story of how Keynesian macroeconomics overreached in the 1960s, leading to high inflation. This led to a series of important policy critiques. Milton Friedman was the key dissident in the 1960s. He argued:
1. The Phillips Curve does not provide a reliable guide to policy trade-offs.
2. Policy should follow rules, not discretion.
3. Interest rates are not a reliable policy indicator.
4. Fiscal austerity is not an effective solution to inflation.
But Friedman’s positive program for policy (monetary supply targeting) fared less well, and is now rejected by most macroeconomists.
Bob Lucas built on the work of Friedman, and developed the “Lucas critique” of using econometric models to determine public policy. Unless the models were built up from fundamental microeconomic foundations, the predictions would not be robust when the policy regime shifted. As with Friedman, Lucas was more effective as a critic than as architect of models with policy implications. It proved quite difficult to create plausible equilibrium models of the business cycle.
New Keynesians had a bit more luck by adding wage and price stickiness to Lucasian rational expectations models, but even those models were unable to produce robust policy implications. Here the problem is not so much that we are unable to come up with plausible models, rather we have many such models, and we have no way of knowing which model is correct. In practice, the real world probably exhibits many different types of wage and price stickiness, making the macroeconomy too complex for any single model to provide a roadmap for policymakers.
Paul Krugman’s 1998 paper (It’s Baaack . . . “) provides another example where the critique is the most effective part of the model. Krugman argues that a central bank needs to “promise to be irresponsible” when stuck in a liquidity trap, although it’s hard to know exactly how much inflation would be appropriate. The paper is most effective in showing the limitations of traditional policy recommendations such as printing money (QE) at the zero lower bound. Just as the work of Friedman and Lucas can be viewed as a critique of Keynesianism, Krugman’s 1998 paper is (among other things) a critique of the positive program of traditional monetarism.
That’s not to say there’s been no progress. Back in 1975, Friedman argued that over the past few hundred years all we had really done in macroeconomics is go “one derivative beyond Hume”. Thus Friedman’s famous Natural Rate model went one derivative beyond Fisher’s 1923 model. There’s no question that when compared to the economists of 1923, we now have a more sophisticated understanding of the implications of changes in the trend rate of inflation/NGDP growth. That’s not because we are smarter, rather it reflects the fact that an extra derivative didn’t seem that important under gold standard where long run trend inflation was roughly zero.
When I started out doing research, I bought into the claims that we were making “progress” in developing models of the macroeconomy. Over time, we might expect better and better models, capable of providing useful advice to policymakers. After the fiasco of 2008, I realized that the emperor had no clothes. Economists as a whole were not equipped with a consensus model capable of providing useful policy advice. Economists were all over the map in their policy recommendations. If we actually had been making progress, we would not have revived the tired old debates of the 1930s. Even if the quality of academic publications is higher than ever in a technical sense, the content seems less interesting than in the past. Maybe we expected too much.
More recently, high inflation has led to a revival of 1970s-era inflation models that I had assumed were long dead. You see discussion of “wage-price spirals”, of “greedflation”, or of the need for tax increases to rein in inflation. And just as in the 1920s, you have some economists advocating price level targets while other endorse NGDP targeting.
Going forward, I’d expect to see a greater role for market indicators such as financial derivatives linked to important macroeconomic variables. In other words, like most macroeconomists I see future developments as validating my current views.
So how should we think about the progress in macro over the past century? Here are a few observations:
1. Both in the 1920s and today, economists have concluded that certain types of shocks have a big impact on the business cycle. The name given to these shocks varies over time, including “demand shocks”, “nominal shocks” and “monetary shocks”, but all describe broadly similar concept. Then and now, economists believe that sticky wages and prices help to explain why these shocks have real effects. In addition, economists have always recognized that supply shocks such as wars and droughts can impact aggregate output. So there is certainly some important continuity in macroeconomics.
2. Economists have made enormous progress in developing highly technical general equilibrium models of the business cycle. But it’s not clear what we are to do with these models. Forecasting? Economists continue to be unable to forecast the business cycle. Indeed it’s not clear that there has been any progress in our business cycle forecasting ability since the 1920s. Policy implications? Today, macroeconomists tend to favor policies such as inflation/price level targeting, or NGDP targeting. Back in the 1920s, the most distinguished macroeconomists had similar views. What’s changed is that this view is now much more widely held. Back in the 1920s, many real world policymakers were skeptical of price level or NGDP targets, instead relying on the supposedly automatic stabilizing properties of the gold standard, which had been degraded by WWI.
3. The shift from a gold exchange standard to a pure fiat money regime allows a much wider range of monetary policy choices, such as different trend rates of inflation. Fiscal policy has become more important. These changes made policymakers much more ambitious, perhaps too ambitious. In my view, the greatest progress in macro is a series of “critiques” of policy recommendations during the second half of the 20th century. Friedman and Lucas provided an important critique of mid-20th century Keynesian ideas, and Krugman’s 1998 paper pointed to problems with monetarist assumptions about the effectiveness of QE at the zero lower bound.
A series of critiques sounds less impressive than a successful positive program to tame the business cycle. But I it is a mistake to discount the importance of these ideas. They have helped to steer policy in a better direction, even as many problems remain unsolved.