Modern macro models are falsifiable

To be honest, I’m having a hard time parsing this post by Arnold Kling. He talks about issues with “structural models” in macroeconomics, but I’m not sure to what he’s referring. Of course models reflect how somebody thinks the economy works. The key is whether or not you can falsify that model. Kling goes on to make is statistical power critique (I think), but I think this misses the point.

Models in modern macro are falsifiable. I know this because I’ve seen some falsified in the literature (this is the whole game!), I’ve been in seminars where a new model is dismissed by some in the audience because it “explains too much” and macroeconomists have developed clear criteria for evaluating models.

A model needs to reproduce time-series data that have statistical properties similar to data generated by the real economy. This often means standard deviations of income, consumption, investment, etc generated by the models match standard deviations in real data. Macro folks also look at how the model reproduces correlations between these data.

An example from recent literature might help. So there’s these new Keynesian models with sticky prices and imperfect competition that predict inflation will spike as soon as a technology shock happens and then peters out over time. The problem with this prediction is that inflation doesn’t react this way to technology shocks in the real economy. A number of empirical studies with a number of different so-called identification assumptions (maybe this is what Kling has problems with?) have found that inflation doesn’t spike right away after a shock. It takes a few quarters for inflation to hit its peak before it then peters out. Macro people affectionately call this “hump shaped” inflation.

This as seen as falsifying the simply new Keynesian model ((Falsifying in the same way the motion of Mercury falsified Newtonian physics. Newtonian physics still does a pretty good job; I’m mindful of walking under trees dropping their fruit. Relativity just explains more facts than Newtonian physics does.)) . And people have been working on alternative models that generate hump shaped inflation. (If you ask me, the issue was successfully resolved in this paper by one of my former teachers.)

The extent of our knowledge

One of the reasons I think modern macro is successful is that it cleanly separates what we know from what we don’t know. Making assumptions about how people make decisions and assumptions about how those decisions interact, modern macro models give expected behavior. In our models we capture all that we don’t know in what are called “exogenous shocks”. They’re “exogenous” because they’re outside the model and they’re “shocks” because they’re exact value is unpredictable even if the economic agents inhabiting the model’s world knows their distribution.

Given the behavior of exogenous shocks and the predicted behavior of the model’s agents to those shocks, we can take the output of the model and compare it to real data. If a model replicates real data then its assumed this is due to the assumptions about shocks and behavior. Because the shocks represent things we don’t know, we’d like to have the most simply shocks possible and therefore have the structure of the model explain as much of the data as possible.

Similarly, we can compare models by seeing which explains the most patterns in the data (or at least the patterns we care about) using the same exogenous shocks. If one model explains more data, we say that one tells us more about whatever is being studied. The shocks, however, continue to reflect our ignorance.

I like this paper by Eggertsson (h/t EotAW) because it shows how important the assumptions on exogenous shocks can be. Cole and Ohanian have several papers (findings summarized here) that show New Deal policies (by which they and Eggertsson mean industrial and union policies increasing monopoly power) were contractionary. This is actually the standard view that follows from microeconomic theory.

Eggertsson shows that after you include some standard modern macro model features (e.g. sticky prices, monopolistic competition) and you change the assumption on the exogenous shocks, these New Deal policies are, to my surprise, expansionary. The reason is at the zero lower bound for interest rates, without these policies, expectation for inflation doesn’t materialize and the economy falls into a deflationary spiral. When the government gives unions and big companies monopoly power, on the other hand, people expect those unions and big companies to use that power to raise prices. In this way, New Deal policies take over traditional monetary policies roles in setting inflation expectations.

But I really, really like the paper because it makes explicit that the reason its results are different from standard results from Cole and Ohanian is because of the assumptions about the shocks. In both sets of research, the shock is to preference for precautionary savings (“animal spirits”). Cole and Ohanian assume the 1929 crash and the bungles of Hoover were the exogenous shock and when Roosevelt took office that shock began to dissipate. Eggertsson, though, assumes the shock persisted through the great depression. Which is the right assumption, nobody knows.

The paper has very nicely separated what we know from what we don’t know. This makes task for future research very clear… what were the nature of those shocks to animal spirits?

To me, this strand of modern macro literature is encouraging. Everyone’s assumptions are laid out on the table, theories are making heavy contact with data and progress can easily be identified. Its almost like this is science.

Snippets from the this week’s NEP-MAC

That’s New Economic Papers – Macroeconomics:

  • “We provide empirical support for a DSGE model with nominal wage stickiness where growth is driven by learning-by-doing and money shocks and their variance are allowed to impact on long-run output growth.”
  • “We extend and estimate the initial model on US data to account for specific shocks that destroy the stock of capital and…”
  • “Our model is calibrated to match the steady-state distribution of price adjustments in microdata; realism calls for firm-specific shocks.”
  • “…we address the causes of the Roaring Twenties in the United States. In particular, we use a version of the real business cycle model to test the hypothesis…”
  • “Household Labor Supply and Home Services in a General-Equilibrium Model with Heterogeneous Agents”

Out of about 30 papers, there was one pure theory paper and the rest were testing models with real live data. Its uncertain if these authors were of the fresh- or salt- water variety.

What is modern macroeconomics?

It is models of the economy using methods of new classical economics with Keynesian price rigidities (and the occasional relaxing of the assumption of rational expectations). That encyclopedia article is great, by the way.

Apropos of nothing, this is from the same encyclopedia’s article on fiscal policy:

The greatest obstacle to proper use of fiscal policy—both for its ability to stabilize fluctuations in the short run and for its long-run effect on the natural rate of output—is that changes in fiscal policy are necessarily bundled with other changes that please or displease various constituencies. A road in Congressman X’s district is all the more likely to be built if it can be packaged as part of countercyclical fiscal policy. The same is true for a tax cut for some favored constituency. This naturally leads to an institutional enthusiasm for expansionary policies during recessions that is not matched by a taste for contractionary policies during booms.

Better macro critics please

Frankly, I don’t feel like writing about this anymore except to register my general dissatisfaction with the recent criticisms of modern macroeconomics. The critics don’t seem to have noticed the progress made in the field in the last several decades.

  • Wilkinson – Macro people have been studying social psychology (aka expectations) for a long time. We know something about it. I do agree, however, that we know very little about how fiscal policy affects expectations.
  • Kling – I’m sorry you hated your graduate macro class in the 70s. There’s been progress since then. Substantive criticism here.
  • Krugman – Fama’s not a macroeconomist so his his writings are a noisy signal from which to glean information about the state of knowledge in macroeconomics and you’re ignoring everything but the first couple of paragraphs of Cochrane’s paper.
  • Waldman – Everyone in macro uses DSGE models with imperfect competition and price stickiness. And, horrors, everybody assumes rational expectations except when, of course, they don’t. The freshwater vs. saltwater distinction is moot. The debate is about what frictions matter, not whether or not there are frictions and its certainly not about methodology.
  • Wolfers – Did you check to see how many references to “phlogiston” there are in the physics literature lately? Does ideological bias explain this trend? Those bastard right- (or is it left-) leaning physicists! Grrr… the answer is fiscal policy doesn’t F-ing work so why F-ing study it!!!111!!!

This isn’t these critics’ fault, of course. Macro people have been bad at getting the message out.

If you really care about the state of knowledge in macroeconomics, there’s a great set of papers in the inaugural edition of the the AEJ: Macroeconomics journal. I like the Michael Woodford paper the best. It is accessible to everybody.

Arnold Kling doesn’t know modern macro

He claims modern macro is only mathematical masturbation, that its not empirically relavent, it ignores unemployment and the modern consensus was reached in the discipline by ignoring critics. He’s wrong, very wrong, on all counts. Also, he seems to misunderstand the modern critique of macroeconometric models — the so-called Lucas Critique ((I’m too lazy to write out what he gets wrong about the problem with macroeconomic models. The fundamental problem with them has little to do with the statistical power of the tests used to test those reduced form models, as Kling claims. Basically, Lucas argued reduced form systems of equations didn’t tell us much with making policy because they only reflected averages of past behavior and they didn’t provide understanding of underlying mechanisms. When the underlying environment changes in some fundamental way, like there’s a change in policy, the past correlation between variables may no longer hold. Its a variation on the theme of correlation not being causation.)).

Macro-theory is math heavy and it has become more so over time. Part of this is pure mathematical masturbation, no doubt, but to a large part this trend is driven by a desire to make explicit all the assumptions that are being made and to more precisely understand the implications of those assumptions.

The need for more math is also related to the increase in the empirical relevance of theory. I’m convinced the only standing legacy of the Real Business Cycle literature, besides method, is its insistence on bringing the models to the data. In modern macro, its simply not enough to identify the existence of some effect or other. For example, real business cycles were relevent because they proved to be quantitatively important… a large chuck of business cycle fluctuations are driven by supply shocks. And RBCs have been supplanted because they didn’t explain enough of the data. The empirical relevance of real shocks couldn’t have been tested without out explicit mathematical models of the phenomenon.

This is what frustrates me about Kling, Krugman, et al’s ad-hoc theorizing. They seem contented to identify that certain macroeconomic features exist, but they don’t bother to quantify the importance of those features. For Kling, he identifies risk preference shocks as important drivers of the business cycle. Fine. How important are they? Well Kling never attempts to answer that question, but these shocks have actually been studied and they’ve been found to be empirically insignificant.

For Krugman, its all about what I call “weak” liquidity traps ((He defines a liquidity trap as a period of zero short term interest rates. This is a “weak” definition because the concern with liquidity traps is that they make monetary policy ineffective. I’ve mentioned before there’s a quite significant literature that shows monetary policy is not ineffective at zero rates of short term interest. Certainly, routine monetary policy where the Fed and banks swap treasuries for cash is ineffective at this lower bound for interest rates, but routine policy isn’t the only policy tool available to the Fed.)). There’s no discussion of magnitudes; its never asked just how important is this fact to our present predicament. Interest rates are zero and to Krugman this is the only salient fact about the macro-economy.

Macro has moved on from real business cycles — even if Kling’s criticisms seem not to have — but it has maintained the discipline of bringing the models to the data. Whole books have been written on how to estimate so-called DSGE models.

Time-series techniques have come a long way since Box-Jenkins — VARs and what have you. More importantly, applied theorists use those techniques to test theory. Its rare to find theory papers that don’t have data sections where there’s at least a nominal attempt at testing the implications of the model.

Real business cycles ignored unemployment. True. Modern macro, on the other hand, has been modeling and testing ideas about unemployment for two decades. My own research is a version of a general equilibrium employment search model. Here’s a DSGE model with unemployment. And look at that, there’s even a data section where the model is tested. How about that!

I’d say, if you had the time, you could pick through Kling’s quite good macroeconomic lecture series and find each element studied somewhere in the modern macro literature. Of course, its not his fault that he not up on the recent literature in macro. I’ve made the case that we in the macro field need to do a better job of evangelism. That said, he should be aware that much research has been done since the 70’s and most of it isn’t real business cycles and lots of it touches on the “fundamental issues of macro.”

Rodrik clarifies the debate?

Rodrik observes, as Wilkinson suspected, the margin on which economists are debating the fiscal stimulus isn’t economics. Instead the Krugman vs. Chicago debate is “philosophical, political, and practical–revolving around the role of government, the extent of rent-seeking and public-choice concerns in government programs, and the right mixture of prudence and boldness that the situation requires.”

I’m not convinced. There’s the issue of whether monetary policy is impotent or not given zero interest rates. Krugman says “traditional” money policy — swapping treasuries for cash — is impotent and he’s right. If the policy choice was just between traditional money policy and fiscal stimulus than Krugman would be right about the necessary policy and Rodrik would be right that this was a debate about ideology. Big government types observe traditional money policy is broken and advocate government expansion. Small government types hate that idea and rail against it.

The choice, though, isn’t between traditional money policy and fiscal stimulus. There are a number of papers showing monetary policy doesn’t have to be impotent when we run up against the zero lower bound of interest rates. The ideological battle dissolves (unless you really, really want fiscal stimulus) because with non-traditional monetary policy we’re still talking about huge government intrusion in markets — albeit temporary and reversible. Non-traditional monetary policy has the Fed intervening in capital markets — running them as Cochrane puts it — but this sort of policy is preferable to fiscal policy for a number of reasons outlined in that Cochrane article, but primarily because its much easier for it to adjust to prevent inflation.

In any case, we’re back to fighting over economics.

Does not compute

Suppose A \Rightarrow B. Steve Waldman says recent events have shown !A:

Economists talk about consumption smoothing, how it may be optimal for a consumer whose income is volatile to borrow during periods of low income and repay (or save) during periods of high income in order to maintain a constant standard of living. That’s very well in models where consumers know the true distribution of their future income, where the spread between borrowing and lending interest rates is not very large, and where consumer preferences are time-consistent. In practice, none of these conditions hold even approximately.

He concludes !B:

consumers ought to borrow only to counter severe downward shocks to income, pay off borrowings quickly, and build buffers of precautionary savings, since the cost of dissaving is much less then the cost of borrowing.

Dude, (A \Rightarrow B) \not \Rightarrow (!A \Rightarrow !B)!

Take out the garbage <rant>

In a comment below, this guy said:

Normal economists worry about the fact that deficits lead to inflation or higher taxes. Neither’s good. Normal economists care about the so-called multiplier in which government expenditures magically increase GDP by a multiple of their value. Normal economists have measured the multiplier and found it to be close to 1, i.e. there is no magic.

Krugman’s basically argues we’re not in normal times so all those normal results don’t apply. We’re living with so-called “depression economics”. With lack of sufficient evidence about what happens with government expenditures in abnormal times — depressions are relatively rare— we’re just suppose to believe Krugman’s conjectures are right. Also, he ignores the few data points we do have and that are contrary to his conjectures, e.g. the fact that Japan had been stimulating the crap out of its economy during its “lost decade” or the fact that when Roosevelt fixed monetary policy the economy started recovering from the Depression (i.e. fiscal stimulus doesn’t do much).

Krugman and team social democrats want a “green economy” and they want nationalized medicine. This crisis has people crying for government to do something. Using Krugman’s depression economics as psuedo-intellectual backing, social democrats have a political window to bring those things about.

As political cover, “depression economics” is brilliant. As a scientific theory, Krugman’s conjectures have little going for them.

Can we all agree that after we’ve spent a few trillion on “green” industries and we’ve got nationalized medicine, we can throw Krugman’s “depression economics” out with the rest of the scientific trash? There’s no solid theory, the evidence for its conjectures is weak at best and there’s not systematic way to test it. The only reason economists are even talking about it is because a prominent social democrat just won the economics Nobel and not because its an idea with merit.

“Depression economics” is brilliant politics… as is mercantilism and nativism. That which politicians use to prey on the ignorance of the masses doesn’t make for interesting topics of scientific economic inquiry.

Sentences of the day

From both the macro evidence and this body of micro–economic work, a large consensus—right or wrong—has emerged:
It holds that modern economies need to constantly reallocate resources, including labor, from old to new products, from bad to good firms. At the same time, workers value security and insurance against major adverse professional events, job loss in particular. While there is a trade-off between efficiency and insurance, the experience of the successful European countries suggests it need not be very steep. What is important in essence is to protect workers, not jobs

— From this neat summary of the experience with unemployment in Europe over the last several decades (and the changing intellectual story-telling that went with those trends)