Wednesday, 28 August 2013

Simon Wren-Lewis absolutely brilliant again

Simon Wren-Lewis is on fire with this article.

I will not comment but merely copy it for every body to read. It is long but so very worth it to read and understand..

Macro workers and macro wars

A long post I’m afraid, even with extensive use of footnotes. But I really think it is much more productive to try and understand someone’s opposing point of view than just be rude about it.

Most academic macroeconomists are just trying to advance the discipline by getting their papers published, and are certainly not consciously trying to defend some ideological viewpoint. As a result, there are lots of macroeconomists producing high quality work in a wide variety of diverse areas. There are many interesting new ideas being explored. Furthermore this work can be appreciated by most fellow researchers. Unlike the 60s and 70s, where members of different schools of thought talked across each other, we now have a shared language as a result of the microfoundation of macro. My own view is that as a result macro today is much more interesting than macro was back then. Furthermore, this work can be useful to policymakers, as Paul Krugman outlines in the case of monetary policy here.

Ahh - that may have made you pause for thought. Isn’t that the same Paul Krugman who says there is something “deeply wrong with economics”. Who talks about how ideology and politics distort the advice that economists - and perhaps particularly macroeconomists - give to policymakers. And who suggests that in many cases policymakers would be better off thinking about good old IS-LM than any of this more modern stuff.

One of the problems when Paul Krugman does this is that it gets on the nerves of many academic macroeconomists, who would much rather identify with the sentiments I express in my first paragraph. Economists like Tony Yates, for example. I suspect they see Paul Krugman’s attacks on the state of macroeconomics as akin to a personal attack on their own and colleagues work, and as a result they can go way over the top in reaction. I think this is understandable, but it is wrong.

As I see it both points of view are correct. As Stephen Williamson argues, in one sense macroeconomics is flourishing. Take the example of financial frictions. There is now a wealth of papers out there exploring different types of friction, in large part responding to events of just the last five years. This is hardly the response of a moribund, out of touch discipline. And as Krugman says, sometimes this work can be useful to policymakers. But that is not the acid test for the integrity of a supposedly scientific discipline. In days of old, policymakers made much use of astrology. The acid test is when the discipline tells policymakers something they do not want to hear, and unites behind this implication of its models and the data.

Nearly fifteen years ago I began working with DSGE models looking at monetary and fiscal interactions. [1] Doing this work taught me a lot about how fiscal policy worked in New Keynesian models. Iunderstood more clearly why monetary policy was the stabilisation tool of choice in those models, but also why fiscal policy - appropriately designed - was also quite effective in that role if monetary policy was absent (individual countries in the Eurozone) or impaired (the ZLB). Why New Keynesian models? Because if you were interested in business cycle stabilisation, that is the framework that most involved in that area (academics and policymakers) were using. So when we hit the ZLB, the reaction of policymakers in using fiscal stimulus seemed logical, entirely appropriate and fully in line with current theory.

I also found that in these models the basics of Barro’s tax smoothing hypothesis continues to apply, so if you needed to reduce debt, you should do so as gradually as possible. This also seemed like as robust a result as one can get in macro.

The acid test for macro came in 2010. Policymakers, for a variety of reasons, went into reverse with fiscal policy. Austerity replaced stimulus around the world. If academic macroeconomists had been true to their discipline, they would have been united in saying that our standard models tell us this will reduce output and raise unemployment. They would have said that if markets allow, the time to reduce debt is when the ZLB comes to an end, and then it should be done gradually. Many did say that, but many did not. This division at least encouraged policymakers to continue with austerity.

So macroeconomists as a collective failed this test, repeating errors made in the 1930s. But unlike the 1930s, it did not have ignorance as an excuse. 

This is a crucial point that many on both sides tend to ignore. In 2010, the standard business cycle model was the New Keynesian model, and the implications of that model for the efficacy of appropriately designed fiscal policy are clear. So to blame the failure of 2010 on the current dominant macro model is just wrong. You may not like that model, but it cannot be blamed for the widespread adoption of austerity, or the ambivalent attitude of many macroeconomists towards that policy change.

So in my view macroeconomists, not the dominant macroeconomic model, failed. Why? The easy answer is to say that macroeconomists were too influenced by ideology, but I think for most (not all) it is the wrong answer, as I said in my opening paragraph. I think for most (not all) it is not simple politics, so in that respect I agree with Stephen Williamson. [2] So what is it?

Here is one possible answer. First I have to appeal to macroeconomists who learnt their trade after the New Classical revolution to consider where the now dominant methodology in macro came from. If you start your history of macro thought in 1980, then everything can seem nicely progressive. First there were RBC models, applying basic micro to macro. The data showed that this framework could not explain how monetary policy appeared to work, so to resolve this puzzle New Keynesian theory was developed.

But it was not like that at all. [3]

Macro started well before 1980. Its theoretical basis may have been shaky, but it was a scientific discipline in the Popperian sense. The New Classical revolution attempted to kill off Keynesian ideas, and deny the importance of aggregate demand. In that respect it now seems clear, because of the dominance of New Keynesian theory, that this was regressive rather than progressive. 
Yet in many who were part of the New Classical revolution, or who were taught by its leaders, there remains a deep antagonism to Keynesian ideas. This was not enough to prevent the emergence of New Keynesian theory, but the NK model built upon rather than challenged the RBC framework, and it could always be dismissed with an assertion about price flexibility. As a result, in certain places NK theory was tolerated rather than embraced, or was quietly marginalised.

This attitude was facilitated by another aspect of the revolution. Although it failed to kill Keynesian economics, it did succeed in changing how macro was done. Microfoundations macro did not just become an important way of explaining the economy, it became the only acceptable way. Now we can debate the wisdom of that. But I think it is very difficult to deny that this methodological revolution reduced the discipline that data can provide on model proliferation. It becomes far too easy to say ‘but in this model something different happens’, even if there is compelling empirical evidence that said model is not applicable. [4] [5]

I think this may help explain why a good proportion of macroeconomists failed to advocate fiscal stimulus in 2009 and call the consequences of subsequent austerity. [6] Now you may disagree with my view that this represented a failure for macroeconomists as a whole. What I want to convince you of here is that my view that it did, and perhaps similar views of others, does not amount to an assertion that modern macro is fundamentally flawed, and that those working within it are wasting their time. While the New Classical revolution may have moved macro many steps forward, in condemning Keynesian ideas it took one large step backwards, and the consequences of that mistake are still with us.

[1] In part this was a reaction to a different policy decision. The work of various economists before the formation of the Euro had suggested that countercyclical fiscal policy should be a key stabilisation tool for individual Eurozone members. That work was ignored by policymakers, and they were backed up by a significant number of macroeconomists, in part using other models that focused on free rider problems and fiscal dominance. How this all turned out is another story, but once again the collective of academic macroeconomists hardly covered itself in glory.

[2] Personally, I do not think the actions of some eminent economists who ignore their economics when batting for their favoured politicians is critical here, regrettable though it is. Nor on its own were the comments of other eminent macroeconomists who appeared not to have kept up with the literature, although as I suggest here I think this was indicative. Both could have been isolated examples, quickly brushed aside. More revealing is this survey, where although 46% of economists agreed that the US stimulus was a good policy, a large 40% were uncertain or did not answer. That is just one survey, but it reflects a similar division amongst the macroeconomists I know, and the views you find on the web. See, for example, the quote from Tom Sargent in Stephen Williamson’s post. Now I think some of this 40% are equivocal about fiscal stimulus (and therefore not too worried about austerity) because of a deep distrust of government intervention or the state. Whatever the merits of this mistrust, this is exactly the ideology and politics that Paul Krugman and I complain about. I think some others of this 40% take a view that monetary policy is still up to the job, even at the ZLB. I have talked about this most recently here. This post provides a third possible explanation, although I am sure there are others. (I tried to be comprehensive here.) 

[3] Getting the history of macro thought right is important for other reasons as well. As I suggested here, the structure of NK models may owe as much to the need to work with rather than against the then dominant RBC paradigm, as to any intrinsic empirical merits of that structure.

[4] I have tried to argue that economists working in central banks take more notice of the data, which helps explain why the NK model is dominant there, but Stephen Williamson disagrees.

[5] Another arguable consequence is that modelbuilding became too conformist. The charge that some element of a model is ad hoc and lacks microfoundations hangs like a Sword of Damocles over modelbuilders. Probably too much intolerance of alternative methodologies came with this revolution as well. I think this is part of any explanation of why so little work was done on financial frictions before 2008, as Mark Thoma suggests. Again I am not arguing that the methodology is wrong, but instead that it may have certain perhaps unintended and unfortunate consequences. In my own experience if you talk to many microeconomists about DSGE in macro they can be quite critical: for example about how narrow the micro used is, or how obsessed with technicalities the analysis can become. Sometimes they can be downright dismissive

[6] You could argue that ideology lay behind the New Classical revolution’s attempt to kill off Keynesian economics, and I do not really know enough to agree or disagree. What I do think is that most of those involved in the revolution thought that they were just exposing deficient theory, which in many cases they were.