« Suppose that printing money were irreversible | Main | What is the most pressing issue in Canadian tax policy today? »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Interesting to see your perspective, Livio.

"Moreover, the ability of modern macro to make itself useful as a forecasting tool has yet to emerge."

But what about as a useful tool for giving policy advice?

Interesting point. I guess its basic policy advice is if there is going to be a policy intervention by government, what is the market economy not doing well and what can be made better with the policy intervention. Given the micro framework underpinning it, what is the market failing at. Oddly enough, based on what I read it seems like modern macro is treading some of the same ground that public finance did years ago in looking at things like externalities but is reformulating it in a macro context. However, that is just my intuition as I do not have a firm enough grasp of what modern macro theory is actually doing.

Moreover, the ability of modern macro to make itself useful as a forecasting tool has yet to emerge.

The obvious question is how this could be possible even in principle if the weak efficient market hypothesis is true. Does the author discuss this problem?

Athreya states that macroeconomics "has failed,in recent years, to be useful in a variety of ways" because the models have not been able to forecast huge downturns. However, he argues that this has resulted in a shift in research priorities to better understand for example the role of finance as sources of macro fluctuations. This suggests to me that Athreya believes that these models eventually should be able to have some ability to forecast the future.

One thing you should keep in mind is that this book received quite a bit of criticism when it first came out. Here's the url for a blogpost from back then by John Quiggin, if you want to know more.

I think Simon Wren-Lewis may have written something about it as well?

Though the book apparerently contains quite a few things of interest and does a good job of explaining stuff, it's problematic because to quote: "The result is that there is almost zero intersection between Big Ideas in Macroeconomics and what I would think of as macroeconomics."

Thanks for the sites Hugo. Some of the ideas in the book I was aware of from my own grad school experience. For example, rational expectations was a new and exciting area when I took macro. I recall another macro course that featured some micro foundations in the analysis and explanations for why it should be important. I suppose the 1980s were an important period in the move to the new macro. However, I must admit that after being away from the stuff for 25 years, modern macro does seem like a pretty alien world.

From the John Quiggin post linked to by Hugo, "The result is that there is almost zero intersection between Big Ideas in Macroeconomics and what I would think of as macroeconomics. It’s not so much that I think Athreya is wrong is that we are talking past each other. As Charles Goodhart said of DSGE, Athreya’s version of macro excludes everything in which I am interested."

I think that summarises Athreya’s book well. It's decidedly focused on published, academic macro and it's tools and not how it's applied, which is still often rather ad hoc as Noah Smith has frequently noted. I think the more clever young researchers see modern macro in the same way you summarised it. The field sort of went back to square one in the 1980's and rebuilt itself from scratch, and it's still building up to a truly realistic model (but that might just me projecting my view).

What books are used to teach macro in Canada?

What kind of knowledge can I expect from a Canadian macro economics M.Sc. ?

Fleming Mundell


what kind of math knowledge ?

"required the leap of faith that the whole was somehow more than the sum of its parts."

...a leap of faith that is made in virtually every other field, including many which have far better empirical records than anything in economics.


I was going to say something similar to UnlearningEcon above. But rather than > than the sum of the parts, I'd say an aggregation at the macro level can be very different than the parts. Take the 19th century gas law: it describes a emergent properties of huge collections of individual gas molecules, whose characteristics on the macro scale might not even make sense to talk about on the micro scale and vice versa: e.g. an individual molecule does not have "pressure." But at the same time, all but a count of the degrees of freedom of movement of the individual molecule survives aggregation to the macro level: i.e. it might be the case that the gas law would survive even if molecules had very different micro properties than they do.

... that should read "nothing but a count of the degrees of freedom" ...

Livio, you write about macro forecasting above, so you might like this: an amateur has a model he's putting up against the Fed. It's not quite a fair comparison because his model only models a few things and the Fed's model produces forecasts of many different variables. However, there is a well defined theory behind his model: he's not just curve fitting to empirical data... although he has a few parameters he does need to fit to the "in-sample" data (when making his "out of sample" forecasts)... and when I say a "few" that's literally true: depending on exactly what his model is doing the parameter count ranges between 1 and 3, whereas the Fed's model has many more times this number of parameters that need to be tuned. I'm a fan of his approach: not just the particular set of hypotheses he's put forward, but his general attitude: he's stated explicitly that he's not interested in adding "epicycles" to his model... so if the evidence suggested that he's wrong, he'll write a postmortem on his blog, and shut it down. So far this hasn't happened yet. I suppose that in being so bold, it helps that it's not his day job!... but I would ***LOVE*** to see a real economist put their ideas on the line like that! Expose their hypothesis to falsification by reality. Now I'm probably being hugely unfair, and for all I know that happens all the time in the academic journals (which I don't read: I'm an amateur who just reads excellent blogs like this one). But still... to pick a target at random, imagine Paul Krugman writing a post in which he states something similar to what Jason has stated on several occasions: "I'll keep my eye on Japan [or Europe, or Russia, or Canada, or the US... Jason has looked at a lot of economies and made predictions] and if the trend I'm predicting here (see figure 2) is X far off, it means my hypotheses are likely false. I won't be adding any epicycles to defend my model from reality. Instead I'll write a postmortem, and wrap it up. Maybe I'll take up gardening." Not in a million years, right? What economist is going to publicly and explicitly lay out numeric criteria to let reality judge if their fundamental hypotheses are wrong (in a blog format anyway), like biologists or chemists do? Jason has a prediction for Canada for 2015 actually... so we won't have to wait long for the first falsification test. In fact Japan is even more pressing than that... that test is has been ongoing and he has a new update now, but so far it hasn't been a "fail" result.

He has a search box in the bottom right so you can look up his predictions for other economies: for example his one for Canada in 2015 (he called it a "Worthwhile Canadian Prediction" in honor of this blog).

Sorry to be so "down on" economists above: obviously the subject fascinates me, and I love to read excellent blogs like this one. And I find John Cochrane and Noah Smith's argument convincing: that macro data is largely uninformative, which has forced economists to become very clever about extracting what data they can from it. This is totally unlike most hard sciences where it's relatively easy to re-run another controlled experiment if the data is suspect. In economics (they argue) all the data is suspect, and there are very few controlled experiments... which has forced economics to be in a bit of a special class by itself... not really a science and not really not-a-science. I think Jason (who's a physicist) is trying to take a wee bit more scientific approach (to this amateur's eyes anyway).

BTW, I love to read bloggers like Nick Rowe. I'm not as familiar with your posts: but in terms of explaining macro concepts in an understandable way to amateurs such as myself it's hard to beat Nick and his tales of haircut based economies, using cows for money, etc. However, one thing I've learned is that Nick is not the only clever story teller amongst macro economist bloggers... and all their stories (while all very convincing sounding) can't possibly all be true. Another thing I've learned from Jason is how difficult it is to distinguish between stories which are legitimate explanations and "just-so" stories. In contrast, Jason's hypotheses do not rely on human scale emotions and motivations and "putting yourself in the shoes of" different actors in a story: I think it's fair to say that in his hypotheses very few human characteristics survive the aggregation process at the macro scale. In fact, he's argued that the famous physicist Richard Feynman may have got it wrong when he stated this:

"Imagine how much harder physics would be if electrons had feelings!" - R. Feynman

Jason says, maybe not! Electron emotions (if they exist) might not have any bearing on our macro observations.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad