Winner of the New Statesman SPERI Prize in Political Economy 2016

Sunday, 15 January 2017

Blanchard joins calls for Structural Econometric Models to be brought in from the cold

Mainly for economists

Ever since I started blogging I have written posts on macroeconomic methodology. One objective was to try and convince fellow macroeconomists that Structural Econometric Models (SEMs), with their ad hoc blend of theory and data fitting, were not some old fashioned dinosaur, but a perfectly viable way to do macroeconomics and macroeconomic policy. I wrote this with the experience of having built and published papers with both SEMs and DSGE models.

Olivier Blanchard’s third post on DSGE models does exactly the same thing. The only slight confusion is that he calls them ‘policy models’, but when he writes

“Models in this class should fit the main characteristics of the data, including dynamics, and allow for policy analysis and counterfactuals.”

he can only mean SEMs. [1] I prefer SEMs to policy models because SEMs describe what is in the tin: structural because they utilise lots of theory, but econometric because they try and match the data.

In a tweet, Noah Smith says he is puzzled. “What else is the point of DSGEs??” besides advising policy he asks? This post tries to help him and others see how the two classes of model can work together.

The way I would estimate a SEM today (but not necessarily the only valid way) would be to start with an elaborate DSGE model. But rather than estimate this model using Bayesian methods, I would use it as a theoretical template with which to start econometric work, either on an equation by equation basis or as a set of sub-systems. Where lag structures or cross equation restrictions were clearly rejected by the data, I would change the model to more closely match the data. If some variables had strong power in explaining others but were not in the DSGE specification, but I could think of reasons for a causal relationship (i.e. why the DSGE specification was inadequate), I would include them in the model. That would become the SEM. [2]

If that sounds terribly ad hoc to you, that is right. SEMs are an eclectic mix of theory and data. But SEMs will still be useful to academics and policymakers who want to work with a model that is reasonably close to the data. What those I call DSGE purists have to admit is that because DSGE models do not match the data in many respects, they are misspecified and therefore any policy advice from them is invalid. The fact that you can be sure they satisfy the Lucas critique is not sufficient compensation for this misspecification.

By setting the relationship between a DSGE and a SEM in the way I have, it makes it clear why both types of model will continue to be used, and how SEMs can take their theoretical lead from DSGE models. SEMs are also useful for DSGE model development because their departures from DSGEs provide a whole list of potential puzzles for DSGE theorists to investigate. Maybe one day DSGE will get so good at matching the data that we no longer need SEMs, but we are a long way from that.

Will what Blanchard and I call for happen? It already does to a large extent at the Fed: as Blanchard says what is effectively their main model is a SEM. The Bank of England uses a DSGE model, and the MPC would get more useful advice from its staff if this was replaced by a SEM. The real problem is with academics, and in particular (as Blanchard again identified in an earlier post) journal editors. Of course most academics will go on using DSGE, and I have no problem with that. But the few who do instead decide to use a SEM should not be automatically shut out from the pages of the top journals. They would be at present, and I’m not confident - even with Blanchard’s intervention - that this is going to change anytime soon.

[1] What Ray Fair, longtime builder and user of his own SEM, calls Cowles Commission models.

[2] Something like this could have happened when the Bank of England built BEQM, a model I was consultant on. Instead the Bank chose a core/periphery structure which was interesting, but ultimately too complex even for the economists at the Bank.


  1. Adam Posen noted on Twitter that the Bank of England, whilst they do make use of DSGE by funding the research into the CCBS, the MPC do not actually use the results in the process of making decisions on monetary policy. My thoughts were that the models at The Bank were more redundant Bayesian estimated standard Real Business Cycle (RBC) models that make no attempt to explain consumption fluctuations during aggregate shocks but are more concerned with explaining macro phenomenon (with the exception of search & matching (SAM) models).

  2. As a response this would be a best satire aplicable:

    "The WHO today warned of a virulent new virus affecting vulnerable groups in the Mid-West and Eastern USA. The outbreak, which began in the Mid-West's extensive Great Lakes ‘Freshwater’ river system, has recently jumped the ‘Saltwater’ barrier, meaning that the entire population of its target species – ‘Mainstream’ economists – is now at risk.

    Speaking on behalf of the WHO, Dr Cahuc explained that the virus works by turning off the one genetic marker that distinguishes this species from the rest of its genus, the Human Race. This is the so-called ‘Milton’ gene (Friedman 1953), which goes dormant in other Humans as they pass through puberty. Its inactivity reduces their imaginative capacity, making it impossible for them to continue believing in such endearing infantile fantasies as the Tooth Fairy and Santa Claus. While regrettable, this drop in imagination is necessary to prepare Humans for the adult phase of their existence.

    ‘Professor Milton Friedman found a way to re-activate this gene during PhD training, using his “as if” gene splicing technique’, Dr Zylberberg elaborated. ‘This enabled a wonderful outpouring of imaginative beliefs by Mainstream Economists, which gave birth to concepts like NAIRU, Money Neutrality, Rational Expectations, and eventually even DSGE models. This wealth of imagination was regarded by Mainstream Economists as a more than sufficient compensation for returning to the child-like phase of the Human species.’"

  3. Amazingly clever stuff by Blanchard. Just one problem though: he was chief economist at an organization (the IMF) which espoused macromedia during the crisis. I therefor do not have time for him.

  4. «the few who do instead decide to use a SEM should not be automatically shut out from the pages of the top journals.»

    And the lamb will lie down with the lion, and the german government will donate an extra 20% of GDP to the greek government in hard-currency, and London property prices will double every seven years forever, and the JB Clark medal will be renamed the T Veblen medal... :-)

  5. One disadvantage with calling them SEM's is that simultaneous equation models already use that acronym. Using the abbreviation in this way is relatively common in both mathematical economics and econometrics (for example when using 2SLS models to estimate something).

    Since the full name (structural econometric models) is a bit long to write out every time, might there be another name that would cause less confusion?

  6. Prof. Wren-Lewis, I'm beginning my PhD and I'm very interested in SEMs. My current work is with SFC models (which you already criticized). If I'm going to widen my research to SEMs in general, where should I start? Do you have a list of recommended literature?

    My best regards from southern Brazil.

  7. I would be very interested in hearing your suggestions for getting journal editors to be more receptive to SEM modeling papers. What are the key barriers to overcome?

    For example, I imagine that part of the attraction of DSGE models is that it is possible to focus a paper on a well-defined question. You start with a more or less accepted methodology and tweak it in some way. Even if there are serious problems with the methodology, the ability to start with an accepted model and add a feature and see how it works allows one to generate the kind of incremental progress that we tend to expect from research (and that helps to build careers).

    I don't know much about modern macro, but I guess that the situation is different with SEM models. Suppose, to build on the point you make in your post, that you estimate a consumption function and find that some variable not part of your theoretical framework seems to be important. It makes perfect sense for you to add it to your model. But since SEMs are built up in this kind of incremental way, it may not make sense for someone else to add it to their model. And a paper published about this result would at least appear to be of less general applicability than a DSGE result.

    So how do you think journal editors should respond? Should they dedicate an issue a year to SEM modeling? Or invite leading SEM builders to report out on their results on a regular (yearly?) basis, perhaps with an established reporting format? Or maybe invite a papers comparing, say, the treatment of investment or consumption across the leading SEM models (maybe if they all include your variable, that gives your approach added credibility).


  8. «‘Professor Milton Friedman found a way to re-activate this gene during PhD training, using his “as if” gene splicing technique’, Dr Zylberberg elaborated.»

    But that's just scratching the surface.

  9. This is my usual comment but with more typos because Ibam typing on a smart phone. I don't see the useful role of the initial dsge model.

    I think the reason it is wise to start with a DSGE model is diplomatic not scientific. It is easier to convince people to go somewhere new (or in this case old) step by step.

    I fear,however that the current unsuccessful DSGE models will surbive this and be considered a useful first approximation.

    I think there are already specialist literatures on roughly ever equation of standard DSGE models which include rejection of the standard specification and improved models if say the firm or of consumption.

    But they remain specialist and academic, and the old rejected equations are still used for forecasting and policy evaluation.

  10. Two questions:

    Q1. If SEMs - at least the traditional type - are known to be vulnerable to the Lucas critique, how can they still be used for policy analysis? Would they have forward-looking behavior?

    Q2. Any comment on the use of DSGE-VAR combinations as a possible third approach?


Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time.