WHEN ECONOMISTS JUST DON’T KNOW
Ol i vier Blanchard — the respected Massachusetts Institute of Technology economist and former chief economist at the International Monetary Fund — recently circulated a short memo with the title, Do DSGE Models Have a Future? It’s written for specialists ( that is, the sort of people who already know what DSGE models are), but Blanchard’s message to macroeconomists deserves a broader audience. Not only do we not fully understand what’s going on in the world economy, our usual tools for figuring things out don’t seem to be working, either.
Notwithstanding some excited claims to the contrary, the Great Financial Crisis ( GFC) of 2008- 09 did not invalidate everything that economists have learned over the centuries: this was neither the first financial crisis we’ve ever seen, nor is it likely to be the last. On the other hand, it is fair to say that the last few years have been difficult for macroeconomics and macroeconomists, but perhaps not for the reasons you might have heard. The problem isn’t understanding how the crisis occurred or the policy response: it’s understanding the world the GFC left in its wake.
Not all elements of the crisis t ook economists by surprise. The U. S. housing bubble of the mid-2000s was a widely acknowledged phenomenon and the anticipated crash in housing prices was expected to reduce household wealth on a scale comparable to the bursting of the dot- com bubble. That’s pretty much what happened at first: the U. S. recession started in December 2007, on a scale comparable to the mild 2001 recession that followed the dotcom crash. And just as in 2001, the U. S. recession hardly affected Canada at first.
The real shock was the financial crisis in the fall of 2008. Again, the possibility that a housing crash could lead to a financial crisis wasn’t something that had never occurred to economists. Indeed, Canada had already experienced a smaller- scale banking crisis when the collapse of the Alberta housing market in the 1980s led to the failure of the Canadian Commercial and Northland banks. As long as financial market regulators were doing their job — and checking on that would require access to confidential data — then there was reason to believe that 2008 would be a rerun of 2001.
As it turned out, U. S. regulators dropped the ball. But since this wasn’t the world’s first financial crisis, policy- makers knew what had to be done: supply financial markets with liquidity and recapitalize the banking sector. Up to about 2010 or so, we were in known — if unpleasant — territory.
The puzzles f acing macroeconomists are understanding why economic growth has been so sluggish since 2010, and figuring out what — if anything — can be done about it. The issue isn’t a lack of imagination on their part: conjectures and policy recommendations abound. The problem is how to choose among the different explanations that have been proposed.
The obvious answer would be to pick the theory that best fits the data, but here’s the thing: macroeconomic data are notoriously uninformative. There are only a limited number of robust correlations in the data that can be used as a filter to reject candidate theories, and some of these correlations appear to have weakened or disappeared in the past five years. Many theories can plausibly claim to fit the available evidence, such as it is. Since the data are so weak, more information must be brought to bear. Theoretical rigour is an obvious source of nonsample information: not all theories are equally solid in their own internal logic.
Blanchard’s topic is DSGE models, which have become the standard theoretical tool in macroeconomic analysis: Dynamic Stochastic General Equilibrium. DSGE models have many attractive features. They are “dynamic” in that they take into account the linkages between current decisions and future outcomes: money spent now is not available in the future, and current investment increases future productive capacity. The “stochastic” part means that they take into account the presence of unpredictable shocks, and these affect peoples’ assessments of uncertainty and risk. And “general equilibrium” basically means that the individual components of the economy add up to a coherent whole.
The problem is that DSGE modelling techniques have, so far, been unable to develop convincing explanations for the post- GFC world, either at the level of the global economy, large economies such as the U. S. or the European Union, or small open economies such as Canada. For example, the Bank of Canada’s standard model has been continually issuing almost identical forecasts of a return to “normal” over the next 18-24 months since 2011.
Unlike some commentators, Blanchard doesn’t recommend abandoning DSGE models entirely: any useful model will have to take into account time, uncertainty and the interactions between different sectors of the economy. As he puts it, “current DSGE models are seriously flawed but they are eminently improvable and central to the future of macroeconomics.”
But until these improvements have been made, macroeconomic debates are going to have to be more open to different sorts of approaches and to different ways of making use of the data we do have. Perhaps most importantly, we’re going to have to accept the fact that we still don’t really understand what’s going on. And until we have a working theory, we won’t really know what should or can be done to improve things.
ECONOMIC GROWTH IS STUBBORNLY REFUSING TO RETURN TO “NORMAL.” AND WE AREN’T REALLY SURE WHY.