Earlier this week, Derek Thomson, a senior editor at The Atlantic, began his article “The Graph That Proves Economic Forecasters Are Almost Always Wrong” with some observations that don’t really require a graph:
“As the saying goes: ‘It’s hard to make predictions. Especially about the future.’ Thirty years ago, it was obvious to everybody that oil prices would keep going up forever. Twenty years ago, it was obvious that Japan would own the 21st century. Ten years ago, it was obvious that our economic stewards had mastered a kind of thermostatic control over business cycles to prevent great recessions. We were wrong, wrong, and wrong.”
In a recent speech, Dennis Lockhart—whom most of you recognize as president here at the Atlanta Fed—offered his own thoughts on why forecasts can go so wrong:
“… you may wonder why forecasters, the Fed included, don’t do a better job. To answer this question, let me suggest three reasons why forecasts may be off.
“While it’s relatively trivial in my view, the first reason involves missing the timing of economic activity. An example of that was mentioned earlier when I explained that GDP for the third quarter had been revised down while the fourth quarter is expected to compensate.
“A second reason that forecasts miss the mark is, in everyday language, stuff happens.
“To be a little more precise, unforeseen developments are a fact of life. In my view, the energy and commodity shocks early in the year had a significant impact on growth in the first half of 2011. The tsunami-related supply disruptions, though temporary, were an exacerbating factor. In fact, a lot of shocks or disruptions are quite temporary and don’t cause one to rethink the narrative about where the economy is likely going.
“Which brings me to the third reason why economic prognostications go off track: we, as forecasters, simply get the bigger story wrong.
“What I mean by getting the bigger story wrong is failing to understand the fundamentals at work in the economy.”
“Getting the bigger story wrong” is Simon Potter’s theme in the New York Fed’s Liberty Street Economics blog post, “The Failure to Forecast the Great Recession“:
“Looking through our briefing materials and other sources such as New York Fed staff reports reveals that the Bank’s economic research staff, like most other economists, were behind the curve as the financial crisis developed, even though many of our economists made important contributions to the understanding of the crisis. Three main failures in our real-time forecasting stand out:
- Misunderstanding of the housing boom …
- A lack of analysis of the rapid growth of new forms of mortgage finance …
- Insufficient weight given to the powerful adverse feedback loops between the financial system and the real economy …
“However, the biggest failure was the complacency resulting from the apparent ease of maintaining financial and economic stability during the Great Moderation.”
Potter does not implicate any of his Federal Reserve brethren, but you can add me to the roll call of those having made each of the mistakes on the list.
Should we have known? A powerful narrative that we should have has taken hold. The boom-bust cycle associated with large bouts of asset appreciation and debt accumulation has a long history in economics, and the theme has been pressed home in its most recent incarnation by the work of Carmen Reinhart and coauthors, including the highly influential book written with Kenneth Rogoff, This Time is Different: Eight Centuries of Financial Folly.
Unfortunately, even seemingly compelling historical evidence is not always so clear cut. An illustration of this, relevant to the failure to forecast the Great Recession, was provided in a paper by Enrique Mendoza and Marco Terrones (from the University of Maryland and the International Monetary Fund, respectively), presented last month at a Central Bank of Chile conference, “Capital Mobility and Monetary Policy.” What the paper puts forward is described by Mendoza and Terrones as follows:
“… in Mendoza and Terrones (2008) we proposed a new methodology for measuring and identifying credit booms and showed that it was successful in identifying credit boom events with a clear cyclical pattern in both macro and micro data.
“The method we proposed is a ‘thresholds method.’ This method works by first splitting real credit per capita in each country into its cyclical and trend components, and then identifying a credit boom as an episode in which credit exceeds its long-run trend by more than a given ‘boom’ threshold, defined in terms of a tail probability event… The key defining feature of this method is that the thresholds are proportional to each country’s standard deviation of credit over the business cycle. Hence, credit booms reflect ‘unusually large’ cyclical credit expansions.”
And here is what they find:
“In this paper, we apply this method to data for 61 countries (21 industrialized countries, ICs, and 40 emerging market economies, EMs), over the 1960-2010 period. We found a total of 70 credit booms, 35 in ICs and 35 in EMs, including 16 credit booms that peaked in the critical period surrounding the recent financial crisis between 2007 and 2010 (again with about half of these recent booms in ICs and EMs each)…
“The results show that credit booms are associated with periods of economic expansion, rising equity and housing prices, real appreciation and widening external deficits in the upswing of the booms, followed by the opposite dynamics in the downswing.”
That certainly sounds familiar, and supports the “we should have known” meme. But the full facts are a little trickier. Mendoza and Terrones continue:
“A major deviation in the evidence reported here relative to our previous findings in Mendoza and Terrones (2008) is that adding the data from the recent credit booms and crisis we find that in fact credit booms in ICs and EMs are more similar than different. In contrast, in our earlier work we found differences in the magnitudes of credit booms, the size of the macroeconomic fluctuations associated with credit booms, and the likelihood that they are followed by banking or currency crises.
“… while not all credit booms end in crisis, the peaks of credit booms are often followed by banking crises, currency crises of Sudden Stops, and the frequency with which this happens is about the same for EMs and ICs (20 to 25 percent for banking and currency for banking crisis, 14 percent for Sudden Stops).”
Their notion still supports the case of the “we should have known” camp, but here’s the rub (emphasis mine):
“This is a critical change from our previous findings, because lacking substantial evidence from all the recent booms and crises, we had found only 9 percent frequency of banking crises after credit booms for EMs and zero for ICs, and 14 percent frequency of currency crises after credit booms for EMs v. 31 percent for ICs.”
In other words, based on this particular evidence, we should have been looking for a run on the dollar, not a banking crisis. What we got, of course, was pretty much the opposite.
No excuses here. Speaking only for myself, I had the story wrong. But the conclusion to that story is a lot clearer now than it was in the middle of the tale.