Illustration by Liz Meyer
Five Years After Crisis, No Normal Recovery
With the U.S. economy yielding firmer data, some researchers are beginning to argue that recoveries from financial crises might not be as different from the aftermath of conventional recessions as our analysis suggests. Their case is unconvincing.
The point that all recoveries are the same -- whether preceded by a financial crisis or not -- is argued in a recent Federal Reserve working paper by Greg Howard, Robert Martin and Beth Anne Wilson. It was also discussed in a recent article in the Wall Street Journal.
It is mystifying that they can make this claim almost five years after the subprime mortgage crisis erupted in the summer of 2007 and against a backdrop of an 8.3 percent unemployment rate (compared with 4.4 percent at the outset of the financial crisis). Our research makes the point that the aftermaths of severe financial crises are characterized by long, deep recessions in which crucial indicators such as unemployment and housing prices take far longer to hit bottom than they would after a normal recession. And the bottom is much deeper. Studies by the International Monetary Fund concluded much the same.
We have suggested that the concepts of recession and recovery need to take on new meaning. After a normal recession (which for the average post-World War II experience in the U.S. lasted less than a year), the economy quickly snaps back; within a year or two, it not only recovers lost ground but also returns to trend.
After systemic financial crises, however, economies of the postwar era have needed an average of four and half years just to reach the same per capita gross domestic product they had when the crisis started. We find that, on average, unemployment rates take a similar time frame to hit bottom and housing prices take even longer. With the Great Depression of the 1930s, economies on average needed more than a full decade to regain the initial per capita GDP.
“After the Fall,” a 2010 paper written by one of the authors of this article and Vincent Reinhart, a former Fed official who is now chief U.S. economist at Morgan Stanley, added evidence that in 10 of 15 severe post-WWII financial crises, unemployment didn’t return to pre-crisis levels even after a decade. It also showed that in seven of the 15 crises there were “double dips” in output.
What’s the best way to accurately calibrate recession and recovery after a deep financial crisis? It isn’t enough simply to establish when per capita GDP growth resumes, as economists have traditionally done to mark the end of a conventional recession. As we emphasized in our book, “This Time Is Different,” it is essential to measure where an economy stands compared with pre-crisis levels of important variables such as output, unemployment and housing prices.
There also is the interesting question of whether, after a deep financial crisis, an economy will ever fully reach its earlier trajectory for trend GDP, or whether some output capacity is lost forever. Researchers at the Organization for Economic Cooperation and Development found that the most likely scenario involves some permanent loss, though extrapolations over long time periods -- a decade or more -- are necessarily subject to a high degree of uncertainty.
What is striking about the Howard, Martin and Wilson Fed study is that it doesn’t really dispute our finding that recessions that follow financial crises are worse in terms of depth as well as duration. Instead, it argues that one should be more interested in a much narrower question: Once they begin, recoveries after financial crises are typically just as strong as those that follow a typical recession.
Moving Goal Posts
We admit that this time is different in one important respect: The goal posts many analysts use to assess economic outcomes seem to shift from data point to data point. When we first identified that financial crises were associated with severe recessions, the rosy-scenario crowd responded that the Great Moderation had smoothed the business cycle. Recessions in the new era were short and shallow.
After the crushing contraction, a new metaphor held that the harder the fall, the more vigorous the bounce back. Nonetheless, what followed was an anemic recovery that has yet to pull per capita annual real GDP back to the level of 2008.
Now, the staff of the Fed hopes to shift the goal posts yet again. Their advice is to forget about the problems of the past few years and focus on the coming expansion that they forecast using their own idiosyncratic interpretation of business-cycle dynamics across 59 countries.
Our results all key off the year of the systemic crisis event itself, a date which is relatively easy to establish. For example, in our research on Japan (like almost all studies of that country’s mega financial crisis), we date the crisis to 1991, though the asset-price bubble burst in 1989. The Fed paper dates the reference turning point for Japan as the second quarter of 1998, when GDP bottomed out. Dating the beginning of the recovery in this way (which also misses the point that 1998 was actually a second trough or double dip) helps support the Fed paper’s main conclusion that recoveries are all alike.
No doubt, 1998 will be remembered as a crisis year for many Asian economies (Malaysia, South Korea or Thailand) -- but Japan had been mired in crisis for nearly a decade. Similarly, Norway’s banking crisis unfolded in 1987; the Fed paper suggests that if instead we begin the comparison four years later, at the end of 1991, the recovery looked the same as other recoveries.
For the current crisis, we use 2007 to mark the beginning of the financial crisis for the U.S. and the U.K., and 2008 for most European economies. Our analysis looks at the grim macroeconomic picture in the years after the crisis year.
Calling the Trough
The Fed study asks a rather different question, one that requires defining and dating the ultimate trough after the fact. Unfortunately, calling the trough is a difficult task to achieve in real time, given the prevalence of double dips that Japan and several other countries experienced after a financial crisis.
For the U.S., the deepest unemployment trough (to date) in the current crisis occurred in the second quarter of 2009, when the jobless rate reached 9.5 percent. For several European countries, it is difficult to tell whether the deepest trough has been reached, despite the relief the European Central Bank has provided.
Looking ahead, does history suggest that the current U.S. recovery will be strong and robust, even if it has been so long coming? We hope this will be the case, but the jury is out. Precisely because there is often no bright line that marks a “recovery” stage after a deep financial crisis -- the kind of halting uneven recovery the U.S. has experienced is the norm -- considerable judgment is required in choosing turning-point dates (especially in real time).
An econometric definition of a turning point after a long financial crisis is an artificial construct, in contrast to our definition that looks at the years following a crisis.
Much work remains to be done to see if economies fully recover to trend potential GDP after a deep financial crisis, and just how many decades it takes (though even in the best case, it is certainly not a matter of a couple of years as would be true after a normal recession).
Is the U.S. on track to an ever brisker recovery in which the jobs numbers -- which have already turned from 100,000 to 200,000 a month -- start showing 300,000 or even 400,000 net new jobs a month?
Perhaps. We hope for good news at the end of the week when jobs numbers are released.
After all, there have been scores of financial, inflation and debt crises in virtually every country over the past 800 years. If financial crises were fatal, the world would hardly have experienced the kind of spectacular growth it has seen over most of the past 200 years, at least. But considering the huge and rising debt levels in the U.S., and the very limited extent to which deleveraging has taken place in the household and government sectors, we would be pretty happy to see a few straight years of trend growth, even if that falls short of the V-shaped recovery that some see around the corner. It might happen, but the historical evidence is at best mixed.
On the point of whether the aftermath of financial crises plays out differently than normal recoveries, the evidence isn’t mixed at all. As our studies and many others have confirmed -- including the misconstrued recent Fed study -- financial crises leave behind deep recessions of long duration and considerable volatility.
A recent paper by the IMF researchers Stijn Claessens, M. Ayhan Kose and Marco Terrones argues further that recoveries characterized by rapid growth in credit and house prices tend to be stronger. But that’s not happening: U.S. housing prices are still falling in many areas, and credit to small- and medium- size businesses is restrained. We hope for better ahead, but let’s not misconstrue the lessons of the past.
(Carmen M. Reinhart is a senior fellow at the Peterson Institute for International Economics in Washington. Kenneth S. Rogoff is a professor of economics at Harvard University. They are co-authors of “This Time is Different: Eight Centuries of Financial Folly.” The opinions expressed are their own.)
Read more opinion online from Bloomberg View.
Today’s highlights: The View editors on Myanmar’s elections and Marco Rubio’s immigration bill; Ramesh Ponnuru on why the U.S. will not go to single-payer health care; Betsey Stevenson and Justin Wolfers on the fallibility of economists; Margaret Carlson on the recall movement in Wisconsin; Carmen M. Reinhart and Kenneth S. Rogoff on why this crisis really is different; Eric Posner and Glen Weyl on creating an FDA for derivatives trading.
To contact the writers of this article: Carmen M. Reinhart in Washington at email@example.com; Kenneth S. Rogoff in Cambridge, Massachusetts, at firstname.lastname@example.org.
To contact the editor responsible for this article: Max Berley at email@example.com.
Bloomberg reserves the right to edit or remove comments but is under no obligation to do so, or to explain individual moderation decisions.