Krugman’s Excellent Column: Who’ll Stop the Pain

As regular readers of this blog know, I often disagree with Krugman. He believes a raft of things about fiscal and monetary policy that I simply don’t. I still think he is one of the best economists of our century and yesterday he proved once again that he understands the world quite well, maybe better than anybody else.

Take a look at this column. You have to read the column yourself to get the full force of his ideas. Basically, he manages in a little over 800 words to summarize both the depth of the current downturn and the seeds of the eventual recovery. Spoiler Alert: It’s not fiscal policy.

Go read the column. Then come back and read my thoughts below.

Since, I trust, you have already read his column, I am going to take Krugman’s points somewhat out of order and rearranged. All of the words belong to him.

To appreciate the problem, you need to know that this isn’t your father’s recession. It’s your grandfather’s, or maybe even (as I’ll explain) your great-great-grandfather’s. Your father’s recession was something like the severe downturn of 1981-1982. Your grandfather’s recession, on the other hand, was something like the Great Depression, which happened in spite of the Fed’s efforts, not because of them. The closest 19th-century parallel [your great-grandfather’s recession] I can find to the current slump is the recession that followed the Panic of 1873. That recession did eventually end without any government intervention, but it lasted more than five years, and another prolonged recession followed just three years later.

This certainly puts the current downturn into perspective. Like me, Krugman believes that the current recession is much worse than 1980 but not nearly as bad as the Great Depression. He picks the recession of 1873 as the appropriate historical comparison.

I am a fan of macro data but I don’t quite have a good feel for 19th century data. So, I went to the NBER and pulled their Index of American Business Activity (you can find it here). In my mind, this series is an excellent proxy for industrial production. The series is monthly and runs from 1855 through 1970 and, over the years where it overlaps with the Fed’s industrial production series, the correlation in the 12-month growth rate of the two series is 0.97.

Plotting the whole series is too noisy. Instead, I plot the cumulative loss in output during every major downturn between 1855 and January 2009. I define a major downturn as any period in which the 12-month growth rate of IP is negative for a sustained period. Cumulative output loss is defined as the trough (the lowest point locally in the series) over the highest level achieved in the previous two years.


The 19th century was not a fun time. There were 15 major downturns between 1855 and 1900 compared with only 8 between 1955 and 2000. On average, in the last half of the 19th century a major decline in output occurred once every three years: when did they find time to grow! Output fell an average of 17 percent in each of these episodes. This is half the output decline experienced in the last half of the 20th century.

There are many reasons for the higher volatility in the 19th century. First, the data is likely not as accurate as in the later period. Although, the long term averages should eliminate a lot of the noise in the data. Second, the economy was much more agriculturally oriented. Agriculture by its nature is more volatile and this volatility would automatically feed into industrial output. Third, the Civil War was a rather major event and may have had a large influence. Finally, the late 19th century was a time of rapid expansion, particularly of geographic expansion. The fits and starts of settling the West may also have had a large influence on the timing and size of downturns. Nonetheless, the mere fact that we are now comparing the current downturn to data from the 19th century is scary.

Returning to Krugman’s editorial, I plot just the four recessions he discusses: 1874, 1930, 1980, and 2009. The Great Depression occurred on an unimaginable scale, simply shocking. I think what surprises most people (as opposed to people reading economics blogs on the internet) is that the current decline in production is already much larger than the declines in the 1980 and 1982 recessions. The cumulative decline in output through January is 13 percent.

Thirteen percent is a large fall but I now believe the odds are exactly even on a cumulative fall in output in this downturn exceeding 20 percent. (Note: In production terms, this meets my definition of depression. However, this does not mean the economy as a whole would meet the definition. So far the overall decline has been moderated by a fall in imports and a relatively stable services sector.)

After thoroughly depressing us, Krugman offers us a glimmer of hope. He points out that the rate of adjustment is indeed sufficient to bring us (eventually) out of the recession. Economies tend to recover (not always – think of the middle ages – but they tend to) and ours is likely to recover as well.

What, then, will actually end the slump? The seeds of eventual recovery are already being planted. Consider housing starts, which have fallen to their lowest level in 50 years. That’s bad news for the near term. It means that spending on construction will fall even more. But it also means that the supply of houses is lagging behind population growth, which will eventually prompt a housing revival. Or consider the plunge in auto sales. Again, that’s bad news for the near term. But at current sales rates, as the finance blog Calculated Risk points out, it would take about 27 years to replace the existing stock of vehicles. Most cars will be junked long before that, either because they’ve worn out or because they’ve become obsolete, so we’re building up a pent-up demand for cars. The same story can be told for durable goods and assets throughout the economy: given time, the current slump will end itself, the way slumps did in the 19th century. But recovery may be a long time coming.

The 2001 recession was characterized by a capital overhang. The capital stock, particularly in things like fiber optics networks, was too high. Investment fell and eventually the stock fell sufficiently that investment recovered, although the recovery was tepid and investment growth never returned to its late 1990s levels. This adjustment was aided by the very high depreciation rate on high-tech capital goods.

This recession, in my mind, is characterized by an overhang of durable consumption goods. There are too many houses, cars, and flat-screen TVs. Just as investment fell in 2001, consumption has to fall to work off this overhang (see my thoughts on consumption here). The fall in building and motor vehicle production are part of this adjustment. We are closer to the bottom now than we were before the last month’s worth of data. Overhangs take a long time to work out.

We are still a long way from the bottom and once we get there, if I am right about the path of consumption going forward, the recovery will be more likely slow and gradual rather than the sharp growth acceleration one typically sees in recoveries.

Who’ll stop the pain? We will. The pain will end when we have worked off our excess durable consumption goods and repaired our balance sheets. Then we will move forward a bit more chastened a bit more cautious and a bit less ready to finance our consumption.