Real shocks/nominal shocks

This is part one of two posts on business cycles.  Both posts will examine one of the greatest mysteries on all of economics:  Why no mini-recessions?

I’ll get into mini-recessions in the next post, but first I’d like to examine another issue; why are US recessions always accompanied by nominal shocks?  And I’ll consider that issue by first examining recent events in Japan.  Here is the unemployment rate since July 2010:

A few issues:

1.  The graph is slightly off, the last two months are September/October, not October/November as it might appear.

2.  The figures from March to August do not include the regions devastated by the quake of March 2011.

Nevertheless, for two reasons I believe the evidence strongly suggests the quake did not significantly impact Japan’s unemployment rate.  First, because when the devastated regions were added to the total in September, the national unemployment rate actually fell from 4.3% to 4.1%.

And second, it was widely believed that the quake would cause lots of unemployment in Japan’s industrial heartland (Osaka to Tokyo), as supply lines were disrupted.  But it clearly did not.

Keep in mind this was a mindbogglingly large real shock.  The death toll was more than 10 times larger than Hurricane Katrina, and Japan is a much smaller country than the US.  The devastation was enormous.   Even today most nuclear plants are shutdown throughout Japan (only 11 of 60 are operating?), and electric power is rationed in some places. If this real shock didn’t affect the unemployment rate, what kind of real shock would?   Industrial production did drop sharply, but quickly recovered.  I’m not arguing that real shocks don’t affect output, I’m arguing they don’t affect jobs (very much.)

Some might argue that Japan is different, that Japanese firms don’t lay off workers.  OK, but let’s see what happens when Japan is hit by a demand shock:

It sure looks like Japanese firms do lay off workers, at least when demand falls.  The unemployment rate rose from 3.8% in October 2008 to 5.6% in July 2009.  That’s a big jump by Japanese standards.

And I’d argue the same for the US.  Almost all of the big jumps in unemployment in US history are due to demand shocks.  I only know of one clear exception in the post-war period: 1959, when a steel strike caused the unemployment rate to rise by 0.8%, and then fall sharply.  And guess what, 1959 also happens to be the only mini-recession in modern American history (that I could find.)  One real shock and one mini-recession.  Coincidence?  I don’t think so, but I’ll examine the mini-recession issue in the next post.

PS.  In fairness, there is one ambiguous case in the US; 1974.  NGDP growth did slow significantly during 1974 (compared to 1973.)  But NGDP growth was still fairly high, so it might be a stretch to call 1974 a nominal shock.  In my view the gradual removal of price controls during 1974 distorts the data, and the negative nominal shock in the latter part of 1974 was much bigger than it looks from reported NGDP data.  Others may disagree.  Thus 1974 remains a possible example of a real shock boosting unemployment sharply.

PPS.  The huge (20%) wage shock of July 1933 sharply reduced industrial output.  But it didn’t seem to affect employment, as it was implemented along with a rule that reduced the workweek from 48 hours to 40 hours, leaving weekly wages unchanged.



15 Responses to “Real shocks/nominal shocks”

  1. Gravatar of Benjamin Cole Benjamin Cole
    20. December 2011 at 08:41

    Excellent post. Again, market monetarism really seems to hold water. The anti-market monetarism arguments appear feeble, strident or politically motivated.

    And some people seem to love gold more than prosperity.

  2. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    20. December 2011 at 10:08

    Off topic (but any chance to make the Housing Cause Denialists squirm):

    The SEC’s disclosure debunks criticism leveled by David Min of the Center for American Progress and others against my analysis of the nature of the mortgage crisis. Min’s central objection was that I used “radically revised definitions for … subprime and Alt-A mortgages.”

    In the Forensic Study, subprime loans were defined as those with a FICO score of less than 660. Min described this as a “radical definition” of subprime as it resulted in Freddie’s loan guaranty portfolio having $243.3 billion in subprime loans at June 30, 2008. It now turns out to be virtually identical to the subprime loan total of $244 billion that the SEC calculated.

    He who laughs last….

  3. Gravatar of Matt Waters Matt Waters
    20. December 2011 at 10:17

    On the 1973-74 shock, the iron-clad COLA’s for strong unions probably throw off the usual NGDP/unemployment dynamics. If a lot of workers have strong COLA’s, then the wage discontinuity happens at the inflation rate instead of at the usual 0%. In that case, RGDP <0% leads to higher unemployment even if NGDP growth is strong. Strong unions with ironclad COLA's (regardless of NGDP growth) may also explain the 1959 shock.

    For Japan, I do know that there's a more cooperative relationship between companies and unions. So if Japan suffered a real shock where NGDP went up 5% but inflation went up 10%, the unions wouldn't demand a 10% wage increase for inflation when their employers' revenue (i.e. NGDP) went up by 5%. American unions of the 50's, 60's and 70's certainly would.

    That's all horribly unempirical, but it made the most sense to me. I never really understood why Friedman and others talked about "rational expectations" for wages during high inflation instead of going straight to strong unions as the cause of high unemployment admist high inflation.

  4. Gravatar of Matt Waters Matt Waters
    20. December 2011 at 10:34


    I have no idea what that link is getting at exactly. I don’t doubt that Fannie and Freddie made a lot of subprime loans through various means. It’s just that private investors/institutions also made a lot of subprime loans. The “cause” is a mix of public and private, as subprime loans allowed people to essentially buy call options on housing values for a very low premium.

    The point remains that whatever caused the housing bubble did not cause 10% unemployment. Whoever caused the housing bubble cost taxpayers a few hundred billion dollars. That sucks, but it’s 2% of one years GDP and 0.66% of three year’s GDP. There should be recriminations about that cost, but it’s not what caused high unemployment.

  5. Gravatar of Luis H Arroyo Luis H Arroyo
    20. December 2011 at 12:35

    Very good post, Scott.

  6. Gravatar of Scott is right: Recessions are always and everywhere a monetary phenomenon – just look at QRPI « The Market Monetarist Scott is right: Recessions are always and everywhere a monetary phenomenon – just look at QRPI « The Market Monetarist
    20. December 2011 at 15:53

    […] Sumner has a couple of fascinating posts on recessions on his blog (see here and […]

  7. Gravatar of ssumner ssumner
    20. December 2011 at 19:27

    Thanks Ben.

    Patrick, You know I always like GSE bashing.

    Matt, That’s a very good observation, although I’d say more for 1974 than 1959. I think 1959 was just the steel strike–it was incredibly short.

    Thanks Luis.

  8. Gravatar of 実質ショック、名目ショック by Scott Sumner – 道草 実質ショック、名目ショック by Scott Sumner – 道草
    20. December 2011 at 23:32

    […] by Scott Sumner // Scoot Sumnerのブログから “Real shocks/nominal shocks“(20. December […]

  9. Gravatar of Bogdan Bogdan
    21. December 2011 at 07:31

    Hmm…I am not convinced by this example of real or supply side shock. If all demand side shocks might be alike, that might not be the case with all supply side shocks 🙂 Specifically, in the case of the recent Japan earthquake, the suplly shock might have triggered an increase in velocity and a reduction in savings, as those affected, relatives and donors elsewhere spent more in order to rebuild the area, but also a small fiscal stimulus in government expenditures. This is different than a suplly shock trigered by a deterioration of the business environment, due to various price controls, burdensome regulations, higher taxes, capital obsolence etc

  10. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    21. December 2011 at 10:17

    ‘The point remains that whatever caused the housing bubble did not cause 10% unemployment. Whoever caused the housing bubble cost taxpayers a few hundred billion dollars.’

    That’s debatable. Remember the story of catastrophe that begins, ‘For want of a nail a shoe was lost…’

    However, Pinto’s talking about, ‘my analysis of the nature of the mortgage crisis’, which has been a topic of discussion here in the past.

  11. Gravatar of o. nate o. nate
    21. December 2011 at 10:26

    ‘The point remains that whatever caused the housing bubble did not cause 10% unemployment. Whoever caused the housing bubble cost taxpayers a few hundred billion dollars.’

    Don’t forget about the existence of a financial system with leveraged exposure to the housing bubble. It’s not too hard to discern a chain of events that over the course of 2 years or so leads from a housing bubble collapse to a collapse in banking firm equity to a financial crisis to high unemployment.

  12. Gravatar of ssumner ssumner
    22. December 2011 at 07:14

    Bogdan, But velocity is a demand side factor, so even if you are right it shows the importance of demand. Keep NGDP on target, and employment will be pretty stable.

    o. nate, It’s very hard to discern a chain of events if the Fed keeps NGDP growing at 5% (stable monetary policy.)

  13. Gravatar of Edwin Herdman Edwin Herdman
    22. December 2011 at 09:07

    Awesome post, thanks.

    In response to the “Japan is different” angle – the historical literature on Japanese business mainly mentions the propensity of companies to offer lifetime employment (essentially) – there is some term along the lines of Wa that I forget, but the idea was that social harmony would be maintained if the dedication of workers was reciprocated by the loyalty of their companies. It is no exaggeration to say that firings have traditionally not been used lightly in Japan.

    I’ve been rooting through my few bits of Japanese business literature to no success on more references to support that point. However, the fact that firing is found as a response to demand shocks “even” in Japan reinforces Mr. Sumner’s point very well.

    I’m not sure how much the change to free firing has upset the system. It is interesting to note that today the earning disparity between top management and the lowest paid assembly line worker was said to be around 10 times only (I think that’s the case for Sony until fairly recently at least, for example), and management – rank/file earnings disparity is still much less than it is in the West generally.

    Matt Waters: “For Japan, I do know that there’s a more cooperative relationship between companies and unions.”

    Just to ram that point home, look at this:

    “Shunto [“spring offensive”] wage bargaining has constituted one of the main foundations of union prestige in Japan. Though it is none too clear that shunto really raises wages much- during its golden age in the 1960s, high growth and labor shortages accomplished that, and it afterward functioned more to facilitate wage restraint-it nonetheless has had a major impact on the consciousness of workers and the public.


    The most interesting sector in 1997’s shunto was the auto industry, whose prosperity made it the clear union choice for chief pattern-setter (unions in other sectors are then supposed to try to match the settlement). Toyota, which is making huge profits, adopted a tough stance about holding down wage costs, then surprised the business world when, just before the settlement deadline, it unilaterally increased its offer by 100 yen. Having already agreed to 9300 yen, the Toyota union representative either “was at a loss for words” or gushed “tears of gratitude,” depending on which newspaper you read. In fact, the Japan Auto Workers Union, the auto industry’s industrial union, was embarrassed, to say the least, since its propensity to demand less than management can afford to pay was a bit too clearly exposed. A new black joke making the rounds of the industry was that the workers now trust Toyota more than they do their union.”

  14. Gravatar of Keith Keith
    22. December 2011 at 10:23

    Are not the demand and supply “shocks” that you describe really two sides to the same coin?

    Starting with the definition of a recession as a decline in nominal income, I’ll assume that unless that nominal decline is purely monetary (increasing value of the unit of measure) that the recession is caused when some quantity of capital and labor can no longer generate income.

    In the case of Japan, some capital and some labor were physically destroyed and others temporary sidelined because of a natural disaster (supply shock). Since the capital was useful and desired, significant effort is being exerted to get the idle resources back up and the destroyed replaced. I suspect this effort minimizes the effects on employment. (Of course, 1000s of people were killed so employment, as opposed to the unemployment rate, cannot return to prior values without immigration, previously unemployed labor moving into the labor force or a significant time lag. But this is small in comparison to the whole of Japan.)

    In the case of the US, some capital and labor were idled because of a lack of demand. But, I assume that the original source of this was the housing sector and that the demand shock was initiated by an idling of supply. A sector of the economy could no longer generate income not because it was physically destroyed but because its economic value was effectively destroyed from over-expansion and production: and not just houses but everything in the housing pipeline. Real and nominal income, and hence demand, declined. And, of course, monetary policy may very well have increased the gain on the downward cycle/snowballing. (In nutshell, there is circularity in the “demand” shock; its source was over-supply/investment on a large scale.)

    So why would one lead to a larger and longer impact on employment? In my opinion, when capital loses value because of bad decision making, it takes time for the marketplace to sort out where replacement investment opportunities lie and then to direct resources toward these opportunities. In a natural disaster, the losses are obvious and replacement can begin immediately.

    Create an environment that hinders the investment needed to replace lost real income, distort the price signals that are needed to direct new investment and you will drag out any recovery from whatever source.

    A mini-recession would require investors and business to know where and how to direct replacement investment basically before the recession began and have the means to implement that investment very quickly: i.e., like a large strike.

    Two cents from a modeler.

  15. Gravatar of ssumner ssumner
    23. December 2011 at 07:27

    Edwin, I agree.

    Keith, You said;

    “Starting with the definition of a recession as a decline in nominal income,”

    Here’s your problem; this is not the definition of a recession. Recessions are when real GDP falls.

    The housing slump did not significantly increase unemployment, it didn’t even create a mini-recession. That’s because during 2006 to early 2008 NGDP kept growing, so workers losing jobs in housing would find them elsewhere.

Leave a Reply