Scott Sumner  

Reallocation and growth

PRINT
Genetic Engineering Is Reprodu... Having Trouble Getting Qualifi...

I was recently chatting with Caroline Baum about "never reason from a price change" and she mentioned an example from 1986 that I had forgotten. Then I realized it also related to the concept of "re-allocation," which is frequently discussed by (former Econlog) blogger Arnold Kling.

In her book (Just What I Said, pp. 210-11) Ms. Baum pointed out that back in late 1985 and early 1986 the price of oil plunged from $32 to $10/barrel. The Fed cut rates by 200 basis points. Many pundits predicted an economic boom in 1986, but were shocked when growth came in at only 1.7% in the second quarter. (Yes, in those days 1.7% was far below trend---that was before the Great Stagnation.) She noted that slowing growth in the oil patch states (such as Texas) was a drag on overall GDP growth.

Of course we've seen another sharp drop in oil prices during late 2014 and early 2015, and thus far growth in 2015 has again disappointed the pundits. It occurs to me that this is one area where Kling's re-allocation theory makes a lot of sense. In much of macroeconomics, things are symmetrical. Thus if a Fed rate cut or a tax cut spurs the economy, we assume that a rate increase or a tax increase slows the economy. That's symmetry.

Re-allocation is different. The act of moving resources from declining sectors to growing sectors is costly and time-consuming--in either direction. Final output declines during the transition (even if individuals are making wise personal "investments" with their time during the re-allocation.) Re-allocation could occur due to falling oil prices, or rising oil prices.

When NGDP continues to grow at a reasonable rate, the impact of re-allocation on RGDP growth is likely to be modest. Thus in 1986 RGDP kept growing slowly and there was no recession. Ditto for the January 2006 to December 2007 crash in residential housing construction. RGDP growth slowed as workers re-allocated out of housing, but no recession. It takes a monetary shock that slows NGDP growth sharply in order to get an outright recession. In the case of the 2008 recession, the key mistake was the Fed's (perhaps unintentional) decision to adopt a tight money policy in late 2007.


Comments and Sharing


CATEGORIES: Macroeconomics




COMMENTS (7 to date)
Njnnja writes:

What is your take on the unit root question for real gdp? The re-allocation theory makes sense but when gdp bounces back does it make it back up to the original trend or are the costs associated with moving resources "lost forever"?

Market Fiscalist writes:

[Comment removed for supplying false email address. Email the webmaster@econlib.org to request restoring your comment privileges. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Ryan writes:

You write about reallocation as if it's a big, occasional event. That's a very odd perspective!

- Flows of resources associated with reallocation absolutely dwarf net numbers in ANY year: http://updatedpriors.blogspot.com/2012/11/facts-about-jobs-that-may-surprise-you.html

- A large chunk of productivity growth is accounted for by daily, pedestrian reallocation: http://updatedpriors.blogspot.com/2014/07/productivity-and-reallocation.html

- A lot of reallocation occurs within narrowly defined industries (Davis, Haltiwanger, and Shuh (1996), and subsequent research).

- The behavior of gross flows is very nuanced and cannot be accounted for by the representative agent framework. For example, job destruction is more cyclical than job creation, total reallocation is typically countercyclical (with the recent recession being an exception), the dispersion and skewness of the firm growth rate have cyclical (and secular) properties, and so on.


Kling's model is much more descriptive of modern economies than the general representative agent framework that you and so many others prefer, useful as the framework can be. I would concede that an understanding of how much reallocation matters for business cycles is at the research frontier. But I think the onus is on the rep agent people to show that macroeconomics is really so simple that we can understand it without paying attention to reallocation.

Becky Hargrove writes:

Ryan,
It depends on the sectors that are experiencing transition - and hopefully reallocation - at any given moment. Sure, reallocation is ongoing. But there were significant shifts for allocation in the mid eighties, when finance assumed more roles for wealth creation in lieu of the earlier traditional manufacture and production.

What's more, the major (needed) reallocation of the present remains in doubt. It has become more difficult to maintain services through the asset formation and consumption capacity which sufficed in recent decades, hence services need to be approached differently. AFAIK, these are the kinds of large transition problems which Scott Sumner is referencing.

Nathan W writes:

Don't most of these models assume that the change is instantaneous? This would lead them to predict immediate changes as opposed to the delays you refer to in reallocation.

Scott Sumner writes:

Njnnja, I haven't looked at that literature in a while, but I seem to recall they lumped together output changes from demand and supply shocks, in which case the estimates are not very useful. I believe the economy does bounce back from demand shocks, but not (most) supply shocks.

Ryan, You said:

"You write about reallocation as if it's a big, occasional event. That's a very odd perspective!"

I don't recall saying that, and I certainly don't believe that, for the reasons you indicate.

I do not rely on representative agent models. I think my model explain the past 10 years far better than any reallocation model.

Nathan, Just the opposite, the whole point of reallocation models is that the process takes time to play out.

Pemakin writes:

It's worth noting that this time around that the US is a high volume and marginal cost producer of oil, thus a decline in the price of oil is actually quite bad for GDP, despite claims of the "tax cut" nature of price declines. We tend to forget that the statistical impact of a change fron a large positive contribution to value added at $100+ per barrel to a modest negative near $50 can be be huge at the "margin".

Comments for this entry have been closed
Return to top