Some views from an economist who is not “advancing the science”.
Reader Joseph commenting on the stimulus bills brings my attention the Cato ad, wherein several Nobel laureates write against a stimulus plan. Ed Prescott is the Nobel laureate who won for work on growth/macroeconomics, so attention must be paid! I looked a bit for his reasoning, and only found this quote from the East Valley Tribune:
“I don’t know why Obama said all economists agree on [the need for a stimulus bill],” Prescott said. “They don’t. If you go down to the third-tier schools, yes, but they’re not the people advancing the science.”
So, I think it’s worthwhile to infer why Prescott is against the bill. I will refer to his very influential 1986 paper “Theory ahead of business cycle measurement”:
Economic theory implies that, given the nature of the shocks to technology and people’s willingness and ability to intertemporally and intratemporally substitute, the economy will display fluctuations like those the U.S. economy displays. Theory predicts fluctuations in output of 5 percent and more from trend, with most of the fluctuation accounted for by variations in employment and virtually all the rest by the stochastic technology parameter. … Theory predicts that deviations [from trend] will display high serial correlation. In other words, theory predicts what is observed … .
…
The policy implication of this research is that costly efforts at stabilization are likely to be counterproductive. Economic fluctations are optimal responses to uncertainty in the rate of technological change … .
The paper lays out in a fairly straightforward fashion the logic. I highly recommend reading it. However, I think it useful to lay out the logic in a simplified manner, such as I would teach it in an intro macro course.
Let output be given by:
Yt = ztkt1-θn t θ
Where z is a technology parameter, or “shock”, k is capital, and n is labor.
Households optimize intertemporally, and both demand and supply jointly determine output (in the crudest Keynesian models — but not the neoclassical synthesis — output is demand determined). Output then only moves in response to changes in z, k and n.
k might change if the user cost of capital changes. n might change as workers decide to alter their labor-leisure tradeoff, perhaps in response to changes in the real interest rate (or perhaps, there is an exogenous change in preferences of leisure against labor). Alternatively, z has changed. Let me focus on this, since that has been a central parameter in real business cycle (RBC) models. In this interpreation, we have since September 2008 experienced substantial technological regress. What the exact nature of this technological regress is remains open to question. One could say that the financial system is part of technology, and we’ve forgotten how to overcome asymmetric information problems previously handled by a banking system combined with regulation. Or perhaps the trend rate of growth in technology has changed. Remember in the original Prescott paper, in order to match the time series pattern of deviation of output from the Hodrick-Prescott filtered data, the autoregressive parameter was set to near unity; but the trend was subsumed into the HP trend.
In this world, output is essentially always at potential (i.e., CBO estimated potential GDP shown in this post is just plain wrong). So government spending would of course be counterproductive, since it would waste resources. Further, in this interpretation, all the current and incipient rise in unemployment is an optimal response to changing relative prices and prospects for technology growth. Hence, Prescott’s anti-stimulus view is definitely internally consistent with his views on the way the world works, and I laud him for being internally consistent (a scarce commodity in this world). But it is important to understand exactly how he thinks the world works before taking his prescriptions to heart.
Now, taking a less extreme view, and allowing for some bigger deviations from potential (say due to irrational expectations, or learning, or some rigidities), one could interpret current events as follows (inspired by a graph in Chad Jones’ Macroeconomics textbook):
Figure 1: An interpretation of the current downturn. Bold line is actual (log) GDP, short dashed line is perceived and long dashed line actual log potential GDP.
Here, what I think is a negative output gap is in reality a positive output gap, and in this case, a stimulus bill with spending and tax cut provisions only worsens the situation.
Now, as Jim has noted, technological frictions, say due to big changes in relative prices, could induce what looks like involuntary unemployment. And we’ve definitely seen changes in relative prices over the past few years. Hence, I don’t rule out an important role for technology, and maybe we can’t get to the pre-shock trajectory. But once one moves away from a model wherein one is always at full employment, then a role for countercyclical policy becomes plausible (note that Jim still thinks transfers to states, and infrastructure investment, makes sense. And I agree — if we could focus most of the spending on these items, and forego the tax cuts, I’d be even more in favor of the stimulus bills.) Nonetheless, I do think at least some of the decline in output in 2008Q4 apparently at a rate of 5% annualized (discussed here) is not attributable to a decline in potential GDP. I leave with a parting picture of the evolution of output over the last two years.
Figure 2: Log first difference of real GDP (blue) and log level of real GDP (red), SAAR, Ch.2000$. Source: BEA GDP NIPA release of 30 January 2009, and author’s calculations.
Technorati Tags: stimulus bill, recession,
real business cycle model, RBC, GDP, Ed Prescott.
I skimmed through the paper and I think I got your lesson. Whatever happened, the households over-corrected, and overestimated future cost of consumption. Did I get it right?
If so then:
We don’t understand if households had any choice in the matter. Given two methods of evaluating and consuming goods, we should expect using both methods simultaneously would be inefficient, trying to find water by taking two paths simultaneously.
If households did change behavior, or change their model; then households would initially have large variances in their estimation of consumption costs. The natural result would be under consumption until they become confident in the new method of consumption.
I don’t understand – is the sugggestion a possibility that the rate of technological change might be slowing down, and so potential GDP growth is slowing?
I do not agree that technological change is slowing down. When Moore’s Law stops, then we can being to consider that possibility.
I will say that most technological change in the world today is enabling developing regions to catch up rather than enabling America to forge to higher heights. This is because it takes large technological gains (like the Internet) for America to make small advancements, while small technological gains (like getting a laptop from $400 down to $200) can enable India and China to make large advancements.
Oh, Professor, Madison is first-tier amongst state schools!
Perhaps a simplistic question, but:
Wouldn’t k, capital, include access to credit? So the downturn needn’t involve any sort of “technology shock”, just a downward shock to the availability of credit, which we’ve certainly had since fall ’08.
“Remember in the original Prescott paper, in order to match the time series pattern of deviation of output from the Hodrick-Prescott filtered data, the autoregressive parameter was set to near unity; but the trend was subsumed into the HP trend.”
So, his true “technology shocks,” as opposed to secular productivity growth—the deviation of the z’s from their HP trend—tend strongly to go up once they are positive and to go down once negative. In English, that IS the business cycle.
So the “theory” explains the cycle by proving that it exists. Last I heard, tautologies were not considered an advancement of anything, much less science. Even the Medieval scholastics at least tried not to go in circles.
My shock theory starts with oil at $150/barrel which changed consumer behavior suddenly.
Zogby documents this in consumer surveys dating to May 2008 in which he found that the consumer adopted smart shopping, mainly planning each shopping trip ahead of time and using the car only for essentials.
Then the second shock hits when the consumer realizes that the fix cost of shopping, using this method, came to about $50/trip; including insurance, DMV fees, monthly car payments.
My narrative is still tautological until it explains why oil peaked and dropped so fast. My answer is that technology changed between this oil shock and the last, the consumer uses the new technology for smart shopping, and that technology did not exist in the last oil shock.
If smart shopping and dumb shopping co-existed then the two different methods, together, would have caused inefficiencies and contributed to an unexpected run up in oil prices, as well as contributed to the unexpected consumer response.
Why focus on technology? Seems irrelevant.
But you could still use the model in another way.
Rather than adapting to changing technology, people are adapting to a couple of things:
1) a change in appetite for risk, resulting in sharply higher risk premiums and thus interest rates.
2) a need for industry (and thus, individuals) to adapt to this new interest rate environment.
Point of fact, the industries that got hit first are the ones that relied on cheap credit: (homebuilding, automaking, etc.) The regions hit first were the ones most reliant on homebuilding for growth (the sun belt).
I have to agree with Matt Young that the oil price spike was the tipping point in consumption. Already consumers were aware that their homes (which many used as ATMs) were not as valuable as they once were; moreover, their 401ks and other investments were sliding rapidly. And the prospects of job loss (if not the reality) suggested limited opportunities for income growth. Household wealth was declining, and income was at risk. At the same time, there debt was extraordinarily high–and costly.
Gas at $4/gal & up was the final straw.
The person who has been badly mugged once–despite have walked the same dark street in a “bad” neighborhood at midnight for years–is unlikely to do that again for quite some time.
It’s behaviorial economics, and our behavior has been modified by a severe economic mugging.
I think the high gas prices lead to more congestion and worse fuel and time economy. We’ve spent more time and fuel getting less done over the past few years. Pile on that increased expenses and flat or falling incomes, you get a lot of uncertainty. James Hamilton’s post on energy shocks and recession is an interesting read.
Here is a paper on how congestion reduces job growth that I found the other day on Physorg.com.
You are being too kind, Professor.
Whenever people ask me about Prescott, I told them that he is the guy who believes that the Great Depression was caused by a sudden surge of workers’ desire to have more leisure and “even at the 1929 peak, stocks were undervalued relative to the prediction of theory.”
jg: Glad to hear you have high regard for the institution.
lilnev: In the model, capital is physical capital.
Buzzcut: Well, it’s not my theory. I was just trying to illustrate why Prescott opposes this stimulus bill. He’d oppose any stimulus bill, if he is working in the context of this model. The factors you highlight don’t show up his model (which, of course, doesn’t mean they’re irrelevant).
It has alway struck me that all these models postulating exogenous technological shocks are the dying gasps of a neoclassical theory that somehow still believes that perfect competition is a reasonable approximation of economic reality. When out there in the real world we have something conforming more closely to monopolistic competition (much of it sanctioned and provided by government and public policy), increasing returns not decreasing returns (if that wasn’t true there wouldn’t be even a hope of justifying CEO comp.) And of course an army of investors engaging in price discovery and hope of locating superior firms. Romer’s endogenous model seems much closer to reality and demonstrates the weakness of RBC models, at least in my opinion. Why else can we have a huge investment banking and investment management industry (where I spent my personal career.)
Look at the bright side. Since there is a positive GDP gap, we don’t have to worry about deflation.
About 2 years ago, Prescott was arguing in the WSJ that basically everything was going swimmingly in the economy, so it’s pretty safe to ignore everything he claims now.
Additionally, Arizona State isn’t exactly a first tier school, last time I checked.
OK, so how did z, the technology parameter change in this round vs the last?
Smarter use of cars because of more accurate goods maps and optimum routes, via the Internet. Consumers are much more efficient demand reducers.
So, we expect a similar response to an oil shock, but get a much more accurate consumer response, the consumer is better than we thought.
The consumer moves fast, demand drops faster than production responds. I say, good work, score one for the consumer.
Like a football game, the producers are on offense and will respond with much more efficient transportation, using the same technology.
Interesting snarky comment from Prescott. Especially coming from someone who did his Nobel Prize winning work while he was at one of those third tier business schools (currently ranked #51 according to the Financial Times). And where does he teach now? I think this just proves the old adage that first tier schools are oftentimes places where aging wunderkinds go to retire after they’ve accomplished their best work elsewhere.
Menzie, in your output function is there are necessary reason that the share parameters have to sum to 1.0? I realize this is often done for reasons of convenience, but is there any reason to a priori discount the possibility of something other than constant returns to scale?
The content is interesting, but I found the post a little whacky. Here’s my take: in most serious macro models, we’d find that frictions lead to departures from what you might call potential output (what you’d get without the frictions). Most of the NK work has this flavor. But typically potential looks nothing like the smooth trend line people like to draw. That’s useful to know: that comparisons with a smooth trend may not tell us anything useful. I don’t see that name-calling adds much to the argument.
If you’d like to see more, check out the Hall and Mankiw links at:
http://www.kc.frb.org/publicat/sympos/2005/sym05prg.htm
Good grief! I have NEVER seen so many 3rd tier schools and cow colleges on one list at one time! (And I speak as a grad of a very very old top tier school.) Appalachian State University? University of Charleston! (Now THERE is scrapping the bottom of the barrel! That place in WV barely hangs onto its accreditation.) Delta College? Spring Hill College? St Vincent? St Michael? Regent University? Western Michigan?
What the vast majority of those signers have in common is that they are Univ of Chicago grads. U of Chi is more than a bit wacky on its undying devotion to their great resident god Friedman – literally off the rails to the point of defying common sense. They are so enamoured of a theory (completely free markets) that has NEVER been accepted and applied in the manner which they theorize in recorded history and so busy debating how many angels dance on a had of a pin that they can not see the forest let alone the tree. The whole theory is based upon the false premise that humans are always totally rational and logical. The one thing that is unalterably true is that human behavior is neither completely logical nor completely rational. Start with a false premise and the the theory falls flat.
The U of C econ dept has this mass email that goes out to their former students to sign these petitions. (And yes, I know of some people who have ended up on the list and hit delete.) Want a reference from a former prof or colleague? It is strongly implied you had better sign.
Aside from the two good critiques of Prescott’s reliability, isn’t Prescott also a supply-side economist?
IMHO, one of the most important lessons from the Bush administration era was that some people treat their reputation as something to be cashed in to promote their viewpoint. Not ‘used’, not ‘leaned upon’, but ‘consumed’. Mankiw is a classical example.
Or, in short, just because they did good work before doesn’t mean that they aren’t lying now.
FU Mr. First Tier Prescott – may you decide not to work for the next 10 years. All of us stupid idiots in the third tier will celebrate with “Good Riddance” and hopefully cling to our jobs as the world forgets you.
What technology have your temporarily forgotten about lately? Your brain?