Brad DeLong mocks Steve Horwitz here for suggesting that the stimulus didn’t create jobs. DeLong then launches a defense of Christina Romer and her thoughtful examination of the data for crowding out caused by the stimulus package. Brad says there isn’t any.
He argues that higher interest rates and wages are how we observe the “unseen” in economics–jobs, for example, might be being lost from the stimulus due to crowding out. We observe those losses via the change in prices.
Because interest rates haven’t risen, there isn’t any crowding out of private investment. And because there’s no “wage inflation” there’s no crowding out of employment.
Then there is a charming analogy between jobs and neutrinos as if Romer and other careful economists were like white-coated physicists working at an accelerator facility.
Then he points out that of course there isn’t any crowding out in employment:
That there hasn’t been any wage inflation-induced crowding-out is, to me, not surprising: the fracking unemployment rate is ten percent, after all. The full-employment world of Bastiat is very very far away. And relative to the magnitude of excess supply in the labor market, the stimulus is simply not big enough to induce any wage inflation-based crowding out. So I am not surprised that there has been no wage inflation-induced crowding out: that is a normal and expected part of our empirical reality.
But there’s a problem with that analysis. The unemployment rate is 10% for every industry in every place. If you don’t have a high school diploma, the unemployment rate in December was 15.3%. If you had a college degree or higher it was 5%. So when the “stimulus” package increases spending on medical research by millions of dollars, it doesn’t do much for the unemployed who were in the construction industry. It does push up the demand and wages for medical researchers. I suspect there’s some crowding out there. It’s hard to measure but I bet we could at least see some of those wage or at least income increases that Brad says haven’t happened.
There is no evidence that the hundreds of millions of dollars that the government has borrowed to spend on the “stimulus” package has created any jobs at all. If Brad has that evidence, I’d sure like to see it. Remember, it was Christina Romer who said that without the stimulus, the unemployment rate would reach 9%. With it, we’re hit 10.2%. Of course it could have been even worse without the stimulus. But what is the evidence for that view, Brad, other than your desire that it be true?
Share
Podcast RSS Feed
Full EconTalk Text
{ 89 comments }
← Previous Comments
Okay. I assume someone will correct me if I'm wrong, but Russ believes:
There is no way AT ALL to measure the effects of the stimulus. There is no way to deterimine a legitimate (that Russ would aprove of) marker and say, “If this number reaches X, the stimulus was a success”? Russ believes that the effects of the stimulus–good or bad–can not be known.
I don't think they can be avoided in a post mortem of past decisions because when one outcome is realized, all other possible outcomes become theoretical. Therefore, if one is to discuss past decisions, counterfactuals are unavoidable. Of course, there are obvious drawbacks to counterfactuals – for instance, unintended consequences of the alternate path can't be compared to the known unintended consequences of the path taken.
Counterfactuals are deeply flawed, but I don't think there is a better alternative.
With logical blunders like this:
“”Even though they cannot see jobs that were never created, they nevertheless can count them–and are trying to do so.”
Delong should rename his blog Counting on Nothing With Both Hands: The Numbers Get Higher. There is a very distinct difference between objects that are difficult to measure and what simply isn't there, absurd appeals to subatomic particles aside.
Oh wow – initially I didn't realize you made that big of a claim! (that Obama's “created or saved” figure doesn't net out jobs lost from the stimulus).
You're quite simply mistaken – the modeling used to produce the “created or saved” number is based on an explicit consideration of net employment change. All multiplier estimates do this too, but for GDP (that's why you can have multiplier estimates of less than one). Are these estimates that can be disputed? Sure. And I've always agreed that any specific number produced while we're still in this should be taken with a grain of salt.
But you are factually in error when you say “What it doesn't consider are the jobs lost due to the very policies that are “saving” jobs”. Unless you have a very different understanding of the modeling that goes into these estimates than I do, they absolutely do consider this. The estimates would be pointless if they didn't.
I'm very concerned that you fundamentally don't understand the models that produce these estimates that you're commenting on.
Now I'm worried about arguing with an economics professor, and perhaps I have it wrong.
Let me ask it this way: what evidence do you have that they are looking at gross employment effects and not net employment effects? Surely you must have some reason for thinking this is truly the case, or else you wouldn't make such a damning accusation. But you don't mention your reason for thinking that, and I'm left confused. I was always under the impression that these estimates are net employment change estimates that do consider jobs lost. Why are you under a different impression?
Perhaps there isn't an alternative (I certainly wouldn't be able to discover one), but those used in economic matters should at least have stronger premises, and conclusions that follow from those premises more tightly. I should qualify this as a personal opinion. I don't find counterfactuals convincing.
The problem is that we can't control the variables as in a scientific experiment. A complex dynamic system has multiple constantly and unpredictably changing variables. Impossible to model.
I didn't say that (or write it). Though, I believe Group 3 is wrong based on my belief of how incentives work. Wasn't even Krugman disillusioned by how the political process allocated the last stimulus?
Why again is it a bad thing to let individuals decide?
“Impossible to model.”
Sam, would you say that most economists would agree with that? I'm a physician and aquatic gardener/aquarist, so I understand the inability to predict complex systems, and how a change in one variable effects dozens (or hundreds) of others.
But would Krugman would say “Impossible to model”? To me, intuitiviely, I agree with you, but I'm still learning the fundamentals of macro.
Truly, I'm stunned, if what you're saying is true–that no one has a way to determine its effects–it it really all politics, ideology, confirmation bias, etc.
Aha – I see. When you ended that sentence with “were wrong” I thought you were refering to the third group. You were refering to their perspective on the policy.
Anyways – my answer still remains that I think they would say it did help just not as much as it could have.
I don't know why it's a bad thing to let individuals decide. Decide what? I think it's a good thing to let individuals decide things, but I'm not sure what you're refering to.
the strength of the premises for counterfactual both varies and is highly subjective. I don't disagree that counterfactuals are very problematic and I agree that conclusions drawn from them are not very useful for that reason.
I should qualify: Impossible to model with any significant degree of assurance that it represents reality.
The real problem is that unpredictable things happen with great frequency. Crops fail because the jet stream shifted a bit. Someone develops an idea into a significant product. Stuff happens that can't be predicted.
When they make models, they are necessarily simplifications of the real world and they certainly aren't prophetic.
They quite literally don't know what would have happened without the stimulus, except that some large, influential institutions would have gone out of business.
The main purpose of the stimulus is to instill some confidence in people that the government is taking care of the problem and we can all go about our business as usual.
What they are doing is borrowing from the future.
DeLong's claims that “Economics isn't rocket science. Rocket science is harder.”
Well, I am a rocket scientist, and I bet Steve could learn my discipline a hell of a lot faster and better than I could learn his.
“The main purpose of the stimulus is to instill some confidence in people.”
Sure, animal spirits. What I'm asking is: Do they–democrats or Keynesians–believe that is the main purpose? I don't know if they think that. They believe it will help in some tangible way, independent of a type of economic-placebo effect, per se. The notion of a “pretense of knowledge” is what Hayek promoted, but not everyone emphasized this, regardless of its validity.
Do Keyneisans THINK that they have worthy models? I saw a poll that said 90% of economists oppose minimum wage. Is the consesus for the inability to predict the effect of the stimulus as high? Or is Russ's perspective also ideological–whether right or wrong? Is the argument that “no one knows” a perspective, or is it a generally agreed upon idea in economics, like free trade and comparative advantage?
A Hit n Run post on stimulus jobs.
I should not have tried to do economics late at night, for, as I’m sure Daniel would tell you, my mind is fuzzy enough in the daytime.
De Long says that since scientists knew of the neutrino even without having observed it, they could count it without having observed it. But the conclusion doesn’t follow from the premise. Knowing of the neutrino didn’t mean you could count it. And even if you could count the neutrino, or jobs created or not, that was still just history and not economics.
The economics of Mises that I referred to above was not exactly right either. Jobs created by the public sector are not necessarily at the expense of jobs in the private sector. If the private sector, paralyzed by fear, wasn’t creating many jobs, and the public sector took up the slack, the jobs it would create would be in addition to and not at the immediate expense of jobs in the private sector.
But that would still just be an artificial stimulus and actual depressant. The only real jobs are in the private sector, and phony jobs in the public are just more of the poison that made us sick in the first place.
The bust is not what caused unemployment. The cause lay in policies preceding it, such as minimum wage laws. Without them, the bust would simply have resulted in some capitalists losing their investments and their displaced workers moving on to other, lower paying jobs.
While inflation had masked the unemployment effect of the preceding policy, the minimum wage law, deflation would expose it again. But that was not the same as causing it. It was still the preceding policy and not the deflation that was the underlying cause.
What neither De Long nor Horwitz understand is that recession and deflation are the market purging mistakes and laying the groundwork for recovery. Whatever the pain, it is like that of withdrawal from drug addiction, not the sickness but the cure. No withdrawal, no recovery, no pain, no gain.
Prof O'Driscoll asked,
“How do debt deflations end?”
That’s like asking how withdrawal from drug addiction ends.
It ends when it’s been completed, when you've purged the system of all the poison, and all the “remedies” that are just more of it.
The underlying fear expressed by O’Driscoll is that the downturn would go on indefinitely without the “gov” stepping in to halt it.
But how could it do so?
Only with money it had taken from that part of the private sector that remained healthy and strong. But why throw its good money after the bad, sacrificing the stronger for the weaker? Why not just let the stronger displace the weaker?
If the aim was to keep the weaker from dragging down the stronger, it made no sense then to drag them down ourselves, nor, to spare fools their soaking, to drown everyone else.
For my more detailed analysis of this, see
http://econotrashtalk.org/#The_Cause_and_Cure_o...
Decide how to spend their own money rather than let Washington do it for them.
George Washington is dead and Washington D.C. is a city, not a decision making entity.
Are you asking me whether people should be allowed to individually decide on how to dispose of their money OR decide come together in a pre-determined manner and place and decide together what they should do with their money? I think individuals should be able to make decisions about their money, and therefore I think both options are good. I don't see them as inherently contradicting each other.
kurlos,
Sullum at Reason has a post today that provides info on how the Administration is playing fast and loose with the numbers. My take from it all is that O and Co. have no idea about jobs saved or created. By their own numbers though it adds up to hundreds of thousands of dollars per job they claim to have created. I can't fathom how that can be a good thing. http://reason.com/blog/2010/02/03/theyre-not-ne...
Of course there’s no crowding out. Brad simply deleted it from his comments section! Poof. No more crowding out.
Prof. Horwitz wrote,
“To be honest, I will not debate DeLong. He has shown himself incapable of civil dialogue and utterly uninterested in the ideas of those with whom he disagrees. He's a dogmatist in the worst sense.”
He can call me any names he wants, so long as he doesn't censor me, as the both of you have done. While his names wouldn't hurt me, and I take your censorship as a compliment, all I want is what I get only here at good old Cafe Hayek, the freedom to give my side of things.
I love the title of the article:
They're Not Necessarily Jobs, and They Weren't Necessarily Created or Saved, but We Gave These Guys Money, and They Gave Us Some Numbers
Earlier tonight I heard Jennifer Granholm giving her last (thank God) state of the state address on the radio. She started naming companies in Michigan that were creating or saving jobs due to stimulus funds or Michigan Economic Development grants. She named company after company after company with rising excitement like it was such a long long list and waited for the applause to drown her out for dramatic effect as she continued naming companies. I think she said there were 110 companies total.
At the time I was driving down 8 Mile Road (yes, the one in the movie) and I was passing hundreds of businesses on both sides of the street. The collision shops, the hair salons, the plumbing supplier, the kitchen remodeling shop. Yeah, I could go on and on too.
I thought, each of these small business people have had to cough up hard earned money to pay into Granholm's scheme to make herself look good. For what amounts to a few hundred or maybe a few thousand jobs in a state of 10 million people. Jobs for the “new Michigan” she said.
Makes me want to puke.
I want so badly to make some memorable comment on this, but I'm too ignorant of the nitty-gritty of the subject to notice much more than that DeLong's argument seems to amount to:
The stimulus worked because it's too small.
Russ, great post. BTW I read DeLong's post before yours, and I was already planning on making your point in a Mises Daily article. In case you see it, I don't want to think I stole it! I also am producing a Keynes/Hayek rap video, which I've been working on for years…
And people do both in the market.
In the public sector, you get dragged into other people's scheme whether you like it or not.
Conscription.
My point is to teach you something, not for you to try to find some way to twist or poke at my argument to make your own point or to take a cheap shot. My point is simple: arguing that because unemployment is 10% means that crowding is plausibly negligible is to misunderstand the nature of unemployment (heavily concentrated in certain industries and educational groups) and the nature of the stimulus (heavily concentrated in different industries and educational groups.). Nothing more. If you don't learn anything from my observation, either I'm a bad writer or you have a different agenda than I have.
“DeLong – often wrong, seldom in doubt.”
– Bertrand Russell
Excuse me? This isn't twisting or poking, or a cheap shot Russ. This is serious stuff. My points and questions are very reasonable, and honestly I'm not the only one out there raising them. I do get a lot out of your posts – that's why I read this blog. But don't accuse me of taking cheap shots when I don't find certain things you write very insightful or convincing.
Bad economics? You mean like the kind of stuff that went into all the bailouts (both Bush and Obama's.)
“Now I'm worried about arguing with an economics professor, and perhaps I have it wrong.”
Maybe you're joking and I've missed the point of that statement, but to my knowledge, you've never worried about this. Take your many posts on this blog as evidence.
I'm not worried about arguing with econ professors in general. I was a little worried because I'm arguing with Horowitz on a very specific point in this case.
He seems to be under the impression that the administration produces gross job estimates. I was always under the impression that these multipliers and job estimates were net job estimates, and that that was very common knowledge. I don't have a great deal of deference to professors simply because they are professors… but sure, I'll hedge when I challenge them on a very specific point like that.
He never cited why he thinks these are gross estimates, though, so I still think it's a good issue to raise with him. Particularly considering it's the entire point of the piece that he wrote.
You're not excused.
Pointing out that the people who benefit from medical spending aren't all PhD's and doctors is irrelevant to my point: citing a 10% unemployment rate as if all sectors can be aggregated into one sector is a mistake. Similarly, citing “stagnant wages” as if all wages are stagnant is to make the same mistake.
You said that “you're struggling to understand my point.” My point is that spending money on medical research doesn't help construction workers. What part of that don't you understand?
You need to check on your series of “as if's”, Russ. Obviously there are as many wage rates to consider as there are workers. And obviously some sectors are going to experience crowding out while others aren't. I've said as much and Brad DeLong said as much, but you keep repeating it as if the point is lost on us.
Since we agree that some sectors may be experiencing more crowding out than others, I tried to advance the discussion by asking you what conditions/data you would look for to assuage your concerns about crowding out. Whenever one makes a claim the first question they should ask themselves is “how would I falsify this claim?”. That's all I'm asking. If you don't want to answer that's fine, but don't keep repeating a point you and I already understand and agree on.
As for your point – again, I understand that point. Mine is simply that you can't evaluate what a stimulus package of hundreds of billions of dollars does for construction workers (who, keep in mind, are only 20% of the job losses anyway) on the basis of a fraction of a percent that is spent on medical research.
These are important questions about crowding out. You're always so quick to cite empirical studies coming up with a multiplier of 0 but then dismiss as ex post story telling empirical studies coming up with a higher multiplier. Well it's this crowding out question that is at the heart of these multiplier estimates. You started out talking about it. I don't think it's too crazy of me to want to press the issue of crowding out a little harder, rather than endlessly repeating a point about a couple million going to medical research that you and I are both on the same page on (the same page regarding the effects at least, if not the policy).
Let me state things just a bit differently.
The fear is that the downturn would go on indefinitely without the “gov” stepping in to halt it.
But how could it do so?
Only with money it had taken from that part of the private sector that remained healthy and strong. But why throw its good money after the bad, and sacrifice the stronger for the weaker? Why not just let the stronger displace the weaker? If the aim is to keep the weaker from dragging down the stronger, it makes no sense then to drag them down ourselves. Rather than keeping the system from collapsing, that just keeps the worst people in charge of it, on the backs of everyone else, and is not economics but The Divine Right of Kings.
Bob, I linked to one of your older posts re: Fed purchases of Treasuries artificially propping up demand in the comments section of DeLong's post, then followed up with some old-school Rothbard quotes and another Mises article on why low interest rates is not evidence of no crowding out.
He deleted them all and sent me a kind of creepy e-mail about how forcing him to “clean up” his comments wastes his time. Let me know if interested and I'll share the originals.
From the post in question he's deleted all the comments by someone called 'James W' – the only one who wasn't simply agreeing with Delong and was actually throwing up polite challenging questions. What a pathetic attitude for an economics professor to have.
And now he's deleted my comment asking why he deleted the others. John's is next, no doubt.
Brad: The crowding out effect is negligible at 10% unemployment and 0% interest rates.
Russ: What is the evidence for that view Brad, I'd sure like to see it?
Commenters: What evidence would you like to see to falsify your claims?
Russ: *Silence*
Brad: The crowding out effect is negligible at 10% unemployment and 0% interest rates.
Russ: What is the evidence for that view Brad, I'd sure like to see it?
Commenters: What evidence would you like to see to falsify your claims?
Russ: *Silence*
← Previous Comments
Comments on this entry are closed.