August 2021

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31        
Blog powered by Typepad

« Some Quick Thoughts on the "Mancession" | Main | The March 2010 Freeman is Available »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Don't you find it problematic that his empirical research covers periods where there was no liquidity trap, where there was no deflationary depression, in other words, where we didn't have conditions where theory that stimulus would work. Indeed - when conditions were significantly different than what they are now?

I'm not surprised that Barro came up with the results he did. I am surprised that he thinks they're at all relevant.

I love how Austrians have enormous problems with macroeconomic modeling, positivism, aggregates, mutlipliers, and empirical studies...

... UNLESS of course those studies confirm Austrian priors.


First, talking about results is not an endorsement of the findings. The same criticisms of macromodeling apply, don't you think?

Second, the criticisms offered are immanent criticisms not transcendent. The actions are being judged against their own claims. If the claims made were correct, then you wouldn't see the results he found.

Third, if Barro is right, then wouldn't that give you pause to believe what you believed to be true in your first paragraph --- if we were in a liquidity trade and deflationary depression, then the stimulus wouldn't have proven to be ineffective, right. So if the crude Keynesianism that justified the stimulus was wrong last year, then how would YOU rethink your discussion of the issues? And, to compound that, if macroeconomic modeling is fundamentally flawed as a picture of the economy, then how would you provide a more microeconomic discussion of the world over the past year and a half and what might have been the appropriate policy choices?

Something to THINK about, not be snarky about.

If Barro is right about what? His empirical findings on the multiplier or his application of that empirically derived mutliplier out of sample to economic conditions that didn't hold true during the period of his empirical work?

I have no doubt Barro is right about his empirical conclusions. What I'm challenging him on his the application to the current situation.

RE: "if we were in a liquidity trade and deflationary depression, then the stimulus wouldn't have proven to be ineffective, right"

But my whole point is it hasn't been proven to be ineffective. Barro is declaring it ineffective by applying a multiplier that is relevant under other conditions and not relevant under current conditions.

RE: "So if the crude Keynesianism that justified the stimulus was wrong last year, then how would YOU rethink your discussion of the issues?"

I'm not sure what you mean. So called "crude Keynesianism" has justified fiscal intervention in the past. I wouldn't call the justification of this stimulus "crude Keynesianism".

RE: "And, to compound that, if macroeconomic modeling is fundamentally flawed as a picture of the economy, then how would you provide a more microeconomic discussion of the world over the past year and a half and what might have been the appropriate policy choices?"

I'm not sure you can say macroeconomic modeling is fundamentally flawed. Certainly there are macroeconomic models that are flawed and need to be reworked. But that doesn't mean macroeconomic modeling is flawed. Just as it's possible to commit aggregation fallacies, it's possible to commit atomization fallacies.

RE: "Something to THINK about, not be snarky about"

Please don't mistake me raising questions about the post for snark. These are important questions for me, not a game. You've pointed out ironies in what you consider to be very problematic perspectives in your posts in the past. That's all I'm doing. No disrespect intended.

RE: "First, talking about results is not an endorsement of the findings."

My mistake on this one. When you wrote this "crowding out combined with inefficiency in operation, not very promising to help us reverse course on the economy", I took it as an endorsement of Barro's findings that there was crowding out.


I don't think we've ever had the conditions for a fiscal stimulus to have a chance of doing any good. Is there a hypothetical scenario in which a fiscal stimulus could have a net benefit? I think so, but we've never been close.

Lee Kelly -
Sure. Now and in the Great Depression are the most obvious candidates. And while I wouldn't call this "fiscal stimulus" in the Keynesian sense, planned pro-cyclicality in the government's budget can help stabilize the business cycle.

Assuming you've accepted that there are conditions under which fiscal stimulus has "a chance of doing any good", what reasons do you have for thinking that such conditions didn't hold in the Depression (or now, for that matter)?


I apologise--my comment was unclear. I was talking specifically about recent years. With regard to the Great Depression, perhaps a fiscal stimulus would have a better chance of doing some good, but given the incompetence of the government at the time, I would probably still opposed it. In any case, the problem was primarily a monetary problem, and it is with monetary policy that it should have been addressed, both before and after the crash.

Barro's piece is, I believe, the third in a series in the Journal. The underlying research covers a much longer time span and, as I recall, no matter how the data are parsed, the expenditure multiplier never rises to 1.

Jerry -
The data covers a wide time-span, but very little if any (does he go before WWII at all?) of that time span covers Keynesian depressionary conditions.

Of course he's going to get low multipliers looking at that period that he does. Even if he does include a period where the conditions are right, most of the identification in the model is going to come from other periods: Korea, Vietnam, etc. - when conditions weren't right, therefore biasing the estimate downwards.

I have no problem accepting the idea of a multiplier of one or less for the period he covers. The problem is that doesn't really tell us much now. As Lee Kelly points out, we're stuck with very little data to draw on. The periods we can look at had wimpy little stimuluses paired with some very destructive policies, and the fact that there are so few of them makes the identification problem very serious. We're largely dependent on theory on this one, in other words.

"Don't you find it problematic that his empirical research covers periods where there was no liquidity trap..."

Your point is well taken regarding the circumstances surrounding his estimates, however, I need to be convinced that liquidity traps exist. A zero lower bound on interest rates does not define a liquidity trap. As Brunner and Meltzer showed in their 1968 paper, a liquidity trap requires that the marginal rate of substitution between money and all other assets is equal to zero.

The simplicity of analysis regarding monetary policy during the recession has been completely abhorrent. Those who believe that we are in a liquidity trap point to two factors: (1) the zero lower bound, and (2) the substantial increase in excess reserves. Now there is a monetarist counterargument to both points regarding substitution effects, unconventional monetary policy, etc. However, what often gets ignored are the conditions under which the excess reserves are being accumulated.

Are excess reserves being accumulated because money is a perfect substitute for Treasury bills? Or could it have something to do with the fact that these injections are, by definition, temporary (at least for the given institution)? In other words, the Fed is using repurchase agreements under which they receive mortgage-backed securities. The agreements are, by definition, temporary. As such, doesn't the bank have an incentive to hold the new reserves on their balance sheet and receive a risk-free return from the Fed until the agreement expires?

We can debate all day long what the Fed should/can do and the effectiveness of stimulus in light of their (in)actions, but so long as the Fed continues to conduct fiscal policy by attempting to control mortgage rates through these purchases, we are unlikely to get a true picture of the current state of monetary equilibrium and the stance of monetary policy. As such, I don't think that starting from the premise of a liquidity trap is necessary for analysis.

As I understand it, Barro's data go back to 1913 and include the Great Depression. I have not read it, but the relevant paper appears to be NBER WP#15369.

I'll stay out of the liquidty trap discussion.

As was noted by many commentators last year when Barro first pushed his argument forward, his results on the defense spending multiplier are heavily driven by WW II (which is when we had the biggest variability of defense spending, so disproporitionately drives the coefficient, despite there being many other years in the sample). This was probably the period more than any other when there was no slack at all in the economy and crowding out occurred not through the usual mechanism of the financial market but by the forcible reallocation of resources through command planning and rationing. Given the absolute lack of slack, it is not surprising that he came up with the very low multipliers he did. Needless to say, the conditions last year and this do not remotely approach those of WW II.

Now it is very difficult to do counterfactuals on this. Given that employment has yet to start rising (although GDP is), it is easy to snipe that the stimulus "did not work," despite various studies claiming that we might have had unemployment greater by two million than we do if there were no stimulus. After all, those studies are based on "crude Keynesian" models that assume larger multipliers than the pretty unrealistic ones estimated by Barro, and so on. Maybe, but this is hard to prove.

I will note two points, one anecdotal, one substantive, although there is no definitive way to resolve this. The anecdotal one is that I do know of jobs created by the stimulus, at least one anyway. That one we just filled by hiring in our department, although that was to replace a retiring faculty memmber. This was stimulus money to support state budgets (about to disappear), but it did lead to one person getting hired whereas that position would not have been replaced otherwise.

Now, it could be argued that because of the inefficiency of governments and so on, that some other jobs would have been created if this one was not, presumably in the more efficient private sector. It is not clear to me where those jobs would have been or what the mechanism of their creation would have been or how the spending of the money to create this one led to those jobs not being created. Jerry thought I was joking previously when I mentioned sectors in North Dakota and Wyoming, but I was not. They are where the unemployment rates are the lowest. So, I put it to anybody here: how would not hiring that person in our department have led to the hiring instead of an oil rig worker in Wyoming and a hamburger flipper in Fargo, ND, just to assume a greater increase in hiring without the stimulus?

The more substantive point is that the usual mechanism for crowding out is through the financial markets, although that is not how it worked in WW II. Two days ago Menzie Chinn on econbrowser looked pretty closely at the most relevant financial markets. No evidence of crowding out, indeed, if anything real interest rates are lower than in past recessions despite the very large deficits being run now. This does not offset the "government always less efficient than the private sector" argument that always does well on this blog, but it does show that probably that is the only argument that can be brought to bear, and also provides more evidence that Barro's model and argument are as silly and irrelevant as they were last year. The failure to hire that oil rig worker and hamburger flipper did not occur because of overly high interest rates driven by the deficit spending.

Barro's estimates are strongly driven by WW II when there was severe crowding out due to command planning and rationing. That is when we had the greatest variability of defense spending. We are not in WW II conditions now.

The crowding out argument seems out of place. I write that not because government borrowing and spending is not crowding out private investment and spending, but because that effect is surely going to be much less during recessions than at other times, right? Or am I missing something?

I believe there are plenty of good reasons to oppose the fiscal stimulus, but the crowding out effect would seem to be one of the least important.


Well, that is Barro's argument, astoundingly enough. I might add that the logic of this would suggest that the crowded out otherwise new jobs would not be in states with the lowest unemployment rates, such as ND and WY, but the highest, and maybe in a related sector. So, perhaps if we did not have that stimulus money to hire a tenure track assistant professor, a two person economic consulting firm would have been started in Michigan, although again, it is very unclear to me just how this would work without some accompanying drag due to the stimulus through the financial markets or some other accompanying action such as increased regs on starting up new firms, which most definitely does not seem to be what has been going on, quite the opposite with tax cuts for new businesses, blah blah blah.


All the points you bring up are good ones. It seems, however, that you aren't focusing too closely on the question of which time period we're talking about. If we're talking about one fiscal year, then your arguments are probably appropriate, but if we're talking about one, two or three years from now, then all of the resources used to fund that tenure-track job (this year and every year after this) will not be available at those later dates. Seems to me that this is the force of Bastiat's argument about the unseen, which has very little directly to do with "slack" in the economy.

Barkley writes:

So, I put it to anybody here: how would not hiring that person in our department have led to the hiring instead of an oil rig worker in Wyoming and a hamburger flipper in Fargo, ND, just to assume a greater increase in hiring without the stimulus?

Didn't unemployment have a recent tick down to 9.7% from 10.0?
Here are some data points from the real world. Our firm, not an oil driller or burger restaurant, grew its revenues 80% yoy, recently hired a bunch of new people, and opened an office in Hong Kong. Does this count? I also received a recent inquiry from another firm about joining it. They are going to hire someone, even if not me. I have also read about other firms hiring new people, and heard about some new hiring activity from people I know at other firms. This is just casual stuff, but once we look beyond oil rigs and fast food places, we see that it has at least as much validity.
What about Keynes' imaginary ditch digger, the guy who would have dug holes and filled them up? Would he have been employed? In my view he would have been anti-employed. His job would not have been created in response to real consumer demand. Ditto for the new hire in Barkley's department. If his "job" was created with taxpayers' money, it's continuation presumably depends on a constant stream of taypayers' money. That's money that might have been spent by a taxpayer on a consumer's good, or saved and possibly invested in creating a job that would have met real consumer demand.
And in any event, it's stolen money.

Before we go down the path of asserting a liquidity trap lets see what Keynes himself said:

"But whilst this limiting case might become practically important in future, I know of no example of it hitherto.” John Maynard Keynes, The General Theory of Employment, Interest, and Money 1936

Dr. O'Driscoll and Prof. Rizzo have been excellent in reminding us what Keynes actually said.


But, of course the hiring by your company occurred in conjunction with the stimulus, including a really big one in China that might have helped with the job in Hong Kong. So, it does not answer the counterfactual of what would have happened if there had been no stimulus and no hiring in our department.

I would like to think that our new hire will be doing something at least somewhat more useful than the infamous ditch digger who fills his holes back up again after digging them, but then, who knows? We do have overflowing classes, which is one of the reasons our department was one of the few that did get to hire this grim year. It would appear that there is a lot of demand for that person's services, even if the result will be an ultimate suckering of the customers.

What I am particularly missing in Barro's WSJ article (and its predecessor) is a STORY about why there is all this crowding out (if that is what it is) now. One of the benefits of a plausible story is that it acts as an additional check of the empirical data.

In cases like this I am always reminded of what George Stigler used to say: "I want to see the econometric results that were thrown in the waste basket."


I imagine that there has been at least one recent start up funded by venture capital (or maybe by the partners in the start up) that was unaffected by the stimulus.
I hope your new hire will teach students the truth about Keynesian hornswoggle.

Perhaps I am misinterpreting some of the comments, but bear in mind that the $900 billion figure assumes that we raise taxes next year to pay for the spending.

What the debate about fiscal stimulus has really taught me is that, in many cases, it is much more about ideology than economic theory.

J. Bradford DeLong's disagreements with Robert Barro are posted here:


The comments to this entry are closed.

Our Books