The size of fiscal stimulus vs. the length of fiscal stimulus

For this blog post, let’s assume Keynesian economics.

For all the talk of a “large stimulus,” you don’t hear much about a “longer stimulus.”

The problem with a “too small” stimulus is that you get an initial economic boost, but when the stimulus expires the economy slumps back down, as indeed happened in mid 2011.  Ideally a stimulus employs some idle labor, stops it from depreciating, and tides those workers over until they can look for other jobs in fundamentally better economic conditions.  Those last few words are important.  If conditions are not improving soon, the ability of the stimulus to “buy time” for those workers isn’t worth much.  The workers get laid off from the government projects and their reemployment prospects are no better than to begin with.  We end up having spent a lot of money to postpone our adjustment problems, rather than achieving takeoff.

Deleveraging recessions last a long time, as shown by Rogoff and Reinhart.  The need for continuing deleveraging implies that even a stimulus twice the size of ARRA won’t turn the tide.

In those cases a well-designed stimulus program should not be so “timely.”  For a given presented expected value sum spent on stimulus, it is better to spread it out across the years.  It is better to help a smaller set of workers for five years (or however many years it takes for most of the deleveraging to end), after which they are reemployable , than to temporarily boost a larger number of workers for two years, and then leave them back in the dust because deleveraging is still going on.

The effectiveness of a stimulus will be measured by how many workers it bridges over until most of the deleveraging is over.  For ARRA, that number is close to zero.

Length may be one reason why WWII was effective stimulus (again, we are operating within the Keynesian worldview here, no need to argue this point in the comments).  The war lasted a while, and in the meantime a lot of balance sheet repair went on.

Oddly, there is not much discussion about the length of fiscal stimulus.  But there should be.


Really interesting thinking, and I hope the comments don't get hijacked by a pro/anti Keynesian debate.

Two thoughts:

First, in political debate, one argument against Keynesian stimulus is that it takes so long to deliver the stimulus, it's no longer needed. That argument was used in 2009, I think.

Second, the government's estimates in 2009 regarding the length and depth of the decline were way too low.

These would seem to imply government spending through automatic stabilizers would be a preferred avenue to deliver stimulus.

Just to second what RZ0 said: I remember the length of the initial stimulus being considered a problem with it--people thought it would take too long to do anything. Does the public/politic simply not have the patience for what's necessary, and want an instant fix?

I think no one expects the instant fix, but the question should be: will it increase the chance of or speed up the recovery.

But isn't the stimulus money still making its way through the channesl? I mean there is a large lag between passing of the simulus bill and the implementation in actual jobs. Most of the construction projects should be under way now, meaning that the stimulus is still working. So isn't it also about the amount of spent money?

At the time nobody expected the crisis to last this long. We're 3 years into this. 3 years ago people would have said 3 years would be way too long. Now Tyler is saying it was too short. The problem seems an order of magnitude larger now than then. Perhaps 3 years from now is still too short. How do we guesstimate?

Dirk, If you go back to the MR archives for last Mar through Apr you will see people saying that there should be stimulus for the coming year, to provide insurance against a stall out or double dip. People excoriated Stiglitz and others (including Krugman) for advocating this. They predicted what we got without additional stimulus.

The answer for how long turns on what you view as the cause. I would argue consumers who increased personal indebtedness from 62% in 2001 to GDP to 112% in 2008 have rationally readjusted their spending habits, particularly when their house equity declined.

dirk, If you would like to see some graphic representations and analysis of magnitude and duration effects, go to the CBO study who link you can find at where there is a similar discussion.

Re: For all the talk of a “large stimulus,” you don’t hear much about a “longer stimulus.”

This post frames the issue as if there was no discussion of the length of stimulus last year, in April, when Obama and the democratic majorities decided that they would not seek additional stimulus because it was unpopular and there was opposition which would make re-election difficult, aka, Tea Party.

Any long term plan can be broken down into short term decisions. The short term decision, made for political reasons, was a long term decision--to shorten stimulus.

Yes, but... In fairness, Obama and the Democrat majorities never presented the initial stimulus as a short term decision for a problem needing a long term plan.

Scott, Do you think it would have mattered. Do you think that if they had requested an extension of stimulus, or a smaller amount later as insurance, they would have gotten it?

One could think of the one-time economic stimulus payments in 2008 as an attempt to shock the economy out of the recession, whereas the Making Work Pay tax credit in ARRA which boosted take-home pay continuously over two years was an attempt to support spending through a slow recovery. At least the administration made some public statements like that. More details on the comparison in the forthcoming AEJ paper:

a lot of ARRA was misdirected; 37% was tax cuts, including the sneaky $5 a paycheck less withholding....most of that went to walmart, & hence employed the chinese...

Tyler, today (August 27, 2011) the fiscal dilemma is whether to focus policy on "consolidation" or on "stimulus". "Consolidation" means an agreement on a strategy that may lead to a budget surplus to reduce the debt. "Stimulus" means a larger deficit (in relation to the latest CBO update) for as long as necessary to accelerate the recovery and reduce unemployment. I believe that "consolidation" is a requisite to accelerate the recovery and that "stimulus" a la Obama is a joke because it will always be aimed at redistributing income (from future taxpayers to Obama's constituencies for his re-election). The problem with "consolidation" is to agree on a strategy that can be effective and credible, and I don't think this sort of agreement can be achieved at least until the November 2012. Therefore, in the meantime rather than discussing terrible "stimulus" ideas (this weekend Keynesians may entertain themselves discussing the economic consequences of Irene), I suggest that economists focus on the credibility and effectiveness of alternative "consolidation" strategies.

Last time we "consolidated" was during a period of growth, creating a surplus, only to see a 10% tax cut when Bush came into office (in fact, it was part of his campaign platform to lower taxes).

The credibility problems lies with the conservatives who see every surplus as an opportunity.

But that is really not true Bill. The 'consolidation' during Clinton was also driven by Republicans (the welfare reform law was rejected twice by the great fornicator and he only approved it by the 3rd time because it was evident that Republicans would not give up). The Bush debt was driven by the recession of 2001 and as far as I know that is what Democrats believe it should be done, right?

We got to stop playing little political games here.

FYI, I'll gladly let the reader decide on who is playing political games here with your history and description versus their own current memory. Seem to have forgotten what happened after 2001 (tax cut, wars) etc..

I am not sure what needs to be decided here. What I mentioned were facts - the Clinton reforms only started and were in many cases initiated by a Republican congress and Bush's spending was justified in its majority as part of stimulus efforts. You can make a moral call on the way Bush spent the money but to deny that it was done as stimulus seems counter-factual.

FYI: It's fair to point out that the mid-90's Republicans (and George H.W. Bush) played a part in balancing the budget.

Bush II, on the other hand, aggressively reversed those gains and left us with a structurally deficient budget, where revenues fall short of expenses even at the peak of the business cycle. I do recall that his tax cuts were sometimes promotes as "stimulus," but long-term regressive tax cuts are almost entirely irrelevant to a demand shortfall. No sane Keynesian would have sold them as such. I certainly don't recall the war spending being described as stimulus. I do recall a specific stimulus-oriented bill in 2001 (?), and there were certainly other policies, but the wars and the tax cuts should account for the majority of Bush's impact on the budget.

Thanks for confirming that (1) ALL politicians are fraudulent clowns and (2) Daniel Kahneman is right when he says that not only are we sometimes “blind to the obvious,” but also we are “blind to our blindness.” (according to S. Levitt this fundamental insight is explained in DK's forthcoming book "Thinking, Fast and Slow").

"The effectiveness of a stimulus will be measured by how many workers it bridges over until most of the deleveraging is over. "

Certainly it would be best if unemployed workers got private sector jobs after the stimulus expired, but doesn't the stimulus itself create value when resources would otherwise be idle?

Case I: A million workers build roads, schools, bridges, skills and intangible capital for one year.
Case II: A million workers sit idle for one year.

After one year, which economy has more real wealth?

You are assuming that those new capital products are worth more than the capital that was used to employ them. The more political the process (i.e., Congress) the more the projects will be chosen based on paying off stakeholders rather than creating wealth.

That's an assertion that assumes what it attempts to prove. All you're saying is that anything the government does is "paying off stakeholders". Good rhetoric. Probably picked it up on Fox News. So, I'm waiting for your prooof from the records of the ARRA that the spending was spent the way you said it was. These are open records. You should be able to offer proof for your assertions.

Partisan much?

Anon, Burden of proof as to the mispending of ARRA falls on those who make the assertion. Otherwise, to make the assertion without support is partisan and just a continuation of Fox News.

From my Twitter feed, a finding from Feyrer and Sacerdote: More stimulus went to states with senior House members.
"Public choice message of Feyrer/Sacerdote study : If your state's House members are senior, you should support stimulus."
I think that GOP tax cuts are more likely to go to GOP supporters, as well.

When you say "are worth more then the capital that was used to employ them", do you mean that some other capital-intensive project was crowded out by the people and equipment used to build the roads, schools, bridges, skills, etc? If the economy were at full employment, I would understand your comment, but with idle resources, it's not clear what you mean.

Focus on the real economy, not the financial accounting, and help me understand what wealth-creation is foregone.

I may have been thinking too much towards the bookkeeping side. But a lot of infrastructure projects require much raw material, which is effectively destroyed and replaced with the infrastructure. Unused labor is wasted, but unused capital generally retains its value.

As a practical matter, stopping the destructive effect of long-term unemployment on the labor pool is probably much more important than "roads, schools, bridges," which are easily overbuilt. I realize this means that I'm arguing for hiring people to bury money in holes and hiring other people to dig it up.

Also, as opposed to Tyler, I think that a big short-term stimulus would have worked better, taking all the spending that was stretched over ~3 years and cramming as much of it as possible into the first year -- even on works of a dubious wealth-creation aspect.

@Dan, wrt "overbuilding roads, schools, and bridges", is there really much danger of that in the present state of US infrastructure? I wouldn't have thought so.

In my area, just 5 years ago they were building schools as fast as possible, and now they're fighting about which ones to close.

I'm not sure it's possible to get a 1) wealth-creating 2) stimulus 3) without significant bad side effects on the economy. You maybe can get 2 of 3.

There are a bunch of competing requirements in designing stimulus. If the government just wants to keep people employed to maintain their skills, and as many as possible, it hires 10,000 workers at minimum wage to build a road. Waitaminute, labor says, you've now depressed wages for all road construction workers -- the labor pool was expecting to build that road at normal wages.

I suspect you want projects that would never be done anyway, which means their wealth-creation properties are dubious. Not impossible, though.

Right on. Particularly if you what you create improves the efficiency or productivity of the private sector (including human capital) in the later period.

Professor Brynjolfsson,
You're assuming your conclusions. Anytime that you assume (1) a zero opportunity cost for a person's time, and (2) an activity in which that person's time can produce a net benefit, it's not a question of logic (rational choice) but of assuming that a can opener is available.

Depends what happens afterwards.

Actually, we know what happens afterwards. The infrastructure spending creates jobs, all kinds of business activity. Suppliers of materials, manufacturers gear up to make the stuff needed. It gets built, The needed bureaucratic infrastructure gets set up to maintain and manage the new stuff.

Then it ends, and all you have is a higher cost structure. The businesses were not responding to competitive pressure and innovating, but simply making stuff as fast as it could be made to fill a politician's order. Contractors learn the skill of fulfilling government contract demands, and find it useless or harmful in other markets. The fiscal pressures caused by the high borrowing required start having an effect, political responses as well as the market distortions caused by the monetary authorities attempting to maintain low interest rates to prevent runaway financing costs from eating up any meager gains in revenues to government.

As time goes on, the schools are half empty and closed, the bridges and other infrastructure in some cases used, others not and everything is poorly maintained.

Stimulus money allows politicians, businesses and the society in general to put off the hard decisions on the issues that created the problems in the first place. The real questions are more fundamental; can you have an EPA and a social safety net? Can you have a world dominating military and a social safety net? Can you structure an economy with support and cost structures for the highly productive 10% and offshore everything else and hope to get enough cash to support by entitlements the other 90%?

But even from a Keynesian perspective, was WW II actually effective stimulus in any way that improved the well-being of Americans? Rationing across many goods and services, half a million dead... Yes, GDP increased and unemployment decreased, but it's pretty easy to argue living standards decreased markedly.

But I do think you make an excellent point in terms of explaining why federal spending has reached 40% of GDP -- times are never good enough to end the "temporary" stimulus, especially given TGS.

FYI US government at all levels approaches 40% of GDP, but federal spending alone is closer to 25%.

I think you mean total spending at all government levels, not just federal.

Yes, thank you, that is correct.

The last sentence of Tyler's post: >>Oddly, there is not much discussion about the length of fiscal stimulus. But there should be.<<

This is indeed odd. At the end of Roosevelt's first deficit-financed stimulus, there was a large "retrenchment" to reduce the deficit. We know that was premature, and the result is also well known.

How quickly we forget!

Inflation crept up to 8.3% in 1937 while unemployment was at 14%. The "retrenchment" was done to stop stagflation before it got out of control. We do forget quickly. Of course, let's not forget that from '36, and to some extent as early as '33, Roosevelt essentially waged a regulatory war on capital, resulting in minimal private investment. By '41, most of the New Deal policy advisers were replaced and that assault lost steam. Real recovery (you can't look at unemployment figures when 14 million are drafted to fight abroad and others join draft-exempt industries to avoid the infantry) didn't start until tariff reductions at Bretton Woods (a trade deal with Great Britain in '38 didn't hurt) and Roosevelt's death (Truman was perceived, rightly, as much friendlier to business).

If by stimulus you mean Friedman's "cash dropping helicopters" or Keyne's "money mines" then longer term affects might be minimal but I would think stimulus spent on public works projects and/or education would add pernanent productive capacity to the economy that could help support demand added by the temporary stimulus.


IIRC, by the end of WWII, private debt was down to 65% of GDP. Since government debt started at very low levels, it was possible to lower the very high levels of private debt while increasing public debt. I am not sure it is possible to carry on a stimulus long enough to get private debt down to acceptable levels given the high levels of government debt with which we started. IOW, part-time Keynes probably does not work.


"...Since government debt started at very low levels..."

Where are you getting your numbers. My references (below) show USG debt at over 117% of GDP in 1945.


Jeez, learn to read.

Don't the last 10 years of ever-growing, monstrous budget deficits effectively represent a long "stimulus"? What has that brought us?

According to Keynesianism (which is assumed for this post), you don't want to stimulate all the time. Only during the downturns.


The largest factors in producing the "monstrous budget deficits" year-over-year are: the Bush tax cuts ($2.5 trillion cumulative lost revenues as of 2010), the Iraq & Afghanistan wars ($1.2 trillion), the DoD in general (roughly 20% of our budget yearly; $689 billion in FY 2010 alone), and TARP ($700 billion). Aside from the tax cuts to lower income brackets in the Bush tax cuts, none of these policies are stimulating to an economy (from either a Keynesian or Austrian perspective).

A large fraction of TARP has been paid back.

Tax cuts and ANY spending are stimulating according to Keynesians.

There's a strawman if I ever saw one. Keynesians don't believe all government spending is created equal, even within spending categories. $25 million in infrastructure improvements in the DC metro area is far more stimulating than the $25 million used to build the Bridge to Nowhere.

And that is called moving the goal posts. I leave it to you to figure out why.

Yancey Ward is correct. All of the things Linguist has mentioned involved increases in C and G. In basically any Keynesian framework, this is going to be stimulative. You'd have to assume a multiplier of 1 or less in order to claim that "none of these policies are stimulating to an economy", which is essentially a rejection of Keynesianism altogether.

Bottles full of money? Enough said.

Yeah but Friedman's Permanent Income Hypothesis says tax cuts are less stimulating.


jdavis, Keynsians don't say tax cuts are equivalent to spending.

What is MPC and MPS.

Did I write "equivalent to spending"? I wrote that Keynesians believe that tax cuts are stimulating (they DO believe this) and that ANY spending is stimulating (they DO believe this). I did not comment on any relative merits of the two strategies, or whether a Bridge to Nowhere is more or less stimulative than repairing roads. My statement was correct.

Actually, Jdavis, your statement is not correct regarding effects of tax reduction v direct spending. I'll find the studies but the tax stimulus is less than unity. I think it was around .80, less than unity, considering that people saved the money, particularly at the high income level. As a high income individual sated in all my appetites I saved my tax refund, and spent some of it on a Korean digital TV.

jdavis, Here is a CBO report on the effects of the ARRA and the range of estimates for the effect on the multiplier of tax cuts for the wealthy: the highest estimate is .6, that is, less than unity, meaning you actually don't get anything, other than repairing a wealthy persons balance sheet for a tax cut directed at that segment.

Jeez, Bill, you are completely incapable of reading anything, aren't you? I leave it to you to puzzle out your mistake here, since I have lost patience to try any longer.

Yancey, You seem to be saying to anyone you disagree with that they should read. See post above.
You may have an opinion, but I think the CBO report provides the facts, even though you may disagree with them.

Statistics are misleading things.

For '05-'07 (Bush tax cuts in effect) federal tax receipts were actually up compared to the higher rates. Receipts fell after the recession began, which is to be expected considering that the economy tanked. If the tax cuts produced deficits, they would've produced them immediately.

I don't know why it's so difficult to grasp that federal tax receipts are essentially uncorrelated to marginal rates. They're only affected by economic performance. Which is to say that over the past 50 years, a marginal tax of 90% on the highest earner inn a boom year has netted the same amount as a tax of 35% on that earner in a boom year. Why? The tax base is inversely correlated to tax rates.

Irene will cause the much hoped for (to the Keynsians) "Alien Invasion" stimulus?

I believe we have plenty of examples of "long term" stimulus. They're more colloquially known as "entitlements." How's that working out?

Well Social Security pays its own way but along with Medicare and unemployment benefits provides a lot of demand that wouldn't otherwise be there -.i.e. it gets spent. That may not spur much growth but at least it helps stop deflation from getting hold.

Now "entitlements" like oil and agricultural subsidies just end up in a bank somewhere with no businessmen wanting to borrow for expansion because they can't use the capacity they have now.


At what point do the policy initiatives become the driver of the economic conditions?

Wasn't the deleveraging over by '35 or '36? From then on it was the accumulation of very bad policies that were finally shed by necessity of the war effort?

Maybe, just maybe all the policy responses in 2008 and 2009 have resulted in exactly what we are seeing right now. The cost of preventing an even more severe downturn is a long period of stagnation. If Japan is any indication, there may be no coming out of it. They are long past the lost decade.

So it's size versus stamina. Well, if I'm guessing, I would think most economists would be forced to go with stamina.

The Keynesian argument is more than just bridging the gap between two good economic states separated by a slump caused by an exogenous shock. Their argument is that the bridge is a boost that makes the gap shorter.

The Keynesian argument isn't merely about the size of the gap but the timing as well. They may not explicitly state that, but it's implied. One way to look at it is that for a given output gap, X, a government expenditure aX where a is a constant between zero and one, will eliminate the gap. Or, temporaly, if the price/wage adjustment to reduce X to zero is T, then the government spending will eliminate the gap at some fraction of T. So spending in the Keynesian argument brings the opposite shore closer. The classical argument of crowding out suggests that spending makes the opposite shore move farther away - preventing readjustment as you say.

So in the Keynesian model, there's no distinction between size and duration of the output gap/stimulus. The failure of the stimulus suggests that the size of the output gap wasn't measured properly, and that the stimulus was either too small, too brief, or both.

I won't discuss the demerits of those ideas because they're obvious and OT.

"Stimulus" is how we got to 2007............clearly, the attempts to "stimulate" the economy since then have only trickled down to the most opportunistic of us. And they are keeping their ill gotten chits under the mattress.

A quibble: you wrote, "The workers get laid off from the government projects and their reemployment prospects are no better than to begin with." Not so. They have a work history (in the government job) rather than a history of unemployment (which is what they would have had absent the stimulus). Their skills have been honed on the job, rather than deteriorating in unemployment.

I agree with this complaint.

I agree that government employment can be better for the worker than unemployment. However, the alternative is not necessarily unemployment. The alternative can be taking the similar work at lower pay, different work, or a period of retraining or education.

Moreover, the government employment in similar skills only slows the adjustment toward developing new human capital. It rewards stagnation. If government embarks on large scale infrastructure projects, it will cause a temporary bubble in demand for those resources. Since the goods are durable, all he resources used to build them will be idled for a very long time. It's a hair of the dog economic policy.

Shouldn't we ask why the private deleveraging takes so long as a prerequisite for designing a policy response? Seriously, one should be deeply worried looking at the proposals on the table and Japan's experience. Such busts were well known long before the Great Depression and the Great Recession. Why have our more modern downturns become so long-lived?

I am deeply worried about the Japan experience coming to America. The unemployment numbers Karl had on Modeled Behavior look similar to America's numbers. The youth unemployment is larger than the rest. I can't help but think that maybe Japan's problems are that the spending is to help the old and the taxes are from the young. As America ages and the boomers retire things will only get worse. Does this sound off base to you?

Prior downturns were induced by interest rate increases, and recovery was stimulated by interest rate cuts, spurring the housing sector.

In this case, housing is dead, consumers are leveraged up the ying yang, and we have a liquidity crisis.

subsitute liquidity trap for liquidity crisis.
Although one did precede the other, the former is our current problem.

I am not asking about generic 20th and 21st century downturns only here, I am asking about the comparison between 19th century banking panics and those of 1930 and forward. Why have those lengthened in the last century? You would think that with our supposed advancement in economic theory, we would have an answer, but I am not seeing this answer from anyone in the field.

The 19th century banking panics were largely regional in effect, impacting agricultural more often than not, but having less effect on manufacturing or manufacturing employment.
Other than the Great Depression, which also had a banking crisis, this current recession is the only one since the 30's that was kicked off with a liquidity crisis and excessive banking leverage.

I wouldn't presume theory advances anything, or if it did, you would also have to presume that people acted on the basis of theory. As we've seen before, if you oppose regulation which is premised on a theory, but do not enforce, theory means nothing.

Yancey, re financial panics, de Long has all of his lectures on the Economic History of the US at the Berkeley site of ITunes U. It is the full Berkeley course, and quite interesting.

Yes, but he doesn't explain why, today, they are longer, and getting longer. Let's take your point about the regionality- are the changes that made them systemic (or, at least, more national) a benefit, or not? Even regionally, they didn't last 1 or 2 decades (and I am assuming 2 decades is an actual limit today). I am having a hard time seeing that we have improved things by understanding them better. Looking at Japan, one might think their funk might well be permanent minus a complete upturning of the policies being followed, or a different diagnosis of what was fucked up in the first place. What I am seeing is a doubling down on the policies that have us in a hole for essentially a decade already, and not actually predicted to produce a real recovery in the next 5 years, and that is taking an optimistic view of what is being proposed. I, first, think we need to answer my original question- why do these things last so long, now? What have we done different? Until I get that answer, I am going to treat any prescription with great skepticism.

Yancey, the comparison to Japan is interesting and provides some answers for why financial crises which are not addressed by fiscal stimulus AND monetary expansion prolong a recession. You might want to look at Bernanke's 2002 Fed speach on the Japanese problem.

Since you do not have an answer, and will treat any proposal with skepticism, I do not expect to see an answer for any of these problems from you and will remind you of your words: "I am going to treat any prescription with great skepticism."

I think we need to stop saying that WW2 was a "spending program." It was a freaking World War. It killed 50 million people and left vast swaths of Earth in runs. And we won it.

The Krugtards of the world will tell you we could have spent those billions on Federal Rabbit Inspectors and the result would have been the same. It ain't so.

The above facts, by the way, are not an argument against Keynesian thinking. I'm just saying Keynesian thinking does not apply to the USA in the 1940s.

"For this blog post, let’s assume Keynesian economics."


Can we now stop assuming it in the real world?

Then again, stimulus can factually depress the rest of the economy, by stealing from productive endeavors the capital and manpower they would have had access to had the stimulus not taken it by force. It also has to be a credible stimulus, in that it is actually accomplishing something of value, and not just degrading the value of the money base, as happens to be the case with the recent stimulus. Almost all of what this stimulus did was to keep unproductive, and in fact in many cases productivity destroying government jobs active. All while saddling the private economy with the threat of future tax increases to pay for the stimulus, tax increases that the politicians aim at the people who might hire someone. Without the tax increases, the other two options the government has is a hard default on its obligations or inflating its way our of the debt problem it created. All three options destroy wealth and cause chaotic reactions in the economy, which makes it hard to determine how to invest and hard to determine if hiring someone will be cost effective or a burden to the fiscal shape of the company.

"stimulus can factually depress the rest of the economy, by stealing from productive endeavors the capital and manpower they would have had access to had the stimulus not taken it by force."

I hear this argument but I don't get it. Right now there is a huge labor surplus, unless we take the Kling view that the surplus exists in fields in which it's not needed and doesn't exist in fields where it is. But even with the Kling view, the argument would have to be that government work is stealing labor from those jobs the private sector already wants, i.e., labor which is already employed. Do infrastructure projects steal labor which is already employed to a material degree? Well, OK, maybe it does. I have no idea. It would be interesting to see the data.


As far as I can tell, most companies that got stimulus dollars weren't idle prior to the stimulus. Most likely shelved other projects for the duration (including other government projects). To be perfectly "stimulating" in regards to employing the idle and idle physical resources, it seems to me you would have actively target exactly those industries that had seen mass layoffs and/or bankruptcies. You would have had to have the government buy cars (they did do that with Cash for Clunkers and the auto bailouts to some extent), and have the government building houses employing construction workers and supplying demand for housing materials (not done so much). My guess is that other than fixed roads and bridges, most of the stimulus stimulated nothing additional.

Yancey Re: your points

1.", most companies that got stimulus dollars weren’t idle prior to the stimulus. Most likely shelved other projects for the duration" That's interesting. How did you determine that?
2. "it seems to me you would have actively target exactly those industries that had seen mass layoffs and/or bankruptcies." Like autos, is that what you mean? Actually, you wouldn't, you'd rely on the multiplier effect to reach business indirectly.
3. "My guess is that other than fixed roads and bridges, most of the stimulus stimulated nothing additional."

Here is what the CBO estimated the effect of the ARRA was for just one quarter in 2010:

CBO estimates that ARRA’s
policies had the following effects in the second quarter
of calendar year 2010:
B They raised real (inflation-adjusted) gross domestic
product (GDP) by between 1.7 percent and
4.5 percent,
B Lowered the unemployment rate by between
0.7 percentage points and 1.8 percentage points,
B Increased the number of people employed by between
1.4 million and 3.3 million, and
B Increased the number of full-time-equivalent jobs by
2.0 million to 4.8 million compared with what would
have occurred otherwise (see Table 1). (Increases in
FTE jobs include shifts from part-time to full-time
work or overtime and are thus generally larger thanincreases in the number of employed workers).
The effects of ARRA on output are expected to gradually
diminish during the second half of 2010 and beyond.
The effects of ARRA on employment and unemployment
are expected to lag slightly behind the effects on output;
they are expected to wane gradually in 2011 and beyond."

Here is the link:

Why do I trust an "estimate", Bill? The model they use is going to show exactly that result with increase spending. It tells me nothing about what actually happened. However, let's go back to my original point- the bulk of the stimulus went to tax cuts (which you claim has no stimulus in comments above), the sustaining of state government spending, which, by definition, is maintenance of spending, not stimulation in and of itself, and additional infrastructure work that is performed by mainly road/bridge/public utility workers, both state employed and private contractors, who had not been laid off prior to January 2009. What I don't see is a lot of stimulus that employed those who had actually been laid off, and the support for that comes from the BLS numbers on employment, both short-term and long-term. Any one can claim a counterfactual for what would have happened without ARRA, and I can cite just as many estimates that counter the CBO's claims about what would have happened without ARRA.

I realize you don't understand the weakness in citing the CBO estimate since you repeatedly return to it despite all the objections others raise to it that you have never once addressed.

Yancey, Tax cuts to persons with high MPC are stimulative (see CBO report) but those with low MPC do not (see CBO report also); regarding states, if the stimulus money wasn't there, there would be no spending by the states, leading to an acceleration of loss but for the spending. You can image people are critical of the non-partisan CBO, but I am not, nor should you be on a good day.

The CBO is tethered to using static analysis approaches. They use specific equations to determine in advance what should happen, and then when all is said and done they use those exact same equations and tell you what "did" happen based on the amount of money spent. It would not matter if the money was spent digging holes with first shift workers and filling them in with second shift workers and inspecting the filled in holes on the third shift, or the money was spent building roads, bridges, hospitals and schools. That is just how they work. If the congress asked them to determine how much money they could get in taxes over the next 10 years if the tax rate were raised to 99%, it would give you an extremely high number, and no one should believe that number. Now tax receipts is something that after the fact they can give you an actual number on. Stimulus effects, not even remotely. There is no way to determine what the effects are, no one has that level of intimacy with the 300,000,000 person economy to figure it out.

Astoni, If you don't use facts or economics to support any position, you are untethered to support any position you choose, and, maybe you have to do that to support your argument. It's funny though that banks, businesses and others use these same statististics and methodologies to make their decisions as well and that CBO doesn't change them for you or for me.

Wealth is the accumulated power of the owner to maintain personal well being and accomplish change. A house is a form of wealth, as is the simple land a person may own with no structures. A pile of gold is wealth as is a computer or a CNC machine. The world only has a certain amount of wealth at any given time. A good education and practical experience in how to create a good is wealth. If one group has the power to take this wealth and use it for their purposes, it follows that there is less wealth left for others to utilize.

For example, when the government creates a huge project, such as a military satellite constellation. It can take upwards of 30,000 different engineers of many specialties to get that program completed. While this project is going, those engineers are not available to any other projects in the private sector. The government use of each is limited and they will pay top dollar for the right engineer to sit on his hands at a computer desk for two years before his talents are really needed. That is two years of destroyed productivity, even though the engineer is employed (even by private sector company), and is probably doing something, it is not anywhere near his potential. I know that the near 30,000 number is correct, because that is the number of engineers that was laid off in Los Angeles and surrounding communities when the government cancelled a single contract with a single company.

The main idea behind stimulus is that it helps recovery to come sooner, but it could help reduce private debt as well. Who us arguing that the main point of stimulus is to temporarily employ workers while the economy recovers at the same speed?

To hear a certain well-known Keynesian tell it, there's no good reason to debate whether a stimulus should put all of our unemployed workers for a short amount of time, or a portion of our unemployed workers for a longer period. Rather, we should implement enough stimulus to employ ALL of our unemployed workers, for as long as it takes. Because, you know, deficits don't matter right now...

You raise a good point.

Keynesians believe you shouldn't be stimulating an economy during a boom. Government should be paying down debt.

But given that we didn't, is debt funded stimulus the correct prescription in a Keynesian world? They seem to be saying it is so, but they don't seem to be checking their models to be sure. Will we enter a debt spiral or must we tough it out with austerity.

Governments around the world have already cast their votes.

- "For all the talk of a “large stimulus,” you don’t hear much about a “longer stimulus.”"
- "We end up having spent a lot of money to postpone our adjustment problems, rather than achieving takeoff."
- "The effectiveness of a stimulus will be measured by how many workers it bridges over until most of the deleveraging is over. For ARRA, that number is close to zero."

On all 3 points: really?
Peter Diamond weeks ago:
"Infrastructure spending is not a vehicle for dealing with a normal recession. But once you recognize this recovery will be slow, you realize this is a time when we should be doing major spending on infrastructure. And a lot of the infrastructure investments are stuff we’re going to have to deal with eventually, so doing it now doesn’t actually add to the trend debt problem, and doing it now means we’re doing it with otherwise unused resources, both in terms of labor and capital, so that makes it cheaper for the economy."

and he doesn't mention that present long-term rates are $2T.

*and he doesn’t mention that present long-term rates are less than 0. engineers' estimated missing infrastructure spending to maintain what's already built is now over $2T.

I'm economically illiterate, so this is a genuine question: Is the whole idea of Keynesianism merely to keep skills from deteriorating until the economy is back to full employment? I thought at least part of the idea was to get the economy back to normal quicker than it would otherwise. If so, it would seem that the greater the size of the stimulus the less the length would need to be.

'Is the whole idea of Keynesianism merely to keep skills from deteriorating until the economy is back to full employment?'

I think the way you ask the question is flawed. Keynes did not believe an economy would get back to full employment 'of its own accord' and that stimulus spending was just a way to tide things over - he believed the stimulus spending was necessary in order to get the economy back to full employment.

Of course, stopping skills from deteriorating is one more argument for it.

While I have disagreements with Keynsians over the nature of countercyclical spending and money multipliers and how every example can be constructed as either for or against stimulus; I have trouble with the framing of the narrative as either or. Either you are a Keynsian or an Austrian. One must be correct. No one is asking really interesting counterfactuals.

What if there really was a continent of Atlantis. And it suddenly raised to the surface this year. What effect would that have on the global economy and why?

What's the conventional economic wisdom on the Reagan years now? Was that Keynesian deficit spending or not? Did the entire deficit spending years of the 80's merely tide us over until the tech bubble years of the 90's? When the hell were things ever normal, as in no asset bubble, no world war, no military deficit spending, etc? I'm starting to take the Vonnegut view that normal is that things suck and we're sick and tired and unemployed and that's normal. Prosperity is a pyramid scheme, no matter how you get to it.

You aren't supposed to do stimulus spending when the economy is doing well.

I understand the theory is that you aren't "supposed to", but it isn't that in fact what we did from 1982 to 1996?

We increased the consumption portion of our economy (which is what governments are best at doing) at the expense of the investment portion of our economy. That continued pretty much on for the next 25 years, and now we're something like 70% consumption-based.

There is a rational argument to be made that the effects of a monetary tightening be alleviated to a certain extent by fiscal loosening. And visa versa. The Bank of Canada had a loose monetary policy during the time that the government was cutting spending substantially to get rid of the deficit. Google David Dodge Calgary where he explains what the central bank policy was at the time and why.

You mention a carry over of 5 years. Will the jobs available in 5 years be the same as those lost? The nature of available new jobs may be drastically different from the old ones. Retraining of displaced workers will be mandatory to give many of them an opportunity, but even that may not work well for some.

What I don't understand: In a de-leveraging or 'balance sheet' recession, spending drops not because of fear or a loss of Keyne's 'animal spirits', but because people are reacting rationally to loss of assets by reverting to savings to repair their balance sheets.

Now... If you engage in stimulus through government borrowing, all you're doing is preventing the process of balance sheet repair by replacing private borrowing with government borrowing. So while the public is repairing its own balance sheets, the government is making its own worse.

If government borrowing is occurring under a situation where the debt is already so high that people believe that it will have to start being paid back in within the period of their long-term planning horizon, then why wouldn't they see the government debt as part of their own balance sheet? If people are saving more because they're worried about their costs in the future, and the government borrows a bunch of money that the people feel means they're going to have their taxes raises to pay it back or lose like Medicare they were counting on in retirement, they'll just save even more.

This seems consistent with the data. One empirical study of the stimulus has found that states that received stimulus money just replaced their own spending with stimulus spending, which just shifting debt from the state to the feds, with no net increase in spending. Other studies have shown that the stimulus may have actually destroyed private sector jobs while creating public sector jobs. Looking at the money supply, as the various quantitative easings injected aggregate money, the M2 money multiplier fell in lockstep with the changes.

In short, I don't think you can fix a balance sheet recession by borrowing money and spending it when the people see the government's balance sheet as part of their problem.

Please point me to those studies.

Conley and Dupor:

From the Abstract:
"This paper uses variation across states to estimate the number of jobs created/saved as a result of the spending component of the American Recovery and Reinvestment Act (ARRA). The key sources of identification are ARRA highway funding and the intensity of state sales tax usage. Our benchmark point estimates suggest the Act created/saved 450 thousand government-sector jobs and destroyed/forestalled one million private sector jobs."

The paper goes on to describe "Appreciable estimation uncertainty" which may affect these numbers, but that's the best estimate.

Another empirical study that gives slightly better results:

From the abstract:
"We use state and county level variation to examine the impact of the American Recovery and Reinvestment Act on employment. A cross state analysis suggests that one additional job was created by each $170,000 in stimulus spending. Time series analysis at the state level suggests a smaller response with a per job cost of about $400,000. These results imply Keynesian multipliers between 0.5 and 1.0, somewhat lower than those assumed by the administration."

In this case, they claim that the low multipliers are the result of the diversion of so much stimulus money into teacher's salaries.

You might like to read this testimony from John Taylor, summarizing his own empirical research:

Lots of interesting charts in that one.

Here's the conclusion:
"In sum, the data presented here indicate that the American Recovery and Reinvestment Act was not effective in stimulating the economy. Despite its large size, ARRA did not result in more than an immaterial increase in government infrastructure and other purchases at the federal level. The large grants to the states did not result in an increase in government infrastructure and other purchases at the state and local level. And finally an analysis of the payments that temporarily increased disposable income shows that they did not significantly affect personal consumption expenditures. In contrast changes in private investment and net exports have been much more of a factor in the recovery. Currently, the increased debt caused by ARRA—both directly through its deficit financing and indirectly through its de-emphasis on controlling spending—is likely a drag on economic growth."

These are studies looking at actual empirical data, not just a re-running of the same models used to justify ARRA in the first place, which is what the CBO estimates really are.

I would also note it would not be unexepected that states WOULD replace stimulus money for their own money, because without the stimulus money they would be cutting their budget to match their state revenue. But, I am really, really interested in seeing those studies which show that stimulus money destroyed private sector jobs, because, my friend, we had high unemployment and a shortage of aggregate demand, and if you are saying that government spending competed away or crowded out private spending, and you believe it, I have a bridge to sell you. And, it goes to nowhere.

Dan, Here is a CBO report on the effects of the ARRA:

One last time, what the models show always depends on the assumptions used. The very fact that there are models that disagree with one another is enough to discount this as strong evidence of the efficacy of ARRA, especially given that many of the models used at the time of ARRA's enactment did not actually predict the outcome we measured, but then those propenents of those models then made the entirely predictable response of claiming that their initial assumptions about the state of the economy in 2009 were wrong, rather than maybe examining whether or not it was the other assumptions that should have been questioned. Why should someone accept that one set of assumptions was wrong while the others were correct in 2009?

Where does the government get money to pay for stimulus or for anything? It taxes or borrows. 60 cents of every dollar spent by Washington right now is taxed, the other 40 cents borrowed.

Let's say government spending inserts $1 trillion into the economy. By the current ratio, it also takes $600 billion out of the current economy and $400 billion through future economies (either through future taxes or declines in purchasing power). The jobs destroyed by those transfers aren't netted under "jobs destroyed" in CBO estimates.

By the CBO's rosiest estimates, the stimulus created "or saved" 3.3 million jobs other than what would've happened without stimulus. This is misleading; the CBO is comparing that to what would have happened with a stimulus but the SAME tax rates. In fact, we need to discount jobs created by jobs destroyed by the taxes that go toward the new policy. We also need to adjust for the fact that $282 billion of the stimulus was tax cuts (of a sort -a lot of these "credits" amount to changes in incentives of production towards whatever government sees fit).

Nevertheless, the inefficiency of the net result ($250,000 spent per job) can't be denied. Granted, 1/3 of that was tax cuts, but that still leaves roughly $160,000 per job created that will have to be paid for through taxes now, in the future, or inflation in the offing. The jobs that will be destroyed by that don't make it into CBO figures and the stimulus seems a net "plus" for the economy.

Judging from the comments of those who dispute economics because it involves theory and statistical methods, we are lost in the swamp. God help the USA.

I want to explain to you more clearly what the CBO report is.

When the ARRA was originally scored by the CBO, it used standard Keynesian models to estimate the impact. That was the source of subsequent claims of how many jobs would be created by ARRA.

As part of the deal, the CBO was asked to revisit the original assessment at various intervals and update it based on changes to legislation and goverment performance in getting the money. And that's what it has done.

In other words, these CBO reports are little more than re-runs of the same models used to justify ARRA in the first place, with changes due to subsequent changes in the legislation, inability to get the money out ('shovel ready' jobs, etc). The models are still based on the same Keynesian assumptions. There is very limited use of actual empirical data from the economy.

If the models themselves are wrong, the CBO report will be wildly innacurate. Since this entire debate is about determining whether the Keynesian models are accurate and the assumptions complete, citing CBO in defense is little more than circular reasoning: "Keynesian stimulus worked! To prove it, here's a report that shows how many jobs must have been created, assuming Keynesian stimulus worked!"

This is my own view of Keynes, so I accept that others see things differently. I think Keynes produced a Theory to explain how and why Govt Borrowing could make sense in certain situations. In the long run, he believed the Govt Debt should be paid down as the economy improved. In many ways, his ideas are similar to the reasoning behind the Chicago Plan of 1933. I don't think there is a clear or obvious a priori answer to how much or how long. Rather, we use Trial & Error and try to adapt as we go along. However, it does seem to me that we should try & get things going ASAP, and I think that would be Keynes view as well, if only based on politics. In the end, the "Stimulus" should last as long as it takes to work, however you define that. For me, Unemployment is unacceptably high, & so we should continue Monetary Actions & Govt Borrowing. & we shouldn't simply make this an Econ Issue. Politics is a large part of reason we find ourselves riding in this slow moving train. It is true that, as we go along, we are, in effect, trading one set of problems for another, but that's life.

Neither size nor length are as important as the skill with which the stimulus is applied. Otherwise you're just getting screwed.

I thought the problem with the high unemployment was depreciation of human capital, which is not a discontinuous function of "completely tiding people over." I don't see how human capital depreciation is different across these scenarios:

1) you keep 2 people both at work for 1 year and then both idle for 1 year,
2) you keep person A at work for 2 years and person B idle for 2 years.

If depreciation of human capital is the same under each, then the length of the stimulus doesn't matter.

You assume that skill degradation is linear. It might be a good first approximation, but with long-term unemployment you have to really examine that assumption closely. There's an old joke that the definition of a man-year is "what IBM says 500 people can do in a morning."

I think the UK has "practice jobs" (probably spelled Britishly) where you do the work like in your real job as part of their version of UI. Seems decent enough.

These sort of things just point out the insanity even more.

Was WWII an effective stimulus for France and England? The WWII interpretation has to feature the comparative advantage provided by the destruction of the productive capacity of the rest of the developed world.

If we have a debt fueled recession, doesn't stimulus create more debt, which must then be "deleveraged"?

Perhaps a more proper form of stimulus when there is a banking crisis is to let corporations borrow from the Fed. We pumped all kinds of money into the banks so they'd still be able to lend, but it seems like a lost cause when they still couldn't lend. We could have just bought corporate bonds...

Certainly, the recipients of such programs are better off than otherwise, and the benefit to society depends on the utility and durability of those goods they produce under it. But the whole premise rests on a weak assumption that there will be private demand for those skills upon completion of the program. For that to be true, policy makers must be omniscient, which is not a human quality.
The fact is the government has had stimulus in the form of deficit spending for almost all of the last 80 years. The end result is a transition from the most prosperous nation to the most indebted nation on the planet. This leads me to believe that the true justification for such a program is for a political return, and not a economic one.
Because the Davis Bacon law, these programs work as a money laundering system between government and unions, which is their true purpose.

That this question is being discussed shows what the author never admits---how little Economics has to be with science.

Because we do not have the tools to measure how much damage was truly done, we have no idea how to determine either the size of length of the stimulus, except trial and error.

Job guarantee programs seem like the best way to have a constant stimulus in a downturn.

'Deleveraging recessions last a long time, as shown by Rogoff and Reinhart.'

They needn't. We should simply write off large portions of the fraudulent debt that was issued.

Comments for this post are closed