Measuring the Effects of the Stimulus

The folks at e21 point out a new study by Daniel J. Wilson of the San Francisco Fed on the effects of the stimulus :

Wilson's study makes an important contribution to this debate by focusing on state-by-state comparisons. A large portion of stimulus funding at the state level was based on criteria that were entirely independent of the economic situation that states faced. For example, the number of existing highway miles was used to calculate additional transportation spending.

The study uses this resulting variation in state-level stimulus funding to determine what impact ARRA funding had on employment — including both the direct impact of workers hired to complete planned projects, as well as any broader spillover effects resulting from greater government spending. Administration economists have repeatedly emphasized the importance of this indirect employment growth in driving economic recovery.

The results suggest that though the program did result in 2 million jobs "created or saved" by March 2010, net job creation was statistically indistinguishable from zero by August of this year. Taken at face value, this would suggest that the stimulus program (with an overall cost of $814 billion) worked only to generate temporary jobs at a cost of over $400,000 per worker. Even if the stimulus had in fact generated this level of employment as a durable outcome, it would still have been an extremely expensive way to generate employment.

Interestingly, federal assistance to state Medicaid programs appears to have decreased local and state government employment. One possibility is that requirements to maintain full Medicaid benefits in order to receive federal aid proved sufficiently expensive that state governments pushed though additional rounds of layoffs in non-health related areas. This finding may suggest a potential pitfall with the Wyden-Brown proposal to decentralize health reform efforts at the state level: if comprehensive insurance requirements are retained, the net effect of reform may only shift safety-net spending towards healthcare and away from other urgent priorities such as education or welfare assistance.

Backers of the stimulus have always had to contend with two big problems: The first is measurement. How do we know how many jobs were created, saved, funded, whatever? Do we count new permanent jobs, or partially funded contractors, or grocery clerks whose paychecks are dependent on added business from stimulus-funded workers across town? And even if you can verify that those jobs are funded by stimulus money somehow, how do you know that the same jobs would not have been created in the absence of the stimulus? It's a thankless task, and the administration has tended to respond by skirting the issue and relying on models that don't really measure output at all. Wilson's study, with its state-by-state comparisons, attempts to partly address this problem.

But his tentative conclusions lead to the second problem, which is value. Even if you find  that the stimulus did create jobs, then the question becomes: Were the results worth the price? The findings in Wilson's study suggest that they weren't.

More on measuring the stimulus here and here. 

Comments

[Legate Damar]

Backers of the stimulus have always had to contend with two big problems: to just repeat their tripe often enough to get people to believe them.

FYP.

[Jimbo]

Even they don't believe it.
When I hear how the "fatcat" bankers and Wall Street goons got bailed out at the expense of Main Street, I laugh. Main Street got a huge bailout: Low interest rates, GM (workers), Obamacare, etc.

If you think the stimulus was expensive, just wait until we get the price tag for the Main Street bailout.

[P Brooks]

And even if you can verify that those jobs are funded by stimulus money somehow, how do you know that the same jobs would not have been created in the absence of the stimulus?

I would contend that those jobs which would have been created anyway are more likely to provide a net long term benefit to the economy.

[x,y]

$400k for each job? That's a bargain for government spending these days.

[John Thacker]

Once more into the breach, eh, Peter? I agree with your articles, but it must be tiresome to address the same data fudging issues over and over again.

At least your wife gets to discuss different topics all the time.

[P Brooks]

And, furthermore, many if not most of the jobs "saved" by the stimulus (like teachers and gas pump inspectors) should have been allowed to disappear into the Schumpeterian maelstrom.

[alan]

+1

[Old Mexican]

Re: P Brooks,

"the Schumpeterian maelstrom"

Catchy! I love it!

[robc]

This paper estimates a bigger impact in the first year of the ARRA, but then a steep drop-off in its employment effects in the legislation's
second year, while the CBO and CEA estimate a continual, near-linear increase in the ARRA's employment impact over time.

So, stimulus money leads to expensive, short term jobs. And, once again, the CBO and CEA show that they have no idea how non-linear systems really work.

[T]

To be fair, rob, most people have a difficult time with non-linear systems. But then again, most people don't have jobs evaluating them.

[robc]

So, if Im to understand what you are saying, not everyone had to take multiple differential equations classes in college?

Huh. Weird.

[R C Dean]

The real real problem with estimating the impact of stimulus spending has to do with the time scale involved.

If you limit yourself to the months the stimulus was being spent, you will of course find jobs associated with the spending.

Once you start stretching out your timescale, though, there is no reason to expect those jobs to turn into permanent jobs once the money spigot turns off.

And, at some point you have to deal with the drag on the economy created by extracting the money necessary to fund those (temporary) jobs.

Its like a credit card. The month you run up your credit card bill with fancy meals and concert tickets, your standard of living is great. But it snaps back to its previous level the next month, and over the months and years you are paying back the credit card (with interest), your standard of living is less than it would have been without the binge.

[American Voter]

Your words are very complicated. Why do you want people lose their jobs?

[Democrat Politician]

No, no, American Voter, that won't do. The question is properly formulated thusly: "Why do you hate working families?"

[robc]

The only purpose to short term stimulus is to win elections.

As this study shows, and was obvious to most of us before, stimulus only helps in the very short term, by the middle or long term, its a negative. The economy wasnt going to collapse without the stimulus, so fuck the short term.

[robc]

And, of course, the Dems screwed even that up. Does explain why they dragged out stimulus spending though. But clearly, not enough, as all the jobs were gone by August.

[alan]

Once you start stretching out your timescale, though, there is no reason to expect those jobs to turn into permanent jobs once the money spigot turns off.

That is what Krugnuts calls 'traction' which he doesn't bother to explain. Similarly, they called the fiscal equivalent 'priming the pump' back in the day, but no explanation for how that works out was never given either.

[alan]

Double negatives! Gasp! The horror.

[Old Mexican]

The results suggest that though the program did result in 2 million jobs "created or saved" by March 2010, net job creation was statistically indistinguishable from zero by August of this year. Taken at face value, this would suggest that the stimulus program (with an overall cost of $814 billion) worked only to generate temporary jobs at a cost of over $400,000 per worker.

How many here are willing to bet the farm that none of those workers ever received anything close to $400,000.00 in salary plus benefits??? I am.

[Jimbo]

They received about $50,000 or so...is that close enough?
So that's a mere $350,000 per job...just chump change.

[Ted]

There is really only one way to figure out if the fiscal stimulus worked.

You can either try to use some type of structural vector autoregression, This is problematic since New Keynesian models would tell you fiscal stimulus is only powerful when there is deflationary pressure and the nominal interest rate is fixed at approximately zero - and we don't have enough data to test a structural VAR on this. If you want to use a neoclassical model, which you can get multipliers above 1 in, you'd have to measure the productivity-enhancing features of government spending - which is incredibly difficult to do.

The next best way to do, which I think will be done sometime in the next 5 years. Someone would need to create a very sophisticated monetary DSGE model with all the bells and whistles (habit formation, capital accumulation, distortionary taxation, capital adjustment costs etc.) Apply the fiscal stimulus, generate impulse response functions and see if it matches the data. This should give us a good idea to what extend fiscal stimulus worked.

Let me make a few points about this paper first.

i) This is an improper methodology to test the effects of fiscal stimulus. Rational expectations tells us that the effects of fiscal stimulus should be immediate - when the projects start isn't completely relevant (though it does affect short-run capital accumulation) - this is also why the whole "a lot of the money hasn't been spent yet" in early 2009 was nonsense. Either you believe in the NK model you used to justify stimulus (which says it starts working immediately) or you don't. Where was his test for this expectational feature? Now, there is actually a much easier test. Since markets must be approximately efficient due to intertemporal arbitrage, all you had to do was watch asset prices, inflation expectations, and the foreign exchange market to figure out whether stimulus was going to "work" - immediately after it passed the chambers. The markets weren't particularly impressed. There was mild improvements, but nothing noticeable at al. This told me immediately the fiscal stimulus result would be very, very mild - in other words, the fiscal stimulus would fail and the budgetary cost wasn't worth the slight gains. No need for this fancy stuff really.

ii) Does he find it curious that it stopped in March? If you believe wages are sticky (which I think they obviously are - prices less so) then the impact of fiscal stimulus is affected by monetary policy. March was when the European debt crisis heated up. This no doubt pushed down the global Wicksellian interest rate and the failure to level target, meant that Federal Reserve policy became de facto contractionary. This would wipe out large potential gains from fiscal stimulus (even though I don't think they were there, it's wrong not to control for this).

iii) I see no attempt to look at state and local government spending. When you include state and local government spending in an ARIMA model, you'll find that due to state and local cut backs that the aggregate increase in government spending was not really above the predicted mean. You need to look at aggregate government spending increases relative to predicted trend - looking at just ARRA tells you almost nothing.

iv) The United State is not an optimal currency area. We are pretty close, but we are not (I think some studies find about 15 or so states would be better off with their own currency). Failure to control for regional asymmetries of monetary policy on the effects of fiscal stimulus in his state-level data is highly problematic.

v) I see no attempt for him to measure inflation expectations. When there is a regime switch from Passive Fiscal / Active Monetary policy to Active Fiscal / Passive Monetary policy, government spending should push up inflation expectations and if wages are sticky this spurs nominal spending. Unless you control for inflation expectations this study is no good.

So, this paper is highly problematic as a matter of testing whether stimulus worked. You need a lot more work before you make your conclusions.

My personally feelings are fiscal stimulus probably helped mildly (as the markets suggested it would), but the problem was the one illustrated by Erceg et al's brilliant paper (they basically took a New Keynesian model with sticky wages and prices and augmented it with financial frictions and distortionary taxation) The marginal multiplier is very different than the average multiplier. Once you push yourself above 1% of GDP in government spending, the multiplier drops to well below 1 (this is because of how it affects the regime switch in monetary/fiscal policy and distortionary taxation to finance the debt begins to matter more). Alternatively, there are New Keynesian models that argue that fiscal stimulus would hurt the economy substantially. If the liquidity trap is caused by sunspot fluctuations, then the fiscal stimulus makes everything worse and supply-side policies become optimal. It's funny that those economist who embrace animal spirits seem to embrace fiscal stimulus so strongly since arbitrary revisions in expectations make fiscal stimulus horrible policy. Anyway, fiscal stimulus shouldn't have been done. Every model where fiscal stimulus works only is true because they assume the Federal Reserve fails to keep nominal expectations stable. If the Federal Reserve committed to level targeting not only would the economy recover much faster, but fiscal stimulus would be pointless.

[Sal Minella]

The only thing that can be said for sure is: Spread evenly over the population, the stimulus cost each and every soul in the US about $3,000. And that can only be said if we didn't borrow or print any of the money. In fact, over time it will cost each and every one of us tens of thousands of dollars in principle, interest, and inflated prices. Nice job, Oblowme!

[Luap Leiht]

We should have just spent it all on hookers and beer. The results would have been more measurable and would have been more fun.

[sloopyinca]

No blow? You amateur!

[Luap Leiht]

I bow to your wisdom, sir. Your plan would, indeed, be more stimulating.

[Mike Bomstein]

The ARRA consisted of 3 parts, only one of which was the stimulus. The stimulus portion therefore was NOT $814 billion. Check it out.

[Luap Leiht]

So what did we get for the other two parts of $814 billion? Obviously, it wasn't jobs or economic growth.

[JohnDavis]

The "stimulus" bill was the biggest waste of tax payer money in the history of the world.

[Linda Lane]

.... next to subsidizing John Davis' education.

[Jeff Perren]

Even if, contrary to history and sound theory, the so-called stimulus created 10 million jobs for a dollar apiece the mere act of the Federal government spending the money would be immoral and unconstitutional.

The Federal government needs to be shaved back to its original intent, and with a very sharp razor.

[Sean Smith]

The reason the stimulus hasn't worked, doesn't work, and won't work is simple: There is no "Keynesian multiplier," and wealth is not created by randomly throwing borrowed or appropriated money at things. See http://estobrevis.squarespace......-side.html

[Andrew P]

This just proves that ARRA was a very badly done stimulus. If you want to create jobs, you build things like FDR did.

[PerryM]

18%+ unemployment and $5+ Trillion blown on buying votes for the next election - what's wrong with this picture?

Do you Libs really believe the Big Government can spend money more wisely than a Billionaire or Millionaire can?

$5+ Trillion down the drain and a generation of new Americans born into slavery to pay back the Libs failed experiment.

Where do those kids go to get their money back on our foolishness? Hopefully that generation will just refuse to pay our bills - then what?????

[DLM]

"Taxation without representation". Future generations will have to pay this back, but they certainly were not represented. I expect they'll be royally pissed when they get the bill.

[valwayne]

Hardly a surprise? Obama, Pelosi, and Reid used most of the nearly trillion dollars in the stimulous bill to pay off their special interests, in particular the public employee unions. The result of all the wasted money and massive debt it created has been massive long term UNEMPLOYMENT, and as we saw UNEMPLOYMENT just skyrocketed up just weeks after the election to 9.8%. Obama, Pelosi, and Reid are still in positions of power and will doubtless continue pushing their extreme left wing agenda at the expense of the economy and jobs! All we can do is hope a Republican House can limit the damage and our economy will finally bounce back despite their efforts against it!

[alaamiah]

only to generate temporary jobs at a cost of over $400,000 per worker. Even if the stimu