Questions that are rarely asked

If the multiplier is 1.4, recovery should have been accelerating pretty rapidly, right? Right? Bueller?

That is me on Twitter.  I am not sure what is the right way to think about the appropriate counterfactual here, but the explanation of “growth has not much accelerated because some other pending catastrophe was in the works” does not seem correct to me.

You can cite various reasons for economic slowness (“they cut state and local spending,” etc.), but the point of course is to explain the second derivative and then make that consistent with a multiplier estimate.  Or is it that government spending is supposed to have a higher multiplier than private economic activity?  Bueller?


Tyler, why do you assume that fiscal policy on net has been eased in the US? There is no evidence of that. There was a big easing in 2009, but as most of it was one-off there was actually net tigthening in particular in 2011. Cyclically adjusted the deficit is about the same as in 2007 - no net tightening. That is one counterfactual.

This is again neglecting the second derivative. There is in any case some positive growth, with a large multiplier that should increase at a decent rate, no? The multiplier presumably applies to all economic activity, not just government spending.

What are you assuming about the unobserved counter factual? If the multiplier was 1.4, given US fiscal policy of late and whatever else you think counts, what is the magnitude of the impact on the second derivative you think ought to be there?

We know what second derivative we are seeing, so I am doubting a multiplier of 1.4.

(I'm not an economics expert) but my question is whether the multiplier could be dampened by private deleveraging?

The government spending multiplier can only have a positive effect when government spending is increasing. The stimulus is over. Government has been shrinking as a share of gdp for a few quarters now.

Tyler this makes little sense to me. The shape of the government impulse has certainly not been uniform. There was a strong boost in 2009 and into 2010 and then some net tightening in 2010 and 2011. That seems to conform pretty well with the acceleration pattern (or second derivative) - pretty strong upturn in 2009 and into 2010 and then sluggishness since then - no?

I am not advocating the 1.4 multiplier. Just saying that the pattern of growth in this economy probably cannot be used to argue against a high multiplier. Perhaps to the contrary, the pattern of growth sits fairly well with an argument that fiscal policy has had an effect. In Europe I think it is quite obvious that the biggest fiscal tightening in 20 years in Southern Europe is part of the reason for the current deep recession. Not the only. But surely part of the explanation. I think it is pretty extreme not to accept that.

I am having trouble understanding the subtle arguments presented here, so let me confirm. TC is pointing out that we have seen positive, but non-accelerating growth, indicating that the multiplier is not 1.4 (or, more generally, not significantly greater than 1?) Valgreen is conceding, or at least not addressing, that point and is, in addition, pointing out that the positive, non-accelerating, and hence low-multiplier, growth that we have achieved was achieved with no net fiscal expansion, i.e., not only is the growth not self-reinforcing through a multiplier, we didn't even need fiscal expansion to get the growth. Is that right?

then maybe your model is wrong?

All the best economic claims begin "all else being equal ..." Does Tyler test such a claim, knowing that many, many, things were different? Does he demand that one big thing be named to explain departure from the pure relation?

In my amateur opinion, the Great Recession was one big mess. I think we are unlikely to extract any detailed metrics from it. At best we can probably only group positive and negative influences.


Economists love to speak in theoreticals, but the reality of the last 5 years should make them all rethink their assumptions.

Is the multiplier the only virtuous cycle in macro business cycles? It does appear that the economy is 'acceleration challenged' currently but it seems like there could be a few different channels for this (slowing trend growth, heightened pessimism, clogged housing channels). The US economy has recovered more quickly from all prior post WWII recessions, why would the forces be so different now? I know about the ill effects of financial crises and leverage, but a) that work is not as slam dunk either and b) the financial crisis hit was not as severe in countries which promptly addressed the crisis, and the US would not be a laggard in the bunch. So if this really all about the multiplier, why did it change so dramatically? And aren't the recent studies claiming it is higher at the ZIRB? One last thought...maybe people aren't framing it this way, but I have read a lot of analyses (and been asked a lot) about why the recovery is so weak...kind of a hot topic, really. But new framings are often a useful path forward.

Are you implying we should be seeing 40% growth? I don't think that is what keynes implies. Please clarify for the sake of your own dignity.

TC is saying here that recent economic activity is showing contradictory evidence to the assumption that the pultiplier falls between 1.4- 2.0.

Also that's not how a multiplier works. If the multiplier is 1.4, to expenditure, a growth in government spending of 10% would expect to see a growth in the economy of 14% (=10% * 1.4).

Sometimes, simple questions miss the mark. This isn't one of those cases.

Can you maybe help us out, Tyler, by saying what you think? I don't understand this query at all or who exactly it's directed at. Spelling things out in a slow way for the rest of us would be very helpful. Like "this guy and that guy and the other guy have claimed that such-and-such is true, but if that's true then ____ should have happened because ____ which isn't the case."

Why not just answer the question?

Simplest take is this: many people claim the multiplier is pretty high right now and recently. When it comes to private sector economic activity, we observe "acceleration problems" rather than dynamic leapfrogging gains. That suggests the multiplier might be pretty low. Of course one still might resort to the "multiplier from G is high but from private activity is low," though that seems weak to me and in any case I would like to see the argument. Contra Seth, no one is saying we should observe forty percent growth.

In all seriousness, "the multiplier" might not be so much "a thing." Macroeconomic BS. We'd all be so much better off if legislatures dived into the nuts and bolts of good and bad programs.

You know, at some point someone decided that an Iraq war was better than that many Pell grants, or whatever. It really boggles that a multiplier would be used to review that choice.

I think an example would help, since I don't think this is just about stimulus multipliers:

Back in January 2010, the central tendency of expected GDP growth in 2012 was 3.5 to 4.5 percent according to the FOMC’s Summary Economic Projection. ( According to the BEA this week, GDP grew 1.5 percent in 2012. One could justify this miss by: 'gee the multiplier is lower now, so we overestimated the add-on effects of stimulus...and all the other income flowing to the economy.' Here is Chairman Bernanke's take on why the FOMC missed at the December 2012 press conference:

"CHAIRMAN BERNANKE. Well, I think it’s fair to say that we have overestimated the pace of growth, total output growth, GDP growth, from the beginning of the recovery and we have had therefore—had to continue to scale down our estimates of output growth. But interestingly, at the same time, we have been more accurate, not perfectly accurate by any means, but we’ve been somewhat more accurate in forecasting unemployment. And how do you reconcile those two things? I talked about this in remarks I gave at the New York Economic Club recently, right before Thanksgiving. And I think the reconciliation is that what we’re learning is that, at least temporarily, the financial crisis may have reduced somewhat the underlying potential growth rate of the U.S. economy. It has interfered with business creation, with investment, with technological advances, and so on. And that can account for at least part of the somewhat slower growth. At the same time, though, what—of course, what monetary policy influences is not potential growth, not the underlying structural growth. That’s for— many other different kinds of policies affect that. What monetary policy affects, primarily, is the state of the business cycle, the amount of excess unemployment or the extent of recession in the economy. And there, I think we’ve also perhaps underestimated a bit the recession, but we’ve been much closer there. And I think therefore that we’ve been able to address that somewhat more effectively with quite accommodative policies. That being said, of course, we have over time, as we have seen disappointment in growth and job creation, we have obviously, as we did in September, have added accommodation and we’ve continued—we continue to reassess the outlook. I think it’s only fair to say that economic forecasting beyond a few quarters is very, very difficult. And what we basically are trying to do is create a plausible scenario which we think is reasonably likely—base policy on that, but we prepared to adjust as new information comes in and as the outlook changes, and inevitably it will."

(pg 22-23 Now the tweet response I got from TC was isn't that consistent with a lower multiplier? I'm not so sure about that. But even if the multiplier were low then SO WHAT? If we assume it was temporarily low then we would not change our forecast going forward. You will note the latest FOMC SEP shows acceleration two years out: So I guess the question back to TC is how is TGS a second moment and a first moment problem? And when will you add the third moment to our woes?

Oh and my sad excuse for a joke at the end was about derivatives, not moments. I've been tormented by the second moment a lot lately so it snuck in and made further nonsense.

So if I translate, Claudia, Bernanke is saying, "Well, our models weren't working, and we're really clueless why, so we've adjusted by dialing back the coefficients."

Did you read my comments about the EIA in our last interchange?

Let me tell you, Claudia, my macro models are working fine, and they're predicting oil demand and price movements accurately, too. And I actually have an explanatory model--a coherent narrative that's consistent with the oil and GDP data.

I have to tell you, I find it enormously dispiriting that four years after an oil shock, and in the face of persistently high oil prices, the Fed seems to have no--absolutely zero--sophistication regarding the impact of oil on the economy.

I disagree with your interpretation, above all your tone.

Good for you that your oil price forecasting model works so well ... you must be making a lot of money betting against energy futures.

Seriously, there are plenty of open questions and the tentative answers may not be very satisfying, but it will require a team effort...not putting people down.

Isn't government spending *obviously* supposed to have a higher multiplier than private economic activity, to the Keynesians? Otherwise debt-financed spending would be an anti-stimulus due to the money it takes out of private investors' hands.

I don't think that's the Keynesian's argument at all. Keynesians suggest that when the private sector is unable or unwilling to borrow to expand the economy, then government should.

They would argue government borrowing isn't crowding out private borrowing because interest rates are incredibly low. Were there more private demand for loans, interest rates would be higher.

There are times when government borrowing appears to crowd out private borrowing - I think that was behind the reasoning behind the tax increases of the early 90s - interest rates were relatively high, and government debt was thought to be part of the reason. Interesting that the government deficit/GDP ratio was much lower then than now.

Also, in the scenario you imagine, government borrowing would only be anti-stimulus if the government multiplier is lower than the private multiplier. That may be, but it's not the discussion here. If the government and private sector multipliers are identical, an economy would be indifferent to who was borrowing.


You say that interest rates are incredibly low, so government spending won't crowd out private investment. I believe the Keynesian view is that while the nominal interest rates are low, they are actually too HIGH for the market to clear, but the interest rates can't be reduced below zero, so we're stuck with a market that doesn't clear and thus high unemployment. The reason for the high multiplier at the zero lower bound of interst rates is that the additional government spending won't require interest rates to be raised to control inflation (since rates are already too high anyway). And since higher interest rates are the mechanism by which government spending crowds out private investment, no increase in interest rates=no crowding out of private sector spending. This alone might suggest a multiplier of only one, but then the extra government spending has knock-on effects in the private sector (government employees spending their paychecks, etc.) which is how the multiplier increases above one.

Are you talking about England?

If it is the US, when you talk about government spending, are you including state and local government spending? And, if you include tax cuts to "stimulate spending" in your model, please recognized that I took the added tax cut and put it into my savings account. Did you treat "tax cuts" from 2007-2009" as equivalent to government spending when the studies showed that the tax cuts did not have a positive other words, are you separating out the two types of stimulus?

The bottom line is that none of you have a good answer.

So saeth the Lord.

I think the bottom line is that your vague allusions to your own internal thought process cannot generate useful responses from others.

You seem to be implying that if there's a multiplier at all, then growth should always accelerate? I would wager that if you write like readers can't read your mind, you'll get more interesting results.

I do. Similar to 2001 when a surprising event caught almost everyone off guard and with their basic thinking severely challenged, the Neocons were the only ones who had anything resembling a rational narrative. Their narrative harkened back to a different time; Germany! Japan! Wilson!, and those in power in 2001 saw this and said, War, We can do this, and did.

In 2008 the basic economic thinking was thrown into disarray with the collapse, the assumptions of a generation were challenged. The only ones with anything resembling a rational narrative, again harkening back to a different time, Roosevelt! Keynes! Great Depression!. The politicians in 2008 saw this and said, Spend, Borrow. We can do this, and did.

Keynesian action didn't bring the US out of the Depression, and had been seemingly thoroughly discredited in the 1970's. I'm amazed that we hear said that there couldn't be inflation with high unemployment, as if ignoring the reality of the past makes it go away.

Stimulus doesn't work. Never has, the multipliers are pure imagination. Anyone in macro economics who uses a decimal point is either deluded or pulling your leg. In fact, the policy decisions of the last few years have actively impoverished people. The job numbers and stagnant economy are characteristics of this type of policy initiative.

"Stimulus doesn’t work. Never has"

I don't even understand what this could possibly mean. You act like there is some big difference between the government spending money and a business spending money. There is no guarantee either entity will do something useful with the money. I have worked for the government and currently work in private industry. In both places there was a similar correlation between my performance and my pay. And in both places I worked very hard not only because I wanted more pay but also because I cared about the products we were producing. The idea that government spending is somehow crowding out private spending is laughable. We could hand Google a billion dollars right now and all they would do is put it in the bank with the rest of their money. Same for a bunch of other very highly profitable corporations. There isn't any lack of money in private industry but there is a lack of desire to spend it.

Possibly you can convince me I am wrong and there is some big difference between a dollar spent by the government and one spent by Google. So far Krugman is much more convincing when he compares countries that have tried to reduce government spending with the US that (somewhat) hasn't. We just had the biggest financial shock since the Great Depression and people are complaining things haven't been fixed fast enough. It appears disingenuous to me. Maybe the multiplier is not 1.4 but you have a LONG way to go to convince me that reducing government spending in a depressed economy is a good idea.

Since government spending is coerced, it makes as much sense to include it in the measure of healthy economic activity as the amount taken in bank robberies.

It has nothing to do with whether or not the spending is coerced. It is whether or not the spending produces anything of value. Possibly you fall into the group that says the government can do nothing right other than the military of course and in that category they are the best in the world.

If I give you money to spend, you spend it, it is economic activity. But where does the money come from?

Government spending makes things happen, no doubt. But there is serious costs. First, do you believe that bending the yield curve on government bonds by market manipulation is healthy? That is one source of where the money is coming from. What happens when the money stops? Usually the economic activity stops because it has not created self perpetuating economic activity. There are exceptions, but not very many. Economic activity comes from individuals or groups of individuals taking their resources to fill what they see as a market demand. Competitive pressures keep the silliness to a minimum, and usually bad ideas are discarded early. Government money has none of the competitive harshness that keeps stupid ideas from sucking up resources. So when the government spigot stops, so does the activity. Another very dangerous, even worse result of government money is that it nurtures skills and structures whose goal is to be effective at getting money from government as opposed to surviving in a competitive marketplace. There was a study in the last few years that looked at the economies of districts where the Senator or member of Congress reached some committee level position, and was then able to funnel lots of cash to his district. The results were opposite of what most would assume; the economies suffered. Rogoff wrote a paper describing the results of high debt levels and found that at a certain point it didn't create prosperity but stagnation. The US is at or past that point.

Krugman assumes that there is no limit to the amount of money available for governments to borrow. From that assumption all kinds of great things can be promised. Unfortunately it isn't true. The situations he is describing as requiring more stimulus are the results of years and years of government overspending, resulting in insolvency or unsupportable debt servicing costs.

"First, do you believe that bending the yield curve on government bonds by market manipulation is healthy?"

What exactly are you referring to? I hope not the Fed:

" Competitive pressures keep the silliness to a minimum, and usually bad ideas are discarded early"

I agree. There is little to stop government spending. The market stops private spending very effectively. I think where we disagree is on whether or not a dollar spent by the government is inherently less efficient than a dollar spent in the private sector. I've worked with idiots in both sectors (probably more idiots in the private sector actually) so from my experience I don't automatically expect a dollar spent by the government to be wasted and I don't expect a dollar spent in the private sector to bear fruit.

"Krugman assumes that there is no limit to the amount of money available for governments to borrow. From that assumption all kinds of great things can be promised. Unfortunately it isn’t true."

This is where the Japan thing comes up. And with borrowing costs so low it does seem like a good time to fix some of our infrastructure issues. I just believe Krugman's narrative more than I believe his detractors. He may just be better with smoke and mirrors.

I'm confused. Aren't these multipliers estimated at an assumed equilibrium state? I never got around to taking ODE, but why would this imply a non-zero second derivative?

As a non-economist I was going to stay out of the discussion but with this challenge I can't resist.

Certain sets of perimeters that overall velocity work within and gain specific patterns from have basically been filled in their current definition and form, which makes the already questionable aspect of multipliers even more murky. The perimeters for economic activity needed to be widened, redefined and made more flexible.

Sorry the reply was to Tyler!

I think most of us are still confused about what exactly you are asking. Looking at FRED, federal government expenditures have been flat for the past two years. I don't see why anyone should expect that to be associated with growth regardless of the multiplier.

The Fed acts second to last, followed by the deleverager?

i don't understand. you are the economist. shouldn't you be answering the question?

He did answer the question "Is the multiplier 1.4?" The answer is no.

The question he asked is to the people who say otherwise. He's giving them an opportunity to explain their position. How does observation reconcile with other observations or their theory?



Damn, I must be an economist. I meant- 1.54.

OK, let's set up the counterfactual and ask how that would have compared.

No stimulus. State and local government cutting spending even more to balance budgets. Myopic humans saving like crazy in banks that are at risk going under. No extended unemployment benefits. No federal spending of around $400B, excluding tax cut or state and local support. Federal government instead reduces spending.

Anyone willing to play this game other than Herbert Hoover?

Government spending was supposed to "jump start" the economy or "prime the pump." It didn't. You are only arguing that we have stopped priming before the water is flowing or whatever metaphor you wish to use. When does the self-sustaining recovery begin? Are we at full employment GDP or not? Do we still have a recessionary gap?

I grant that state spending might be a factor. But then where is state government spending in Y = C + I + G + NX?

I know a couple of places it could be, but where is it measured and what is its multiplier?

The self sustaining recovery begins when there is enough demand in the economy that the Fed has to start raising interstate rates to keep inflationary pressures in check. As it stands, the Fed can't lower interest rates enough to stimulate the economy to the point of full employment (does anyone doubt that if the Fed's rate was currently 2% instead of the current 0% that they'd be lowering it further to help stimulate the economy?). As long as the Fed is stuck with interest rates higher than they'd actually like, no self-sustaining recovery. Until that time, more government borrowing means higher employment and higher GDP.

Why didn't you just say "liquidity trap?"

How does state spending (which is supposed to be on a balanced budget) figure into this?

Does the current debt level, absolute or relative to GDP (including future spending obligations) figure into optimal policy at all? Is there no upper bound for deficits and debt?

You also haven't resolved the multiplier question.

And regardless of the lower bound on nominal interest rates, the Fed still has substantial instruments of monetary stimulus.

Even back in the bad old days, when a fair-to-poor stimulus was being designed, people argued separate multipliers for separate components. Even stimulus boosters called some elements, bad, with poor multipliers. It's it kind of a cheat to go back and say "hey you promised 1.4 for everything, black swans be damned!"

So you're admitting we wasted a large portion of those trillions in stimulus?

The argument isn't over whether individual components of spending separately had multipliers of 1.4 but whether the joint multiplier was 1.4.

So you bring up the black swans, apparently saying it's George Bush' fault the stimulus isn't working, Obama and the Bernanke and Timmy gave it the good ole college try, that things would have been worse if they didn't, and we just need to do more of the same.

Let's assume that's all correct. Before we waste another few trillion dollars, can you tell me what the multiplier will be, how it will be spent, when it will work, what condition our balance sheet will be in at the end, who will pay for it, and how we will get the money from them?

That's a pretty messy answer, assigning to me all kinds of positions I do not have.

For what it's worth, my input (as if congress cared) at the time of the stimulus was that we should do true counter-cyclical spending, not inventing needs, but moving forward "things we were going to do anyway." I wanted it to be program based. I wanted it to be centered on a search for good programs. I wanted ROI.

So now you are attempting attach the "bucket o' stuff" stimulus to me, and ask why not 1.4?

Gawd, I hope Tyler has a better agenda that that. Hopefully he really is calling BS on multipliers, and the idea that "generic spending" is a good measure or focus.

I apologize John. Apparently I inferred too much.

Your suggestions are facially reasonable, and I addressed precisely that in another response on this thread.

The trouble with bringing projects forward that we were going to do anyway is that we scrap the remaining useful life of the existing asset. This is waste of the national balance sheet. I can live with that, especially if interest rates are so low that projects should be brought forward anyway.

But this creates artificial demand for the resources used in the project, and those resources will be displaced later. Resources such as labor will be repurposed and repriced. That creates adjustment problems later, unless you plan to build lots of bridges and dams every four years.

The public projects also raise resource costs for private enterprise. This is a form of crowding out not captured in standard macro models that use interest rates as the mechanism through which this occurs.

A very sharp V shaped downturn and recovery. We have historical precedents.

The depression lasted for a long long time and was made worse by policy decisions of Roosevelt. Things turned around when his active antipathy towards capital changed when he needed them to win a war.

I think this recovery looks more like a mirror image of the square root symbol, √

The recession is more like a U with the right leg leaning far to the right and rising gradually from a very low depth. I'm surprised no one has offered the counter factual as a possibility that could have worked. Why don't we try it? Why is it that the government needs to make the pain less for its citizens? How is it that the government knows how much each citizen can tolerate in the aggregate? What would have happened if the market for supply and demand actually found it's true clearing price? I would surmise that yes, it would be painful, but upon finding a clearing price for supply/demand of all products, buyers would step in as they saw value. This animal instinct would multiply rapidly as others noticed that yes, risk is in fact good and rewarding. Very quickly prices would recover and the pain would subside. Government intervention distorts animal instincts, the value of risk taking and provides alternative incentives that therein distort the market place. It would be nice if for once, someone would recognize that if you're putting no money down on a $1million McMansion in an average suburb that both you and the bank should lose your money because both of you didn't know the value of the asset. A better investment of government resources would be to implement and education policy that required economics to be taught at every level of primary school. Perhaps in a few decades we'd have an educated society that knows value and understands how supply/demand works.

Yes. All recessions are exactly alike.


I think you make a mistake by assuming > zero private sector growth would have a positive (accelerating) multiplier effect and < zero growth would have a negative (decelerating) effect. We would never apply that standard to rates of government spending. You must compare both government spending and private growth rates to some (probably unobservable) expectations of spending and growth prior to the crisis. If spending/growth come in below these expectations they are exerting a drag on the economy; if they come in above these expectations they are exerting a thrust.

I think the explanation is the other variable, time, that you have to account for when calculating a rate. The stimulus effect COULD have been that high for a very short period of time. The short run is perhaps a very short period of time when the financial system is in chaos.

Now we have to face the long term issue of increasing competitiveness/efficiency and pay off the debts caused by the stimulus attempts.

Stimulus right now is probably less than 1 since the economy is stable.

Growth happens either from newly-discovered rent yielding assets or new efficiencies. I haven't seen much of either in the US except perhaps for increased natural gas production.

I doubt that will be enough to counter the effects of our changing demographics.

"Growth happens either from newly-discovered rent yielding assets or new efficiencies. I haven’t seen much of either in the US except perhaps for increased natural gas production."

In other words, the great stagnation.

Western civilization is failing.

No, Western civ isn't failing, it's changing.

If failure is ever increasing global standards of living and lifespans, which is all we've had, then bring on more failure!

Government spending can have a higher multiplier than private spending if it goes to people with a higher propensity to spend. Food stamps are an excellent example: they get to skip one more step before the propensity to save kicks in and cuts growth.

Additionally, government spending cuts uncertainty in a way private spending does not, since the government makes credible commitments to spend over the short term. If Europe craters, government spending will remain unaffected but private spending will probably go with it.

You regurgitated textbook theory. You didn't answer his question. TC knows how fiscal policy is supposed to work. If the multiplier is 1.4, why aren't we seeing stronger growth? If we aren't seeing strong growth, is there a reason for it or is the claim that the multiplier is 1.4 incorrect?

So is Tyler trying to convince us the multiplier is 1.39 or .9? He is such a troll on things like this. Doesn't say anything you can actually pin him down on. Isn't stimulus effective as long as the multiplier is > 1? That seems like the much bigger issue as opposed to arguing its absolute size.

TC is provoking thoughtful discussion, not "trolling."

His argument would be identical if the claim was that the multiplier is 1.39, 1.2, or 1.01. With a multiplier greater than 1, we should be having an increasing rate of growth.

He is saying that if the multiplier is 1.4, then what part of your model jibes with a declining growth rate. If you have no such explanation, then he calls into doubt your claim that the multiplier is 1.4 or 1.2 or even 1.01.

If I got that wrong, I hope TC or someone else corrects me.

I don't know anyone that runs a business that has both government contracts and private contracts that thinks the government contract is more certain.

She's talking about the animal spirits of investment versus the steady hand of government spending. The government's hand trembles when the debt is too damned high.

Why believe the multiplier is 1.4? I don't really see don't really see what path it would take to give that large a bang for the buck.

As for the y'' to make sense out of what it is and if it makes sense don't we need to know where we are relative to the local extrema?

Wish there was an edit button.

I wouldn't have noticed the error if you hadn't said anything about it. :)

It's like that sentence with a word that is repeated at the end of one line and the beginning of the next, and you're asked to find the error.

I agree with what you said, and I think we all understand tour statement.

"The" multiplier is a parameter of a model of the economy. To answer Tyler's question we neeed to see the model and how well it fits the data. And yes "the" multiplier will vary according to the compoition of the postulated autonomous change in spending.

TC is a chess player, nay, a chess master. Do you know what chess players are good at? Retrograde analysis. This is where you construct a position on the chess board and the describe what moves lead up to it. It's very hard to do, but it's the basis for winning in chess (you imagine what position will win for you in the future, given your present position, then play the moves to get to that future position). None of you can answer TC's question because all of you--with a few exceptions--are all intellectual patzers. Me? I have no clear idea what TC is talking about either, but second derivative means the acceleration of the rate of change, so if Multiplier is > 1, which means for every dollar spent by government more than one dollar is ultimately spent by non-government, then it implies the second derivative is positive, akin to a geometric progression (see where r > 1, which is an infinite open loop--hence acceleration (the curve is concave down or 'exponential'/ 'run-away'). However, if r < 1 you will get a finite sum, or concave down, and hence the second derivative is less than zero. So TC is pointing out, via a rhetorical question, that the multiplier is less than one, otherwise we'd have seen unemployment at 3% by now. Speaking as a strong Class B or weak Class A chess player (and I'm getting better; I got game from reading this blog for some reason). 1. e4 (best by test). Your move!

I think your description is on target, but I don't think you made it more clear for us patzers.

A couple of points:

1) the multiplier could be > 1, but small (this seems unlikely)
2) the multiplier is a step function and for some periods of time it's > 1 and at other times it's <= 1 (which kind of fits the data, better)

If 2 is correct and assuming that the multiplier has dropped to < 1, then government stimulus spending is, in aggregate, lowering economic growth potential.

Of course not all government spending is equal. Some types of spending (extended unemployment benefits for example) probably have a high multiplier. Other types (investments in solar production firms) probably have a low multiplier.

As I say above, I think the caveat "not all government spending is equal" is the real story, as well as the fact that it has played out against a very, very, confused background. I personally disliked cash-for-clunkers, but even something that "simple" would be horribly complex to decode ... with follow-on impacts on vehicle retention 10 years from now.

Yes, of course there are many nuances of the economic responses to stimulus programs. I've done research on that as have many, many other economists. I don't think 1.4 is being set as a benchmark here. But the question is where's the acceleration? Stimulus is supposed to support a recovery (not necessarily long term growth). And if there's been no acceleration that means something was either getting worse or stimulus did not have the usual add on effects. I don't know the answer but I do think the question is legitimate.

Is the question one of historical curiosity? What if we cynically assume that future congresses would be about as bad as that one, and that economic crises would be as muddy as the last one?

It makes me think we should ask congress (such as we can) to buckle down, to look for good stimulus plans, with positive ROI. After that, let another generation of economists fight over "the multiplier."

I think you frame the question nicely. And if I understand correctly, Tyler is requesting an explanation of why the stimulus has been so ineffective. Are there factors exogenous (or perhaps endogenous) to the spending that diminished its efficacy. What are those factors and how do we fix it? If the situation would have been worse without the stimulus, present your counterfactual and defend it.

How does your model jibe with the observed spending and the effect on GDP and unemployment?

BTW, thanks for posting the Bernanke response. I had missed that and you saved me the trouble of searching for it. The Bernake's response was....predictable. it has the virtue of being plausible but the vice of being the usual excuse for failure. If I were in his position, I'm not sure I would act or say differently, so my critique is that of the armchair quarterback.

john - OMG, none of this is a historical curiosity. Well-designed, effective government spending will always be important. I just don't think that's the question here.

willitts - I actually don't think (could be wrong) that Tyler is attacking stimulus, in particular, here. Sure stimulus was part of the income flowing to the economy in the recovery, and it's something concrete to discuss, but it was not the only support by any means. I think he's calling more broadly into question the recent multiplier on all economic activity. A deeper question than effective stimulus. Bernanke was noting that recession seemed to have temporarily reduced our economic capacity and Tyler is suggesting that our economy also lost its virtuous circles. While the former is clear this is temporary, I suspect the latter might lean on permanent, though I don't know.

willitts, if you disbelieve "immaculate spending" then you must accept that debt financed stimulus created some demand. How much do you want to buy? What kind? I think that's always going to be more arbitrary than economists would dream. But certainly if we'd bought less stimulus we'd have had less demand. I can conceive no reasonable human behaviors to make the opposite happen (that stimulus so frightens the peoples that they spend even less) en masse.

@Claudia and John

I don't disagree with you.

My point is this: if the government spending temporarily fills a gap in C or I, and the recession ends before the extra G burns off, that's fine.

But the extra G is being sold as "stimulus" to get C and I on a permanent track of higher levels.

I don't dispute that a boost of G increases Y. But if it does not create a self-sustaining increase in C or I, it is NOT stimulus.

And even if the G is just spending as a stop gap measure to ease the pain, there will be crowding out through resources costs, and it is more likely than not that the G will be spent on precisely what which we do not need because that is where the pain is. It is a hair of the dog policy. It has the likelihood of prolonging the recession, making the pain worse, and wasting the future revenues that will pay down the debt.

There is a valid economic argument to do nothing at all. Of course, that is not politically feasible. I would transform unemployment insurance into a fully funded program of individual accounts or policies instead of bogus social insurance. State unemployment funds were bone dry because the source of its revenues were tied to current economic conditions. This is procyclical policy.

Claudia, I think that Tyler is suggesting that the virtuous cycles could be the reason why the multiplier isn't as high as the standard models predict. He wants an explanation of how you can stick to your models,claim a multiplier of 1.4, and have a growth rate that is declining. It is anomalous. But I withdraw from claiming to read TCs mind, so invite his reply.

I think this is a correct view. If we think government stimulus is efficacious, I think it is now clear that how we spend it really matters. It is, after all, intended to fill gaps in investment and consumption. The danger is that investment and consumption contracted for rational reasons, and there is no particular reason for boosting spending on what we spent before, particularly if those resources contributed to a glut of long lived assets.

The trouble for the central planner is to take an educated guess as to what the most productive investments would be, and avoiding the standard public choice problem of rent seeking. Clearly our spending on alternative energy was a mistake. Cash for Clunkers was an enormous waste of resources. In fact, all spending intended to draw future production to the present implies wasting still-valuable resources today. For example, if we move up the replacement time on a bridge for the same of employing laborers until the project is completed, we toss away the remaining useful life of the old bridge. We are depleting assets in order to increase our income statement. This might be good policy, but we must understand our losses. We must also understand that when we employ workers to build a bridge today, there will be no bridge for them to build tomorrow, and we have deprived every private user of that labor and natural resources of their use. We distort private incentives for resource allocation through space and time. This is not without consequences.

To amplify what JP said, size matters. If we could prove that stimulus had a multiplier of 2, but the stimulus was $20, we would see no measurable effect.


"all spending intended to draw future production to the present implies wasting still-valuable resources today" Not if there a plus-one multiplier and the economy is below potential!! There is a huge benefit to giving the economy a kick in the pants when it is operating below capacity (high unemployment). I think there is a fair bit of confusion here between marginal propensity to consume out of stimulus and multipliers. Even with a low MPC a plus-one multiplier would assume an add-on effect...and give you acceleration. (And note, MPCs and multipliers are not just stimulus concepts.) We need a virtuous cycle to start as it has in past recoveries. As I am told, macro forecasters are always surprised when the economy 'pops' in a recovery, sadly I got to learn forecasting when the surprise is the lack of a pop.

I don't believe they commonly throw away bridges. In most cases refurbishment creates a waste which is just a few years (the most you'd want to "pull forward") of use. You save maintenance costs on the worn bridge in those "last years." And importantly you avoid all that "crowding out" you'd do, working on the bridge during a private sector construction boom. I think that's true of maintenance-infrastructure projects in general.

Cash for clunkers couldn't really be that because it was essentially a bid for people to find cars worth less that the incentive. The system overpaid for the utility still in those old cars, which were scrapped.

" Even with a low MPC a plus-one multiplier would assume an add-on effect…and give you acceleration"

Confused. Stimulus is not a perpetual motion machine. Obviously something is slowing the growth. Terminal velocity exists even though gravity is constant. The multiplier does not have to cause acceleration.

I don't recall calling stimulus magic. It provides a temporary boost to the economy. When introduced it supports private spending (boost to GDP growth) and when it expires that support is withdrawn (drag on GDP growth). In ideal stabilization policy, stimulus is withdrawn when the economy is back on a firm footing, solid/strong GDP growth. And we think, in a plus-one-multiplier world that stimulus helps to not just plug a hole, but also generate other add-on spending. Through 2012 stimulus has been a net positive on growth, true to direct boost to growth is less, but there should have been sizeable add on effects if you believe in a 1.4 multiplier.

Oh and more simply the .4 part of the multiplier are the second, third, and later round effects not the on impact effects. That's why you'd expect acceleration, though I agree it would matter on the pattern of the stimulus impact (and the lack of any other other shocks) to see the add on effects.

Sorry I didn't mean to imply you called stimulus magic. My point is wondering where the add on effects are is the equivalent of questioning gravity when the skydiver reaches terminal velocity. It seems more likely to me that the multiplier is greater than one yet something we cannot see (similar to air) is negating the effect.

We once had like one honest-to-goodness Keynesian who used numbers and stuff that commented here. We probably ran him off.

When God commanded, "Be fruitful and multiply," I'm fairly certain he didn't mean government spending and open market operations. :)

If all the economists would disappear tomorrow would the US be a better place? Would anybody care? Those are perhaps the more relevant questions.

I've wondered if one were to regress various economic stats -- median income, % under poverty, size of government, and the like with number of economists if the coefficient on the economists variable would be positive, negative or 0.

Read the IMF Working Paper, "What Determines Government Spending Multipliers?" by Giancarlo Corsetti, Andre Meier, and Gernot J. Müller.

Thanks for the paper. I read it, and it says: "The multiplier is a fudge factor". Ergo, from a science point of view the multiplier is either a tautology or meaningless, unless you can somehow figure out how it varies over time and by country. Excerpt: "From a theoretical perspective, however, there is no such thing as the multiplier. Instead, fiscal multipliers are likely to depend on a number of factors which vary both across countries and time."

I'll say it again: there is also a multiplier from an initial round of private economic activity. Some of you are writing as if the only multipliers are from government.

Were the government stimulus immediately tax funded it would be a 1:1 opportunity cost. We can't worry about that and spending the grandkids' inheritance at the same time.

And I certainly don't think unfunded government spending crowds out private spending.

I'd guess that they are more or less equivalent. A tax cut or cash transfer is the same as a private cash gift. Public construction is the same as private construction. A grocery store gift card is the same as food stamps. However, we don't care about the private sector multiplier because we can't collectively control it. Given a set of conditions, people are going to spend what they are going to spend. What we can do is to change those conditions via monetary and fiscal policy.

Tax policy was one of the things in stimulus mode before 2007.

And if private corporations were spending money rather than saving it we would not be discussing this at all.

You didn't read it.

Just to clarify, are you saying that the multiplier on that initial round of private activity differs from the multiplier associated with government spending or that any initial activity will be multiplied?

By the way, I've always believed that the US government was running stimulus as it hit the 2007 troubles. As a general believer in counter-cyclical spending I've always found that really unfortunate. No one could possibly "start stimulus" in 2009. They could only "heap more on." In my opinion the difficulty you have "heaping" doesn't really undermine the counter-cyclical principle.

I believe the answer is - as you alluded to - that the multiplier in the private economy is currently less than 1.4, and I would suggest less than 1. Just as there are different estimated multipliers for the various spending that the government could undertake, there are different multipliers for the various transactions undertaken in the private economy, and many of these multipliers are less than 1.
Let me clarify your hypothesis. If the multiplier for all actors in the economy in aggregate was 1.4, then for every transaction there would be a multiplier on that transaction’s value and the growth of the economy would accelerate at a greater and greater rate. GDP would be growing exponentially.
Now let’s consider a normally functioning economy growing a constant rate, say 3% GDP per year. If the growth is constant, then the second derivative is zero, it is neither accelerating nor decelerating. I would suggest that in an economy like this, in aggregate, the multiplier is about 1, and that the Fed uses monetary policy to fine tune that multiplier to get the desired growth. If, in aggregate, the economy wide multiplier is less than one, then the economy would begin to shrink, but the Fed injects some liquidity to keep growth on track. Conversely, if the multiplier was much more than one, growth in the economy would be accelerating and the Fed would be reducing liquidity to keep growth in check. (My theory assumes that multipliers are basically a proxy for velocity of money resulting from a given transaction).
Now consider the current economy. If the government decides to borrow money and spend it, it’s increasing the velocity of money: Right now the Fed isn’t reducing the money supply in response to additional government borrowing, the private economy isn’t reducing their spending in direct response to additional government spending since interest rates and taxes aren’t being increased in response to the additional borrowing (some would say this assumption is incorrect, but as long as the change is small enough, the multiplier theory can still hold). The additional government borrowing and spending alone would suggest a multiplier of 1, but once that money is out in the private economy, there are going to be further knock-on transactions which increase the multiplier beyond 1.
However, as I suggested, I believe that in aggregate the private economy currently has a multiplier of less than 1. Many actors in the economy aren’t increasing the velocity of money, instead, when they receive more money they are increasing their savings, pulling that money out of circulation, reducing the velocity of money (think corporations stockpiling cash, or a retiree trying to get their retirement plans back on track after the value of their house fell by half). Sure, there are some actors spending everything they get, some companies products are still in great demand so they’ll re-invest everything that comes in on additional equipment, or new hires. Likewise, a person earning minimum wage may be spending everything they earn, just to get by. But I believe that right now the savers are overwhelming the spenders in the private economy which is how you’d end up with a private economy multiplier of less than 1, and that would explain why growth in the economy as a whole isn’t accelerating.
Of course the normal response to this situation would be for the Fed to reduce interest rates to so that savers would have less incentive to save, and spenders would have more incentive to borrow and spend, but with Fed rates stuck at the ZLB they’ve lost that ability and now have to rely on less effective mechanisms.

Isn't this the whole point? The stimulus was too small, we heard, ( obvious implication being that the multiplier is really we needed much much more)

If the multiplier was < 1, then we'd be shooting ourselves in the head.

Not necessarily. A positive multiplier indicates that GDP is higher than it would have been without the stimulus. As long as what we are spending the stimulus on is useful, we will be richer than we otherwise have been.

Discussions about the multiplier are like discussions on the existence of God.
I, for one, welcome your comment, Tyler.


I for one still don't understand how a stimulus is supposed to work when corporations are already experiencing profits and banking cash at record rates. If profits are not in enough to get business investing how is handing the poorest 20% of Americans 200 extra buck each (or whatever your favorite stimulus plan is) supposed to change that?

It changes the return on the marginal dollar. Profit isn't a sign of health if it is the result of sitting on cash rather than reinvesting it.

If there are unemployed resources in the economy, companies can employ those resources to capture the additional money in the economy and have that money become production rather than inflation.

I could understand that if it was true, but it is not. With interest rates so low profits not the result of sitting on cash. The cash piles are the result of the not wanting to invest to capture more profits. Your argument would make more sense to me if profit margins were at historic lows. Then I could see business not wanting to reinvest.

Can't tell if you are asking a serious question... in general poor people spend their money they do not save it. Was just reading something about how saving stimulates the economy. Had this great line in it "Sometimes we buy stocks, which provides companies with money to expand." Ha!

It was a serious question but I don't think you understood it. Believe me I know from first hand experience that most poor people don't save money and I know that private companies are not spending money even thought they have it. The question is whether the money spent by poor people that is given to them by the government will have a multiplier affect. And I don't understand how it will when poor people will just give that money to companies that already don't seem to be responding to profit signals.

Put it another way: How is giving money to poor people to give to corporations more likely to cause a multiplier effect than the government giving money to corporations directly. I understand why the second option does not work. I just don't understand why first is considered to have a better multiplier.

Would you consider infrastructure spending giving it directly to the companies? I do see your point now.

I can understand in theory how infrastructure spending might have a multiplier effect over the long term. But I thought the consensus among the great and good was that infrastructure projects take too long to get off the ground to be very effective as a stimulus. Certainly in the government construction projects that I am involved in there is usually at least a year between when bids are solicited and any money is spent. And that does not count the time it takes to make a decisions on soliciting bids.

Sounds like the complaints about projects not being "shovel ready" are real. My belief was that the multiplier was the impact on GDP and Tyler is saying if the multiplier is greater than one then not only should a constant stimulus increase GDP that increase should accelerate over time. So where is the acceleration? The question seems somewhat ridiculous to me. Clearly there are things working against increasing GDP (have you tried getting a loan recently for example). For Tyler's question to be relevant it implies just give the US sailboat a little breeze in the global economic sea and it will soon be cruising like a speedboat.

If the multiplier is a constant, any constant, the second derivative is zero, no? Am I being stupid or is something here not very clear?

f=ma. A constant force doesn't just move an object it accelerates it. Of course there is the minor matter of friction. But certainly there wouldn't be any of that in the economy.

What are the variables we are talking about here? Which one are we taking derivatives of, please?

I love when you talk mathy to me, but I'm getting confused by the terminology. So, if GDP is X, and GDP growth is X', then I'm guessing that since X' has been pretty constant that means that we haven't accelerated X' (e.g. X'') so as to catch up with trend GDP. But we don't 'observe' the second derivative, we observe the integral.

I observe the second derivative of GDP, in particular consumption, all the time. Look at the 2010 SEP that I linked to above. It has forecasts for GDP growth in three years...see how they get bigger? A forecast of a pick up in growth (acceleration), then compare to the latest published numbers. No acceleration. I think the multiplier is probably too simplistic a way to diagnose the problem, but it's fair game to ask since conceptually it should have been providing some acceleration. Something else had to be decelerating at the same time to offset it.

It's conversations like this that make me wonder if maybe the French aren't wrong to teach Economics as a philosophy.

What if -and I'm just spit balling here- the size of "the multiplier" depends on the type of spending, and factor of production being employed relative to its slack in the economy. So there is a larger multiplier for employing, say, unemployed construction workers or under employed education majors, then there is if Deloitte hires more actuaries and accountants. You could have a large private multiplier if, somehow, the debris from the housing crash were cleared up, and demand for new housing increased.

In that universe, the second derivative would be positive for some time, and then negative after that, as the underemployed factors of production are gainfully employed. Maybe we could summarize that relationship as saying that a 1 unit increase government spending results in a 1.4 unit change in GDP for people who aren't thinking about it as a differential equation.

shorter TC: i don't believe in a multiplier for government spending because private sector spending is anemic. and therefore i'll talk about first versus second derivatives.


^ 'where's the acceleration?' ftfy

I don't understand your complaint. If a multiplier for government spending does not cause private sector spending to cease to be anemic, then what good is a multiplier? You might argue that private sector spending would be even worse without the multiplier. But if the Multiplier is real, the only place to look for it is in private sector spending and Tyler is saying he cannot see it there. To me, you are asking Tyler to prove something when really the burden of truth should be on those arguing that a multiplier exists.

In other words, by what methodology do you think we should use to discern the multiplier in private sector spending if even you agree that private sector spending is anemic?

When the government bumps up against a debt wall (e.g. downgrade by credit ratings agencies) then marginal government spending increases may produce marginal private retrenching.

Some seem to suggest no difference between the private sector spending and having the government spend that money.

But surely there is a difference between a decentralized private entity spending on what it sees as a recurring market for its goods, and a faraway government spending on a recurring cost, which much of the stimulus was in the form of public workers and their salaries.

The former stimulates markets and growth while the latter creates fear with deficits and debt.

Registered No – 1026167
Registered office
1 Churchill Place,
London E14 5HP.

Barclays is a trading name of Barclays Bank PLC and its subsidiaries. Barclays Bank PLC is registered in England and authorized and regulated by the Financial Services Authority (FSA No. 122702). We’re one of the largest financial services providers in the world, Also we engage in retail banking, credit cards, corporate and investment banking, wealth and investment management.

We are here to introduce a loan program that will help improve you financially and our only focus is on providing you with great service and helping you meet your financial needs.

We offer a low rate at 2% interest. We do however receive commission from payday lenders and brokers when customers enter into a consumer credit agreement with them, having been introduced via our service.


*Home Owner Loans
*Graduate Loan
*Debt Consolidation
*Professional and Career Loan

Applicants interested in this loan offer are to SEND the details information to the account overleaf.


Full Name:
Loan Amount Needed:
Duration of Repayment:
Marital Status:

Enjoy Best Services Online with Barclays

Tyler, some of us have a very good answer. Keynes, as explained by M.Pettis a few days ago, implicitly, at least, asserted there is no multiplier when you have a current account deficit. We have been running a multi-year experiment proving him correct.

What we needed was big deficits in China, big enough to reduce the current account deficit here.

Go back and read what you then agreed with

I would like to know if Tyler believes the multiplier in the private economy can ever go negative, i.e. a transaction taking place today actually causes GDP to be lower tomorrow than it would have been had the transaction not occurred.

I think the fact that an economy can go from positive growth to negative growth, with no change in government/Fed policy, explicitly shows that in aggregate the multiplier in the private economy must at some times be negative. How else to you get negatively accelerating growth?

And if it's true that the multiplier can be negative at the onset of a recession, is it hard to imagine that the multiplier can still be negative while an economy is in the depths of a recession? Can't it still be negative right now?

The impact multiplier for all income together must always=0, if you're doing a same period autoregression of, say, GDP to GDP, or income to income; a one-period lagged impact multiplier should give you a slightly positive multiplier. A cumulative multiplier would be the sum of the AR coefficients of some set of lagging periods' GDP changes, but the second derivative of the resulting curve (the acceleration rate of the economy) would not have to be high in cases of a high (>1) cumulative multiplier, if the number of periods which the shock affects are large (the effect is more spread out).

You may also not be seeing a rapid recovery from the positive fiscal shock from increased G and (X-Im) specifically or improved spending in some segments of the private economy due to offsetting negative shocks from somewhere else (state and local G, decreased I or C), of a size that partially counteracts the positive stimulus.

The 2012 IMF paper by Baum, Poplawski-Ribeiro, and Weber presents cumulative multipliers that are in the range of 1.4 or more for US fiscal increases in down times. Onset here and elsewhere appears to be relatively slow, suggesting that a high second derivative isn't in the cards, even with a high multiplier.

But what about 2012 issue of AEJ: Economic Policy? E.g. Auerbach and Gorodnichenko paper: multipliers seem to be less than 1 in expansions and most of the time higher than 1 in recessions. They have a nice follow-up on OECD.

Comments for this post are closed