Measuring stimulus

Thanks for the clarification.

I disagree that the CBO numbers are not estimates. They are estimated magnitudes of ARRA, based on theoretical models with calibrations for the magnitudes of predicted effects. I think it's the only procedure possible for answering questions of what effects ARRA had.

It's not measurement, and it is theory laden. It's also not fundamentally different from what we do when we ask what the effect of some microeconomic even on a particular market. A flood occurs in Iowa and we claim to see corn prices go up, but we cannot measure against what price would have been w/o the flood.

Our micro theory is much less controversial, but the exercise remains theory laden.

I'm posting additional thoughts on my own blog for anyone interested.

When a "counterfactual" model says that the stimulus worked, such models are not to be trusted. no way. snake oil.
When another model somewhat indicates that the central bank creates financial booms and busts, this modelled evidence is hailed as the ultimate truth.

Just two post next to each other.


Is it hypocrisy? Confirmation bias?

If the multiplier was anywhere near 1.5, governments the world over would constantly be enacting stimulus plans that were 10,000x the size of their entire economies.

A >1 "multiplier" exists only in the minds of the alchemists who call themselves Keynesians. Since their theory cannot possible be proven or refuted, and since they provide intellectual cover to power-hungry politicians and since most people arne't interested in economics, they persist in influence. Santa Claus polls quite well, it seems.

I'e read DK state, essentially, that the estimate is a net impact analysis, which has considered "all" costs and benefits, seen and unseen. The problem with this view is that it takes what a model purports to do and treats it as if that is what the model actually does, while acknowledging some identificaiton problems as almost a throw-away. This is folly. It is impossible to know and account for everything, even if you move the goal-posts when challenged and try to persuade us you only think macro stabilization is what we can rationally hope to achieve and predict/forecast/estimate. No one on this planet fully comprehends or can capture or measure the destabilizing effects of the potential malinvestment created by the stimulus, and the stimulus does and did not exist in a vaccuum - it was enacted and in process during an administration that has lashed out at various sectors of the economy, and much uncertainty has been created by all of the actions of the previous two administrations, so it is impossible to measure this and even if we could measure it, we'd have to attribute portions of that uncertainty to different things; Bush's stimulus, Obama's, TARP, auto bailouts, health-care and cao and trade and all of the anti-business rhetoric, the economic downturn itself, global macro conditions, etc.

It's interesting as a academic exercise, but people who work a few models or spreadsheets or regressions and come up with some absurdly specific multiplier effect, like 1.3567 or something, are cartoon characters who should not be taken seriously.

While I am not on the Keynesian side of this debate, I don't think this is a valid argument. What you are essentially saying is that, if its true that this stimulus spending had a 1.5 multiplier, then why not constantly stimulate the economy all the time? If that isn't a valid strategy, then the 1.5 multiplier cannot be real.

As I understand it, the Krugman side of this argument is that stimulus spending is effective, *when used during a deep recession and over short periods*. I don't think they are claiming that stimulus is a good thing to do all the time during normal circumstances.

In other words, just because calcium is good for my bones, doesn't mean that it would then follow that I should consume huge amounts of calcium all day long to have really strong bones. Too much calcium can actually hurt me, so taking a supplement only makes sense when I'm not getting enough calcium in my regular diet.

As a side note, I just want to say that I really appreciate the comments that danielkuehn adds to these discussions. While I am biased toward libertarian slanted economics, it is important to have people like danielkuehn on here who help keep us honest and point out when our biases are getting the best of us.

I think it is also important to make a distinction between the policies that the neoKeynsian economists advocate, and those that the left-leaning politicians advocate when they misunderstand the Keynesian economists. Given enough evidence, I could be convinced that stimulus spending can actually help during a recession if done responsibly by government. But I think even if Keynesians are right about that, the theory could never really be put in to practice because left-leaning politicians can't resist the urge to go crazy with it and over-do it.

In other words, I think there is a greater chance of the neoKeynesians being right in theory than government ever being responsible enough to put that theory into practice.

Even so, the krugman-keynesiac claim of economic stimulation during times of recession *presumes* that government spending during those times is better for the economy than the lack of spending that would occur otherwise.

No, actually they have a very scientific theory, not a presumption. They theorize that the ghosts of dead animals, apparently dumb suicidal animals--lemmings I suppose--come to possess the minds of investors during a recession.

You see, while most of mankind enters a state of heightened awareness during times of stress, with the usual self-interested behavior, investors are deprived of that aspect of human nature by these invaders from the animal spirit world.

So while an unscientific boob like you or me might think that a recession is a time when productive investments are hard to find, therefore requiring the careful scrutiny by the self-interested investor so as not to squander precious capital, what is really happening is that EVERY IMAGINABLE investment from sinking new battleships to digging and refilling holes to breaking windows has become suddenly productive.

Fortunately, though while the ghosts of lemmings prevent investors from reaping the rewards of all these abundant easy-picking investments, the lemmings do NOT prevent the possessed bodies of investors from loaning their money to the US Treasury.

That means the talented Mr. Reid can use those investment funds to invest in those profitable endeavors with a productive return not only sufficient to pay back the unexorcised investors, but also lift the economy out of its irrational depressed state.

The real problem, is that we have to contend with right-wing extremists who imagine some mystical concept of "squandering capital". As though it is even possible for spending to be wasteful. Those are the same right-wing nutjobs who invented the notion of "negative ROI", as though that even makes any sense.

Somehow I don't think other folks commenting like danielkuehn would describe their view that way, or as quite that simplistic. You have a lot of insightful comments on here vikingvista, but I think this one is a little more toward the straw-man side.

If I were a neo-Keynesian, which I am not, I might argue the theory by accepting the efficiency of a market. As I understand their side of the story, they don't think everything is actually determined by "animal spirits", but rather that uncertainty about the future simply has consequences on investor behavior. That is very different than saying all investor behavior is determined by nothing other than gut feeling.

For example, we know that incentives play a huge role in the economic behavior of individuals, but we also know that incentives aren't the only thing at work. Psychological studies have shown people to act irrationally on the surface in many different game theory type scenarios where it is clear that incentives aren't the explanation for all human behavior. A good "free market" guy knows this, and doesn't try and argue something as simple that we are all purely driven by incentives and that explains all human behavior. Likewise, it is foolish to accuse the Krugmans of the world of thinking that animal spirits are the only thing at play when it comes to investors.

Anyway, if I were a neo-Keynesian, I would probably something like, "government spending is ineffecient, which is why we don't advocate for it at all times. The reason why it works in short term crises, is because it quickly gets money into the hands of the people in the free market. And once they have money to spend or invest or whatever, the amazing efficiency of the private sector will more than make up for the short term inefficiency of government stimulus spending."

I think what they are banking on, is that all they need to do is get some money into the hands of the next Sam Walton during a depression, and just by giving that money out, the next Sam Walton will take that money and use it in a way to help rebuild all of the lost wealth of handing that money out in the first place. In a sense, I think they may actually be trusting in the free market more than they care to admit. On one hand, they have a very socialist attitude of redistributing wealth, but on the other hand, they seem to have a lot of faith in the free markets ability to produce people who will take that wealth that was distributed to them and make the economy grow.

I could be completely wrong about this. I'm still very much a student in these areas, but this is how I understand their side of the story so far.

So, while I don't really agree with that line of thinking, I think it is a little more rational than we try and make it out to be in these discussions sometimes.

"I think what they are banking on, is that all they need to do is get some money into the hands of the next Sam Walton during a depression"

It is the same dollar, just different spenders.

In one case the dollar is in the hands of a private investor who may be a functioning paranoid schizophrenic, or experiencing some other mental diversion or random event, but ON AVERAGE, across all investors in the economy, will act as humans act ON AVERAGE--in his own self-interest.

In the other case, that very same dollar is taken from the hands of the investor, and put into the hands of a government bureaucrat who is ALSO a human, with all of the foibles and unexpected behavior patterns that other humans, like investors, are subject to. But, ON AVERAGE, the bureaucrat will behave as humans ON AVERAGE behave--in his own self-interest.

In this mental experiment, we have controlled for the fact that both participants are human and subject to all kinds of potentially self-destructive or even inexplicable behaviors that individual humans have been seen to express. So that has been removed from consideration.

So then what you are saying, is that the dollar spent by the human bureaucrat, is on average more likely to find the next Sam Walton, then the very same dollar spent by the experienced professional investor--who by the way, is not so crazy or self-destructive as to miss out on the Treasury-buying opportunity.

My question then is, what is it about the incentives of the bureaucrat, than makes bureaucrats better ON AVERAGE at finding Sam Walton II, than the incentives of the investor?

I completely agree with you here. And I don't really have any kind of an answer for you, nor have I been convinced by the Keynesians when they try and answer this. When all is said and done, I think they inevitably keep relying on the "philosopher kings" to rule the world. Just that in this case, Plato's philosopher kings are replaced with economic kings, who lead the Federal Reserve and ultimately know best about where money should go because they are really really smart.

I have yet to see any Keynesian perspective that could not ultimately be reduced to this idea.

Brilliantly written Viking. You've poignantly outlined some of the difficulty I encountered while learning “dead theory” in Intermediate Macro and Monetary Policy as an undergrad.

Right, Keynesians (be the neo, post, or just plain old) generally do say multipliers are larger during recessions, but the logic they put forth to explain why that may be doesn't follow (to me at least it makes no sense). In my view, fiscal (as well as monetary) multipliers should be far higher in times of normalcy, even considering things like crowding out. To demonstrate why I think this is the proper way to think of multipliers, consider the following thought experiment:

Would additional spending be more effective in times where, though production capacity is near it's limits, the will to produce and demand for production is high or would spending be effective in times were there is surplus production capacity but the will to produce and demand for production is low? Seems to me it is far easier to increase production capacity (growth) in times of normalcy, than it is to encourage entrepreneurism and consumerism in bad times. That isn't to say I disagree with safety nets, but I do not buy the high fiscal multiplier story.

So, depending on the audience, his argument makes a lot of sense.

That does make sense what you say. What it comes down to is the fact that not only is not all spending the same in the keynesiac multiplier sense, but spending really can be counterproductive--not just wasteful, but actively destructive. And in a recession, rational investment declines not because of "animal spirits", but because counterproductive spending (usually left over from an asset bubble) prevails and the wise investor knows to stay away from it. But government, on the other hand, indiscriminantly dumps MORE capital into to it.

It is not just diminishing, but ignoring, the fact of counterproductive spending that is one of the distinguishing and intellectually damning traits of the keynesiacs.

...they provide intellectual cover to power-hungry politicians...

"Keynes did not add any new idea to the body of inflationist fallacies, a thousand times refuted by economists… He merely knew how to cloak the plea for inflation and credit expansion in the sophisticated terminology of mathematical economics." -- Mises

By the way, does anyone know where to find this model, I'd have much more confidence if I could play with it and figure out its mathematical assumptions.

"I would start by saying that I’m not sure we can gauge the effects of the stimulus."

And there is far too little public debate about what economics thought predicts that stimulus effects would be. Apparently analyzing hopelessly confounded, incomplete, and inaccurate data using historically wrong models is preferable than reasoned economic analysis.

This is the new age of scientism. Reason has been supplanted.

Or, as High Priest Krugman, et al say:

Look around at this world we've made
Equality our stock in trade
Come and join the Brotherhood of Man
Oh, what a nice, contented world
Let the banners be unfurled
Hold the Red Star proudly high in hand

We are the Priests of the Temples of Syrinx
Our great computers fill the hallowed halls (and tell us of Global Warming! And Economics!)
We are the Priests, of the Temples of Syrinx
All the gifts of life are held within our walls

[prescient album, 1976 libertarians I might add!]

Classic!

CBO is forced to make projections on absurd assumptions like 8% GDP growth (OMB, Peter Orszag). Asinine.

They’re also forced to toe the line to whatever Congress gives them, which is usually budgetary crap (CBO, Peter Orszag)

And guess who is most responsible for prostituting CBOs role in the largest entitlement expansion in the history of budgetary economics?

That’s right, Petey Orszag. Genius.

CBO to OMB: A Tangled Tale

I would start by saying that I’m not sure we can gauge the effects of the stimulus.

Viagra sales? Perhaps?

Do you people (the ones in the Comments section - all of you!) ever pause to reflect on the absolute GIBBERISH you espouse? Do any of you have real jobs that matter in the real world? Do all of you practice some sort of "think tank" profession? Do you have any idea how stupid you sound? Multipliers and statistics and mathematics and faith in estimates are somehow supposed to represent real life individuals. So all the "public policy" makers, the movers and shakers of the political elite can point to your estimates while they CONTINUOUSLY mess with people's lives. Sorry for the rant, but sometimes I wonder if any of you do anything meaningful in this world.

And you come here to read the comments section because....?

Because in spite of it all I like you guys.

Nongovernmental professional independent contractor. There are stretches of time when I only have enough work to tie me up for a measly 8 hours per day, so I read and learn and post during that "time off".

"the political elite can point to your estimates while they CONTINUOUSLY mess with people's lives."

Explain how you think keeping ourselves uninterested and in the dark would address this problem.

First, vista, let me apologize for my rant. Yesterday I had to lay off a man who had been with me for 8 years. There were tears and apologies and heartfelt handshakes all around. He is a good man, has a wife and three kids, and it was the most gut-wrenching experience in my life.

So I get home and find a bunch of yahoos (remember my mindset) nitpicking with each other over how fast a cheetah can run, whether the stimulus was working, whether the CBO submits estimates or whether they're really projections of numbers handed to them by the politicians, whether the multipliers are valid or at what point do the multipliers become invalid, admissions that the whole thing is a crap shoot, and on and on about statistics and numbers. And it became apparent to me that none of those yahoos had any realization that those numbers represented people, real, live people.

One guy even said he wished he could get the "numbers so I could play with them." And I was screaming at the computer that THEY'RE PEOPLE, YOU IDIOT!!! I couldn't help but wonder if there had been some economist/policy wonk/bureaucrat type that worked for the Nazis who ran numbers on how best to exterminate the Jews. (Remember my mindset.)

You see, I think all this talk about graphs and slides and multipliers and theories is a giant sham. The argument shouldn't be about whether the stimulus is a success. Who cares if it's a success? It's an implementation of evil, success or not. I think when you get caught up in "arguing the numbers" you are ceding a moral validity to an opponent that has no moral validity. You've compromised. And the instant you compromise with him, he's won. It's a moral argument, not a mathematical one. And as long as the collectivists have you arguing over the numbers you'll never win.

Again, sorry about the rant.

Sorry about your employee.

And you are right. The moral argument makes everything else irrelevant, and maybe even a distraction. And even if you revealed to the keynesiacs that they are wrong even on their own terms (and they are), they would just switch to some other excuse to seize control of our lives, and destroy our futures.

It's like the global warming debate--what difference does it make what the scientific merits are? Nothing justifies a new world fascist order. Are the antistatists going to hold a victory party if the whole climate change program turned out to be a total fraud? There will just be some new nonsense tomorrow with the same end in mind. Those activists never care about the validity of their claims, only their persuasive power. And they happily jump from one proxy cause to another.

But people are persuaded that it is necessary to accept a little more state violence, give up a few more freedoms, because the world's best & brightest have shown scientifically and mathematically that they must. Some of those folks are deprogrammable.

Raise children, design electronics, and write embedded programs.

IMO the multiplier might well be less than zero.

Rounding figures we've got a 10% of GDP public spending stimulus for a 2% GDP growth return in its year of operation. Then there's the diverted investment that debt servicing will cost and the hellish risks that come from printing money. The range of multiplier estimates may as well be -0.4 to +1.56.

Yeah. It might.

http://www.imf.org/external/pubs/ft/wp/2002/wp0...

Canned response to that goes something like this, "Lol, are you seriously trotting out a biased study from the extreme right wing think tank that is the IMF?"

Let's Counterfeit Our Way to Wealth

Team Obama claims that every dollar in stimulus creates $1.50 in wealth. This is economic fiction. The costs of the stimulus reduce future growth. As we see, the so-called stimulus isn't very stimulating, because it has gone or will go to government consumption, and the threat of future taxes puts people out of work now.

The 1.5 wealth multiplier is ludicrous. It is the Keynesian myth of distributing money to promote a recovery. As a counterexample, if it were true, then the government could license counterfeiting and we would all become rich. Actually, the government attitude toward printing money is very close to counterfeiting.

Consider that the government has been spending money wildly for 20 years. If government spending produced lasting prosperity, then we should all be rich already. The housing boom, financed by government borrowing and guarantees, was a period of substantial, increased spending through housing loans. Where is our lasting prosperity now?

How much far can we trust CBO estimates? Meaning how much do a priori biases, that stimulus MUST do something, play a part in the estimates?

It seems to me like we are trusting the Fox to watch the Hen house on this one. The CBO is government, and government has a HUGE incentive for ever increasing stimulus and government intervention. When the Fox keeps telling us that the hens are running away and that we need to keep a steady supply of hens going in to the hen house, how long before we start to question what is REALLY going on?
I mean look how excited Pelosi is to see what's in her HC bill! We have to pass it before we can see whats in it!

Can anyone tell me if the CBO predicted, or projected the financial crisis we just experienced?

I am sure the Stimulus saved and added some jobs. I highly doubt it is the 2 million that Obama claims. Heck I cannot identify but a small handful of stimulus jobs here in Maryland. Even state/local layoffs that may have been prevented barely number 1,000.

Forgot to add, a recent required regulatory mailing, which none of our customers will ever read nor understand much of the 12 pages, stopped us from adding 3 full time employees.

Does the CBO score all the jobs lost because of regulation?

The unseen goes unscored. Otherwise the stimulus recommendations would be for an emergency trillion dollar CUT in government spending.

There are no economists employed by the government, because economics does not favor the state.

Calculation in the fiat economy produces fiat results.

GIGO. Especially when what it passes through is also G.

What exactly does this mean? This is the same point you mentioned above.

A multiplier is a net impact estimate. It nets out all costs and all benefits, seen and unseen. It's not like they figure out what every good impact is and subtract what every bad impact is and if they forget one then the estimate is biased. It is a net impact estimate. It incorporates all of this. The question is - how well does the estimate identify the effect, and how much error is in the estimate?

Guess would be more like it; the term estimate dignifies it far too much.

And, I would add, how long you should wait to measure the effect. If you are overstimulating the economy, the effects may not show up for quite some time in the form of CPI inflation, a second bubble and/or a government budget crisis.

Since before they can spend the money, the government must borrow the funds out of the hands of people who are much better at finding productive investments, the negative economic impact occurs even BEFORE the first government dollar is spent.

The GDP may transiently rise (relative to if they had not borrowed the money), but that only reflects government spending on things the expert private investors judged to be unproductive. It's a GDP boost measuring how quickly the government blew the economy's wad. This compared to the wiser investors who would've moved more cautiously so they could nab truly productive endeavors as they arose.

Truly government spending has NOTHING good to offer to economic recovery, but rather impedes it from day zero.

56% return on investment... mmmm... I will go for that! Just let government make ALL of our future investments! Can you imagine how wealthy we all be in a few short years?!

I find this comment: "that outcome reflects greater-than-projected weakness in the underlying economy rather than lower-than-expected effects of ARRA." particularly interesting.

If they can't estimate the performance of the underlying economy, how can they differentiate the effects of the stimulus?

Seems to me what they are doing is announcing the results of their calculations and nothing more.

This hits a little too close to home:
http://www.theonion.com/content/infograph/one_y...

There's also a time factor (the long run). The full data on the effects of extreme borrowing aren't going to be available for decades (we hope). At the time that Social Security was implemented, who knew with any degree of accuracy what its long term costs would be. The short term positive effects of borrowing and spending a few trillion dollars can be viewed as a multiplier, but the long term result is likely to be very, very, negative.

The only way the short-term effects can be deemed positive, is by wearing blinders that hide everything in the economy except the short-term effects on the immediate recipients of the government funds.

Let's dash the myth of short-term stimulus spending benefits. It is exceedingly destructive.

The real multiplier should be a function of what the stimulus was spent on. If you spend government money on more bureaucrats that are inhibiting business development and growth (ie. more planners allowing more stakeholders to have veto power over more projects) you should get a lot smaller multiplier that you would building a new world trade tower. The damage that can be caused by spending on parasitic activities is real and only the lawyers seem to benefit.

In Calif. we can't build large solar, nuclear power or wind plants or build power lines to transmit the power. Remember, this country build the empire state building in one year, but now, we have spent 9 years and 7 billion dollars and still have a hole in the ground where the world trade centers stood. One activity wastes wealth, another creates value and they have drastically different multipliers.

Even building a new world trade tower can be destructive, if nobody wants it. Any spending that you can think of can be destructive of achieving economic growth. Determining what people want and will want is not an easy task.

Investors who spend their lives and put up their fortunes looking for just some small part of that story, even in a good economy, frequently fail. So how much more likely is it that dollars borrowed away from those investors and instead spent by Pelosi/Reid/Obama will find productive ends? And since P/R/O have to pay an interest premium to those investors, AND recoup the principle from taxes, the P/R/O investments have to be that much BETTER than what professional self-risking investors could find.

Logically, government spending multipliers of course have to be negative on net. If estimations don't come up negative, then they just haven't included all the direct or opportunity costs.

The CBO might as well measure the effects of the stimulus using a ruler.

It wouldn't surprise me if an accurate estimation, including all the costs that keynesiacs are careful to ignore, would not just return a multiplier <1, but <0.

Something like taxation for the waging of war is a classic example. The multiplier is almost certainly negative. You could also lump in putting people in prison for drug crimes or tax evasion as net negative multipliers.

For less insidious projects I don't see how it can be negative. Taking money and building a bridge might be a waste compared to what would have been built, but it provides some benefit, so the multiplier is positive.

"Taking money and building a bridge might be a waste compared to what would have been built, but it provides some benefit, so the multiplier is positive."


Actually, if the bridge spending is sufficiently less beneficial (or has the same multiplier) than how the money otherwise would've been spent, then the multiplier IS negative. From the keynesian view, the multiplier is positive because the money would NOT otherwise have been spent. At so-called "full employment", where all capital is being maximally used, government spending at best is just a transfer from some other potential expenditure, resulting in zero net change, so m=0 (assuming the same multiplier).

So you are already imagining a negative multiplier.

But to be fair to the keynesians, they are supposing that during a recession the government is spending money that would not otherwise be spent. That brings new money into the expenditure accounting, with a corresponding boost in income, and therefore positive multiplier.

I would argue that even that probably isn't true. That is, they like to assume m>1, but it isn't hard to imagine m<0. This is because:

1. It actually DOES decrease other expenditures.

The signals government spending sends to investors and consumers are bad, not good. It complicates rather than eases the quest for sustainable profitable investments, because of the unsustainable politically-motivated shifts in resources and price structures that it causes. Investors need to know what consumers want and will want, not what politicians want during the time of the stimulus. This is the investor uncertainty argument you hear from people skeptical of keynesian policies.

The analytic multipliers would be wrong, because they too optimistically model the interactions between the different expenditure aggregates.


2. Spending is not, by definition, better than nonspending.

The sustainable economy, the one you and I want rather than the one Washington wants, must emerge. That means productive investment opportunities arise over time. If investors' funds are all lured away by more attractive government bonds, then not only will those investors lose incentive to find those emerging opportunities, but they will not have the capital on hand even if they do see them. An ACTUAL stimulus would improve the ability of expert self-interested investors to find these productive opportunities. It is rediculous to think that government has either the expertise, or the incentives to do that as well (or at all). Therefore, government "stimulus" actually diverts funding from stimulus, and muddies the waters for real investors.

The multipliers would be wrong because the models seek the quickest boost in GDP, without being able to model the effect of the myriad different kinds of spending which are different for every recession, not all immediately available, and not even knowable ex ante.


3. Spending is not, by definition, economic growth.

Long-term GDP is a reasonable correlate of economic growth, but only because you need economic growth to be able to have sustained increases in spending. GDP is only a measure of spending--of pushing money around--without regard for whether that money pushing is productive or destructive. If I take $10 out of your wallet at gunpoint, and hand you something you don't want in return, I've just increased the GDP by $10 (assuming you don't report me to the police). Even destructive spending can boost GDP (though not indefinitely). In other words, you can have a GDP boost which is merely a harbinger of our future destruction. As a hyperbolic example along the continuum, imagine a huge increase in spending on entertaining gasoline powered crop circle machines. Big GDP spike, then no food, no fuel, no GDP.

Here it isn't the multipliers that are wrong. The mistake is using GDP as a proxy for economic growth. That works when everybody is making only productive (i.e. voluntary) transactions. But government transactions are a completely different animal, and change the very nature of GDP into a different variable. That is, the act of manipulating GDP with fiscal policy changes it into something else.



So, excluding the transient money pushing effects on GDP, you should expect m<0, because government

1. squanders the capital needed to do what a government is incapable of doing--investing in real economic growth; and

2. actually invests that squandered capital into economically destructive ends (not the least of which is the political economy).

I think you are confusing the Keynsian definition of the multiplier. Generally it's assumed to be economic activity created by spending divided by economic activity lost due to taxation/inflation/etc.. It can very well be "positive," but if it's less than 1, we are looking at a net drag on the economy.

Once again, I agree with your sentiments, the government taking money and spending it overall leads to net lower, and even negative growth, but the multiplier is not less than zero unless the government is using money to actively destroy other property.

Thoroughly outlining these definitions I think we can empirically show a "multiplier" much less than one, and thus eliminate the case for more stimulus. Saying dogmatically that there can be no help to the economy and that the multiplier is always less than zero convinces no one because it can not possibly be true. Even if a bureaucrat taxes 100 dollars, waste 99 on useless activity, an spends 1 in a way similar to the market would have, it has a "positive" multiplier, while being a net "negative" on the economy.

Is it your opinion that keynesian economists think negative multipliers are "impossible", even at full employment, or just that the algebraic expressions derived from the usual textbook models are never negative?



I do find it hard to believe the multipliers are positive. But let me explain in a different way...


Assume aggregate expenditures and income are equal, and conceptually partitioned so that

Y = C + I + G

We want to know how Y changes with G, or deltaY/deltaG, which is the multiplier, m. We can conceptually partition deltaY however we want, so imagine it were possible to separate it into a portion due to deltaG and a portion due to deltaI. Assume for simplicity that the multipliers are equal, then

deltaY = deltaYa + deltaYb

deltaYa = m*deltaG
deltaYb = m*deltaI

Now if deltaG is entirely at the expense of deltaI--a mere transfer without additional economic costs--then deltaG = -deltaI, and

deltaY = deltaYa + deltaYb
= deltaYa - deltaYa
= 0

So, in terms of total GDP,

deltaY = 0 * deltaG

The gist of it is, if increased government spending detracts equally from other expenditures, then GDP is unchanged, so the multiplier is zero.

In the usual Keynesian scenario, other expenditures are not modeled with that kind of dependence on G, at least not during a recession. Therefore the algebraic expression for the multiplier, built from those models, results in a value always greater than 0. But models are for prediction, and posthoc empirical estimations of multipliers are usually lower--and I argue still not nearly low enough.

Because government spending is almost guaranteed to be less productive than private spending or private **non-spending**, a multiplier of zero is a best case scenario, and an unlikely one at that.

The positive multiplier presumption only works if you think government spending is better than no spending at all. But if that were true, the spending would happen directly from investors without the intermediate step of government first borrowing that money from the investors.

This is where the "animal spirits" comes in, but that does not hold up to scrutiny. You can't base a theory on large economic movements on behavioral anomalies--especially when we're talking about the same species of mammal spending the same money, whether bureaucrats or investors. So the difference in spending behavior really comes down not to "animal spirits", but to the different incentives faced by bureaucrats vs investors. And clearly--CLEARLY--government spending cannot be justified on those grounds. And at any rate the "animal spirits" assumption is utter nonsense. The investor behavior during a recession is quite rational, and well-suited to restoring economic growth.



In my last post to you I said "excluding the transient money pushing effects on GDP". Let me further explain that in terms of the deltas. It also explains why I think empirical estimates must be overestimates.

Notice that any delta requires a time frame. In this case, I'm assuming a time frame sufficient to capture the full effects of deltaG on deltaI (I being just an example, as I think it may affect C as well). The money that the government spends rapidly (and recklessly), is the very same money that private investors would've spent cautiously over a longer period of time. So all of deltaG is completed in the beginning of the longer interval of deltaI. If your interval for post hoc estimating the multiplier is shorter than deltaI, you are excluding a part of the negative effect, and the multiplier will appear larger than it should.

You've essentially assigned your own definition for the "multiplier." You won't convince anyone in the other camp of your view because you redefine terms they made up.

You are confusing 1 for zero in your analysis. If increased government spending is the same as reduced market expenditures the multiplier is 1. This is essentially a worse case scenario for Keynsians. The actual multiplier is much less.

NathanS,

It sounds like you might have some formal training in this area that I lack. I would be grateful, if you pinpoint my error. Let me summarize to save you time:

deltaY/deltaG *IS* the government spending multiplier, by definition.

If deltaY<0 for deltaG>0, then the multiplier is negative.

The algebraic expressions for this multiplier that are derived from keynesian models, do not (at least typically) allow for negative multiplier values. The reason is because those models do not allow for interactions between G and the other expenditure components that could result in a negative multiplier (typically the multiplier>1 effect is a result of the keynesian consumption function model, while the other components, G and I, are modeled simply as autonomous).

It is imaginable, and has been observed, that GDP can decline (deltaY<0).

It is imaginable, and has been observed, that governments can spend money (deltaG>0).

It is imaginable, and has been observed, that both can happen simultaneously (deltaY/deltaG < 0).

It is therefore imaginable that the government spending multiplier can be negative (whether or not typical keynesian models accommodate that possibility).

Where am I wrong?

Great comment! Why no DK response?

One of the better comments I've read on here in the three plus years I've been reading.

What do you mean by "all the costs that kenyesiacs are careful to ignore"? How could a "keynesiac" exclude those costs from the estimates even if he wanted to?

Not properly identify the effect? That's possible. But when they run the models it's not like they pick and choose what costs to include and what benefits to include. These are net effects - net of all possible costs and benefits. There may be identification concerns, but I don't understand how there could be concerns about selective omission of different costs. Could you explain what you're thinking of?

Also - read the CBO publications. Steve Horwitz raised this same point recently and I directed him towards the CBO report that produced this estimate. They spent an entire page of it talking about why stimulus might not work and they hit on every single issue that Steve raised with me in our correspondence, and many that Steve didn't raise. People that work at the CBO aren't dumb. They're as insightful as anyone is on here. None of this is news to them.

It would be helpful if economists, politicians, and agencies like the CBO had official "stats" similar to athletes, so we'd know if the person actually knows how to hit the ball, or just looks good (i.e. talks a good game). Consumer Reports could verify that so-and-so has a 90% percent accuracy within a reasonable SD. A "Good Buy".

Officially record numerous basic economic forecasts. What will the unemployment rate be in 6 months? 12 months?

Just basic stuff...become an economics star based on results, not appearances (See Moneyball.)

It would be helpful, but I don't see how that can be accomplished. As far as I can tell, the CBO does not produce estimates which can later be verified. They do not predict what the unemployment rate will be in 6 months or in 12 months. Rather, the project that a specific piece of legislation will cause x change in number of jobs - not from current levels of jobs but from the level that would exist in 6 months without the legislation. As I see it, there is no way to hold them accountable.

"Given that the range of multipliers in the profession these days is from .4 to 1.5 (at least), it is clear that there is no consensus as to how to predict the impact of the stimulus."

I don't think this is clear at all. The specification that produces the low estimates and the specification that produces the high estimates are very different. You can't lump them together any more than you can say that biologists have no way of knowing how fast a cheetah can run because sometimes they are measured to be very slow and sometimes they are measured to be very fast.

There are huge identification problems. I agree with you that we'll never have a strongly conclusive estimate for the magnitude of the multiplier. It will always be a rough estimate. But I don't think it's right to just point to the range of multipliers and throw up your hands.

I think you are exactly on target with your assessment of the so-called "estimate" produced by the CBO. It is an estimate from the past, but it is just a projection of what's going on now. I think a lot of the confusion over this is a result of how it's reported in the media. If you read the CBO's Economic Outlooks it's clear as day. But it's usually "clear as day" a couple pages into the report, and if you know something about what they do. That doesn't always translate well - so it's good that you highlight this the way you do.

"I don't think this is clear at all. The specification that produces the low estimates and the specification that produces the high estimates are very different. You can't lump them together any more than you can say that biologists have no way of knowing how fast a cheetah can run because sometimes they are measured to be very slow and sometimes they are measured to be very fast."

Actually, that's exactly what biologists would have to do. If cheetahs are sometimes measured to be very slow and sometimes measured to be very fast, then you have no way of knowing how fast cheetahs are in general. If fact, if cheetahs behaved in such a way, the question, "how fast can a cheetah run?" wouldn't have a simple answer; at best you could reply, "some cheetahs are fast and some cheetahs are slow." If you had no way of telling fast and slow cheetahs apart before they started running and someone asked you, "how fast is that cheetah," you would have to throw up your hands and say I don't know. Just giving someone a rough estimate of how fast you think the cheetah could run would be irresponsible.

If the multipliers range from 0.4 to 1.5 because economists can't agree on what factors should be included or to what extent, then the whole multiplier concept is useless. On the other hand, if the stimulus multiplier is 0.4 under condition A and 1.5 under condition B (or somewhere in between under condition C) then you should be able to make a convincing case as to what conditions apply and consequently what the multiplier should be.

If I understand Russ's point, the CBO and other pro-stimulus economists are arguing that the multiplier is always 1.5 (or at least >1) without bothering to see if the economy is under condition A or B. Furthermore, because we CAN'T every truly know whether we are in state A or B, saying we know what the multiplier is makes no sense. As Russ has pointed out before, people on both sides can point to some piece of data to support their own viewpoint, but system as a whole is too complex for any of us to truly know effective stimulus spending is.

I really don't understand your cheetah paragraph - could you rephrase it?

My point is simply that we get a range of observations on cheetah speed too. The fact that we have a wide range doesn't suggest we're bad at measuring cheetah speed. What it suggests is that in some situations cheetahs run fast, in some situations they run slow, and certain individual cheetahs have traits that make them faster or slower than others. But it doesn't mean we're bad at measuring cheetah speed. Our measurements are simply (1.) conditional, and (2.) liable to a certain degree of sampling error.


"On the other hand, if the stimulus multiplier is 0.4 under condition A and 1.5 under condition B (or somewhere in between under condition C) then you should be able to make a convincing case as to what conditions apply and consequently what the multiplier should be."

Yes, precisely.

"If I understand Russ's point, the CBO and other pro-stimulus economists are arguing that the multiplier is always 1.5 (or at least >1) without bothering to see if the economy is under condition A or B."

If he is arguing that then (1.) I missed it, and (2.) I think he's wrong. After all - just read any garden variety stimulus advocate's response to Barro's WSJ articles. They talk at great length about the conditionality. The ones I haven't heard talk much about conditionality are the opponents, but that may be the bias that I bring to what they write (not Russ here - I think he probably pleads ignorance more than he needs to, but I don't think he ignores the conditions involved).

"My point is simply that we get a range of observations on cheetah speed too"

The exact speed of the cheetah is not as important as whether it can overtake the antelope. If the biologist advances a theory that a class of predators are successful because they can overtake prey in short sprints, and that the cheetah belongs to this category, and subsequently found that cheetahs do not generally sprint faster than antelopes, I would expect the biologist to revise the theory.

Likewise, when periods of Keynesian stimuli are characterized not by wealth and prosperity, but prolonged, deep recessions, I would expect a candid Keynesian to revisit the theory. Instead, what we hear is "it would have been worse without the stimulus."

To be fair to Keynesians, I am sure they work the theories and the models as new data come to light, but publicly they reiterate the "stimulus is good" mantra.

Is it so hard to understand why we are not impressed?

That sort of situation wouldn't impress me too. I suppose it's your misdiagnosis of the situation that's hard for me to understand, not the underlying logic you provide here. I think we should revise what we think based on observed evidence too.

"Our measurements are simply (1.) conditional, and (2.) liable to a certain degree of sampling error."

Exactly. And I would argue that we are (1) bad at figuring out what conditions are are important, (2) how much weight to give to the conditions we think are important and (3) the sampling error is prohibitively large. My guess is that you disagree. Russ seems to be arguing in his past few CBO posts that he disagrees with not just their numbers, but how they come up with their numbers. I happen to agree with him. I'm not sure that a model sophisticated enough to capture the relevant aspects of the stimulus and the economy wouldn't itself be too complicated to understand.

Simple models are easy to understand, but small errors can lead to huge problems. Adding detail minimizes the errors, but increases complexityl and obscures understanding. I don't know that we can add enough detail to make the errors acceptably small without losing all ability to understand what we've modeled.

You're just capturing a net impact - you don't need to model all the complications of the economy. I would agree we aren't up to that task - that's why the farthest I go in terms of planning is macro stabilization. We can't plumb the depths of the complications of the economy. But we can do a much better job at getting a net impact. The biggest obstacle is the identification problems, not the complexity.

I'm a little short on insight as to why you or he thinks the models are bad. The only point seems to be "it's too complicated". What's so complicated about dY/dG? If I knew exactly what it was that you or Russ objected to, perhaps I would understand better. But the biggest problem I see is a sample size problem and an identification problem. And I fully submit that those are big issues to deal with.

To continue your cheetah analogy, isn't the question here not how fast the cheetah "can" run but how fast it "will" run? If the cheetah has been observed running at many different speeds, how does one predict how fast it will run the next time?

You can stick a cattle prod up its rear as stimulus and there are still too many things you cannot know. And good luck with that cattle prod and unintended consequences.

How could we get an estimate of the effect of cutting government spending?

They don't want to go there.

Ya - I'd definitely agree with that.

If I knew something about how fast a cheetah would run and why and when they run that fast, though, I'd probably be willing to at least hazard a guess at how fast they'd run if a gazelle walked by.

And really, what is the alternative? Claiming ignorance and doing nothing is no different than estimating a zero impact and acting on that. I don't have unparalleled faith in a 1.56 multiplier by any stretch of the imagination - but I have considerably more faith in that than in a 0 multiplier.

Hmmm...well I do agree it is a matter of faith. Since estimating multipliers isn't exactly accurate, how small of a multiplier will you go down before you lose your "faith?" What makes you so faithful of the estimate anyway? Are you saying that you just plain wouldn't believe a multiplier estimates of less than zero?
Most of the Multipliers in this paper are very very small, some 2 and 3 year multipliers are less than zero. Anything multiplier less than 1 is completely worthless and shouldn't even be tried....notice almost all of them are less than 1.
http://imf.org/external/pubs/ft/spn/2009/spn091...

What do you mean "measuring" and what do you mean "accurate"? I would say we don't measure multipliers - we estimate them. And for an estimate, "accuracy" isn't an on/off switch - it's a range of certainty.

I could certainly imagine the existence of multipliers that are less than zero. I doubt such a thing would be that likely in the U.S., but I don't think it's implausible. My faith in the estimate is dependent on how the estimate was derived. Although I probably shouldn't have used the word "faith".

Yes estimate is a better word, but there isn't exactly a good reference to see how valid our estimates are. It's more pixie dust economics.
That's the problem with econ, with Biology or Chemistry, we can actually test to see if your estimates are valid. With Econ, not so much, so we can't justifiable have any faith in the estimates at all, because nothing is ever ceteris paribus.

I have never, and probably will never, understand this perspective that because non-experimental empirical work is trickier it is somehow fake or suspect. Are astronomers "pixie dust" astronomers? Are meteorologists "pixie dust" meteorologists? Are evolutionary biologists "pixie dust" biologists? Not from my perspective. Yes, it's different. Yes we need to be careful and very cognizant of identification issues. But who cares if it's not experimental? How do you make your leap from my "conditional/qualified faith" to your "we can't have any justifiable faith in the estimates at all"?

Can you give me some examples where economists have accurately predicted the future outcome of anything. And by this I mean more than mere vague assertions. Despite all the math and modeling I have yet to see any difference between the predictive power of historians and that of economists. Economics seems to me largely a bunch of grandiose fumbling in the dark. Which leads me to conclude that if economists were even slightly less arrogant they might use what little they know to greater effect.

They do quite well predicting the future. CBO evaluates the accuracy of it's own forecasts - http://www.cbo.gov/ftpdocs/104xx/doc10484/07-30...

They do considerably better than meteorologists, for example. See, you say "economics seems to me largely a bunch of grandiose fumbling in the dark" - but have you even TRIED to assess the accuracy of economic forecasts? No - you have a predetermined notion of what they're like, so you just prognosticate about it. You have no clue how the accuracy of economic forecasts compare to the accuracy of meteorological forecasts. And you can't even compare economics to evolutionary biology because evolutionary biologists so rarely make forecasts - there's nothing to compare it to!!!! But again - that's not because meteorology is unscientific. Meteorologists can forecast only a couple days ahead. Economists can forecast several months ahead. It's the nature of the different systems they look at. But you certainly can't say that economists are less accurate than meteorologists. Astronomy is even easier. They worked out the basic behavior of astronomical bodies centuries ago because the behavior was so simple. Meteorology is harder. Biology is harder. Economics is harder. That doesn't mean any of those are unscientific.

You've just made one declaration of what it "seems to you" something is after another. Sit down and read the CBO report where they review the accuracy of their forecasts. That way, you don't have to rely on what something "seems like to you" - you can actually have facts to form an opinion on.

"They do quite well predicting the future. CBO evaluates the accuracy of it's own forecasts "
Ha ha ha, I hope you see the contradiction there.

No, I don't.

Do you think they can't master a proceedure for testing forecast accuracy that's taught in high school math classes? I mean, seriously - I know you guys like to point out the fallibility of government employees. And I agree - that needs to be regularly pointed out. But give it a rest, Justin. Where's the contradiction? What problems do you forsee?

How often do you trust self-interested organizations, when they tell you that they are on the up and up? Aurthur Anderson, Enron, WorldCom all come to mind. It's just another version of appeal to their own authority in my book. Just because they have procedures in place doesn't mean they follow them all the time, especially if a number is too far out from their estimates...outliers anyone?

Self interested organizations in the private sector have to deal with competitive forces, not staying on the up and up = failure.

Predicting the economic effects of socialist policies is extremely easy as they invariably follow the same pattern. Price controls? Shortages. Government takeover of economic sector? Lower quality, reduced services. Democracy vs Republic, people voting themselves money (in former), leads to enormous deficits and eventual devaluation, inflation or a reversion back to capitalism. Now we are seeing in western europe the start of generational warfare, the entitlement class has sucked so much money out that they have made wage slaves out of the new workers.

You don't need a formula to prove these results, they are so uniformly consistent that I'd easily wager on any of them.

BTW, it doesn't take any sort of radical epistemology to be highly skeptical of economics as a science.

I have to say, it is rather humorous that you are working with this rather Gouldian notion of what evolutionary biology is about. I know he is pretty famous in "Hens Teeth & Horses Toes" for arguing that evolutionary biology is a historical science not a predictive one, but that was something contested even when he was alive and has been I would argue overturned.

No, I'm not saying that. You're trying to insist that I'm saying evolutionary biology doesn't make predictions, but I've never said such a thing. What I've said is that all science does a mixture of explanation and prediction. Evolutionary does less prediction - that's not to say it's not predictive. This is your M.O., mommsen - you did it with our talk on slavery too. You're intimately aware of some great controversy that happened decades ago because you heard about it in some class, and you're chomping at the bit to show off that you know it, so you convince yourself that someone is committing the fallacy just so you have an opportunity to shout it from the rooftops and chuckle to yourself at how sophisticated you are and how ignorant everyone else is. I have to say, at least you add some variety to it. Usually on here the "gotcha" that gets repeated over and over again by twisting someone's argument is Bastiat's broken window fallacy. You add a lot of fresh ones to the list, but the story is the same - you twist what other people say so you have an opportunity to show off what you know.

Ahh, meteorologists, for example, do a far better job predicting the paths of hurricanes than economists do predicting the subjects that that they study.

"And you can't even compare economics to evolutionary biology because evolutionary biologists so rarely make forecasts..."

This is just flat our wrong. Evolutionary biologists make predictions all the time; indeed, that has been one of their most powerful tools in their toolbox over the past couple of decades (particularly when it comes to things like genetics). When was the last time you actually read anything in this field?

"Meteorologists can forecast only a couple days ahead. Economists can forecast several months ahead."

Yes, but meteorologists forecast far, far more accurately than economists do. There is nothing like the accuracy of hurricane prediction in the field of economics. If is please do show us such.

Biology and meteorology are more difficult disciplines than economics is.

"Yes, but meteorologists forecast far, far more accurately than economists do."

You keep SAYING this, but you don't PROVE anything you say. I provided you documentation of forecasting accuracy for economics. stop just repeating this over and over. Prove it. I'm curious what the root mean squared error of meteorological forecasts are. Generally speaking scientists will make forecasts about as far into the future as they can be accurate. The fact that economists make forecasts considerably farther into the future alone is highly suggestive. Stop just repeating this - you've repeated it several times now. Provide some evidence, mommsen. Everything is always about your impressions.

"If is please do show us such."

See the link.

"Biology and meteorology are more difficult disciplines than economics is."

What is your experience with any of these three?

Meteorologists actually have a fair amount of experimental empirical evidence to draw from. The same is true with astronomers and evolutionary biologists as well (indeed, evolutionary biologists have made a whole bunch of very accurate predictions of late*). Economics remains a very crude, infant science in comparison to these fields of study.

*See here: http://ncse.com/rncse/17/4/predictive-power-evo...

What experimental evidence do meteorologists have? Like physics/aerodynamics experiments? If you're calling that meteorology, economists have comparable experiments to draw on, including the burgeoning field of experimental economics - but also psychology. My whole point, though, is that you could do good meteorology without any experimental evidence. Good meteorology can be done with non-experimental empirics and theory (just like good astronomy - another one). Everyone has some experimental evidence available, but some sciences have less experimental evidence than others. There's nothing special or more mature about experimental evidence. It's just much more convenient for deriving conclusions than non-experimental evidence.

As for evolutionary biology - don't even try to match up the predictive power of evolutionary biology with the predictive power of economics. It's not even close. But again - that doesn't make evolutionary biology "non-scientific". You're priveleging "prediction" over "explanation" unnecessarily and inexplicably, just like you priveleged "experimental" over "non-experimental" unnecessarily and inexplicably. Why are you priveleging these things?

I'm not sure what you mean by "very crude". The systems that economists study are much more complex and have considerably more feedback loops than the systems that astronomers study. That doesn't make it "crude" - that makes the subject matter different. The same can be said of meteorology. An economist can predict phenomenon farther into the future and with more accuracy than a meteorologist, but it's not because meteorology is "crude" - it's because meteorology is studying a different kind of system that we should have different expectations about.

Stop demagoging science. Science is empirical and it is descriptive. The degree to which it is experimental/non-experimental or predictive/explanatory all depends on the subject of scientific study. These things say nothing about whether it is an "infant" science or whether it is "crude" or not.

Or, to be more succinct, we've learned of late that the Emperor Has No Clothes.

"What experimental evidence do meteorologists have?"

All sorts; it depends on numerous the sub-fields. Both in the field and in the lab meteorologists conduct all manner of experiments on manner as diverse as tornadoes, hurricanes, low and high pressure fronts, lightning, etc. For example, meteorologists create their own lightning.

"As for evolutionary biology - don't even try to match up the predictive power of evolutionary biology with the predictive power of economics. It's not even close."

Yes, economics has little predictive power, whereas evolutionary biology has a great deal of predictive power.

"Why are you priveleging [sic] these things?"

Because one is clearly superior to the other.

"Stop demagoging [sic] science."

Well, since economics is not a science, I am not quite sure how I would be doing that.

Great comments. I agree with you on just about everything. My whole point, which is lost to Dan, is that no matter what we have to be skeptical of any claim; whether it's from a physicist, chemist, biologist, sociologist or economist.
I think the modern partisan political environment has invaded academia, actually I think it all started in academia. As a result there is an overriding tendency for people to believe anything 100% so long as their scientists claim it, like AGW.
AGW is a great example of this, but I think the phenomenon is quite prevalent in economics as well. There is no skepticism for hypothesis or theories that conform to a priori biases. They are taken at face with nary a peep. Anything that doesn't conform to previously held biases, are not subject to skepticism, but out right rejection! Again look at AGW, and to a big extent Austrian Economics.
It's not hard to see the damaging effects this has to our knowledge over time.

No where did I say "we can't have any justifiable faith in the estimates at all".
My only point is that we can't accept it as 100% fact. We have to be skeptical of what anyone tells us, even if they have a nice PHd on the end of their name.

It's pixie dust because at some point it all hinges on "faith" like you said in another post. The empirical work isn't conclusive and always based on some assumptions that may or may not hold true all the time. How do you know in vivo, whether or not your in a situation where the assumption holds? You almost never do until ex post.
That's part of my point, knowledge is always incomplete, always.
Now that doesn't mean we should always do nothing and never bother to learn anything. Quite the contrary, we always should strive to improve our knowledge, but we also need to be aware of our limitations.
My problem with pixie dust Economists are that they never seem to acknowledge that they are even aware that they could be wrong. For all we know, they could be trying to prove a Heliocentric solar system!
We just don't know, so we should always be skeptical.
Basically take all your skepticism of anything an Austrian economist would say about something and hold all the rest of the economics profession to the same standard.

Science is always about skepticism.

"No where did I say "we can't have any justifiable faith in the estimates at all". "

Aside from some grammatical unmangling I quoted you verbatim, Justin.

"We have to be skeptical of what anyone tells us, even if they have a nice PHd on the end of their name."

I do that every day on here, don't I! I couldn't agree with you more.

Umm, no your trying to put words in my mouth, grammatical un-mangling or not. I did not say that at all.

Yes you do, do that everyday. We'll just agree to disagree. Because your never going to get me to agree with you on most Keynesian econ.

I just wish more people read Economic Consequences of Peace and less General Theory.

OK - I'll just copy and paste. You wrote: "we can't justifiable have any faith in the estimates at all"

I really don't think my slight modification changed the meaning of that. That was what you said.

Okay I found it. Well like you are always saying to the people here in relation to Krugman, what I really meant was....

I think my comment above explained in more details what I mean when I said that. The same one you said you agreed with.

Yes, "faith" is the key word.

It seems to me there are plenty of arguments to be made (and are often made here) that any stimulus bill has negative impact. Certainly from the standpoint of individuals wanting to be allowed to spend their own money (or future earnings) on things they individually value which only they can know, the impact is entirely negative.

You can't lump them together any more than you can say that biologists have no way of knowing how fast a cheetah can run because sometimes they are measured to be very slow and sometimes they are measured to be very fast

If you ask a Keynesian, "how fast will your cheetah deliver the message," the reply will be, "it will move at its top speed, in excess of 40 mph, the entire way, and it will make no detours."

Spot on Henri !!!

Hmmmm... ya, I'm not so sure about that.

But by all means - let's trade in caricatures. That's extremely productive.

Well, you started. I do not see a meaningful comparison between cheetahs and multipliers.