Brad DeLong says:
Megan McArdle writes:
All you have to do is believe . . . - Megan McArdle: The real question, I think, is how close the permanent income hypothesis is to being true. The basic idea is that people are forward looking, and they try to smooth their consumption over time. So if you give them a "temporary tax cut", they save most of it, knowing that eventually they will have to give the money back. But of course, this should also be true of "temporary government spending"--if people think the money won't be there next year, they'll salt as much of the money away as possible. This is a topic very underexplored in the various estimates of the stimulus multiplier...
No it isn't. This is a topic that economists have been exploring for fifty-five years. It is a topic that has been very thoroughly explored in all of the estimates of multipliers.
Indeed--and I see that I have been unclear. All I meant was that I would like to see more economists talking about how consumer savings affects their current estimates of the multiplier for Obama's stimulus plan. It seems to me that the rather extraordinary credit binge American consumers have been on for the last seven years has got to affect the knock-on effects of direct government spending--the marginal propensity to consume seems to be dropping rapidly as people focus more on their future income stream. Perhaps my intuition is wrong, but if so, I long to have better heads than mine explain why.
Professor DeLong's misunderstanding was my fault for not being clear--the penalty of a form that is written quickly. But I confess to being purely puzzled as to Professor Krugman's addendum:
Brad DeLong links to Megan McArdle saying something wrong about the effects of a temporary increase in government spending. But he fails to note that it's not just wrong, it's 180 degrees wrong: a temporary increase in government spending should have a larger impact on demand than a permanent increase, not a smaller impact.
And that's actually an important point: one way to explain why government spending is better than tax cuts as a stimulus is to say that temporary tax cuts aren't effective at increasing demand, but temporary spending increases are.
Here's the logic (which follows directly from Milton Friedman's permanent income hypothesis, by the way): suppose that the government introduces a new program that will cause it to spend $100 billion a year every year from now on. To pay for this, it will have to raise taxes by $100 billion a year, permanently -- and if consumers take this into account, they might well cut their spending enough to offset the increase in government purchases.
But suppose the government introduces a one-time, $100 billion program to repair bridges over the next year. The government will have to issue debt to pay for this, and will have to service that debt, requiring higher taxes -- say, $5 billion a year. That's a much smaller impact on consumers' future after-tax income than the permanent program. So much less of the spending rise will be offset by a fall in consumer demand. (I'm not considering the effect of the spending in raising income, which would probably cause consumer demand to rise rather than fall.)
So economic theory -- Milton Friedman's theory! -- says that spending is a more effective form of stimulus than tax cuts.
Which is extremely interesting and informative--but it has nothing to do with the content of my post, which did not address the relative virtues of permanent and temporary spending increases or tax cuts. I was simply interested in how much of the spending people will save to cover the future taxes needed to pay for it.
It does make me wonder, though: if Professor Krugman believes this
to be true, why isn't he making more fuss about the likely-to-be-permanent components of the stimulus package, like the new Cobra
provision to pay for many of the health care costs of laid-off workers
over 55?
I believe Business Week had a story last year, showing that people in their late 50's early 60's without health insurance, end up costing the Government more in Medicare Benefits. I can't image not treating heart disease, hypertension, diabetes etc.. is cheaper in the long run for the taxpayer, unless these people die before being eligible for Medicare
The extent to which I can be wrong makes me laugh with the type of hysterical shame that makes one even more deeply ashamed. However, the capacity of other economists to get it backside before the left hand can drive me to despair.
Megan,
It all depends on the type of tax cut or spending increase. Given a strong permanent income hypothesis, a temporary cut in consumption taxes or in payroll taxes or a temporary increase in unemployment benefit will only feed through to matching increased saving with a lag. During that lag, demand is stimulated.
Megan and Prof. De Long,
To put it simply, consumers' estimates of the likelihood of unfavourable outcomes have increased. Their propensity to save has therefore risen and their propensity to consume fallen. The multilpliers have therefore changed.
Prof. Krugman,
Broadbrush thinking is rarely, very rarely applicable to concrete policy choices. Consider first a temporary cut in consumption taxes (e.g. the current British temporary cut in Vaue Added Tax). To improve her welfare, a rational consumer spends more during the cut and less later. The stimulus is direct and quick. Second, consider your temporary bridge repairs program. That will reduce consumer costs over the longer term, so reducing marginally needs to save; but will feed through to actual spending and demand multipliers almost entirely after the British VAT cut has been reversed and ceased to have positive effect.
Further, please, please bear in mind that changes in the current balance between saving and comsumption (under a permanent income hypothesis) refect expectations of what the government will do in future, not politicians' proclamations. Confidence is vital if spending (or tax cutting) is not to be offset immediately by precautionary saving. (Arguably, confidence in the competence and effectiveness of the political establishment was what was missing during the failed stimuli of 1990s Japan.)
Megan,
I really don't think Krugman has a solid argument. The argument about the one time spending doesn't sound any more/less solid than the usual argument against one time tax cuts. Both are financed with "future taxes", so why is it any more/less logical to believe that consumers will save the entire tax cut to pay for it in the future than to believe that consumers will save the amount needed to provide the income to pay the "future taxes" for the one time spending. The two arguments appear to be the same, but Krugman believes only one of them?
If I am right, and I don't see where I am wrong (maybe some of your readers can point to the flaw in my argument), then Krugman is either being dishonest or is a moron.
Sigh. If a permanent gov't spending increase reduces permanent consumer spending by a lot, and a temporary gov't spending increase reduces consumer spending less, guess what would REALLY help consumer spending? A reduction in government spending, coupled with a permanent tax cut!
I miss George W. Bush already.
Didn't take long to find support for my argument. Robert Murphy addressed exactly this argument on Wednesday of last week- you will have to scroll down to find it since he doesnt' appear to provide individualized pages for his posts. Below, I have excerpted the relevant parts:
The arguments on both shot-terms spending (stimulus) and tax-cuts both look pretty shaky. The only advantage to tax-cuts is the ample evidence in past adventures of fiscal policy. They do seem to hold up Milton Friedman's Permanent income hypothesis. Whereas pure stimulus packages are almost solely the domain of the 1930's. Which, by virtue of the economic situation of the time, makes the conclusions suspect.
Thus the overall academic debate seems handicapped. But not to worry:) we're going to embark on the guess and check fiscal policy model. Brace yourelf.
...if Professor Krugman believes this to be true, why isn't he making more fuss about the likely-to-be-permanent components of the stimulus package, like the new Cobra provision to pay for many of the health care costs of laid-off workers over 55?
Perhaps because he believes that permanently strengthening the safety net is a worthy and critical goal based on the merits -- sufficiently worthy, that is, to justify the diminution of the marginal effect of the nth dollar of the stimulus package (especially when some of the gains can be clawed back by making the package larger in absolute terms).
The argument about the one time spending doesn't sound any more/less solid than the usual argument against one time tax cuts.
Yancey: That's because it's not supposed to. Krugman is simply explaining the benefits of temporary over permanent stimulus. He's not discussing here the virtues of spending vs. tax cuts. I think Krugman (and I know plenty of other liberals) have made the argument in favor of spending and against tax cuts (or stimulus checks, which are the same thing). But that argument usually has to do with the economic status of the recipients: a certain percentage of broad based tax cuts (or government checks) for stimulus purposes inevitably ends up in the hands of people who are in a position to save the money (or pay down debt) whereas money for things like food stamps or unemployment insurance nearly all gets quickly spent.
The arguments on both shot-terms spending (stimulus) and tax-cuts both look pretty shaky.
Really? I think it looks overwhelmingly logical and compelling. A three year stimulus component (whether tax cuts or an expenditure) costing, say, $300 billion+ (100 billion/year x 3 years plus interest) is obviously smaller in financial terms than a permanent (let's call it 30 years for our purposes) component costing $3 trillion (100 billion/year x 30). A logical taxpayer will put aside more money (by curbing spending) to pay for the latter than to pay for the former. Right? Or am I missing something?
I generally belive that the Permanent Income Hypothesis is sound as a first approximation in most cases. However, one thing that it does assume is well-functioning credit markets that allow people with relatively little savings but good future income prospects (like, say, college students and people in their first job) to borrow against their future income. Of course, if there's one thing that virtually everybody agrees about right now, it's that our credit markets aren't working too well.
So, it seems to me that there is some scope for certain types of temporary stimuli that would benefit people facing credit constraints (more help for student financial aid, for example) that could have a significant impact on aggregate demand, even assuming a strong form of the Permanent Income Hypothesis. That said, it seems that only a small fraction of the stimulus is target towards those areas.
Jasper,
Krugman and others always bring out Ricardian equivalence when arguing against temporary stimulus checks- that is the argument for the well off saving the stimulus check rather than using it for present consumption, but then he denies the same argument is valid against one-time spending. He is being inconsistent, or dishonest (I pick dishonest since I don't think he is stupid).
Krugman and others always bring out Ricardian equivalence when arguing against temporary stimulus checks- that is the argument for the well off saving the stimulus check rather than using it for present consumption, but then he denies the same argument is valid against one-time spending.
Ricardian equivalence is the theory that rational taxpayers should act the same whether a government finances its spending on a pay now (taxes) or pay later (bonds) basis, correct? But applying the theory in the real world obviously requires that these rational taxpayers can actually afford to eat and pay rent, and that there remains a surplus they can put aside to pay for the future higher taxes that will eventually be required. But unemployed or poor people -- the primary recipients of stimulus-related jobs, or increased unemployment insurance benefits, or increased food stamps spending -- presumably are less likely to be in a position to set money aside for future tax hikes than all taxpayers as a group. So, temporary spending (which disproportionately targets poor and/or unemployed persons) surely does pack a bigger stimulative punch than temporary broad-based income tax cuts or stimulus checks. If Krugman indeed argues what you say he does, I therefore reckon he's on firm ground.
To simplify, poor and/or unemployed people on average possess a greater marginal propensity to consume. Increased government spending -- which on average benefits such persons more than it benefits taxpayers in the aggregate (most of whom, after all, are both employed and ineligible for government safety net programs) -- therefore stimulates consumption -- and by logical extension the economy as a whole -- to a greater degree than broad-based tax cuts do.
re: if Professor Krugman believes this to be true, why isn't he making more fuss about the likely-to-be-permanent components of the stimulus package, like the new Cobra provision to pay for many of the health care costs of laid-off workers over 55?
This is actually an early reform which will be rolled into a more generic healthcare overhaul as we (finally) move to a universal system. And those who prefer keeping healthcare as privatized as possible should support it since the alternatives are either enrolling these people in Medicare, or (as we do currently) compelling them beggar themselves if they suffer a serious health problem and then enrolling them in Medicaid. IMO, we should be paying COBRA for all unemployed people, period.
Short-term stimulus and tax cuts (Also short-term). The "and" apparently didn't connect the dots well enough. Basically short-term tax-cuts have a proven effect of people solely saving the money. The comparison effect I was going for is that we are fairly sure what happens with short-term tax cuts (there is a human element so its never a 100%). Stimulus on the other hand does not have a large record to pick from, and the best examples come from the 30's, which due to its own extreme circumstance, is not fundamentally sound. Trying to sum it up think of it as a doctor who has a patient with a disease. He is looking at two ways to cure this disease, one (tax-cut) has been done before and it hasn't always worked. The other (stimulus) has been done before albeit less frequently, but its results are skewed and the effects have never been concretely connected. Which do you go with?
Personally I'm a stimulus fan, but since this is a credit problem I fear the only way to solve it is going to be Bank cannibalism. Which means that stimulus is necessary to keep the shakeup at the top from reverberating too hard on the bottom. Bon Appetit!
Jasper,
If you can show me where he makes that argument, I will let it go, but I can't find that. We aren't talking about jobs for the unemployed, remember- the new infrastructure spending will employ the already employed for the most part- there are no proposals for pure make work.
There are perfectly good, independent justifications for giving more extensive safety net payments to the unemployed, but then I don't see the multiplier any longer since the funds must still arise from the private sector and you aren't getting any "investments" from the additional spending.
Krugman can no longer be taken seriously. His economic opinions vary depending on which party is in the White House. Krugman has become an agenda driven pundit with an Economics degree. His Nobel was awarded because of his politics.
there remains a surplus they can put aside to pay for the future higher taxes that will eventually be required.
I don't understand the arguement that borrowing now will increase taxes in the future. This seems to assume that any money spent by the government will have no positive aspects in terms of growing the tax base or reducing costs.
As a business owner, I can borrow say $100,000 dollars now to expand to my restaurant, and while I will have to pay interest on this loan, any such interest would be offset by the fact that i can get more profit from the expansion than i would pay in servicing the debt (otherwise why expand my restaurant?) It seems this should be the same with government. If the Government gives a high school graduate say $12,000 dollars in pell grants, and she graduates, earns far more than she would have otherwise and pays more than $12,000 dollars (+interest) in taxes than she would have otherwise paid over the course of her career, this is a profitable move for the government, and they wouldn't have to increase taxes to cover the debt. This would seem to be true for all sorts of interventions (broadband internet, better electrical grid, healthcare).
If the Government gives a high school graduate say $12,000 dollars in pell grants
Unless colleges just raise their tuition by $12,000 because everyone now has more money to pay for college. Then nobody gets in that couldn't afford it before, and we give $12,000 gifts to colleges. Secondly, the more people that go to college the less those people will make relative to a high school degree. You can't say that somebody who goes to college and gets a BA in philosophy has better income prospects than somebody who does an apprenticeship in plumbing. Especially if the job market is flooded with BAs.
Megan,
I think Brad's critique is essentially valid, but only in the most limited sense. You can be excused in the name of brevity and the form. I would add that the fact that they have explored some potential effects of the permanent income hypothesis does not mean they actually have a good answer to the extent that it actually affects the multiplier. They have a bunch of plausible theories, but no very little experimentally confirmed results.
As for Krugman, he is an embarrassment not only to his school and to the Nobel committee, although the Nobel committee certainly doesn't need any help embarrassing itself these days. I don't envy you having to waste time refuting his arguments. Only commenting on what is interesting is a luxury only the non-pundits get to enjoy.
vis a vis Krugman's post, Megan says "but it has nothing to do with the content of my post...I was simply interested in how much of the spending people will save to cover the future taxes needed to pay for it."
Wow, um I've been exposed to PIH theory for all of five minutes now but even I understand the gist of Krugman's argument. Basically, PIH theory says the consumer spending offset will be small relative to spending because of the "smoothing" affect of the applied future tax cuts. This is the answer to your query, no?
Unless colleges just raise their tuition by $12,000 because everyone now has more money to pay for college.
but is this really what happens? University educations arent pefectly elastic, and Prices arent going to magically increase by 12,000 dollars, really what your doing is providing funding to schools with the added benefit of giving someone training (and while the school has increased costs for the student, because of economies of scale, the increases are relatively low).
but can't say that somebody who goes to college and gets a BA in philosophy has better income prospects than somebody who does an apprenticeship in plumbing. Especially if the job market is flooded with BAs.
but you're comparing two people with different post-secondary educations, not a person with and a person without secondary education. Even the person with BA in philosophy is likely to have better career prospects than a girl who only has a high school diploma.
But unemployed or poor people -- the primary recipients of stimulus-related jobs, or increased unemployment insurance benefits, or increased food stamps spending -- presumably are less likely to be in a position to set money aside for future tax hikes
The poor don't pay income tax and aren't likely to in the future. So they don't need to put aside money to pay for the later increases.
It should be obvious there's a "happy minimum" size of government in which it provides just the minimum level of externals that are cost-effective. Any increase in government spending over this must damage the overall economy.
The notion there can be a positive multiplier at current levels of spending is highly dubious.
It seems everyone here is forgetting the "invisible foot" of government.
If you let taxpayers keep more of their money, they will save it (helping the banks improve their capital, and thus clean up their balance sheets) or else spend it on things they consider worth the money.
If you let the government spend the money, it will spend it on things that politicians like Jack Murtha and Robert Byrd think are worthwhile.
How often is it that "Other People's Money" is spent as carefully as "Your Own Money"?
The GDP goes up the same whether I pay someone $200 to dig a hole and then fill it in (or to work on a Bridge to Nowhere), or if I pay someone $200 to landscape my back yard (or build a workbench). But the nation's real wealth increase a lot more when I do the later, because in that case I'm getting something useful for the money.
Put the money into tax cuts, and what is bought with it will be useful to the people who buy it. Put the money into government spending, and the "usefulness" of the spending will be judged based on "how does this help Congressman X win re-election" rather than "is this the best way to spend the money?"
Lots of people here don't seem to know the first thing about macroeconomics. And yes, that means Keynes. You talk as if the aggregate economy is the same as your private household budget, only bigger. Well, it's not.
1) There's a shortfall in aggregate effective demand. That means productive resources are left idle: Workers unemployed, physical capital underutilised. Liquidity piling up in bank vaults.
2) Normally you'd loosen monetary policy to lower the interest rate and thus stimulate investment and consumption. Well, The Fed has done just that and loosened all it can´- and then some. The interest rate is at zero - but it doesn't help any. (Or enough).
3) This situation is worse - much, MUCH worse - than your standard recession. The reason is that rational economic behaviour on the part of individual firms and households reinforces the problem in the aggregate via:
- The paradox of thrift - increased savings means falling demand, falling production, falling employment, falling incomes. The increased savings rate applied to a now lower income results in less than the initial desired actual saving. Thus, consumers increase their savings rate even more, and on it goes.
- The paradox of deleveraging - the need to pay your creditors and limit your exposure in an uncertain situation means its rational to sell off your assets. But when everybody does this at once, asset prices fall, and the situation worsens, and on it goes.
- Debt deflation - when faced with falling demand for your produce or your labour services, it's rational to lower your asking price or wage. But, when everybody does so at the same time, you get general deflation which means whatever debts you have become an even heavier burden around your neck. Deflation also means it's rational to postpone your purchase and hiring, thus lowering employment, investment and consumption, and on it goes.
In short, we're in Depression territory. The risk is it will just get worse and worse with absolutely no end in sight. The normal self-correcting mechanisms of falling prices, wages and rational adjustment of economic plans become self-defeating in the aggregate.
We're basically f***ed.
The only available remedy is fiscal expansion: Put idle resources to use by government spending. And the risk is you do too little rather than too much. If you do too much, you'll be able to reign in your overshooting quickly by raising interest rates. If you do too little, well, monetary policy will still not be available to stimulate demand.
Krugman is right. Keynesianism is here to save capitalism from itself. Get it and get over it.
I noticed on Oct 15 of last year you posted a very snarky commentary on N. Roubini. Mr. Roubini on Oct 8 last year stated on in an interview on Yahoo Finance that the Dow would be at 7000 in 2009, It closed at 7062 today. None of the last None recessions. Hmmm.