Keith Hennessey critiques the stimulus package. Some points make sense, some points, well, I wonder. For instance, Hennessey argues the stimulus is not timely. As I’ve noted before, it’s not timely only if you think this will be a relatively short recession, characterized by a rapidly dissipating negative output gap. [0] [1] [2].
That scenario hardly seems to be the consensus, as indicated in CBO’s most recent analysis.
And, according to CBO, the output gap would have been even deeper and more persistent in the absence of the stimulus package, as constructed.
But I do agree that the package could have been better crafted. For instance, I did not see why the Republican moderates at the time of negotiation demanded a reduction in transfers to the states [3]. That seemed to me crazy on the efficiency and timeliness grounds that Hennessey highlights (see also Jim’s take on support for states). And yet, now we see the fiscal situation of the states deteriorating further, requiring additional cutbacks on government spending exactly when spending is most needed. [4]
On the other hand, I do disagree with Hennessey’s attraction to tax cuts (reminiscent of this view of the world). He bases his approach on the implications of the Hall version of the Permanent Income Hypothesis (essentially a Euler equation linked up with rational expectations). Well, as Campbell and Mankiw documented in a series of papers back in the early 1990s (and subsequently confirmed by subsequent research), the PIH seems to be a very poor characterization of consumption behavior. There seems to be a large share of “rule-of-thumb” consumers, who consume a constant share of income. If this behavior is being driven by liquidity constraints that arguably have become more binding with the financial crisis and economic recession, then the preference for permanent tax cuts over temporary is less well-founded (see discussion of the PIH here).
So, while I think that while it would be preferable to have the spending quickly implemented, the single-minded concern for speed is based upon a highly unlikely scenario. In any case, according to CBO, most of the spending will take place within the next 16 months (19.5 months from when the bill was passed).
An aside: I was amused by Hennessey’s lauding of the 2003 tax cut as effective fiscal stimulus. I think we know that we are partly where we are — hamstrung with massive inherited Federal debt, much owed to the Chinese, and an overextended household sector — exactly because of that ill-advised measure [5] (stacked on top of the 2001 tax cuts).
I do not see how the prospects for a long recession provide an excuse for a lag in spending. I will grant that in a long recession, the probability that spending a year from now will kick in while there is still slack in the economy is high.
But even if you expect slack, the sooner the spending kicks in, the better. The whole concept of a multiplier is that you get spending up now, employment up now, and those people who are working now start to increase spending later. The multiplier is larger the sooner you get the process started.
The statement “it’s not timely only if you think this will be a relatively short recession” is incorrect in any basic Keynesian model, unless by “timely” all you mean is that the recession is not over by the time the stimulus is spent.
To me, “timely” should have to do with getting the most bang for the buck. From that perspective, it is a huge mistake to have delayed stimulus. The fact that we are in a deep and probably prolonged recession only worsens the mistake.
I also do not understand the priority with giving money to state and local governments. The unemployment rate among government workers is the lowest of all worker categories. If the goal of stimulus is to put money where the slack is, then government should not be the priority. (By the same token, government construction projects would be helpful, since unemployment is high among construction workers.)
Arnold Kling: Thanks for your insightful comments. It’s fine for you to redefine timely as equivalent to efficacy, but that isn’t the standard way the people have debated the issue. Efficacy for me refers to the size of the multiplier, while timely refers to whether the additional stimulus takes place during a time when there is slack. Since the maximal output gap is still to come, according to CBO, it’s not a big deal to me if the spending occurs in FY2010, as opposed to FY2009.
Staying in the Keynesian framework, transfers to the state are particularly efficacious since we know they’ll be spent (to retain teachers, police, fire services). Tax cuts — well we have heard a lot about how the cuts are going to rebuild balance sheets rather than consumption. I do agree that more allocated to infrastructure would have been better — but this is exactly where Hennessey critiqued the stimulus package — he thought too much was going there. So you should take that issue up with him.
Arnold Kling on 06/05/09 @ 04:05 PM asks:
“I also do not understand the priority with giving money to state and local governments.”
Here’s the answer (I am sure you know this already).
http://online.wsj.com/article/SB124227027965718333.html
How do we get to full employment without tax cuts?
“An aside: I was amused by Hennessey’s lauding of the 2003 tax cut as effective fiscal stimulus. I think we know that we are partly where we are — hamstrung with massive inherited Federal debt, much owed to the Chinese, and an overextended household sector — exactly because of that ill-advised measure…”
My word, if effect on the size of the national debt (and knock on effects) is the measure to be used, let’s remember those Bush deficits 2003 to 2008, after that tax cut, averaged 2.7 points of GDP, for six years 16 points.
CBO puts this year’s deficit alone at 13 points, projects the four year total deficit for Obama’s first term (2009-12) at 33.3 points of GDP, then for his presumptive second term adds 17.6 points more. That’s 51 points of GDP in deficits that Obama’s budgets are planning to give us in eight years. An average of 6.4 points.
For perspective, all the infamous deficits of Ronald Reagan (who started off with a recession as bad as today’s if not worse, 10.8% unemployment) totalled 34 points of GDP in eight years, 1981-1988.
Yes, Obama’s projected eight-year deficits are as large as all the historic Reagan deficits and those Bush deficits combined! Obama is planning in just two budget-busting terms to match the four most budget-busting terms in our history combined.
Obama’s smallest projected deficit over eight years is 4.1% of GDP (in 2013, then they start growing again, forever). Bush’s biggest deficit in his eight years was 3.6%.
In short, it’s well nigh time for people who affect concern about the size of the national debt, and about how much of it the Chinese are going to wind up owning, to stop wagging their fingers backward at the Bush deficits and start wagging them forward at the Obama deficits.
Jim Glass: I’ve discussed before the concept of full employment budget balance. I think it would be useful for you to review that concpet, and to keep in mind the shift in balance that occurred from big surpluses to big deficits.
I also think, in considering the lingering legacy of the Bush Administration, you should keep two words in mind: contingent liabilities. A large portion of the current fiscal year’s deficit is due to the bailout; and the remainder, the declining tax revenue associated with the financial crisis bequeathed by the Bush Administration. Please see (again) this post.
Menzie Chinn on 06/06/09 @ 09:36 PM
“I’ve discussed before the concept of full employment budget balance. I think it would be useful for you to review that concpet [sic], and to keep in mind the shift in balance that occurred from big surpluses to big deficits.”
What I’ll keep in mind are the costs of this new employment (taxes, subsidies, lower productivity, etc…)
Delayed implementation of the “stimulus” is totally a political calculation. When the “stimulus” that has been injected does not work the politicians will point to the future “stimulus” and say, “It hasn’t been injected yet. Give it time.”
Arnold Kling is right in one sense. If the stimulus actually works as advertised it should be done as quickly as possible before more people lose their jobs and enter into starvation. But Kling has it wrong in another sense. This kind of “stimulus” simply does not work and makes the bad worse.
There is Keynesian stimulus that does work but not because of the reasoning used by Keynes. The Kennedy-Johnson tax cuts were Keynesian tax cuts to inject money into a sagging economy. This type of Keynesian stimulus works precisely because it conforms to Say’s Law which is anathema to Keynesians.
Our current money crank stimulus schemes would have made Keynes sick to his stomach.
Menzie wrote:
An aside: I was amused by Hennessey’s lauding of the 2003 tax cut as effective fiscal stimulus. I think we know that we are partly where we are — hamstrung with massive inherited Federal debt, much owed to the Chinese, and an overextended household sector — exactly because of that ill-advised measure [5] (stacked on top of the 2001 tax cuts).
To use Menzie’s reasoning:
An aside: I think we all know that Menzie is full of it when he assumes he knows what we all know.
DickF: Say’s Law is anathema not just to Keynesians, but to people who want to base their reasoning on observations of reality. So I say yes to Walras’ Law, no to Say’s.
Regarding your second comment: Thank you; in the future please be so kind to document your assertions, as I try to with respect to discussion of your commentary. Restraint from engaging in ad hominem attacks would also be welcome (recalling our exchange surrounding the October 31, 2006 post).
The current stimulus package is wrong for the same reason that the 03 tax cuts were wrong. They both occur well after the recession has ended. If the 03 tax cuts hampered our ability to deal with this crisis, what do you think this spending proliferation will do to our ability to handle the next crisis?
Menzie wrote:
So I say yes to Walras’ Law, no to Say’s.
So what’s the difference? Walras’ Law quantifies Say.
I know that Menzie has already addressed Jim’s error, but let me offer other terms. The usefulness of a single-figure metric of recession “badness” is just about zero. The rise in the jobless rate, bottom to top, during the Reagan recession, was 5.2%, from 5.6% to 10.8%. The rise so far in this recession is 5.0%, from 4.4% to 9.4%, and there is good reason to believe the rise is still underway. That rise is far more telling than a single level. Don’t we kinda suspect that any given jobless rate means something different at different times? Haven’t we noticed that NAIRU seems to move around?
We have more than one jobless rate. We have more measures of economic performance than just the jobless rate. We have many, many economists on record saying this is the worst recession since the Great D, but Jim G wants Reagan’s recession to be worse, based on a single, static figure. Better to try that sort of argument somewhere there’s a chance people will fall for it.
Now, on to the central issue. To argue for more rapid implementation, more rapid implementation has to be reasonably likely that more rapid implementation is possible. China has shown that more rapid implementation is possible – in China. Do we have examples of faster implementation of government spending on a large scale in the US than has been the case this year? If not, then isn’t complaining about the delay rather sterile?
I really don’t know whether large-scale spending has gotten underway faster, but that seems a reasonable issue to address before claiming the current effort was too slow.
Recessions are demonstratable failures of PIH.
DickF,
Say’s Law says “is”. Walras’ says “if”. There is a substantial difference between those two.
“a glut can take place only when there are too many means of production applied to one kind of product and not enough to another”
kharris,
Who wrote the above?
DickF,
Jean-Baptiste did. But you have chosen one of several statements of his “law” which seems to allow room for an “if”. Most other statements do not. This is what is known as cherry-picking.
If you prefer Say when he is closest to allowing that a general glut is possible, please say so outright. You have claimed their ideas are equivalent. They are if you restrict Say to that instance you cite. If you only toss it out to avoid recognizing that, as Menzie suggested, Say and Walras are generally seen to have said rather different things, that is not entirely honest.
Note that Say often talks of money as if it has a very limited role, that goods production and consumption cannot be affected by money. There can be local disequilibrium, but not a general glut. Walras says that there cannot be disequilibrium in only one market, but allows that there can be disequilibrium, there can be a general glut. Money plays a role there. If you don’t understand that, say so and I’ll explain at greater length.
Anon,
I know that Walras were different on some issues especially central planning and welfare (Walras defended), but on the issue of a general glut they were basically saying the same thing.
Say talks of money as if it SHOULD have a limited role. Changes in the quality of money are always destructive whether they are inflationary or deflationary and when you gyrate from one to the other as we do under a floating currency regime the errors do not cancel out but compound one another. For example deflations restrict investment in infrastructure so that when an artificial boom is created by inflation there are serious dislocations in price increases and we get “bubbles.” The best example of this was the Greenspan deflation of the late 1990s preventing oil infrastructure investment followed by the international boom of the 2000s where storage capacity could not keep up with demand and prices of oil shot through the roof.
What is missed in both Keynesian and monetarist analysis is that changes to the quality of money poison the interaction between traders. One will reap a windfall profit while the other a loss because of monetary manipulations. This creates friction in the economy as traders attempt to hedge against potential windfall loses.
There are definitely glut situations but these are created by intervention in the free market, monetary or fiscal. Say’s law tells us that, given the conditions traders have to live with, there can be no glut, or shortage for that matter, because all production not consumed by the producer is used in exchange, but when the rules of the game are changed after the fact a supply that was proper, based on the market demand, is suddenly changed into a glut.
For example think about real estate tax law. Homes are built with the consideration that real estate taxes will be deducted from income tax. If that is changed and taxes are no longer deductable demand will instantly decline and there will be a glut. But this is not because of Say’s law but because of government intervention.
DickF: If you believe Says’ Law and Walras’ Law are the same, than indeed you are in a different universe than I am.
Okay, I am open. Educate me on the differences between Say and Walras concerning glut.
Okay, I will stop playing with you. The differences between Walras and Say are that Walras’ attempted to subject Say to econometrics and created what Rothbard called a “Walrasian Box.” Walras required mathematical equilibrium and that created a problem between his threory and the real world. Schumpeter (one of my favoriets btw) faced the same problem of being locked in the Walrasian Box (again Rothbard) because he had to contradict some of his own principles, such as the time preference theory of interest, to maintain Walrasian equilibrium.
This is the problem with econometrics. Too often it drives theory when it should be considered a tool to validate theory.
There are many who believe that Walras and Say are saying the same thing. After all Walras was in business with Say’s grandson so he was probably intimately involved with Say’s economics.
But I would still like to read what you believe are the differences between Walras and Say.
DickF: Walras’ Law, short version, sum of excess demands for commodities must equal zero. No econometrics necessary; I learned it in a macro course.
I don’t understand how government spending potentially improves the economy. If the problem is debt, how does the government generating more debt get the United States out of debt? Weren’t these kinds of ‘stimulus’ tactics tried during the depression of the 1930’s which only led to the longest and worst depression in United States history? If government investment in and influence over the economy led to prosperity and stability wouldn’t the Soviet Union still be alive and flourishing? Can somebody direct me to resources that will clear my confusion?