Wren-Lewis seems to refute his own argument
In a previous post I criticized Wren-Lewis’s claim that John Cochrane’s denial of a positive balanced budget multiplier was wrong because of consumption smoothing. Of course there will always be debate about the balanced budget multiplier itself, but I hope now we can at least all agree that consumption smoothing has nothing to do with it, as it is part of the classical model. In a comment at my blog he argued that saving doesn’t equal investment. In a new post he repeats that claim, then contradicts it, but by doing so undercuts his own argument:
This post is for first year undergraduate students (and the occasional blogger) who appear confused.
Q: If consumers spend less and save more, does this mean investment must increase?
A: Absolutely not. Someone increasing their saving does not automatically imply that some firm will decide to buy more capital goods.
Q: But surely savings equals investment by identity in the national accounts.
A: Indeed. Total output = total income = total expenditure = Y. In the most simple model of a closed economy without government, income (Y) = consumption (C) + saving (S), but also expenditure (Y) = consumption (C) + investment (I). So S=I by definition. But here investment includes what is called ‘stockbuilding’ or ‘inventory accumulation’, which includes goods that firms wanted to sell but could not. To make this clear, lets split measured investment (I) into these two components: I=DK (buying new capital goods) +DS (stockbuilding). So if people consume less (C falls), but investment in new capital (DK) stays the same, measured investment rises because firms accumulate inventories of the goods that consumers did not buy (DS rises).
Q: But this situation cannot continue, as firms may be losing money.
A: Exactly. They will cut back on their output, incomes will fall, consumption may fall further, and savings will also fall, cutting back on the initial increase that we started with.
The first Q&A seems clear. Saving can increase without any increase in investment. Hence contrary to the claim of Mankiw and Krugman, S=I is not an identity.
But then he claims S=I is an identity in the second Q&A. Wren-Lewis is a Professor of Economics at Oxford, so I’m going to follow Tyler Cowen’s advice and give him the benefit of the doubt. I must have misunderstood the first Q&A. Let’s continue.
In the rest of the second Q&A Wren-Lewis says investment is composed of both inventory accumulation and the construction of new capital goods. Technically, inventories are a part of the capital stock, but he’s right that they aren’t generally viewed as ‘capital goods’ in terms of everyday speech, so I’m completely fine with that. It still makes the first Q&A oddly worded (the question should have asked if fixed investment increased), but I’ll pass over that.
So now let’s revisit his consumption smoothing argument. He suggested that if after-tax income fell, then due to consumption smoothing C would fall much less. Let’s assume it only fell 20% as much as after-tax income fell. The next question would be; so what? How does that in any way undercut Cochrane’s argument that private spending (on C+I) would fall? If after-tax consumption fell by 20 and after-tax income fell by 100, then saving would obviously fall by 80. Suppose we hold fixed investment (DK) constant, as it’s hard for businesses to build new capital good at the drop of a hat. I’m fine with that as a short run assumption. But in that case isn’t Wren-Lewis claiming that both saving and new inventory accumulation have to fall by 80? In that case Cochrane is right.
Yes, this inventory accumulation is likely to be temporary, but that’s the assumption for the fiscal stimulus as well. If the new spending program was permanent, and funded by permanent tax increases, then there’d be no obvious need to assume consumption smoothing, people would have to adjust to the full impact of the tax increase. And of course fixed investment might also fall in the long run.
Astute Keynesians will notice that there is a way out of this box. Perhaps the attempt to save less causes velocity to rise, so after-tax income does not in fact fall by 100. That’s a mechanism I discussed in the earlier posts, but it has nothing to do with consumption smoothing. The basic problem with Wren-Lewis’s argument is that he implicitly led his readers to assume that total spending would fall by less than after-tax income fell, because consumption fell by less than income fell. But even using his assumptions, the total decline in spending on consumption plus inventory accumulation will equal the decline in after-tax income. So consumption smoothing doesn’t get you anywhere, you need to make a separate argument for why these events will raise national income, perhaps via less demand for money and/or more velocity. My hunch is that Wren-Lewis understands this, as later in the post he presents an excellent explanation of just such an effect, and makes no mention at all of consumption smoothing. Indeed it’s pretty much the sort argument that I suggested Krugman and Wren-Lewis make when I suggested they use this argument:
Cochrane ignores the fact that tax-financed bridge building will reduce private saving and hence boost interest rates. This will increase the velocity of circulation, which will boost AD.
So I still think I was right that his use of consumption smoothing was a fairly basic mistake, akin to what he and Krugman accused Cochrane and Lucas of having made.
Or maybe I’m just a confused blogger.
HT: CA
Tags:
15. January 2012 at 14:26
http://www.themoneyillusion.com/?p=9960
this reminds me of something from this old post that stuck in my mind
15. January 2012 at 14:30
Tim, Yes, that was one of my better posts.
15. January 2012 at 14:41
“Yes, this inventory accumulation is likely to be temporary, but that’s the assumption for the fiscal stimulus as well.”
You’re playing a game with the words “temporary” and “permanent” here (just as Wren-Lewis is playing a game with the word “investment”). The horizon for inventory accumulation is clearly shorter than the horizon for fiscal stimulus.
15. January 2012 at 16:06
It’s hard for a layman to follow this discussion. Still, it is a very interesting discussion nonetheless. It’s funny to think back that a quip by Lucas began all this.
I also am pretty sure that as smart as everyone is, no one is coming at this from a purely objective perspective. If the Keyenesians give up and admit that S=I, their stimulus argument goes up in flames. So, they see this is as something they cannot back down from. As Krugman himself points out, it’s hard to argue with people when what you are arguing about is implicitly their reputation and sometimes their income.
All I would say for you and Nick Rowe is don’t back down. Don’t let them beat you down into submission. (Btw, I think you are very close to flipping Matt Yglesias to a conservative. Read some of his comments and see his readers bash him for as they say “spewing conservative talking points”)
Speaking of Krugman, he pisses me off again. In this post, he complains that the composition of the top 0.1% is all just executives and managers, or as he puts it “corporate suits”. Who does he expect to be in the top 0.1%? Janitors and mechanics?
http://krugman.blogs.nytimes.com/2012/01/15/but-the-top-0-1-percent-isnt-diverse/
15. January 2012 at 16:10
“So now let’s revisit his consumption smoothing argument. He suggested that if after-tax income fell, then due to consumption smoothing C would fall much less. Let’s assume it only fell 20% as much as after-tax income fell. The next question would be; so what? How does that in any way undercut Cochrane’s argument that private spending (on C+I) would fall? If after-tax consumption fell by 20 and after-tax income fell by 100, then saving would obviously fall by 80. Suppose we hold fixed investment (DK) constant, as it’s hard for businesses to build new capital good at the drop of a hat. I’m fine with that as a short run assumption. But in that case isn’t Wren-Lewis claiming that both saving and new inventory accumulation have to fall by 80? In that case Cochrane is right.”
A fall of 100? In response to what? When exactly did Wren-Lewis suggest that after-tax income had to fall by 100? It could fall by 50 with only a fall in consumption by 10, resulting in a fall in investment by only 40.
C-10
I-40
G+100 <– Note the balanced government budget
T+100 <– Increased G balanced by increased T
Y+50 <– Increased Y
Yd-50 <– Decreased Yd
Done.
15. January 2012 at 16:16
Or maybe…
C-0.01
I-0.04
G+100
T+100
Y+99.95
Yd-0.05
How about that?
15. January 2012 at 16:27
“Astute Keynesians will notice that there is a way out of this box. Perhaps the attempt to save less causes velocity to rise, so after-tax income does not in fact fall by 100.”
Or, you know, it’s the GOVERNMENT SPENDING $100 MILLION A BRIDGE which causes velocity to rise.
Just sayin’.
15. January 2012 at 16:28
“Astute Keynesians will notice that there is a way out of this box. Perhaps the attempt to save less causes velocity to rise, so after-tax income does not in fact fall by 100.”
Or, you know, maybe it’s the GOVERNMENT SPENDING $100 MILLION A BRIDGE which causes velocity to rise.
Just sayin’.
15. January 2012 at 16:29
Sorry for the near-dup. Had a network error.
15. January 2012 at 16:48
Heh this is pretty funny, Scott. I literally was halfway done, typing up a scathing comment in which I explained that Krugman et al. had been right on this, and Cochrane had been a moron. But then at the crucial point in the argument, it suddenly clicked what you have been saying in this debate. I’m glad I read this, because I actually have a draft blog at my site, titled, “Krugman Was Too Easy on Cochrane.” I think maybe I have to change it now…
Let me paraphrase what happened in this debate, Scott, and tell me if you agree:
1) Cochrane in several places, addressing the layperson, challenged even the theoretical possibility that fiscal stimulus could boost AD, by basically saying if the the government borrows and spends an extra $1 billion, it has to come from somewhere. It’s either taxed or borrowed, in which case current taxpayers adjust their lifetime after-tax income estimates accordingly.
2) Krugman et al. went nuts, saying that this is an illegitimate flipping of Ricardian Equivalence. R.E. says that if government holds spending constant and cuts taxes, then households don’t consume any of it. But, Krugman et al. continue, it doesn’t work the other way. If gov’t holds taxes constant in the present, and boosts spending, then it can raise total spending because we wouldn’t expect households to cut back their consumption spending all in this period; instead they would smooth the blow out over time.
3) Scott Sumner says this is goofy. The smoothing element just shows how much the household makes consumption vs. investment spending get dinged in the present, but it doesn’t alter the fact that households would reduce their total spending to perfectly offset the increase in government spending.
That’s what it sounds like you are arguing. But, (3) can’t be quite right, since you personally believe fiscal policy can stimulate aggregate demand. So what am I missing? Are you saying that yes, Cochrane was an idiot, but not for the particular reason Krugman et al. listed?
15. January 2012 at 17:28
Bob: If I follow your presentation correctly with respect to point 3, consumers not only undertake consumption smoothing but also savings smoothing. In other words, how do consumers smooth their consumption without affecting/altering their savings.
with respect to your question about how then fiscal policy affects aggregate demand, I think Scott would argue that it is through the monetary channel. As Scott argued:
“Astute Keynesians will notice that there is a way out of this box. Perhaps the attempt to save less causes velocity to rise, so after-tax income does not in fact fall by 100.”
15. January 2012 at 18:12
Scott
The issue now is that you didn´t figure Krugman/Wren-Lewis were doing a comparative static exercise!
http://krugman.blogs.nytimes.com/
15. January 2012 at 18:15
Thanks for taking the time to try to sort out some of these issues.
It seems that you and Wren Lewis agree that the building the bridge can lead to an increase in GDP but you disagree with the best way to explain the move from one equilibrium to another one. It’s not clear that your two approaches are really that different; after all they do lead to the same place, and do it via increasing demand.
“Wren-Lewis claiming that both saving and new inventory accumulation have to fall by 80”
Yes, the inventory fall is a signal to firms of increased demand. This will lead firms to increase output and hire workers. When there’s unemployment, they can do that without having to hire them away from other firms, so total output increases. Equilibrium is restored, but at a higher level of GDP. [If we’re already at full employment, then output can’t increase, only prices, but that’s another story.]
In fact, as you note:
“this inventory [reduction] is likely to be temporary,”
Presumably one way this can happen is that firms will increase output to restore the missing inventory.
You seem to prefer another route to the same conclusion:
“the attempt to save less causes velocity to rise”
This is a signal to firms the demand is higher (just like falling inventories).
As before, this leads firms to increase output, and a new equilibrium is reached with higher GDP.
Am I missing something, or are you guys essentially arguing vehemently about whether you should do long division or short division to get to the same answer? Everyone has their favorite tools, and perhaps you guys just aren’t used to working things through with one another’s toolkits.
15. January 2012 at 18:41
Scott,
I seriously cannot figure out what you, Wren-Lewis, and Krugman are arguing about at this point. What is the issue here???
15. January 2012 at 18:41
Wow. Nobody does snarky like Paul Krugman.
15. January 2012 at 19:00
rob,
Why does Scott have to turn to the monetary channel? What happened to the bridge? Didn’t producing the bridge raise after-tax income? Why is Scott only counting the tax?
15. January 2012 at 20:40
D R: I am not a macroeconomist. I am just trying to follow the arguments presented by Scott and Wren-Lewis. So please bear with me.
1. Wren-Lewis makes the point that the multiplier is positive and equal to or greater than 1. To support this assertion, he made the following argument.
“If you spend X at time t to build a bridge, aggregate demand increases by X at time t. If you raise taxes by X at time t, consumers will smooth this effect over time, so their spending at time t will fall by much less than X. Put the two together and aggregate demand rises.”
So based on the above statement, one could conclude that income smoothing is central to his assertion about the multiplier effect.
2. Wren-Lewis says the aforementioned point of consumption smoothing refutes the Cochrane argument that the multiplier is negligible. Just to be clear, this is what Cochrane said and is quoted by Wren-Lewis:
Before we spend a trillion dollars or so, it’s important to understand how it’s supposed to work. Spending supported by taxes pretty obviously won’t work: If the government taxes A by $1 and gives the money to B, B can spend $1 more. But A spends $1 less and we are not collectively any better off.
3. I think Scott’s point is this: the consumption smoothing argument employed by Wren-Lewis does not refute or negate Cochrane’s argument as to why there will not be any multiplier effect. Cochrane states the dollars for tax financed projects must come from someone. If you tax A to hire/employ B to build the bridge, then A’s consumption will go down by an equal amount that is paid to B. Hence, there is a wash. Wren-Lewis says that we need to distinguish between consumption and saving. Consumption will not go down by an equal since the consumer would undertake income smoothing. However, Scott says that this argument is not sufficient to refute Cochrane. Not only will consumption go down in this setting, savings will also go down. How else is A able to smooth his consumption?
4. I don’t think Scott’s point is that fiscal policy will not positively impact aggregate demand. I think what he is trying to say is that merely resorting to consumption smoothing is not sufficient to refute Cochrane’s claim that tax financed government spending is a wash in that it merely transfers dollars from A to B.
5. D R: I hope I am making sense to you. If its muddled and incoherent, I hope you will set me straight. Scott: I hope I have correctly understood the point that you were trying to get across.
15. January 2012 at 22:14
rob,
There is a lot to cover there. Let me focus on items 3 and 4 to start.
Cochrane never actually addresses the question of tax-financed bridge building. He only addressed a forced transfer from person A to person B and then implied that the argument carried over to any kind of stimulus.
Leaving aside myriad reasons why his argument about transfers might not apply in the real world, there are very very good reasons why the argument would not carry over to the bridge-building.
Think of it like this. Cochrane argues that if the government taxes A and gives to B, there is no increase in production. Cochrane *implies* that if the government taxes A and gives to B on the condition that B produce a bridge, there is still no increase in production.
Which is a little strange, to say the least.
One may legitimately argue that B would have produced the bridge (or been otherwise just as productive) anyway. Which I accept. But there are lots of people who are looking for work, and Wren-Lewis makes an argument that if the government taxes in order to pay these otherwise unproductive workers to actually produce something, then more stuff will be produced overall.
Now. I know that Sumner does not disagree with that. What I don’t understand is why he is worked up about consumption smoothing. Consumption smoothing is irrelevant to the dispute with Cochrane. I know that Sumner does not disagree with that either.
The fact is, Cochrane’s “proof” never laid out an argument against tax-financed government spending on currently-produced goods and services. He merely asserted that it could not increase overall production. Thus, any case to the contrary– however flimsy– refutes Cochrane. At least, as long as the case is consistent.
Now, Sumner has chosen to ignore this, and is worked up about consumption smoothing. I think this is surely trivial compared to the refutation itself, but it’s his blog and that’s his choice. But I’m still not clear on exactly what statement of Wren-Lewis’ bothers him so.
15. January 2012 at 22:41
rob,
“So based on the above statement, one could conclude that income smoothing is central to his assertion about the multiplier effect.”
Welllll… strictly speaking there’s nothing wrong with making unnecessary assumptions in your argument. It frequently makes it harder to understand, but not every statement in a proof needs be “central”
I do agree though, Wren-Lewis says he employed consumption smoothing in his model and so we have to use consumption smoothing if we want to say anything consistent with Wren-Lewis.
“Wren-Lewis says the aforementioned point of consumption smoothing refutes the Cochrane argument that the multiplier is negligible. Just to be clear, this is what Cochrane said and is quoted by Wren-Lewis”
I am not ready to accept the first sentence there. That sounds like Sumner’s claim, but I don’t understand why Sumner specifically picks out smoothing as the crux of the thing. Then again, I’ve been at this so long, maybe I’ve forgotten.
I do accept that consumption smoothing is part of Wren-Lewis’ argument that Cochrane is wrong. But I’m not sure why Sumner seems to think it all hinges on it.
15. January 2012 at 22:43
(dangit….)
I had put fake tags in there after the first quote. They must have been taken out in processing. *sigh*
15. January 2012 at 22:44
(double dangit)
Fake HTML tags which read “pedant” and “/pedant”.
16. January 2012 at 02:14
This entire debate is stupid. I’m disappointed that Wren-Lewis couldn’t dismember it in one foul swoop.
>>First recall that C + I + G = AD = GDP = gross income in a closed economy. Because the problem involves a tax-financed increase in G, we can assume that any changes in after-tax income and C + I are identical. Checkmate in four.<<
This assumes that GDP doesn't change after the tax-financed increase in G. Which is what you're trying to prove. Circular argument. Oops.
To be a bit more descriptive, the taxes could come out of money hoarding, rather than C or I.
16. January 2012 at 07:54
“But in that case isn’t Wren-Lewis claiming that both saving and new inventory accumulation have to fall by 80? In that case Cochrane is right.”
Cochrane is wrong and Simon is right. Fall in inventories has positive effect on spending (some firms receive cash from selling off their stock and others want to spend on it so it generates spending) so if you add government spending plus decrease in consumption (less than government spending) the overall effect is positive.
16. January 2012 at 09:30
Simon is arguing that increase in savings equals increase in investment production but may not equal increase in investment spending. In the same way we can argue that decrease in savings doesnt have to equal decrease in invetment spending (because of stock liquidation) so I cant see error in his reasoning.
16. January 2012 at 10:14
Andy, Fine, but as long as saving falls, then some combination of inventory investment and fixed investment must fall. I have a new post that explains several possibilities.
Liberal Roman, You said;
“If the Keyenesians give up and admit that S=I,”
Don’t look now, but they have given up. Wren-Lewis first denied this, but now accepts that it’s an “identity.”
And yes, I’m seeing some interesting arguments from Yglesias. But in fairness it’s possible for him to accept a good bit of my monetary approach without fundamentally changing his progressive views.
DR, You said;
“A fall of 100? In response to what? When exactly did Wren-Lewis suggest that after-tax income had to fall by 100? It could fall by 50 with only a fall in consumption by 10, resulting in a fall in investment by only 40.”
Doesn’t help. The problem is he assumes that a $10 fall in C is a $10 fall in “spending,” but it isn’t. Spending falls by $50.
You said;
“Or, you know, it’s the GOVERNMENT SPENDING $100 MILLION A BRIDGE which causes velocity to rise.”
Yup, and if he’d explained why it’s be a fine argument.
Bob, You said;
“Scott Sumner says this is goofy. The smoothing element just shows how much the household makes consumption vs. investment spending get dinged in the present, but it doesn’t alter the fact that households would reduce their total spending to perfectly offset the increase in government spending.”
Close, but one very important mistake. Replace “would” with “might.”
Marcus, Yes, idiotic.
Erik, There are lots of ways of looking at this dispute, which I try to lay out in a new post. But I’d say the main problem is that Wren-Lewis tries to claim that consumption smoothing refutes Cochrane, whereas at best you might claim it’s consistent with the Keynesian model (but not necessary.)
Noah Smith, Check out my newest post (today’s post.)
Rob, Yes, I’m not arguing Wren-Lewis’s conclusion is wrong, I’m arguing that his argument seems to forget investment spending, and implicitly treat “spending” as C + G.
DR, As I said many times, Cochrane’s “argument” is just as weak as Wren-Lewis’s “argument.” Indeed both basically just assume they are right. I happen to agree with Cochrane’s conclusion (although in a much less dogmatic fashion–I concede I may be wrong) but not his reasoning.
Daoist. You said;
“This entire debate is stupid.”
I notice the most insulting comments are usually the weakest. No, I do not assume GDP is fixed.
Tom, You made the same mistake, you forget the effect on investment.
16. January 2012 at 10:15
Tom, Wren-Lewis says it’s an identity, I agree. There’s no way around it.
16. January 2012 at 10:34
“As I said many times, Cochrane’s ‘argument’ is just as weak as Wren-Lewis’s ‘argument.'”
I find your false equivalence disturbing. At least Wren-Lewis gave us something to argue over. Cochrane did not.
“I happen to agree with Cochrane’s conclusion (although in a much less dogmatic fashion-I concede I may be wrong) but not his reasoning.”
Oh? Cochrane’s “reasoning” you say? Show me where in Cochrane’s proof lies reasoning that a tax-financed increase in G results in a zero multiplier.
16. January 2012 at 10:40
“Tom, Wren-Lewis says it’s an identity, I agree. There’s no way around it.”
Ok but from production or spending point of view? Because it cant be both true if we allow frictions (stock changes).
16. January 2012 at 10:50
Scott, we know that investment spending=investment productions – change in inventories.
And Simon said:
“So S=I by definition. But here investment includes what is called ‘stockbuilding’ or ‘inventory accumulation'”
So what do you mean by saying identity? Because Simon is referring to production point of view: “measured investment rises because firms accumulate inventories of the goods that consumers did not buy”
16. January 2012 at 17:14
[…] couldn’t raise (nominal) Aggregate Demand. Krugman & Co. have been going nuts, and Scott Sumner has defended Cochrane on grounds that (I think) contradict Sumner’s own view that fiscal policy can indeed raise […]
16. January 2012 at 21:11
>>I notice the most insulting comments are usually the weakest. No, I do not assume GDP is fixed.<<
Apologies if that was taken the wrong way…but you say you assume that changes in G must be exactly balanced by changes in C + I ("we can assume that any changes in after-tax income and C + I are identical"). This means that C + I + G must remain constant. C + I + G happens to equal GDP. So what you are assuming is indeed equivalent to GDP being fixed – your conclusion.
17. January 2012 at 01:46
^Also, just so it doesn’t look like I’m fiddling with mathematical symbols, or playing with words, it is not tautological to say that taxes must come from C + I. They can also come from horded money. I probably wouldn’t be making this point had you not recognized the fundamental importance of hoarded money later in that same post.
17. January 2012 at 05:37
Daoist:
“it is not tautological to say that taxes must come from C + I. They can also come from horded money.”
Daoist, the government taxes transactions, not cash balances. If you accumulate cash over time, then the government taxed the transactions you made in the past that added to your cash balances in exchanges. Once you have the cash, they don’t tax it any more, so no, taxes cannot come from hoarded money.
Yes, taxes are sometimes paid out of cash holdings, but this is due to taxes being accrued over time in transactions, but were not paid at the time the transactions were made, so the tax payment was deferred to the future. But this is not the same thing as taxing hoarded money.
17. January 2012 at 15:30
DR, Have you read the longer Cochrane paper that DeLong attacked, not the one Wren-Lewis cites?
Tom, I address those issues in your comments to more recent posts.
Daoist, I don’t recall which argument. I might have used it to make a narrow technical point somewhere, but certainly not to criticize Wren-Lewis. I accept that fiscal stimulus might change GDP.
It’s probably best to view money as government debt, so it’s not net saving.
18. January 2012 at 13:12
[…] touch off the sort of cyberspace fireworks triggered by his series of posts (this, this, this, this, this and this) about Paul Krugman and Simon Wren-Lewis and their criticism of Bob Lucas and John […]