There is something very strange about a counterfactual conditional in which the counterfactual is logically impossible. There are no possible worlds in which something logically impossible happens, so we can't ask which of the possible worlds in that subset is most plausible. The subset is empty.
"If saving were bigger than investment, what would happen to income?"
That's a bit like asking: "If 6 were 9, what would happen to income?".
(Assume closed economy, "saving" means "national saving", etc.)
People's actions have to all add up, but their planned actions don't have to. I can plan to buy 6 apples from you, and you can plan to sell 9 apples to me. And if our plans don't add up, at least one of us will be unable to carry out his plans. We can ask what would happen next if that were the case. But we can't ask what would happen if I buy 6 apples from you and you sell 9 apples to me. It doesn't make sense. It's logically impossible.
Now, if you asked instead: "If planned saving were bigger than planned investment, what would happen to income?", that question makes sense. We are not being asked to imagine something that is logically impossible.
"Planned investment" is a simple concept. It's how many newly-produced investment goods we plan to buy. "Desired investment", or "quantity of investment goods demanded" are equally good ways of saying the same thing.
In a Keynesian model with deficient aggregate demand, where quantity sold is determined by quantity demanded, "planned saving" is a trickier concept. It means "the income we expect to earn from the sale of newly-produced consumption and investment goods, minus the quantity of newly-produced consumption goods we plan to buy". It can be slightly misleading to talk about "desired saving" rather than "planned saving", because the income we expect to earn in these models is (usually) less than the income we desire to earn. Because we expect we can't sell as many newly-produced consumption and investment goods as we desire to sell.
This is what I think Paul Krugman should have said, or meant to say, with [my edits]:
"We’re not forgetting that in the end [planned] savings must equal [planned] investment; we’re just doing the first step in a comparative statics exercise. Suppose that I want to ask how, say, a fall in housing wealth affects the economy. First I ask how much this would increase [planned] savings, holding GDP constant; then I ask how much GDP has to fall to restore the equality between [planned] savings and [planned] investment.
......
In the end, of course, [planned] savings must equal [planned] investment, which is why GDP must fall from Y1 to Y2."
Actual saving and actual investment are always equal. They aren't just equal in the end, or when GDP adjusts to make them equal. They are the same thing. But planned saving and planned investment only become equal over time if people adjust their expectations and plans to make them equal.
No big deal.
Please, let's all just stop talking about "saving". Talk about the demand for stuff, and its flip-side, the demand for money. That's where the action is.
I'm beginning to forget where this argument all started.
Personally, I think Krugman's post is clear enough without the edits.
Posted by: Kevin Donoghue | January 16, 2012 at 11:29 AM
I liked the video.
Posted by: Bill Woolsey | January 16, 2012 at 11:57 AM
Kevin: try this: "First I ask how much this would increase [C+I-C] savings, holding [C+I] GDP constant; then I ask how much [C+I] GDP has to fall to restore the equality between [C+I-C] savings and [I] investment." Do you find that clear?
Posted by: Nick Rowe | January 16, 2012 at 12:00 PM
Nick,
No, I don't find that at all clear. Is it supposed to be a paraphrase? Honestly, I've no idea what you're getting at.
Before he even mentions the dreaded S-word, Krugman explains the thinking behind comparative statics:
"It’s often helpful to do the analysis in two stages. First, you ask how some desired quantities would change holding the equilibrating variable constant; then you ask how that variable has to change to restore equilibrium."
Note the careful reference to desired quantities. He also supplies a diagram, so that we know he's thinking of the equilibrium as the point where two curves intersect.
In order to misunderstand him you really have to work at it. I'll say this for Scott Sumner: he works bloody hard at it.
Posted by: Kevin Donoghue | January 16, 2012 at 12:10 PM
Kevin: "Note the careful reference to desired quantities."
Yep. I did notice that. And he was doing fine up there. Then, in the bit I quoted, he suddenly forgot to include the word "desired". I can understand him fine if I stick the word "desired" back in. (I prefer "planned", but no matter).
"Is it supposed to be a paraphrase?" It is an exact paraphrase. The things I wrote in [..] are a literal translation of the word that immediately follows.
Posted by: Nick Rowe | January 16, 2012 at 12:32 PM
I see the words "Increase in saving" on the diagram, with a line indicating that this refers to the vertical shift in the S-curve. Not only that, but he spells it out: Here the savings curve shifts up from S1 to S2; the upward shift of the curve at a given level of GDP is what Wren-Lewis and I mean by the “increase in savings”.
The offending phrase "in the end" comes after all that.
Posted by: Kevin Donoghue | January 16, 2012 at 12:39 PM
Maybe if 30 years ago we hadn't stopped teaching Keynesian cross and injections-withdrawals , more people would be clearer on the concepts. It may have been barbarian times but we weren't confused in 1975. I feel so old...
Posted by: Jacques René Giguère | January 16, 2012 at 12:41 PM
Wrong. Planning is about the arrangement of production processes across time -- not simple either analytically or in practice. It's not simply about how much is spent. It's about which goods to use and which goods to stop using, whih goods to maintain, and which goods not to maintain, which lroduction length to extend and which to shorten, exectera.
All of the stuff at the heart of the problem of collectivist economic planning ....
Nick wrote.
""Planned investment" is a simple concept. It's how many newly-produced investment goods we plan to buy. "
Posted by: Greg Ransom | January 16, 2012 at 12:48 PM
In my view, "planned" is always understood in these discussions and the identity between realized saving and investment is ignored.
The arguments are always about how something brings planned saving and investment into equality. Anyone who says, but saving is always equal to investment, so why does output or interest rates have to adjust to bring them into equalty is being obtuse.
If there is a shortage of gasoline, the price needs to rise to bring quantity supplied and demanded back into balance.
We know that quantity supplied and demanded refer to desired amounts.
No one says that this all means nothing because purchases and sales are always equal by definition.
Posted by: Bill Woolsey | January 16, 2012 at 12:52 PM
Jacques: I feel old too. And I think we understood it better in 1975.
Kevin: Just like a demand curve doesn't show quantity bought as a function of P. It shows desired quantity bought as a function of P. And those two curves show desired saving and desired investment as a function of Y.
Posted by: Nick Rowe | January 16, 2012 at 12:52 PM
"And those two curves show desired saving and desired investment as a function of Y." Correct. But is there anyone who doesn't know that? If so, it's hardly Krugman's fault.
Posted by: Kevin Donoghue | January 16, 2012 at 01:48 PM
"I" is such an artificial construct and so unattached from the whole economic process, no wonder basing macroeconomics on it lead to pathology and predictive incompetence.
Posted by: Greg Ransom | January 16, 2012 at 03:16 PM
Yes, let's debate this all again!
It is necessarily the case logically that income always equals expenditure, and actual savings always equals actual investment. It is *not necessarily the case logically, although it may well be so in practice, that a reduction in desired consumption, relative to income, is the same as increased demand for money.
I can postulate an economy with no investment, where people's desired consumption is equal to A + bI, where I is income and b is less than one. As we all know, that means that total output will be A(1-b) in equilibrium. I can say this without bringing money into the discussion at all.
But what about the (1-b) share of the marginal dollar? If people don't want to consume it, they must want to hold it in the form of some asset, right? Not necessarily. It actually doesn't matter what people would do with that non-consumed share of income, because it never actually occurs. In equilibrium, all income is consumed. Out of equilibrium? Lots of things are possible. Maybe there's a gap between people's expected and actual income. Maybe there's a gap between their desired and actual consumption. In real-world economies we don't think people can be surprised by consumption above what they intended, but there's nothing logically impossible about it.
Now, it may also happen, as in the monetarist story and Keynes story (well, one of his stories) that in practice when there's a divergence between desired and actual expenditure and/or expected and actual income, that we will also see excess demand or supply for money. But that's a distinct claim. The statements "output fell because desired consumption fell" and "output fell because demand for money rose" are not logically equivalent (tho they may often be equivalent in practice.)
At least, that's how it looks to me.
Definitely agree that mid-70s macro handled this stuff better than "modern" macro, tho I'm much too young to remember it myself...
Posted by: JW Mason | January 16, 2012 at 03:29 PM
Doing this in a "barter economy" / pure goods construct which lays out coordination across time is completely different from doing this in a "measured with dollars" circular flow Keynesian or "macroeconomic" construct.
Case in point.
Alternative choices in HOW production goods are used across time implicate DIFFERENT streams of output and consumption (as I showed here in comments on another post).
How much is available for planned savings, planned consumption at different points in time -- and WHAT IS AVAILABLE difference depending on alternative production plans _given_ exactly the same starting stock of "savings goods" or "goods for investment".
THEREFORE, how "savings" comes into equality with "investment" depends on the alternative _uses_ production goods have been put, which is to say, what choices have been made dedicating theme to alternative production plans.
The "identity" depends on which plans are chosen -- it does _not_ depend on past aggregate measures. In other words, streams of savings cross time come need to come into equilibrium with any number of alternative possible stream of production and production output -- output which NECESSARILY differs depending on which plans have been chosen, e.g. in making the choice for longer production processes/superior output vs choosing shorter production processes/inferior output.
The Keynesian "S = I" construct gives the tenured economist the equivalent of a stroke making it impossible for him to even think these thoughts, just as stroke victims lose capacity to recognize faces -- just as a stroke victim can't recognize his wife, a professor of economics cannot recognize the logic of choice/problem of choice across time in the domain of production.
Bill writes
"In my view, "planned" is always understood in these discussions and the identity between realized saving and investment is ignored.
The arguments are always about how something brings planned saving and investment into equality. Anyone who says, but saving is always equal to investment, so why does output or interest rates have to adjust to bring them into equalty is being obtuse.
If there is a shortage of gasoline, the price needs to rise to bring quantity supplied and demanded back into balance.
We know that quantity supplied and demanded refer to desired amounts.
No one says that this all means nothing because purchases and sales are always equal by definition."
Posted by: Greg Ransom | January 16, 2012 at 03:31 PM
Kevin: "But is there anyone who doesn't know that?"
I would confidently guess there are people who will be reading this blog post (or who read Paul Krugman's blog post) who didn't know that. Yes.
Bill: "We know that quantity supplied and demanded refer to desired amounts. No one says that this all means nothing because purchases and sales are always equal by definition."
Again. Some people don't know that and do say that. And even more people will be unclear on whether a binding price ceiling increases quantity bought because it increases quantity demanded.
Greg: OK, you want to unpack "planned I", because you have a different model. And it's not at all simple in that model. Fair enough.
Posted by: Nick Rowe | January 16, 2012 at 03:32 PM
Spam filter is playing up again. Just rescued JW and Greg's latest. I will check periodically.
Posted by: Nick Rowe | January 16, 2012 at 04:44 PM
After trying to follow these arguments, I'm still stuck with the way my non-economist stubborn mind wants to see it, so here goes. Within a given set of GDP, two things are happening. One, production is taking place in ways that include the wealth of new elements (extraction, manufacturing, etc). Each income from that particular product reflects the places where wealth enters the picture, perhaps that is a wealth residual. Our income is our consumption capacity, and it does not change in monetary terms because it is derived from a monetary constant. What I want to understand is the path that set amount of money is able to take. Granted, the proces starts over again, but what's important is that it is constrained by the existing supply of money.
Posted by: Becky Hargrove | January 16, 2012 at 04:50 PM
JW: "The statements "output fell because desired consumption fell" and "output fell because demand for money rose" are not logically equivalent (tho they may often be equivalent in practice.)"
I think that's correct. Here's a very weird and non-Keynesian model in which it is true. The two goods are apples and leisure. You can't consume your own apples and have to barter them for someone else's apples. The harder you work producing apples, the more apples you want to eat.
Posted by: Nick Rowe | January 16, 2012 at 04:57 PM
Greg: in the simplest one good model, where you can either consume that good or else use it as a capital good at the immediate flick of a switch, we only have one production constraint. Cd+Id cannot exceed desired Y. Yep, with a more complex technology there will be many production constraints. And there will be long lags between flipping the switch and the mix of produced goods changing.
Just to get some very very basic idea, just assume that there is still one good, but a one-period lag in changing the level of production. That by itself would give some dynamics to the equilibrating process.
Posted by: Nick Rowe | January 16, 2012 at 05:10 PM
It is winter. If we are thinking of savings in value terms.
Consider:
(A) I put out bucket of water, let it freeze, then set it by the back porch, to use the next day for some lemonade.
(B) I set out a bucket of water, let it freeze, then store that bucket in the cave out back, packed in hay, to use during the summer to help keep food from going bad in the ice box.
I've saved the same thing -- a frozen bucket of water. I've invested the same thing. A frozen bucket of water.
The output is completely different, the time dimension is completely different, and the value of what has been produced with my investment is different.
Two different methods of "saving". Two different ways of "investing". Two different streams of output, two different value streams.
But both represent "S = I" ....
Posted by: Greg Ransom | January 16, 2012 at 05:28 PM
Yes.
Nick wrote,
"Just to get some very very basic idea, just assume that there is still one good, but a one-period lag in changing the level of production. That by itself would give some dynamics to the equilibrating process."
Posted by: Greg Ransom | January 16, 2012 at 05:30 PM
The virtue of my model is that if we want to understand a world like our world -- and the key coordination problem in our world -- a multi-production good model with a time structure helps us see what is going on in such a world and what the essential coordination problem is that the market must solve.
Does a one production good model do that? Does a representative agent model do that?
Does a Keynesian model do that? Does "S = I" help with that?
To say that those who control tenure and advancement in macroeconomics say it does, does not decide the matter, it merely begs the question .. and motivates another.
Nick wrote,
"Greg: OK, you want to unpack "planned I", because you have a different model."
Posted by: Greg Ransom | January 16, 2012 at 05:41 PM
OK, I give it a go.
I=S, that is the starting assumption, actually the definition of macro economics. The planning that the agent does is the measuring norm, it generates units of measure absolutely designed to make I=S.
When I != S, then the model is broken by definition. That is not bad, not the end, it is quite normal. It don't mean the economist is dumb, it means the economist told the accountant: "This is the model you use, use it until it breaks, then call me"
Can we make a model that never breaks? Dunno.
Posted by: Matt Young | January 16, 2012 at 05:54 PM
Nick, I've got to admit I don't understand anything of all this. So I apologize upfront. In the world where I dwell, savings is the lack of consumption and investment: A creature of the difference between income and all expenditures. As income goes, up or down, so to do consumption, investment and savings follow, with a lag in time and seldom in proportion to the change in income.
All private economic activity is an effort to profitably satisfy today and tomorrow's consumption. Public economic activity is not so constrained. It need not make a profit nor spend only up to it's income. In a nation with it's own currency, there is literally no constraint whatsoever. There is never a lack of work nor a lack of capital to fund that work.
Anyway none of this is original to me. Most of it comes from reading Keynes. So if this discussion fits my version of economic reality, I'd like to know. Not that I'm going to study economics to understand the econoomist's language. I'd just like to know. Am I being a heretic here?
And nothing is ever in equilibrium. Ever.
Posted by: beezer | January 16, 2012 at 07:32 PM
I see Greg Ransom, JW Mason and the other Austrians are arguing the Austrian position.
They have entirely missed the point, unfortunately.
Greg, first of all, an intertemporal model can still be subject to monetary shocks. Actually it is extremely easy to demonstrate a Keynesian money/demand led recession and an Austrian supply contraction in the same model. Actually making a model do this is what conviced me that recessions and contractions are both valid and distinct forms of economic malaise, with vastly different causes and solutions though.
To see what I mean, take the excellent graphical model of an economy with a nice Hayekian triangle from Time and Money by Roger Garrison, Production Possibilities and Loanable Funds Market, with Savings on the vertical axis and Investment on the horizontal axis.
Garrison assumes that the Loanable Funds Market always clears. I don't. Here's why.
Draw an arc on the Loanable Funds Market graph beyond the intersection of S and I. The area under this curve is the Money Supply. Now add the constraint that the S-I intersection has to be within or at the Money Supply; it can't go beyond that curve (no money for the S-I equilibrium).
Now reduce the money supply so that the S-I point is how beyond that curve, violating the constraint I just defined. Consumption and investment will contract so that Investment is now within the Money Supply curve, yet the economy is not in equilibrium, savings and investment are not equal, and investment and consumption have fallen. All due to a contracting money supply. This is a Keynesian, demand-led recession illustrated with an Austrian framework and fully consistent with an Austrian framework.
This is what an economy looks like when it is strangled by its own money supply, in other words it is a balance-sheet recession.
Nick's right that S isn't always equal to I, but he's wrong in saying that all recessions are demand-led and there cannot be supply-led contractions. You can invest your way out of a supply-led contraction, you can't invest your way out of a demand-led recession, you have to expand the money supply in order to exist a recession.
Posted by: Determinant | January 16, 2012 at 08:04 PM
Matt: "The planning that the agent does is the measuring norm, it generates units of measure absolutely designed to make I=S."
I'm afraid you lost me there.
beezer: "Am I being a heretic here?"
You have a different (heretical?) definition of the word "saving", but in my view yours is closer to being a useful definition if we are interested in short run macroeconomics. Your definition of "saving" comes closer to "hoarding money", which is what I think is important.
Posted by: Nick Rowe | January 16, 2012 at 08:10 PM
Beezer:
I also think that you are defining investment differently.
In personal finance income is used for either consumption, the purchase of risky assets, or the purchase of safe assets. Accumulating unspent money balances is an accumulation of safe assets.
Personal Finance Economese
Consumption = Consumption
Investment = Saving--purchase of risky assets
Saving = Saving--purchase of safe assets or
accumuliation of money
Posted by: Bill Woolsey | January 17, 2012 at 10:10 AM
Nick, thank you. You guys have been beating this to death, but it's paying off. I've often wondered about what S=I actually meant. Now I understand. It's true, by definition, always and immediately. Krugman's graphs are graphs of intentions ('planned' investment and saving) -- much like demand schedules.
Maybe we all knew that back in 75 but this bemused citizen had sure forgotten.
Posted by: Ian Rae | January 17, 2012 at 09:24 PM
Ian: And I thank you, for confirming what I guessed. I did say to Kevin and Bill (in earlier comments) that this did need saying, and did need repeating, precisely because not everyone knew this!
Posted by: Nick Rowe | January 17, 2012 at 09:33 PM
Late to this, but very interested, as you can imagine. Definitions!
Planned
Desired
Expected
Demanded
"=" = "synonymous"
"Planned investment" is a simple concept. It's how many newly-produced investment goods we plan to buy. "Desired investment", or "quantity of investment goods demanded" are equally good ways of saying the same thing.
So Planned = Desired = Demanded
"planned saving" ... means "the income we expect to earn, minus...
So Planned = Desired = Demanded = Expected
But:
"the income we expect to earn in these models is (usually) less than the income we desire to earn"
So Expected ≠ Desired
??
JW Mason: "It is *not necessarily the case ... that a reduction in desired consumption, relative to income, is the same as increased demand for money."
Seems right. Two scenarios:
1. Booming economy. People consume a smaller percent of income because they're investing instead -- starting new businesses and remodeling their houses.
2. Declining economy. They're consuming less because they want the future certainty/security of money "in the bank."
Nick replying to Beezer: "You have a different (heretical?) definition of the word "saving", but in my view yours is closer to being a useful definition if we are interested in short run macroeconomics. Your definition of "saving" comes closer to "hoarding money", which is what I think is important."
I like it. As long as it's understood that in aggregate we can't "hoard money". The stock is unaffected by the spending/saving proportion. Only banks and the government can "hoard" money, by declining to create it -- lend it or spend it, respectively. (Ignoring Fed machinations for the moment.) More personal/business "hoarding" just means reduced velocity. There's no need to talk about saving, just spending.
Posted by: Steve Roth | February 07, 2012 at 10:21 AM
Steve: I think "planned" is the better word.
When we come to planned saving, it is planned income minus planned consumption.
1. In a world where people expect to be able to sell less output than they desire, but are able to consume as much as they desire, conditional on that level of output sold, (because of a deficiency of aggregate demand), it can be misleading to talk about *desired* saving.
2. In a world where people expect to be able to sell as much output as they desire, but are unable to consume as much as they desire, (because of excess demand and shortages of goods, like in Cuba), it can also be misleading to talk about *desired* saving.
3. Only in a world of market clearing, where people can both sell and buy all they desire, can we talk about "desired" saving in the full sense of the word.
Posted by: Nick Rowe | February 07, 2012 at 10:33 AM
@Nick:
Just to add to the collection, I see that PK now has yet another modifier:
"we have an “incipient” excess supply of savings"
I don't call those "scare quotes," btw. I call them "so-called" quotes, and they can suggest various things. In this case I think they're saying "this is a vernacular term, not a term of art, and I'm acknowledging that here without really defining what I mean by it, just assuming that properly knowledgeable people will map it onto the body of theory that I prefer, and others will just let it slide."
I'm still befuddled by the seeming synonymy contradiction above.
Posted by: Steve Roth | February 08, 2012 at 08:53 PM