I was surprised at the amount of resistance to my recent claim that consumption is not a part of GDP, i.e. that consumption is not production. Here I’ll respond to two claims that don’t mean what people think they mean:
1. The BEA measures consumption when it is computing GDP.
2. Without consumption there would be no production.
Both are true, but these facts have no bearing on the issue. Yes, the BEA does measure consumption, but consumption is not included in GDP. Here’s how it actually works:
The BEA measures the production of all consumer and investment goods, which is GDP. (Actually, they separate out private consumer (C) and investment (I) goods production from public C and I production, with the latter called “G”, but that minor point doesn’t change anything.) Basically, GDP is the total domestic production of consumer and investment goods in a given period. That’s the textbook definition, and that’s what the BEA measures.
In order to determine the production of consumer and investment goods, they look at domestic consumption and investment spending, and then adjust for inventory changes and trade imbalances. That converts domestic spending on C and I goods into domestic production of C and I goods. But make no mistake, GDP does not include consumption; it includes the domestic production of consumer goods plus the domestic production of investment goods.
The second point is also misleading. Yes, consumption is the primary motivation for production, indeed almost the only motivation. But that doesn’t make consumption a part of production. Some commenters tried to imagine what would happen to production if there were no consumption. Yes, it would likely decline. But it’s also true that production would decline if there were no oxygen in the atmosphere. That doesn’t mean oxygen is a part of GDP.
All this is true regardless of whether or not you accept my critique of Keynesian economics. But if we are going to discuss the merits of Keynesian economics, it’s important to start with some facts on which everyone should agree—such as that consumption is not a part of GDP. GDP measures production, not consumption.
Peter Navarro and Wilbur Ross made the opposite mistake when they claimed that imports are a negative part of GDP. Actually, imports are not a part of GDP at all, as GDP measures domestic production.
This identity:
GDP = C + I + G + (X-M)
does more harm than good. It leads people into all sorts of fallacies, including protectionism and fiscal stimulus.
I much prefer this identity:
GDP = M*V
Neither identity has any causal implications, but at least the equation of exchange is less likely to lead to erroneous public policies.
READER COMMENTS
Jon Murphy
Aug 15 2020 at 1:36pm
This follow-up makes things crystal clear. Thank you!
Jon Murphy
Aug 15 2020 at 1:40pm
By the way, these two posts were very enlightening to me. I am altering how I explain GDP to my classes because of these posts. I’m also sharing them with my colleagues and encouraging them to be more accurate as well.
Thanks
Scott Sumner
Aug 15 2020 at 5:38pm
Thanks Jon, you made my day!
Dylan
Aug 15 2020 at 1:52pm
Thanks for the follow-up. This does help clear things up for me, but unlike Jon, not quite crystal yet.
Two questions:
1. I understand that GDP is supposed to be a measure of production (I mean, it’s right there in the name), but it’s one of the primary methods for deriving it the expenditure method, which is basically measuring consumption?
2. Still not clear on services like a haircut where consumption and production seem like they have to be equal?
Scott Sumner
Aug 15 2020 at 5:44pm
Yes, it’s estimated by first measuring consumption, but then subtracting out all consumption that is not produced domestically during the current period. That turns it into production.
Suppose I estimated housing production as follows. I started with housing sales. Then I subtracted out housing sales of homes that were not newly built. Then I added in newly built homes that were not yet sold. That’s analogous to how GDP is estimated.
On haircuts, I was making a different point. My getting a haircut doesn’t create income for me. It is produced by someone else, as you say, and thus produces income for that person. Whether it boosts aggregate income is another question. It depends on whether my decision to get a haircut “crowds out” some other type of production.
dylan
Aug 16 2020 at 2:54pm
Thanks Scott. Very helpful.
Matthew Waters
Aug 15 2020 at 2:33pm
Private investment only makes sense in expectation of future consumption. Or more precisely, investment goods are produced in expectation of producing future consumption goods.
Private investment as share of GDP shows remarkable consistency around 17%. The two most recent peaks above 19% had significant negative-NPV investment. The time in the 80s above 19% was followed by large amounts of bank failures.
https://fred.stlouisfed.org/graph/?g=ufcO
Private investment spending has this hard limit. So extraordinary propensity to save over consume has to find other outlets. Pure financial loans, such as credit cards or auto loans, can be used for outlets. Easiest outlet is government debt.
I agree with previous post that savings vs consumption only matters at zero bound. But high propensity to save has made zero bound much more likely in, say, a 2% inflation targeting regime.
Warren Platts
Aug 15 2020 at 5:13pm
I guess you could carry it on your inventory for a while, but if nobody consumes your production, you haven’t produced anything. One is tempted to say that while consumption per se is not production, it is consumption that transforms production into production.
I dunno.. Looks like if only we printed a lot more money, we could increase GDP! 😉
I am not sure about that. Consider the following formula:
Force = mass X acceleration
Does it have causal implications? Hume would probably say no. He might even call it an accounting identity. Indeed, this is the first thing that pops up on google:
On the other hand, the vernacular sense of ‘causality’ is if you do one thing and something else happens as a result, then we say the first thing “caused” the second happening. So we press the pedal to the metal, increasing the force on the automobile, and the acceleration happens to increase. Are we not entitled to say that increasing the force caused the increase in acceleration?
Similarly, a recession happens, roughly, when V slows down because of a paradox of thrift. Why can’t we say that increased saving caused V to slow down that in turn caused Y to decline? Meanwhile, increased government spending increases V, and the Fed prints more cash thus increasing M and then Y happens either grow again or at least stops declining. Why can’t we say the government and Fed interventions caused the recession to not be as bad as it otherwise would have been?
Warren Platts
Aug 15 2020 at 5:15pm
Actually, now that I think about it some more, I should have written:
Consumption transforms inventory into production.
Scott Sumner
Aug 15 2020 at 5:47pm
In that equation, I mean nominal GDP. And printing lots of money is more likely to increase nominal GDP than is running budget deficits.
Michael Sandifer
Aug 15 2020 at 5:14pm
I think your point in your first post was very clear.
James
Aug 15 2020 at 6:16pm
Scott,
Your original post had a seven word sentence for a title: Consumption is not a part of GDP. That sentence is false. The people who defined the procedure for computing GDP as an accounting variable came before us and decided consumption is a part of GDP.
Now if you actually meant to say something different, fine. Maybe you mean one or more of these:
The expenditure formula for GDP should not be understood as a causal relationship.
The production of consumption goods is a part of production but the consumption of those goods is not an act of production at all.
GDP as taught in school has flaws that lead to confusion.
National income accounting rules should not be taken as economic models.
All of these are entirely reasonable, but your initial claim was false and you cannot rescue a false claim by defending some different claim.
Jon Murphy
Aug 15 2020 at 6:38pm
The sentence is not false. As I think Scott proves rather conclusively, the claim “The people who defined the procedure for computing GDP as an accounting variable came before us and decided consumption is a part of GDP” is false.
I thought, and taught, like you did up until these two posts. But now I see that consumption is no more part of GDP than imports.
Scott Sumner
Aug 15 2020 at 10:35pm
Again, it’s simply false to claim that consumption is a part of GDP. GDP measures production, not consumption. They are two different concepts.
You said:
“The people who defined the procedure for computing GDP as an accounting variable came before us and decided consumption is a part of GDP.”
No they did not, as I explained in this post. They said the production of consumer goods is a part of GDP, and that’s correct. Production is growing a tomato. Consumption is eating a tomato.
Ike Coffman
Aug 17 2020 at 11:04am
Maybe the dictionary is wrong, but this definition certainly seems to support James view:
From Brittanica ( https://www.britannica.com/topic/gross-domestic-product )
In economics, the final users of goods and services are divided into three main groups: households, businesses, and the government. One way gross domestic product (GDP) is calculated—known as the expenditure approach—is by adding the expenditures made by those three groups of users. Accordingly, GDP is defined by the following formula:GDP = Consumption + Investment + Government Spending + Net Exports or more succinctly as GDP = C + I + G + NX where consumption (C) represents private-consumption expenditures by households and nonprofit organizations, investment (I) refers to business expenditures by businesses and home purchases by households, government spending (G) denotes expenditures on goods and services by the government, and net exports (NX) represents a nation’s exports minus its imports.
Scott Sumner
Aug 17 2020 at 1:30pm
But the equation you cite does not in fact measure the expenditure of these three groups, it measures total production. It it were measuring expenditure then why would it subtract imports, or add change in inventory?
James
Aug 17 2020 at 3:54pm
Scott,
The BEA says you are mistaken about whether consumption is a part of GDP.
Of course, that takes nothing away from the other very fine points you have made.
marcus nunes
Aug 15 2020 at 9:31pm
Still in the Gross Deceptive Partitioning spirit.
https://www.themoneyillusion.com/c-i-g-nx-grossly-deceptive-partitioning/#comments
Scott Sumner
Aug 15 2020 at 10:35pm
Always good to revisit that post.
Danny
Aug 15 2020 at 10:01pm
1) Does a barber produce a haircut and does the customer consume that haircut?
2) So when people say that GDP is 70% private consumption, doesn’t that just mean that 70% of the goods and services produced domestically were consumed by US citizens?
Scott Sumner
Aug 15 2020 at 10:38pm
Yes.
It means the amount we consume is 70% as large as GDP. Here’s an analogy. One might say that Social Security is 7% of GDP. But no one would claim that Social Security payments are a part of GDP, as they are not production. Saying something is “x% of GDP” is just a way of indicating size, not a claim as to what it is.
Daniel R. Grayson
Aug 16 2020 at 7:16am
“domestic spending of C and I goods”
“on”
“GDP does not includes consumption”
“include”
Warren Platts
Aug 16 2020 at 10:26am
It seems to me that at least part of the confusion is merely semantic, hinging on what ‘C’ stands for. We call it “consumption” and say “consumption is part of GDP”. But everyone knows that C = Cd + Cf, and so when we want to get pedantic we say “Only Cd is part of GDP”.
And everyone knows that when we say “only Cd is a part of GDP”, that is simply a shorthand, if imprecise, way of saying “only the goods & services produced domestically that are consumed by households are a part of GDP.” As long as that linkage is known, I don’t see the harm in saying “C is a part of GDP” where it is understood that the intended meaning is “the goods & services consumed by households, less the imported goods & services consumed by said households, are a part of GDP.”
Scott Sumner
Aug 16 2020 at 1:20pm
Thanks Daniel, I fixed those typos.
Warren, No, it’s not just semantics. Consumption is not production. Consider a domestically built good that is manufactured in Q1 and consumed in Q2. That good is not a part of Q2 GDP, it’s part of Q1 GDP. It’s the production of a good that makes it part of GDP, not the consumption of a good.
robc
Aug 16 2020 at 1:47pm
What happens if a good is produced but never consumed? If it is written off in Q3 is it subtracted from Q3 gdp? I assume they dont go back and correct Q1, which would be the correct thing to do as it was actually never production, just waste.
And a bigger issue I have is with home production. If I homebrew and consume a beer, that isnt counted in gdp (as far as I know), but it is domestic production ( and consumption). But if Bell’s brews it and I consume it, it counts. WTF!?! I would happy to be wrong about that, but I have never had my homebrewing surveyed.
Jon Murphy
Aug 17 2020 at 10:50am
Home production isn’t counted in GDP. GDP relies on market transactions of final goods and services. Since your home-brewed beer isn’t sold on the market, it’s not counted.
For this reason (and others), I do not refer to GDP as a measure of economic well-being. It can be proxy for such, but GDP does not measure well-being. It just measures domestic production of goods and services.
robc
Aug 18 2020 at 9:31am
But shouldnt it? It is domestic production. Why should it not be counted? A huge amount of production is missed due to this.
robc
Aug 18 2020 at 9:37am
The grains, hops, yeast, and water that are being counted in GDP are **NOT** final goods. They are intermediate goods, the final good is the beer I produce. But it is the intermediate goods that are getting counted in GDP.
Scott Sumner
Aug 17 2020 at 1:32pm
Stuff produced but never consumed eventually ends up as depreciation. That’s why it’s called “gross” domestic production, NDP subtracts depreciation.
robc
Aug 18 2020 at 9:34am
It seems that NDP is a better measure. Why all the focus on GDP?
sk
Aug 16 2020 at 1:53pm
Since the C=I=G =GDP is wrong, why does it continue to be taught? Not a trick question !
Also, we continually here we are a consumer based economy
Scott Sumner
Aug 17 2020 at 1:33pm
It’s true for a closed economy. Most books include the export sector.
art andreassen
Aug 16 2020 at 4:43pm
Scott: Aren’t both equations: GDP=C+I+G+(X-M) and GDP+M=C+I+G+X equal? Doesn’t the second equation better describe the true picture of the economy? C+I+G+X all contain imports, GDP does not. Did I miss something, aren’t exports domestically produced? Subtracting M from X removes M from calculated GDP and from X but not from the other components. Thus you are saying GDP is a measure of domestic consumption not domestic production and you are right. Saying C is 70% of GDP is an arithmetic sleight of hand that makes it appear that M has no impact on C or the non adjusted components. Actually C adjusted for imports divided by GDP is really 60%, as is C/(GDP+M), but who really cares?
Jon Murphy
Aug 16 2020 at 4:55pm
No. The second would double-count. M is a factor in C, I, and G.
art andreassen
Aug 16 2020 at 6:00pm
Jon: Don’t you see that the first equation removes M from GDP but only from X and not from C, I, and G. Neither double counts. The seconds add M to GDP and keeps X on the component side. Netting imports and exports does not remove imports from the remaining components which you admit they still have embedded in them. In the second you are dividing an input adjusted GDP into import containing components thus comparing apples to apples. Exports are part of domestic production.
Jon Murphy
Aug 16 2020 at 6:26pm
Not quite. C, I, and G all have import components to them. C is Cd (Consumption of domestically-produced goods) and Cf (Consumption of imported goods). Same with I and G. Since we only really have total trade data (exports and imports) and nothing more precise to say what goods consumed were made domestically and which were imported, we subtract out imports. I have a short handout I give to my classes on this.
art andreassen
Aug 17 2020 at 7:44am
Scott, Jon: I see my error. Subtracting (X-M) makes GDP a measure of production and consumption in total but not by individual component. In order to get the correct percent that each component is of total production and of total consumption you have to remove the imports in each component. Took me a while to realize that. It’s the constant use of the incorrect expression that PCE is 70% of GDP that gets me since imports have been removed from the GDP total but not from the components..
Capt. J Parker
Aug 17 2020 at 6:50pm
@Jon Murphy
I’d be very curious to know if any of your students raised either of the following objections to your handout on how trade deficits affect GDP:
1) What is it about opening the border to trade that would suddenly enable consumers to double their total consumption expenditures, C ?
Or
2) If one is asking the question: how does GDP change if we introduce a trade deficit, all other things being equal, then how, exactly, would doubling consumption expenditures at the same time as we introduce a trade deficit qualify as all other things being equal?
Jon Murphy
Aug 18 2020 at 9:40pm
Parker-
Neither question is particularly relevant. They are both hypothetical. Thus the footnote in the handout where I state these are absurd assumptions but just done to make the point. Including more realistic numbers would just serve to add no insight while obscuring the lesson, something that is never good in pedagogy.
Capt. J Parker
Aug 19 2020 at 4:03pm
@Jon Murphy
My gripe is that you make one point reasonably well namely, the dollar value of imports are not something that is part of the definition of gross domestic product.
But when you say:
You are alluding to a different and important point which is GDP=C+I+G+(X-M) is an accounting identity and not theory about how the economy works. Giving a numerical example where M and C increase by the same amount and then declaring: see M has no primary effect on GDP seems like a poor way to teach this latter point.
Ike Coffman
Aug 17 2020 at 9:41am
I do not think the equation GDP = M * V is correct because those two things are related. If we print money that is not spent velocity goes down, although by what ratio I have no idea. I am not saying that the equation is totally wrong, but because those two variables are not independent the equation is certainly not complete. We have a healthy economy when GDP is growing.
I agree with all the other commentators who say that consumption is the point of production regardless of how it is counted. A bigger problem that I have is when investment is counted with the same weight as other variables. If there are only two things you can do with income, spend it or save it, then you have to look at what is more productive. If the point is to increase GDP so we can increase incomes, which of those two (savings vs spending) is more productive in doing so?
Now I know that savings is not the same as investment, but it is generally assumed that putting money in the stock market is considered an investment. Increasingly, that is what people are doing with their savings, and what I see is that the increasing conversion of savings to stock market investment is leading to share price inflation. Investment is supposed to lead to increased opportunities for economic growth (however you measure it), and at some point there is just not much real growth to be found. And right now there are these huge capital flows into the economy driven by the Fed.
It seems to me we are entering a phase of economics where we are trying to justify the huge growth in government spending by saying that since we don’t see the inflation we predicted then government spending is not as bad as we thought.
I propose that a curve exists, similar to the Laffer Curve, where peak economic growth occurs at some ratio of spending vs. savings. If no spending occurs GDP has to be zero which implies that a reduction in spending will reduce GDP, if no investment occurs growth will be zero. Call it the Coffman Curve. We have been assuming that more and more government debt leads to inflation, but I think we have been looking at the wrong type of inflation. Instead of consumer goods and consumption inflation, I think we need to look at stock market inflation. My belief is that we are way, way beyond the point where savings leads to investment, and I think the inflationary effect is being driven by government debt. This has to be bad.
Jon Murphy
Aug 17 2020 at 10:48am
“Investment” in the GDP formula is business purchases of domestically produced goods and services. If you put $100 in the stock market (which is to say, you give a company $100 to become a part-owner), that does not show up in GDP. When that company buys a computer for $100, then it shows up in Investment.
Ike Coffman
Aug 17 2020 at 11:14am
You might be right, but if Scott can redefine C then I get to redefine I. Anyway, Wikipedia kind of supports my position: In some research, investment is modeled as an increasing function of the gap between the optimal capital stock and the current capital stock. Here the optimal capital stock is modeled as that which maximizes profit. (https://en.wikipedia.org/wiki/Investment_(macroeconomics)) and it is still my contention that the money in the economy chasing investment is much, much larger than than the amount which would maximize profit.
Jon Murphy
Aug 17 2020 at 11:34am
He’s not redefining C. He’s clarifying what “C” stands for. There is no redefinition going on.
Scott Sumner
Aug 17 2020 at 1:35pm
You said:
“I do not think the equation GDP = M * V is correct”
False; V is DEFINED as GDP/M
So it’s correct.
Ike Coffman
Aug 18 2020 at 12:50pm
I am not really here to argue with you, so don’t misunderstand me, my point was that I do not think GDP = M * V is a good alternative to the traditional equation because V is not an independent variable. With that definition you are basically saying GDP = GDP. How does that help?
Ike Coffman
Aug 18 2020 at 2:42pm
I apologize. I came here for a specific reason, and I got sidetracked. I would delete a bunch of my entries if I could.
I came here because I want to get expert (or at least knowledgeable) information about the government (Fed, executive, legislative) response to the economic shock of the corona virus. My position is that the government is totally screwing this up, but what do I know?
What makes me really mad is not that the response is ineffective, it is that the response is effective in supporting the already wealthy, who don’t really need help, without helping those who are really badly affected. Those low wage service economy workers have gotten the short end of the stick for what, 30 years now, and government is basically saying sucks to be you.
Using the same formula Scott uses (and in this specific case there is not only nothing wrong with it, it works perfectly for what I want to say), in my opinion a much better government response would be to prop up GDP by increasing the velocity of money. Unfortunately this would take a legislative solution, which is currently deadlocked. The FED has powerful tools, and because those powerful tools (in this case more quantitative easing, raising M) are effective in propping up the stock market, the rest of government says they don’t have to do a thing. My contention is that this is lowering V, which makes the Fed response inefficient.
Money needs to go more directly to the people who need it, who will spend it quickly, raising V without lowering M, and the Fed doesn’t have any good ways to do that.
If I am wrong, please tell me.
Capt. J Parker
Aug 17 2020 at 2:07pm
Economists flit effortlessly between the real and the nominal and then when we mere mortals become confused, as does the magicians audience when they lose track of which cup the real (or is it nominal) ball is hiding under, in they swoop to reveal to us our foolish error. Should we then believe we now understand how we were fooled, they stand ready with an even grander illusion, the misdirection for which they have been curating from the beginning.
Capt. J Parker
Aug 17 2020 at 9:24pm
Regarding Navarro and his supposed basic error in using an accounting identity to determine causal influences in the economy:
Navarro argues, in part, that the U.S. Tax structure, with higher taxes on corporate income, labor and personal income but not on consumption, put US firms at a trading disadvantage relative to trading partners that rely primarily on VAT taxes. His argument isn’t novel and there is plenty of support for it.
Dr. Sumner calls Navarro’s ideas on trade a complete mess. But Sumner doesn’t address Navarro’s claim about how the US tax code together with trading partner VAT taxes hurt US exporters and aid foreign importers. Instead Dr. Sumner addresses a different claim, namely that a VAT in and of itself is an export subsidy if it is levied on imports but not on exports. That claim about a VAT is true but it is distinctly not what Navarro was arguing.
Capt. J Parker
Aug 19 2020 at 4:01pm
@Jon Murphy
My gripe is that you make one point reasonably well namely, the dollar value of imports are not something that is part of the definition of gross domestic product.
But when you say:
You are alluding to a different and important point which is GDP=C+I+G+(X-M) is an accounting identity and not theory about how the economy works. Giving a numerical example where M and C increase by the same amount and then declaring: see M has no primary effect on GDP seems like a poor way to teach this latter point.
Comments are closed.