Consumption is not a part of GDP
By Scott Sumner
I used to teach my students that consumption was roughly 2/3rds of GDP. Actually, consumption is not a part of GDP.
Consider this analogy. You have an annual income of $100,000. You might say:
1. Income = wages + capital income
You might also say:
2. Income = consumption + saving + taxes
And you can make the second equation more complex by breaking consumption down into spending on clothes, haircuts, restaurant meals, etc.
It makes sense to say wages and capital income are both a part of income. On the other hand, to me it doesn’t make much sense to say spending on haircuts is “a part of” income.
Of course that’s a question of semantics. So let’s look at some statements with real implications:
1. My decision to earn more wages will cause my income to rise.
2. My decision to get more frequent haircuts will cause my income to rise.
The first claim is obviously plausible (assuming I find employment), while the second claim would require some pretty fancy footwork to justify.
In the textbooks, GDP = C + I + G + (X-M)
That makes it look like consumption (C) is a part of GDP. But it isn’t—consuming adds to C and subtracts from I (inventories). GDP measures aggregate production, and consumption is obviously not production. If I go and spend $100 on a new product, it doesn’t cause GDP to rise by $100, even if the good is domestically built. Rather investment falls by the value of the good, as inventories fall. (GDP may rise slightly, as my decision to buy the good boosts retailing services.)
The decision to consume more only causes GDP to rise if it causes production to rise.
Now it’s certainly possible that my decision to get up off the couch and go buy something does in fact cause GDP to rise. But if GDP does rise, it is for reasons that are completely unrelated to the myth that consumption is a part of GDP. Consider the following analogy. Getting a haircut might make my income go up because I look more presentable and find it easier to get a job. But in that case the reason my income goes up has nothing to do with the claim that my consumption of haircuts is “a part of my income”. It’s a use of my income. There has to be some other indirect effect, which needs to be explained.
In order for consumption to boost GDP, we need two assumptions to be met. First, more consumption must boost nominal GDP. Second, more NGDP must boost real GDP.
The first assumption requires an incompetent central bank. The Fed is supposed to keep NGDP growing at rate that best allows it to meet the dual mandate. Even if the optimal NGDP growth rate is not constant, it surely doesn’t depend on whether you decide to go get a haircut. There’s some growth rate in NGDP that best meets the Fed’s mandate. For your decision to get a haircut to boost NGDP, the Fed must respond incompetently, allowing aggregate demand to go off course. That might happen, but in that case NGDP is not rising because “C is part of GDP”. Perhaps velocity rises by more than M falls—they screw up.
Even if NGDP does increase, there’s no necessary increase in RGDP. If more NGDP always did cause RGDP to rise, then Zimbabwe would be the world’s richest country. For an increase in NGDP to boost RGDP you also need the assumption of sticky wages.
So it’s just possible that consumption does boost GDP, but not for the reason you were taught in school. If we were honest with students we’d say, “Spending more might cause Jay Powell to screw up while doing his job at the Fed. This causes the total dollar value of spending (NGDP) to rise. And because people have money illusion, they are fooled into thinking that work has become more lucrative, and so they decide to work more. Output (RGDP) also rises.
I think you see why we don’t teach students this way. It’s confusing. So we make up a fairy tale. But the fact that students are taught the wrong explanation means that they never really learn macro, even if they are absolutely straight A students, the best in their university. How can they learn what actually happens if they are taught that consumption is part of GDP, so spending more causes GDP to rise?
You might argue that Jay Powell doesn’t know when I go out to buy something, so the lack of monetary offset is a plausible assumption. But the consumption shocks that matter are not the whims of individuals (which balance out due to the law of large numbers), they are the collective decision of millions of people to spend more, perhaps due to a federal tax rebate. In 2008, the tax rebate did boost GDP a bit in the second quarter, but the Fed responded with tighter money and it led to lower GDP in the 3rd and 4th quarters. For the year as a whole, the tax rebate added nothing to GDP. Students wouldn’t know that based on the models they are taught in macro. It seems like a policy that gets people out shopping should boost GDP. It doesn’t.
The only even halfway plausible argument that consumption boosts GDP occurs at the zero bound, where there is some debate as to whether the Fed offsets consumption shocks.