Get It Straight: Consumer Spending is *not* 70% of GDP

Ok, now I’m getting aggravated. On the front page of the NYT this morning, Peter Goodman wrote:

Given that consumer spending has in recent years accounted for 70 percent of the nation’s economic activity, a marginal shrinking could significantly depress demand for goods and services, discouraging businesses from hiring more workers.

And Martin Crutsinger of the Associated Press wrote

Especially in the U.S., consumer spending is essential: It drives about 70 percent of economic activity — more than for most European nations and well above the rates in developing countries such as China.

Both of these fine economics writers have fallen into a subtle but very important trap. They look at the category of GDP which the BEA calls ‘personal consumption expenditures’ and assume that it means what it sounds like: The money that persons, like you and me, spend on consumption.

But in fact, ‘personal consumption expenditures’ in the U.S. is a grab-bag category which includes all sorts of money—like Medicare spending by the government—which never passes through the hands of households. PCE also includes all the consumer goods imported into the U.S.—cars, computers, clothing, and the like—which create very little economic activity in this country.

In fact, by my very rough calculations, the money that people actually pull out of their paychecks and bank accounts to pay for domestically-produced goods and services drives about 40% of economic activity in this country. That’s still large—but the U.S. is nowhere near as dependent on consumer spending as people think.

Okay. Let’s look under the hood. The conventional thinking about the importance of consumer spending comes from this chart (all numbers are taken from the latest GDP release).

In this chart, PCE dwarfs everything else. If you just look at the BEA’s numbers you see that personal consumption expenditures in the second quarter were running at a $10 trillion annual rate, 70.7% of the $14.1 trillion GDP figure.

But when you actually look at the detailed breakdown of PCE, you get a much different picture. I divided PCE into five categories, like this.

Let’s start from the bottom of the bar. The first category includes household spending on goods and services which are primarily domestically-produced. That would be things like food, recreation, haircuts, utilities, legal fees, airplanes, auto repair, and so forth.

This category—roughly $4.3 trillion, or 30% of GDP—is all ‘pocketbook’ expense. Households lay out money, which primarily goes to support domestic production and employment.

But then we come to the second category: Import-intensive goods. These are items such as clothing, personal computers, cell phones, televisions, toys, sporting goods, cars, gasoline, and so forth. These are items where a substantial amount of production is done abroad, either directly or indirectly.

For such import-intensive goods, a $1 of consumer spending does not correspond to a $1 of domestic activity. If you buy a shirt or a laptop which is made overseas, much of your money supports economic activity in China or Taiwan, not the U.S. This category is worth $1.7 trillion, or 12% of GDP.

[Important note: This is a very very rough breakdown]

Now we come to the third category of PCE—“imputed services.” What this means is that the BEA assigns a number to certain economic activities, even though no money actually changes hands. The two most important imputed services are “imputed rental of owner-occupied nonfarm housing” and “financial services furnished without payment”. Respectively, these are the money you supposedly pay yourself to live in your own home, and the money you supposedly pay the bank for such services as free checking (by accepting lower or no interest on your demand deposits).

This category—worth $1.5 trillion or 11% of GDP—does represent real economic transactions. However, households don’t have any direct short-term control over this spending. You can’t say to yourself, oh, I’m going to pay a little less rent to myself this week, and use less of the house. Or you can’t say to the bank, oh, I’m not going to write any checks this week, so why don’t pay me a bit more interest.

To put it another way, payments for imputed services don’t directly drive economic activity. They are basically entries which make the government’s economic books balance.

Now we come to a wonderful category—healthcare goods and services, including hospitals, drugs, doctors, nursing homes, and health insurance. Because of the vagaries of national income accounting, most of the money that the government pays for Medicare and Medicaid, and that businesses pay for employer health insurance, shows up in the PCE category.

To put it another way—if Medicare pays the hospital $25 K for your father’s knee replacement, that money shows up as personal consumption expenditures. If your company health plan pays $30K for the birth of your son—that counts as PCE, even though you never see the money.

The healthcare category totals roughly $2 trillion, or 15% of GDP. But in fact, only about (roughly!) 15% of healthcare spending is “out of pocket”. The rest comes from government or through employee health plans.

And now we come to the final catch-all category, which I have labeled “social services, religious activities, R&D, and other similar activities.” This category includes spending by religious groups, such as the Catholic Church. It includes community food and housing relief. It includes R&D spending by private educational institutions, like Harvard. It includes social advocacy groups, like Greenpeace. It includes (I’m relatively sure) spending by political parties—Democrats and Republicans alike.

In other words, this wonderful category—totaling about $400 billion, or 3% of GDP—includes all sorts of spending which could be described as “social” rather than “individual”. And it’s funded by individuals, government, charitable contributions, and investment income.

So how does this all add up? The stories I cited at the beginning of this post wanted to argue that cutbacks in consumer spending was going to hold back growth. But in fact, the portion of PCE which is both under the control of households and drives domestic production, is really quite smaller than 70% of the economy.

How much smaller? Let’s make a back-of-the-envelope calculation. Let’s assume that $1 spent in category 1 corresponds to $1 in GDP. Simple enough.

But what about the category of import-intensive goods? I’m going to assume that every $1 spent in this category corresponds to only $0.50 of GDP. [Important note: This number is a very rough estimate, based on my read of input-output tables, import penetration ratios, and sunspots. Anybody who wants to come up with a better estimate is welcome to give a try]

Category 4, healthcare, I estimated that out of $1 in spending, only $0.20 was coming directly from consumers (that’s 15%, plus a fudge factor).

Category 3, imputed services, don’t correspond to actual household outlays at all. And category 5 may have some household outlays, but there’s no reason to think of religion donations and charitable contributions to universities, say, as part of consumer spending. In fact, as households cut back on ostentatious consumption, we could actually see an increase in charitable giving.

So when I added this all up, I got that households actually lay out about $5.5 trillion a year which drives domestic economic activities—about 40% of GDP.

Now, I had to make an insane number of heroic assumptions to get here. However, even if we make different assumptions, the conclusion is still true. Household outlays are an important driver of economic activity, but healthcare and ‘social’ outlays—driven by a different dynamic—are important as well.

Before it's here, it's on the Bloomberg Terminal.