Bloomberg Anywhere Remote Login Bloomberg Terminal Demo Request


Connecting decision makers to a dynamic network of information, people and ideas, Bloomberg quickly and accurately delivers business and financial information, news and insight around the world.


Financial Products

Enterprise Products


Customer Support

  • Americas

    +1 212 318 2000

  • Europe, Middle East, & Africa

    +44 20 7330 7500

  • Asia Pacific

    +65 6212 1000


Industry Products

Media Services

Follow Us

Bloomberg Customers

What Really Happened to Consumer Spending

Posted by: Michael Mandel on April 02

Here’s a more detailed version of the analysis in my latest BW story. Right now it looks like personal consumption expenditures have fallen by only 0.4%, or $40 billion, over the past year (that’s nominal dollars). For an economy the size of the U.S., $40 billion is a rounding error.

But that number didn’t seem right, given the devastation in the economy, so I took a closer look. I started with an advantage—I’ve written about these numbers before. When we hear the words ‘consumer spending’ or even ‘personal consumption expenditure’, the image that comes to our mind is a household buying food, clothing, televisions, haircuts, cars, etc etc—all the things we buy in our daily lives. Let’s call these ‘pocketbook’ expenses—we make the decision about whether to reach into our pocketbook or wallet and pay for them.

However, there are a lot of categories in the BEA’s definition of personal consumption expenditures which are not ‘pocketbook’ expenses. In particular, there are four important categories in PCE which shouldn’t arguably shouldn’t count as personal consumption.

1. Spending by religious groups, foundations, and other nonprofits. The BEA puts these into PCE because they don’t fit anywhere else, but they are clearly not household spending. For example, if your church or synagogue around the corner spends more on helping the poor, that counts as consumer spending. Another example: the Bill and Melinda Gates Foundation is raising its spending by 15%, or $500 million this year. To the degree that the money is spent in the U.S,, that shows up in consumer spending.

2. “Imputed” spending, where no cash changes hands, but the statisticians book an expenditure for technical reasons. The biggest of these is “owner-occupied nonfarm dwellings—space rent” –in effect, the rent you supposedly pay yourself for the privilege of living in your home. To put it another way, the consumer spending numbers charge Americans $1.1 trillion for living in their own homes—and that number is rising

3. Money spent by Medicare, Medicaid, and employer health care is mostly counted as part of PCE, even if the payments go directly to the hospitals or doctors and never pass through the hands of households. So to put it, if a senior citizen goes to the hospital and gets a heart transplant, which is completely paid for by Medicare, that is counted as part of consumer spending.

4. Spending for education and research. This has some overlap with category 1, but my main problem is that spending on education on R&D should really be counted as investment, not consumption. Treating this as part of consumer spending gives you the weird result that if a private nonprofit university steps up research, that counts as an increase in consumer spending and a fall in the savings rate.

I added together these four categories of PCE. For healthcare. I took 85% of the total medical spending in the data, because about 15% of health care spending is “out-of-pocket” for consumers.

Altogether, these four categories account for $3.7 trillion, or more than one-third, of so-called consumer spending. Remarkably, these non-pocketbook expenses rose by 4.5% over the past year.

Meanwhle pocketbook spending—all the things that we normally think of as consumer spending—is down by 3.1%, or $200 billion, over the past year. This is by far the biggest year over year plunge since 1959, when the current data series starts. In fact, measured on a year over year quarterly basis, the fourth quarter was the first time that pocketbook spending fell, in nominal terms.

What about the savings rate? That’s a bit trickier. If we want to calculate the actual saving rate of households, we have to remove all the spending that doesn't really belong. But because of the way the numbers are constructed, we have to remove the corresponding income from the personal income side as well. For example, Medicare is counted both as a transfer payment to households--adding to their income--and spending by households. Similarly, the rent of owner-occupied housing shows up both on the spending side, and then buried deep in the numbers on the income side as well, because you are paying rent to yourself.

So what I did was subtract non-pocketbook spending from both the personal outlays, and disposable personal income. Total savings, in dollars, stays the same, but the amount of income goes down. Hence, the savings rate in February rises from 4.2% to 6.4%.

The savings calculation should be thought of as a rough approximation. I didn't worry about the effect on taxes. And starting with this summer, the BEA will be breaking out data on nonprofits on a regular basis (right now the breakdown is annual with a lag). That will enable better calculations.

But there's no doubt that in terms of the money households actually control, they have cut back a lot more than the raw PCE numbers show.

I welcome any and all comments on anything I should have done different.

TrackBack URL for this entry:

Reader Comments

Tom E.

April 2, 2009 08:32 PM

Wow, good detective work. If they are working with these kind of numbers no wonder why government leaders seem clueless about the economy.

I suppose you would also have to look at whether people are paying down debt (which would be like saving) or adding to it.

This is a good test to see if government numbers are correct. The last couple of months most people stopped buying anything not essential. If the numbers hardly budge, then something is wrong.

Noni Mausa

April 3, 2009 07:52 AM

I am curious-- how much of the remaining consumer spending, the pocketbook spending, goes to things which cannot be reduced, or not easily? I am thinking especially of rent/mortgage, utilities, and transportation.

Something no-one discusses, or not that I ever hear, is that the ability of people to spend is more or less fixed. There is a finite collective pocketbook out of which all these purchases and payments are made.

Quite a lot of the pocketbook is already spoken for the instant the paycheques clear the bank, leaving a smaller liquid stream of income which people can truly choose to allocate, whether through purchases or savings. Most businesses, I assume, are competing for this much reduced stream of currency.

As the percentage of the consumer dollar which is spoken-for rises, that stream necessarily shrinks. It cannot be increased, only diverted this way or that (unless wages rise or fixed expenditures are somehow reduced.)

To a conservative viewpoint, the solution to this bottleneck seems paradoxical. Three efforts would increase the cash available for discretionary spending immediately and permanently -- single payer health care, public transit, and increased wages. All three together would add an effective $8000 to $15,000 to a median household income, plus allowing people the security to relax and spend the newly available income rather than squirreling it away.



April 3, 2009 09:57 AM

I'm very curious by what means the 4 categories you pointed out have increased by 4.5%. But these kinds of findings tend to really reduce my faith in reported statistics. Were consumption figures rolled up in this way to tell a particular story? If so, by whom and what was their angle? And no wonder that personal consumption is reported as 2/3rds of GDP!


April 3, 2009 10:47 AM

Tom E.

Read what administration economists are saying. They mention weakness in household spending repeatedly. The fact that you were unaware of how the data are put together doesn't mean that Summers and Romer are unaware.

Given that you were unaware of what Mandell points out, and made a very bad assumption about who else might have been aware, I would warn you away from the sort of simplistic rule-of-thumb you suggest in your third paragraph. Unless you have a grasp of the accounting involved, you aren't going to be able to put together a reasonable test of that accounting.


You say that government spending on household consumption arguably shouldn't count as household consumption. This is a matter of whether we want to count money spent on or by households. There are any number of ways in which government economic accounting isn't what any one of us may have chosen to do, but the accounts close (in theory and with allowance for error). Getting the books to balance is an important feature in accounting, so I wouldn't be hasty with the "shoulds". It would be a valuable addition to break out spending "by" households, but from a welfare point of view, spending "on" households is also an important concept. I'd recommend doing the exposition and leaving out the rest.

Tom E.

April 3, 2009 01:25 PM

Kharris: by government officials being clueless, I guess I meant more the previous administration telling us everything was fine when you could look around and see things falling apart.

There seems to be two sets of numbers, one for government insiders and another sugar coated set for the public.

I have the feeling that we have been fed misinformation on the economy in order to protect ideology (free trade, no regulation is always good) and to protect an economic model that appears to benefit the wealthy at the expense of everyone else in the US.

I am not an economist, just someone who does not like being told its sunny outside, when I can look for myself and see that it is dark out.

Mike Mandel

April 3, 2009 02:22 PM

Kharris: The key is who makes the decisions. If we don't know who makes the decision, we don't know where the incentives are applied.

CompEng: For example, Medicare spending is increasing, and that's how it shows up in PCE


April 3, 2009 03:20 PM

I guess where this line of thought is leading me is whether the rise is a stimulus effect, part of our safety net (increased Medicare spending could signal), people skipping the workforce to go back to school, or if some areas such as the imputed spending make this partially an accounting anomaly.
Knowing those answers, you could make a pretty good guess how well that part of spending will hold up, and therefore how much worse that part of the economy will get and when. It might also help to figure out the bang for the stimulus buck we'll be getting.


April 3, 2009 06:25 PM

Are the savings rates you quote derived values based on your mathematical formulas and assumptions, or actual value based on deposits? The reason I ask is that perhaps the drop in consumption is due to a drop in household income. In that case, people aren't forgoing expenditures and saving that money, they are forgoing expenditures because the money no longer exists in their household revenue. In that case, the actual savings rate is unchanged. The money isn't there to spend or save.


April 3, 2009 06:27 PM

Are the savings rates you quote derived values based on your mathematical formulas and assumptions, or actual value based on deposits? The reason I ask is that perhaps the drop in consumption is due to a drop in household income. In that case, people aren't forgoing expenditures and saving that money, they are forgoing expenditures because the money no longer exists in their household revenue. In that case, the actual savings rate is unchanged. The money isn't there to spend or save.

Patricia Shannon

April 3, 2009 09:31 PM

The previous administration(s) have made definitions that make the economy look better than it is. If the current or future administration starts using more realistic definitions, it will make them look worse.

The Mad Hedge Fund Trader, San Francisco, CA

April 7, 2009 10:25 PM

Call me an optimist, but I am starting to see crocuses of economic recovery busting out all over. Long side traders now have a spring in their step after a 23% rise in the Dow in three weeks, the best move since 1938. If it is true that the stock market anticipates moves in the real economy by six months, then a lot of managers are going to come back from their summer vacations in September to find a surprising batch of new orders. Commodities have been on an absolute tear this year, especially oil and copper, classic harbingers of future business activity. Just look at my favorite, Freeport McMoran (FCX), which soared 170% from the November lows. The downward momentum of a whole range of economic indicators is slowing. The durable goods number was actually up last week! The mother of all inventory adjustments is almost over. Retailers are still offering the deals of the century, but there is very little left in the back room. The Baltic Dry Shipping Index has tripled off of its November low, hinting that international trade may come out of its comatose condition. And they are no longer looking for organ recipients for the major airlines. Now I hear that semiconductor makers are expected to make their Q1 targets. Maybe it’s because spring has arrived, and the girls on the Embarcadero have shed their overcoats for low cut tank tops. Or maybe the $4 trillion in global stimulus is starting to have its desired effect.

John K

April 8, 2009 04:58 PM

Michael --

Interesting analysis, but the adjustment for owner-occupied rent has me a bit confused. Are you simply subtracting the entire amount of "Owner-occupied nonfarm rent" (NIPA Table 2.5.5 Line 24) from both disposable income and PCE? I had thought that only imputed rent net of expenses (e.g. mortgage interest) and capital consumption ended up on the income side in "Rental income of persons with capital consumption adjustment" (NIPA Table 2.1, Line 12). If that is the case, subtracting the entire amount of imputed rent (rather than just the net) from income would significantly overstate the saving rate, would it not?

Have you been able to run this analysis all the way back to 1946 or 1950? Looking at the longer-term averages would be particularly useful if you think the last 15 months might be enough to scare our consumers back to even average thrift. Based on the BEA stats, PCE (even properly focused on "pocketbook" expenses) could have another big leg down if average saving rates prevail any time soon.

John K

April 13, 2009 11:51 AM

Michael --

After taking a minute to look into this, I have to disagree with your treatment of imputed rent. Take a look at Mayerhauser & Reinsdorf, "Housing Services in the National Economic Accounts" published by the BEA, updated 9/11/07. Imputed owner-occupied rent is comprised almost entirely of items that are difficult to characterize as anything other than "pocketbook" expenses, namely: property taxes, mortgage interest, and maintenance goods and services. Furthermore, the only number that seems to show up in disposable personal income is, appropriately, a net figure that takes the imputed owner-occupied rent and backs out all the expenses noted above (once netted, it ends up being an very small number, relatively speaking). The only arguable housing-related adjustment I can imagine to get to a more intuitive saving figure might be to remove the amounts for consumption of fixed capital (depreciation) from PCE.


June 23, 2009 05:09 AM

October 2007...I do not have a Harvard education, but if I were looking for answers I might look at the source of the problem...duh!

Thank you for your interest. This blog is no longer active.



Michael Mandel, BW's award-winning chief economist, provides his unique perspective on the hot economic issues of the day. From globalization to the future of work to the ups and downs of the financial markets, Mandel-named 2006 economic journalist of the year by the World Leadership Forum-offers cutting edge analysis and commentary.

BW Mall - Sponsored Links

Buy a link now!