How Americans spend money: A look at the latest Consumer Expenditure Survey

When I discuss American spending habits, I try to cite specific numbers. Sometimes people write to ask where I get my info. Simple. Whenever I cite figures about American earning, saving, and spending, I get them from the U.S. government. In particular, I use the Consumer Expenditure Survey (or CEX) from the U.S. Bureau of Labor Statistics.

Here's how the BLS website describes the Consumer Expenditure Survey:

The Consumer Expenditure Survey program consists of two surveys, the Quarterly Interview Survey and the Diary Survey, that provide information on the buying habits of American consumers, including data on their expenditures, income, and consumer unit (families and single consumers) characteristics. The survey data are collected for the Bureau of Labor Statistics by the U.S. Census Bureau. The CEX is important because it is the only Federal survey to provide information on the complete range of consumers' expenditures and incomes, as well as the characteristics of those consumers.

The Consumer Expenditure Survey is the only reliable source I've found about actual spending habits. Most similar projects have much smaller sample sizes and/or provide theoretical numbers. The CEX is a great way to develop a descriptive budget (one that deals with real behavior) instead of a prescriptive budget (one that pushes an agenda).

Naturally, the CEX has its drawbacks. As always, averages (and medians) only provide a limited view of a dataset. Plus, what might be true for an entire population (a country, in this case), probably isn't true for a small subsection (your state or city, for instance). Still, for looking at the Big Picture, nothing I've found beats the Consumer Expenditure Survey.

Because I'm a money nerd, I get very excited when the new Consumer Expenditure Survey numbers are released each year. And guess what! The 2018 data was released two weeks ago. I spent some time yesterday sitting in the hot tub and geeking out over U.S. spending stats on my iPad. Then I updated my personal CEX spreadsheet. (What? You don't have one of your own?)

Let's take a closer look at the Consumer Expenditure Survey — and what we can learn from it.

Decades of Data

If you visit the BLS Consumer Expenditure Survey page, you'll likely be overwhelmed by the amount of information available. When I first found the site, I had to sort through the various reports until I found the one most useful for my work: the “age of reference person” table, which splits spending info based on the age of the person surveyed. (This is the base report I use when I collect stats.)

I also like the “Library” section at the Consumer Expenditure Survey website. It contains decades-worth of research papers and reports related to personal spending. Some of my favorites include:

To a money nerd like me, this stuff is golden!

Like most federal government agencies, the BLS has an excellent website. For starters, you can access all past CEX data on one of two pages: tables from 1984 to 2011 (plus 1961 and 1973), tables from 2012 to 2018.

Turns out the government performed this survey once during the early 1960s, once during the early 1970s, then made it an annual thing starting in 1984. That means there are now nearly fifty years of stats for money geeks to sort through. (Plus, the site provides access to a “forgotten” expenditure survey from the early 1940s!)

I've created the following table, which points to individual reports for each year. Yes, this is mainly for my own personal nerdery, but I hope that at least a few of you money bosses will enjoy it also.

The BLS Consumer Expenditure Survey summaries by year

1941196119731984198519861987198819891990
1991199219931994199519961997199819992000
2001200220032004200520062007200820092010
2011201220132014201520162017201820192020

If you'd prefer, you can make your way to the Get Rich Slowly file vault. I've created a directory that contains a PDF summary for every year of the Consumer Expenditure Survey. In the future, I'll add various other CEX-related downloadables here.

Crunching the Numbers

A few years ago, I went through past editions of the Consumer Expenditure Survey to grab data for a spreadsheet that tracks how American spending has evolved over time. Naturally, I just updated it for this year.

Rather than bury myself (and you) in a flood of information, I decided to look at roughly ten-year intervals. I figure by skipping a decade we get a more meaningful look at shifting habits. That means I used expenditure tables from 1961, 1973, 1988, 1998, 2008, and 2018 — the most recent year for which data is available. For fun, I also included data from the “forgotten” 1941 survey (which is much less comprehensive than later reports).

When I get time this winter, I'm going to update the spreadsheet to contain data for every year of the CEX.

Note: For our purposes, I cut a lot of cruft when transferring these numbers to spreadsheet format. These are complex statistical reports, and they're filled with numbers we don't care about (do we need to know the standard deviation for everything?) and jargon that just confuses, such as “consumer unit” (which we'll call “household”) and “reference person” (which we'll call “head of household”).

Here's a complete glossary of CEX terminology. Also, I'm not going to break down every little subcategory. The “Food at home” subcategory has 25 sub-sub-categories beneath it, including “bakery products”, “fish and seafood”, and “processed fruits”. I'm just going to stick to the primary categories and subcategories.

The following tables provide a run-down of the data I collected. After I show you the numbers, I'll discuss a few things that stood out to me about how American spending has changed with time.

Table 1: Cost of living adjustment and sample size
CEX - AssumptionsTables 2 and 3: Attributes of the average American household
Note that estimated market value of home is inflation-adjusted
CEX - AveragesCEX - Adjusted PercentTable 4: Income, taxes, and spending (actual)
CEX - Raw Income

Table 5: Income, taxes, and spending (inflation-adjusted)
CEX - Adjusted Income

Table 6: Spending by category (actual)
CEX - Raw Spending

Table 7: Spending by category (inflation-adjusted)
CEX - Adjusted Spending

Take some time to browse this info. Let it soak in. Look for patterns. Look for changes. What surprises you about how Americans spend their money? Where do we spend less than you thought we would? Where do we spend more?

Before I point out some of the things I find interesting, you should know a couple of things:

  1. First, the tables from the 1961 Consumer Expenditure Survey contain wildly contradictory information. Different tables supply different numbers. Sometimes when you add subcategories, the total is greater than the parent category. Where there's contradictory info, I used table B-17 as my guide (since it most closely resembles later CEX reports). Another problem? Sometimes the scanned pages are difficult to decipher. Is that a three or a five? A six or an eight? I think the numbers here are accurate, but there could be errors.
  2. The CEX tax calculation changed in 2013. Before then, tax numbers were getting corrupted by “non-response errors”. So, the BLS changed how they calculated taxes to get greater reliability. But, as a result, taxes for 2013 aren't like any other year in the survey's history. And taxes in years from 2014 on can't really be compared to 2012 or before. (For more info, check out the Consumer Expenditure Survey FAQ.

With those notes out of the way, let's dive into the numbers! Let's look at how Americans spend their money — both now and in the past.

A note on inflation: Using 2018 as a base, I calculated a cost-of-living adjustment for each year in order to factor out inflation. (I used the U.S. government CPI inflation calculator to get these numbers.) Looking at Table 1, you can see that prices in 1988 were roughly 48% of what they were in 2018. After computing this number, I converted it to a multiplier to get inflation-adjusted expenses. If 1988 prices were 48% of 2018 prices, that means we have to multiply numbers from the former by 2.08 to get equivalents for today.

How Americans Spend Their Money

We could spend hours sifting through this information to find patterns and trends. Instead, let's just hit the highlights.

First up, note how the size of the average household has dropped from 3.2 people in 1961 to 2.5 people in 2018. Almost the entirety of that drop comes from the fact that we're having fewer children. In 1961, the average household contained 1.2 children; today it contains 0.6 children.

Meanwhile, automobile ownership has boomed. Since 1973, we've gone from owning 1.3 cars per household to owning 1.9, an increase of 46% during the past 45 years. Other notable demographic changes:

  • Today, there are more full-time earners per household than there were in 1961 (1.3 vs. 0.8).
  • We, as a society, are much more educated than we were forty years ago. In 1973, an amazing 21.2% of Americans didn't have a high school diploma. Today, that number is down to 3%. Meanwhile, far more people are attending college (28% in 1973 vs. 67% in 1998).
  • Home values have far outpaced inflation. The average home was worth $76,156 in 1973 (inflation-adjusted); today, it's worth $198,612. But if you look at the tables, you can see the housing bubble in there. Look at the numbers for 2008!

Moving on to Tables 4 and 5 (Income, taxes, and spending), you can see that even when adjusting for inflation, household income doubled between 1941 and 1973. In the past 45 years, household income has only increased another 24%. It's fun to speculate about causal relationships for these numbers. I suspect that household income skyrocketed after World War II due to women entering the workforce. (But admit I could be wrong.)

As much as income has increased, spending has grown more quickly. In 1973, Americans spent 85.0% of their after-tax income. In 2018, they spent 91.1% of their after-tax income. (Admittedly, spending as a percentage of income seems to bounce around.)

To make it easier to visualize some of these changes, I created another table to calculate some important ratios and percentages.

Table 8: Derived numbers
CEX - Derived

In 1941, the average American family spent 31% of its budget on food. In 2018, the average family spent 12.9% of its budget on food. That's a huge decline! From 1973 to 2018, restaurant spending (food away from home) increased 48% while spending on food at home declined 36%. Overall, food spending declined 30% in in the past 45 years.

For me, it's more interesting to compare numbers today with numbers thirty years ago. (Thus the 30-year change column in the spreadsheet.)

I'm surprised that when you adjust for inflation, a lot of American spending has stayed relatively flat — or even dropped. We're spending about as much of our income on housing today as we did thirty years ago. There's a clear downward trend on our transportation expenses. Our food spending remains flat — even restaurant spending! We spend much less on clothing and (unfortunately) reading. (Although that reading number is probably down because we do most of our reading online.)

Meanwhile, health insurance costs are out of control. Between 1973 and 1988, the average American family decreased its spending on health insurance by 9.3%. But in the past thirty years, rates have shot up 245%. That is insane. No other spending category comes even remotely close. And it's why I make an exception to my no politics policy to advocate for scrapping our U.S. healthcare system. Any other option would be better and cheaper than what we have right now.

Looking at these tables, what numbers stand out to you? How does your spending compare to the average American family? What worries you about your spending — and the spending of the country as a whole? (And are there other similar surveys and reports I should take into account when writing about consumer spending?)

More about...Economics

Become A Money Boss And Join 15,000 Others

Subscribe to the GRS Insider (FREE) and we’ll give you a copy of the Money Boss Manifesto (also FREE)

Yes! Sign up and get your free gift
Become A Money Boss And Join 15,000 Others
guest
44 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
PhysicianOnFIRE
PhysicianOnFIRE
4 years ago

Great post, Boss! Thanks for breaking it all down for us. I heart numbers. The health insurance and health care numbers certainly stand out. I hope we start to see those numbers stabilize. The increase in the big-ticket spending and workload on each household go hand in hand. This is why Keynes’ prediction of the 15-hour workweek won’t come to fruition. Through marketing or greed or whatever it is, we want so much more than we did back when he made the prediction in 1930. The improvements in efficiencies are happening, but he failed to predict our increased appetite for… Read more »

Dividendsdownunder
Dividendsdownunder
4 years ago

I think all of these are great to analyse and look at J D, and once you take out inflation, there are still some scary increases in there, however I think most of it almost makes sense, or needs a bit more analysis. Housing – This has increased a lot, and in some cities is very unaffordable. However, I would be interested to know if they amount of square feet that people are getting for their money is increasing, about the same, or decreasing. Because if it’s changed, we’re not quite comparing apples with apples. Health (insurance) – Health expenditure… Read more »

Dan
Dan
10 months ago

There’s little in here that I take issue with. If we take a free market view of the world, things will cost what people are willing to pay. That’s how markets work. And everything you list is a good that can be purchased on the free market. Given this premise, one could easily take the opposite view: Historically, things were under-priced because the markets weren’t efficient. Take college: Assume I can get a job paying minimum wage with no-post secondary education. (What’s that, $7.50 in most places?) Now assume I can get my current job, which is paying me almost… Read more »

plain_jane
plain_jane
4 years ago

Clothing stood out to me. We’re spending less, likely on more pieces, and at what human and environmental cost?

Anne
Anne
10 months ago
Reply to  plain_jane

I wanted to reply to this. I was a child in the 1950s. Clothing was much more expensive then because almost all of it was made in the USA. Now, very little is. Getting some new clothing was a major event. Now, most of us can afford piles of it, far more than we need. (Guilty here.)

Jack Oatmon
Jack Oatmon
4 years ago

I’m amazed by how much the average annual expenditures are right now. Especially because it makes up such a high percentage of the average income. I wonder whether consumer debt is included here as well. The gap between income and expenditures is almost $5000, however I bet most people aren’t saving $5000 per year. My annual household income is almost three times the average amount quoted here, yet we spend around the average . . . but with that said I feel like we spend way too much! To know that people with average incomes are spending just as much… Read more »

TheSouthernStache
TheSouthernStache
4 years ago

Another informative post.

Is there anywhere on here that states the change in the number of households? If due to divorce, there is an uptick in number of households, that could help explain some of the numbers. Reduced number of children per household, for example.

Mysticaltyger
Mysticaltyger
4 years ago

I will get on my soapbox here and preach that we probably need to spend more on quality food and generally cut back on meat & poultry and increase intake of organic fruits and vegetables. If we would do that, health care costs would plummet. Incidence of diabetes, heart disease, some cancers, and even Alzheimer’s would be greatly reduced.

El Nerdo
El Nerdo
10 months ago
Reply to  Mysticaltyger

Meat doesn’t cause diabetes ¯\_(?)_/¯

chacha1
chacha1
4 years ago

I’m going to respectfully disagree with mysticaltyger: improving the nutritional quality of the American diet will not cut health care costs, because the vast majority of individual healthcare spending is not on actual healthcare services purchased, it’s on insurance premiums paid. My family insurance premium – including the share paid by my employer – is $1300/mo. I am a healthy 50 yr old woman and my husband is a healthy 56 yr old man. Our purchases of actual healthcare services averaged less than $600 a year – retail value – for the ten years 2003-2013, a period during which I… Read more »

Sandi K.
Sandi K.
4 years ago

One category that is missing from my quick scan is childcare. For some households, that is an enormous expense.

Sam Smith
Sam Smith
4 years ago
Reply to  Sandi K.

I wonder if that is captured under education? But, yeah that is a huge number for us right now.

Barry N.
Barry N.
10 months ago
Reply to  Sam Smith

I agree, I feel like this an enormous expense that is not being clearly reported. I know that childcare takes up the lion’s share of my monthly expenses.

Steve S
Steve S
10 months ago
Reply to  Barry N.

Childcare costs are really only very large for the first 5 years of your child’s life…putting that in as an average across all households would really just capture the snapshot of people going through that particular rough period before their kids started kindergarten.

FoxTesla
FoxTesla
10 months ago
Reply to  Steve S

Disagree, as the cost is significant until the parents decide to trust a child at home alone.

Example: Post-school care costs us 66% of the full-time daycare rate, and school-age summer full-time rates were higher than our pre-schooler’s full-time rates.

Steve Henderson
Steve Henderson
4 years ago

This is a very interesting and good look at people’s spending. Thanks for writing it. Yes, the Consumer Expenditure Survey changed the income tax calculation process with the 2013 tables. Instead of trying to get people to guess what their federal and state income taxes would be, the survey started using the National Bureau of Economic Research TaxSim estimator. Here’s a link to the details in FAX #23 on the BLS Consumer Expenditure Survey website: http://www.bls.gov/cex/csxfaqs.htm#q23

The BLS online tables do show the average mean family size: http://www.bls.gov/cex/22015/midyear/age.pdf
It is the row labeled “Average number in consumer unit”

Sam Smith
Sam Smith
4 years ago

Interesting data. Surprised that number of earners has been flat for more than 20 years. Most of the spending seems relatively flat too. The biggies are health insurance, education and retirement. Sadly you really can’t control health insurance costs and if you want to retire these days you have to put a lot of money away instead of your employer doing it for you. You can control, in part, education costs but even doing the state university route is still way more expensive, adjusted for inflation, than it used to be. The other two, housing and transportation are interesting to… Read more »

LennStar
LennStar
3 years ago

I think the comparison of X to cloth is not a good one, because clothes have become a lot cheaper. Its not as in the olden days where the average people had one normal cloth and then one sunday cloth and that was it.
If course for every 5$ T-Shirt you can still buy the 100$ and more variant.
Similar with food.

Dave @ Accidental FIRE
Dave @ Accidental FIRE
10 months ago

“From 1973 to 2018, restaurant spending (food away from home) increased 48% while spending on food at home declined 36%. Overall, food spending declined 30% in in the past 45 years.”

So overall food spending is the same but we’re eating out more than ever. Seems to me we could be spending way way less on food than 45 years ago by eating out less.

Love the data, I mine the BLS stuff all the time

Klaus
Klaus
10 months ago

I’d love to see a “Spending by Category (adjusted for inflation and household size)”

Food about the same, but paying for fewer mouths, so has gone up per person….

S.G.
S.G.
10 months ago

What I want to know is how you posted it today, but updated it two days ago.

El Nerdo
El Nerdo
10 months ago
Reply to  J.D. Roth

Ah haha, I was pleasantly surprised to see Chacha back. Unfortunately, it was lies…

IMO, just keep the original dates on, like Don Bigotes does. Then just add the update date.

What’s wrong with comments… spanning time? (Buffalo 66 reference, hahaha).

Sue
Sue
10 months ago

I think you’ve transposed some data in the taxes columns for 2018 – looks like the inflation-adjusted numbers have been used in both table 4 and 5.

Must be a money nerd, because I find this really interesting too!

Sue
Sue
10 months ago
Reply to  J.D. Roth

Did taxes really go up that much between 2008 and 2018? Still referring to table 4.

Sue
Sue
10 months ago
Reply to  J.D. Roth

Thanks, J.D., that makes a bit more sense. I’m a Canadian, so I’m also looking at this data from the outside.

Dan
Dan
10 months ago

JD,

As a data analyst, this is a bad, post sorry 🙁 In general, just throwing numbers at people and expecting them to figure things out for themselves is considered bad form.

Here’s what gets my attention: You say health care costs are out of control. But your last derived table calls “health care share” 5.0% in 1941 and 8.1% in 2018. I look at that say “looks fine to me, what’s the problem?”

Dan
Dan
10 months ago
Reply to  J.D. Roth

1. Charts, not tables. Most people can’t do numbers very well, so bar charts or line graphs (as appropriate) will illustrate relationships much more clearly. I’m at the point now where if someone is showing me a table full of numbers, I assume they’re hiding something. (And BTW, I’m good with tables full of numbers…) 2. For a high-level top post, just show the inflation-adjusted numbers. You can provide links to the un-adjusted data. These tables are long, and your presentation starts with the “raw” data. People dig right into it, formulate some relationships, and then realize the numbers are… Read more »

Luke Calvano
Luke Calvano
10 months ago

Have you written recently about the health care system or education? These two systems feel incredibly broken. Health Care and Education are the fastest-growing expenditures on an inflation-adjusted basis over the 30-year period, and they are both places where we culturally diverge from European models. They are probably the two most complex problems the country faces and have the highest consequences, which makes them important things to discuss. I know you prefer to focus the blog on tangible advice you can provide readers, in the area of personal finance you have expertise. This is an all-positive, no negative approach to… Read more »

bogart
bogart
10 months ago

On the household size / fewer children point, I’d guess that the fact that we’re living (lots) longer plays into this too — many more older householders who have children, but whose children are adults who’ve moved on to their own households (the fact that we’re having children later may also play into this, though I’d guess less so).

RNP
RNP
10 months ago

As with others, the number one factor that stands out is health insurance, especially the ratio of health insurance to health care. In 1973, health insurance was at 37% of the health care cost, 1988 – 36.5%, 1998 – 47.9%, 2008 – 55.5%, 2018 – 68.5%. This reflects the evolution of medical insurance evolving to more of a cost sharing system than a true insurance product. It may also reflect the increased shifting of medical insurance premiums to workers vs employers. Education costs have meaningfully increased as we have evolved from a lower skilled manufacturing economy to a higher skilled… Read more »

Abby
Abby
10 months ago

I’m surprised that retirement savings is roughly 10% of net income in 2018. Not bad. And why wasn’t this category part of the derived numbers chart? Is this category not significant?

El Nerdo
El Nerdo
10 months ago
Reply to  J.D. Roth

Retirement savings increased when 401(k) enrollment was made automatic rather than people having to fill out endless forms.

See:

http://money.com/money/4974462/thaler-nobel-economist-retirement-savings-nudge/

Social policies matter 😛

El Nerdo
El Nerdo
10 months ago
Reply to  J.D. Roth

oh! i’ve been meaning to watch that for ages…

Steve S
Steve S
10 months ago

Two things I notice: 1) I can really see the high inflation rates of the late 70’s/early 80’s in the data. Look at the nominal values for 1973 vs. 1988. It looks like everything went way up – relative to later time intervals – but knowing that inflation was high during those years explains away some of the disparity. 2) “Tobacco” as a budget line item was interesting. Most household budgets have exactly $0 budgeted for tobacco. Why average out a line item that only affects 15-25% of households across the entire country? The numbers for what the average smoker… Read more »

Greenbacks Magnet
Greenbacks Magnet
10 months ago

I have to agree that healthcare and education costs are out of control. I think college should get straight to the point and be only about 8 months. Why do you need 5 electives ? Just decreasing college by 30 credits knocks off a year of schooling. Allow people to get into the workforce sooner so they can pay off debt and save in a 401k faster.

Engaging Data
Engaging Data
9 months ago

JD,

thanks for the interesting article on the CEX which I am also interested in. I like the historical look at expenses and taxes. I made some interactive data visualizations (i.e. Sankey diagram) of some of these numbers for the 2017 data. And it is disaggregated across certain demographic characteristics, like age, education level, marital status and children, etc. . .

https://engaging-data.com/household-spending-income/

I’m continuing to go through the CEX data and make additional graphs as I find interesting bits of data to share. I’m also currently beefing up my stats so I can access and analyze the micro-use data.

shares