Greater Greater Washington

See how housing and transportation costs hold the poor back

Everyone's spending habits are basically the same. But rising housing and transportation costs hit low-income households hardest.


Consumer spending by income. All images by the author.

Data from the Bureau for Labor Statistics show that people of all income levels tend to spend similar percentages of their budgets on each expenditure category, with some exceptions. For example, as income rises, Consumer Units (defined as families living together, financially independent individuals, or groups of unrelated individuals who budget jointly) dedicate an increasing percentage of their budget to personal insurance and pensions.

Consumers at the lower end of the income spectrum spend a disproportionately higher percentage of their budget on housing costs. But on most other measures, including transportation, health care, and entertainment, the percentages across income levels are fairly equivalent, as shown in the graph above.

But these percentages represent share of total expenditures, and not all Consumer Units are operating with a balanced budget. A comparison of income and expenditures shows that lower-income families and individuals tend to spend more than they earn while higher-income units are able to stash some of their earnings away. The graphs below attempt to illustrate this:


Low-income households spend more than they earn.

When you change the denominator in the first graph from Total Expenditures to Annual Income, a more accurate depiction of our spending habits is revealed. Consumers earning between $5,000 and $30,000 per year spend 62% of their income on Housing, 24% on Transportation and 23% on Food. That's 109% of their income gone just on these three basic necessities.


Low-income households are burdened by high housing and transportation costs.

This modified version of the first graphic presented gives a fuller representation of household finances at different income levels in the US. It paints a pretty bleak picture for low-income families and individuals.

A version of this post originally ran at R. U. Seriousing Me?


Chris Dickersin-Prokopp spends his days in Anacostia and nights in Petworth. He studied Latin American Studies and Urban Planning. He runs the blog R.U. Seriousing Me? and occasionally contributes to the Washington City Paper

Comments

Add a comment »


So, low income people don't have discretionary purchases.

Pretty curious how you spend 100% on housing making 5000 a year. That is close to 400 a month, which seems like a lot.

by charlie on Jan 9, 2014 11:16 am • linkreport

not sure why we are expecting people making $5000 a year to be able to afford all of their expenses on what they make. They aren't working, at least not full time.

7.25 an hour, 40 hours per week 50 weeks a year(includes a vacation) is $14,500.

If you aren't working full time, you are going to have to dip into saving, that is what savings is for or you arent going to be able to live a "typical" lifestyle.

by Richard on Jan 9, 2014 12:27 pm • linkreport

but look at the numbers for the 20 to 30k column. or the 14 to 20k column. Not pretty there either.

and of course lots of people are involuntarily woking less than 40 hrs a week - either because they cannot find full time employment, or they are single parents with few good child care options.

by AWalkerInTheCity on Jan 9, 2014 12:30 pm • linkreport

I would assume that the lower end people are mostly retired/ on public assistance / social security/other fixed income. On the otherhand, I'm in the 40-50 column and that is pretty accurate except the remaining 10% is paying off student loans.

by BTA on Jan 9, 2014 12:41 pm • linkreport

@ AWalkerInTheCity; those numbers aren't individuals -- they are "consumer units" which is a apparently a type of household.

So, yes if you household is making less than 30K a year, you're pretty damn tight.

by charlie on Jan 9, 2014 12:43 pm • linkreport

I would be interested to see transportation separated from discretionary travel. I would guess that wealthier households spend more on travel but at a certain point your transportation costs can't realize get much higher than a car payment and gas and/or taxis.

by CBGB on Jan 9, 2014 1:03 pm • linkreport

Eh you can buy a car for 15k or 50k, you can buy more than one car, you can get cut rate nearly useless insurance or premium insurance. Most wealthier people I know have at least one car for every adult/teenager in the family. But yeah I imagine once you hit the $10k range of so most of the additional transportation spending could be classed as discretionary.

by BTA on Jan 9, 2014 1:22 pm • linkreport

@AWalkerInTheCitybut look at the numbers for the 20 to 30k column. or the 14 to 20k column. Not pretty there either.

Oh, I agree, there are a lot of people working full time, very hard, in the 20-30k range and that requires exploration and discussion. For a family making that, the choices on where to live, how to get around, what to consume, and how to care for their health and future are all very difficult.

I just hate when these conversations start revolving around a ultra low end that doesn't make much sense. You wouldn't expect someone to be able to cover all the expenses of life on $5k a year and more than you would expect them to be able to do it on $0 a year.

by Richard on Jan 9, 2014 1:36 pm • linkreport

One thing that is not mention is what type of job is paying the $7.25

In a retail job you will never work 40 hours and your hours change season by season and sometimes every month. The same in hospitality, your hours could change during different seasons.

The transportation could also be a big factor with the following ways

The time of day you work will determine if you can take bus or rail; The bus runs later than the train for the most part.

The first trains leave the out about the start of the line at 5:00am; the first trains get to Metro Center about 5:30am on Weeekdays.

If you work weekends and have to be to work before 9:30am Metrorail is out of the question

The last trains leave out the end of the lines at about 11:20pm something except for the Green Line which leaves at almost 11:40pm for some reason.

If you work past about 11 you wont be taking the train unless you work in DC where the last trains reach Metro Center/Lefant Plaza & Gallery Place about 12am.

If you have to take a bus before or after rail add between 30minutes to 1 hour before the first or last train.

Many of the Metrobuses in DC run until 2am, some RideOn and Fairfax Connectors run until 1am

If you work on a weekends and have to be at work by 8 or 9am mostly hospitality, retail or restaurants your screwed or it you get off after 8pm in Virginia or Maryland and take buses you are also screwed.

by kk on Jan 9, 2014 1:43 pm • linkreport

@kk
One thing that is not mention is what type of job is paying the $7.25
In a retail job you will never work 40 hours and your hours change season by season and sometimes every month. The same in hospitality, your hours could change during different seasons.

Certainly it can be tough to get 40 hours at minimum wage in nice 8 hour per day blocks M-F. I know I have been there. You end up usually having to work 2 jobs ~20 hours each and then your transport cost is even worse and there is a lot of dead time in between.

by Richard on Jan 9, 2014 1:53 pm • linkreport

@ Richard:I just hate when these conversations start revolving around a ultra low end that doesn't make much sense.

Yeah, let's ignore the people who have the least. It's very hard to help them anyway, and besides, do you know how much a nanny is these days?

by Jasper on Jan 9, 2014 2:02 pm • linkreport

@Jasper
Someone making $5000 a year at minimum wage would be working 13 hours a week.
Should that someone be able to afford a large house, car, and still have enough left over for a hefty entertainment budget?

If someone is working 13 hours a week, probably only 2 shifts a week, then I would fully expect housing and transportation to fully consume everything they make. The only way I see someone only working 13 hours a week is if they have free housing/transport or don't expect to be in a steady state(not spending more than they earn)

by Richard on Jan 9, 2014 2:17 pm • linkreport

Around the office:

"Should we include the whole graph, all the way to the left?"

"sure, we have data for that, its statistically significant, why would we imply that there are no data points at less than 5k income?"

"because its going to jump out and some idiots are going to start quibbling about who makes below 5k a year, which will detract from the main takeaway of the chart"

"nah, no one would do that. Publish as is"

by DataAnalyst on Jan 9, 2014 2:21 pm • linkreport

@DataAnalyst: Statistical significance in and of itself isn't much of an argument for including something in a visualization. A more useful metric here would be the percentage of respondents that fall in that range (directly related to the significance of the results from that category). I imagine that it's not a large percentage of households, so yes: because it's such an outlier, and has little real significance, it detracts from the rest of the chart.

by Gray on Jan 9, 2014 2:28 pm • linkreport

That exclusion is easily done by the reader, which allows those readers who are interested in the entire spectrum of income to get the benefit of inclusion.

I am not saying it might not be wise for the analyst to exclude that data - but that they are making a concession to the - how shall we say this within the terms of use - the lack of acuity of some readers. Note one speaker in the dialog takes your POV - but she still treats the reader who would quibble as an "idiot".

by DataAnalyst on Jan 9, 2014 2:38 pm • linkreport

The figure on expenditures by income shares is interesting. The spending components that vary are: food (decreasing), saving/pensions (increasing), and housing (decreasing). Declining food expenditure shares in income is sometimes known as Engel's Law, known for 150 years or more. The saving/pension category is not like the others: it's deferred consumption. And that brings us to housing.

My first thought is that these consumption categories include a wide variety of quality. A BMW transportation unit is nicer than a Chevy transportation unit (or a Surly is better than a Huffy); a white tablecloth restaurant meal is nicer than ramen and Kraft dinner; a visit with a Dupont ortho specialist is nicer than the WHC ER; etc. So maybe income shares are the same across incomes in most categories because people smooth across quality. So why don't people smooth housing consumption this way? Maybe there is an insufficient supply of low-cost, low-quality housing? That is part of Matt Yglesias's "Rent is Too Damn High" thesis. It fits with gentrification ideas too. Although the observing group housing (and multi-generation cohabitation?) shows that smoothing is possible to a degree.
Also: The housing+transportation share category varies over income, so "drive until you qualify" is over.
Also: home ownership and savings both have tax advantages. I don't see those in the data.
Also: mortgage amortization means that homeownership is saving, to a degree. So that makes the divergence in shares by income even more stark.
TL;DR: I like graphs.

by JLK on Jan 9, 2014 2:59 pm • linkreport

Do the expenditures include spending that's funded by government benefits such as food stamps, welfare, housing vouchers, EITC, medicaid, etc.? If so or not, it would be interesting to see what expenditures are excluding/including benefits.

Does personal insurance expenditure include employer health insurance subsidies, employer payroll tax contribution, and/or individual payroll tax contribution?

by Falls Church on Jan 9, 2014 2:59 pm • linkreport

@DataAnalyst: Hey, if you want to rely on calling readers idiots rather than improving a chart, I guess that's one approach.

My point is that a chart is meant to convey certain information. If you include other information for no particular reason other than that it exists, then it's rather silly to complain that people don't focus on what you wanted them to see.

Excluding less important information is not making a concession because everyone else isn't as smart as you, it's making sure that your chart shows what you want it to show.

by Gray on Jan 9, 2014 3:10 pm • linkreport

I don't think excluding a valid data point, because the n is fairly small (is it? I'm not sure it is in this case) is "improving" the chart. For readers who are capable of understanding the meaning of that leftmost column, its not an improvement at all - they can see the leftmost column and STILL see the principle takeaway. Its an improvement only because many readers are not able to both see the leftmost column AND focus on the main takeaway.

Whether in the privacy of the office analysts refer to those readers as idiots, people with below average IQ, or "the target audience" will vary from time to time and place to place.

by DataAnalyst on Jan 9, 2014 3:21 pm • linkreport

by all means, show the entire graph. I find it all very interesting. The income vs expenditure graph with the red and gray lines are very interesting although I am a little currious why the far left hand entry for income is $-2,153 and not 0. I guess someone had some sort of debt that is being calculated as income distorting the data.

Anyway looking at that curve we can see that people who make basically nothing have higher expenditures than those who are making around $15,000.

I would assume that a lot of the data of people making $5-10k a year are either students, who are spending more money now than they make hoping to make more later or seniors who are spending more now than they make because they have some money saved and are deciding not to die rich.

by Richard on Jan 9, 2014 3:27 pm • linkreport

JLK, housing and food (and to a lesser extent transportation) are the two most basic consumptive necessities so they are naturally less elastic at the bottom of the scale. It's not really surprising that housing is relatively inelastic since the alternative is not have a roof over your head. Food is subsidized in this country meaning minimum calories if not nutrition or quality are relatively cheap to attain. I think gentifrification is a red herring compared to major barriers like zoning which can have a huge effect on disorting supply.

by BTA on Jan 9, 2014 3:28 pm • linkreport

Negative income could be investment losses that offset actual income?

by BTA on Jan 9, 2014 3:29 pm • linkreport

@DataAnalyst:
Whether in the privacy of the office analysts refer to those readers as idiots, people with below average IQ, or "the target audience" will vary from time to time and place to place.
Right. And other analysts will focus on creating charts and figures that focus on what they want readers to see, instead of spending time convincing themselves of their superior intellect. To each his own, I guess.

by Gray on Jan 9, 2014 3:31 pm • linkreport

Indeed.

Some will adapt to the limits of the audience with no frustration, and some will be frustrated at having to shape and reshape charts to avoid misinterpretation. I suppose those who are able to take joy in the adaptation may consider themselves superior - certainly our society rewards them better.

I would have to admit, the frustration of data nerds with the need to reshape charts for innumerate audiences is one of the lesser of society's problems - not only behind poverty, disease, global warming - but even beneath say the need to clear snow from bike trails, or to better regulate DC taxis.

I do look forward to the day when no one is innumerate, and no one needs charts to spoon feed them the main point. I can't see what we would lose if that were the case.

by DataAnalyst on Jan 9, 2014 3:43 pm • linkreport

This modified version of the first graphic presented gives a fuller representation of household finances at different income levels in the US. It paints a pretty bleak picture for low-income families and individuals.

I'm not sure that the picture painted is bleak. To me, it would be far more bleak if someone making $10K a year was only living $10K. That would be a pretty bleak existence.

Assuming these expenditure numbers include consumption made possibly by government benefits, it really goes to show how important benefits are to raising living standards and keeping people from a truly bleak existence. Another graph worth showing is one showing real income over time vs. real consumption over time. Real income for the poor and middle class has been stagnating for decades. But, thanks to government programs, real consumption has been on a steady rise for decades for the poor and middle class.

by Falls Church on Jan 9, 2014 3:44 pm • linkreport

Each income bin has a mean (maybe median, I forget) income reported by BLS for all consumer units in that bin. The lowest-income range includes everyone earning $5k or below, and the mean income for that group is -$2,153. That same group spent an average of $23k. To say that they spent -1070% of their income that year would not have been meaningful, and would have screwed up the third graph, so I excluded it.

by Chris Dickersin-Prokopp on Jan 9, 2014 3:46 pm • linkreport

Some will adapt to the limits of the audience with no frustration, and some will be frustrated at having to shape and reshape charts to avoid misinterpretation.

And, some will be frustrated by the inability of some data nerds to present large amounts of data in a way that they can be interpreted within a few seconds with little chance of misinterpretation.

I do look forward to the day when no one is innumerate, and no one needs charts to spoon feed them the main point.

I'm plenty numerate but would much rather have a chart spoon feed me the main point in a matter of seconds rather than digging through the raw or summary data myself and spending far more time figuring out the main point.

by Falls Church on Jan 9, 2014 3:57 pm • linkreport

You are asked to give a presentation comparing the projections for last year to actual year-end values, by company unit. So you create a straightforward bar graph. For each unit, you show four bars: 2013 projected values, 2011 actual values, 2012 actual values, and 2013 actual values. You then use this chart to illustrate how projections matched observed results by unit.

Someone asks why the 2011 and 2012 actuals are included. You reply that they are not the focus of the graph and are irrelevant to your point, but are included because you had them and they could be of interest to some.

Is the correct interpretation that the audience is stupid or innumerate, or that you are bad at figuring out which data are important to present?

by Gray on Jan 9, 2014 4:14 pm • linkreport

Should a chart only have one main point? The full data provide other interesting subsidiary information. In this case the breakdown of expenditure by those with 5k or under incomes (whether those are students or retirees or whomever). There is usually a tradeoff - (when there is not, there is no frustration with spoonfeeding.) The more people can see the main point without exclusion of data, the more useful the table can be to all.

by DataAnalyst on Jan 9, 2014 4:15 pm • linkreport

Naturally a comparison of predicted to actual is only meaningful for dates where there is a projected value. And of course any time series must have a starting date.

But the natural beginning for an income chart is zero income. There is no logical reason to exclude below 5k income from a chart of expenditure categories. And there is no good reason why it should confuse someone numerate who puts a few seconds thought into it. IMO.

by DataAnalyst on Jan 9, 2014 4:19 pm • linkreport

Should a chart only have one main point?

A really good chart makes multiple points, all of which can be readily identified with little scope of misinterpretation in a very short period of time.

The full data provide other interesting subsidiary information.

And, may provide a level of detail that is extraneous and/or take longer to interpret than a good chart.

There is usually a tradeoff - (when there is not, there is no frustration with spoonfeeding.)

The frustration a lot of presenters have with "spoonfeeding" is that it can take a lot more time and effort on the part of the presenter to develop a good chart instead of just putting a bunch of data in a table. However, if a few hours on the part of a presenter can save a few minutes for thousands/millions of people looking at the chart, it's time well spent.

by Falls Church on Jan 9, 2014 4:49 pm • linkreport

Nice work as usual, Chris.

by MLD on Jan 9, 2014 5:02 pm • linkreport

The tradeoffs can be in the form of time, or they can be in the form of cutting out useful information. When the former the kind of analysis you suggest makes sense. But quite often its more like the case above - the tradeoff is leaving out potentially useful information. Or creating other head scratchers "Why did he leave out the under 10k income block - is he hiding something?"

by DataAnalyst on Jan 9, 2014 5:29 pm • linkreport

@Falls Church
The income line includes almost all government benefits (unless the expense the benefit is intended to cover is also excluded from the expenditure line). The one exception is that refundable tax credits, such as the EITC, do not show up here, and that would help close the gap between income/expense. Despite that additional assistance, a typical household making $15k and spending $25k, for example, still has a sizable deficit that the government does not subsidize directly. Rather the gap must get filled by some combination of savings, loans, gifts not reported as income, and credit.

by Chris Dickersin-Prokopp on Jan 9, 2014 6:20 pm • linkreport

It is possible to work full time (40 hours a week or more) and earn less than 5,000 if all of your income is self-employment income.

by George Costanza on Jan 9, 2014 6:26 pm • linkreport

very nice analysis and figures

by Tina on Jan 10, 2014 10:22 am • linkreport

@Gray -i agree w/DataAnalyst: it was better to include the low end data rather than exclude it. It tells us something. I find these figures very easy to read.

by Tina on Jan 10, 2014 10:28 am • linkreport

@DataAnalyst
Its an improvement only because many readers are not able to both see the leftmost column AND focus on the main takeaway.

Whether in the privacy of the office analysts refer to those readers as idiots, people with below average IQ, or "the target audience" will vary from time to time and place to place.

Thats kind of ungracious. I see it (inability to understand data) as indicative of the overall US weakness in math and science education.

e.g. I had the same IQ before and after learning in statistics. So did you, no?

The discussion of whether or not to include the left-hand data is more a discussion of style, and in this I agree w/you: include it. It tells us something. Its useful. Its interesting. Its the biggest picture.

by Tina on Jan 10, 2014 10:38 am • linkreport

@Falls Church
I'm plenty numerate but would much rather have a chart spoon feed me the main point in a matter of seconds rather than digging through the raw or summary data myself and spending far more time figuring out the main point

heehee

by Tina on Jan 10, 2014 10:41 am • linkreport

Add a Comment

Name: (will be displayed on the comments page)

Email: (must be your real address, but will be kept private)

URL: (optional, will be displayed)

Your comment:

By submitting a comment, you agree to abide by our comment policy.
Notify me of followup comments via email. (You can also subscribe without commenting.)
Save my name and email address on this computer so I don't have to enter it next time, and so I don't have to answer the anti-spam map challenge question in the future.

or