Source: Census Transportation Planning Projects. Chart by BikePortland.
Last week, we shared some new Census data showing that people who bike to work in Portland have quicker commutes than you might expect. This week, let’s look at a different question: who bikes?
“We have to prioritize investments in communities that have not been prioritized for investments in the past.”
— Gerik Kransky, BTA
It turns out that in the Portland metro area, people of every household income level bike for transportation. But the lower your household’s income, the more likely you are to use a bike to get to work.
That fact — which national data has shown for years but had never been available at the local level — is part of the thinking behind a rising focus in the bicycle advocacy community on the ways that biking can help underprivileged groups.
“People living with multiple resource constraints are in the best position to benefit from increased access to healthy, active transportation options,” Mychal Tetteh, CEO of Portland’s Community Cycling Center, said Thursday. “If we want to see bike mode share increase, a focus on historically underesourced populations will result in the greatest return on investment.”
According to the new estimates, which are based on Census surveys that include margins of error, the poorest 25 percent of Portland-area households are home to about 34 percent of the metro area’s bike commuters. The other three quartiles are quite evenly split, suggesting that bike commuting is both a useful necessity for some and a desirable choice for most.
A bill introduced yesterday in the U.S. House of Representatives reinforces this concept. The bipartisan measure championed by the League of American Bicyclists would create a low-interest long-term loan program for communities to build biking and walking networks, with one quarter of the cash set aside to be used in low-income communities.
H.R. 3978 is worth just $11 million for the whole country — about enough for each state to get either one new stoplight, several blocks of sidewalk, a few bike share kiosks or a few miles of bikeways.
On the other hand, it’s showing (yet again) that the appeal of active transportation can cross party lines in a deeply divided Congress.
“It’s a good idea, it’s a good bill and we should certainly support it,” Bicycle Transportation Alliance Advocacy Director Gerik Kransky said in an interview Thursday. “It takes a small but important step toward acknowledging that we have to prioritize investments in communities that have not been prioritized for investments in the past. … These are the kinds of policy decisions we’re going to have to make to let low-income communities make their own decisions.”
Kransky said the risk that a bill like this becomes an excuse for politicians to avoid bigger changes is “always out there.” But he hopes the small amount of money could prove that there’s public support for further shifts, including increased “self-deterimination” by poorer communities of the transportation investments in their neighborhoods.
“A bill like this passes, the process is set forth, money is spent, the outcome is fantastic and the community support is there, then all of a sudden we have a working model for engagement and communication,” Kransky said. “If this works, we can use it as a model at the state or local level.”
Well, of course. People with less income can’t spend as much on cars so they bike or take mass transit. Yet, infrastructure like good mass transit and safe bike routes are often not in poorer areas. Hmmmm….
I’m amazed how it’s relatively evenly split (with an advantage to lower quartile).
Michael, since I’m lazy, can you list the cut offs for each quartile?
Good question, and one that matters to how we think about the data. I went back and forth on whether to put this in the story proper but decided it was already getting complicated. Instead, what I’ll do is add these to the image up top.
– The poorest quartile tops out at about $46,000.
– The second quartile ends just under $75,000.
– The third quartile ends around $112,000.
– The top quartile is everything above that.
I was pretty surprised by such high income figures in general. Then I realized that the figures are for households, and a strong majority of working households in the Portland area have two workers.
So there’s an important lurking variable here: family size. It’s clear that bike transportation is more common among the single, and probably also among those without kids.
That said, it’s also true poor people are disproportionately single, and single people are disproportionately poor.
Also important to keep in mind: these figures apply only to people who currently have jobs to commute to. Obviously such people are wealthier than the public at large.
Many factors at play here.
Great point. I hadn’t thought about the fact that many single people who would still be middle/upper middle would fall into the lowest quartile category. With that in mind, I’m even more surprised that the lowest 25% limit is at $46K, as there are a lot of single people in this town!!!
Would household income per capita be a better metric? Not sure if the data access allows it….
It would be in my opinion, at least for this question. Sadly, this isn’t in the data.
Michael, any data about the average or median age of a bike commuter? While I certainly see all ages out there, it seems the majority of people are under 40. Certainly younger people are not going to be as established in their careers yet and probably would not be making over $46K/year yet.
Hm. My understanding, which I’ve just confirmed via the census website, is that the median household income in Portland is closer to $50K, and was lower during the years of this survey, so I’m confused by this quartile breakdown. Do you happen to know why this doesn’t match census data?
That seems exceptionally low. My neighborhood is supposedly around $46k and we’re one of the lower ones in the city.
And doesn’t the quartile breakdown in this graph mean that the median is $75k? That actually sounds about right to me.
I guess if you looked at the census data (I couldn’t make any sense of the above linked site), you must be correct.
The difference is that the Census data you’re looking at is based on a wider set of people (a different “universe,” in the Census jargon): all households. The above data applies only to households with workers, a group whose household incomes will trend somewhat higher since it excludes people who are unemployed, on disability, on social security, full-time students, etc.
Thanks for the clarification.
Depending on the data – and I assume it is American Community Survey – the cross tab of income and commute type is likely to cause larger error bars in the higher income ranges due to sampling. I also suspect there is more of a story in the lower quartile since that is not too far from a median household income number. (So the chart over represents the wealthy as three slices, and the rest in one.) Fun with stats!
Paul, you’re right that this is ACS data. But in what sense would the chart overrepresent the wealthy? The ACS categories for higher household incomes are much less finely grained than those for lower household incomes, but it’s true that about 25% of Portland metro households have total incomes over $112k. Is this a technical issue in ACS collection that I’m unaware of?
The old saying, ‘correlation does not imply causation,’ comes to mind; the data presented here does not prove the statement, “the lower your household’s income, the more likely you are to use a bike to get to work.”
For example, let’s suppose most of the people that ride to work are under 25. People under 25 don’t, on average, make as much money as someone over 50. Therefore the statement could be, “the lower your household’s AGE, the more likely you are to use a bike to get to work.” But that doesn’t mean if you are 55 years old and make under $46k, you are more likely to bike than another 55 year old that makes $60k.
I am not claiming that the above example is correct, either. I’m just trying to illustrate that more statistical analysis is needed to prove causation… and maybe that has been done but simply not presented in this article.
Perhaps a better example: Statistics may show that 76% of cyclists in Portland are white. However, since 76% of Portland is white, clearly that would not support the statement, “If you are white, you are more likely to use a bike to get to work.”
Jeff, I appreciate the pushback and I think I take your point, but I don’t think what you’re describing is a correlation/causation problem. (If I’d written “being poor makes you more likely to ride a bike,” that’d be a correlation/causation problem.) What you’re describing is a lurking variable. And you’re right, younger people are more likely to bike and more likely to have lower household incomes, both because they’re usually less experienced and because they’re more likely to be single. I’m sure that explains some, though probably not all, of the discrepancy shown above.
See my note to Davemess above.
Moreover, as Tetteh points out in the post, the fact that bicycling works for many low-income people suggest a lot of potential for increasing biking among other low-income people.
I am not understanding how “But the lower your household’s income, the more likely you are to use a bike to get to work” (from your initial post) is different than “being poor makes you more likely to ride a bike”. Aren’t those the same conclusion? I absolutely support bike infrastructure. But this is an incorrect interpretation of statistical data.
A conclusion that is actually demonstrated by your data is “as a percentage of the total bike commuting population, more bicyclists are lower income.”
Unless we are looking at what percentage of people within each income quartile bike, then we cannot conclude anything about income making one more likely to bike. I mean, how do we know that 100% of people in the 2nd or 3rd quartile do not bike? Maybe they are just smaller groups, so their piece on your pie is smaller.
Anyway, thanks for posting, and I do support this bill.
I feel I need to clarify, since my my 100% example is hyperbole for the sake of illustration. I am interpreting income quartiles by household. Eg, census data is showing households with AN individual who bikes, and income data is by household. So, if 5 people in a high income household bike, and 1 person in a low income household bike, they each are one data point on this pie. So thus, we cannot say the poorer people are more likely to bike just because their household stat says SOMEONE bikes.
Actually, it doesn’t work that way – the workers you describe would represent six points in this data. The question is essentially: “Did you bike to work most days last week?” Then we map all the people who said “yes” according to their household’s income.
They’re not the same. The first statement (“the lower…”) implies only correlation. The second statement (“being poor…”) implies causality.
However, you are right that I’m generalizing somewhat by phrasing it as “the lower your income, the likelier you are to ride.” There are definitely peaks and valleys inside each quartile. But there are peaks and valleys in any trend if you slice it thinly enough, and given the relatively high margins of error in this data, I don’t think it’s safe to look much more finely grained than quartiles.
Using a bigger data set, like the national commute rates, makes this safer. And check out the income distribution of bikers at the national level: it’s very clearly shifted to lower household incomes than the general commuting population. Given this fact, I think my level of generalization is safe. But you can feel free to disagree.
Michael – I did not claim that young people are more likely to bike. That was an illustration that I made up to explain that the data presented in the article neither proves nor disproves the statement, “the lower your household’s income, the more likely you are to use a bike to get to work.”
That statement may or may not be true, but the data presented does not make that conclusion and it certainly should not be presented as a “fact,” as stated in the following paragraph. It could, however, be presented as a hypothesis that may be validated with additional studies.
The only thing the data says is, if a person is bike commuting in Portland, there is a 34% chance that their household income is <$46k. That is a very different statement than if your household income is <$46k, you are more likely to bike commute.
Otherwise, the data and article bring up some interesting questions.
Jeff, I agree that these statements would have different truth values if we didn’t also know that only 25% of commuters at large have household incomes below $46k. But we do know that. Therefore, both statements are true. No?
Again, I’m not claiming causality. Both statements regard correlation.
The data suggests income *may* be a contributing factor, but does not prove it to be anything but a correlation. For example, let’s assume the underlying cause is actually determined by a person’s age and 50% of commuters are <25 years old (I am making this up). Since younger people tend to make less money and commute by bike more often, we see a higher percentage of bike commuters with lower income – again, I am making up this example to show that we cannot be certain that income levels drive choice of transportation, based on the data presented.
“For example, let’s suppose most of the people that ride to work are under 25. People under 25 don’t, on average, make as much money as someone over 50. Therefore the statement could be, “the lower your household’s AGE, the more likely you are to use a bike to get to work.”
Also, it would be interesting to see a breakdown of bicycle riding using years of education. I suspect that high-education, low-income folks are disproportionately represented among bike riders.
An excellent piece, Michael. Thanks!
How would that chart compare to a household income-only chart, regardless of transportation mode?
I’m Rich because I ride 🙂
Every time some one tries to blame gentrification on cyclists or vice-versa, THROW THESE STATISTICS IN THEIR FACE! The drumbeat of well-meaning but ill-informed citizens against bike transportation can be fought with the truth that low income people benefit from bike safety improvements just as much (or more) than rich people.
As Michael added above, this data is based on household income. I don’t think a single person making $46K/year would be considered low income at all, but that’s how they are categorized on this chart.
But I completely agree with your underlying meaning.
so can we stop hearing the term “gentrification” now when we want to improve bike facilities?
In New Orleans, a city marked by high rates of low income families, the bike transit movement underway is meaningful and inspiring. Thank you for shedding light on the important topic of income and how transit barriers can be somewhat improved with good access to biking.
You’re analysis is really flawed. From this data we don’t know which poor people are commuting by bikes. It could be mostly young college kids living off of loans… not a bad theory considering the extent and impact of gentrification in the inner neighborhoods.
That does nothing to help advocates for bike commuting get insight into the plights of the working poor – including people with young families, eldercare issues, disability issues, and job insecurity. As we know, there are fewer and fewer options for affordable places to live around easily commutable routes for people who fit into the above categories. The outer metro areas are the most difficult to get around in by bike, and the places where Portland’s poor have been pushed to concentrate into.
Lisa, you’re right – we don’t know what sort of poor the poor people in the chart above are. (Though because it includes only workers, young college kids living off of loans wouldn’t actually be included.) The post above doesn’t claim to know this.
If we’d only had this when the Williams/Vancouver conflict was at its peak.
Poor people already know this, Bikeportland.org. It’s not news.
The news is the criminal abuse of transportation development funds dedicated to providing landing strips in far-flung rural counties, providing tax incentives for Walmart to pave over hectares of land for parking lots and the pièce de résistance: six-lane highways through cities.
All kinds of people ride bikes, despite the conspiracy against them.