Press Release: Spare Partners with Winnipeg Transit Plus to Make Specialized Transit Services More ConvenientLearn More

Go Home

Income and on-demand transit: Modal split

We uncover trends about how mobility relates to income, based on a treasure trove of mobility and economic data we’ve collected in our biannual travel surveys.

Jerome Mayaud

Twice a year, Spare sends out travel surveys to thousands of riders all across the world. These surveys are immensely valuable to planners, because they reveal the true needs and behaviours of riders, and ultimately empower our partners to deliver more sustainable, profitable and equitable transportation for everybody.

With three travel surveys under our belt, we turn our attention to a delicate issue: the relationship between income and transportation.

Research shows that ride sharing is a preserve of the wealthy. Meanwhile, people on lower incomes are disproportionately dependent on conventional public transit, even though it often fails those who need it most. On-demand transit falls on the spectrum between ride sharing and conventional public transit, yet little research has been published on the link between income and on-demand transit use. It’s time to change that!

In this first of two articles, we’ll explore how income affects the modal shift to on-demand transit (the topic of a previous blog post). Specifically, we will:

  • Examine which income groups ride on-demand transit most often;
  • Show how alternative travel opportunities are intricately linked to someone’s income;
  • Assess whether income affects the service quality a rider experiences.

Data incoming!

Income data is often not collected in travel surveys, because it’s a sensitive issue for many people. However, income disparity sits at the heart of transportation inequity – a phenomenon whereby the ‘goods’ and ‘bads’ of transportation are unevenly distributed among different groups of people.

Since many transportation organizations are actively seeking to redress transportation inequity, we thought we’d help them out by collecting that data in our surveys.

We asked our survey respondents from North America (USA and Canada) to anonymously disclose their median household income, alongside other socioeconomic factors. We only asked about income in our second and third surveys (March and November 2021), and a quarter of respondents declined to answer. Still, we’ve so far gathered income data from more than 1,500 respondents!

We linked each respondent’s answers to the specific trip they took at the time, and aggregated the data to respect data privacy issues (see the end of this article for more information).

This allowed us to disaggregate travel choices and behaviour by income band, and to uncover all sorts of tasty data morsels along the way. So let’s dive in!

Come in, whatever your income!

First, let’s explore how income is distributed across our ridership. Of all the respondents who reported a household income across our two surveys, 66% of riders were low-income (household earnings <US$25k per year), 28% were medium-income (US$25k – US$75k), and 6% were high-income (US$100k+). We based our thresholds on the US Census Bureau’s definition of poverty.

We find that low-income people are benefiting disproportionately more from on-demand transit, compared with medium- and high-income people. The distribution of household incomes among our North American riders is almost inverse to the income distribution of the general population in the US. In other words, our riders skew much lower on the income scale than the general population. This is a big win in terms of transportation equity!

Modal shift is intricately linked to income

Whenever on-demand transit is introduced in a neighborhood, it tends to displace other modes – from private cars and carpooling to buses, walking and cycling. To understand this impact, we asked our riders two related questions. First, would they have taken their trip even if our on-demand service hadn’t been available? If so, which transportation mode would they have used instead?

Income has a strong bearing on our riders’ responses to this question. Over 40% of low-income riders said that their on-demand service enabled them to take a trip they wouldn’t have otherwise taken. We call this phenomenon ‘inducing a trip’. In contrast, only 21%–23% of medium- and high-income riders took induced trips.

This finding is hugely encouraging. When marginalized groups are empowered to access more services and opportunities thanks to on-demand transit, it has a hugely positive impact on mental and physical health, recreation, and employment prospects. It makes us happy at Spare, because we’re helping towns and cities become fairer places to live, work and play.

For trips that were not induced, it is crucial to understand the alternative travel modes riders would have taken, because this has implications for the net benefits of on-demand transit.

We found that income affects the alternative transportation modes riders rely on. Low-income riders formed the vast majority of those whose alternative was carpooling, walking, or riding a bus. In contrast, less than half of riders who would have used a private car or a taxi were low-income.

This is not surprising, since wealthier people tend to have better access to private cars and taxis. In fact, it’s encouraging to see that people who would use private vehicles are switching over to pooled, shared transportation. By luring wealthier people away from cars, convenient on-demand services are reducing congestion, greenhouse gas emissions, pollution and noise.

Nonetheless, it’s important to consider how we can best support riders’ transitions away from their usual mode of travel – especially if our services displace active modes such as cycling and walking, which typically have a better environmental and health impact than vehicular transport.

A great outcome, whatever the income

An important clue that you are achieving transportation equity is if the travel experience of marginalized people is the same, or better, than average. As a proxy for quality of experience, we analyzed wait times, walking times and total journey times by income band.

In general, the quality of experience is comparable across the three income bands. We conducted T-tests and found that there was no statistically significant difference between any of the income bands for wait times and walking times, and only a slight statistical difference between the journey times of high-income people and the rest. This is highly encouraging, because it suggests that everyone receives the same quality of service! Nonetheless, some interesting discrepancies stand out.

Low-income riders on average have ~1 minute longer wait times than medium- and high-income riders, and a larger proportion of low-income riders have to walk for more than 1 minute in their trip. To compensate for this slight difference, more vehicles could be artificially rerouted towards low-income neighborhoods at strategic times, ensuring that low-income residents have a better chance of experiencing shorter wait times and walking times.

Strikingly, the bulk of high-income riders have much longer total journey times than the other income groups. Given that the majority of high-income riders use Spare-powered services to commute, these longer journey times likely result from wealthier riders living in suburban areas that are further from job centres, which in turn means the distances they travel are greater than for urban-dwelling, low-income riders.

Arguably, high-income riders benefitting from longer trips and paying the same as low-income riders do for shorter trips is a form of transportation inequity. Means-tested pricing, or distance-based pricing, could also allow operators to recoup fare revenue from high-income riders to subsidize shorter trips by low-income riders.

Transportation doesn’t have to be taxing

As we’ve shown in this blog, your income has a strong bearing on how you choose to travel, which in turn impacts the collective travel experience of riders. Of course, it is incredibly difficult to control the vast range of travel needs, expectations and behaviours, but that’s precisely the point: we shouldn’t be designing one-size-fits-all solutions for transportation.

In fact, we believe that it’s important to recognise the historic failings that arise from unintentionally biased planning decisions, and to try to correct them. Unless we actively choose to build things differently, we continue to shut out under-represented groups from opportunity.

Low-income people have long been forced to rely on cheap, yet often inconvenient, public transit, and this perpetuates inequality. On-demand technology can overhaul access to transportation for low-income riders, and if carefully balanced with the needs of other income bands, it has the potential to address – and redress – fundamental issues of transportation equity.

A final note on bias and privacy

Surveys can be laced with bias if we’re not careful, so we mitigate against it wherever possible. For example, not everyone books trips through an app, so we worked with partner agencies to conduct some surveys by phone – although this is expensive and logistically difficult, so only a small proportion of surveys were conducted this way. Income is a sensitive topic for some people, and around a quarter of our respondents chose not to disclose their income, so their data is missing from our sample. Ultimately, we could only collect data from riders who chose to answer our survey in the first place, so the usual caveats around bias apply.

And last but not least: privacy. Since respondents shared private information in our surveys, we never linked survey responses to individual Rider Identification numbers in our database, nor their addresses or names. This allowed us to maintain user privacy while collecting meaningful insights at the population level.