Is high rent bad for your health?

The data comes from the Canadian Housing Survey, a large government survey where Statistics Canada asks thousands of households across the country about their housing and their lives. The answers are anonymous, and a public version is free for anyone to download, which is the version I used. Among many other things, the survey records how much of a household's income goes to rent and utilities, and asks people to rate their own mental health.

I sorted 14,110 renter households into four groups, from those spending less than 30% of their income on housing (the usual definition of "affordable") up to those spending more than everything they earn, which happens when people live off savings, loans, or family help.

How to read the first chart: each row is one of those four groups. The bar shows how many renters out of 100 in that group rate their own mental health as fair or poor. A survey can't ask everyone in the country, so there is always some wiggle room around each number; the thin whisker under each bar marks the range the true number very likely falls in.

To separate the two, I used a statistical model that compares like with like: renters with the same income, age, gender, education, and household type, who differ mainly in how big their housing bill is. I also followed Statistics Canada's own recommended method for calculating how certain we can be about each estimate.

How to read the second chart: each row asks "compared with the baseline group, how much do the odds of fair-or-poor mental health change?" A value of 1× means no difference. A value of 2× means the odds are doubled. The whisker again shows the range we can be fairly confident in; if it crosses 1×, the difference could plausibly be zero.

Once income is taken into account, the rent-share groups all land close to 1×: spending half your income on rent, by itself, no longer predicts worse mental health. What stands out instead is the top row. Renters who actually missed or delayed a housing payment in the past year, about 9 in 100, have more than double the odds of rating their mental health poorly, and that holds at every income level. That last phrase is a testable claim rather than a figure of speech, and the technical notes test it.

The same answer, in plain units. Odds ratios take practice to read, so the third chart asks the question again in the units the first chart used: people out of every 100. Each group appears twice. The light dot is the raw share from the first chart. The dark dot is the model's like-for-like estimate: what that group's share would be if all four groups contained the same mix of people by income, age, gender, education, and household type.

Compared like for like, the staircase flattens to about 26, 26, and 28 out of 100 across the first three groups. The raw climb mostly reflected who sits in each group: renters with heavy housing costs tend to have lower incomes, and low income travels with poorer mental health for reasons of its own.

Why does the highest-burden group drop below everyone else? Paying more than 100% of income on shelter is rare (269 households here), and it usually means the household's measured income was unusually low that year: a student living on savings, someone between jobs, one bad year for a business. Incomes that low predict poor mental health all by themselves, and this group's raw 37 in 100 is roughly what such incomes produce. Once income gets that credit, the ratio itself stops adding risk, and the like-for-like estimate even dips below the other groups. The dip is too uncertain to lean on: with so few households, its plausible range runs from about 11 to 26 in 100.

In plain words: the survey suggests the harm shows up less in the size of the bill and more in the moment the bill stops being payable.

How to read this chart: each row compares renters in 2018 (light) with renters in 2022 (dark). Only one row moves. The share of renters rating their mental health fair or poor rose from 16% to 26%, a ten-point deterioration in four years. Meanwhile the housing side stood remarkably still: roughly one renter in eleven missed a payment in both cycles, and almost exactly one third spent past the 30% affordability line in both cycles. Whatever drove the mental-health decline, it was not a wave of new housing hardship.

And the penalty attached to hardship held steady too: renters who missed a payment carried about double the odds of poor mental health in 2018 (2.1 times) just as in 2022 (2.3 times), and the difference between those two numbers is well within survey noise. The technical notes treat it as no change. Put together, the two cycles say the hardship-health link is structural rather than a pandemic artifact: the pandemic raised the floor of distress for everyone, and missing a rent payment cost roughly the same, large amount of wellbeing on top, in both worlds.

This survey is a snapshot of one moment. It can show that missed payments and poor mental health go together; it cannot show which causes which. Struggling with rent can wear a person down, and poor mental health can also make it harder to keep up with bills. Both stories fit this data, and the write-up says so plainly.

A detail I caught while building this: the survey question I first picked as the measure of "housing hardship" turned out, on a careful read of the questionnaire, to measure general money trouble instead. I switched to the housing-specific measure before running anything. Reading the documentation is part of the job.