What public data says about Toronto's housing squeeze
Each study on this site answers one question. Read together, they sketch a system: where homes leak out of the long-term market, where new supply waits, where the harm lands, and where the pressure comes from. Five findings, each pointing at a different lever with a different owner.
1 · Homes are leaking sideways out of the long-term market
My thesis linked entire-home Airbnb activity to rents about 5.6% higher citywide than they would otherwise be, with listings outside the by-law's reach carrying 2.5 times the impact of regulated ones. Two years later, study 04 found the escape route had become the main route: 58.7% of entire homes now sit past the 28-night line where the rules stop, priced at roughly $3,671 a month for a one-bedroom against $2,365 for an average rental condo. These are furnished, lease-free, month-minimum units: housing-shaped supply held at prices most tenants cannot reach. The lever this points toward is the by-law's 28-day boundary itself, which is a city council decision.
2 · New supply waits longest at the stage the city controls least
Study 05 finds 5,065 open applications with a median age of 4.1 years. Among open rezonings, 54.6% have an appeal status, compared with 8.7% of open site plans. This supports investigating appeal-stage processes. Without decision dates or time in each stage, the study cannot estimate how much faster a different process would be.
3 · The harm concentrates at the moment payment breaks
In study 01, the share of renters reporting fair-or-poor mental health climbs from 23% to 37% as the rent bill's share of income grows. But once income itself is held constant, the ratio stops mattering and one thing keeps mattering: renters who actually missed or delayed a housing payment carry 2.3 times the odds of poor mental health, at every income level. If that association runs even partly from hardship to health, the efficient lever is not general affordability rhetoric but the specific moment of arrears: rent banks, eviction prevention, emergency supports. Those programs are mostly municipal and provincial.
4 · Building quality is surprisingly even; heat risk is citywide
Study 03 finds similar average audit scores across neighbourhood income groups and widespread reports of no building-provided air conditioning. Scores rose under the earlier evaluation scheme. Reinspection, changing building coverage, and regression to the mean complicate attribution to enforcement. Scores and cooling provision warrant separate investigation.
5 · Shelter demand is set upstream of the city that pays for it
Study 02 shows refugee shelter counts rising before the 2023 crisis, peaking in 2024, then declining. The model estimates an August 2023 level contrast of +21% relative to non-refugees. Its later trend changes, and the refugees-only model has a near-zero level change. These results motivate closer investigation of intake, capacity, and arrivals; they cannot allocate responsibility to one policy or level of government.
Reading the five together
The squeeze is not one problem with one owner. The data points at a provincial tribunal, a municipal by-law boundary, arrears-moment supports, a citywide heat standard, and a federal-municipal pipeline, five levers held by different governments. That is an argument for precision: knowing which finding belongs to which desk. The same decade also stressed the systems next door; my opioid study tracks one of them.
Honesty carries over from the underlying studies: every claim above is an association measured in one city's public data, each with limits stated on its own page. "Points toward" is meant literally; none of these findings proves a policy will work. What they support is a shortlist worth testing, and the methods on this site are how I would test them.