Why housing, and why this shape of it
We did not pick this model because it sounded good. We picked it because the arrangement — a small, sober, well-run house in an ordinary neighborhood — is one of the better-studied interventions in the field, and because the gap it fills in Whatcom County is documented rather than assumed.
Every claim below is sourced and dated. Follow the links; we would rather you check than take our word for it.
For a lot of people, the thing standing between prison and home is a bedroom
Washington will not release someone to an address it has not investigated and approved. That is a reasonable rule with an unreasonable consequence: people who have served their time and earned release stay inside because no approvable address exists. The state pays to hold them there.
$6,298
Washington requires an investigated and approved release address before someone can leave prison. People who cannot produce one stay incarcerated past their earned release date, at a cost to the state of roughly $6,298 per month, for want of a bedroom.
$3.82
Washington's Department of Corrections reports that every dollar spent on reentry programming returns $3.82 to the state. Housing is the part of reentry that everything else depends on.
Give someone a stable address and the other numbers move
Washington has studied this on its own population, and the specific model we operate — a small shared house with rules and no clinical staff — has been through randomized trials. Both point the same direction.
Washington's own research institute found that people who received housing vouchers on release were less likely to experience homelessness, more likely to find work and earn more from it, and less likely to be rearrested.
24 months
Self-run, single-family shared recovery homes are among the most rigorously studied housing models there is, including randomized trials funded by the National Institutes of Health. At two years, residents showed lower substance use, higher employment, and lower incarceration than people receiving usual care, with the benefit growing past the six-month mark.
These findings describe a class of housing, not a guarantee about any one address. A house produces results because of how it is run, which is why our rules and our staffing are published rather than summarised.
Small and well-run turns out to be the whole variable
The fear that housing like this drags down a street is the most common objection we hear, and it is the one with the most research behind it. The consistent finding is that scale and management quality determine the outcome — and that single homes at neighborhood scale are the most benign version of it.
That is not an argument for trusting us. It is an argument for holding us to the conditions the research identifies: stay small, stay well-managed, and stay answerable.
A study of 123 supportive housing developments found that property values within 500 feet did not drop when a development opened, and grew steadily relative to other properties in the neighborhood. Only very large developments showed localized crime effects; small ones did not.
Federally commissioned research on siting housing in residential neighborhoods found that well-managed housing built to the scale of the street it sits on seldom produces negative neighborhood impacts, and that scattered single homes are the most benign arrangement of all.
Whatcom County counts the shortfall every year
This is not a problem imported into the county to be solved by outsiders. The county publishes the numbers, and the people in them are mostly from here.
1 in 5
Whatcom County's supply of affordable and supportive housing is enough for roughly one in five people who need it. 840 people were counted as homeless in the county in 2026.
2 in 3
About two-thirds of people counted as homeless in Whatcom County last listed a Whatcom County address as their last stable housing. These are not people arriving from elsewhere. They are neighbors who lost their footing here.
One house does not close that gap and we will not claim it does. It houses six people at a time who would otherwise be counted in next year's figures, and it is built to be repeated.
We have not opened yet, so we have no results to show you
It would be easy to fill this section with other people's outcomes and let them read as ours. Instead, here is what we will track from the first resident onward, and publish in aggregate once there is enough of it to mean something.
- Housing stability
- Length of stay, and whether a resident left to permanent housing, to another programme, or to nothing.
- Employment
- Employment at intake and at exit, and whether earnings were enough to carry rent independently.
- Programme completion
- How many residents complete a planned stay, and the reasons the others did not. Both halves get reported.
- Returns to custody
- Whether residents remained in the community during their stay and afterward, for as long as we can responsibly follow up.
- Neighborhood relations
- Concerns raised by neighbors, how fast we responded, and what we changed. A number nobody wants to publish is exactly the one worth publishing.
- Property condition
- Maintenance response times and the condition of the building and grounds, inspected on a schedule rather than when somebody complains.
Resident data is reported in aggregate only. No resident is identified, photographed, or written about without specific, informed, written consent that they may withdraw at any time.
See how the model actually runs
The evidence is the argument for this shape of housing. How it works day to day — the counts, the rules, and who answers the phone — is a separate and more concrete question.
See how it works