When I moved from London to a Southern California suburb, the first thing I noticed was how clean the streets looked. The second thing I noticed was where everyone had gone.
London is not a city that has solved homelessness. Rates are rising. Rough sleeping has increased five years in a row. The British social housing system is straining. There is real poverty and real visible suffering. Anyone who walks through Soho or Camden in February can see it.
But you can see it. That is the part that struck me when I came back.
In the Southern California county where I now live, you mostly cannot. You find encampments behind freeway sound walls, in flood control channels, in patches of dead grass behind big-box parking lots. The rough sleeping happens in cars in church parking lots after dark. Shelter beds sit in industrial parks the bus does not reach. The visible street homelessness that exists concentrates in a few neighborhoods most county residents do not drive through.
The streets look clean because the policy choice was to make them look clean. Not to end homelessness. To make it less visible.
I want to be careful here, because I spent six years working on homelessness in U.S. counties as an improvement advisor with Community Solutions, helping cities and counties reduce chronic homelessness. The Built for Zero methodology that drove that work is real, and the reductions were real. The communities that committed to a quality, by-name list of every person experiencing homelessness, and worked the system around that list, did get people housed.
What I want to draw a line around is something separate. The federal counting infrastructure that measures broader success in this country, the Point-in-Time Count, the visible-encampment metrics that local press track, the political reward a city gets when downtown "looks cleaner," is not the same thing as the underlying outcome. By-name list methodology counts who is actually housed. Local political incentives often count what the median voter sees. The two diverge, and the divergence is widening.
Part of the U.S. policy infrastructure for homelessness is, in effect, a visibility management system: tent enforcement bans, sit-lie ordinances, closure of public restrooms, hostile architecture, displacement of unhoused residents from downtown corridors to areas where the median voter does not have to encounter them.
These are not anti-homelessness policies. They are anti-visibility policies. They reduce what the median voter sees without changing how many people sleep outside.
This is the part I think we should be more honest about in the American public conversation. We have been treating two different problems as if they were the same. One is the suffering of people without homes. The other is the discomfort of housed people who would rather not see that suffering. The second problem is much easier to solve.
In the UK, the visibility-management infrastructure is weaker. Public encampments often go unenforced. City centers are not zoned away from poverty. The London Underground is a public service, not a wealth segregator. The result is that an unhoused person and a housed person share more space. The housed voter is more likely to encounter the policy failure and demand a fix.
I am not making the argument that British homelessness policy is good. It is, in many places, getting worse. I am making the argument that visibility is a design parameter. The U.S. has designed for invisibility while pretending we were designing for solutions.
What this means for funders, agencies, and policy makers: when you read that a U.S. city has reduced street homelessness by some percentage, ask the next question. Did the count go down because people exited homelessness? Or did the count go down because they were pushed out of where the counters were counting? Built-for-Zero communities can give you the first answer. Most municipal headlines cannot.
The Point-in-Time Count is a federal HUD requirement. Volunteers conduct it on a single winter night, mostly on foot, mostly in visible-to-the-counter areas. It was never built to find the encampments tucked behind the sound wall. The federal data is, in effect, also a visibility-managed dataset.
The honest work, the work I think we should be doing more of in this country, is the work of refusing to let visibility stand in for outcomes. That means counting where the counters do not go. It means insisting that the metric for success is not "what does Main Street look like" but "where did this person actually sleep last night." It means letting funders and electeds see what they are paying for.
In our practice, we work on this through Program Assessment sprints designed to find what the system has stopped measuring. Often what the system has stopped measuring is exactly what the population it serves needs the agency to see.
Visibility is a policy choice. We made one. We can make a different one.
