The Homeless Population Is Aging Faster Than the Country Is. That Changes What Works.

By Common Ladder · July 30, 2026 · 9 min read

Most of the debate about homelessness is framed around a person in their thirties: someone who could work, who might be talked into treatment, whose life could be turned around with the right push. That person exists. But the fastest-growing group in the homeless population looks nothing like him. She is 68, she worked for forty years, she has arthritis and a walker, and she lost her apartment when her husband died and half the household income went with him.

She is not an anecdote standing in for a trend. She is the trend. And a system built around the imagined thirty-year-old — bunk beds, stairs, long intake lines, "housing readiness" — does not work for her.

The numbers, and what they actually say

HUD's 2025 Annual Homelessness Assessment Report, released in May 2026, counted about 745,652 people experiencing homelessness on a single night in January 2025, and adults 55 and over made up nearly 20% of them.¹ The prior year's count, which published age brackets in detail, found more than 104,000 people aged 55–64 and more than 42,000 aged 65 and older — roughly 146,000 people, more than the population of Tempe.¹

The projection is the part that should hold a planner's attention. In 2019 a team led by Dennis Culhane at the University of Pennsylvania modeled aged homelessness in Boston, New York City, and Los Angeles County, and projected the national 65-and-older homeless population rising from about 40,000 in 2017 to roughly 106,000 by 2030. The driver is specific and demographic: a cohort born roughly between 1955 and 1966 that has carried elevated homelessness risk across its entire adult life and is now crossing into old age all at once.² HHS reached the same conclusion in 2023, calling people 50 and older the fastest-growing age group among those experiencing homelessness.⁵

State-level data confirms this isn't a counting artifact. California's 2023 statewide study — the largest representative survey of homeless adults since the 1990s — found a median age of 47, with 44% of participants 50 or older and a median age of 49 among single adults.⁴

Two different groups, one age bracket

It would be easy to assume older homeless adults are simply younger homeless adults who never got out. Some are. But UCSF's HOPE HOME study, which followed 350 adults aged 50 and older for years, found 43% had their first adult episode of homelessness at 50 or later — close to half the group arriving in the system for the first time in the last stretch of their working lives.³

The two groups look different. Those who became homeless before 50 had more of what the literature calls adverse life experience: earlier mental health and substance use problems, incarceration, fewer completed adult milestones. The later-onset group looks much more like the general low-income older population, and their pathway in tends to be a single economic shock rather than a long unraveling — a death or divorce that halves a household income, a layoff at 58 that never turns into a rehire, a rent increase a fixed check cannot absorb.³ The distinction points at different levers: the first group needs deep, long-term support; the second, in many cases, needed a few hundred dollars a month and a landlord who would rent to them.

Their bodies are fifteen years ahead of their birthdays

The most operationally important finding in this literature is this one: the HOPE HOME cohort had a median age of 58 but showed geriatric conditions — falls, mobility limitation, difficulty with daily activities, cognitive impairment — at rates comparable to housed adults in their 70s.³ Homelessness accelerates aging by roughly a decade and a half.

Translate that into a shelter floor plan and the problem becomes obvious. A person with a walker cannot climb into a top bunk. A person who cannot stand for ninety minutes cannot hold a place in an intake line. A person with early cognitive impairment cannot reliably navigate a coordinated entry process built on remembered appointment times, intact paperwork, and callbacks at a specific hour. None of these are refusals of service. They are exclusions produced by design assumptions that were reasonable when the median client was 35.

The same logic runs through the housing side. Permanent supportive housing was designed for adults with disabilities who can live independently with services. A growing share of older homeless adults sit in a gap: too impaired for standard PSH, not impaired enough — or not documented enough — for a nursing facility, and unable to pay for assisted living. HHS's review named that gap explicitly, and framed the population as needing distinct service pathways rather than one model.⁵

Being precise about what this evidence does and doesn't establish

A few honest caveats, because the numbers here get stretched in both directions.

Projections are models, not prophecies. The 106,000 figure comes from a three-city model extrapolated nationally, and depends on assumptions about mortality, housing markets, and program capacity that could move. The cohort logic is the durable part; the decimal places are not.

Point-in-Time age data is a one-night snapshot. It undercounts people who avoid both shelters and visible outdoor locations — and older adults are disproportionately likely to be doubled up, in a car, or hidden. The true 55-plus share is probably higher than the count shows, not lower.

Cost offsets are real but not automatic. Housing an older adult with high medical needs reduces emergency department visits, inpatient days, shelter nights, and sometimes nursing home stays. The National Academies' review found those offsets consistently present and sometimes substantial, but not always enough to cover the full cost of the housing.⁸ "It pays for itself" is the strong version of the claim, and it isn't always true. "It costs less than the crisis version of the same care, and produces a better life" is the defensible one.

And none of this is an argument against serving anyone else. Families, youth, and working-age adults have not stopped needing help. The point is narrower: a system that treats age as an incidental demographic detail rather than a design input will keep building programs that older adults cannot physically use.

What this looks like in Maricopa County

The January 2026 Point-in-Time Count found 9,726 people experiencing homelessness in Maricopa County — 1,417 aged 55–64 and 642 aged 65 or older, about 2,059 people, roughly 21% of the total. People 65 and older were 9% of the sheltered population but only 4% of the unsheltered.⁶ Read one way, that is the system correctly prioritizing frail people for beds. Read another, it is a measure of how fast the desert sorts out who can survive outside.

Phoenix supplies the harshest version of that arithmetic. In Maricopa County's 2025 heat-related deaths report, about 60% of the people who died were 50 or older, and roughly half were experiencing homelessness.⁷ Older bodies regulate heat worse, common cardiac and blood-pressure medications impair thermoregulation further, and a person who cannot walk a mile cannot reach a cooling center a mile away.

The prevention math is equally stark. The federal SSI benefit for an individual is $994 a month in 2026. HUD's FY 2026 fair market rent for a one-bedroom in the Phoenix–Mesa–Chandler metro is about $943.⁹ No amount of budgeting discipline closes that gap: the whole check is roughly the rent. Every senior housed on that income here is housed because of a subsidy, a family member, or a below-market unit — and all three can disappear in a month.

What follows from this

Nothing here calls for a new theory of homelessness. It calls for applying the existing one with a different set of bodies in mind. Prevention has an unusually high return for late-onset older adults, because the shock that displaces them is often small, one-time, and visible in advance — and because once they are outside, their health declines fast. Shelters need accessible beds, ground-floor placement, and intake that does not require standing. Supportive housing needs a tier that can absorb real functional impairment, which in practice means bringing home- and community-based services into it rather than pretending the two systems are separate. And every discharge from a hospital to "unknown" is a person who will be in the count next January.

The demographic wave is not a forecast anymore. Most of the people who will be 65 and homeless in 2030 are already 60 and precarious today.

Have a correction or an updated figure? Email corrections.commonladder@gmail.com.