How Oʻahu's Point-in-Time count works, and what it misses

Counts & data · Published 2026-10-19

Illustration of a clipboard with a pen and a column of ticked checkboxes — The count is one night's tally, completed with surveys and shelter records. Illustration: Pixabay contributor, CC0.

Almost every number quoted about homelessness on Oʻahu traces back to one exercise: the Point-in-Time count, a tally of everyone found experiencing homelessness on a single night in January. It shapes federal funding, headlines and political debate. Knowing how it is produced is the best protection against reading too much, or too little, into it.

A federal requirement, run locally

The count exists because the U.S. Department of Housing and Urban Development (HUD) requires it. Communities that receive HUD homelessness funding are organized into "Continuums of Care", and each one must count the people experiencing homelessness in its area. Hawaiʻi has two. Partners in Care is the Continuum of Care for Oʻahu; Bridging the Gap covers Hawaiʻi Island, Maui and Kauaʻi. The state's homelessness office describes the count as a federally mandated census and notes that its results help secure HUD funding as well as state, local and private money.

HUD's rule has two parts. People staying in shelters must be counted every year. People sleeping outside, in vehicles or in other places not meant for habitation must be counted at least every other year. The state's 2026 report to the Legislature puts it plainly: HUD requires a count of unsheltered individuals at least every other calendar year, and the data from every state is combined into a report to Congress.

Two methods for two populations

People in shelters

The sheltered count does not rely on volunteers knocking on doors. According to Partners in Care's description of the count, sheltered people are counted from the Homeless Management Information System (HMIS), the shared database that emergency shelters and transitional housing programs use to record who is staying with them. If a program reports a person in a bed on the night of the count, that person is counted as sheltered.

People outside

The unsheltered count is the part most people picture: teams of outreach workers and volunteers canvassing the island, from beach parks and stream banks to bus stops and industrial streets. Partners in Care describes it as an island-wide canvass accompanied by a survey that asks about demographics and personal history. Those answers feed the sub-reports on veterans, kūpuna (older adults), Native Hawaiians and other groups that are published alongside the headline figure.

Terms you will see in every report

Sheltered: staying in an emergency shelter, transitional housing or a safe haven on the night of the count.
Unsheltered: sleeping on the street, in a park, a vehicle, an abandoned building or another place not meant for people to live.
Chronically homeless: a HUD category for people with a disabling condition who have been homeless for a long period or repeatedly. The state notes that this status is often underreported.
People in families: members of households with at least one adult and one child.

Why 2025 had no street count

Because the unsheltered count is required only every other year, Partners in Care and Bridging the Gap counted only sheltered people in January 2025. The most recent full counts on Oʻahu therefore took place in January 2024 and January 2026. When an article says the number of unsheltered people "fell by a fifth", it is setting 2026 against 2024, two years apart, not against the previous winter. Our summary of what the 2026 count found keeps those years side by side.

The gap matters for anyone tracking trends. Street homelessness can shift a great deal in two years as encampments are cleared, new shelters open or rents change. A two-year interval smooths over all of that.

What the count does well

Consistency over time. The same basic method has been used for years, so long-run trends, such as the rise of unsheltered homelessness on Oʻahu over the past decade, are meaningful.
Detail. The surveys produce breakdowns that no other source provides at island level: age groups, household types, veteran status, how long people have been homeless and where they were living before.
Geography. Results are reported by region of the island, which helps service providers decide where outreach is needed.

What the count misses

A single night in January cannot capture everyone. The main blind spots are well known to the people who run the count:

People who are hard to find. Someone sleeping in a car on a quiet residential street, deep in a valley or moving between places may never meet a volunteer. Partners in Care itself notes that results depend partly on the number of volunteers.
Doubled-up households. Families staying with relatives or friends because they have nowhere else to go are not counted as homeless under HUD's definition. The state Department of Education uses a broader definition under federal education law, which is why its numbers look very different; see our guide to students without stable housing.
Movement through the year. Many more people experience homelessness at some point in a year than on any given night. The Point-in-Time count is a snapshot, not an annual total.
Changes in method or coverage. A year with more volunteers, better maps or a new survey app may find more people without homelessness actually rising.

Disasters can also distort the picture. After the August 2023 Maui wildfires, thousands of displaced residents were living in non-congregate shelter programs on Maui. The state's report explains that its headline 2024 statewide figure excludes those households, and that including them would add 2,328 households, or 5,245 people.

How to read a new count responsibly

Check which years are being set side by side. On Oʻahu, sheltered-only years such as 2025 cannot be used for unsheltered trends.
Look at sheltered and unsheltered separately. A stable total can hide a large shift between the two, as happened in 2026.
Read the footnotes. Changes in shelter capacity, new programs being added to HMIS or missing survey questions can move numbers on their own.
Use other data too. School records, 211 call data and housing program enrollments each show a different slice of the same problem. Our piece on what happens after shelter covers the system data that track people over a whole year.

Taking part

The unsheltered count depends on volunteers, and Partners in Care recruits them each winter before a street-count year. If you are interested, watch its website in late autumn; our guide to helping well covers what volunteers are usually asked to do and how to approach people with respect.

Sources

Partners in Care, Point in Time Count (method and reports).
State of Hawaiʻi, Statewide Office on Homelessness and Housing Solutions, Point in Time Count.
Statewide Office on Homelessness and Housing Solutions and Hawaiʻi Interagency Council on Homelessness, Report to the 2026 Legislature (PDF).