On paper, Alameda County’s preliminary 2026 Point-in-Time (PIT) homeless count provides local policymakers with a triumphant headline. Official data reveals a countywide total of 8,201 unhoused individuals—a 13% drop from 2024, marking the largest overall reduction since federal mandates began and bringing totals roughly back in line with pre-pandemic baselines. County materials also emphasize a rising sheltered rate, noting that the proportion of the unhoused population in shelters climbed to 37% in 2026, up from 21% in 2019.
However, a closer look at city-by-city metrics exposes an uneven reality. While a massive drop in Oakland pulled down the countywide average, homelessness crept upward in several neighboring municipalities. This fragmentation raises critical questions about whether the ongoing crisis is genuinely being resolved or merely displaced across municipal borders by more aggressive local enforcement.
Methodology Flaws Questioned
The framework producing these metrics remains structurally vulnerable to criticism. Mandated every two years by the U.S. Department of Housing and Urban Development (HUD), the PIT count relies on roughly 1,300 volunteers conducting visual tallies during a single early-morning window in January.
While intended to standardize regional data, experts widely recognize PIT counts as inherent undercounts. By relying on a brief snapshot, the methodology systematically misses individuals who are couch-surfing, doubling up precariously in unstable households, or hiding in industrial corridors to evade enforcement. Rather than an absolute census, the data functions as a conservative estimate easily skewed by morning weather or shifting migration patterns.
Emeryville’s Marginal Dips and the Border Matrix
Hundreds of millions of dollars have been funneled into East Bay interim and supportive housing since the pandemic. In the immediate corridor, two prominent projects came to life to address the shortage: the 90-unit Nellie Hannon Gateway Supportive Housing project on San Pablo Avenue (offering 39 units for those exiting homelessness), and Mandela House, a 105-unit converted Extended Stay America hotel operating right on the Oakland-Emeryville border.
Yet despite these localized capacity boosts, Emeryville’s reductions effectively flatlined—dropping by just four individuals from 38 down to 34, a tiny fraction of the county total. (The city’s unhoused footprint previously peaked at 178 in 2019 before plummeting to 91 in 2022). Meanwhile, Mandela House itself highlights the fragility of hotel-turnkey solutions: as its initial interim shelter contracts approach a May 2026 expiration to transition into permanent supportive housing, unhoused residents have launched public protests, alleging that systemic failures in case management are threatening to push them back onto the street.
The Oakland Displacement Effect and Regional Ripple
The county’s overall decline was overwhelmingly driven by Oakland, which saw its footprint drop nearly 20%, falling from 5,485 in 2024 down to 4,410 in 2026. The causes of this contraction are fiercely debated. Watchdogs at The Oakland Report questioned the timing of the census, noting it occurred directly ahead of major scheduled sweeps. Concurrently, reporting from The Oaklandside noted that aggressive enforcement didn’t necessarily eliminate homelessness but scattered large, visible encampments into smaller, decentralized pockets.
This shifts the crisis into surrounding jurisdictions. Empowered by the Supreme Court’s 2024 Grants Pass ruling—which granted cities the legal authority to enforce anti-camping ordinances regardless of shelter availability—Oakland utilized its encampment management policies with renewed leverage.
Consequently, while the combined “North County” zone (Berkeley, Albany, Emeryville) held nearly flat with a marginal 0.3% net increase, Berkeley’s individual footprint crept up 4% (844 to 880). South of the border, San Leandro saw a 4% increase (284 to 296), and Hayward jumped roughly 6% to 852 individuals. The most dramatic spikes occurred in East County, where Livermore’s unhoused population surged 24% (277 to 343), Pleasanton’s rose 37% (70 to 96), and Dublin’s surged 60% (25 to 40).
Evaluating the complete regional dynamic remains difficult without evaluating adjacent Contra Costa County (including Richmond, El Cerrito, and San Pablo). Because Contra Costa breaks from the traditional biennial cycle to conduct its PIT counts annually, tracking the fluid, real-time migration of unhoused communities across county lines remains a shifting target.
Great article, well put. I oversaw a Homeless Outreach program for the City of Berkeley for years, and it became very clear to me that until we adopt a regional, or even statewide, approach to addressing homelessness we are just kicking the can down the road. Thanks especially for these key paragraphs: “This fragmentation raises critical questions about whether the ongoing crisis is genuinely being resolved or merely displaced across municipal borders by more aggressive local enforcement.”
“Evaluating the complete regional dynamic remains difficult without evaluating adjacent Contra Costa County (including Richmond, El Cerrito, and San Pablo). Because Contra Costa breaks from the traditional biennial cycle to conduct its PIT counts annually, tracking the fluid, real-time migration of unhoused communities across county lines remains a shifting target.”
I find the interpretation of the numbers misleading. Given that the methodology is vulnerable to a lot of happenstance and that Emeryville’s absolute numbers were low, it’s hard to use count-to-count variations to look at a small geography. Looking at more years of data can help smooth out the noise.
Plus, 38 to 34 is not necessarily “flat-lined.” It’s still a 10.5% drop. And given that there was a 1,000 drop in homelessness in Oakland, it’s not like those 1,000 people moved to Emeryville. Not that you should take any individual number too confidently, though the higher numbers in Oakland mean that margin of error probably still gives you a high drop. Again, there is a lot of noise and homelessness is a dynamic and not a static status.
The regional drop is great news, but I agree with the broad themes of your article that we still have a long way to go in terms of getting people housed and preventing people from losing their homes. Plus it is terrible that Contra Costa doesn’t line up their data collection with the rest of the area.