Vacancy and Rental Yield: Why the Official Bay Area Figures Can't Be Used
The Census rental vacancy rate for San Jose reads 0.5% with a margin of error of 0.9 points. What a vacancy assumption is actually worth, and how much one empty month costs a Bay Area landlord.
On this page
Vacancy is the cost line people most want a real number for, and it is the line where the real numbers are least usable. This page gives the official figures, shows why they cannot carry an underwriting decision, and then shows what one empty month actually costs.
The official figures
The Census Bureau's Housing Vacancy Survey publishes rental vacancy rates for the 75 largest metropolitan areas. For 2026, the two metros covering the nine Bay Area counties:
| Metro | Q1 2026 | Q2 2026 | Q3 2026 |
|---|---|---|---|
| San Francisco-Oakland-Fremont | 5.5% ±2.1 | 4.2% ±1.9 | 4.6% ±1.9 |
| San Jose-Sunnyvale-Santa Clara | 1.7% ±1.9 | 0.5% ±0.9 | 1.8% ±1.9 |
Source: Census Bureau, Housing Vacancy Survey, Table 4, Rental Vacancy Rates for the 75 Largest Metropolitan Statistical Areas.
Those ± figures are the margins of error, and they are the point of this page.
Why these cannot be used as an assumption
The San Jose Q2 reading is 0.5% with a margin of ±0.9 points. The confidence interval therefore runs from below zero to 1.4%. A vacancy rate cannot be negative, so what the survey is really saying is "very low, and we cannot resolve it further".
The quarter-to-quarter moves are smaller than the error bars. San Jose reads 1.7%, then 0.5%, then 1.8%. That looks like a sharp dip and a full recovery. With margins of ±1.9, ±0.9 and ±1.9, every one of those three readings overlaps every other. There is no measured movement here — only sampling noise that looks like a story.
The San Francisco metro is wide too. 4.6% ±1.9 means somewhere between 2.7% and 6.5%. Those 3.8 points of spread are worth about $1,258 of annual rent on an Oakland property — 0.18 percentage points of net yield, or about a third of the entire gap between Oakland's net yield and Hayward's. The error bar on the input is a material fraction of the differences the number is used to judge.
Both metros are also metros. Santa Rosa, Vallejo and Napa are separate metropolitan areas and are not among the 75 largest, so the HVS publishes no figure for them at all — an absence worth knowing rather than filling in with the San Francisco number.
What one empty month actually costs
Vacancy is easier to reason about in months than in percentages. One vacant month is 1/12 of annual rent, so 8.33% — already above this site's 5% assumption.
Here is what a single empty month does to net yield, holding every other cost constant:
| City | Net yield at 5% vacancy | Net yield with one empty month | Cost of the month |
|---|---|---|---|
| Oakley | 3.22% | 3.02% | $1,318 |
| Antioch | 2.93% | 2.74% | $1,126 |
| Oakland | 2.18% | 2.02% | $1,103 |
| Hayward | 1.63% | 1.50% | $1,061 |
| San Jose | 1.18% | 1.08% | $1,433 |
| San Mateo | 0.96% | 0.86% | $1,587 |
The absolute cost rises with rent, as you would expect. But look at the share of net yield it destroys: one empty month costs Oakley 6% of its net yield and San Mateo 10% of its. The same disruption is twice as damaging in the low-yield city, because there is less income for it to come out of.
That is the same compounding that makes expensive markets worse on net yield than on gross — see gross vs net rental yield.
Vacancy is not the same as turnover cost
A vacancy assumption covers lost rent. It does not cover the costs that arrive with a turnover, and those are frequently larger:
- Make-ready: paint, cleaning, carpet or floor repair. A few hundred to a few thousand dollars.
- Leasing commission: in the Bay Area often half a month to a full month's rent, whether paid to a broker or absorbed by a management company.
- Concessions: a free half-month is common in a soft quarter and is a rent reduction by another name.
A property that turns over every two years with one month vacant and a half-month leasing fee is losing about 6.25% of rent a year to turnover alone — before the make-ready. So 5% is a floor for a property with normal turnover, not a cushion.
The one thing that genuinely reduces it
Tenant tenure. Every cost above is a per-turnover cost, so the single largest lever on effective vacancy is how long tenants stay — which is why below-market rent on a long, reliable tenancy frequently beats market rent with annual turnover, and why a buyer inheriting a long-standing tenant is buying something different from an empty unit.
In a rent-stabilised city that tradeoff is not a choice but a structure. See San Francisco rental yield.