San Francisco Property Tax Appeals & Prop 13
Who may be overassessed in San Francisco?
Fourteen percent of the 124,859 homes we analyzed had an assessment above our modeled market range. Condos and peak-era buyers carried most of the exposure.
San Francisco's 2026 property-tax appeal window closes on Tuesday, September 15. Every year, homeowners discover that the city's assessed value is higher than the market evidence for their property. Many never check.
We analyzed 124,859 San Francisco condos and single-family homes using the city's 2025 closed assessment roll and comparable-property evidence available before the valuation cutoff. That covers 65% of all residential tax parcels and 85% of the city's standard condos and single-family dwellings.
Three findings
- 14% had an above-range assessment signal. That is 17,667 homes worth a closer look.
- Condos carried much more exposure. The signal appeared in 25% of condos and 7% of single-family homes.
- When you bought mattered most. Exposure rose sharply among buyers from the second half of the 2010s through the 2021–2022 market peak.
Most homes sit below the modeled market range
Across the analyzed homes, 62% had an assessment below our modeled market range. That position is consistent with a homeowner benefiting from Proposition 13's limit on annual assessment growth. Another 23% were roughly in line. The remaining 14% were above the modeled range.
Assessment position across 124,859 analyzed homes
Most homes sit below the model's market range. Fourteen percent sit above it.
These categories describe the distance between an assessment and a modeled market range. “Above range” is a reason to investigate comparable sales. It does not establish legal eligibility or predict what the Assessment Appeals Board will decide.
Condos are much more exposed
Condos and single-family homes occupy very different positions. One quarter of condos had an assessment above the modeled range. The rate among single-family homes was 7%.
Condos are more exposed to an above-range assessment
Share of each property type in the three modeled assessment positions.
The favorable side shows the reverse. Seventy-seven percent of single-family homes sat below the modeled range, compared with 41% of condos. Long-held houses are more likely to carry a low Prop 13 base. The condo stock is more exposed to recent purchase cycles.
When you bought matters most
The pattern is clearest when we show each transfer year. Condo exposure rose from 21% among 2013 transfers to 40% in 2015, then remained between 40% and 49% through 2022. Single-family exposure rose later and reached roughly 40% among 2021 and 2022 transfers.
Above-range assessment signal by last transfer year
Peak-era buyers show the clearest assessment-gap signal.
The decline for 2023 and 2024 transfers also makes sense. Those owners bought after the earlier market peak, so their assessed values were more likely to begin close to recent market evidence.
Prop 13's advantage compounds with time
Older cohorts show the mirror image. Among 2010 through 2012 transfers, 54% to 70% of condos and more than 91% of single-family homes had assessments below the modeled range. Among 2021 transfers, only 9% of condos and 13% of single-family homes did.
Below-range assessment position by last transfer year
Older cohorts are much more likely to have an assessment below the modeled market range.
Holding longer does not guarantee a favorable assessment gap. The data still explains the mechanism: assessed values generally grow within Proposition 13's cap while market values can move much faster.
The modeled annual tax difference is $58.8 million
Across all 17,667 homes above the modeled market range, the modeled annual tax difference totaled $58.8 million. The median was $1,769 per property. For 96% of these homes, the modeled difference exceeded San Francisco's $120 filing fee.
Modeled annual tax difference per above-range home
The median modeled difference is $1,769. The total across 17,667 homes is $58.8 million per year.
The fee comparison covers the filing fee alone. A homeowner should also weigh the quality of the evidence, preparation time, and the possibility that the Board reaches a different value.
What a homeowner should do
- Check the assessment. Use the free address check or compare the assessment with current market evidence.
- Find genuinely comparable sales. Focus on similar homes sold near the January 1 valuation date. Verify price, sale circumstances, size, location, and material differences.
- File by September 15. The city's regular 2026 appeal period runs from July 2 through September 15. The filing fee is $120.
- Keep the evidence ready. California law generally gives the Board up to two years to hear and decide a timely application, unless the parties agree to an extension.
Our step-by-step filing guide explains the city's form and online portal. The free tool provides an initial screen. A $10 report provides the full comparison set and downloadable evidence for properties with a strong signal.
Before paying someone a percentage of the savings
Several services advertise contingency fees based on first-year savings. Compare that fee with the work involved. A 30% fee on the median modeled difference in this study would be about $531. At the 75th percentile it would be about $1,049.
For an ordinary residential decline-in-value appeal, the self-service path asks you to review the evidence, complete the application, and participate in the process. Learning the process once can also help if you need to file again.
SF Property Tax Appeal is a self-service evidence tool. It does not file an appeal, represent a homeowner, provide an appraisal, or guarantee a reduction.
Method and important limits
The study uses the 2025 closed assessment roll because a complete 2026 city roll with observed subject assessments was not available when the model ran. The citywide rates describe that 2025 snapshot. The deadline and filing instructions describe the 2026 appeal window.
The screen contains 49,374 condos and 75,485 single-family homes. It represents 84.7% of the 147,356 standard condo and dwelling parcels on the official roll and 65.2% of all 191,389 assessor-designated residential parcels. Multi-family homes, TICs, mixed-use property, and other classes sit outside the model.
“Purchase year” is the last recorded transfer year. Some transfers are partial, family-related, or otherwise not open-market purchases. The model cannot observe interior condition, renovations, view, noise, parking, HOA quality, or every other feature that affects value.
Read the full methodology and data sources. Filing facts are from the San Francisco Assessor-Recorder's 2026 guidance. The two-year rule is described by the California Board of Equalization.