JL Scoring Engine · Analysis
Over-assessment is a neighborhood problem.
In East Point, nearly one in three homes is taxed on a value the sales evidence doesn't support. In Roswell, it's one in sixteen. And city lines are only the start: neighborhood by neighborhood, the rate runs from 0% to 92%.
About 1 in 11 homes in Fulton County is over-assessed — taxed on a value higher than what similar homes actually sold for. That's the number we've published, and it holds. But it's a county-wide average, and almost nowhere in Fulton County looks like the average.
In East Point, nearly one in three of the homes we can value is over-assessed — 2,286 homes. In South Fulton, more than one in five — 5,341 homes. In Roswell, one in sixteen. And cities are still averages: at the neighborhood level, the rate runs from 0% in some neighborhoods to 92% in others.
Over-assessment isn't a light rain falling evenly across the county. It pools. Where it pools has less to do with any individual house than with how much sales evidence exists around it — the same mechanism we wrote about in the first article in this series, now visible on a map.
The gradient, city by city
Line up Fulton County's cities by over-assessment rate and the county stops looking like one place. One housekeeping note first: the rates below are computed on the 159,870 homes where the sales evidence lets us state a value with confidence — 24,440 of them are over-assessed, which is 15.3% of valued homes and the “1 in 11” of everything we examined.
| City | Homes valued | Over-assessed | Rate |
|---|---|---|---|
| East Point | 7,232 | 2,286 | 31.6% |
| South Fulton | 23,379 | 5,341 | 22.8% |
| Union City | 4,371 | 855 | 19.6% |
| College Park | 1,387 | 236 | 17.0% |
| Fairburn | 2,369 | 370 | 15.6% |
| Atlanta | 58,685 | 8,852 | 15.1% |
| Milton | 6,504 | 940 | 14.5% |
| Johns Creek | 16,817 | 2,100 | 12.5% |
| Hapeville | 1,230 | 152 | 12.4% |
| Alpharetta | 10,953 | 1,333 | 12.2% |
| Sandy Springs | 9,013 | 733 | 8.1% |
| Roswell | 17,276 | 1,111 | 6.4% |
| Fulton County | 159,870 | 24,440 | 15.3% |
At the top, East Point: 31.6% of the homes we can value are over-assessed — 2,286 homes in a city of under 40,000 people. At the bottom, Roswell: 6.4%, or 1,111 homes. A homeowner in East Point is nearly five times as likely to be over-assessed as one in Roswell — under the same tax authority, valued by the same office, in the same year.
The middle of the table matters too. Atlanta sits almost exactly at the county rate — 15.1% — which is what you'd expect from a city large and varied enough to be most of the county average. Hapeville and Alpharetta land at nearly the same rate despite being, by most other measures, very different places. The gradient isn't a north-south morality tale; it's a map of where the county's valuation model has good evidence to work with and where it doesn't.
Share of each city's confidently valued homes found over-assessed; the number beside each bar is the count of over-assessed homes. Dashed line: the county rate, 15.3%.
Neighborhoods, where the spread gets extreme
Cities are still averages, and averages still hide things. East Point's 31.6% is not spread evenly across East Point, any more than the county's 15.3% is spread evenly across the county.
The honest unit of analysis is the neighborhood. Across the hundreds of Fulton neighborhoods large enough to measure, the over-assessment rate spans from 0% to 92%. There are neighborhoods where we found essentially no over-assessment at all, and neighborhoods where nine homes in ten are taxed above what the sales evidence supports. No other way of cutting this data — by home value, by home condition, by anything — produces a spread anywhere near that wide. Geography dominates.
Why? The short version of the first article in this series: the county has to value every home, so it uses a model calibrated to be right on average across the whole market. That works well where sales are dense and homes are alike — the model has plenty of nearby evidence to lean on. It works worst where sales are sparse, homes are older and more varied, and the model has to reach further afield for comparisons. The neighborhoods at the top of the over-assessment range are, overwhelmingly, the second kind of place. The concentration isn't a decision anyone made. It's what “right on average” looks like from the places the average fits worst.
Over-assessment rate (% of the neighborhood's confidently valued homes) — each column counts neighborhoods
485 neighborhoods with at least 100 confidently valued homes. 15 sit at exactly 0%; at the far end, the highest reaches 92%.
Where the over-assessed homes actually are
The rate table tells you where over-assessment is most likely. It does not tell you where most over-assessed homeowners live — and the answer is different, and worth being honest about.
Atlanta, at a rate barely off the county average, contains 8,852 over-assessed homes — about 36% of all the over-assessment in Fulton County, more than South Fulton and East Point combined. South Fulton holds 5,341 (about 22%), East Point 2,286 (about 9%). The highest-rate places and the highest-count places overlap, but they are not the same list, and any serious account of this problem has to carry both numbers at once: East Point has the county's worst odds; Atlanta has the most affected homeowners.
What is all of it worth? Across the 24,440 over-assessed homes, we estimate roughly $18.7 million a year in property taxes above what the sales evidence supports — a typical (median) overcharge of $371 per home per year, with a mean of $765. Those figures account for Fulton's floating homestead exemption, which freezes much of a homesteaded home's bill and so shrinks what an over-assessment actually costs. These are estimates, built from publicly available sales data and subject to verification home by home; individual results vary, and some of these homes, on closer inspection, will turn out to be fairly valued. But the shape of the finding doesn't depend on any single home being right. Over-assessment in Fulton County is real, it is large, and it is anything but evenly distributed.
Notes
- All figures are computed on the JL Scoring Engine v2.3 residential universe as of August 12, 2026: 262,883 homes examined, 159,870 valued with confidence, 24,440 over-assessed (15.3% of valued homes; 9.3% of all homes examined). Homes we could not confidently value are excluded from every figure in this article — exclusion is not a judgment that a home is fairly assessed. Methodology: jasminelane.app/methodology.
- The city table includes the 12 Fulton cities with at least 300 valued homes, covering approximately 99.8% of the valued stock. One further city sat above that floor on earlier substrates and no longer does (note 3).
- Chattahoochee Hills is excluded from the table on data-quality grounds, and on the current book it also falls below the 300-valued-home floor (275 valued). Its measured rate (32.7%) rests on only 90 flagged homes, and our address-matching fails for 61% of homes there — too unreliable a base to publish alongside the others.
- Neighborhood-level rates are computed on neighborhoods with enough valued homes to publish a stable rate; the 0%–92% range refers to those publishable neighborhoods.
- Savings figures are estimates based on publicly available sales data, account for Fulton's base-year floating homestead exemption, and are subject to verification. They describe the over-assessed cohort in aggregate and are not a prediction of any individual appeal outcome.
- Previously in this series: “Right on average. Not on your house.” — the mechanism behind these numbers.