What Actually Happens When You Hire Someone: 426,336 Harris County Protests

The finding

In 2025, Harris County homeowners filed 426,336 protests on residential property. About four in five were filed by a paid agent. The rest were filed by owners themselves.

Both groups got their appraised value cut by the same percentage.

That is not a rounding coincidence — it holds to three decimal places (4.453% and 4.454%), across 426,000 cases. The median dollar reduction differs by less than 1%.

Agents won more often. They did not win bigger.

The 6.6-point gap in success rate is real and it matters. But on the measure most homeowners care about — how much value came off — representation and self-filing produced the same result in aggregate.

Read this before you conclude anything

This is a single county, a single year, and a comparison the data cannot fully control.

Agents choose their cases. A protest firm working on contingency has every reason to take cases it expects to win and decline the rest. A homeowner protests the house they own. Those are different populations, and no public file separates skill from selection.

But the selection gap is smaller than you might assume. The average agent-filed property was appraised at $490,386; the average owner-filed property at $428,997 — about 14% higher, not multiples. Agents are not working an entirely different market.

Aggregate is not per-case. The 4.45% figure is total value removed divided by total value protested. It is weighted toward larger properties. A per-property average could differ.

One year, one county. Harris publishes this data; the DFW districts do not. Whether the pattern holds elsewhere or in other years is unknown. HCAD’s public data page offers bundles back to 2005, so a multi-year version of this analysis is possible and we intend to build one.

What this data does establish, cleanly, is descriptive: this is how represented and self-filed protests actually resolved in Harris County in 2025, according to the district’s own records.

The part that should change what you do

Buried in the outcome codes is the most actionable number in the file.

Owner-filed protests ended in a no-show 17.7% of the time. Agent-filed protests, 2.9%.

Nearly one in five homeowners who filed a protest did not appear, and got no change as a result. The single biggest difference between hiring someone and doing it yourself, in this data, is not the quality of the argument. It is that somebody shows up.

If you file your own protest, the most valuable thing you can do is attend — in person, by phone, by video, or by written affidavit, all of which Texas law permits.

Two different systems, in the same county

Agents and homeowners do not travel the same path through the protest process.

Homeowners settle online and informally. 64.5% of owner-filed protests resolved informally. About 45% went through HCAD’s iSettle system — the online settlement track where an owner accepts the district’s offer or the district accepts the owner’s opinion of value. Those four iSettle outcome codes account for 41,073 owner protests and almost no agent protests.

Agents go formal, and take a route homeowners effectively cannot. Only 2.6% of agent protests resolved informally. Instead, 66.4% ended with a single outcome code HCAD calls “topline value changed.”

99.9% of topline outcomes belong to agents. It is the dominant resolution path for represented protests and is essentially unavailable to a homeowner filing alone.

Topline resolutions happen without a hearing

The file records an actual hearing date for each protest where a hearing took place. 63.5% of all protests have no hearing date at all — 345,966 of 544,902 were resolved without anyone appearing before the board.

Topline outcomes are two-thirds of that group:

Topline is therefore not an outcome of adjudication. It is a settlement concluded before a hearing occurs — at a scale of 229,421 residential accounts in one county in one year.

What the mechanism appears to be

HCAD’s published code list defines TC as “TOPLINE VALUE CHANGED” and does not explain how such a settlement is reached. We have asked the district and will update this page when they respond.

Texas law provides a mechanism that fits the shape of what the data shows. Tax Code §1.111(e) allows a written agreement between a property owner’s designated agent and the chief appraiser, and the Tax Code elsewhere treats values established by such an agreement as a distinct category — §41.43(a-3)(2) specifically carves them out. An agent representing thousands of accounts can negotiate at portfolio scale in a way an individual owner cannot.

The data is consistent with this: topline outcomes are agent-filed 99.9% of the time, and they conclude without a hearing. That is the shape a negotiated portfolio agreement would take.

This remains an inference from the pattern, not an established fact. HCAD does not publish what a topline settlement is. It should be confirmed with the district directly before publication, and stated as an open question if it cannot be.

What this means for a homeowner deciding

Hiring someone buys three things this data can measure: a somewhat higher chance of getting a reduction, near-certainty that someone shows up, and access to a settlement route that is not practically available to you.

It did not buy a larger percentage reduction in Harris County in 2025.

Filing yourself costs nothing but your time, and the district’s own online settlement system handled nearly half of owner-filed protests. The main way self-filers lost was by not participating.

Most contingency agreements charge a share of first-year tax savings. Whether that share is worth the difference measured here is a judgment about your own time, your own property, and your own willingness to attend a hearing. This page does not answer it, and cannot.

How this was computed

Source: Harris Central Appraisal District public data downloads, arb_hearings_real.txt, tax year 2025, retrieved August 2026. Filtered to state class code A1, single-family residential. 426,336 records.

The file identifies whether each protest was filed by an agent or by the owner. It does not name the agent — HCAD redacts agent identity in the public file, so no firm-level comparison is possible from this source and none is offered here.

Hearing-date analysis uses Actual_Hearing_Date across all property classes, not only A1. Reduction figures use Initial_Appraised_Value and Final_Appraised_Value. Records with non-numeric or zero initial values were excluded: 3 agent records and 14 owner records out of 426,336.

Outcome code definitions are HCAD’s own, from desc_r_25_conclusion_code.txt in the same download.