Zillow Zestimate versus assessed value: which is more reliable?

Last updated 2026-07-10

What is a Zestimate and how does Zillow calculate it?

A Zestimate is Zillow's automated estimate of a home's current market value. Zillow builds it from public records, user-submitted data, and a neural network trained on past sale prices. The model weighs factors like square footage, lot size, year built, number of bedrooms and bathrooms, recent nearby sales, and local market trends. Zillow updates the figure daily for most homes and publishes the methodology on its research pages ^[1].

The word 'automated' is doing a lot of work there. No appraiser walks through the property. No one checks whether the county record showing three bedrooms is actually three bedrooms today. If the public record has an error, the Zestimate inherits that error. If your neighbor did a kitchen gut-renovation last year, the model may or may not have captured it.

Zillow covers roughly 104 million homes across the United States ^[1]. That scale is impressive and also the source of its biggest weakness: the model has to generalize across wildly different neighborhoods, property types, and data-quality environments. In dense urban markets with lots of recent sales, it performs reasonably well. In rural areas or markets with thin sales volume, the error rates climb fast.

What is assessed value and how is it different from market value?

Assessed value is the dollar figure your county assessor assigns to your property for tax purposes. In many states it is not meant to equal market value. Most jurisdictions apply an assessment ratio, sometimes called an equalization rate, that sets assessed value as a fixed percentage of estimated market value ^[2].

Say your state uses a 100% assessment ratio and your home's estimated market value is $400,000. Then the assessed value should be $400,000. But if your county uses a 70% ratio (common in states like New York and Illinois), the assessed value should be $280,000 on the same property. Your tax bill is then calculated by multiplying that assessed value by the local mill rate or tax rate.

The key phrase is 'should be.' Assessors reassess properties on different schedules: annually in some counties, every two or three years in others, and as rarely as every six years in parts of the country ^[3]. Between reassessments, a property's assessed value can drift far from current market value in either direction. A hot market like Austin in 2021-2022 saw assessed values lag behind sale prices by tens of thousands of dollars per property. A cooling market can leave assessed values above current market value, which is exactly when you want to appeal.

Cook County, Illinois, assessed residential property at 10% of market value for most of its history before moving toward 100% under state statute, but implementation has been uneven ^[4]. Montgomery County, Maryland, reassesses every three years [see /articles/assessments-explained/montgomery-county-property-tax]. New York City uses a fractional system where Class 1 residential properties are assessed at 6% of market value [see /articles/commercial-property/nyc-property-tax]. The variation is enormous, which is why you cannot compare assessed values across county lines without knowing each jurisdiction's ratio.

How accurate is the Zestimate? What does Zillow's own data say?

Zillow publishes its own accuracy statistics, which is unusual and worth reading carefully. As of its most recent published figures, the nationwide median error rate is 2.4% for on-market (listed) homes and 7.49% for off-market homes ^[1].

Median error means half of all Zestimates fall within that percentage of the eventual sale price, and half fall outside it. For an off-market home valued at $400,000, a 7.49% median error translates to roughly $30,000 in either direction just to hit the median. A quarter of estimates are worse than that.

State-level performance varies a lot. Zillow's own accuracy table shows some states achieve median errors below 3% for off-market homes while others exceed 10%. Markets with infrequent sales, unique properties, or older public records tend to perform worst. Rural areas in states like West Virginia, Mississippi, or parts of the Mountain West routinely see higher error rates than coastal metros with deep, well-maintained MLS data.

The practical implication: a Zestimate is a reasonable first-pass sanity check on whether your assessed value is in the right ballpark. It is not evidence you can bring to a board of equalization and win with. Assessors and appeal boards know its limitations, and many will say so out loud if you cite it.

How accurate is assessed value compared to actual market value?

Assessed value accuracy depends almost entirely on how recently the assessor updated the roll and how well the assessor's office is funded and staffed. A 2020 study by the Lincoln Institute of Land Policy found that assessment uniformity varies dramatically across jurisdictions, with coefficients of dispersion (a standard measure of assessment uniformity) ranging from under 5% in well-run offices to over 30% in poorly-run ones ^[3]. A coefficient of dispersion under 15% is generally considered acceptable by the International Association of Assessing Officers.

The more important question for a property tax appeal is not whether assessed value matches market value in absolute terms, but whether your property is assessed at a higher ratio than comparable properties in your jurisdiction. That's the legal standard in most states: you're entitled to uniform treatment, not necessarily a perfect appraisal ^[2].

So an assessed value that is 10% above market value may still be legally correct if every other home in your neighborhood is also assessed 10% above market value. The appeal wins when your assessed ratio is higher than the median ratio for comparable properties. That's why comparable sales data matters far more than either the Zestimate or the assessed value in isolation.

Which is more reliable for a property tax appeal?

Neither one wins outright, but for different reasons.

The Zestimate updates frequently and captures recent market movement. If your assessed value was set 18 months ago and the market has dropped 12% since then, a Zestimate trending downward is at least consistent evidence that something has changed. But it's a black box. You cannot explain to an appeals board exactly why Zillow's model produced that number, and you cannot audit or replicate it.

Assessed value has the advantage of being the legally operative number. Your tax bill starts there. But it can be stale, it can reflect errors in the property record, and it can be applied inconsistently across your neighborhood.

For an actual appeal, what you need is real comparable sales: arms-length transactions of similar properties that closed within the past six to twelve months (most state statutes specify the relevant date range). Recent sales are primary evidence. Everything else is secondary.

A Zestimate can tell you whether to bother looking at comps. If your assessed value is $350,000 and Zillow says $340,000, the gap probably doesn't justify spending a Saturday pulling sales data. If your assessed value is $350,000 and Zillow says $295,000, that's a flag worth chasing down. Think of it as a free screening tool, not as evidence.

The Lincoln Institute found that overassessment is most common for lower-value properties and that Black and Hispanic homeowners face higher effective tax rates on average than white homeowners in the same jurisdiction, in part because of assessment inequities ^[3]. That context matters. Systematic errors in assessed values are real and documented, which is exactly why the appeals process exists.

Can you use a Zestimate as evidence in a tax appeal?

In nearly every jurisdiction: no, not on its own. Boards of equalization, assessment review boards, and tax courts want documented sales of comparable properties, certified appraisals, or both ^[2]. A printout of a Zestimate is not a certified appraisal and it's not a sale record. Some hearing officers will look at it as background context. Most will give it little or no weight as standalone evidence.

That said, a Zestimate screenshot alongside three or four real comparable sales that support the same value conclusion can help your narrative. It's corroborating, not controlling. Don't lead with it.

What actually works in most jurisdictions: 1. Pull sold comparables from your county assessor's public sales database or from a free MLS tool like Redfin or Realtor.com (mark them as sold, not listed). 2. Find three to five homes similar to yours in size, age, condition, and location that sold in the past six to twelve months. 3. Calculate the assessed-value-to-sale-price ratio for each comparable property. If those ratios are consistently lower than your own ratio, you have a uniformity argument. 4. Submit those comps on the official appeal form with the filing.

If you want step-by-step guidance on structuring that evidence package, the TaxFightBack DIY appeal kit walks through the exact form structure used in the most common state appeal systems, including how to present comps for maximum clarity.

For specific local processes, see how Cook County handles evidence submissions [see /articles/appeal-process/cook-county-tax-assessor-tax-bill] or how Gwinnett County evaluates residential appeals [see /articles/appeal-process/gwinnett-county-tax-assessor].

What does it mean when your Zestimate is lower than your assessed value?

It means you might have a case worth investigating. If Zillow's model, which is trying to estimate market value, thinks your home is worth less than what your assessor says it's worth, that's a sign your assessed value may be set too high relative to current market conditions.

The gap has to be meaningful to justify an appeal. A $5,000 difference on a $400,000 home probably isn't worth the time. A $40,000 difference is. Run the math on what the tax savings would be if you got your assessed value reduced to match the Zestimate. Take your local mill rate, multiply by the proposed reduction, and see what you'd save annually. If it's $600 or more, the appeal is usually worth filing.

Then verify with real comps before you file. The Zestimate might be wrong. If you pull five comparable sales and they cluster around $395,000 on your $400,000 assessed home, the Zestimate was misleading you and the appeal won't succeed. But if the comps cluster around $340,000, you have both the Zestimate and the sales data pointing the same direction, and that's a strong position.

Bexar County, Texas, has seen rapid assessment increases in recent years, and many homeowners found significant gaps between Zestimates and assessed values during the 2022-2024 period [see /articles/appeal-process/bexar-county-tax-assessor]. The same pattern showed up in Gwinnett County, Georgia and across much of Los Angeles County [see /articles/appeal-process/los-angeles-county-property-tax].

What does it mean when your assessed value is lower than your Zestimate?

It usually means you're in a favorable position and you should not appeal. If the assessor's value is below current market value, your effective tax rate is lower than it would be at full market value. Appealing and winning would reduce your taxes, but the assessor could also respond to new information by raising your value on the next assessment cycle.

There's a more important caveat. In some states, most famously California under Proposition 13, assessed value is capped at the purchase price plus a small annual inflation adjustment regardless of market appreciation ^[5]. Californians who bought a decade ago often have assessed values far below current Zestimates. That's the law working as intended, not an error. Santa Clara County homeowners, for instance, routinely see Zestimates two or three times their assessed value [see /articles/commercial-property/santa-clara-property-tax]. Appealing would accomplish nothing and could theoretically invite a reassessment trigger.

Other acquisition-value states include Michigan, which uses a 'state equalized value' capped at 50% of market value with annual increases limited to 5% or CPI, whichever is lower ^[6]. Know your state's rules before assuming any gap between Zestimate and assessed value is a problem.

How do assessment ratios affect the comparison between Zestimate and assessed value?

This is where most homeowners make the comparison wrong. They look at the Zestimate ($350,000), look at the assessed value ($210,000), and think the assessor is being generous. But if the county's assessment ratio is 60%, then the assessor's implied market value estimate is $350,000 ($210,000 divided by 0.60). The two numbers agree exactly.

To make a fair comparison between what the assessor thinks your home is worth and what Zillow thinks it's worth, you have to convert assessed value to its implied market value. The formula is simple:

Implied market value = Assessed value / Assessment ratio

Your county's assessment ratio is usually published on the assessor's website or in the state department of revenue's equalization study. In states that do annual equalization studies (most do), the current ratio is a public document.

Jurisdiction	Assessment Ratio	Assessed Value	Implied Market Value
Example: 100% ratio state	100%	$350,000	$350,000
Example: 70% ratio state	70%	$245,000	$350,000
Example: 60% ratio state	60%	$210,000	$350,000
Example: 10% ratio state	10%	$35,000	$350,000

Once you've done that conversion, you can compare the assessor's implied market value to the Zestimate and to your actual comp sales on equal footing. Skip this step and you get one of two bad outcomes: false alarms (panic over an 'inflated' assessed value that's actually just a 100% ratio jurisdiction) or missed appeals (overlooking an overassessment because the raw assessed number looks small).

How often is the Zestimate updated versus how often assessed value changes?

Zestimates update daily for most properties. The model ingests new sales, new listing data, and market trend signals continuously. A Zestimate from today reflects the market as it looked roughly in the past few weeks.

Assessed values change on whatever schedule your jurisdiction uses. Annual reassessment is common in states like Texas, North Carolina, and Ohio. Triennial reassessment (every three years) is standard in Illinois and Maryland ^[4]. New York City reassesses annually for most classes. California reassesses only on change of ownership or new construction, with that 1-2% annual cap otherwise ^[5].

Here's the practical result. In a rapidly appreciating or depreciating market, Zestimates and assessed values can diverge sharply within a single reassessment cycle. The 2020-2022 run-up in home prices left many assessed values well below Zestimates in markets that reassess every three years. The partial cooling of 2022-2023 then left some recently reassessed properties with assessed values above current Zestimates. Timing matters a great deal.

For homeowners in states with less frequent reassessment, the comparison is most useful right after a reassessment notice arrives in the mail. That's your legal window to appeal, and the Zestimate from that date is your first free data point.

Are there better free tools than Zillow for checking your assessed value's accuracy?

Yes. A few are more reliable for the specific task of a tax appeal.

Your county assessor's own website is the primary source. Most counties now publish their full sales database online, searchable by address or parcel number. This gives you actual transaction prices for real comparable properties in your neighborhood, which is the gold standard for appeal evidence. Hennepin County, Minnesota, for example, publishes its sales data publicly through the county's GIS portal [see /articles/commercial-property/hennepin-county-property-tax].

Redfin's sold listings are another solid free option. Redfin pulls from MLS data and lets you filter by sold date, property type, square footage, and location radius. The sale prices are real MLS-recorded transactions, not estimates. That's the data you bring to an appeal board.

Realtor.com sold listings work the same way. Zillow's sold listings (not the Zestimate, but the actual sold price history on each listing) are also reasonably accurate for MLS-recorded sales.

For certified appraisals, expect to pay $400 to $700 for a residential appraisal from a state-licensed appraiser ^[7]. A certified appraisal carries far more weight at a formal hearing than any AVM (automated valuation model) including the Zestimate. If the potential tax savings are large enough to justify the cost, a certified appraisal is the strongest evidence you can bring.

The NBER working paper by Avenancio-León and Howard (2022) documented that AVM errors are not random: they tend to be systematically higher for lower-value properties, which may compound assessment inequities already documented in the Lincoln Institute research ^[8].

What does the research say about AVM accuracy versus appraisals?

Automated valuation models (AVMs), which include the Zestimate, have been studied fairly extensively because mortgage lenders use them for origination decisions. The Consumer Financial Protection Bureau has noted that AVMs can introduce bias and inaccuracy that affects loan decisions, and it issued a proposed rule in 2023 to add quality controls for AVMs used in mortgage origination ^[9].

A 2021 study in the Journal of Housing Research found that AVM accuracy degrades significantly for properties that differ from the norm in their market: unusual architecture, large lots in dense areas, properties with accessory dwelling units, or homes with deferred maintenance that doesn't show up in public records ^[10]. Those are precisely the kinds of properties where a Zestimate is least trustworthy.

For tax appeal purposes, the key finding from the literature is that AVMs are calibrated to the median, which means they systematically underestimate high-end properties and overestimate low-end properties in the same market. If your home is on the lower end of your market, the Zestimate may actually be too high, which would lead you to underestimate the gap between assessed value and true market value. If you own a high-value property, the Zestimate may be too low.

None of this means AVMs are useless. It means you use them as a first screen, then replace them with actual sales data before making any decisions.

The TaxFightBack approach is the same one most experienced tax consultants use: Zestimate to decide whether to look further, comp sales to decide whether to file, certified appraisal if the stakes are high enough to justify the cost.

Frequently asked questions

Is a Zestimate admissible as evidence in a property tax appeal?

Almost never on its own. Boards of equalization and tax courts want documented comparable sales or a certified appraisal. A Zestimate is an automated estimate with a published median error of 7.49% for off-market homes. It can support your narrative alongside real comp data, but if you submit it as your primary evidence, most hearing officers will give it little or no weight.

Why is my assessed value so much lower than my Zestimate?

Two likely reasons: your state may use a fractional assessment ratio (so a $200,000 assessed value on a 60% ratio implies a $333,000 market value estimate, not $200,000), or your jurisdiction hasn't reassessed recently and market values have climbed faster than the assessment roll was updated. Convert assessed value to implied market value by dividing by your assessment ratio before comparing to the Zestimate.

Why is my assessed value higher than my Zestimate?

This is the situation worth acting on. It may mean the assessor overestimated your home's value, or that the market has softened since your last reassessment. Pull three to five comparable sales from the past six to twelve months. If those comps cluster near or below the Zestimate, you have grounds to appeal. File before your jurisdiction's appeal deadline, which is typically 30 to 90 days after your assessment notice.

How often does Zillow update the Zestimate?

Zillow updates Zestimates daily for most properties. The model ingests recent sale data, active listing data, and market trends continuously. Assessed values, by contrast, change only when the assessor reassesses your property, which can happen annually, every three years, or as rarely as every six years depending on your state and county.

What is the Zestimate's median error rate?

Zillow's own published accuracy figures show a median error rate of 2.4% for on-market (listed) homes and 7.49% for off-market homes nationwide. State-level performance varies. Markets with thin sales volume, unique property types, or weak public records consistently show higher errors. These figures come from Zillow's research page and are updated periodically.

Can I use Redfin's estimate instead of Zillow for a tax appeal?

Redfin's AVM has similar limitations to the Zestimate: it's automated, not auditable, and carries no legal weight as standalone evidence. Redfin's sold listings, however, are excellent comp sources because they pull from MLS transaction data. For appeal purposes, use any AVM as a screening tool and rely on actual sold comps for your evidence package.

What is an assessment ratio and how do I find mine?

An assessment ratio is the percentage of market value at which your jurisdiction is supposed to assess properties. It ranges from 100% in states like Texas to as low as 6% for some NYC residential classes. Your county assessor's website usually publishes the current ratio, or your state department of revenue's annual equalization study lists it. Divide your assessed value by that ratio to get the assessor's implied market value estimate.

How do I know if my property is overassessed?

Compare your assessed value (converted to implied market value using your assessment ratio) to recent sale prices of similar homes nearby. If comparable homes sold for 10% or more below your implied market value, you likely have an overassessment. Also check whether comparable properties in your neighborhood carry lower assessment ratios than yours, which is a uniformity argument. Both grounds are valid in most states.

Does California's Prop 13 mean my Zestimate is always higher than my assessed value?

Often, yes. Under Proposition 13, assessed value is set at the purchase price and can increase only 2% per year or CPI, whichever is lower, until a change of ownership triggers reassessment. Homeowners who bought years or decades ago frequently have assessed values far below current market value and Zestimates. In those cases, the Zestimate being higher is by design, not an error or an opportunity to appeal.

What is a better way to estimate market value than the Zestimate for appeal purposes?

Pull actual sold comparable sales from your county assessor's public database or from Redfin or Realtor.com filtered to sold (not active) listings. Find three to five homes similar in size, age, and location that closed in the past six to twelve months. Those are real transactions an appeal board will accept. A state-licensed appraisal ($400-$700 typical cost) is the strongest single piece of evidence you can submit.

How much can a property tax appeal save me?

Savings depend on your local mill rate and the size of the assessment reduction. A $30,000 reduction in assessed value in a jurisdiction with a 2% effective tax rate saves $600 per year. Most successful residential appeals reduce assessed value by 5% to 15%, according to general industry experience, though hard national data is scarce. The Lincoln Institute documents that assessment errors can be substantial, particularly for lower-value properties.

Do I need a lawyer or professional to appeal my property taxes?

No. Most jurisdictions design their first-level appeal (the informal hearing or board of review) specifically for homeowners to handle themselves. You need a clear comp analysis and the right forms filed before the deadline. Contingency firms take 30% to 50% of the first year's savings. A DIY approach keeps the full savings. Hiring counsel only makes sense if the assessment is very large or the case goes to tax court.

Is the Zestimate more accurate in cities or rural areas?

Cities, by a significant margin. Dense urban markets have more recent sales of similar properties, stronger MLS data, and more consistent property types, all of which improve AVM accuracy. Rural areas have fewer comparable sales, more unique properties, and weaker public record quality. Zillow's own state-level accuracy tables confirm this pattern: states with large rural populations show higher median error rates.

Can the assessor raise my value if I appeal and lose?

In most states, no: the principle of no-increase-on-appeal protects homeowners who file in good faith. But in some jurisdictions, assessors can review your file and raise the value if they find your property is underassessed. Check your state statutes before filing. This risk is uncommon in practice but worth understanding, particularly if your Zestimate is significantly higher than your current assessed value.

Sources

1. Zillow Research, Zestimate Accuracy: Nationwide median Zestimate error rate is 2.4% for on-market homes and 7.49% for off-market homes; Zillow covers roughly 104 million homes.
2. International Association of Assessing Officers (IAAO), Standard on Ratio Studies: Assessment ratio is the legally defined fraction of market value used to set taxable assessed value; uniform treatment is the standard for appeal.
3. Lincoln Institute of Land Policy, Measuring Up: The Assessment Burden Across the United States (2020): Coefficients of dispersion in assessment quality range from under 5% to over 30% across jurisdictions; overassessment is most common for lower-value properties and disproportionately affects Black and Hispanic homeowners.
4. Cook County Assessor's Office, Illinois: Cook County historically assessed residential property at 10% of market value; triennial reassessment cycle is standard in Illinois.
5. California State Board of Equalization, Proposition 13 Overview: Under Proposition 13, California assessed value is set at purchase price with annual increases capped at 2% or CPI, whichever is lower, until change of ownership.
6. Michigan Department of Treasury, Property Tax Estimator and Millage Rates: Michigan caps taxable value increases at 5% or CPI annually; state equalized value is set at 50% of market value.
7. Appraisal Institute, Guide to Residential Appraisals: Typical cost for a state-licensed residential appraisal is $400 to $700.
8. NBER Working Paper No. 29695, Avenancio-León & Howard (2022), The Assessment Gap: AVM errors are not random; they are systematically higher for lower-value properties, which compounds documented assessment inequities.
9. Consumer Financial Protection Bureau, Proposed Rule on AVM Quality Control Standards (2023): CFPB proposed quality control rules for AVMs used in mortgage origination in 2023, citing bias and inaccuracy concerns.
10. Journal of Housing Research, AVM Accuracy for Non-Standard Properties (2021): AVM accuracy degrades significantly for properties with unusual architecture, large lots, ADUs, or deferred maintenance not captured in public records.
11. New York City Department of Finance, Property Tax Rates and Assessment Classes: NYC Class 1 residential properties are assessed at 6% of market value under the city's fractional assessment system.
12. National Association of Realtors, Appraiser and Valuation Issues: Comparable sales (arms-length transactions of similar properties) are the primary evidence standard in most residential appraisals and tax appeal proceedings.