What data assessors use for mass appraisal models

Last updated 2026-07-11

What is mass appraisal and why does it affect your tax bill?

Mass appraisal is the process of valuing thousands of properties at once with a standardized statistical model instead of visiting each home. Your county assessor doesn't walk through your living room every year. Software takes in data about your property and every nearby sale, runs it through a formula, and produces a value.

The International Association of Assessing Officers (IAAO), the professional body that sets standards for assessors, defines mass appraisal as "the process of valuing a universe of properties as of a given date using standard methods, employing common data, and allowing for statistical testing." ^[1] Every word in that definition is a potential failure point. The data can be wrong. The method can be a bad fit for your neighborhood. The statistical testing can be weak.

The model produces your assessed value. The assessor's office multiplies that by the local assessment ratio, then by the mill rate, and out comes your tax bill. A $50,000 error in assessed value on a property taxed at 2% costs you $1,000 a year. Over a five-year reassessment cycle, that's $5,000 gone.

Knowing what feeds the model is how you find the error that's costing you money.

What are the main types of mass appraisal models assessors use?

Three model types dominate American jurisdictions, and most counties mix them.

The sales comparison (or comparable sales) model is the most common for residential property. It finds recent arm's-length sales of similar homes, adjusts for differences in size, age, condition, and features, and applies the result to unsold properties. It's easy to defend at a hearing because you can trace every adjustment. ^[1]

Multiple regression analysis (MRA) is the statistical engine behind most modern mass appraisal systems. It treats sale price as the dependent variable and uses 20 or more property characteristics as independent variables to build a pricing equation. Apply that equation to every property in the jurisdiction, and every property gets a value. The IAAO's Standard on Mass Appraisal of Real Property calls MRA "one of the most widely used techniques in mass appraisal." ^[1]

Cost approach models handle new construction, special-use properties, and markets with few sales. The assessor estimates the cost to replace the structure, subtracts depreciation, and adds land value. Commercial and industrial properties often get this treatment, and so do rural parcels.

Many counties layer these methods. A county might use MRA to set neighborhood adjustment factors, then run a sales comparison grid on individual parcels. You can find out which model your county uses. Most assessor offices publish their methodology in a document called the assessment technical manual or mass appraisal report. Ask for it. ^[2]

What specific data inputs go into a mass appraisal model?

This is the list you actually need. The IAAO and the Uniform Standards of Professional Appraisal Practice (USPAP) Standard 6, which governs mass appraisal, both require assessors to document the data they use. ^[3] Here are the major categories.

Physical property characteristics are the foundation. Every model uses gross living area (square footage of heated and cooled space), lot size, year built, number of bedrooms, number of bathrooms, garage type and capacity, basement finish percentage, number of stories, exterior material, roof type, and HVAC type. These come from building permits, field inspections, and property record cards. If your property record card says 2,100 square feet when you actually have 1,850, the model over-values you from the first line.

Property condition and quality ratings carry enormous weight. Most assessors assign a quality class (A through F, say, or 1 through 6) and a condition rating (excellent, good, average, fair, poor). These ratings are subjective and often stale. A home rated "good" in 2015 with a $40,000 foundation crack today is still "good" in the model. ^[1]

Location and neighborhood variables include school district, flood zone, proximity to commercial uses, census tract, and assessor-defined neighborhood codes (sometimes called market area codes). The model uses sales inside a neighborhood to set value multipliers. Get your neighborhood code wrong and you're compared to the wrong sales.

Sales data calibrates the model. Assessors collect arm's-length sale prices, filter out non-market transactions (foreclosures, family transfers, estate sales), and use what's left to build or update the pricing equation. The IAAO recommends a sale ratio study using at least 30 usable sales per stratum. ^[1] In thin markets, assessors reach back across multiple years, which produces stale calibrations.

Income and expense data adds a fourth category for commercial property: gross scheduled income, vacancy rates, operating expenses, and capitalization rates. If you own a small rental or commercial building, check the rent and vacancy numbers the assessor used. They're usually generic market assumptions, not your actual figures.

Table: Major data inputs in a residential mass appraisal model

Data Category	Common Variables	Where Errors Hide
Physical characteristics	GLA, lot size, year built, beds, baths	Incorrect square footage; wrong bedroom/bath count
Quality & condition	Quality class, condition rating	Outdated inspection; wrong class for the neighborhood
Location/neighborhood	School district, flood zone, market area code	Wrong neighborhood code; stale flood zone designation
Sales data	Sale price, sale date, arm's-length filter	Non-arm's-length sales included; outdated sales used
Cost data	Replacement cost, depreciation tables	Wrong depreciation schedule; outdated cost tables
Income data (commercial)	Gross rent, vacancy, cap rate	Generic assumptions vs. actual property performance

How do assessors collect this data, and how often is it updated?

Data comes in through four channels, and each has its own error rate.

Building permits trigger updates to physical characteristics. Pull a permit for an addition and the assessor's office gets a copy and updates your record. This works fine for major additions. Smaller projects done without permits never show up, and the reverse happens too: demolitions and removals sometimes go unrecorded, so you can end up taxed on a structure that's gone.

Field inspections send an appraiser to the property to measure the exterior and sometimes photograph the interior. Most jurisdictions run full inspections every 4 to 10 years, and cash-strapped counties stretch it further. ^[2] Between cycles, the physical data is frozen. A kitchen gutted by a fire in year two of a six-year cycle stays coded "good" until the next visit.

Self-reporting shows up in some jurisdictions, mainly for personal property and income-producing property. Some counties mail income and expense questionnaires to commercial owners every year. Skip the return and you can draw a penalty assessment.

Geographic Information Systems (GIS) and aerial imagery are now widespread. Many large jurisdictions use aerial and satellite imagery, sometimes paired with machine learning, to spot construction and update records between inspection cycles. Cook County, Illinois has publicly described using computer vision to identify property improvements. ^[4] These tools shrink the lag but add new error types. A pergola gets read as a room addition.

Sale ratio studies compare assessed values to recent sale prices and tell assessors whether the model is accurate. The IAAO standard calls for a median ratio between 90% and 110%, meaning the median assessed value should land at 90 to 110 percent of actual sale price. ^[1] If your jurisdiction's study shows consistent over-assessment in your neighborhood or property type, that's documented evidence for your appeal.

What statistical standards must a mass appraisal model meet?

The IAAO's Standard on Mass Appraisal of Real Property and USPAP Standard 6 set the bar. ^[1][3] These aren't academic footnotes. Many state statutes require assessors to follow IAAO standards, and missing them can be grounds for an appeal.

Three numbers matter: the median assessment ratio, the coefficient of dispersion (COD), and the price-related differential (PRD).

The median ratio should fall between 0.90 and 1.10, meaning assessed values cluster around 90 to 110 percent of market value. Anything above 1.10 means the jurisdiction is systematically over-assessing.

The COD measures uniformity. IAAO recommends a COD below 15.0 for residential property, and below 10.0 for newer, homogeneous neighborhoods, meaning individual ratios should cluster tightly around the median. ^[1] A high COD means some properties are wildly over-assessed compared to their neighbors. That's your opening. A property assessed at a 1.30 ratio in a jurisdiction with a 1.00 median and a COD of 20 is an outlier, and an outlier is an argument.

The PRD tests whether low-value homes are assessed more heavily than high-value ones, a problem called regressive assessment. An IAAO-compliant PRD falls between 0.98 and 1.03. Studies of big cities have found real regressivity. A 2021 working paper from the Lincoln Institute of Land Policy found evidence of regressive assessment across many large U.S. jurisdictions, meaning lower-value homes often carry higher effective tax rates than expensive ones. ^[5]

States codify these thresholds. Minnesota requires assessors to maintain a sales ratio between 90% and 105% under Minn. Stat. § 273.11, and the Department of Revenue runs annual sales ratio studies by jurisdiction. ^[6] Texas requires that assessed values not exceed 110% of market value under Tax Code § 41.43 and lets taxpayers use the appraisal district's own ratio study as evidence. ^[7] Know your state's statute. It changes what you can put in front of a hearing officer.

Where do mass appraisal models make the most errors for homeowners?

Four error types drive most successful residential appeals.

Wrong square footage tops the list. Assessors measure from the exterior or from permit records, and their figure often sweeps in space that shouldn't count as gross living area: unfinished basements, detached garages, screened porches. Pull your property record card, measure your actual heated and cooled interior yourself, and compare. A 200-square-foot discrepancy at $150 per square foot is $30,000 in assessed value.

Wrong quality or condition rating comes second. These ratings are subjective, often entered by an inspector who spent four minutes on your driveway. Deferred maintenance, an aging roof, dead systems, functional obsolescence: the model probably reflects none of it. Document everything with photos and contractor estimates.

Wrong neighborhood code drops your property into the wrong market area and compares your value to sales in a stronger submarket. This hits transitional areas and homes near neighborhood boundaries hardest.

Stale or incorrect sales data skews the calibration. If the model was built on sales from a peak year before a local decline, every property in that neighborhood carries an inflated base. That's a live issue after any market downturn.

Larger and commercial buildings add two more: wrong depreciation estimates and bad income assumptions. If you own a gwinnett county tax assessor property or you're tracking a cook county tax assessor tax bill, these errors show up often in those jurisdictions' appeal records.

How can you find the exact data the assessor used on your property?

Start with your property record card, also called a property data sheet or field card. This is the source document for your property's characteristics in the model. It's a public record in every state, and most assessor offices post it online. The card shows the square footage, bedroom count, quality rating, condition rating, and effective year built the model used. ^[2]

If your county doesn't post it, request it under your state's public records law. Most states require disclosure within 5 to 10 business days. Take the card home and walk through it room by room.

Next, request the mass appraisal report or assessment technical manual. It explains which model type the county used, which variables are included, how much each variable weighs, and the results of the most recent sale ratio study. Some counties post it. Others require a formal request. Either way, it's public record.

For comparable sales, ask the assessor which sales were used to value your property, or run your own search. Most county websites have a sales search tool. Pull sales from the past 12 to 24 months in your neighborhood or market area. If the model's sales are mismatched in size, age, or location, that's evidence.

For big-county examples, la county property tax and montgomery county property tax both publish detailed methodology documents. The Los Angeles County Assessor publishes its Assessment Practices Survey results, and Montgomery County, Maryland posts its appraisal model parameters online. Look in the assessor's resources or publications section.

TaxFightBack's DIY appeal kit walks through how to pull and annotate a property record card and find the discrepancies the model used to over-value your home. Same process the contingency firms use. No fee.

How does the model handle land value separately from improvements?

Every mass appraisal model splits your total assessed value into two parts: land and improvements (the structure). That split matters for your appeal because the arguments for each are different.

Land value comes from a land residual method or comparable land sales. The assessor finds recent sales of vacant lots, adjusts for size and location, and applies a value per square foot or per acre to your parcel. Land values within a neighborhood are often uniform or run off a simple frontage or size formula.

Improvement value is what the model calculates for the structure using the physical and condition variables above. In most jurisdictions the assessor sets land value first, then derives improvement value as total value minus land.

Land allocation errors matter most for commercial property. If the assessor loads too much value onto the land (which usually can't be depreciated or argued on condition), it's harder to cut your assessment with condition arguments. Check whether your land value tracks recent vacant land sales nearby.

States with Proposition 13-style acquisition value systems, like California, run the model differently. Your base year value is locked at purchase price and can rise only by 2% a year (or CPI, whichever is lower) under California Constitution Article XIII A. ^[8] There the mass appraisal model mainly flags properties where market value has dropped below the factored base year value, which triggers a temporary reduction. santa clara property tax shows how this plays out in a high-value California county.

Does the model treat new construction differently from existing homes?

Yes, and the differences can push errors in either direction.

New construction usually gets a cost approach, especially in the first few years after completion. The assessor uses cost tables (Marshall & Swift or a similar cost service) to estimate replacement cost, then applies minimal depreciation because the building is new. This runs high in markets where construction costs are rising faster than sale prices, because the model treats cost as value. Cost isn't always value.

Existing homes are more likely to go through sales comparison or MRA. But new construction sales can distort the MRA model when they land in the calibration data without enough adjustment. A neighborhood mixing 1960s ranches with newly built infill homes will drag the older homes toward infill prices if the model doesn't control well for effective age and quality.

Effective age (sometimes called effective year built) is how the model applies depreciation to existing homes, and it's different from actual age. A 1975 home fully renovated might carry an effective age of 2010, so it gets far less depreciation than an untouched 1975 home. A neglected 1990 home might carry an effective age of 1975. These ratings are subjective. Check yours.

Can you use the assessor's own model data against them in an appeal?

You can, and it's often your strongest argument.

The assessor's own sale ratio study is the clearest case. If the study shows your neighborhood stratum assessed at a median ratio of 1.18 (18% over market value), you walk into the hearing and point at the assessor's own document as proof the assessment is systematically too high. Texas Tax Code § 41.43 explicitly allows this. ^[7] Minnesota's Department of Revenue publishes local assessor sales ratio studies taxpayers can use the same way. ^[6]

Property record card errors are even stronger because they're provable facts, not opinions. If the card says 2,400 square feet and your interior measures 2,050, the assessor used wrong data. There's no counterargument. You present the measurement, and the assessor either accepts the correction or defends a number they can't support.

Model parameter documents, when you can get them, let you check whether your characteristics were coded correctly. If the model adds a 10% premium for a three-car garage and your card lists a three-car garage when you have a carport, you can calculate the exact dollar impact of that mistake.

For income-producing property, the assessor's income assumptions are public record in most states. If the assessor assumed a 5% vacancy rate and your documented vacancy was 22%, bring the rent rolls and expense statements and override the model's generic numbers.

Working a bexar county tax assessor appeal or a nyc property tax challenge draws on the same category of evidence. The procedural rules for introducing it change by jurisdiction.

What are the limits of mass appraisal models, and what do assessors acknowledge?

Assessors, through the IAAO, openly admit mass appraisal can't match the precision of individual appraisals. The IAAO standard states that mass appraisal "is not expected to achieve the same level of accuracy as individual appraisals." ^[1] The acceptable COD threshold of 15.0 for residential property means the IAAO treats it as acceptable for individual assessments to swing roughly 15% around the median. On any property, that's real money.

Sparse sales markets are a known weakness. Rural counties, unique properties, and thin submarkets don't produce enough sales to calibrate reliably. There, assessors lean harder on cost approach estimates that may miss local buyer behavior.

Rapid market changes create lag. The model is built on past sales, so it always trails current conditions. In a falling market, assessments stay high. In a rising market, they stay low. The 2008 to 2012 housing downturn produced widespread over-assessments because models were still calibrated on 2006 and 2007 peak sales.

Unique properties break the model. Unusual architecture, heavy deferred maintenance, a non-market location, or features buyers systematically discount: the generic adjustments won't capture the full hit. That's exactly when an individual appraisal, used as appeal evidence, earns its cost.

Here's the honest read. The model is a reasonable tool for managing millions of assessments, and it's not infallible. The data going in is imperfect, the ratings are subjective, and the statistics permit a range of individual error. Your job in an appeal is to find where the model went wrong on your specific property and prove it.

Frequently asked questions

What is a property record card and how do I get mine?

A property record card (also called a field card or data sheet) is the document the assessor's office uses to store your property's physical characteristics: square footage, bedroom count, quality rating, condition, and construction type. These are public records in every state. Most county assessor websites let you search by address and download the card as a PDF. If yours isn't online, request it under your state's public records law.

How many comparable sales does an assessor need to build a valid mass appraisal model?

The IAAO Standard on Mass Appraisal of Real Property recommends at least 30 usable arm's-length sales per stratum for statistically reliable results. In thin markets, assessors combine multiple years of sales or use broader geographic strata, which adds uncertainty. If your neighborhood had fewer than 30 sales in the study period, the calibration for your area is weaker and more open to challenge.

What does 'arm's-length sale' mean in the context of mass appraisal?

An arm's-length sale is a transaction between unrelated parties, both acting in their own interest, with no undue pressure on either side. Assessors filter out sales between family members, foreclosures, estate sales, short sales, and deals with unusual financing because these don't reflect open-market value. The remaining arm's-length sales calibrate the model. If too few survive the filter, the model turns unreliable.

What is a coefficient of dispersion (COD) and why does it matter for my appeal?

The COD measures how uniformly properties are assessed relative to market value. The IAAO recommends a COD below 15.0 for typical residential neighborhoods. A high COD means wide variation in individual accuracy. If your county's COD is 20 or above, individual assessments are highly variable, and you have a reasonable statistical basis to argue your property landed on the over-assessed side of that spread.

Can an assessor use automated valuation models (AVMs) or AI tools?

Yes. Many jurisdictions run computer-assisted mass appraisal (CAMA) software that folds in machine learning, aerial imagery analysis, and automated valuation models to update property data between inspection cycles. These tools flag new construction or improvements, but they add new error types, like reading a covered patio as living space. If your assessment jumped with no actual improvements, check whether an automated update wrongly added square footage or changed your condition rating.

How often does the assessor update the mass appraisal model?

It varies by state law and local practice. Some states require annual reassessment (California triggers annual review for market value drops). Others reassess every two years (Texas), every three years (many counties), or every four to six years. Between reassessments, the physical data is largely static, though calibration may update more often using rolling sales. Check your reassessment schedule with the state department of revenue or department of taxation.

Does my property's assessed value have to equal its market value?

Not necessarily. Many states apply an assessment ratio below 100%. A state might assess residential property at 85% of market value. Your assessed value should equal market value times the applicable ratio. What matters for fairness is that all similar properties are assessed at the same ratio. If yours sits at a higher effective ratio than comparable homes, that's the basis for an appeal, whatever the nominal ratio.

What is the price-related differential (PRD) and does it affect homeowners differently based on home value?

The PRD measures whether low-value properties are assessed more heavily (as a percentage of market value) than high-value ones, which is called regressive assessment. An IAAO-compliant PRD falls between 0.98 and 1.03. Research, including work from the Lincoln Institute of Land Policy, found evidence of regressivity in many large U.S. cities. Owners of lower-value homes may be systematically over-assessed relative to owners of expensive homes, and a local PRD above 1.03 supports that argument.

How does the assessor decide what neighborhood code my property gets?

Assessors divide jurisdictions into market areas or neighborhood codes based on geographic boundaries, school districts, land use patterns, and how sales cluster. Properties sharing a code are assumed to move with the market together and are compared to each other in the model. Those boundaries are drawn by the assessor and can be arbitrary. If your property sits near a boundary and is coded into a stronger market area than where it actually competes, its value gets pulled upward by that area's sales.

What is USPAP Standard 6 and does it apply to my county assessor?

USPAP Standard 6 covers mass appraisal and is issued by The Appraisal Foundation, the congressionally authorized body for appraisal standards. It requires assessors to document their data, methodology, and quality controls. Whether your county is legally bound depends on state law. Many states require it directly or by reference to IAAO standards. Check your state's property tax code or the state department of revenue's guidance to see if Standard 6 is mandatory where you live.

If the model clearly has wrong data on my property, how do I correct it before appealing?

Contact the assessor's office and request a data review or informal review. Bring your property record card, your measurements, photos of the discrepancy, and supporting records like permits or appraisals. Many jurisdictions allow an informal correction without a formal appeal. If they correct the data, ask for a revised notice of value before your appeal deadline, and keep the original notice. If they refuse to fix an obvious error, document the refusal and use it as evidence in your formal appeal.

Do commercial properties use the same mass appraisal model as residential homes?

Generally no. Commercial properties usually run through an income approach (capitalizing net operating income) or a cost approach rather than the sales comparison or MRA model used for homes. The income approach requires the assessor to estimate market rent, vacancy, operating expenses, and a capitalization rate. Those assumptions are often generic and may miss your property's actual performance. Presenting real income and expense statements is usually the strongest commercial appeal move.

Is there a way to see the weight the model assigns to each property characteristic?

Sometimes. Mass appraisal reports and assessment technical manuals may include regression coefficients or adjustment tables showing how much each variable adds to or subtracts from value. The report might show that each additional bathroom adds $12,000 in your market area. If you can get these parameters, you can calculate the dollar impact of any error on your record card with precision, which makes for a far more persuasive appeal.

Sources

1. International Association of Assessing Officers, Standard on Mass Appraisal of Real Property (2017): Definition of mass appraisal; MRA as a widely used technique; COD thresholds of 10 and 15; median ratio standard of 90-110%; statement that mass appraisal is not expected to achieve the same accuracy as individual appraisals
2. IAAO, Standard on Property Tax Policy: Field inspection cycles and the role of permit data in updating property records; mass appraisal report as a required documentation element
3. The Appraisal Foundation, Uniform Standards of Professional Appraisal Practice (USPAP) Standard 6: USPAP Standard 6 governs mass appraisal and requires documentation of data, methodology, and quality controls
4. Cook County Assessor's Office, Residential Modeling Methodology: Cook County has described its use of computer vision and machine learning to identify property improvements between inspection cycles
5. Lincoln Institute of Land Policy, Working Paper on Inequity in Property Assessments (2021): Evidence of regressive assessment in many large U.S. jurisdictions, with lower-value homes carrying higher effective tax rates than high-value ones
6. Minnesota Department of Revenue, Sales Ratio Study: Minnesota requires assessors to maintain a sales ratio between 90% and 105% under Minn. Stat. § 273.11; Department publishes annual local assessor sales ratio studies
7. Texas Tax Code § 41.43, Texas Legislature Online: Texas Tax Code § 41.43 allows taxpayers to use the appraisal district's own ratio study as evidence; assessed values must not exceed 110% of market value
8. California State Board of Equalization, California Constitution Article XIII A Overview: Under California Constitution Article XIII A, base year value is locked at purchase price and increases are capped at 2% per year or CPI, whichever is lower
9. IAAO, Glossary for Property Appraisal and Assessment: Definitions of arm's-length sale, effective age, price-related differential, and neighborhood code as used in mass appraisal
10. National Taxpayers Union Foundation, Property Tax Assessment Uniformity Research: Research supporting that assessment errors and lack of uniformity are widespread in mass appraisal jurisdictions
11. Los Angeles County Assessor's Office, Assessment Practices Survey: Los Angeles County Assessor publishes Assessment Practices Survey results documenting model methodology and ratio statistics
12. Montgomery County, Maryland, Department of Assessment and Taxation, Appraisal Manual: Montgomery County, Maryland posts appraisal model parameters and methodology documentation online for public access