What Is Regression Analysis
Regression analysis is a statistical method that assessors use to estimate property values by identifying relationships between sale prices and property characteristics like square footage, lot size, age, and location. The assessor plugs in data from recent comparable sales, then the regression model calculates how much each feature contributes to the final value. This produces a formula that assigns a predicted value to properties in the jurisdiction.
How Assessors Use It in Mass Appraisal
Most county assessors rely on regression analysis as the foundation of mass appraisal systems. Rather than individually appraising 50,000 properties, they use computer assisted mass appraisal software to run regression models against all properties at once. The model identifies which comparable sales best predict value, then applies those relationships across the entire jurisdiction. This approach keeps appraisals consistent but introduces systematic errors when the model doesn't fit your property type well.
For example, if an assessor's regression model underweights waterfront location because only 15 sales had waterfront access, your waterfront property may be undervalued relative to its comparable sales. Conversely, if the model assumes all garages add $8,000 to value but your garage is detached and rarely used, you may be overvalued.
What You Can Challenge at Board of Review Hearings
During a board of review hearing or formal appeal, you can contest how the regression model valued your property. Effective challenges focus on three areas:
- Comparable sales selection: The model may have excluded recent sales of truly comparable properties or included sales that don't match your property type. If a foreclosure or distressed sale skewed the data, that's worth flagging.
- Feature weightings: Challenge specific assumptions. If the model assigns equal value to all renovations regardless of quality, or doesn't account for your property's unique condition, you have grounds to argue the assessment ratio is too high.
- Assessment ratio accuracy: Your state publishes assessment ratios showing what percentage of market value the assessor used. Illinois requires a 33% assessment ratio, but if your property was assessed at 45% of its market value while neighbors were assessed at 28%, regression analysis didn't work fairly for you.
Building Your Case with Comparable Sales Data
Request the assessor's comparable sales list and regression model documentation. You're entitled to this under most state freedom of information laws. Then find your own recent comparable sales from the past 6 to 12 months, preferably arm's length transactions in similar neighborhoods. If market data shows your property should trade at $320,000 but it's assessed at $400,000 (at a 33% ratio, that implies a $1.2 million value), the regression model failed to capture market reality.
Document any property features the model appears to have ignored. Missing pool, additional lot acreage, or architectural significance all factor into regression weightings. Bring photos, recent appraisals, and repair estimates if your property condition differs materially from the model's assumptions.
Common Questions
- Can I see the actual regression formula the assessor used? In most states, yes. Request the regression coefficients and model statistics from the assessor's office. Look for the R-squared value, which indicates how well the model explains price variation. An R-squared below 0.70 suggests the model has significant predictive problems, strengthening your appeal argument.
- What if the assessor's comparable sales are old or cherry-picked? File a formal objection before your board of review hearing. Provide documentation showing the assessor ignored recent comparable sales or included outliers. Courts have reduced assessments when regression models were based on stale or unrepresentative data.
- Does regression analysis apply to agricultural or industrial property? Regression models work poorly for specialized property types with few comparable sales. If your property falls outside typical categories, the assessor may have forced an inappropriate model. This is one of the strongest appeal arguments because regression analysis has documented limitations for non-standard properties.