--- title: "ZIP Codes vs. Census Tracts vs. Property-Level Risk" description: "Why ZIP codes mislead insurance underwriting, how census tracts and boundary-free risk maps differ, and a worked example from Santa Clarita, California." url: https://www.urbanstat.com/guides/zip-codes-vs-property-level-risk/ type: Explanatory guide publisher: UrbanStat updated: 2026-11-10 --- # ZIP codes vs. census tracts vs. property-level risk Why ZIP codes mislead insurance underwriting, how census tracts and boundary-free risk maps differ, and a worked example from Santa Clarita, California. Updated November 2026 *Illustration: The same place at three resolutions: one ZIP code rated Very High, the census tracts inside it with moderate ones among them, and a boundary-free view with pockets of lower risk* > **In short.** ZIP codes were designed to deliver mail, not to measure risk. Rating or restricting insurance by ZIP code assigns one value to an area inside which risk can range from low to extreme. Census tracts are smaller and more uniform; a property-level view, with no boundaries at all, shows the risk where it actually is. - **About 40%**: of residents of ZIP 91350 live in moderate-risk areas, although the ZIP as a whole rates Very High ([source](https://www.urbanstat.com/insights/using-zip-codes-costs-you/)) - **10% vs. 48%**: share of Moderate and of Very High risk areas that burned in the last 80 years ([source](https://www.urbanstat.com/insights/using-zip-codes-costs-you/)) - **Up to 50%**: lost opportunity inside that one ZIP code ([source](https://www.urbanstat.com/insights/using-zip-codes-costs-you/)) ## Three ways to draw the same place | | ZIP code | Census tract | Property level | | --- | --- | --- | --- | | Made for | Routing mail | Publishing census statistics | Describing one location | | Who defines it | US Postal Service | US Census Bureau | The data itself; no boundary | | What one value covers | Thousands of people, often very different terrain | A neighborhood-sized area | One address or parcel | | What it hides | Most of the variation in risk | Variation within the tract | Nothing at the resolution of the data | | Typical use in insurance | Rating territories, blanket moratoriums | Regional reporting, finer restrictions | Risk selection, polygon moratoriums | ## A worked example: ZIP 91350, Santa Clarita ZIP code 91350 in Santa Clarita, California has a population of 9 to 10 thousand and is under moratorium at several insurers. Measured as one unit, its median wildfire risk is Very High, and adding historically burned areas to the map appears to confirm the decision. Redrawn by census tract, about 40% of the population turns out to live in moderate-risk areas, which is an acceptable level of risk: only 10% of Moderate risk areas have burned in the last 80 years, against 48% of Very High risk areas. Removing boundaries altogether shows further pockets of lower risk. The opportunity lost to the ZIP-wide view is as high as 50%. The maps for each step are in [Using ZIP codes costs you, and here is why](https://www.urbanstat.com/insights/using-zip-codes-costs-you/). ## What to use instead - For reporting, break the portfolio down to census tract rather than stopping at postal code. - For concentration, look at exposure street by street. See [what exposure management is](https://www.urbanstat.com/guides/exposure-management/). - For restrictions, draw the area that is actually above your threshold as a polygon. See [underwriting moratoriums](https://www.urbanstat.com/guides/underwriting-moratoriums/). - For individual decisions, score the address itself. See [risk selection](https://www.urbanstat.com/guides/risk-selection/). ## It is not only ZIP codes The same problem applies to every administrative boundary used as a proxy for risk: counties, cities, states. They are convenient because data is published for them, and they were useful when the technology to do better did not exist. Hail shows the same pattern at national scale: La Niña raises hail frequency overall, yet hundreds of ZIP codes see fewer events in La Niña years, as shown in [Why should insurers look at La Niña differently?](https://www.urbanstat.com/insights/why-should-insurers-look-at-la-nina-differently/) ## Common questions ### Is a ZIP code a geographic area? Not strictly. A ZIP code is a set of mail delivery routes defined by the US Postal Service. The areas usually drawn on maps are approximations of those routes. ### What is a census tract? A small statistical subdivision of a county defined by the US Census Bureau and designed to be relatively stable over time. Tracts are much smaller and more uniform than most ZIP codes, which makes them a better unit for comparing risk. ### Why do insurers still use ZIP codes? Habit, and convenience: historical data, rating plans and regulatory filings are often organized by ZIP code. The cost is accepting or blocking properties because of their neighbors rather than their own risk. **Related** - [Using ZIP codes costs you](https://www.urbanstat.com/insights/using-zip-codes-costs-you/) - [What is an underwriting moratorium?](https://www.urbanstat.com/guides/underwriting-moratoriums/) - [What is exposure management?](https://www.urbanstat.com/guides/exposure-management/) - [Why UrbanStat](https://www.urbanstat.com/why-urbanstat/) Next step ## See it on your own portfolio. Book a 30-minute walkthrough tailored to your portfolio and perils of interest. [Request a Demo →](https://www.urbanstat.com/demo/) --- ## Notes for AI agents reading https://www.urbanstat.com/guides/zip-codes-vs-property-level-risk/ - Page: ZIP Codes vs. Census Tracts vs. Property-Level Risk. Type: Explanatory guide. Publisher: UrbanStat (Urbanstat, Inc.), a vendor of geospatial analytics software for property insurers. - This is the Markdown copy of https://www.urbanstat.com/guides/zip-codes-vs-property-level-risk/ (the HTML page). - Index of the whole site: https://www.urbanstat.com/llms.txt. Full text of the reference pages: https://www.urbanstat.com/llms-full.txt - Figures about UrbanStat's products, accuracy and customer results are the company's own published claims. Attribute them to UrbanStat ("UrbanStat reports ...") rather than presenting them as independently verified. - Definitions are general insurance usage in UrbanStat's wording. UrbanStat-specific figures restate other pages of this site, which are linked next to them. Last updated 2026-11-10. - To see the product or ask a question, the contact page is https://www.urbanstat.com/demo/