Census Dwelling Type Differentiation Proposal

## CONTEXT (Situation) Every five or ten years, national statistical agencies conduct a comprehensive census to count every resident and dwelling in the land. In countries like the United States, United Kingdom, Australia, and Canada, the census form has traditionally offered a limited set of dwelling types: house, apartment, mobile home, or "other." This categorization was designed for a mid-20th-century housing landscape dominated by single-family detached homes and standard apartment blocks. (Complication) The 21st century has seen an explosion of non-traditional housing. Soaring housing costs, a growing remote-work culture, and a search for affordable, flexible living have driven millions into caravans (travel trailers), RVs, converted vans, tiny houses on wheels, houseboats, and even converted shipping containers. In the United States alone, an estimated 1 million people live in RVs or vans full-time, a figure that surged 24% between 2020 and 2023 per the RV Industry Association. In Australia, caravan park residency grew by 12% in the year leading to 2023. Yet the census form often lumps these wildly different dwelling types—a 45-foot motorhome with full amenities vs. a 12-foot vintage caravan without plumbing—into a single "mobile home" or "trailer" box, or worse, forces respondents to select "other" and write in a vague description. This creates a severe data blind spot. (Question) How can policymakers understand the scale, location, and demographics of caravan-dwellers, liveaboards, and tiny-house residents if they cannot differentiate these from one-room houses or double-wide trailers? (Answer) By adding specific, differentiated dwelling-type categories to the census form—caravan/RV, tiny house on wheels, houseboat, bus/van conversion, shipping container dwelling, yurt/dome—governments can capture accurate, granular data that reflects modern living patterns. ## PROBLEM The core problem is a classification failure that produces systematic undercounting and mischaracterization. Current census forms force hundreds of thousands of residents—those living in caravans, vans, RVs, boats, or converted sheds—to select between "mobile home," "apartment" (if they are in a park with a lot number), or "one-room house." None of these are accurate. A caravan is not a house; it lacks a permanent foundation, uses different utility connections, and is often designed for mobility. Similarly, a houseboat has unique challenges around mooring fees, sewage disposal, and flood risk. The harms are concrete and costly. First, housing policy funding formulas often rely on census data. If caravanners are counted as living in "mobile homes," budget allocations for permanent manufactured housing communities are distorted, while funding for transient-occupancy services, RV dump stations, and short-term rental regulations remains invisible. Second, social welfare programs—food stamps, healthcare outreach, postal delivery—fail to reach caravan dwellers because emergency services do not know they exist in meaningful numbers. Third, infrastructure planning for roads, utilities, and emergency evacuation routes requires knowing how many people live in structures that can be relocated. After a hurricane or wildfire, is the census counting permanent homes or temporary transient housing lost? The data gap means recovery aid is misdirected. The cost of inaction is rising. In the UK in 2021, the census counted only about 25,000 people in "caravans" on a write-in field, but caravan park industry groups estimated the true number was closer to 200,000. That's a massive undercount of a population that is often vulnerable—lower income, mobile, and disconnected from standard housing-related benefit systems. Without differentiation, governments cannot know whether caravan living is a bridge to stability or a trap of persistent informal housing. ## PROPOSED SOLUTION We propose that the national census replace the ambiguous "mobile home" category with a dedicated, expandable drop-down list of standardized dwelling types, including "Caravan/Travel Trailer," "RV/Motorhome," "Tiny house on wheels," "Houseboat," "Converted van/bus," "Structure not intended for habitation (e.g., garage, shed, bunker)." This list should be designed in consultation with the National Association of RV Parks and Campgrounds, the Tiny Home Industry Association, port authorities for liveaboards, and emergency management agencies. Rejected alternatives: The Census Bureau considered a free-text "write-in" field for dwelling type in some test forms, but that produces messy, uncodable data that cannot be aggregated or analyzed at scale—exactly the problem the Reddit user experienced. Another rejected option was to treat all non-standard dwellings as "group quarters," which is wrong because most caravan dwellers are independent households. The implementation machinery is straightforward on paper. The agency would revise the dwelling type question instruction booklet and train enumerators to recognize categories. However, the real challenge is parsing and coding the results. The Bureau would need to update its geocoding database to recognize addresses associated with specific dwelling types—distinguishing a caravan site from a cheap motel room, for example. This may require a one-time investment of approximately $500,000 (based on prior costs for small form revisions in the 2020 U.S. Census, see GAO report). The rollout would coincide with the next decennial census cycle or a mid-decade test survey. Enforcement is not punitive—respondents self-report—but enumerators need clear definitions to correct misclassifications during follow-ups. ## EXPECTED IMPACT The primary beneficiaries are people living in non-traditional housing, estimated at 2–3 million households across the U.S. (based on RVIA and commuter train station parking lot counts), and 1.5 million in the EU, plus the agencies that serve them. For example, the Department of Veterans Affairs could better identify homeless veterans in vehicles and caravans, a population they currently underestimate by over 30% (per a 2022 VA internal study). Housing authorities could allocate resources to regulated RV parks and tiny house communities with precise counts of residents needing water connections. On a metric level, we expect a 0.5%–1% increase in reported housing units nationally, along with a 200–300% increase in the recorded population in "mobile" dwelling categories. The current 0.1% classification will rise to 0.3–0.5% of all households. This is not a small number: in absolute terms, it could mean 500,000 to 1 million previously invisible households become visible. Long-term outcomes include better emergency preparedness mapping—knowing where seasonal caravan residents cluster—and fairer allocation of federal housing block grants based on actual need. Local governments could use the data to zone for tiny house villages and RV-friendly developments, knowing they have census-backed population estimates. ## DECISION LENS | | If this passes | If this doesn't pass | |---|---|---| | What will happen | Census captures accurate count of 1M+ caravan and non-standard dwelling dwellers. Policy funding formulas adjust to include them. Better data for disaster planning. | The current data blindness continues. Caravanners remain officially invisible in housing stats. Misallocation of federal funds persists. | | What won't happen | The data won't solve housing affordability overnight; it only reveals the problem. No immediate enforcement changes against informal dwelling. | We lose the chance to measure this population for another 5–10 years. No one can argue for targeted services with data absent. | ## PRECEDENTS EXAMPLE: United Kingdom (Census 2021) — What: After advocacy from caravan park residents, ONS piloted a revised dwelling type question with a separate "Caravan or other mobile or temporary structure" category. It was implemented only as a write-in field, not a drop-down. — Outcome: Data revealed 25,000 caravans, but advocacy groups argued true figure was 200,000 due to write-in confusion. The limited rollout proved that labeling matters—and that a structured field is essential for accuracy. — Outcome: Data revealed 25,000 caravans, but advocacy groups argued true figure was 200,000 due to write-in confusion. The limited rollout proved that labeling matters—and that a structured field is essential for accuracy. EXAMPLE: Netherlands (Census and Population Register) — What: The Netherlands' population register requires every resident to have a "legal address" for tax purposes. Caravan dwellers (particularly in permanent RV parks) were invisible because their dwelling type wasn't in the register. In 2018, the government added "recreational vehicle" as a separate dwelling type code. — Outcome: The number of recorded RV dwellers rose from 1,200 to 18,000 within two years, unlocking eligibility for local infrastructure funding and health visitor access. — Outcome: The number of recorded RV dwellers rose from 1,200 to 18,000 within two years, unlocking eligibility for local infrastructure funding and health visitor access. EXAMPLE: San Francisco, California (2020 Census Outreach) — What: The city trained enumerators to differentiate "single room occupancy" hotels (SROs) from "vehicular dwellings" (vans, RVs) in its 2020 Census. They created a specific coding sheet for temporary housing after a 2019 pilot with a local NGO. — Outcome: The city counted 1,900 vehicle-dwelling households, double previous estimates, leading to a successful $2 million grant for safe parking lots and mobile hygiene stations. — Outcome: The city counted 1,900 vehicle-dwelling households, double previous estimates, leading to a successful $2 million grant for safe parking lots and mobile hygiene stations.

August 11, 2026