Housing tenure shapes food access in ways that are not immediately obvious. Renters are less likely to own cars than homeowners, making transit access to groceries particularly consequential for this group.1 If we want to understand who faces the greatest barriers to healthy food, we need to examine how housing status intersects with transportation options.
The article reports a comparison across 9,039 residential census tracts, including shorter distances (0.58 versus 0.94 miles), lower food-desert rates (8.4% versus 18.2%), and lower mobility-desert rates (9.8% versus 13.6%) in renter-dominated tracts.2 Those values, their direction, and the proposed urban-geography explanation are under reconciliation and are not publicly reproduced.
The article also reports that the difference narrows within income groups and that renter-dominated tracts have a higher vulnerability index. These are published claims awaiting a matched rerun, not settled comparisons.
The article reports lower distance and food-desert measures, 2.6 times more transit stops, and a higher vulnerability index in renter-dominated tracts. The tract inventory, values, direction of the comparisons, and urban-geography interpretation are not publicly reproduced.
The Descriptive Pattern
The table preserves the article-reported comparison of renter-dominated (>50% renter) and owner-dominated (>50% owner) tracts. No matched public output derives these values.
| Metric | Renter-Dominated | Owner-Dominated |
|---|---|---|
| Tracts | 3,847 | 5,192 |
| Mean distance to grocery | 0.58 mi | 0.94 mi |
| Traditional food desert rate | 8.4% | 18.2% |
| Mobility desert rate | 9.8% | 13.6% |
| Mean transit stops within 0.5 mi | 8.4 | 3.2 |
| Mean vulnerability index | 0.358 | 0.287 |
How the Article Interprets the Table
Article-reported locational comparison: 0.58 versus 0.94 miles and 8.4% versus 18.2% food-desert rates. The values and direction are not publicly reproduced.
Article-reported transit comparison: 8.4 versus 3.2 nearby stops and 9.8% versus 13.6% mobility-desert rates. The values and direction are not publicly reproduced.
Article-reported vulnerability comparison: 0.358 versus 0.287. The proposed poverty and economic-instability explanation is a hypothesis, not a reproduced decomposition.
A Proposed Urban-Geography Explanation
The article proposes residential location and urban form as possible explanations for its reported comparison. It does not publicly reproduce or causally test those mechanisms.
Urban vs. Suburban Geography
One hypothesis is that renter housing is more common in urban cores while owner-occupied housing is more common in lower-density suburbs. Testing whether density accounts for the article-reported access direction requires a matched tract inventory and model output that are not currently public.
Development Era
Prior work motivates a second hypothesis: older mixed-use neighborhoods may place retail closer to multifamily housing.3 This article does not reproduce a development-era test for the California tract sample.
Self-Selection
Renters may sort into neighborhoods with better amenity access. Households that cannot afford cars may choose locations with transit access. Households prioritizing walkability may choose rental apartments over suburban homes.
Self-selection is therefore another possible explanation, but the article does not estimate its contribution.
The Income Interaction
The article reports that its renter-owner difference narrows when stratified by income. The table values are not publicly reproduced.
| Income Quintile | Renter Mobility Desert Rate | Owner Mobility Desert Rate | Difference |
|---|---|---|---|
| Q1 (Lowest) | 14.2% | 18.4% | -4.2 pp |
| Q2 | 11.8% | 15.2% | -3.4 pp |
| Q3 | 9.6% | 13.1% | -3.5 pp |
| Q4 | 8.1% | 11.8% | -3.7 pp |
| Q5 (Highest) | 6.4% | 10.2% | -3.8 pp |
Within each income quintile, the article reports lower mobility-desert rates for renter-dominated tracts. The narrower gap and the proposed income-composition explanation remain under reconciliation.
Accounting for Density
The article reports the following linear-probability-model specification and coefficients. No matched public model output reproduces them.
Mobility_Desert_Rate = β₀ + β₁(Renter_Dominated) + β₂(Income) + β₃(Density) + ε
| Variable | Coefficient | Clustered SE | p-value |
|---|---|---|---|
| Renter-Dominated (dummy) | -0.012 | 0.006 | 0.08 |
| Median Income (per $10K) | -0.006 | 0.002 | <0.01 |
| Population Density (per 1000) | -0.008 | 0.002 | <0.01 |
The article reports an adjusted difference of 1.2 percentage points with p = 0.08 and attributes much of the raw difference to density. The estimate and interpretation are not publicly reproduced.
Article-reported inference: The page states that standard errors were clustered across 58 counties.4 The model code and fitted output are not public, so that implementation cannot currently be checked.
The Vulnerability Paradox
The article reports a higher vulnerability index in renter-dominated tracts and proposes economic components as an explanation. Neither the component table nor that decomposition is publicly reproduced.
| Component | Renter-Dominated | Owner-Dominated |
|---|---|---|
| Food access score | 0.15 | 0.22 |
| Poverty score | 0.24 | 0.14 |
| Renter score | 0.72 | 0.32 |
| Minority score | 0.58 | 0.42 |
| Sprawl score | 0.28 | 0.54 |
In the article-reported table, renter-dominated tracts have lower food-access and sprawl component scores and higher poverty, renter, and minority component scores. Those component values are not publicly reproduced.
The article interprets the net index as higher vulnerability despite lower reported access measures. It cannot establish from these unreproduced components that food access is secondary or that economic factors dominate.
Regional Patterns
The article reports that its renter-owner difference varies by region. The table is not publicly reproduced.
| Region | Renter-Dominated % | Mobility Desert Diff (R-O) |
|---|---|---|
| Bay Area | 48% | -4.2 pp |
| Los Angeles | 54% | -3.8 pp |
| San Diego | 46% | -3.1 pp |
| Central Valley | 38% | -2.4 pp |
| Other | 35% | -2.8 pp |
The article describes the Bay Area and Los Angeles as having the largest reported differences and the Central Valley as having a smaller difference. Those directions and the urban-suburban explanation await a matched regional output.
What This Analysis Cannot Determine
Causation
The article-reported correlation does not establish causation. Urban form, residential sorting, and tenure are candidate explanations, but the current materials do not identify which mechanism accounts for the reported direction.
Individual Experience
Tract-level analysis captures neighborhood averages. Not all renters in a tract have the same access. A renter near a transit stop and a renter far from one both count toward the tract's renter percentage. Within-tract variation may be substantial.
Stability
Gentrification, rent increases, and displacement can change neighborhood composition. Even if reproduced, a tract-level pattern from one ACS period need not persist as housing markets evolve.
Limitations
Binary Classification
Treating >50% renter as "renter-dominated" loses information. The article reports a continuous-share model in which each 10 percentage-point increase in renter share is associated with a 0.8 percentage-point decrease in mobility-desert probability (p = 0.03). That model is not publicly reproduced and therefore does not yet confirm a monotonic relationship.
Household-Level vs. Tract-Level
Not all renters in a tract have the same access. Tract-level analysis cannot capture within-tract variation in individual household circumstances, car ownership, or proximity to specific transit stops.
Ecological Fallacy
Tract-level correlations do not necessarily apply to individuals. A renter in an owner-dominated tract may have different access patterns than the tract average suggests.
Data and Methods
Material status: Article only. No matched public output currently derives this page's tract population, tables, or model estimates.
Data sources:
- Housing tenure: ACS 2019-2023, Table B25003
- Transit stops: Cal-ITP statewide GTFS
- Grocery distances: Calculated from population-weighted centroids
- Income: ACS 2019-2023, Table B19013
Classification:
- Renter-dominated: > 50% of occupied units are renter-occupied
- Mobility desert: Grocery within 1 mile, transit stop > 0.5 miles or < 2 stops within 0.5 miles
Statistical analysis:
- Descriptive comparison of means
- Income-stratified analysis
- Linear probability model with income and density controls
Software: Analysis conducted in Python 3.11 using statsmodels for regression estimation and clustered standard errors.
Notes
[1] Blumenberg, E., & Pierce, G. (2012). Automobile ownership and travel by the poor: Evidence from the 2009 National Household Travel Survey. Transportation Research Record, 2320(1), 28-36. Documents lower car ownership rates among renters compared to homeowners. ↩
[2] Renter-dominated defined as census tracts where renter-occupied units exceed 50% of total occupied housing units. Based on ACS 2019-2023 Table B25003. ↩
[3] Levine, J., & Inam, A. (2004). The market for transportation-land use integration: Do developers want smarter growth than regulations allow? Transportation, 31(4), 409-427. Discusses historical zoning patterns and their effects on land use mix. ↩
[4] Cameron, A. C., & Miller, D. L. (2015). A practitioner's guide to cluster-robust inference. Journal of Human Resources, 50(2), 317-372. https://doi.org/10.3368/jhr.50.2.317 ↩
This concludes the Food Security series on transit-based food access in California. Public-material status for this article: article only. The repository named in the original version no longer resolves.
Tags: #FoodSecurity #Housing #Renters #TransitAccess #Demographics #California #UrbanPlanning #Equity
Frequently Asked Questions
Do renters or homeowners have better grocery access in California?
The article reports shorter distances, lower food-desert rates, and more transit stops in renter-dominated tracts. Those values conflict with a related statewide page and are under reconciliation.
Why do renter neighborhoods have better food access?
The article proposes urban geography as an explanation. That interpretation is provisional while the underlying populations and measures are reconciled.
How much of the renter advantage remains after accounting for income and density?
The article reports a change from 3.8 to 1.2 percentage points after controls. The model estimate is under reconciliation.
Do renters have higher or lower food security vulnerability despite better access?
The article reports vulnerability indices of 0.358 and 0.287. Those values and their interpretation are under reconciliation with a related statewide page.
How to Cite This Research
Cholette, V. (2025, December 21). Renters vs. owners: Housing tenure and grocery access. Too Early To Say. https://tooearlytosay.com/research/transit-equity/housing-tenure/Copy citation