The Food Desert Myth: Why Proximity Alone Doesn't Explain Food Insecurity

The Conventional Wisdom

Food deserts have become the dominant framework for understanding food insecurity in America. The logic is intuitive: if people lack nearby grocery stores, they can't access healthy food. Federal policy, from the Healthy Food Financing Initiative to USDA mapping tools, focuses on geographic proximity as the key barrier.

But what if proximity is the wrong measure?

We analyzed grocery store access in Santa Clara County, home to Silicon Valley and one of America's wealthiest metropolitan areas. What we found challenges fundamental assumptions about the relationship between food access and food security.

"Zero census tracts in Santa Clara County meet the federal definition of a food desert, yet SNAP participation varies 4.7x across similar communities."

— Analysis of 6,613 grocery store locations

Methodology

We compiled a comprehensive dataset of 6,613 grocery stores across Santa Clara County, including supermarkets, grocery stores, and convenience stores with fresh food. Each store was geocoded and verified against multiple sources.

For each of the county's 372 census tracts, we calculated:

• Distance to nearest supermarket (straight-line and road network)
• Number of stores within 0.5, 1, and 3 miles
• Transit travel time to nearest store using 2.7 million route calculations
• SNAP enrollment rates from California Department of Social Services
• Demographic and economic indicators from the American Community Survey

This approach allows us to test whether geographic proximity actually predicts food security outcomes, controlling for income and demographic factors.

Findings

Finding 1: Zero Food Deserts

By the USDA definition (low-income census tract where at least 500 people or 33% of the population live more than 1 mile from a supermarket), Santa Clara County contains zero food deserts. Every resident lives within reasonable geographic proximity to a grocery store.

This finding alone would suggest food access is not a problem. But other indicators tell a different story.

Finding 2: Massive SNAP Variation

Despite similar geographic access, SNAP participation rates vary dramatically across census tracts. Among tracts with similar poverty rates and demographic profiles:

• Lowest participation: 12% of eligible households enrolled
• Highest participation: 56% of eligible households enrolled
• Variation factor: 4.7x

This variation cannot be explained by distance to grocery stores, which is similar across all tracts.

Finding 3: Transit Access Matters More

When we calculated actual transit travel times (not just distance), a pattern emerged. Census tracts with poor transit access to grocery stores had significantly lower SNAP enrollment, even controlling for income.

We call these areas "mobility deserts" — locations where stores are technically nearby but practically unreachable for residents without cars.

Policy Implications

These findings suggest a reorientation of food security policy:

1. Geographic proximity is necessary but not sufficient. Having a store nearby matters, but it's not enough.
2. Transit access deserves equal attention. Improving bus routes to grocery stores may be more cost-effective than subsidizing new store construction.
3. Enrollment friction matters. The 4.7x variation in SNAP enrollment suggests administrative barriers, not just physical access.
4. One-size-fits-all solutions won't work. Different communities face different barriers, requiring targeted interventions.

Limitations

This analysis has several limitations:

• Santa Clara County is wealthier than most U.S. counties; findings may not generalize to rural or high-poverty urban areas.
• We use SNAP enrollment as a proxy for food security; not all food-insecure households are SNAP-eligible.
• Transit calculations assume weekday daytime travel; evening and weekend access may differ.
• We cannot distinguish between enrollment friction and deliberate non-enrollment.

Conclusion

The food desert framework, while intuitive, may be insufficient for understanding food insecurity in metropolitan areas. Geographic proximity to grocery stores is nearly universal in our study area, yet food security outcomes vary dramatically.

Future research should examine the relative contributions of transit access, enrollment friction, and other barriers. Policy interventions focused solely on store placement may miss the larger picture.