What would improve food access?

Guest Lecture for AREC 222

Dr. Lauren Chenarides

October 13, 2026

The issue: limited food-shopping options

Some communities have few nearby places to buy a wide selection of affordable food.

For households, that can mean longer trips, higher shopping costs, or fewer choices. A supermarket may offer more products, but households still need to be able to reach it and afford what it sells.

How can food policy improve the options people have?

2 minutes including the title. Introduce your work on food retail markets. Define food access in terms of obtaining food, including availability, affordability, and the ability to reach retailers. Avoid a general food-security overview. This is the issue that connects all of today’s examples: a new supermarket is one possible response, but existing stores and assistance with purchases are other possible options.

Vote: Would you fund a new supermarket?

A neighborhood has few nearby grocery options.

The city proposes a grant to bring in a supermarket, with the goal of improving food access and residents’ diets.

Would you support the proposal?

Yes   No   I need more information

2 minutes. Preserve Lauren’s initial vote and request for context. Take a show of hands and two reasons, then leave the decision open. Ask students to remember their initial position. We will return to precisely this proposal at the end. Access and diet are two objectives; explain that the city needs to evaluate both rather than assume they move together.

How could a new supermarket improve food access?

A closer store could reduce the time and cost of shopping and offer foods that were difficult to obtain nearby.

For the proposal to improve diets, residents would also need to buy and eat more nutritious food.

To understand whether this will happen, we need to look at two decisions:

1. where households shop
2. whether a retailer can keep the store operating

2 minutes. The next two slides explain why a grant might help a supermarket open and why the store still needs enough sales to stay in business. Make that connection aloud: a new building helps only if it creates an option households can use and the retailer can sustain. No outcome is guaranteed by the opening alone.

Why might an area struggle to attract a supermarket?

A retailer may expect too few sales to cover its costs because there are too few customers, limited purchasing power, or strong competition for those customers.

The store must pay for inventory, labor, space, refrigeration, and deliveries. Fresh food also creates a risk of spoilage if it does not sell quickly enough.

Residents can need better food options even when expected sales are too low to support a supermarket.

Bitler & Haider (2011); Bonanno (2012).

3 minutes. This answers why an area might need help attracting a supermarket, not why households have poor diets. These are possible explanations to investigate rather than a diagnosis of every underserved neighborhood. Purchasing power means the ability to pay, not the importance of the need. Connect production and distribution to assortment: refrigeration, frequent deliveries, and sufficient sales volume make perishables feasible. Transition: a grant could reduce the initial investment, but recurring costs remain.

A grant can help a store open. What keeps it open?

A grant can cover part of the upfront investment, such as equipment or refrigeration, and make a location financially feasible.

Once the store opens, it still needs enough sales to cover ongoing costs such as inventory, wages, rent, and electricity.

The city therefore needs to ask both whether the store can remain in business and whether residents will use it in ways that improve food access or diets.

Bitler & Haider (2011); Bonanno (2012).

2 minutes. This is the direct policy point. Distinguish the initial investment from the ongoing operating budget. A grant can lower the first without guaranteeing the second. We are explaining retailer viability with the economic framework, not attributing a profitability result to Cummins. Transition: the next study evaluates what happened to residents after a supported supermarket actually opened.

What happened after a new supermarket opened?

Researchers studied a supermarket opening in Philadelphia supported by the Pennsylvania Fresh Food Financing Initiative - a program that helped fund the opening of a new store in a “food desert.”

Residents reported better perceptions of food access, but the study found no detectable improvement in reported fruit-and-vegetable intake or body mass index.

Opening the store improved how residents viewed their options. That did not automatically translate into a change in what they ate.

Cummins, Flint & Matthews (2014).

3 minutes. This pilot evaluates resident outcomes following a supermarket opening, not store profitability. Perceived access and dietary behavior are distinct outcomes. Do not interpret a failure to detect change as proof that no supermarket intervention can work. Transition: Allcott and colleagues ask the related, broader question of how much differences in products and prices explain nutritional purchasing differences across income groups.

Would better shopping options change what people buy?

Another set of researchers wanted to understand whether higher-income households bought more nutritious groceries, on average, than lower-income households. The comparison concerns the nutritional quality of groceries purchased.

They asked: If lower-income households could buy the same products at the same prices as higher-income households, how much of that difference would disappear?

A. Less than one-quarter   B. About half   C. Most of it

Allcott et al. (2019).

2 minutes. Explain the difference before asking for predictions: the researchers compared the nutritional quality of grocery baskets across income groups, not the amount of money spent or the number of stores. The hypothetical changes products and prices, not household income, time, cooking resources, or every other influence on diet. Take an A/B/C vote and one explanation. This is a prediction about a research result, not a judgment about people with low incomes.

Researchers tracked households’ grocery purchases

They examined purchases before and after a new supermarket opened nearby, and when households moved to a different neighborhood.

They asked whether households changed what they bought when their shopping options changed.

They also used a model of purchasing decisions to estimate what would happen if lower-income households faced the same products and prices as higher-income households.

Allcott et al. (2019).

2 minutes. Emphasize the actors in the title: researchers follow households’ purchases. Explain why comparing two neighborhoods once would be less informative: they may differ in income, transportation, and many other conditions. Changes around openings and moves provide evidence about the supply explanation; the next number is from the model that equalizes products and prices. There is no need to teach the estimator in AREC.

Most of the nutritional difference in purchases remained

In the model, giving lower-income households the same products and prices as higher-income households closed about 9% of the difference in the nutritional quality of their purchases.

Most of the difference remained. Access to the same products and prices alone did not explain why grocery baskets differed.

A nearby supermarket can still make shopping easier. But opening it does not, by itself, establish that residents will eat more nutritious food.

Allcott et al. (2019).

3 minutes. Restate the denominator: the difference between the two income groups’ nutritional purchasing measures. This is not the percentage of households who changed behavior. The paper reports that many people in food deserts already buy groceries at supermarkets farther away; a new nearby store can redirect those trips rather than change the basket substantially. Its discussion explicitly distinguishes nutritional effects from travel-cost benefits (introduction). The previous shorter-trip slide was an instructor-created illustration of this distinction, not a separate empirical estimate. Do not interpret the remaining model gap as unconstrained choice. Transition: policymakers have several ways to respond, depending on which part of food access they want to change.

Policymakers can address food access in several ways

Approach	What it tries to change	Example
Help retailers serve an area	Which foods are available nearby	A store-opening or refrigeration grant
Help households afford food	What people can buy with their budget	SNAP benefits or produce incentives
Help households reach food	The time and cost of shopping	Transit, delivery, or mobile markets

A community may need more than one approach. The supermarket proposal addresses only part of this picture.

Examples organized using the supply-and-demand framework in Bitler & Haider (2011) and Bonanno (2012).

3 minutes. This is the explicit bridge from Allcott to food-assistance policy. Explain supply-side support as changing retailers’ costs or offerings and demand-side support as changing household purchasing power. Transit is another way to connect existing supply to people. These options are not ranked by the Allcott finding. The next local example makes the affordability approach concrete.

Helping households afford produce: Double Up Food Bucks

Colorado’s Double Up Food Bucks matches SNAP purchases dollar for dollar at participating locations, up to $20 per day, for fresh fruits and vegetables.

An eligible $10 SNAP purchase earns $10 in Double Up benefits for produce.

This approach helps a household buy more food from participating sellers. It does not require the city to build another supermarket.

Double Up Food Bucks Colorado. Program rules checked October 8, 2026; purchase and redemption procedures vary by location.

2 minutes. SNAP is the Supplemental Nutrition Assistance Program. Double Up is the familiar local example of a produce incentive. Eligibility and qualifying purchases depend on location, so this is a teaching illustration of the match rather than individual enrollment advice. It is related in mechanism to HIP, but it is not the same program or incentive rate. Transition: we can observe that a program offers more purchasing power; research asks whether that changes what people eat.

Researchers tested a produce rebate in Massachusetts

The Healthy Incentives Pilot (HIP) offered a 30% rebate on eligible fruit and vegetable purchases made with SNAP benefits.

Researchers randomly assigned 7,500 SNAP households to receive the offer. Other SNAP households in the county received their usual benefits.

They compared fruit and vegetable intake between the groups to test whether the incentive changed what people ate.

Olsho, Klerman, Wilde & Bartlett (2016).

2 minutes. Introduce the paper explicitly: this is the randomized evaluation of USDA’s Healthy Incentives Pilot, not a study of Colorado Double Up. A 30% rebate returns 30 cents per eligible dollar, whereas the local example matches dollar for dollar. Random assignment makes the offer, rather than a prior interest in produce, the systematic difference between groups at the outset. Dietary recall measured intake. This is a different outcome from Allcott’s grocery purchasing measure.

The rebate increased fruit and vegetable intake

Adults offered the rebate consumed about one-quarter cup more per day of the fruits and vegetables targeted by the program, compared with adults receiving usual SNAP benefits.

The estimated increase was 0.24 cup-equivalents per day.

This trial provides evidence that making eligible produce more affordable can change consumption. It does not establish the effect of Colorado’s Double Up program.

Olsho et al. (2016).

2 minutes. Make the practical meaning accessible, then give the exact unit. This is an average difference, not an increase for every participant. Explain that the studies address different questions: equal local shopping options versus an incentive tied to particular purchases. Do not rank their effects numerically. The trial does not measure a long-term health effect. The analysis included 2,009 adults with follow-up, 72% of baseline respondents; keep that detail for questions. Transition: both approaches still depend on the places where households can shop.

Could other stores help meet the same food-access need?

Our opening problem was limited access to affordable food. A new supermarket is one response, but households may already use small grocers, dollar stores, or several types of stores together.

Dollar stores may offer a nearby place to buy some foods, even when their assortment differs from a supermarket’s.

Do they fill a gap where supermarkets are absent, and how do they fit into households’ shopping?

2 minutes. This is the explicit transition from policy interventions to the retail options already available. We have examined bringing in a supermarket and increasing household purchasing power. Now ask whether other retailers can serve some of the same need without treating every format as interchangeable. Return to the first slide’s food-access problem. The Applied Geography paper examines locations and persistence; the next papers examine competition and purchases.

Do dollar stores fill a gap in food access?

We tracked U.S. store locations from 2000 to 2017 and identified areas with low incomes and limited supermarket access.

We examined where dollar stores entered and whether those areas later gained supermarket access.

This allowed us to distinguish places with a temporary lack of access from places where supermarkets remained absent.

Chenarides, Cho, Nayga & Thomsen (2021), “Dollar stores and food deserts”.

2 minutes. Use we for this coauthored study. The analysis links Nielsen TDLinx store records with Census/ACS data and measures access annually at the block-group level. Explain food desert as the paper’s low-income, low-supermarket-access classification, not the absence of every place selling food. Dollar stores do not count as supermarkets in that measure. We observe geographic relationships; store location is not randomized.

Dollar stores can remain where supermarkets are absent

Dollar stores were not disproportionately entering low-income, low-access areas overall.

But among low-income, low-access areas where a dollar store entered, more than 90% remained without supermarket access for the rest of the study period.

Dollar stores also had lower exit rates in those areas. This suggests they may fill part of a food-access gap where supermarkets do not locate.

Chenarides et al. (2021).

3 minutes. This is the central finding to feature. More than 90% refers to the selected group of low-income, low-access block groups with dollar-store entry, followed for the remaining observed years; follow-up length varies with entry year. It is not 90% of all neighborhoods or a causal effect of entry. A store can add access to some foods without changing a supermarket-based access classification. The paper cannot determine whether entry prevented a supermarket from opening. The implication is that banning an available retailer need not cause a supermarket to appear. These associations do not show that dollar stores caused supermarkets to stay out. Transition: where both formats operate, are they necessarily competing for the same shopping trips?

Dollar stores and supermarkets can complement each other

We then modeled Dollar General expansion, and found that rival small-format stores lost net revenue, while supermarkets and supercenters gained net revenue and store density.

The results suggest dollar stores compete more directly with other small formats and can complement supermarkets, which offer a different assortment.

That raises a household-level question: could shopping at both help people stretch their food budget?

Chenarides, Çakır & Richards (2024).

2 minutes. State the result as complementarity in this model rather than saying the formats never compete. The paper discusses different assortments and the indirect benefit to supermarkets as smaller rivals weaken. It does not test whether a household spends dollar-store savings at a supermarket. Explicit exit is not modeled; likely exit is inferred when simulated returns cannot cover costs. Connect this finding to the proposed shopping mechanism on the next slide, but keep market-level findings and household behavior distinct.

Shopping at two stores could stretch a food budget

Hypothetical example: the same size and quality of packaged item

	Supermarket	Dollar store
Price for a 16-ounce item	$2.40	$1.60
Price per ounce	15¢	10¢
Cost of five items	$12	$8

Buying those items at the dollar store saves $4, which could be spent on other foods at the supermarket—if the extra trip does not use up the savings.

Illustration of a possible household response, not an estimate from either Chenarides paper. Dollar stores do not necessarily have the lowest unit price.

3 minutes. Invite students to calculate the savings before explaining. Holding size and quality constant is essential; a smaller package can have a lower sticker price but a higher unit price. This is an increase in what a fixed budget can buy, not an increase in earned income. Savings could be spent on supermarket foods, other needs, or saved. Transport and time can offset a price advantage. We have not tested this budget reallocation in the entry paper. The Tufts study provides related evidence on purchases across retailers, rather than a causal test of this exact mechanism.

Households use several stores to fill their grocery basket

Households can buy some items at dollar stores and complete their grocery basket elsewhere.

Food-access policy needs to consider the combination of stores households use.

Feng, Page, Sun & Cash (2025).

1 minute. Keep the focus on the full basket and connect directly to the preceding unit-price example. The study’s nutrition comparison is consistent with households making complementary purchases elsewhere. It does not directly show that dollar-store savings pay for specific supermarket purchases, or that shopping at dollar stores improves a household’s diet. Leave those details for questions rather than adding them to the slide. Transition: return to the city’s proposal and consider all the stores residents already use.

Let’s revisit the city’s supermarket proposal

The city wants to fund a supermarket to improve food access and residents’ diets.

Would you support it?

• Would it make food easier to reach, more affordable, or both?
• What are residents already buying at other stores?
• Would a produce incentive or support for existing stores address the problem more directly?
• What evidence would show that the grant helped residents?

Begin this slide by minute 45. Give pairs two minutes to reconsider their initial vote, then take a second vote and invite contrasting reasons. From minute 50, keep this discussion slide on screen for the 25-minute Q&A, including students’ prepared questions. There is no single correct answer without local information about prices, assortment, transport, use of existing stores, and store viability. Ask what changed from the opening vote. Do not add a separate recommendation template or final takeaway slide. The following slide lists all citations; additional source notes are in notes/AREC222-sources.md.

References

Bitler, M., & Haider, S. J. (2011). An economic view of food deserts in the United States. Journal of Policy Analysis and Management, 30(1), 153–176. DOI.

Bonanno, A. (2012). Food deserts: Demand, supply, and economic theory. Choices, 27(3), 1–4. Article.

Cummins, S., Flint, E., & Matthews, S. A. (2014). New neighborhood grocery store increased awareness of food access but did not alter dietary habits or obesity. Health Affairs, 33(2), 283–291. DOI.

Allcott, H., Diamond, R., Dubé, J.-P., Handbury, J., Rahkovsky, I., & Schnell, M. (2019). Food deserts and the causes of nutritional inequality. Quarterly Journal of Economics, 134(4), 1793–1844. DOI.

Double Up Food Bucks Colorado (2026). Program description and rules. Website. Accessed October 8, 2026.

Olsho, L. E. W., Klerman, J. A., Wilde, P. E., & Bartlett, S. (2016). Financial incentives increase fruit and vegetable intake among Supplemental Nutrition Assistance Program participants: A randomized controlled trial of the USDA Healthy Incentives Pilot. American Journal of Clinical Nutrition, 104(2), 423–435. DOI.

Chenarides, L., Cho, C., Nayga, R. M., Jr., & Thomsen, M. R. (2021). Dollar stores and food deserts. Applied Geography, 134, 102497. DOI.

Chenarides, L., Çakır, M., & Richards, T. J. (2024). Dynamic model of entry: Dollar stores. American Journal of Agricultural Economics, 106(2), 852–882. DOI.

Feng, W., Page, E. T., Sun, B., & Cash, S. B. (2025). The healthfulness of foods purchased at dollar stores: Insights from household scanner data from 2008 to 2020. Journal of the Academy of Nutrition and Dietetics, 125(11), 1736–1745.e6. DOI.

All nine sources cited in the current lecture. The supplied Bitler PDF is dated 2010; the published article is cited as 2011. The city proposal and unit-price comparison are hypothetical examples. Stay on the preceding proposal slide during discussion; these references are available for students afterward.