The local food environment is a determinant of food access and diet quality (1–3). This relationship is of particular concern in low-income urban communities where there are few supermarkets and an abundance of fast food outlets and corner stores. In Philadelphia, residents of low-income neighborhoods are half as likely to have access to grocery stores as residents of high-income neighborhoods (4).

Numerous strategies have been implemented to improve access to healthful foods in low-income communities, including supermarket financing initiatives (5), efforts to provide healthy foods at corner stores (6–8), and expansion of farmers markets (9,10). However, less attention has been devoted to improving affordability of healthful foods. During the past 4 decades, prices of healthful foods and beverages have increased relative to unhealthful items (11). Programs that increase the purchasing power for low-income consumers to buy healthful foods are promising approaches (9,12,13). However, the impact of bonus incentive programs has not been extensively evaluated.

Farmers markets may be particularly conducive for bonus incentive programs because they offer predominantly healthful foods (14,15) and increasingly are located in low-income communities and are equipped to accept Supplemental Nutrition Assistance Program (SNAP) benefits through the use of wireless technology (16,17). Philadelphia, Pennsylvania, is one of several cities and counties nationwide that have implemented bonus incentive programs targeting SNAP recipients as part of the Communities Putting Prevention to Work (CPPW) initiative of the Centers for Disease Control and Prevention, which aims to reduce chronic diseases through policy, systems, and environmental changes (18).

From 2010 to 2011, The Food Trust partnered with the Philadelphia Department of Public Health CPPW-funded Get Healthy Philly initiative to implement Philly Food Bucks, $2 bonus incentive coupons that could be redeemed only for fresh fruits and vegetables at farmers markets. The Philly Food Bucks program aimed to bring new customers to markets in low-income communities, to increase purchasing power for fruits and vegetables, to increase fruit and vegetable consumption among low-income customers, and to increase use of SNAP at farmers markets. Philly Food Bucks were distributed in 1 of 2 ways. They were distributed onsite at farmers markets along with SNAP purchases: a $2 bonus incentive coupon was received for every $5 in SNAP benefits used. (Original SNAP purchases did not need to be fresh fruits and vegetables to qualify.) Philly Food Bucks were also distributed by community organizations that serve SNAP-eligible populations to promote farmers market access among low-income residents. Those coupons could be redeemed without making a SNAP purchase.

Our study presents data from 2 sources: customer surveys and objective sales data from 22 farmers markets in low-income areas of Philadelphia. Ten of 22 markets were newly opened in 2010 to 2011 through Get Healthy Philly. The other 12 markets were in operation for 3 to 14 years. At schools near the 22 markets, more than 70% of students are eligible for free or reduced-price school meals (range, 72% to 95%), indicating that more than 70% of households near these markets are at or below 185% of the federal income poverty level (19,20).

This study addresses the following questions: What are the characteristics of Philly Food Bucks users? Among farmers market customers, are Philly Food Bucks users more likely than nonusers to report increased fruit and vegetable intake? What is the association between the Philly Food Bucks bonus incentive program and SNAP sales at farmers markets?

The study protocol was determined to be exempt from review by the Institutional Review Board of the Philadelphia Department of Public Health.

This evaluation lends support to the effectiveness of bonus incentive programs in improving nutrition behaviors of low-income residents. Our results suggest that among customers of farmers markets in low-income communities, participants in the Philly Food Bucks program were significantly more likely than nonparticipants to report eating more fruits and vegetables and trying new fruits or vegetables since becoming customers at the market. Markets participating in the Philly Food Bucks program demonstrated larger increases in SNAP sales per market than were observed before the bonus incentive program. Associations between Philly Food Bucks and SNAP sales were stronger than expected given the $2 for $5 bonus incentive model (or $1 in Philly Food Bucks for every $2.50 of SNAP sales).

The findings related to increased consumption of fruits and vegetables are consistent with those of prior studies assessing bonus incentive programs (12,13); however, there are key differences in the type and size of the incentives offered, as well as study settings and designs. Anderson and colleagues (12) found that women participating in WIC in Genesee County, Michigan, increased their fruit and vegetable consumption from baseline when provided one $20 farmers market coupon during a 5-month period. The effect was even greater for those receiving a brief, interactive nutrition education session. In another study, Herman et al (13) demonstrated that women participating in WIC in Los Angeles, California, purchased a wide variety of fruits and vegetables when given $40 per month in vouchers during 6 months for use in supermarkets and farmers markets.

In our study, participants were customers of farmers markets in low-income neighborhoods in Philadelphia. SNAP recipients received coupons based on SNAP purchases at farmers markets or through community organizations with a high percentage of SNAP-eligible clients. In contrast to studies cited above, in our study the size of the incentive was not fixed. Overall, our study population was neither exclusively female nor exclusively comprised of WIC participants. Our study also had a higher percentage of African Americans and a higher mean age compared with other studies. These characteristics may enhance the generalizability of our study to other large urban areas.

Assessment of increased fruit and vegetable consumption in our study was based on self-report at a single point in time. We also explored SNAP redemption levels at farmers markets before and during the Philly Food Bucks intervention. The growth in SNAP sales attributable to Philly Food Bucks may indicate that the incentive bolstered farmers market use among SNAP recipients. The bonus incentive program could have attracted low-income customers to markets when they otherwise would not have patronized them or it could have prompted them to “spend more to earn more.” Other factors also likely contributed to increased SNAP sales, including temporal increases in SNAP eligibility and farmers market use citywide, expansion of farmers markets in low-income neighborhoods, community-based promotion of markets, and nutrition education offered at markets.

Our findings are subject to several limitations. First, the study involved a convenience sample, which may not accurately reflect a representative sample of low-income residents who did or did not use Philly Food Bucks at farmers markets. Second, Philly Food Bucks participation was not randomly assigned, so differences in fruit and vegetable consumption may reflect unmeasured confounders or that customers who want to increase fruit and vegetable consumption are more likely to participate in the Philly Food Bucks program. Third, the measure of fruit and vegetable consumption was self-reported, based on 1 survey question, and cross-sectional in nature, limiting causative conclusions. Social desirability bias may have led to over-reporting of increased fruit and vegetable consumption. However, this measure was supplemented by self-report of trying new fruits and vegetables, which was also positively associated with Philly Food Bucks use. Fourth, there could be a seasonal effect on our findings because the customer survey was conducted toward the end of the farmers market season. Fifth, increases in SNAP sales beyond baseline could be attributable to Philly Food Bucks and other factors, such as increases in the SNAP-eligible population that were accounted for in our analyses or increases in food prices at our farmers markets that were not accounted for in our analyses. Lastly, these findings may not be generalizable to rural communities.

Despite these limitations, the initiation of the Philly Food Bucks program was associated with a significantly higher likelihood of self-reported increases in fruit and vegetable consumption and trying new fruits and vegetables in addition to increased per-market SNAP sales at farmers markets in low-income communities. This pilot program highlights the potential health impact of bonus incentive programs for low-income populations.

Future work can build on successes of bonus incentive programs at farmers markets. Pilot programs, such as the US Department of Agriculture’s Healthy Incentives Pilot in Massachusetts, will explore how integrating bonus incentives into the Electronic Benefits Transfer system can further increase fruit and vegetable consumption and SNAP use among low-income urban populations (23). Public and private organizations should prioritize investing in food access initiatives that address affordability as well as availability. Federal food programs should fund and further evaluate bonus incentive programs to decrease barriers to fruit and vegetable consumption among low-income populations.