The Magnitude and Determinants of Partial Redemptions of Food Benefits in the Special Supplemental Nutrition Program for Women,Infants and Children (WIC)

Abstract

Purpose:

WIC provides food assistance to low-income pregnant and postpartum women, infants, and children ages 1- 4. A partial redemption (PR) occurs when a participant redeems only a portion of the prescribed benefit, thereby not obtaining the full nutritional benefit. We study the magnitude and determinants of PRs using electronic benefit transfer data.

Design/Setting:

Statistical analysis of all WIC transactions from Feb 2016 to Nov 2018.

Subjects:

Oklahoma WIC participants.

Measures:

The probability of a WIC household fully redeeming a food category. A random effects probit model is used to study determinants of PRs.

Analysis:

Estimate the marginal effects of key variables on households’ likelihood of full redemption: location—urban/rural, number of members in WIC, duration in the program, number of shopping trips, shopping venue, and prescribed foods.

Results:

Overall 18.5% of $ value of benefits are unredeemed, 29.3% excluding formula. Some foods have PRs > 40%. Only 17.3% of households fully redeem benefits in a given month. PRs increase with number of household members in WIC and duration of participation. PRs are lower for participants in rural locations, who shop more frequently, and who shop at WIC-specialized stores.

Conclusions:

Food packages with high PRs fail to impart prescribed benefits. Results pinpoint products and household characteristics associated with PRs, enabling targeted nutritional counseling and suggest food package choices need to be made with participant acceptance in mind.

Keywords

Women Infants and Children (WIC) Program partial redemptions

Purpose

The Special Supplemental Nutrition Program for Women, Infants and Children (WIC) provides food access, health care, and nutritional education to low-income pregnant and postpartum women, infants, and children less than five years of age. WIC is the third largest federal nutrition assistance program in the United States, with roughly $6 billion appropriated in 2019, of which two-thirds was spent on food benefits.¹ The program covers over one-half of all newborn infants and one-quarter of all children under the age of five.²

The U.S. Department of Agriculture, Food and Nutrition Service (USDA, FNS) administers the program at a federal level. FNS has established seven core food packages. Each eligible participant receives one of the packages based on his/her own characteristics such as pregnancy status, breastfeeding practices, and age of children or infants. Variants of these packages exist for participants with special dietary needs. State WIC agencies have discretion in deciding the specific brands and products to include in approved WIC food packages, within the bounds of federal regulations.³

The food packages currently available under WIC emerged from a process that began in 2004 when FNS requested that the Institute of Medicine (IOM—now the National Academy of Medicine) convene a panel of experts to review and update the food packages. Most recommendations in the IOM’s report were implemented including introduction of a cash-value voucher (CVV) for purchase of fruits and vegetables, prescription of low- or non-fat milk for most participants, addition of several grain options, and reduced allocations of milk, eggs, and juice. Further revisions in the food packages were proposed in a 2017 report based on a review conducted by an expert committee convened by the National Academies of Sciences, Engineering, and Medicine.

Research on the nutritional impacts of WIC has generally found a positive effect on the health status of women and infants.⁴ WIC participation has been shown to mitigate iron deficiency anemia in children,⁵ positively influence birth weights and other birth outcomes,^6

-12 reduce incidence of attention deficit hyperactivity disorder in children,¹³ reduce food insecurity,^14
-16 and improve participants’ overall nutritional status.^17,18

Given the program’s documented health benefits, a key policy concern regarding WIC is its low and declining participation rate — currently about 55% — among the eligible population relative to other food-assistance programs.^19

-22 A related, but much less well-understood concern that we explore in this paper is that many WIC participants receive less than the full benefit from their participation in the program because the prescribed food benefits are only redeemed partially or not at all.

The 2017 report of the National Academies review committee provides no direct evidence on the partial redemption problem but recognizes its potential importance: “for many participants, however, some foods (e.g., fluid milk, ready-to-eat cereals) may not align with personal or cultural preferences, leading to reduced redemption and consumption” (p. 2).²³ Further, the review committee’s first recommendation for funded research includes study of “recommended revisions to the WIC food packages on participant satisfaction, participation in the program, redemption of WIC foods, and participants’ diets and health.”

Partial Redemptions in the WIC Program

Several factors could prevent or discourage WIC participants from fully utilizing their food benefits. As the National Academies report indicated,²³ prescribed foods may not coincide with a participant’s personal or cultural preferences. Focus groups interviewed by Phillips et al.²⁴ suggest that partial redemptions may also be related to difficulties in locating WIC foods within stores and keeping track of unused food benefits under the EBT system. Food accessibility and availability may also be an issue in rural areas or urban areas designated as food deserts. However, Wu et al.²⁵ found similar shopping patterns for WIC participants in food deserts and non-food-desert areas of Greater Los Angeles.

Partial redemptions can occur under either the traditional food-voucher system or the electronic-benefit-transfer (EBT) system. Understanding the magnitude of the partial-redemption problem was difficult prior to the implementation of EBT. Under the paper voucher system, WIC participants were required to redeem all food benefits on a given voucher in a single shopping experience, and only the total value of the redemption was recorded; the specific foods purchased or left unredeemed were unknown. The total value of a redemption was not dispositive because a low value could represent a partial redemption or a recipient who simply purchased a low-cost bundle from among the available alternatives.

EBT makes it possible to know precisely which products WIC recipients are purchasing, and whether a participant household is fully redeeming its benefits. However, we are aware of only two studies that have examined partial redemptions using EBT data. Hanks et al.²⁶ examined redemption behavior pre- and post- EBT transition in Ohio based on loyalty-card data from a single grocery chain. They found evidence of increased WIC redemptions occurring two to four months after EBT transition, although a similar increase recorded in non-WIC food purchases during the same period casts doubt on that result, as do subsequent FNS data that compare average redemption costs in EBT vs. non-EBT states.¹

Phillips et al.²⁴ used EBT redemption data in Kentucky, Michigan, and Nevada from January through March 2012 and found a high rate of partial redemptions, with only 12.6% of participants in the three states fully redeeming their food benefits over the study period. The authors were also able to adduce some evidence regarding the impacts of household demographics on partial redemptions and foods most likely to be partially redeemed. However, the short, 3-month time window of their study precluded the authors from analyzing participants’ dynamic behavior (e.g., whether the incidence of partial redemptions is impacted by the duration of participation). Moreover, Phillips et al.²⁴ could not control for unobserved, time-invariant heterogeneity within the participant population. Our empirical analysis demonstrates the importance of controlling for this factor.

In this paper we seek to extend understanding of the WIC partial redemption problem in several dimensions based on analysis of EBT redemption data for the state of Oklahoma. In addition to documenting the extent of the partial-redemptions problem by food category, we study determinants of partial redemptions by estimating a dynamic, random-effects probit model.

Methods

Eighteen total food categories were authorized for redemption by state agencies during the study period. Under the EBT system, all participants in a household share one EBT card for food benefit issuance and redemption. The authors received the data directly from the Oklahoma WIC State Agency in an anonymized fashion; institutional review board review and written informed consent were not required.

The study focused on the period from February 2016 through November 2018 (34 months), covering 113,339 households and 17,583,888 total redemptions. For each WIC food category we classified WIC participants as having purchased the full amount prescribed (full redemption), or less than the full amount (partial redemption or non-redemption). Because the smallest WIC-eligible breakfast cereal box size is 12 oz, we defined the cereal benefit as fully redeemed when the difference between issuance and redemption was less than 12 oz.

Given the binary nature of the dependent variable, we specified a dynamic random effects (RE) probit model:

P (y_{i, j, t} = 1 | y_{i, j, t - 1}, \dots y_{i, j, 0}, x_{i t}, c_{i}) = Φ (β x_{i, t} + ρ y_{i, j, t - 1} + F_{j} + c_{i}),

where i denotes households, j denotes food-product category, t denotes time in months, and $Φ (\cdot)$ is the standard normal cumulative distribution function. The binary dependent variable, $y_{i, j, t},$ indicates whether a household i fully redeemed product category j in time t, $y_{i, j, t - 1}$ is its lagged value, $x_{i, t}$ are the set of explanatory variables, $F_{j}$ is a fixed-effects variable to indicate the specific food category, and the $c_{i} \sim N (0, 1)$ account for household-specific unobserved heterogeneity that is assumed to be independent of the model’s random error term.

The RE probit model posits that households that fully redeemed their WIC benefits have certain observed characteristics represented by the $x_{i, t}$ and unobserved characteristics (e.g., tastes, cultural food preferences, access to vehicles or public transportation) represented by the c_i that create systematic differences compared to households that partially redeemed their benefits. This specification allows us to distinguish between household heterogeneity and “state dependence,” for example, whether having redeemed benefits partially in the past period affects the probability of being in that state in the current period.

Since the unobserved effect, $c_{i},$ is correlated with $y_{i, j, t - 1}$ , the model has an endogeneity problem if the initial conditions $y_{i, j, 0}$ are correlated with the c_i . To solve this initial-conditions problem, we use Rabe-Hesketh and Skrondal’s²⁷ extension of the Wooldridge²⁸ approach that includes using the initial-period explanatory variables as additional regressors.

The model includes the following household-specific independent variables, $x_{i, t}$ , to control for observed heterogeneity:

geographic location of the WIC household in terms of urban vs. rural based on Federal Office of Rural Health Policy classifications,

average number of household monthly shopping trips,

number of WIC participants in the household,

duration of the household’s participation in WIC, and

type of retail outlet where primary WIC shopping is done, specifically whether the household does its primary WIC shopping at a vendor specialized to serving the WIC program.

The expected impact of the urban vs. rural designation is unclear ex ante. Urban consumers tend to have more shopping options than their rural counterparts, but are less likely to own personal vehicles to facilitate shopping. Further, LaVarnway and Craven²⁹ indicate that 54 of 77 counties in Oklahoma, including counties containing the major urban centers Tulsa and Oklahoma City, contain low-income and low-food-access areas based on USDA definitions.

We use average number of shopping trips across the sample instead of number of trips in a given month to avoid endogeneity concerns. We expect the average number of shopping trips to be positively correlated with full redemptions because multiple trips create more opportunities to redeem WIC benefits under EBT. The expected effect of number of WIC participants in the household is unclear ex ante. More participants imply a greater value of total benefits and the need to arrange shopping to redeem them. On the other hand, benefit packages for multiple household participants often contain similar foods and the total may exceed what the household can consume in a given month.

The impacts of participation duration on partial redemptions are also unclear ex ante. Participants may not fully redeem benefits at the outset of joining WIC because they are unfamiliar with the redemption process or have trouble locating authorized foods within a store. On the other hand, participants who are in the program for a prolonged period of time, may tire of some WIC foods, causing lower redemption. In addition, participants who transition into different WIC participant categories (e.g., from pregnant woman to postpartum woman, or infant to child) receive benefit adjustments that could impact their redemption behavior. We defined three indicator or “dummy” variables to study impacts of duration of program participation on redemption behavior: 1 year or less (the default variable), 1 to 2 years, and more than 2 years.

State agencies at their discretion may authorize vendors whose primary function is to provide WIC-program foods to participants. Any vendor for whom over half of total food sales are through the WIC program is classified as an “above 50” or A-50 vendor, and special program regulations apply to these vendors. In particular, A-50 vendors’ prices are constrained to be no greater than the statewide average price for any particular food product. Oklahoma is among 15 U.S. states that authorize A-50 vendors.³⁰ Because these vendors’ business models are geared toward serving WIC participants, redemption rates may be higher for households that patronize A-50 vendors. We classify as household as an A-50 shopper in a given month if over half of the household’s WIC purchases were made with an A-50 vendor.

Results

Table 1 lists the total redemptions, issuances, and unredeemed value for each WIC-authorized food category in Oklahoma during the study period presented in rank order from most redeemed to least redeemed.

Table 1.

Oklahoma WIC Food Redemption and Issuance: February 2016-November 2018.

WIC food	Unit	Redemption	Issuance	Redemption/Issuance	Unredeemed ($)
Special Formula	CAN	245,052	249,127	0.98	96,741
Infant Formula	CAN	3,172,528	3,231,304	0.98	1,036,809
Pediasure Shake	PACK	198,445	235,472	0.84	131,446
Milk-whole	GAL	772,603	930,792	0.83	572,644
Eggs	DOZ	1,265,593	1,588,925	0.80	430,032
Juice-64 oz	CON	1,579,572	1,997,674	0.79	1,166,505
CVV	$	10,828,718	13,706,214	0.79	2,877,496
Cheese/Tofu	LB	1,145,719	1,493,977	0.77	1,459,201
Milk—Fat Reduced	GAL	2,898,437	4,202,373	0.69	4,550,737
Infant F& V	OZ	21,499,082	33,813,512	0.64	1,847,165
Whole Grains	OZ	22,538,084	37,155,672	0.61	2,484,990
Cereal	OZ	32,181,040	53,518,896	0.60	4,694,328
Legumes	CON	1,097,093	1,825,728	0.60	1,675,861
Yogurt	OZ	24,051,872	40,594,336	0.59	1,323,397
Fish	OZ	1,371,271	2,340,457	0.59	193,837
Infant Cereal	OZ	3,207,224	5,952,264	0.54	741,161
Juice-12 oz	CON	680,244	1,297,741	0.52	1,099,145
Infant Meats	OZ	775,440	2,091,297	0.37	526,343

Notes: CVV = cash value voucher, CON = container; redemption/issuance proportions exclude 2,295 households who redeemed no benefits.

Sample Characteristics

In total, 2,295 households did not redeem any of their issued food benefits in the entire data period. The unredeemed value for each product category was calculated by multiplying the unredeemed quantity by the average redeemed price per unit. The unredeemed value estimates both the value of the food items participants “left on the table” and the food costs not incurred by the WIC program for all foods except infant formula, for which states receive rebates. Breakfast cereal and milk had the largest unredeemed value, over $4 million for each over the 34-month data period. In total $27 million in food benefits were unredeemed in Oklahoma during the study period.

Among households who did redeem benefits, almost all issued formula benefits were fully redeemed (98%), while at the other extreme only 37% of issued infant meats were redeemed fully. Whole milk and 64 oz bottled juice have much higher redemption rates when compared with reduced-fat milk and 12 oz frozen juice.

Women and children are issued 36 oz of breakfast cereal benefit each month. A household may be issued multiples of 36 oz depending on the number of WIC-participating members. Figure 1 shows the household monthly redemption of breakfast cereal. The redemptions spike at 0, 36, and 72 oz, showing that most WIC participants either fully redeem their cereal benefits or redeem nothing. Approximately 37% of WIC households did not redeem any of the issued cereal benefits and 32% of households purchased all of the cereal benefits that they were issued.

Figure 1.

Monthly redemption of breakfast cereal.

Cash value vouchers (CVV) can be used to purchase fresh, canned, and frozen fruits and vegetables that meet specific federal guidelines. The National Academies committee report²³ recommends allowing participants to receive more CVV value in place of specific foods such as juice or jarred infant foods. During the study period, CVV monthly values were $10 (breastfeeding women), $8 (pregnant, formula feeding, and breast & formula feeding women), and $6 (children). Figure 2 shows the household monthly redemption pattern.

Figure 2.

Monthly Redemption of CVV.

Although 79% of CVV value was redeemed during the study period, 19% of WIC households did not redeem any of their CVV benefits. Among WIC households that did redeem CVV benefits, the spikes in the histogram at 8, 10, 16, and 18 dollars show a tendency to redeem the total value of the benefit issued. The relatively high rate of CVV redemption provides some support for the National Academies committee’s recommendation for greater use of CVVs.

Figure 3 plots per participant food costs for seven regions within the state that transitioned to EBT at different times over a 6-month period. Food costs are uniformly lower post transition compared to pre-transition. Regression results show that statewide food costs were reduced by about $8.24 per participant per month or 13% after EBT conversion, evidence suggestive, although not determinative, that partial redemptions are more prevalent under EBT than the voucher system.

Figure 3.

Oklahoma Per-Participant WIC Food Costs. Note: the EBT transition period is from month 0 to month 6 indicated by the 2 vertical lines.

Econometric Results

Results for the probit regression models are contained in Table 2. The results for the probability of fully redeeming prescribed WIC benefits using the dynamic RE probit model are presented in column 1, and results for a pooled probit model without random effects are presented in column 2. This second model was estimated to enable comparison with the approach of Phillips et al.²⁴ and to demonstrate the importance of accounting for time-invariant household heterogeneity. Each of the explanatory variables has a significant effect on the likelihood of full redemption in both models, although the magnitudes of the effects differ, illustrating the importance of controlling for household-specific unobserved heterogeneity.

Table 2.

Regression Results and Marginal Effects.

Variables	Coefficient estimates		Marginal effects
Variables	RE Probit(1)	Probit(2)	RE Probit(3)	Probit(4)
Full redemption in previous month	0.371***	0.862***	0.111***	0.268***
	(0.00118)	(0.000911)	(0.000355)	(0.000241)
Cheese	−0.642***	−0.502***	−0.180***	−0.145***
	(0.00503)	(0.00289)	(0.00132)	(0.000781)
Eggs	−0.576***	−0.450***	−0.160***	−0.128***
	(0.00499)	(0.00287)	(0.00130)	(0.000772)
Cereal	−1.076***	−0.890***	−0.317***	−0.270***
	(0.00503)	(0.00287)	(0.00134)	(0.000785)
Legumes	−0.976***	−0.831***	−0.286***	−0.250***
	(0.00504)	(0.00287)	(0.00135)	(0.000784)
Fish	−1.094***	−0.910***	−0.323***	−0.276***
	(0.00966)	(0.00589)	(0.00292)	(0.00185)
Infant Cereal	−1.268***	−1.087***	−0.377***	−0.334***
	(0.00624)	(0.00396)	(0.00174)	(0.00117)
Infant F & V	−1.230***	−1.089***	−0.365***	−0.334***
	(0.00624)	(0.00397)	(0.00175)	(0.00117)
Infant Meats	−1.824***	−1.516***	−0.533***	−0.464***
	(0.0160)	(0.0109)	(0.00396)	(0.00307)
Whole Grains	−1.103***	−0.929***	−0.326***	−0.282***
	(0.00511)	(0.00292)	(0.00137)	(0.000801)
Special Formula	−0.144***	−0.134***	−0.0362***	−0.0355***
	(0.0158)	(0.0103)	(0.00411)	(0.00281)
Pediasure Shake	−0.637***	−0.559***	−0.179***	−0.163***
	(0.0485)	(0.0267)	(0.0150)	(0.00847)
Yogurt	−1.024***	−0.858***	−0.301***	−0.259***
	(0.00510)	(0.00292)	(0.00137)	(0.000804)
Whole Milk	−0.806***	−0.697***	−0.232***	−0.207***
	(0.00613)	(0.00364)	(0.00174)	(0.00106)
Fat Reduced Milk	−1.219***	−1.017***	−0.362***	−0.311***
	(0.00516)	(0.00294)	(0.00138)	(0.000808)
Frozen Juice	−1.173***	−0.954***	−0.348***	−0.290***
	(0.00564)	(0.00328)	(0.00156)	(0.000941)
Bottled Juice	−0.621***	−0.527***	−0.174***	−0.152***
	(0.00528)	(0.00299)	(0.00141)	(0.000819)
# of WIC Members = 2	−0.0882***	−0.166***	−0.0265***	−0.0517***
	(0.00137)	(0.00101)	(0.000413)	(0.000314)
# of WIC Members = 3	−0.149***	−0.254***	−0.0448***	−0.0792***
	(0.00201)	(0.00148)	(0.000605)	(0.000461)
# of WIC Members ≥ 4	−0.260***	−0.381***	−0.0782***	−0.119***
	(0.00406)	(0.00303)	(0.00121)	(0.000938)
A-50	1.851***	1.188***	0.557***	0.369***
	(0.00252)	(0.00141)	(0.000711)	(0.000393)
12 < participation length ≤ 24	−0.196***	−0.0993***	−0.0590***	−0.0308***
	(0.00117)	(0.000967)	(0.000350)	(0.000300)
Participation length > 24	−0.360***	−0.204***	−0.108***	−0.0634***
	(0.00185)	(0.00156)	(0.000552)	(0.000486)
Average # of Shopping Trips	0.112***	0.112***	0.0336***	0.0348***
	(0.000732)	(0.000370)	(0.000218)	(0.000113)
Urban household	−0.0222***	−0.0869***	−0.00667***	−0.0270***
	(0.00195)	(0.000950)	(0.000587)	(0.000295)
Constant	0.0306***	0.0755***	9,693,143	9,693,143
	(0.00497)	(0.00287)
Observations	9,693,143	9,693,143

Notes: Standard errors in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1.

Columns (3) and (4) of Table 2 display the marginal effects of the relevant variables from these two models. Marginal effects show the change in probability when the independent variable increases by one unit. For continuous variables this represents the instantaneous change, while for binary variables the change is from 0 to 1.

Our discussion of results is based on the RE probit model given its superior statistical properties. The probability of full redemption varies considerably across food categories, confirming the evidence presented in Table 1. The reference group, infant formula, is most likely to be redeemed by WIC households, so all food categories have negative coefficients, indicating they are more likely to be partially redeemed relative to formula. Other food categories with relatively high probabilities of being fully redeemed include eggs, cheese, and bottled juice, while infant meats and infant fruit & vegetables are least popular among WIC participants.

The predicted probability of full redemption is 0.111 greater for households that fully redeemed their WIC benefits in the previous month than those who chose to redeem partially. This suggests that WIC participants’ redemption behavior persists systematically over time to some degree.

The hypothesis that households that on average have more shopping trips are more likely to fully redeem benefits is also supported in the estimation. An additional shopping trip is associated with a 0.034 greater probability of full redemption. The more WIC participants in a household, the less likely is the household to fully redeems all of the prescribed WIC benefits. Each additional member participating in the program is associated with a reduced probability of full redemption of from 2 to 3 percent.

In Oklahoma, rural households are slightly more likely to redeem benefits fully than their urban counterparts. This result is contrary to the finding of Phillips et al.²⁴ A-50 shoppers are much more likely to fully redeem benefits than participants who do their primary shopping at other types of vendors. An A-50 shopper, as defined for this study, is 55.7% more likely to fully redeem benefits than a participant whose primary WIC shopping is done elsewhere.

Figure 4 shows that the average probability of full redemption changes nonlinearly with the length of time that a WIC household remains in the program. The probability of full redemption is relatively stable and near 0.5 in the first 15 months of WIC program participation, then fluctuates over the subsequent 10-month period (15th to 25th month), and declines thereafter.

Figure 4.

Probability of Full Redemption as a Function of Participation Duration.

Regression results are consistent with Figure 4. Households in a second year of program participation are 6% less likely to fully redeem benefits than households in year one of participation. Households remaining in the program into a third year are about 11% less likely to fully redeem than year 1 participants.

Discussion

This study reveals a high incidence of partial redemptions in Oklahoma. During the 34-month study period, on average only 17.3% of WIC families in Oklahoma fully redeemed their WIC food benefits in each month, 71.6% redeemed partially, and 11.1% did not redeem any benefits in a given month. The rate of partial redemptions exceeded 40% for some food items.

An important question is whether the transition of WIC programs to EBT has increased the incidence of partial redemptions. The question is difficult to answer because, as noted, it is nearly impossible to gauge the magnitude of partial redemptions with data available under the paper voucher system. The average monthly WIC food cost per participant for EBT states in 2018 was $37.52, compared to $44.87 for (non-EBT) states.¹ Our analysis shows similar results for Oklahoma. One cannot, however, state conclusively that the reduction in program costs is due to more partial redemptions under EBT. One confounding factor is that EBT could cause participants to shop more at lower-cost supermarkets and supercenters because use of a debit card reduces or eliminates any stigma associated with making WIC redemptions compared to the paper-voucher system.^24,31,32 Li,³² however, found little evidence that EBT conversion had caused WIC shoppers to redeem benefits more extensively at supermarkets and supercenters in Oklahoma.

A second confounding factor is possible reduced prevalence of participants redeeming ineligible and more expensive products with EBT compared to the voucher system. Approved product lists containing the UPCs and other food-item codes are loaded into retailers’ inventory management systems under EBT, making it difficult or impossible for more expensive, unauthorized products to be purchased. Conversely, under the voucher system it was mainly left to participants and retailers to visually identify WIC-approved food items while shopping and at checkout.

Our results provide considerable insight into foods most and least likely to be fully redeemed. WIC foods with higher market value, infant formula in particular, and those that can be shared with the whole family, such as cheese, eggs, milk, and juice, are most likely to be fully redeemed. The econometric results also confirm that WIC participants on average prefer whole milk over reduced-fat milk and bottled juice over frozen juice. Among the foods least likely to be fully redeemed are jarred or boxed foods for infants—meats, fruits and vegetables, and cereal. These results may be useful to federal officials in contemplating future redesigns of the WIC food packages and to state WIC administrators in exercising their discretion regarding brands and products to include in WIC food packages,

Our results suggest that state agencies that do not authorize the A-50 vendor model could increase benefit redemptions significantly by introducing this type of authorized vendor. A-50 vendors are essentially cost neutral in terms of prices charged for WIC foods based on FNS regulations, so A-50 vendor authorization likely has little direct effect on program costs. In fact, McLaughlin et al.³³ showed that A-50 vendors in the greater Los Angeles area of California were associated with a small reduction in program costs.

Our results also identify some characteristics of WIC shoppers associated with partial redemption of benefits that may help guide nutritional counseling to encourage full redemption of benefits. In particular, participation in the program for a second or third year is associated with more partial redemptions. Notably, participants who stay in the program 24 months or more are exclusively children, suggesting that a high rate of partial redemption is associated with childrens’ benefit packages. A promising area of future research could involve surveying WIC participants in order to determine the facets of the program (e.g., vendors) and food benefits (e.g., product restrictions) that deter or encourage full redemption of benefits.

Conclusion

This paper demonstrates that partial redemptions are widespread across many categories of WIC supplemental foods in the Oklahoma WIC program. Broad failure of participants to fully redeem their food benefits makes it difficult for the program to achieve its goals and for partially redeeming participants to attain the demonstrated health benefits associated with participation in the WIC program. Although nearly all formula benefits are redeemed, the rate of full redemption was 60% or lower in Oklahoma for seven WIC food groups during our study period. Significant shares of participants don’t redeem any benefits in some months.

Based on the RE probit model, we find that participants are more likely to partially redeem benefits (i) the more household members that participate in the program, (ii) the longer the time period a household is engaged in the program, and (iii) the fewer shopping trips taken on average by the household. Further, participants who fully redeemed benefits in the prior month are 11% more likely to fully redeem in the current month, and participants who do a majority of their WIC shopping at specialty WIC vendors (i.e., A-50 stores) are substantially more likely to fully redeem benefits than other participants.

Several policy implications flow from this analysis. First, the very high rate of partial redemptions for several product categories suggests food packages and specific choices of eligible products and brands need to be made with an eye toward participant acceptance. The National Academies committee recommendations regarding greater use of CVVs as a substitute for jarred infant fruits or vegetables or juice may be one answer, given the high rate of full redemption for CVVs relative to these other products. Second, the nutritional education component of WIC could be targeted to overcome barriers that participants may face in acquiring and consuming products identified as having low redemption rates. Third, state WIC agencies presently not authorizing A-50 vendors could consider allowing these specialized vendors as a cost-neutral way to increase redemption of benefits.

So What?

What is already known on this topic?

WIC participation yields demonstrable health benefits, but partial redemptions are a significant concern. Participants’ demographic characteristics may impact the probability of full redemption.

What does this article add?

The present findings demonstrate the high overall rate of partial redemptions—nearly 72% of households did not fully utilize their food benefits—and indicate redemption rate by food category. New results are provided regarding persistence of partial redemptions, impact of duration in the program on partial redemptions, importance of A-50 vendors in encouraging full redemptions, and possible role of EBT in increasing partial redemptions.

What are the implications for health promotion practice or research?

Design of WIC food options must consider participants’ likelihood of purchasing and consuming the prescribed foods. WIC nutritional counseling may be able to address barriers to full redemption. Authorizing A-50 vendors may improve redemption rates.

Footnotes

Authors’ Note

The findings and conclusions in this publication are those of the authors and should not be construed to represent any official USDA or U.S. Government determination or policy.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Received funding from the U.S. Department of Agriculture, Economic Research Service

ORCID iD

Tina L. Saitone

References

U.S. Department of Agriculture Food and Nutrition Service. WIC Data Tables. Published 2019. Accessed February 9, 2021. https://www.fns.usda.gov/pd/wic-program

U.S. Department of Agriculture Food and Nutrition Service. Women, Infants, and Children: WIC Eligibility and Coverage Rates. Published 2017. Accessed February 9, 2021. https://www.fns.usda.gov/wic/wic-eligibility-and-coverage-rates

Code of Federal Regulations. Special Supplemental Nutrition Program for Women, Infants, and Children; 2011.

Bitler

Seifoddini

. Health impacts of food assistance: evidence from the United States. Annu Rev Resour Econ. 2019;11(1):261–287. doi:10.1146/annurev-resource-100518-093823

Miller

Swaney

Deinard

. Impact of the WIC program on the iron status of infants. Pediatrics. 1985;75(1):100–105.

Buescher

Larson

Nelson

Lenihan

. Prenatal WIC participation can reduce low birth weight and newborn medical costs: a cost-benefit analysis of WIC participation in North Carolina. J Am Diet Assoc. 1993;93(2):163–166. doi: https://doi.org/10.1016/0002-8223(93)90832-6

Kowaleski-Jones

Duncan

. Effects of participation in the WIC program on birthweight: evidence from the National longitudinal survey of youth. Am J Public Health. 2002;92(5):799–804. doi:10.2105/AJPH.92.5.799

Figlio

Hamersma

Roth

. Does prenatal WIC participation improve birth outcomes? New evidence from Florida. J Public Econ. 2009;93(1):235–245. doi:10.1016/j.jpubeco.2008.08.003

Hoynes

Page

Stevens

. Can targeted transfers improve birth outcomes? Evidence from the introduction of the WIC program. J Public Econ. 2011;95(7):813–827. doi:10.1016/j.jpubeco.2010.12.006

10.

Rossin-Slater

. WIC in your neighborhood: new evidence on the impacts of geographic access to clinics. J Public Econ. 2013;102:51–69. doi:10.1016/j.jpubeco.2013.03.009

11.

Currie

Rajani

. Within-mother estimates of the effects of WIC on birth outcomes in New York City. Econ Inq. 2015;53(4):1691–1701. doi:10.1111/ecin.12219

12.

Blakeney

Herting

Zierler

Bekemeier

. The effect of women, infant, and children (WIC) services on birth weight before and during the 2007-2009 great recession in Washington State and Florida: a pooled cross-sectional time series analysis. BMC Pregnancy Childbirth. 2020;20(1):252. doi:10.1186/s12884-020-02937-5

13.

Chorniy

Currie

Sonchak

. Does prenatal WIC participation improve child outcomes? Am J Heal Econ. 2020;6(2):169–198. doi:10.1086/707832

14.

Black

Cutts

Frank

, et al. Special supplemental nutrition program for women, infants, and children participation and infants’ growth and health: a multisite surveillance study. Pediatrics. 2004;114(1):169 LP-176. doi:10.1542/peds.114.1.169

15.

Metallinos-Katsaras

Gorman

Wilde

Kallio

. A longitudinal study of WIC participation on household food insecurity. Matern Child Health J. 2010;15(5):627–633. doi:10.1007/s10995-010-0616-5

16.

Kreider

Pepper

Roy

. Identifying the effects of WIC on food insecurity among infants and children. South Econ J. 2016;82: n/a-n/a. doi:10.1002/soej.12078

17.

Basiotis

Kramer-LeBlanc

Kennedy

. Maintaining nutrition security and diet quality: the role of the food stamp program and WIC. Fam Econ Nutr Rev. 1998:11(1):4–17.

18.

Kropf

Holben

Holcomb

Anderson

. Food security status and produce intake and behaviors of special supplemental nutrition program for women, infants, and children and farmers’ market nutrition program participants. J Am Diet Assoc. 2007;107(11):1903–1908. doi:10.1016/j.jada.2007.08.014

19.

Oliveira

Frazao

. The WIC program: background, trends, and economic issues, 2015 edition. SSRN Electron J. 2015;(134):88. doi:10.2139/ssrn.2709086

20.

Oliveira

. WIC participation continues to decline. Amber Waves. 2017.

21.

Martinez-Schiferl

WIC Participants and Their Growing Need for Coverage; 2012. https://www.urban.org/sites/default/files/publication/25306/412549-WIC-Participants-and-Their-Growing-Need-for-Coverage.PDF

22.

Johnson

Giannarelli

Huber

Betson

. Project Officer: Grant Lovellette. U.S. Department of Agriculture, Food and Nutrition Service, Office of Research and Analysis. National and State-Level Estimates of Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) Eligibles and Program Reach. 2014.

23.

National Academies of Sciences and Medicine E. Review of WIC Food Packages: Improving Balance and Choice: Final Report. The National Academies Press; 2017. doi:10.17226/23655

24.

Phillips

Bell

Morgan

Pooler

. Transition to EBT in WIC: review of impact and examination of participant redemption patterns; 2014. https://altarum.org/sites/default/files/uploaded-publication-files/Altarum_Transition%20to%20WIC%20EBT_Final%20Report_071614.pdf

25.

Saitone

Sexton

. Food access, food deserts, and the women, infants, and children program. J Agric Resour Econ. 2017;42(3):310–328.

26.

Hanks

Gunther

Lillard

Scharff

. From paper to plastic: understanding the impact of eWIC on WIC recipient behavior. Food Policy. 2019;83:83–91. 2–4. doi:10.1016/j.foodpol.2018.12.002

27.

Rabe-Hesketh

Skrondal

. Avoiding biased versions of Wooldridge’s simple solution to the initial conditions problem. Econ Lett. 2013;120(2):346–349. doi:10.1016/j.econlet.2013.05.009

28.

Wooldridge

JM.

Simple solutions to the initial conditions problem in dynamic, nonlinear panel data models with unobserved heterogeneity. J Appl Econom. 2005;20(1):39–54. doi:10.1002/jae.770

29.

LaVarnway

Craven

. An Overview of Food Deserts in Oklahoma. Regional Food Bank; 2017.

30.

Tiehen

Frazao

. Where Do WIC Participants Redeem Their Food Benefits? An Analysis of WIC Food Dollar Redemption Patterns by Store Type. United States Department of Agriculture; 2016. https://www.ers.usda.gov/webdocs/publications/44073/57246_eib152.pdf?v=0

31.

Moffitt

. An economic model of welfare stigma. Am Econ Rev. 1983;73(5):1023–1035. https://econpapers.repec.org/RePEc:aea:aecrev:v:73:y:1983:i:5:p:1023-35.

32.

. WIC Program Policies and WIC Participants’ Redemption Patterns. 2020. Ph.D. Dissertation, University of California, Davis.

33.

McLaughlin

Saitone

Sexton

The economics of food vendors specialized to serving the women, infants, and children program. Am J Agric Econ. 2019;101(1):20–38. doi:10.1093/ajae/aay054