The Healthfulness of Food and Beverage Purchases after the Federal Food Package Revisions

Year: 2015

Research Center: Institute for Research on Poverty, University of Wisconsin-Madison

Investigator: Andreyeva, Tatiana

Institution: University of Connecticut

Project Contact:
Tatiana Andreyeva
University of Connecticut
Rudd Center for Food Policy and Obesity
1 Constitution Plaza
Hartford, CT 06103
Phone: 860-380-1014


Federal food assistance programs support one in four American households in achieving adequate nutrition. The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) provides a set of nutrient-dense foods, nutrition education, and medical referrals to 8.0 million participants, including 51 percent of infants born in the United States, 28 percent of children age 5 and under, 29 percent of pregnant women, and 30 percent of postpartum women. Because of its broad reach and targeted impact at a critical age of human growth and development, WIC has considerable potential for early intervention to establish healthier eating habits in low-income populations.

In 2009, WIC implemented new food packages to improve dietary intake among WIC participants. The changes included the addition of whole-grain products; fruit and vegetable cash-value vouchers; reductions in milk, cheese, and juice allowances; restrictions on milk fat content; and incentives to encourage breastfeeding. The revisions provided a unique natural experiment to assess the ability of food assistance policy to improve diet quality in low-income populations. This study examines how this major WIC reform influenced the healthfulness of food purchases among low-income consumers.

The study is based on point-of-sale, loyalty-card-based data for Connecticut and Massachusetts from a regional supermarket chain. A unique feature of these data was information on the source of funds used to pay for every purchase, including WIC benefits, which identified program participation. The study examined all purchases made by a panel of 2,137 households that shopped at the chain using WIC benefits at least quarterly before and after implementation of the WIC revisions: January to September 2009 (pre), and January to September 2010 (post). A control sample of 1,303 low-income households was selected from former WIC households that stopped participating in WIC prior to the revisions and were not expected to be affected by the new changes.

The sample of WIC-participating and -control households purchased a total of 29,204 unique food and beverage products. Nutrition information, product/container size, and ingredient list for all purchased products was extracted from the syndicated nutrition databases via online search. The healthfulness of purchased foods and beverages was determined per their saturated fat, sugar, and sodium content. Each food and beverage item was classified as a Healthy, Neutral, or Moderation (less healthful) product, based on criteria in the USDA, Economic Research Service study, “Are Healthy Foods Really More Expensive? It Depends on How You Measure the Price.”

A pre- and post-assessment of the product-basket healthfulness was completed using generalized estimating-equation models. The outcomes were monthly household spending (measured in dollars) and monthly amount/volume purchased (measured in ounces) for healthy, neutral, and less healthy products, estimated using generalized linear models from the Poisson family with a logarithmic link function. Beverages and food were examined separately, in addition to total purchases. Household-level controls included measures of Supplemental Nutrition Assistance Program (SNAP) participation and cash assistance receipt, shopping frequency, and month and State indicators. Sociodemographic characteristics of the areas where households shopped were added to account for differences in prices, product selection, and marketing across stores. Data were analyzed in 2015.

At baseline, healthful products accounted for most of the food volume purchased by WIC participants, but the beverage category was dominated by moderation (less healthful) items. With new subsidies for fruit, vegetables, and whole grains, the WIC revisions increased the volume of healthy food purchases of WIC-participating households by 3.9 percent and expenditure by 3.1 percent. The share of healthful foods in the overall food basket increased even further. Reductions for moderation foods were fairly small in volume (1.8 percent), but larger in expenditure (4.7 percent), driven mostly by products such as cheese (reduced WIC allowances) and substitution of refined grains with whole grains. In contrast to WIC-participating households, control subjects had a reduction in purchases of healthful foods in 2010 (down by 4 percent).

The biggest improvements were seen for moderation beverages (down by 24.7 percent in volume and 23.4 percent in expenditure). This is consistent with the changes in the WIC packages that introduced restrictions on milk fat content and replaced whole milk (a moderation item) with low-fat/skim milk (a healthful item) for women and children ages 2 through 4. As low-fat milk purchases replaced whole milk, the share of less healthful beverages declined from 56 percent at baseline to 49 percent post-revisions. This improvement would likely have been even more substantial if reductions in WIC allowances of 100 percent juice (a healthful item) did not counteract the growth of healthful beverage purchases. No similar improvements were seen for beverage purchases of control households.

Households shopping in Massachusetts had a larger increase than Connecticut households in healthy and neutral beverage volume and a greater decrease in moderation beverage volume in the post-period. There was no differential effect for households shopping in stores located in lower socioeconomic neighborhoods. WIC households participating in SNAP did not decrease beverage volume purchased (all health types) at the level of WIC households not participating in SNAP. Larger food budgets among SNAP households may insulate them from the effects of the WIC revisions.

In conclusion, the healthfulness of the product basket improved post-WIC revisions; mainly caused by a reduction in the volume of moderation food and beverages purchased (down by 15.5 percent) rather than growth in healthful products (up 1.9 percent). Lack of similar changes in purchases of control low-income households suggests a causal impact of the WIC revisions. Efforts to encourage healthful eating by people receiving Federal food assistance are paying off. Healthier food purchases are likely to reflect more healthful eating patterns and diet improvements. If sustained, these changes may have significant long-term effects on health care costs and productivity. Given the cost-neutral design of the WIC revisions, the current and future benefits to WIC participants could produce an impressive return on investment for taxpayers.