Costs, Reach, and Benefits of COVID-19 Pandemic Electronic Benefit Transfer and Grab-and-Go School Meals for Ensuring Youths’ Access to Food During School Closures

This economic evaluation evaluates the reach, costs, and benefits of the Pandemic Electronic Benefit Transfer and grab-and-go meals programs for students during COVID-19–related school closures in the US from March to June 2020.


Monthly individual program benefit for eligible children, in cash value
7) $150,000 -$199,999 8) $200,000 and above To classify households as either FRPM-eligible or non-eligible, we used the income and household size variables to estimate the household's income-to-poverty ratio using guidance for calculating poverty thresholds from the U.S. Census (https://www.census.gov/data/tables/timeseries/demo/income-poverty/historical-poverty-thresholds.html). We then set all households with a ratio ≤1.85 and reporting that they had a student attending school as eligible for FRPM.
To classify households as receiving grab-and-go meals or not, we coded all families that selected first that they had received free groceries or a free meal in the last 7 days and then selected that they had gotten the meals through the school or other programs aimed at children as having receive grab-and-go meals. While there may be some misclassification in this variable due to the phrase "other programs aimed at children," this is likely to be quite small, given that the listed options for other named sources were comprehensive and likely to cover any non-school programs.
We then calculated a) the proportion of FRPM-eligible students receiving grab-and-go school We calculated these metrics for six weeks, from April 23 -June 9 th , stratified by state. Because reach was variable across weeks (possibly due to real differences in reach or initial difficulty getting the program running, but also possibly due to imprecision in state-specific estimates), we selected the highest reach estimate for each state among these weekly estimates to represent spring 2020 reach overall. eAppendix 3. Decision process for selecting P-EBT reach data source

SOURCES FOR ESTIMATING THE NUMBER OF BENEFITTING CHILDREN PER STATE, P-EBT, IN ORDER OF PREFERENCE (WITH SOME EXCEPTIONS)
1) State press releases or state-released data from FOIAs submitted by investigators 2) Taking the total amount of cash benefits planned to be disbursed from March-Sept 2020 and dividing by the product of the maximum daily benefit amount times the number of days the state was issuing benefits, using states' initial P-EBT plans submitted to and approved by USDA as the estimate of cash benefits 3) Summing the "people" estimates from USDA's data release on P-EBT across March-Sept 2020, then dividing by the number of issuances for that state 4) Taking the total amount of cash benefits disbursed from March-Sept 2020 and dividing by the product of the maximum daily benefit amount times the number of days the state was issuing benefits, using data from the USDA data release (note: this is secondary to option 2 because we learned that many states had mixed in SNAP benefits in P-EBT data. However, there are a few instances where the planned estimates from option 2 are much too high and these option 4 estimates were more reasonable). 5) After reviewing monthly reimbursement data, identifying any patterns where it appeared that a single month's number of "people" getting benefits might actually represent the entire population getting benefits, i.e. the states reported serving the same number of people each month

DECISION FLOW FOR SELECTING THE MOST RELIABLE SOURCE
Do we have an estimate from a state-specific agency press release or website announcing the number of children who received benefits for spring 2020?

Costs to school food authorities
To estimate the costs to school food authorities to operate grab-and-go school meals, we used data collected in a prior study of large urban school food districts in the country (Kenney et al 2021), which included cost data for 7 districts, combined with a survey we administered to a convenience sample of school food service directors from 17 districts across 5 states. The data collected included: • Number of students enrolled in the district • Number of lunches served per month during school closures in 2020 • Number of breakfasts served per month during school closures in 2020 • Whether or not the district was charged with also serving meals to adults in need • Total monthly labor costs including wages, overtime, hazard pay, overhead and benefits during school closures in 2020 • Breakfast food costs per month during school closures in 2020 • Lunch food costs per month during school closures in 2020 • Recurring equipment and supply costs, such as food packaging materials (e.g., carryout containers, bags, disposable silverware), staff personal protective equipment (e.g., face masks, face shields, gloves, sanitizing and cleaning supplies), on-site communication materials (e.g., signage, floor stickers), mailers, fliers, or other communication materials, and any other (per month during school closures in 2020) • One-time equipment and supply costs, such as freezers, folding tables, or carts that were purchased to set up grab-and-go meal distribution (per month during school closures in 2020) • Whether or not the district delivered meals or distributed via stationary sites • (If delivery): Driver wages, benefits, and overtime if not already included above in 'labor costs.' • (If delivery): Costs associated with renting or operating delivery vehicles.
• (If delivery): Fuel expenses, if not included in the vehicle expenses above.
• Other -any other costs associated with meal delivery From these data, we then calculated the total school food authority operating cost for each surveyed district per month during COVID-19 school closures, and divided this by the sum of the reported lunches and breakfasts served per month during COVID-19 school closures to calculate a per-meal cost. To calculate a national weighted average cost per meal, we first calculated cost per meal for each district providing data. We then created an average cost per meal within each of the nine participating states, weighting each district's cost per meal by the total number of students in that district. Lastly, we calculated the average cost per meal for school food authorities across all nine states, weighted by the total number of FRPM-eligible students in that state. Treating these nine states as a sample of the U.S. states, we thus used this weighted average cost across the nine sampled states as the national weighted cost per meal.

Time and travel costs for families (uncompensated):
We calculated the time spent and travel costs expended on average for families to pick up grab-and-go school meals. We estimated that the average distance from a child's home to their school was 4.8 miles; that this trip lasted 18 minutes on average(31); and that families made four trips per month (assuming a weekly pickup of meals). We used the 2020 federal mileage reimbursement rate to estimate the cost of a car trip and valued adult caregivers' time in FRPM-eligible households using a market substitute based estimate developed by Davis & You (2010)(32) and inflating to 2020 dollars. We then weighted this estimate by the proportion of school districts that used a parent-pickup model as opposed to a delivery model using unpublished data collected via survey by Julia McCarthy and colleagues at the Laurie M. Tisch Center for Food, Education & Policy.