New USDA estimates show household food insecurity held constant in 2020 during the pandemic

 USDA yesterday released its annual food security report, showing that 10.5% of U.S. households were food insecure in 2020. Surprisingly, the 2020 estimate was unchanged from 2019 despite the COVID-19 pandemic. Observers expected the 2020 statistics to show a jump in household food insecurity.

The annual food security report uses data from a December food security supplement to the monthly Current Population Survey (CPS), the same data source used for federal statistics on unemployment and poverty. The report asks 18 survey items about hardships experienced in the past year.

The USDA report this year includes an interesting and thoughtful discussion of differences in methodology between the CPS supplement and the Household Pulse Survey (HPS), a special federal survey designed to measure household conditions during the pandemic year. The Pulse survey used a simple food insufficiency question about hardships experienced in the past 7 days. As the pandemic worsened in 2020, this higher-frequency food insufficiency measure in Pulse showed sharp increases in hardship, followed by a return to the milder starting levels.

It will take researchers a while to sort through the distinct results in the new USDA annual report. The CPS survey has some advantages, including a higher survey response rate. It has a broader time reference period covering the entire year. The Pulse survey also has some advantages, including more frequent survey administration and a shorter reference period, which could be useful for understanding short-term changes in hardship.

Currently, along with Irma Arteaga at the University of Missouri, I am helping to oversee a small grants program for USDA, looking at innovations in household food security measurement on the occasion of the 25th anniversary of USDA's official measure. You can imagine that the new 2020 annual food security report will be a big topic of discussion as this Food Sec 25 project proceeds.