# #FoodSecurityFridays — Week 11: Unemployment & Food Security Reproducibility Files **"Even at full employment, jobs couldn't fix food insecurity."** Unemployment Fell to Historic Lows. Food Insecurity Kept Climbing. This folder contains a Stata program to reproduce all statistics in the #FoodSecurityFridays LinkedIn post and infographic analyzing the disconnect between declining unemployment and rising food insecurity — showing that employment alone doesn't guarantee food security when wages don't keep pace with food prices. --- ## Files Included | File | Description | |------|-------------| | `cpsdec2024.do` | Reads raw CPS-FSS ASCII data and creates Stata dataset | | `fsf_week11_reproducibility.do` | Reproduces all post statistics + full analysis | | `fsf_week11_README.md` | This documentation file | --- ## Requirements - **Stata** (version 14 or later recommended) - **December 2024 CPS Food Security Supplement** microdata --- ## Data Access Download the December 2024 CPS Food Security Supplement from the U.S. Census Bureau: **URL:** [https://www.census.gov/data/datasets/time-series/demo/cps/cps-supp_cps-repwgt/cps-food-security.html](https://www.census.gov/data/datasets/time-series/demo/cps/cps-supp_cps-repwgt/cps-food-security.html) Download the file `dec24pub.dat` (the raw ASCII data file). --- ## Instructions ### Step 1: Prepare the Raw Data 1. Open `cpsdec2024.do` in Stata 2. Update the directory paths at the top of the file: ```stata local indir "YOUR_PATH_HERE" local outdir "YOUR_PATH_HERE" ``` 3. Run the program to create `cpsdec2024.dta` ### Step 2: Reproduce Statistics 1. Open `fsf_week11_reproducibility.do` in Stata 2. Update the two `"YOUR_PATH_HERE"` entries near the top: ```stata if "$rawdata" == "" { global rawdata "YOUR_PATH_HERE" } if "$projdir" == "" { global projdir "YOUR_PATH_HERE" } ``` 3. Run the program 4. Review the output log (`fsf_week11_analysis.log`) to verify statistics --- ## Statistics Reproduced ### Headline Numbers (Tables 1–2) - 13.7% of U.S. households food insecure (18.3 million) - 66.5% of FI households have at least one employed adult - 12.2 million working food-insecure households - 9.9 million with a full-time worker ### FI Rate by Employment Status (Table 3) | Employment Status | FI Rate | VLFS Rate | FI Households | |---|:-:|:-:|:-:| | Full-time worker | 11.8% | 4.3% | 9.9M | | Part-time only | 19.1% | 7.9% | 1.7M | | Unemployed | 39.5% | 16.2% | 0.9M | | Not in labor force | 14.8% | 6.3% | 5.8M | ### Working FI Households — FT vs PT (Table 4) Among 12.2 million working FI households: - 9.9 million have at least one full-time worker - 1.7 million have part-time workers only ### Food Spending by Employment Status (Table 5) | Employment Status | Need More | SNAP Rate | |---|:-:|:-:| | Full-time worker | 52.4% | 19.5% | | Part-time only | 62.7% | 46.2% | | Unemployed | 57.4% | 62.8% | | Not in labor force | 60.5% | 52.5% | **Note:** SNAP participation is self-reported in the CPS-FSS (`hesp1`) and may understate actual participation due to reporting error. ### FI by Employment x Income (Table 6) Cross-classification showing that within each employment category, FI rates track income — the mechanism is wage adequacy, not employment status per se. ### SNAP Participation by Employment (Table 7) Working FI households: 24% self-reported SNAP participation vs. 57% for non-working FI. SNAP participation is self-reported in the CPS-FSS (`hesp1`) and may understate actual participation due to reporting error. ### Historical Context (Table 8) Published data from BLS (unemployment, median earnings, CPI-U) and USDA ERS (FI prevalence), 2001–2024: - Unemployment 2019–2024: 3.7% → 4.0% (essentially flat) - FI 2019–2024: 10.5% → 13.7% (+3.2 pp) - CPI Food at Home 2019–2024: +25.5% - Real food wage index (2019=100): 99.3 in 2024 (wages didn't keep pace) - Correlation (unemployment, FI): r = 0.53 - Correlation (real food wage, FI): r = -0.43 ### Profile: Working vs. Non-Working FI (Table 9) | Statistic | Working FI | Non-Working FI | |---|:-:|:-:| | Median income/poverty | 1.99 | 0.91 | | SNAP participation (self-reported) | 24% | 57% | | Households with children | 46% | 19% | | Median food gap | $100/week | $80/week | ### VLFS Among Working Households (Table 10) - VLFS rate among FT-worker households: 4.3% - VLFS households with a FT worker: 3.6M - VLFS households with any worker: 4.6M ### Hours Worked — Working FI Households (Table 11) - Max usual weekly hours (working FI HHs): P25=40, P50=40, P75=40 - Total HH hours (working FI HHs): P25=40, P50=40, P75=79 - Share of working FI households where at least one member works 40+ hours: 78.3% - Comparison: 87.0% among working food-secure households ### Single-Earner vs. Dual-Earner Vulnerability (Table 12) - FI rate among single-earner working households: 15.7% - FI rate among multiple-earner working households: 9.6% - FI rate among single-earner working households with children: 23.3% - FI rate among multiple-earner working households with children: 11.7% - Share of working FI that are single-earner: 62.9% - Single-earner FI HHs with children: 3.2M ### Food Spending Paradox (Table 13) - Median self-reported usual weekly food spending: $160 working FI, $100 non-working FI - Share reporting need to spend more: 53.5% working FI, 61.5% non-working FI - Median weekly spending gap: $100 working FI, $80 non-working FI - Median household size: 3 working FI, 1 non-working FI **Note:** Food spending (`hets8ou`) and spending gap (`hets8co`) are self-reported amounts, not verified expenditures. They reflect households' perceptions of their usual spending and shortfall. --- ## Methodology ### Employment Aggregation Employment status is aggregated across all adult household members before restricting to reference persons: - **pemlr** (Monthly labor force recode): 1–2 = employed, 3–4 = unemployed, 5–7 = NILF - **pehruslt** (Usual hours worked): ≥35 = full-time, 1–34 = part-time - Household categories assigned hierarchically: FT worker > PT only > Unemployed > NILF **Note on hours-missing cases:** Employed adults with missing or zero values for `pehruslt` are counted in `hh_employed` (via `pemlr`) but are not classified as either full-time or part-time. This creates a small number of households (~620K weighted) where `hh_anyemp == 1` but `hh_empstat` is coded as "Unemployed" or "NILF." These cases represent less than 5.2% of working food-insecure households and do not materially affect headline statistics. ### Real Food Wage Index Constructed from published BLS data: - **Numerator:** Median usual weekly earnings (Series LEU0252881500, Q4 annual) - **Denominator:** CPI-U Food at Home (Series CUUR0000SAF11, annual average) - **Index:** (earnings / CPI-FAH) × 100, normalized to 2019 = 100 ### Variable Construction - **Food security status** from `hrfs12m1` (1=food secure → 0, 2=LFS → 1, 3=VLFS → 2) - **Income** via midpoint imputation of `hefaminc` brackets; top bracket (≥$150K) assigned $160,000 (conservative midpoint; ERS reports typically use $175,000) - **Poverty thresholds** are 2024 Census thresholds. - **SNAP receipt** from `hesp1` (self-reported; -1=screened-out nonrecipients, 1=Yes, 2=No) - **Food spending** (`hets8ou`) and **spending gap** (`hets8co`) are self-reported and topcoded by Census ($500 and $300, respectively). Screened-out households (-1, not in universe) are set to $0. Topcoding does not affect medians for the FI subpopulation. - **Weights**: raw `hhsupwgt` ÷ 10,000 ### Survey Design All estimates use household supplement weights (`hhsupwgt`) via `svyset [pw=hhsupwgt]`. Standard errors for proportions are computed under a simple random sampling assumption — CPS design variables (stratum, PSU) and replicate weights are not incorporated. This approximation tends to understate standard errors; reported confidence intervals should be interpreted as lower bounds on true sampling uncertainty. Median estimates (income-to-poverty, food spending, household size) are computed via `_pctile` and do not have associated standard errors or confidence intervals. --- ## Key Variables | Variable | Description | Values | |----------|-------------|--------| | `pemlr` | Monthly labor force recode | 1–7 (see above) | | `pehruslt` | Usual hours worked | 0–99 | | `hets8ou` | Usual weekly food spending | 0–500 (topcoded) | | `hes8b` | More/less/same to meet needs | 1=More, 2=Less, 3=Same | | `hets8co` | Weekly dollar gap | 0–300 (topcoded) | | `hesp1` | SNAP receipt past 12 months (self-reported) | -1=Screened out, 1=Yes, 2=No | | `hrfs12m1` | Food security status (3-cat) | 1=FS, 2=LFS, 3=VLFS | | `hefaminc` | Family income bracket | 1–16 (see Variable Construction) | | `hhsupwgt` | Household supplement weight | Rescaled (÷10,000) | --- ## Sources ### Primary sources (statistics reproduced in this package) Rabbitt, M.P., Reed-Jones, M., Hales, L.J., Suttles, S., & Burke, M.P. (2025). *Household Food Security in the United States in 2024* (Report No. ERR-358). USDA, Economic Research Service. [https://www.ers.usda.gov/publications/pub-details?pubid=113622](https://www.ers.usda.gov/publications/pub-details?pubid=113622) U.S. Bureau of Labor Statistics. Unemployment Rate, Series LNS14000000; Median Weekly Earnings, Series LEU0252881500; CPI-U Food at Home, Series CUUR0000SAF11. [https://data.bls.gov/](https://data.bls.gov/) U.S. Census Bureau. (2025). *Current Population Survey, Food Security Supplement, December 2024*. [https://www.census.gov/data/datasets/time-series/demo/cps/cps-supp_cps-repwgt/cps-food-security.html](https://www.census.gov/data/datasets/time-series/demo/cps/cps-supp_cps-repwgt/cps-food-security.html) ### External citations (findings cited in the post but not reproduced here) Restrepo, B.J., Rabbitt, M.P., & Gregory, C.A. (2021). The effect of unemployment on food spending and adequacy: Evidence from coronavirus-induced firm closures. *Applied Economic Perspectives and Policy*, 43(1), 185–204. [https://doi.org/10.1002/aepp.13143](https://doi.org/10.1002/aepp.13143) *Cited for:* Causal effect of job loss on food outcomes — involuntary unemployment reduced food spending 15%, increased free food receipt 36%, reduced food sufficiency 10%, and reduced confidence in next month's food adequacy 21%. Mabli, J., Monzella, K., Franckle, R.L., & Lavallee Delgado, P. (2023). Food insecurity transitions and changes in employment and earnings. *American Journal of Preventive Medicine*, 64(3), 368–376. [https://doi.org/10.1016/j.amepre.2022.09.028](https://doi.org/10.1016/j.amepre.2022.09.028) *Cited for:* Asymmetric employment transitions — job loss OR 2.67 for becoming food insecure; re-employment showed no statistically significant association with becoming food secure. Wells, W., Jackson, K., Leung, C.W., & Hamad, R. (2024). Food insufficiency increased after the expiration of COVID-19 emergency allotments for SNAP benefits in 2023. *Health Affairs*, 43(10), 1464–1474. [https://doi.org/10.1377/hlthaff.2023.01566](https://doi.org/10.1377/hlthaff.2023.01566) *Cited for:* SNAP emergency allotment expiration led to 8.4 percentage point increase in food insufficiency among recipients. Rothbaum, J. & Bee, A. (2021). *Coronavirus infects surveys, too: Survey nonresponse bias and the coronavirus pandemic* (Working Paper SEHSD-WP2020-10). U.S. Census Bureau. [https://www.census.gov/library/working-papers/2020/demo/SEHSD-WP2020-10.html](https://www.census.gov/library/working-papers/2020/demo/SEHSD-WP2020-10.html) *Cited for:* CPS pandemic nonresponse was correlated with income, which may have modestly understated food insecurity in the 2020 CPS-FSS. Winkler, M.R., Clohan, R., Komro, K.A., Livingston, M.D., & Markowitz, S. (2025). State minimum wage and food insecurity among US households with children. *JAMA Network Open*, 8(3), e252043. [https://doi.org/10.1001/jamanetworkopen.2025.2043](https://doi.org/10.1001/jamanetworkopen.2025.2043) *Cited for:* A 10% increase in state minimum wage associated with 0.39 pp reduction in food insecurity among working households with children (95% CI: −0.74 to −0.04). --- ## Contact **Matthew P. Rabbitt, PhD** Email: matthew.p.rabbitt@gmail.com LinkedIn: [linkedin.com/in/matthew-p-rabbitt](https://www.linkedin.com/in/matthew-p-rabbitt/) --- *#FoodSecurityFridays — Measuring What Matters*