Using time series data on municipal-level household food scrap drawn from a city-wide food scrap recycling program from 2007m1 to 2018m8, we proxy the household food wastage rate (FWR) by the edible-to-inedible food scrap ratio and then apply both linear and semiparametric varying-coefficient cointegration tests to examine the municipal-level long-run relationship between socioeconomic factors and household FWR. The linear model shows that food price, old-age population share, and average household size are positively related to FWR, whereas working-age population share is negatively related to FWR. Consistent with the environmental Kuznets curve hypothesis, the varying-coefficient model further reveals that the direct relationship between income and FWR is inverted-U shaped. Moreover, income alters the relationship between socioeconomic factors and FWR. The FWR-increasing effect of population aging (food price) likely aggravates (diminishes) with economic growth. With these in mind, for policymakers trying to reduce food waste, the steady increases in food price and old-age share pose challenges, whereas the declining trend in average household size brings relief. Given the results, one way to reduce food waste is subsidizing both the design of smart packaging and the development of semi-prepared convenience food with flexible portion sizes and portion sizes suitable for older adults.
