Answers are grounded in this paper and cite their sources.
In shrinking cities, population decline and aging influence local food retail environments in ways that GIS-based accessibility measures cannot fully capture, because these measures record where outlets exist rather than how they are used. Using a balanced monthly panel of 205 sub-districts in Busan, South Korea (January 2023 to December 2025, 7,380 observations), this study estimates sub-district fixed effects from two-way panel models across three outcome dimensions, including total transaction value, transaction count, and mean transaction value, separately for food-specialized (Food) and general retail (Retail) establishments. Sub-districts in the lower 10th percentile of fixed effects are classified as consumption-vulnerable and cross-tabulated across channels into four categories, including Type A (Neither), Type B (Retail-only), Type C (Food-only), and Type D (Both). The results show three distinct formation pathways, including a simultaneous supply and demand collapse in Type D sub-districts concentrated in the old urban core of Busan, a demographic contraction pathway in Type C sub-districts, and a supply infrastructure gap in Type B sub-districts. These vulnerability patterns are spatially clustered rather than randomly distributed. This diagnostic framework provides shrinking cities with a screening tool to monitor early warning signals of food access vulnerability before conventional food desert conditions occur.
No takes yet. Share an insight, caveat, or question.
Kim et al. (2026) studied this question.