Purchased resilience in the food industry: how COVID-19 deferred structural vulnerability in Italian agri-food SMEs (2019–2023)

Purpose To examine whether SMEs in Italian food and beverage processing were disproportionately affected by COVID-19, and by what means short-run outcomes were sustained, developing purchased resilience: performance maintained through external resources whose cost is deferred. Design/methodology/approach Accounting panel for 2,661 Unionalimentari-affiliated SMEs (2019–2023) against 199 enterprises of 250 or more employees. Fixed-effects estimates of differential changes in profitability and leverage, plus a test of whether leverage accumulated under the support schemes predicts later performance. Findings Profitability shows no detectable differential shift between size classes (ROE β = −0.262; ROI β = −0.082; both n.s.). Leverage diverged: SMEs deleveraged more slowly (β = +2.061, p = 0.022), widening an eighteen-point pre-existing gap. Unit labour costs rose 16.6% for micro-enterprises against 6.6% for large firms, and firms accumulating more leverage recorded weaker 2022 ROI (−0.184, p < 0.001). Research limitations/implications With one pre-treatment year, parallel trends cannot be verified; estimates are differential changes, not causal effects. The groups differ in more than size, and 2023 coverage reaches 56.0% of 2019 reporters. Practical implications Crisis instruments delivered as debt require an exit design. The change in the debt ratio over that period is observable in filed accounts and could screen firms for post-crisis measures. Originality/value Introduces purchased resilience as a construct distinct from absorptive, adaptive and financial resilience, with firm-level evidence linking leverage divergence to later performance in the context of broad crisis-support mechanisms.