Food security, and vulnerability of Slovak households during the global crisis, 2022–2023

Summary

This analysis of Slovak Household Budget Survey data (2020–2023) used machine learning to identify households most vulnerable to food price inflation during the post-pandemic cost-of-living crisis. Vulnerability was concentrated among single-parent households, large families, the unemployed, disabled persons, pensioners, and tenants, with pronounced regional disparities favouring the capital over rural and eastern areas. The findings suggest that during severe shocks, universal support is necessary but must be complemented by targeted interventions for structurally disadvantaged groups.

Regional applicability

While conducted in Slovakia, the findings are relevant to United Kingdom policy analysis insofar as they demonstrate how machine learning can identify food price vulnerability clusters during inflation shocks—a methodology applicable to UK household survey data. However, direct policy transfer requires consideration of UK-specific welfare systems, regional inequality patterns, and demographic structures, which differ materially from Slovakia.

Key measures

Ratio of expenditures on staples to disposable income (top decile as vulnerability proxy); Categorical Boosting machine learning with SHapley Additive exPlanations; weighted logistic regression benchmarking

Outcomes reported

The study identified socio-economic characteristics of households vulnerable to food price inflation, measured as the ratio of staple food expenditure to disposable income. Machine learning analysis revealed that vulnerability was concentrated among specific demographic groups and regions, with disparities intensifying during peak inflation (2022–2023).

Topic tags