Chilean household food-demand MDCEV outputs and price-scenario predictions, stratified by socio-economic status (EPF IX, 2021-2022)

This study investigates how Chilean households would adjust their food purchasing patterns—and how their welfare would change—if they moved toward the healthy-sustainable diet recommended by the EAT-Lancet Commission. We merge nationally representative micro-data from the 2021-2022 Family Budget Survey (EPF IX, 5 682 urban households) with nutrient conversion factors and EAT-Lancet intake targets for 17 food groups. A Multiple-Discrete Continuous Extreme Value (MDCEV) demand system is estimated separately for three socio-economic strata (low, middle, high income). The model captures both the discrete choice of which food groups are purchased and the continuous decision of how much to spend on each, while allowing for corner solutions.

Using the fitted MDCEV coefficients, we run virtual-price simulations that make the EAT-Lancet basket just affordable for the median household in each stratum. These counterfactual prices generate predicted quantities and expenditure shares, from which we calculate Laspeyres, Paasche, Fisher and a novel “Naïve” price index. Compensating‐variation measures link the indices to consumer welfare, revealing which income groups would gain or lose from the dietary transition once realistic adoption frictions are acknowledged.

The accompanying data files include:

• γ, δ and β parameter estimates (with robust standard errors) for each income stratum
• simulated price predictions under the EAT-Lancet scenario.

All R scripts required to clean the raw EPF data, estimate the MDCEV models (via the apollo package), and reproduce every figure and table will be released in an open repository upon publication. The work provides the first demand-system-based affordability and welfare assessment of the EAT-Lancet diet in Latin America, offering actionable insights for nutrition and fiscal-policy design aimed at equitable, sustainable dietary change.