Mechanisms Linking Economic Forecasting Access to Household Welfare in Kenya: Multilevel Causal Mediation Analysis

Economic forecasting is typically evaluated for its accuracy, yet its developmental value depends on whether forecast information reaches households and improves their welfare. The framework treats forecasting access as a district-level intervention, welfare as a household-level outcome, and the three mechanisms as mediators operating across levels. I argue that conventional single-level mediation analyses misattribute welfare gains because they ignore the institutional conditions under which forecasts are disseminated and the household capabilities required to act on them. The proposed approach integrates a counterfactual definition of indirect effects with hierarchical models that accommodate cross-level mediation, and it specifies identification assumptions that can be tested with observational data. The framework is intended to guide empirical research on the welfare returns to forecasting investment in agrarian economies.