Post-Shock Forecasting Workflow
Overview | Package data structure | Donor balancing weights with dbw() | Mean forecasting with SynthPrediction() | ConocoPhillips example | Fit a one-step post-shock forecast | Inspect forecasts | Compare the forecasts | Inspect donor weights and the combined adjustment | Structured donor pools | Define the donor pools | Fit the MultiPool specification | Inspect the pool-specific adjustments | Volatility forecasting with SynthVolForecast() | Load the IYG example | Produce a variance forecast | Inspect the selected GARCH-X orders | Inspect donor weights and variance adjustments | Compare the variance forecasts | Interpretation and practical guidance | Align episodes consistently | Separate donor matching from model regressors | Choose donors using substantive information | Inspect weights and convergence | Interpret empirical variance forecasts carefully | Avoid post-shock information leakage | References