Forecasters often make judgmental adjustments to exponential smoothing forecasts to account for the effects of a future planned change. While this approach may produce sound initial forecasts, it can result in diminished accuracy for forecast updates. A proposed technique lets the forecaster include policy change adjustments within an exponential smoothing model. For 20 real data series representing Virginia Medicaid expenses, initial forecasts and forecast updates are developed using the proposed technique and several alternatives, and they are updated through various simulated level shifts. The proposed technique was more accurate than the alternatives in updating forecasts when a shift in level occurs approximately as planned.
Williams, D. W., & Miller, D. M. (1999). Level Adjusted Exponential Smoothing: A Method For Judgmentally Adjusting Exponential Smoothing Models for Planned Discontinuities. Paper presented at the Decision Sciences Institute, 1999, Athens.