Abstract

The Early Warning System project aimed at developing a system that can generate short-run forecasts. Nine commodities were included in this activity. These were palay, corn, onion, mango, banana, coconut, broiler, eggs and swine. Three methodological approaches were explored to develop the tool for forecasting. These were the survey-based method, modeling and Expert’s Opinion Survey. The survey-based method sought to get direct forecast estimates and/or indicators from which forecasts can be based. This involved either a modification of an existing BAS survey or development of an entirely new activity. Modeling was used to construct simple statistical models fitting a dependent variable, usually production of a commodity against some independent variables relevant for the commodity. The Expert’s Opinion Survey was used to validate the survey-based and model-based forecasts.