Adaptive Expectations, the Exponentially Weighted Forecast, and Optimal Statistical Predictors: A Revisit

Relationships between adaptive expectations, the exponentially weighted moving average, and optimal Univariate statistical predictors are reviewed We show that the behavioral-based adaptive expectations are a subclass of both the exponentially weighted moving average and the (0,1,1) ARIMA model The applicability of the adaptive expectations model to 25 empirical price and quantity series is investigated The adaptive expectations behavior and the optimal statistical forecasts are equivalent for 13 senes-11 on Yields and 2 on prices Numerous price series, while exhibiting the general form o[ the adaptive expectations (a (0,1,1) ARIMA process), did not have a coefficient o[ expectations within the originally hypothesized range The behavior consistent with the model underlying these price series would be trend extrapolation rather than averaging (averaging the most recent observation and Its forecast) Series measured at monthly or quarterly Intervals were not adequately modeled by adaptive expectations or as a (0,1,1) ARIMA process