International FinancialManagement 3国际财务管理课件.ppt

* Forecast Bias The following regression model can be used to test for forecast bias: realized = a0 + a1 ′ forecast + m Where a0 = intercept a1 = regression coefficient m = error term * Forecast Bias *If the predictor is unbiased, the intercept should equal zero, and the regression coefficient should equal 1.0. *If a0 = 0 and a1 is significantly less than 1.0, this implies the predictor is systematically overestimating the spot rate, for example, if a0 = 0 and a1 = .90, the realized value is estimated to be 90 percent of the forecasted value. Vice versa. *If a predictor is found to be biased, the estimated a0 and a1 values can be used to correct the systematic error. * Graphic Evaluationof Forecast Performance Forecast performance can be examined with the use of a graph that compares forecasted values with the realized values for various time periods. * Graphic Evaluationof Forecast Performance Region of upward bias (overestimating) Region of downward bias (underestimating) Perfect forecast line $.10 $.12 $.14 $.16 $.18 $.20 $.22 $.24 $.26 $.28 $.30 $.28 $.26 $.24 $.22 $.20 $.18 $.16 $.14 $.12 $.10 Predicted Value (in U.S. dollars) Realized Value (in U.S. dollars) . . . . . . . . * Graphic Evaluationof Forecast Performance If the points appear to be scattered evenly on both sides of the perfect forecast line, then the forecasts are said to be unbiased. Note that a more thorough assessment can be conducted by separating the entire period into subperiods. * Comparison ofForecasting Techniques The different forecasting techniques can be evaluated graphically - by comparing the distances from the perfect forecast line, or statistically - by computing the mean of the absolute forecast errors, and then us