Tourism Forecasting: Error Magnitude, Direction of Change Error, and Trend Change Error

Forecasting accuracy can be assessed in various ways. The most popular accuracy measures are based on the magnitude of error, but directional accuracy and trend change accuracy are also important. Seven quantitative forecasting methods are used to forecast one-year-ahead international tourist flows, and the consistency of accuracy rankings across the various accuracy measures is analyzed. The "no change" extrapolation model is ranked most accurate in terms of error magnitude, econometrics is ranked first in terms of direction of change error, and exponential smoothing is ranked first in terms of trend change error.