Trend Stability and Structural Change: An Extension to the M1 Forecasting Competition

The global linear trend with autocorrelated disturbances is a surprising omission from the M1 competition. This approach to forecasting is therefore evaluated using the 51 non-seasonal series from the competition. It is contrasted with a fully optimized version of Holts trend corrected exponential smoothing. It is found that an adaptation of Holts method, in which the growth rate is restricted to be constant, performs almost as well as its traditional counterpart and usually out-performs the global linear trend with autoregressive disturbances.