ABSTRACT
export instability is defined here as short-term fluctuations in export earnings corrected for trend. Some form of trend correction is needed to avoid interpreting a constant year-to-year increase or decrease as indicating instability. The number of methods of trend correction possible is almost as numerous as the number of studies on export instability. As long, however, as each index is calculated for the same variable over the same period of time, the results are invariably highly correlated. Most techniques assume a linear time trend. With countries where no single linear trend fits, the index will tend to exaggerate the amount of short-term instability. For this reason I prefer an index of instability which is measured as the average percentage deviation of the dollar value of export proceeds from their five-year moving average centred on the mid-year. This has the disadvantage of losing two years from the beginning and end of the time series and of being less convenient for computer calculation. The indices employed in this study are specified in the footnotes as they are used.
