A new approach to the evaluation and selection of leading indicators

Leading indicators are typical constructs used in macroeconomics to guide decision making in several areas of economic activity, including policy formation and long term investment. Researchers often evaluate and select leading indicators on a seemingly ad hoc basis involving OLS regression, which does not take into account the fact that perhaps the most important property of a good leading indicator lies in its ability to anticipate the turning points of the time series of interest. We propose an alternative assessment of leading indicators, based on the turning point significance transform, which weights each observation of the original time series according to how much it functions as a turning point. This new construct is then used to evaluate the accuracy and timeliness of several German and American macroeconomic time series as leading indicators for GDP growth.