Price Discovery via Long-run Forecast

We demonstrate the superior performance of the price discovery measure recently developed by Kim and Linn (2022), termed the Long-run Forecast Share (LFS). Our examination involves a comparison of LFS with existing measures and highlights its wide applicability across various data generating processes. Recent studies, such as Shen et al. (2024) and Lautier et al. (2024), have overlooked reporting the uncertainty arising from finite sample estimation of price discovery measures. Our empirical investigation reveals that estimation uncertainty is significant in many cases, highlighting the importance of accurately quantifying this uncertainty. We introduce a novel approach for implementing the calculation of LFS based on its structural interpretation and demonstrate how our method allows quantification of the uncertainty associated with the measure. Our primary conclusions are based upon extensive simulation experiments across numerous data generating processes. We also present an in-depth investigation of price discovery in the spot and futures markets for key metal and energy commodities and find that LFS provides consistent conclusions across a variety of assumptions.