Measurement errors in index trader positions data: Is the price pressure hypothesis still invalid?

In this paper, we examine whether the repeated rejection of Masters's price pressure hypothesis is robust with respect to measurement errors in index trader positions data. We allow for autocorrelated errors and a potential impact of index trader positions on both the level and volatility of commodity returns. The resulting state-space model is estimated via particle Markov chain Monte Carlo (PMCMC). The empirical investigation relies on weekly data for 12 commodities listed in the SCoT and DCoT reports from 2006 to 2020. Our empirical findings show that the rejection of the price pressure hypothesis is robust concerning the inclusion of measurement errors in index trader positions data.