By term structure I mean a non-stationarity in the pattern of intraday adverse selection as a given instruments approaches its expiry. Note that I am interested in the adverse selection on the instrument itself, not the underlying.
On the empirical side, I've seen myself a significant uptick of the price impact on S&P500 E-mini futures in the week leading into expiry. I am interested in understanding if that empirical observation is linked to a general decrease of liquidity as people rotate out of the instrument, or if there is something else in play.
References on models, whether theoretical or empirical would be greatly appreciated!