We present a high-frequency based method for analyzing a one-period futures hedging problem. The realized hedge ratio is constructed by applying the realized regression with the R-squared coefficient as an ex-post performance measure. The asymptotic theory enables us to assess the parameter estimation risk of the hedge ratios. An empirical study is conducted on the S&P 500 index and their hedging performance is compared to the conventional hedge ratios. Moreover, impacts of the market microstructure effect on the realized hedge ratio and the effect of the parameter estimation risk on the corresponding hedging performance are
discussed.