Bridge homogeneous volatility estimators

We present a theory of bridge homogeneous volatility estimators for log-price stochastic processes. Starting with the standard definition of a Brownian bridge as the conditional Wiener process with two endpoints fixed, we introduce the concept of an incomplete bridge by breaking the symmetry between the two endpoints. For any given time interval, this allows us to encode the information contained in the open, high, low and close prices into an incomplete bridge. The efficiency of the new proposed estimators is favourably compared with that of the classical Garman--Klass and Parkinson estimators.