Title: An inside look to the spectrum of realized betas
Authors: Malvina Marchese - Cass Business School (United Kingdom) [presenting]
Rodrigo Hizmeri - Lancaster University (United Kingdom)
Marwan Izzeldin - Lancaster University Management School (United Kingdom)
Abstract: The impact of the generally ignored log-price drift in estimating realized betas is investigated. The presence of non-negligible drift results in poor measures of the realized betas, mainly because the non-negligible drift negatively impact the realized (co)variance measures. Our Monte Carlo experiment confirms the dramatic impact of the drift-term not only in the generic realized beta, but also in the continuous and jumps betas. The main distortions occurred via three channels: i) high levels of the drift-term in the stock and/or the index; ii) when the drift-term of the stock and the index differ greatly; iii) as the sampling frequency decreases. We propose an alternative approach that mitigates the impact of the temporary non-negligible drift, yielding more accurate estimates of the realized betas and their risk-premia. We illustrate the usefulness of our new approach using all the Dow Jones constituents for a period of 17 years.