Zero-intelligence realized variance estimation. (Q2430259)

This paper compares a comprehensive set of ``second generation'' realized variance measures using a simulated ``zero-intelligence'' order book market. The emphasis of this investigation lies on statistical efficiency, implementation, and robustness. An important aspect of the analysis is the focus on data sampling. The authors consider three different sampling schemes that can all be implemented in practice and give rise to price series with radically different properties: the series of trade prices, series of mid-quotes, and the series of micro-prices formed by linear weighting of the best bid and ask price by market depth. The limit order book model allows them to implement these sampling schemes in an internally consistent manner and to gain insights into their properties. Performing an extensive amount of tests and simulations, the authors give their recommendations on the use of the compared measures, and their conclusions seem to agree with those of other authors.