Non-Microstructure Noise

High resolution Ultra High Frequency (UHF) quotes and trades are examined to detect patterns which do not correspond to plausible market activity as in Brownlees and Gallo (2006). Non-microstructure noise is identified as a process. Methods of treatment are evaluated. A resultant paradigm of data handling is presented. Empirical evidence from a sample of NASDAQ stocks demonstrates that the removal of non-microstructure noise from the limit order book significantly improves estimation across techniques and all levels of market depth