Author: Mingyuan Kong
Degree: M.S. in Financial Engineering
Advisory Committee: Dr. Dragos Bozdog, Dr. Rupak Chatterjee, Dr. Ionut Florescu
Abstract: Liquidity describes the degree to which an asset or security can be quickly bought or sold in the market without affecting the asset’s price. Evaluating market liquidity levels and analyzing liquidity measures are valuable. In this thesis, existing liquidity measures are studied and their distribution characteristics are analyzed during the Brexit.
ETF (Exchange-Traded Fund) components are tested in this thesis. The symbols are
28 components of Utilities Select Sector SPDR Fund (XLU). The time period analyzed is between June16, 2016 and June 30, 2016, including the Brexit result release date June 24, 2016 for a total of 11 trading days. Both trade quote data and limit order book data are high-frequency tick level data.
TAQ (Trade and Quote) liquidity measures and limit order book liquidity measures are studied in this thesis. Besides mathematical definitions of these measures, true ranges and other statistical features are given which help to understand liquidity measures in the real market. A phenomenon is found by observation that the aggregate correlation measures have corresponding price changes, and after further studies, it is concluded that low correlation between TAQ measures and LOB measures indicates high probability to have large price change. This helps to understand the relationship between TAQ measures and LOB measures giving a prediction of price change range. GAMLSS framework is employed to find the best fitted distribution for different measures in each day, and by scoring method, the best distribution is selected. Clustering based on best fitted distributions is implemented, and by changing the criteria level, some liquidity measures are persistent in particular clusters. The results are examined through diagnosis test. Distribution parameters are analyzed under the best distribution framework to find the performance on special events. Furthermore, empirical distributions are studied, and their distribution characteristics demonstrates that Brexit caused fatter tails on liquidity measures distributions. This indicates that infrequent (low) liquidity condition occurs more frequently on the Brexit day.