Essays on the empirical characteristics of market dynamics and order submission aggressiveness in the Stock Exchange of Thailand (SET)

This thesis contributes to the existing literature by undertaking three empirical studies addressing crucial issues within the market microstructure domain of the Stock Exchange of Thailand (SET). Following the introduction, Chapter 1 investigates how differences in market returns affect the price impact of trades in the SET, and how turnover and market capitalisation influence this relationship. Vector autoregression (VAR), impulse response function (IRF), cumulative impulse response function (CIRF), and panel regression are employed. The results show that entities listed in the SET100 index have low information asymmetry, low transaction costs, and high liquidity. The results also show that an escalation in turnover correlates with a heightened impact of trades on prices.
Chapter 2 examines the order submission aggressiveness of different types of trader in the Thai stock market - retail, foreign, and institutional - and the ways in which market conditions influence these order submission aggressiveness. Ordered logit regression methodology is employed, and the findings indicate distinct order submission aggressiveness among different trader types. Foreign traders are the most strategic traders implementing order submission aggressiveness, reflected by their intense activity in cancelling pending orders. Expressing a high willingness to cancel their existing orders underscores their sensitivity to the risk of non-execution, prompting them swiftly to adjust their pending orders with cancellations, and transition to more aggressive resubmissions.
Chapter 3 investigates the learning effects in traders arising from three separate market-wide applications of circuit breakers (MWCBs) during the first COVID-19 lockdown of March 2020. By using panel regression and difference-in-differences estimations, the results show consistent evidence across a wide range of different indicators that even though the first two MWCBs failed to restore market quality, their application nevertheless conditioned a dissimilar response to the third MWCB, allowing the stock exchange to curb volatility. This learning effect speaks directly to the individual effectiveness of MWCBs when used sequentially.