Browsing Finance by Author "BAE, KYOUNG HUN"
Now showing 1 - 1 of 1
Results Per Page
- ItemESSAYS ON MARKET MICROSTRUCTURE(2015) BAE, KYOUNG HUN; Kyle, Albert S.; Business and Management: Finance; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This dissertation includes two essays on topics related to market microstructure. In the first essay, we analyze algorithmic trading in the Korean Index Futures market. We document that short-term traders consistently anticipate the order flow of large traders that build large positions within a short period of time. We study trade-by-trade data around 36,164 trades by large traders among the largest 1% of all active trades during 66 trading days in 2009 from the Korean Index Futures market. We find that large traders manage their orders first by executing small, positively correlated trades, which are followed by a single large trade. While the small trades are executed, short-term traders gradually increase their inventories in the direction of the forthcoming large trade. After the execution of the large trade, short-term traders unload their inventories to other traders. We find that short-term traders correctly anticipate the direction of large trades 56.06% of the time. Furthermore, the aggregate positions of short-term traders are statistically significant predictors for the direction of large trades that will arrive within 120 seconds. In the second essay, we explore market microstructure invariance in the Korean stock market. We define the number of buy-sell “switching points” based on the number of times that individual traders change the direction of their trading. Based on the hypothesis that switching points take place in business time, market microstructure invariance predicts that the aggregate number of switching points is proportional to the 2/3 power of the product of dollar volume and volatility. Using trading data from the Korea Exchange (KRX) from 2008 to 2010, we estimate the exponent to be 0.675 with standard error of 0.005. Invariance explains about 93% of the variation in the logarithm of the number of switching points each month across stocks.