Abstract

ABSTRACT
Three Essays on the Microstructure of the BIST
Osman Ulas Aktas, Ph. D.
Concordia University, 2016
This thesis consists of three essays. The first essay (chapter 2) examines the accuracy of five algorithms for classifying trades as buyer- or seller-initiated for BIST-30 index constituents over a period including the Lehman collapse. The highest classification accuracy rate (over 95%) is for the one-second lagged Lee & Ready (LR) algorithm. The LR’s classification accuracy is highest (lowest) for trades representing mixed agency and principal (pure principal) relations between clients and executing brokers. Unlike for U.S. markets, almost all trades are classifiable with accuracy rates of 90-plus percent for both long and short trades. As for U.S. markets, higher misclassification rates occur for trades in the first versus last 30 minutes of the trading day, as the time between consecutive trades decreases, and for decreasing trade sizes.
The second essay (chapter 3) examines the trade price effects and their determinants for BIST-30 index constituents for a period that includes the Global Financial Crisis and the Lehman collapse. Consistent with theoretical predications, we find that informed trades in the BIST tend to be large. Our findings that price discovery appears to be fairly rapid on the BIST and that the average multi-sample stock trade price effect of less than 30 basis points is competitive with other markets have important implications for the purchase and execution decisions of investors. Our finding of positive mean price effects for short trades that are larger for seller-initiated trades and larger than for long trades has implications for the ongoing debate about the regulation of short sales since it suggest that the average short sales does not depress prices. Furthermore, the higher price effects of (especially buyer-initiated) trades in the last minutes of a trading session and the variation in price effects with whether the client-broker relationship is agency, principal or mixed have important implications for market regulators in terms of refining their surveillance systems to better control any inappropriate stealth trading or end-of-session price manipulation.
The third essay (chapter 4) examines the price-limit hits for members of the BIST-50 index during the March 2008 through March 2009 period. The effects of price-limits are not homogeneous for upper and lower hits when they occur and if they continue in a subsequent trading session. Our results are supportive of the no-, dampening and spillover effects on volatility hypotheses, overreaction and no-effect price hypotheses, magnet price effect hypothesis, and greater informational asymmetric effect on market-quality hypothesis. They are not supportive of the price-delay hypothesis, and trading interference hypothesis. The results are robust using equi-distant and trade-by-trade returns and volatility measures ccounting for the autocorrelations in these return series. The results have implications for regulators contemplating the introduction of similar mechanisms or fine-tuning their current mechanisms.