The roles of circuit breakers in Malaysia capital market : an in-depth analysis

Abstract

This thesis investigates an oft-implemented intervening instrument in financial markets: circuit breakers. Proponents claim it endows a propitious time-out when asset (or market) prices are stressed and persuades traders to make rational trading decisions. Empirical works examining performances of regulatory rationales, such as deterring volatility, enhancing price discovery, interferences in trading, and a self-fulfilling gravitational pull (dubbed the magnet effect) seem to mushroom soon after headline-grabbing financial crises or flash crashes (as in May 2010 in the US). Though Asian markets such as Korea, Taiwan, Tokyo, Shanghai, and Shenzen dominate the studies with mixed results, the only major study undertaken in Malaysian market by Chan et al. (2005) found evidence of worsening market quality, using transaction data from 1995. Since then, KLSE has become BM, trading platforms have grown sophisticated, and circuit breaker regime in the bourse underwent tweaks. Despite the changes, the limit of ±30 percent is intact since 1989, albeit with some qualifications. This is rather puzzling, for (a) most exchanges prefer a very tight collar, (b) many exchanges experiment with the limit in tranquil times, presumably in quest of an optimal collar, (c) advanced exchanges commission studies corroborating the efficacy of the proposed regime and make the results known, and (d) most exchanges play with the circuit breaker around market crash periods to forestall the crash or reinforce confidence. Somehow, Bursa Malaysia did none for nearly 3 decades. The investigations of this thesis report result that are mixed to varying degrees of statistical robustness with no clear indication of improvement in stopping volatility to spill over the subsequent trading days and delay in emergence of equilibrium price. The limits, however, do not appear to interfere with trading activities, an indirect proxy for liquidity. The third empirical investigation utilizes proprietary high-frequency intraday data from January 2015 to September 2017 to examine existence of magnet effect and finds mixed evidence supporting a weak form of magnet effect and nearly as much repellant effect. Vast majority of stocks, however, exhibited neither. Moreover, price acceleration in magnet-esque scenarios were severe for most stocks with disproportionate lack of support from underlying order flow and volume. Lastly, the overall performance of different circuit breaker regimes in promoting efficient pricing via random walk is examined for affected stocks through a battery of parametric and nonparametric tests. The evidences favor the regulatory practice and indicate a liberal band corresponds with propensity for greater random walk.