The financial market operates as an ecosystem, involving diverse yet interconnected marketplaces and participants. Market design, intricately interacting with technology, regulation, and competition, shapes how participants adapt their trading behavior and therefore influences market performance. This dissertation investigates how market microstructure impacts the behavior of fast and slow traders, the incentive for liquidity provision and liquidity demand, and the competition between exchanges for gaining order flow.

Article I examines a strategic solution to the concern that fast arbitrageurs make liquidity provision more costly by picking off quotes before market makers have time to revise them. The solution in question was to prohibit proprietary traders from engaging in liquidity taking, which prevents fast arbitrageurs from sniping market makers’ stale quotes. The results reveal that the trading ban mitigated adverse selection costs and narrowed the bid-ask spread. Market makers experienced higher profits and quoted higher volumes at better prices. The ban successfully eliminated over half of cross-exchange toxic arbitrage trades.

Article II evaluates the market quality effects of market fragmentation. The study leverages a quasi-natural experiment, which occurred when the Swiss stock markets suddenly transitioned from fragmentation to centralization in July 2019, following the breakdown of EU-Switzerland equivalence rules. Because this event was unrelated to technological developments, and did not disrupt trading, it establishes a firm ground for identifying the effect of fragmentation. The key finding is that greater market fragmentation improves market liquidity, as captured by bid-ask spreads and depth, while it does not impact market efficiency. These results align with theoretical predictions stating that market fragmentation improves liquidity through quote competition across exchanges.

Article III studies the role of restrictions on the minimum tick size and the minimum lot size for determining transaction costs at the futures market. The arrangement of minimum tick and lot sizes by regulators constrains trading at discrete prices and quantities. The study demonstrates, both theoretically and empirically, a trade-off between the restrictions of discrete price and discrete quantity. Given this tradeoff, a futures exchange can minimize futures transaction costs by choosing the optimal futures price. That is, when the futures price is high, it is also relatively more continuous, and the futures quantity is relatively more discrete, compared to the case when the futures price is low. In other words, when the futures price is high rather than low, the lot size restriction causes relatively more friction than the tick size restriction and vice versa. The empirical results strongly support the model.