روند های بین المللی در طراحی بازار: اثرات درون زای شفافیت سفارشات محدود بر نوسانات، گسترش، عمق و حجم

Following other leading international securities markets, the Tokyo Stock Exchange (TSE) has adopted a publicly displayed but anonymous limit order book, and we ask: how is market quality affected? Accounting for fixed effects and endogeneity, we find increased volatility and higher order book depth at the best bid and ask prices, while total depth is not significantly impacted. This predicts more competitive order strategies in a trading system with anonymous orders but with more visible price levels. Spreads are found to be unaffected by the market design change, in contradiction to previous literature. Complementing the literature, we find volume increases, indicating that the aggregate effect of the design change is positive.

A recent trend in international securities markets has been to cease disclosing pre-trade broker identification and at the same time increase the visibility of the now anonymous order book.2 This paper investigates the impact of this common change in pre-trade transparency on market quality. This is an important question as the change in pre-trade transparency may affect market quality in several ways. Anonymous trader identities may potentially have negative effects as the search for counterparties becomes more costly. The ability for brokers perceived as informed to remain anonymous may again have positive effects. A greater transparency of the depth of the order book is expected to have positive effect, as trades can be executed with greater accuracy when the exact volume on several price levels is known. In this paper we measure market quality as volatility, bid-ask spreads, limit order book depth and volume as we investigate the Tokyo Stock Exchange (TSE), the second largest stock market in the world by market capitalization and fourth by traded value up to 2009.3 The TSE ceased to display broker identity to exchange members in June 2003, and simultaneously increased the level of limit order book information by providing the three best bid- and ask-price levels and volumes to the public. This provides us with an excellent opportunity to study the impact on market quality of a common change in market design, and to do so using data from a highly significant stock market that also represents a pure limit order book market rather than a hybrid of dealer- and order-driven markets. The studied change in market design is common for limit order book markets, while hybrid markets tend to provide some transparency of dealer identity. Although we recognize that the market design change investigated in this paper is complex, it is the design of choice for most limit order markets, and hence it is important to provide a more complete picture of the impact of such a common design change. In one previous study, where the 2003 market design change at TSE is included, Comerton-Forde et al. (2005) focus on the impact on spreads, which does not provide a complete picture of the impact of the market design change studied here. In addition, we contribute to the literature by addressing the endogeneity problem of previous empirical research that typically examines one statistic for market quality and uses several others as exogenous controls.4 We use several alternative metrics for market quality, and methodology that accounts for fixed effects and endogeneity in the included statistics for market quality. We find that volatility increases after the TSE market design change. The depth of the limit order book increases at the best bid and ask prices while total depth is not significantly impacted. Accounting for the interaction between spreads, volatility and volume, we find that spreads are not affected by the design change. Previous literature does not consider the volume effects of transparency. In our paper, we find that volume increases as a result of the TSE design change, which is consistent with the higher volatility.

ConclusionsWe investigate the impact of a common change to pre-trade transparency on market quality. Manyexchanges have recently ended disclosure of broker identification and, over time, increased the levelsof price and volume displayed from the limit order book. Effectively this is the market structure thatNYSE-Euronext, Nasdaq, Nasdaq-OMX, most European exchanges and Australia have moved to. Wemeasure the effect of the introduction of this policy on volatility, transaction costs, LOB depth andtraded volume in the Tokyo Stock Exchange (TSE). Foucault et al. (2007) confirm that volatility co-moves with a variety of other variables, such as the number of trades, when opacity is introduced.Further, there were two conflicting events taking effect simultaneously in the TSE, which may causeambiguous effects on liquidity and order aggressiveness. This is why we implement a two-stage least-squares model with fixed stock effects to examine the impact of the change in pre-trade transparencyat the TSE.We find that volatility increases after the TSE market design change which we attribute to a sloweradjustment to new information when transaction counterparties are unknown. The depth of the limitorder book increases at the best bid and ask prices while total depth is not significantly impacted.This supports earlier studies in the literature that predict an increase in depth as a result of morecompetitive order strategies in an opaque system. Our results also confirm findings in earlier studiesthat an increase in reported LOB depth is associated with a greater depth of the LOB. Accounting for theinteraction between spreads, volatility and volume, we find that spreads are not affected by the TSEmarket design change, which is in contradiction to results in previous literature. We show this to be anexpected result of the combined effect of the removal of broker identity and more visible LOB price and order size information. Previous literature largely ignores the volume impacts of transparency alteringmarket design changes. We find that volume increases as a result of the TSE market design change,which is consistent with the higher volatility. Overall, we conclude that the TSE market design changeincreases the welfare of traders, as it promotes higher traded volume, the most important indicatorof gains from trade within a trading system.Interestingly, we find that the general combined effect of the removal of broker identity and morevisible LOB price and order size information is greater for smaller stocks and close to neutral for largerstocks. As institutional traders that benefit most from anonymous broker IDs are active mainly in largestocks, it appears that the market design change has not affected large stocks negatively, while it hassome positive effects on smaller stocks. Displaying more information about orders in the LOB, albeitanonymously, has previously been shown to improve liquidity (Boehmer et al., 2005). The good newsis that while the search for counterparties becomes more costly when broker IDs are anonymous, inthe Tokyo case, the benefits from a more transparent LOB appear to outweigh any negative effects.