Research

Publications


Working Papers

We study the dynamics of bid-ask spread and trading volume using a multi-period trading model with asymmetric information and oligopolistic market makers.   Market makers optimally make offsetting trades in "bid" and "ask" markets by adjusting bid and ask prices/depths to avoid holding inventories.  We find that when market makers have significant market power, other traders optimally smooth out their trading even though they are not strategic.  Consequently, trading persists even after arrivals of information and liquidity shocks.  Traders trade quickly to compete on their private information while postponing their hedging trades until later periods.  As a result, both trading volume and bid-ask spread may exhibit U-shaped patterns. More trading rounds mitigate market makers' market power, and thus decrease bid and ask spreads and increase total trading volume. Therefore, while a higher trading frequency benefits traders, it may hurt market makers. 

 Presented at    2019 Western Finance Association Meetings                             
                              2018 NYU Stern Microstructure Meeting
                              2018 Finance Theory Group  Meeting at MIT

        2019 Best Paper in Investments, FMA Asia/Pacific Conference


We study  trading among strategic informed traders who may have incorrect beliefs about market crowdedness defined as the correlation of traders' private signals about fundamental values. These mistakes distort equilibrium strategies and prices. When traders underestimate market crowdedness, they target larger inventory levels, trade more aggressively, and provide more liquidity to others.  The actual profits are lower than expected because traders underestimate the amount of information already impounded into prices. Crowded markets are more vulnerable to abrupt crashes. The magnitude of price crashes  and the speed of recovery during  off-equilibrium fire-sale events can help infer traders' misperception of market crowdedness. 


We study how  asset managers'  benchmarking affects market efficiency under two learning technologies: separative and integrative. With integrative learning, investors process portfolio-wide signals, optimizing information allocation across  assets instead of focusing on individual ones. Therefore, increased benchmarking on an asset with greater uncertainty can enhance its price informativeness as investors direct more attention to it---a stark contrast to the results under separative learning. Moreover, our study indicates that benchmarking could increase overall market efficiency, with each learning technology presenting distinct implications for asset pricing. These findings emphasize the critical role of learning technology in understanding the effects of benchmarking.





We study how the frequency of information disclosure affects the welfare of market participants in different market conditions. In competitive markets, we show that both informed and uninformed traders benefit when firms gradually disclose information, as opposed to releasing it all at once. Gradual disclosure reduces uncertainty over time, enhancing risk-sharing and improving welfare for all traders. This suggests that, in competitive markets, firms  should release information as it becomes available, rather than infrequently. Conversely, in non-competitive markets where informed traders possess market power, increasing the frequency of information disclosure continues to benefit uninformed traders but may negatively impact informed traders who trade for both private information and liquidity shocks. This is because increased frequency of information disclosure may increase the cumulative costs associated with hedging.  Our paper suggests that,  firms might opt for less frequent information disclosure if their decisions are influenced by institutional traders with monopoly power.



 

Existing literature concludes that binding short-sale constraints limit the revelation of negative information but overlooks negative information in informed sales. Our study, based on data on informed institutional and illegal insider sales, reveals that informed sales can significantly weaken the impact of short-sale constraints. We provide evidence of a lead-lag information transmission mechanism, with informed sales unidirectionally leading short-sales, other institutional sales, and the stringency of short-sale constraints. These findings are supported by a rational expectations equilibrium model. Our analysis underscores the undervalued role of informed sales in understanding overpricing and market bubbles.


We analyze the impact of indexing when investors can endogenously acquire information. We find that the impact critically depends on what drives the rise of indexing. If indexing is driven by reduced  participation costs or decreased  liquidity trading in the non-index market, then:  (1) price informativeness and relative price efficiency increase, (2)  the welfare of non-liquidity traders decreases,  and (3) both stock correlations and the expected market capitalization increase.  The opposite is true if indexing is driven by  increased   liquidity trading in the index market. Our analysis  highlights the importance of identifying  the driving forces behind the rise of indexing.




Using Trade and Quote (TAQ) data from 2000 to 2021, we find that NASDAQ stocks exhibit pronounced U-shaped patterns in bid-ask spreads and trading volumes. Traders often trade aggressively at market open to capitalize on their private information and rebalance their portfolios at market close, leading to concentrated trading at market open and close. Our findings suggest that the U-shaped pattern in bid-ask spread arises from imperfect competition among liquidity providers, who may increase spreads in response to heightened trading incentives. The U-shaped pattern is more pronounced for smaller stocks and those with larger order imbalances, consistent with our hypothesis and existing theoretical predictions. Moreover, the spikes in bid-ask spreads at market close help explain the observed trend of higher overnight returns compared to intraday returns for certain stocks on NASDAQ.


We present a one-period model of oligopolistic strategic trading among symmetric traders who agree to disagree about the precision of their private signals. We derive several invariance relationships relating the number of firms, number of firm's employees, average trade size, price impact, and pricing accuracy to dollar volume and returns volatility. Since a substantial part of order flow is often internalized within firms and does not reach the marketplace, invariance relationships can be modified to account for internalized order flow. We also clarify important conceptual issues arising when one seeks to use a one-period model to generate predictions about dynamic markets.