We model two methods of executing segregated retail orders: brokers’ routing, whereby brokers allocate orders using the market maker’s overall performance, and order-by-order auctions, where market makers bid on individual orders, a recent U.S. Securities and Exchange Commission proposal. Order-by-order auctions improve allocative efficiency, but face a winner’s curse reducing retail investor welfare, particularly when liquidity is limited. Additional market participants competing for retail orders fail to improve total efficiency and investor welfare when entrants possess information superior to incumbent wholesalers. Our results hold when new entrants are less informed or the information structure differs. We also examine the cross-subsidization of brokers’ routing.

Postcrisis bank regulations raised market-making costs for bank-affiliated dealers. We show that this can, somewhat surprisingly, improve overall investor welfare and reduce average transaction costs despite the increased cost of immediacy. Bank dealers in OTC markets optimize between two parallel trading mechanisms: market making and matchmaking. Bank regulations that increase market-making costs change the market structure by intensifying competitive pressure from nonbank dealers and incentivizing bank dealers to shift their business activities toward matchmaking. Thus, postcrisis bank regulations have the (unintended) benefit of replacing costly bank balance sheets with a more efficient form of financial intermediation.

We study a dynamic stopping game between a principal and an agent. The principal gradually learns about the agent’s private type from a noisy performance measure that can be manipulated by the agent via a costly and hidden action. We fully characterize the unique Markov equilibrium of this game. We find that terminations/market crashes are often preceded by a spike in manipulation intensity and (expected) performance. Moreover, due to endogenous signal manipulation, too much transparency can inhibit learning and harm the principal. As the players get arbitrarily patient, the principal elicits no useful information from the observed signal.

U.S. retail brokers provide “best execution” for client orders, though how this actually works remains unclear. We model the interaction between brokers and wholesalers executing orders as imperfect competition, where privately-informed wholesalers compete for future order flow through current price improvement, with brokers designing allocation rules to enhance competition. Using data from three large retail brokers, we document three main findings. First, brokers allocate more orders to wholesalers with better past performance. Second, wholesalers recognize this and respond strategically to broker routing criteria and competitive pressures. Finally, there are differences among stocks and brokers in managing competition.

Retail Liquidity Programs (RLPs) allow on-exchange, sub-penny price improvement for retail orders, yet capture only a negligible fraction of retail flow. We find RLPs deliver less price improvement than off-exchange wholesalers, wholesalers route more orders to RLPs when market conditions deteriorate, and wholesalers internalize even when an RLP offers a superior price. Evidence from the SEC's Tick Size Pilot further shows that off-exchange effective spreads are insensitive to wider quoting increments, whereas RLP effective spreads narrow. We rationalize these findings with a model of wholesaler routing built around three frictions: non-displayed prices, absence of order protection, and wholesaler routing discretion.

A lender uses a predictive model to assess borrower quality based on manipulable data. The model is unobservable to the borrower, but the lender can disclose information about it. Both full disclosure and no disclosure lead to excessive data manipulation. Under full disclosure, the borrower can perfectly "game" the model. Under no disclosure, the lender's equilibrium use of data exceeds the ex ante optimal level, which also exacerbates the borrower's incentive to manipulate. We fully characterize the optimal policy, which preserves significant uncertainty in the borrower's posterior by introducing gaps in beliefs. When restricted to monotone policies, the optimal disclosure policy always involves discrete signals, and we provide conditions under which no disclosure becomes optimal. Our results are robust to extensions with commitment to lending decisions and costly verification.

[Abstract]

[Abstract]

[Abstract]