Robin Yifan Luo

Misallocation in Banking

Solo-authored · Draft available upon request

Uneven bank pricing power can direct deposits and loans away from the banks where they create the most surplus at the margin. Banking makes this misallocation problem two-sided because demand and cost complementarities link the two sides of bank balance sheets. Using a two-sided oligopolistic model and U.S. bank data from 1986 through 2021, I estimate that annual welfare losses from misallocation range from $1.3 to $3.1 billion in 2019 dollars. Their share of sample banks’ aggregate net income declines from an average of 7 percent through 2006 to 2.2 percent thereafter. Loans contribute more to the loss, while the deposit-loan interaction offsets a substantial share because efficient reallocations often expand one side of a bank and contract the other. After merger completion, surviving acquirers move farther from efficient marginal-surplus benchmarks, yet the merging parties’ combined welfare loss declines by $3.7 million on average. Holding the structural estimates fixed in 2021, I find that deposit reallocation increasingly moves against the efficient direction during the 2022 monetary tightening and 2023 bank stress, while loans move with it. The misallocation framework extends beyond market power to cost wedges across banks and beyond banking to industries in which demand or production links multiple products.

with Felix Feng and Beatrice Michaeli

Journal of Accounting and Economics, 2024

presented at the 2023 Journal of Accounting and Economics conference

We study the optimal dynamic contract that provides incentives for an agent (e.g., SPAC sponsor, VC general partner, CTO) to exploit investment opportunities/targets that arrive randomly over time via a costly search process. The agent is privy to the arrival as well as to the quality of the target and can take advantage of this for rent extraction during the search process and the ensuing production. The optimal contract provides the agent with incentives for timely and truthful reporting via a time-varying threshold for investment and an internal charge for the time spent on search. In the equilibrium, as time elapses, the charge becomes progressively higher while the investment threshold is progressively lower, resulting in overinvestment at a time-varying degree. Our model generates empirically testable predictions regarding investments (such as M&As, hedge fund activism, VC investing, SPACs, and internal innovations), linking the degree of overinvestment to observable firm and industry characteristics.

The Optimal Schedules of Incentives and Cash Flows

with Felix Feng and Mark Westerfield

We model delegated management when an agent can invest in durable asset quality or raise instantaneous cash flow. Both actions add value, but only aggregate output is observable, so incentivizing durable investment necessarily also incentivizes instantaneous effort. The flexibility to schedule incentives makes durable investment relatively cheaper to induce, and the optimal contract begins back-loaded. As promised future incentives accumulate, further back-loading becomes increasingly costly, and the contract can eventually become front-loaded. Negative cash-flow surprises raise promised future incentives. The resulting dynamics generate mean reversion in cash flows and other empirical implications for the duration of cash flows and incentives.