"Essays on Agency Costs of Financial Intermediation"

A financial intermediary is a delegated monitor that produces information and adds value to capital allocation between market agents. A borrower, an entrepreneur or a firm obtains capital investments, and in return, the lender or investor profits through interest payments or equity shares. These transactions are facilitated by or implemented under intermediaries, either commercial banks, investment banks, credit rating agencies, venture capitalists or platforms. As economic agents, financial intermediaries may not have their interest aligned with that of clients, which can result in inefficient capital allocation, market failure and financial instability. This dissertation studies conflicts of interest in different types of intermediation and from various aspects.

In the first chapter, I study how financial intermediaries balance between market share and reputation under competition, using unique datasets on loans originated and declined on peer-to-peer lending platforms. Using a platform entry event that intensifies market competition on borrowers and lenders, I document less prudent borrower screening, credit rating inflation and aggravated loan performance. In particular, post-entry borrowers are more likely to obtain financing, equally creditworthy borrowers receive better credit ratings and their average loan performance deteriorates significantly. It distorts platforms' incentive on truthfully reporting borrowers' risk to compete for market making. The incumbent platform lowers interest rates to encroach on creditworthy borrowers, indicating aggressive undercutting behavior. Raising interest rates on subprime borrowers maintains lenders' participation while accounting for competition-induced adverse selection.

I further document that, as disincentive for platforms' credit inflation, lenders exit the platform upon their borrowers' underperformance. In particular, with vintage loan performance deterioration, the number of lenders on a newly originated loan decreases, credit crunches emerge and capital flows slow down. The magnitude of the 'punishments' mitigates significantly post-entry, arguably because the market size expands with the entry event, eliciting new and unfamiliar lenders to enter, which intensifies borrower competition and credit inflation.

By fuzzy matching borrowers' identities between the platforms, I identify the incumbent's first mover advantage, where incumbent-rejected borrowers get financing at the entrant but rarely vice versa. As a dominant player, the incumbent gets high-quality borrowers and induces severe adverse selection for the entrant. I contend that this first mover advantage is endogenized by the incumbent's active borrower screening beyond "hard information" and its capital provision for borrowers facing credit crunches.

The second chapter, coauthored with Zhaohui Chen, Alan Morrison and William Wilhelm, examines and identifies the underlying mechanism of the decline of investment bank-client relationships from 1960 --- present. As investment banks know superior information about their clients in security underwriting, the internal agents, investment bankers, often face conflicts of interest and thus, have incentive to misuse the information against clients. Without contractibility, banks' internal governance and monitoring provide incentives for bankers to harness their relationship with clients by making agency problems costly.

The adoption of computing technology started in the 1960s has caused investment banks' internal governance to evolve. Advances in technology and novel financial economic theory make it profitable to be engaged in the trading and risk-taking business, which induces investment banks to get bigger in scale and more complex in financial innovation. The increasing internal liquidity dampens senior bankers' incentive to train and monitor younger partners by easing the mobility of their stake. These changes on the internal governance endogenize the breakdown of the investment bank-client relationship. We provide a causal econometric model to test how increasing bank complexity affects the propensity for their clients to switch underwriters in the succeeding deal. Measured by investment banks' capital, partnership size and an event study on investment banks' decision to go public, we find that the increasing complexity induces clients to switch out of their relationship bank.

In the last chapter, I study how a venture capitalist's information production induces an entrepreneur's effort. In particular, I design a contingent contracting mechanism where the principal's (a venture capitalist) private monitoring induces the agent's (an entrepreneur) effort and adds value to the project through the capital investment from the principal. Featuring double-sided moral hazard, the optimal contract subsumes a menu that entitles the principal to punish the agent upon negative information. Also, it is incentive compatible to prevent the principal from falsely punish to expropriate a bigger equity stake. Compared to the "second best" under "pay-for-performance" mechanism, this scheme grants the principal high ex ante equity stake. The project value and capital investments are commensurate with a higher marginal return on the investments. The optimal monitoring intensity increases with the value added by the agent's effort but decreases with the cost of monitoring.