Working Papers

Arbitrage Effectiveness and Stablecoin Run

Arbitrage is one of the most critical mechanisms in well-functioning financial markets. Stablecoins, designed to maintain dollar parity through arbitrage, provide a natural laboratory to study this mechanism under stress. This paper uses extremely granular data to study Terra stablecoin’s arbitrage failure, which occurred 48 hours before the May 9, 2022, 5 PM depeg. I develop a generalized methodology applicable to all safe assets, using stablecoin pricing data, to measure arbitrage effectiveness in stablecoins. I further show that the declining collateral value and increasing marginal trading costs in Terra’s blockchain acted as frictions to arbitrage effectiveness. I use order book data to show the microstructure of the run dynamics that followed. I show that liquidity vanished first on smaller exchanges and persisted longest on Binance, the deepest market. Results are consistent with arbitrage–run tradeoff models under extreme arbitrage concentration: unlimited participation supports price correction but amplifies run risk. My results have important implications for the stability of safe assets in general.

Information Dynamics and Run on Stablecoin

Stablecoins can rapidly transition from information-insensitive to information-sensitive assets during periods of stress. During these periods, market makers demand higher compensation for quoting, while traders pay premiums for immediacy when liquidity is most critical. Using granular trade data from major crypto exchanges, I estimate hourly adverse selection for various cryptocurrencies during the May 2022 Terra collapse. I show that Terra’s hourly adverse selection costs surge to 101.9 bps from 0.34 bps during the stress period, a few hours after the arbitrage breakdown, indicating severe information asymmetry. I document flight-to-safety dynamics analogous to those observed in money market mutual fund runs. High-frequency Twitter data traces information diffusion: sentiment remains stable until May 9 morning, then turns sharply negative. Investor disagreements, as captured by the standard deviation of polarity, remain elevated.

The Internet and the Housing Market: The spillover effect of the digital divide on home buyers

I study the effect of broadband access on the price and availability of mortgage credit. The results of my study suggest that with the increase in broadband access, the interest rate decreases while the loan amount increases. I also analyze whether the broadband usage will increase the competition among the mortgage lenders, but find no support for this idea. I describe several mechanisms that could be driving the results. Overall, my results shed light on the correlation between broadband access and the housing market.