Research

We study recurrent patterns in volatility and volume for major cryptocurrencies, Bitcoin and Ether, using data from two centralized exchanges (Coinbase Pro and Binance) and a decentralized exchange (Uniswap V2). We find systematic patterns in both volatility and volume across dayof-the-week, hour-of-the-day, and within the hour. These patterns have grown stronger over the years and are presumably related to algorithmic trading and funding times in futures markets. We also document that price formation mainly takes place on the centralized exchanges while price adjustments on the decentralized exchanges can be sluggish. 

We reconsiders the question of third-degree input price discrimination assuming a Mussa-Rosen style vertical differentiation between downstream firms. We show that, in contrast with the traditional analyses with homogeneous goods, input price discrimination can improve allocation efficiency of differentiated products and with this additional gain total welfare may increase even without an expansion of total quantity. Also the effect on consumer surplus is quite different from the previous result obtained for price discrimination in a final-good market. These results shed new light on public policy toward input price discrimination. 

We analyze the welfare effect of input price discrimination in a horizontally differentiated final-goods market. Previous studies assuming downstream firms competing with homogeneous goods have shown that input price discrimination lowers social welfare by reducing production efficiency. On the other hand, Hahn and Kim(2018) recently showed that input price discrimination can improve social welfare by increasing consumption and production efficiencies when the final-goods market is vertically or horizontally differentiated and consumers have heterogeneous preferences for the products. In contrast, this paper shows that even if the final goods are spatially differentiated, input price discrimination always reduces social welfare as long as all consumers participate and the participation constraint is binding in the middle of the consumer distribution in equilibrium. This result highlights that the welfare effect of input price discrimination in the differentiated final-goods market critically depends on the type of consumer participation constraint, which provides useful implications for competition policy towards the third-degree price discrimination in intermediate-goods markets. 

A Decentralized Exchange (DEX) is the most popular Decentralized Finance (DeFi) protocol. DEXs adopted a unique system called Automated Market Makers (AMMs), consisting of two types of participants: liquidity providers and traders. Liquidity providers deposit a pair of assets into liquidity pools, and AMMs define the trading rules of the liquidity pools. Therefore, the sustainability and potential growth of the system hinge on the profitability of AMMs. I estimate the expected return of liquidity providers using the realized kernel and compare the expected return of ETH/USDC liquidity pools in Uniswap, the most popular DEX. Although those liquidity pools are based on the fundamentally same algorithms, the expected return in the version 2 (V2) pool is estimated to be strictly positive at 9.15% per annum, but it is insignificantly different from zero in the version 3 (V3) pools. This difference presumably relates to the lower gas fees in V2. I also find that the expected returns in the V2 pool decreased after the launch of V3. It implies that high liquidity can undermine the profit of liquidity providers.

Private property rights play a crucial role in economic development by providing incentives for capital investment but can also distort efficient resource allocation by granting exclusive rights to owners. For example, if a property owner insists on a selling price much higher than their own utility value, intending to sell at a higher price, it can result in a loss to society as a whole. The Harberger Tax, discussed in Posner and Weyl (2017), is a mechanism where asset owners pay taxes proportional to the self-assessed value of their assets, and others can forcibly purchase ownership if they are willing to pay that price. In this case, asset owners are incentivized to report a reasonable value. If they assess the value too high, their tax payments increase, and if they assess it too low, they risk losing ownership at a lower price. Additionally, when investments increase the value of the asset, the owner can sell the asset at a higher price, preserving the incentive to invest in the asset from a long-term perspective, while also preventing the distortion of resource allocation through the abuse of exclusive ownership.

When the expected utility from an asset follows a stochastic process, the sum of expected utilities over a specific period is not necessarily equal to the asset's value. If the asset’s expected utility increases, the likelihood of others purchasing the ownership rises, while if the expected utility decreases, the owner must still pay taxes based on the previously assessed higher value. Therefore, the asset’s valuation must consider both of these effects. We propose a method to estimate the value of ownership in this case by combining Monte Carlo simulations with the reinforcement learning algorithm, DDQN (Deep Double Q-Network).

Hansen, Kim, and Wade (2024) report periodic patterns in volatility and volume observed in high-frequency data from the cryptocurrency market. Using trade data, we propose a method to develop a profitable high-frequency trading strategy utilizing periodic patterns. From January 2022 to July 2024, this strategy generate 2.5 annualized Sharpe ratio after accounting for transaction costs.