Asset heterogeneity is widely believed to restrict liquidity in many markets involving important fixed-income assets. We model the impact of quasi-consolidated (QC) trading---a design that allows sellers to deliver heterogeneous assets for identical payments---on over-the-counter (OTC) markets involving assets with varying values. We show that allowing for QC trading reduces market fragmentation but introduces a cheapest-to-deliver (CTD) effect. In consequence, although QC trading increases total trading volume and social welfare, it hurts liquidity for sellers who do not switch to QC trading and lowers profits for both these sellers and some other sellers who switch to QC trading. Consolidating multiple QC contracts increases (decreases) total trading volume and social welfare if the contracts cover assets with similar (distinct) values.

Repo markets suffer from funding misallocations and funding runs. We develop a rollover risk model with collateral to show how repo trading and clearing mechanisms can resolve these inefficiencies. In over-the-counter markets, non-anonymous trading prevents asset liquidations but causes runs on low-quality borrowers. In central-counterparty markets, anonymous trading provides insurance against small funding shocks but causes inefficient asset liquidations for large funding shocks. The privately optimal market structure requires central clearing with a two-tiered guarantee fund to insure against both illiquidity and insolvency. Our findings inform the policy debate on funding crises and explain empirical patterns of collateral premia.

We examine how intermediary capitalization affects asset prices in a framework that allows for intermediary market power. We introduce a model in which capital constrained intermediaries buy or trade an asset in an imperfectly competitive market, and show that weaker capital constraints lead to both higher prices and intermediary markups. In exchange markets, this results in reduced market liquidity, while in primary markets, it leads to higher auction revenues at an implicit cost of larger price distortion. Using data from Canadian Treasury auctions, we demonstrate how our framework can quantify these effects by linking asset demand to individual intermediaries' balance sheet information.

We show in a model of over-the-counter trading that customers in equilibrium may choose to contact very few dealers to incentivize maximum liquidity provision—“less is more.” This happens when dealers’ liquidity supply is sufficiently elastic to competition. This mechanism is orthogonal to conventional concerns, such as contacting or search cost, private information, and relationship. A social planner would mandate even fewer contacts than the market outcome, where customers induce excessive dealer competition. The model predicts endogenous market power, yields implications for regulation and design of electronic platforms, and speaks to customers’ search behavior and their execution quality.

This paper rationalizes the LASSO algorithm based on uncertain fat-tail priors and max-min robust optimization. Our rationalization excludes heuristic learning or restrictive prior assumptions in the original interpretation of LASSO (Tibshirani (1996)). In our setting, economic agents (arbitrageurs) face ambiguity about fat-tail shocks and in equilibrium, they ignore a reasonable range of ambiguous signals but respond linearly to almost unambiguous signals. With this LASSO equivalent strategy, arbitrageurs can amass extra market power which induces a “cartel” to protect their aggregate profit from being competed away. This result shows a new mechanism for limited arbitrage.

Information arrives at securities markets through price quotes and trades. Informed traders impose adverse-selection costs on quote suppliers. This creates incentives for the latter to identify relatively uninformed groups and trade with them off-exchange. The marketplace turns hybrid, at the cost of thinner, highly informed (toxic) volume at the center. This pattern has largely eluded econometricians, because the standard approach to measuring information shares is biased against finding it. We show why this is the case, and design a bias-free approach. The novel approach shows that, indeed, the conjectured pattern is strongly present in the data.

Using a comprehensive multi-year U.S. dataset, we show that off-exchange (wholesaler) executions tend to benefit retail investors by separating their flow from the more toxic non-retail flow. Although the wholesale industry is concentrated, three findings suggest that wholesalers may not abuse market power. Firstly, brokers reward wholesalers who offer lower liquidity costs with more order flow. Secondly, the largest wholesalers offer the lowest costs, due to economies of scale. Finally, the entry of a new large wholesaler does not result in a reduction of liquidity costs.

We investigate differences in execution quality and payment-for-order-flow (PFOF) across asset classes. In equities, retail trades receive meaningful price improvement, particularly in tick-constrained stocks, and PFOF is small. In single-name equity options, problematic market structures lead to worse retail price improvement, and PFOF is large. While all option trades execute on-exchange, option exchange rules facilitate internalization. We exploit variation in designated-market-maker (DMM) assignments, minimum tick size, and auction allocation rules, showing that option internalization is imperfectly competitive. Option market structure gives rise to a second potential incentive conflict of brokers: encouraging customers to trade assets offering higher PFOF.

We investigate the effects of market fragmentation on price impact. Using a newly launched exchange as a quasi-natural experiment, we find that an exogenous increase in market fragmentation leads to a higher price impact of equity trading in the primary U.S. equity exchanges. Our IV estimates suggest a 1.6% increase in market fragmentation of a stock will induce approximately 2.9 bps increases in NBBO-based price impact and much more significant increases in exchange-based price impact for trading that stock. We attribute the increases in price impact to both the mechanical channel and the informational channel. Our results are providing supportive evidence to the recent theories such as Chen and Duffie (2021) that the introduction of a new lit exchange changes the landscape of trading in a multi-market setting, therefore leading to an increase in the price impact of trading.

We study the optimal execution problem in a principal-agent setting. A client (e.g., a pension fund, endowment, or other institution) contracts to purchase a large position from a dealer at a future point in time. In the interim, the dealer acquires the position from the market, choosing how to divide his trading across time. Price impact may have temporary and permanent components. There is hidden action in that the client cannot directly dictate the dealer’s trades. Rather, she chooses a contract with the goal of minimizing her expected payment, given the price process and an understanding of the dealer’s incentives. Many contracts used in practice prescribe a payment equal to some weighted average of the market prices within the execution window. We explicitly characterize the optimal such weights: they are symmetric and generally U-shaped over time. This U-shape is strengthened by permanent price impact and weakened by both temporary price impact and dealer risk aversion. In contrast, the first-best solution (which reduces to a classical optimal execution problem) is invariant to these parameters. Back-of-the- envelope calculations suggest that switching to our optimal contract could save clients billions of dollars per year.

We show that the trading-fee breakdown (fee pricing) depends on the distribution of investor gains-from-trade relative to the tick size. Absent price discreteness, an increase in investor gains-from-trade increases the total fee proportionally, but the fee breakdown has no effect. With price discreteness, the fee breakdown can mitigate the loss of welfare due to difficulty in trading when gains-from-trade are small relative to the tick size. However, when gains-from-trade are large, the exchange fee breakdown plays only a small role and exchanges extract rents from investors gains-from-trade by increasing total fee. The resulting gap between welfare relative to fees set by a Social Planner can be large. Consequently, a regulator can improve welfare substantially by imposing a cap on exchange fees.

I study a futures market model with imperfectly competitive traders, some precluded to trade spot (financial traders), some not (physical traders). I first show that, suprisingly, introducing futures makes physical traders worse off without financial traders, because physical traders seek to influence futures payoff by trading spot, and choose negative hedging ratios. Financial traders improve futures market liquidity, so that physical traders adopt positive hedging ratios when liquidity is sufficiently improved. However financial traders also raise prices when they are long, which benefits high-inventory physical traders at the expense of low-inventory physical traders. Overall, physical traders with high or very low inventory are better off with financial traders than without futures, while traders with intermediate inventory and trading in the same direction as financial traders lose. I also show that imperfect competition makes futures and spot market imperfect substitutes, implying a spot-futures basis.

We compare the market quality of centralized crypto exchanges (CEXs) such as Binance and Kraken to decentralized blockchain-based venues (DEXs) such as Uniswap v2 and v3.

After discussing the microstructure of such exchanges, we analyze two key aspects of market quality: transaction costs and deviations from the no-arbitrage condition.

We find that CEXs and DEXs operate on roughly equal footing in terms of transaction costs, particularly in light of recent innovations in DEX protocols.

Moreover, while CEXs provide superior price efficiency, DEXs eliminate custodian risk.

These complementary advantages may explain why both market structures coexist.

In contrast to centralized exchanges (CEXs) which match orders continuously following a price-time priority rule, decentralized exchanges (DEXs) process orders in discrete time and require traders to bid a blockchain priority fee to determine the execution priority of their orders. We employ a structural vector-autoregressive (structural VAR) model to provide evidence that blockchain fees attached to DEX trades reveal their private information, contributing to price discovery. We show that informed traders bid higher fees not only to avoid execution risk resulting from blockchain congestion but also to compete with each other. Using a unique dataset of Ethereum mempool orders, we further demonstrate that informed traders mostly compete on DEXs through a jump bidding strategy.

How do cryptocurrency exchanges compete with each other? We show that small and large crypto exchanges appear to be complements, rather than substitutes, as traditional oligopoly theory would predict. When large exchanges list new tokens, trade volumes on small exchanges increase, and small exchanges become more likely to list. We rationalize these facts in a model where small exchanges have captive customer bases, and rely on arbitrage trade with large exchanges for liquidity provision. Our results imply that large exchanges’ listing decisions play a systemically important “leader” role in determining trade volumes and listings on other exchanges.

Financial literacy affects wealth accumulation, and pension planning plays a key role in this relationship. In a large field experiment, we employ a digital pension aggregation tool to confront a treatment group with a simplified overview of their current pension claims across all pillars of the pension system. We combine survey and administrative bank data to measure the effects on actual saving behavior. Access to the tool decreases pension uncertainty for treated individuals. Average savings increase especially for the financially less literate. We conclude that simplification of pension information can potentially reduce disparities in pension planning and savings behavior.

We provide evidence of a positive causal link between financial knowledge acquired through business education and returns on stock investments. Using exogenous variation generated by admission thresholds to university business programs in Sweden, we document that early investments in financial sophistication causes individuals to invest significantly more in the stock market, to earn higher portfolio returns, and to end up accumulating higher levels of wealth. Investments in financial sophistication at the launch of economic life thus significantly alters the life cycle wealth profiles of individuals.

We study how bankers affect financial decisions of their social connections. Register data from Finland allow us to relate professional transitions into the banking industry to the financial decisions of the newly appointed banker’s family members. This within-individual identification strategy reveals a strong positive effect on financial market participation. The banker effect declines in social distance, emanates largely from nonparticipating individuals, and is stronger for riskier assets. Market participants appear unaffected. Additional results suggest bankers’ sales skills are instrumental for effecting change rather than their financial knowledge solely. These insights are relevant for understanding the design of policies attempting to improve financial literacy, the impact and value of financial advice, and the nature of expertise in financial markets.

How do financial frictions affect the type of human capital investments that students make in college? To study this question, I build a novel dataset covering more than 700,000 U.S. students, merging commencement records with address histories, credit bureau records, and professional resumes. I document that students trade off initial earnings against lifetime earnings when choosing college majors and that students from low-income families are more likely to choose majors associated with higher initial earnings but lower lifetime earnings. I provide causal estimates of how student debt affects this trade-off using the staggered implementation of universal no-loan policies across 22 universities from 2001 to 2019. I find that students who are required to take on more student loans to finance their education choose majors with higher initial earnings but lower lifetime earnings. Furthermore, student debt affects students differentially depending on their family backgrounds: Students from low-income families display greater sensitivity to changes in student debt. Finally, I show that differences in student debt amounts lead to different job profiles and earnings later in life. Combined, these findings highlight the role of financial frictions in human capital investments and subsequent labor market trajectories.

Using repeat-sales data on apartments in Sweden, we estimate the gender gap in housing returns. We confirm that single women’s returns gross of renovations are lower than single men’s by more than 2pp, that half of this gap is due to market timing, and that it is concentrated in short holding period. Adding administrative data on renovation expenses and traders’ background, we find that women are much less likely to undertake renovations and to specialize in real estate professional activities. Once these differences are accounted for, we do not find any gender gap in real estate returns.

How do shocks to property taxation affect lifetime wealth accumulation and investment? We examine a unique 18th-century tax reform in Holland which resulted in large and unanticipated changes in the effective tax rates on real estate wealth, plausibly exogenous to the owners and different for each property. We find the reform capitalized into house values in the short-run and had large impacts on household wealth that grew substantially over time. On average, a one percent shock increased wealth at death by 3.5 percent. We show these effects are consistent with the fact that households do not update housing consumption in response to large tax changes: large positive or negative shocks had few impact on the likelihood of selling voluntarily, even in a liquid market with low transaction taxes. Instead, changes in taxation primarily affected annual saving. The shock had a very large impact on foreclosure rates and still affected property-level vacancy and owner-occupancy rates 70 years after the reform. Our findings suggest that shocks to property taxation have large and persistent effects on household wealth and the housing stock.

In this paper, we characterise the liquidity provision and price discovery roles of dealers and HFTs in the FX spot market during the sample period between 2012 and 2015. We find that they have different responses to adverse market conditions: HFT liquidity provision is less sensitive to spikes in market-wide volatility, while dealer bank liquidity is more robust ahead of scheduled macroeconomic news announcements when adverse selection risk is high. In periods of extreme levels of volatility, such as the ‘Swiss De-peg’ event in our sample, HFTs appear to withdraw almost all liquidity while dealers remain. In normal times, we also find that HFTs contribute to market liquidity by passively trading against the pricing errors created by dealers’ aggressive trade flows. On price discovery, HFTs contribute the dominant share, mostly through their high-frequency quote updates which incorporate public information. In contrast, dealers contribute to price discovery more through trades that impound private information.

We let “Algorithmic Market-Makers” (AMs), using Q-learning algorithms, choose prices for a risky asset when their clients are privately informed about the asset payoff. We find that AMs learn to cope with adverse selection and to update their prices after observing trades, as predicted by economic theory. However, in contrast to theory, AMs charge a mark-up over the competitive price, which declines with the number of AMs. Interestingly, markups tend to decrease with AMs’ exposure to adverse selection. Accordingly, the sensitivity of quotes to trades is stronger than that predicted by theory and AMs’ quotes become less competitive over time as asymmetric information declines.

We find that clients with stronger past trading relationships with a dealer receive consistently better prices in corporate bond trading. The top 1% of relationship clients face a sizeable 67% drop in transaction costs relative to the median client - an effect which is particularly strong during the COVID-19 turmoil. We find clients' liquidity provision to be a key driver of relationship discounts: clients to whom balance-sheet constrained dealers can turn to as a source of liquidity, are rewarded with relationship discounts. Another important motive for dealers to quote better prices to relationship clients is because these clients generate the bulk of dealers' profits. Finally, we find no evidence that extraction of information from clients' order flow is related to relationship discounts.