Working Papers

Asset Allocation and Returns in the Portfolios of the Wealthy (with Cynthia Balloch, LSE)

Despite accounting for a large amount of total wealth, there is little direct empirical evidence of the investment behavior of wealthy households. Based on a proprietary database of investment portfolios and returns, we document three new facts about ultra-high net worth portfolios. First, asset allocations change strongly with total wealth, as super-wealthy households hold a much larger share of alternative investments, such as private equity and hedge funds, and a lower share of liquid assets, such as public equities. The data includes a significant number of portfolios large enough to explore allocations and returns within the top percentile of the wealth distribution, including the top 0.01 percent. Second, while realized returns are increasing with wealth, Sharpe ratios are broadly similar across the top of the wealth distribution. This suggests that investment skill does not differ among investors in upper portions of the wealth distribution, but that risk tolerance increases with total wealth. Third, we use the data to explore whether returns differ within narrow asset classes, and find that returns on alternative assets, in particular, are increasing in wealth. This indicates that access and manager selection play a large part in determining returns and raises questions about the benefits of broadening access to delegated investing in private assets. Taken together, these findings substantially improve on existing empirical evidence on return heterogeneity in the U.S., which is increasingly understood to be critical in both macroeconomic dynamics and asset pricing

UIP Violations and the Cost of Capital: Firm-level Evidence

This paper establishes cross-currency differences in risk-free interest rates as a key determinant of the cost of capital of firms. I introduce a new security-level dataset of primary market prices of corporate bond issuance and find that violations of uncovered interest rate parity (UIP) directly pass through to firm borrowing costs. As a result, firms that issue debt in currencies with high risk-free interest rates face higher effective funding costs, and, consistent with this finding, firms in countries with higher interest rates have higher returns on assets (ROA). When risk-free interest rates are relatively high in the local currency, firms are more likely to issue bonds in foreign currency, and when they do so, they appear to be more insulated from the local interest rate environment. This suggests that firms use foreign currency bonds as a way to alleviate domestic financial constraints. In contrast to the role of UIP violations, differences in sovereign risk and violations of covered interest rate parity (CIP) do not exhibit a statistically significant relationship with firm borrowing costs.

International Portfolio Holdings and Return Differentials

This paper examines the relationship between international portfolio holdings and asset returns. When foreign investors own fewer assets in a particular country, currency returns, interest rates and stock returns are consistently higher in the data. This finding establishes a connection between two major puzzles in the literature, the carry trade and portfolio home bias, that have mostly been studied in isolation. Measures of capital market accessibility can jointly explain empirical variation in foreign ownership rates and return differentials. A portfolio sorting approach motivated by market accessibility gives rise to a carry trade strategy with higher Sharpe ratio and lower downside risk, both in currencies and equities. Motivated by the finding that countries with lower levels of financial integration also have a stronger link between consumption and domestic output dynamics, I develop an international asset-pricing model with agency frictions. This gives rise to a simple mechanism: When frictions constrain the amount of local assets that foreigners can hold, local investors are limited in their ability to diversify internationally. This generates cross-country variation in risk premia in line with the empirical evidence. The underlying mechanism suggests a new fundamental explanation for the existence of the carry trade, rooted in limited financial integration.

On the Edge of Doom: Optimal Bank Recapitalization with Sovereign Default

This paper addresses the optimal structure of bank recapitalization policy when sovereign debt is risky. I propose a model that combines a classic sovereign default model with private sector financial frictions, which generate fully endogenous and time-varying default costs. When the sovereign lacks commitment, I find that the impact of bank recapitalization on sovereign default risk follows a Laffer curve: Public capital infusions can decrease sovereign spreads when domestic banks are weak, even when transfers are fully financed by external borrowing. At the same time, if transfers are excessively large, recapitalization increases sovereign credit risk. Government bond holdings of domestic banks amplify the mechanism and can generate virtuous and vicious ("doom loop") cycles. This mechanism has implications for macroprudential regulation, optimal bank funding structure, and the workings of a banking union.