Working Papers / Published Papers / Data 

The International CAPM Redux 
 With Francesca Brusa and Tarun Ramadorai. July 2014, pdf (appendix)

Abstract:This paper presents a threefactor international asset pricing model, using a global equity factor denominated in local currencies and two currency factors, dollar and carry, which are constructed from a large set of currency excess returns. In a comprehensive set of equities from 46 developed and emerging countries spanning value, growth, and country index returns from 1976 to the present, we find evidence that, accounting for timevariation in risk exposures, currency risk is priced in the crosssection of international equity returns. Our model outperforms standard models proposed in the international asset pricing literature.

The Term Structure of Currency Carry Trade Risk Premia 
 With Hanno Lustig and Andreas Stathopoulos. First version: May 2013; This version: May 2014, pdf

Abstract: Investors earn a large carry trade premium by taking long positions in shortterm bills issued by countries with high interest rates or flat yield curves, funded by short positions in bills issued by countries with low interest rate or steep yield curves, but the average returns to the carry trade decrease as the maturity of the foreign bonds increases.
The absence of arbitrage implies that foreign currency risk premia are high when there is less overall risk in the foreign countries' pricing kernels than at home, but the foreign bond risk premia in dollars at long maturities are higher only when there is less permanent risk in those foreign countries' pricing kernels than at home. If the permanent innovations are fully shared and the permanent risk is identical across borders, then the long bond returns in dollars are equalized (uncovered long bond return parity). Our empirical findings suggest that the bulk of risk borne by currency investors is less persistent than the overall risks borne by stock investors, because there is more crossborder sharing of permanent risks.

The Share of Systematic Risk in Bilateral Exchange Rates 
 February 2013, pdf (appendix, data)

Abstract: Two economically motivated factors account for 20% to 90% of the daily, monthly, quarterly, and annual exchange rate movements in developed countries and in emerging and developing countries with floating exchange rates. These two factors  a carry and a dollar factor  are risk factors: the former accounts for the crosssection of interest ratesorted currency returns, while the latter accounts for a novel crosssection of dollarbeta sorted currency returns. The different shares of systematic risk across currencies are related to financial and macroeconomic measures of world integration. These results have direct implications for asset managers, present a new set of precisely measured currency characteristics, and offer new insights into international macroeconomics and finance models, pointing to large shares of global shocks in the dynamics of exchange rates.

Sovereign Risk Premia 
 With Nicola Borri, March 2012 pdf (appendix, data).
 Winner of the WRDS Best Paper Award, EFM 2010.

Abstract: Emerging countries tend to default when their economic conditions worsen. If harsh economic conditions in an emerging country correspond to similar conditions for the U.S. investor, then foreign sovereign bonds are particularly risky. We explore how this mechanism impacts the data and influences a general equilibrium model of optimal borrowing and default. Empirically, the higher the correlation between past foreign bond and U.S. market returns, the higher the average sovereign excess returns. In the model, sovereign defaults and bond prices depend not only on the borrowers' economic conditions, but also on the lenders' timevarying risk aversion.


Crash Risk in Currency Markets 
 With Emmanuel Farhi, Samuel Fraiberger, Xavier Gabaix and Romain Ranciere, May 2014, pdf (appendix).

Abstract: Outofthemoney currency options point to large expected exchange rate depreciations for high interest rate currencies, suggesting that disaster risk is priced in currency markets. To study the price of disaster risk, we propose a simple structural model that includes both Gaussian and disaster risk and can be estimated even in samples that do not contain disasters. Estimating the model over the 1996 to 2013 period using
monthly exchange rate spot, forward, and option data, we obtain a realtime index of the compensation for global disaster risk exposure. We find that disaster risk accounts for almost a third of average currency carry trade excess returns in advanced countries over the period. We document that after the fall of 2008 options prices exhibit a sharp increase in tail risk, similar to the emergence of the smirk in equity options after the 1987 crash.

Published Papers

Countercyclical Currency Risk Premia 
 With Hanno Lustig and Nikolai Roussanov, Journal of Financial Economics, March 2014, Vol. 111 (3), pp. 527553, pdf.

Abstract: We describe a novel currency investment strategy, the ‘dollar carry trade,’ which delivers large excess returns, uncorrelated with the returns on wellknown carry trade strategies. Using a noarbitrage model of exchange rates we show that these excess returns compensate U.S. investors for taking on aggregate risk by shorting the dollar in bad times, when the U.S. price of risk is high. The countercyclical variation in risk premia leads to strong return predictability: the average forward discount and U.S. industrial production growth rates forecast up to 25% of the dollar return variation at the oneyear horizon. The estimated model implies that the variation in the exposure of U.S. investors to worldwide risk is the key driver of predictability.

The WealthConsumption Ratio 
 With Hanno Lustig and Stijn Van Nieuwerburgh, Review of Asset Pricing, June 2013, Vol. 3 (1), pp 3894, pdf.

Abstract: We set up an exponentially affine stochastic discount factor model for bond yields and
stock returns in order to estimate the prices of aggregate risk. We use the estimated risk prices
to compute the noarbitrage price of a claim to aggregate consumption. The pricedividend
ratio of this claim is the wealthconsumption ratio. Our estimates indicate that total wealth
is much safer than stock market wealth. The consumption risk premium is only 2.2 percent,
substantially below the equity risk premium of 6.9 percent. As a result, the average US
household has more wealth than one might think; most of it is human wealth. Nearly all of
the variation in total wealth can be traced back to changes in longterm real interest rates.
Contrary to conventional wisdom, we find that events in bond markets, not stock markets,
matter most for understanding fluctuations in total wealth.

International Risk Cycles 
 With Francois Gourio and Michael Siemer, Journal of International Economics, March 2013, Vol. 89, pp. 471484, pdf (appendix).

Abstract: This paper studies a twocountry real business cycle model with two novel features: (1) exogenous shocks to worldwide uncertainty, (2) heterogeneous exposures of countries to a world aggregate shock. When world risk increases, investment falls and a worldwide recession ensues, even in the absence of technology shocks. Equity prices and interest rates fall, especially in the more risky country, and the high interest rate currency depreciates. The model generates a low correlation between relative consumption growth and exchange rates, a large forward premium, and "excess comovement" of asset prices relative to quantities. Empirically, we measure the response of macroeconomic aggregates and exchange rates to an aggregate volatility shock, and nd support for the model mechanism.

Business Cyclical Variation in the RiskReturn Tradeoff 
 With Hanno Lustig, Journal of Monetary Economics, December 2012, Vol. 59, pp. 3549, pdf (appendix).

Abstract: We measure the variation in realized returns in bond and equity markets over the business cycle. In each of these markets, the equity, government bond and corporate bond markets, the realized Sharpe ratio is much higher in recessions than in expansions. However, the variation in realized Sharpe ratios during expansions (recessions) differs across securities markets. The realized Sharpe ratio on equity increases monotonically during recessions to reach a maximum of 1.1 at the trough of a business cycle, and it declines monotonically during expansions to reach a minimum of 0.2 at the peak. In bond markets, the realized Sharpe ratios are lowest (highest) at the trough (peak) and increase (decrease) almost monotonically during expansions (recessions), with values ranging from 0 to 0.7. These results are consistent with large increases (decreases) in actual equity risk premia and Sharpe ratios during recessions (expansions).

Common Risk Factors in Currency Markets 
 With Hanno Lustig and Nikolai Roussanov, Review of Financial Studies, November 2011, Vol. 24 (11), pp. 37313777, (pdf, data).
 Winner of the Terker Prize in Investment Research, 2009.

Abstract: We identify a `slope' factor in exchange rates. High interest rate currencies load more on this slope factor than low interest rate currencies. As a result, this factor can account for most of the crosssectional variation in average excess returns between high and low interest rate currencies. A standard, noarbitrage model of interest rates with two factors  a countryspecific factor and a global factor  can replicate these findings, provided there is sufficient heterogeneity in exposure to global or common innovations. We show that our slope factor identifies these common shocks, and we provide empirical evidence that it is related to changes in global equity market volatility. By investing in high interest rate currencies and borrowing in low interest rate currencies, US investors load up on global risk.

The CrossSection of Foreign Currency Risk Premia and US Consumption Growth Risk: A Reply 
 With Hanno Lustig. American Economic Review, December 2011, Vol. 101, pp. 34773500 (pdf).

Abstract: The consumption growth beta of an investment strategy that goes long in high interest rate currencies and short in low interest rate currencies is large and significant. The price of consumption risk is significantly different from zero, even after accounting for the sampling uncertainty introduced by the estimation of the consumption betas. The constant in the regression of average returns on consumption betas is not significant. In addition, the consumption and market betas of this investment strategy increase during recessions and times of crisis, when risk prices are high, implying that the unconditional betas understate its riskiness. We use the recent crisis as an example.

Information Shocks, Liquidity Shocks, Jumps, and Price Discovery: Evidence from the U.S. Treasury Market 
 With George J. Jiang and Ingrid Lo, Journal of Financial and Quantitative Analysis, Vol. 46, No. 2, Apr. 2011, pp. 527–551 (pdf).
 Abstract: In this paper, we identify jumps in U.S. Treasurybond (Tbond) prices and investigate the causes of such unexpected large price changes. In particular, we examine the relative importance of macroeconomic news announcements versus variation in market liquidity in explaining the observed jumps in the U.S. Treasury market. We show that while jumps occur mostly at prescheduled macroeconomic announcement times, announcement surprises have limited power in explaining bond price jumps. Our analysis further shows that pre announcement liquidity shocks, such as changes in the bidask spread and market depth, have significant predictive power for jumps. The predictive power is significant even after controlling for information shocks. Finally, we present evidence that postjump order flow is less informative relative to the case where there is no jump at announcement.

LongRun Risk, the WealthConsumption Ratio, and the Temporal Pricing of Risk 
 With Ralph Koijen, Hanno Lustig, and Stijn Van Nieuwerburgh, American Economic Review, Papers and Proceedings, May 2010, Vol. 100, No 2, pp 552556 (pdf).
 Abstract: We show that the longrun risk model, which is successful at matching the wealthconsumption ratio, high equity risk premium and the nominal yields at
short maturities implies too little (much) variation in the martingale component of the nominal (real) pricing kernel. This is because the nominal bond risk premium at infinite horizon is too high, or in other words because the real bond risk premium at infinite horizon is too low and thus the inflation risk premium too high. We conclude that the wealthconsumption ratio, the equity risk premium, and the long horizon bond risk premium impose tight restrictions on dynamic asset pricing models.
 Appendix (pdf).

A HabitBased Explanation of the Exchange Rate Risk Premium 

First Version: November 2003, This Version: October 2008, pdf.

Journal of Finance, February 2010, Vol. 65, No 1, pp 123145.

Abstract: This paper presents a model that reproduces the uncovered interest rate parity puzzle using on endogenous countercyclical risk premia and procyclical real interest rates. Investors have preferences with external habits. During bad times at home, when domestic consumption is close to the habit level, the pricing kernel is volatile, and the representative investor is very riskaverse. When the domestic investor is more riskaverse than her foreign counterpart, the exchange rate is closely tied to domestic consumption growth shocks, and the domestic investor expects a positive currency excess return. Since interest rates are low in bad times, expected currency excess returns increase with interest rate differentials.
 Appendix (pdf) : In this appendix, I derive the optimal foreign and domestic consumption
allocations in a twocountry model where agents are characterized by
habit preferences. Agents can trade, but incur proportional and
quadratic trade costs. This model replicates the empirical forward
premium and equity premium puzzle and the interest rates and
exchange rates' volatility. Finally, I investigate the role of non
tradable goods on the correlation between exchange rates and
consumption growth rates.

The CrossSection of Foreign Currency Risk Premia and US Consumption Growth Risk 
 With Hanno Lustig. American Economic Review, March 2007, vol. 97, No 1, pp 89117, pdf.

Abstract: Aggregate consumption growth risk explains why low interest rate currencies do not appreciate as much as the interest rate differential and why high interest rate currencies do not depreciate as much as the interest rate differential. We sort foreign currency returns into portfolios based on foreign interest rates, and we test the Euler equation of a domestic investor who invests in these currency portfolios. We find that domestic investors earn negative excess returns on low interest rate currency portfolios and positive excess returns on high interest rate currency portfolios. Because high interest rate currencies depreciate on average when domestic consumption growth is low and low interest rate currencies do not under the same conditions, low interest rate currencies provide domestic investors with a hedge against domestic aggregate consumption growth risk.
 Data.
 Reply to Burnside, American Economic Review, December 2011, Vol. 101, pp 34773500.

Investing in Foreign Currency is like Betting on your Intertemporal Marginal Rate of Substitution. 
 With Hanno Lustig. Journal of the European Economic Association, Papers and Proceedings, AprilMay 2006, Vol. 4, No. 23, pp 644655, pdf.

Abstract: Investors earn positive excess returns on high interest rate foreign discount bonds, because these currencies appreciate on average. Lustig and Verdelhan (2005) show that investing in high interest rate foreign discount bonds exposes them to more aggregate consumption risk, while low interest rate foreign bonds provide a hedge. This paper provides a simple model that replicates these facts. Investing in foreign currency is like betting on the difference between your own intertemporal; marginal rate of substitution (IMRS) and your neighbor's IMRS. These bets are very risky if your neighbor's IMRS is not correlated with yours, but they provide a hedge when his IMRS is highly correlated and more volatile. If the foreign neighbors that face low interest rates also have more volatile and correlated IMRS, that accounts for the spread in excess returns in the data.
 