Tommy von Brömsen
ESSAYS IN INTERNATIONAL RISK SHARING
CURRENCY STRATEGIES AND CREDIT CYCLES
Tommy von Brömsen ESSAYS IN INTERNATIONAL RISK SHARING
ISBN 978-91-7731-109-6
DOCTORAL DISSERTATION IN FINANCE
STOCKHOLM SCHOOL OF ECONOMICS, SWEDEN 2018
Tommy von Brömsen
ESSAYS IN INTERNATIONAL RISK SHARING
CURRENCY STRATEGIES AND CREDIT CYCLES
Tommy von Brömsen ESSAYS IN INTERNATIONAL RISK SHARING
ISBN 978-91-7731-109-6
DOCTORAL DISSERTATION IN FINANCE
STOCKHOLM SCHOOL OF ECONOMICS, SWEDEN 2018
Essays in International Risk Sharing
Currency Strategies and Credit Cycles Tommy von Brömsen
Akademisk avhandling
som för avläggande av ekonomie doktorsexamen vid Handelshögskolan i Stockholm
framläggs för offentlig granskning tisdagen den 11 december 2018, kl 15.15,
Swedish House of Finance, Drottninggatan 98, Stockholm
ESSAYS IN INTERNATIONAL RISK SHARING:
CURRENCY STRATEGIES AND CREDIT CYCLES
Essays in International Risk Sharing:
Currency Strategies and Credit Cycles
Tommy von Br ¨ omsen
Dissertation for the Degree of Doctor of Philosophy, Ph.D., in Finance
Stockholm School of Economics, 2018
Essays in International Risk Sharing: Currency Strategies and Credit Cycles
SSE and Tommy von Br¨c omsen, 2018 Front cover illustration:
Pushish Images/Shutterstock.com, 2018c ISBN 978-91-7731-109-6 (printed)
ISBN 978-91-7731-110-2 (pdf)
This book was typeset by the author using LATEX.
Printed by:
BrandFactory, G¨oteborg, 2018 Keywords:
Carry trade, Exchange rates, FX, Interest rates, Predictability, Risk factors, Risk sharing, Time-varying correlation, Cash holdings, Business cycles, Macro-finance, Financial frictions, Credit cycles
iii
To Isak and Inez
Foreword
This volume is the result of a research project carried out at the Department of Finance at the Stockholm School of Economics (SSE).
This volume is submitted as a doctoral thesis at SSE. In keeping with the policies of SSE, the author has been entirely free to conduct and present his research in the manner of his choosing as an expression of his own ideas.
SSE is grateful for the financial support provided by Bankforskningsin- stitutet which has made it possible to carry out the project.
G¨oran Lindqvist Magnus Dahlquist
Director of Research Peter Wallenberg Professor of Finance Stockholm School of Economics Head of Department of Finance
Stockholm School of Economics
Acknowledgements
This doctoral thesis has been written with the support and encouragement from a large number of incredible individuals.
I am eternally grateful to my supervisor Magnus Dahlquist for his sup- port and guidance throughout these years. He has persistently motivated and inspired me to excel as a researcher. Perhaps most importantly, I have always stepped out of his office feeling less stressed than when I stepped in.
It has been an honor and a true privilege having Magnus as my supervisor.
My gratitude is extended also to other faculty members of the Depart- ment of Finance, some of whom have advised me directly on my research while others have simply been a source of inspiration. This goes to the entire faculty, but in particular Ramin Baghai, Bo Becker, Tomas Bj¨ork, Mike Burkart, Michael Halling, Jungsuk Han and Irina Zviadadze. I am also indebted to the administrative staff of the department, especially Elisa- beth Kempe, Anki Helmer, Hedvig Mattsson, Anneli Sandbladh and Jenny Wahlberg Andersson who have always provided excellent support. You are the oil in this machine called SHoF.
I am grateful to a number of former and current co-workers at the Riks- bank. I specifically want to thank Tor Jacobson and Kasper Roszbach, and everyone else at the Research Division, for convincing me that true hap- piness can only be achieved by first getting a PhD. I am also thankful to Mattias Erlandsson for making it possible for me to work part-time at the bank. I am highly grateful to Daria Finocchiaro and Karl Walentin for pro- viding me with invaluable advice on my macro-paper (the third paper in the thesis). Without their help it would have taken me at least four years to finish the paper, instead of three. Finally, my gratitude goes to Jens Iversen and the rest of the MFA team for allowing me the time necessary to finish this thesis in the best way possible.
viii ESSAYS IN INTERNATIONAL RISK SHARING
I thank my fellow PhD students for sharing the pain. These friends in- clude, but are not limited to Rafael B. de Rezende, Vasilis Dedes, Andrejs Delmans, Jieying Li, Mahamadi Ouoba, Henrik Petri, Yingjie Qi, Laszlo Sajtos, Valeri Sokolovski, Viktor Thell and Qing Xia. Special thanks go to Mats Levander and Matias Quiroz for never turning down a ping-pong match at the Riksbank.
Now to matters of true importance. I owe a tremendous gratitude to the Japanese place across the street, for providing me with a steady stream of spicy, fried chicken, albeit at a hefty cost. On countless occasions have I stood there in a thick cloud of evaporated Bibimbab, waiting for my order, when suddenly the equation that had been bugging me makes perfect sense, or a new research idea pops up. So, Arigatou!
I am grateful to my family and friends back home in Gothenburg, for supporting me from afar.
With unconditional love, I thank my most amazing wife, Irma, for her endless encouragement and support over these years. Now I can finally get a
‘real’ job. Now is also the time to reveal whether the debating-with-myself- while-hand-waving-in-the-air has been a product of the doctoral program or if it’s something of a more permanent nature. Don’t worry honey, these things are perfectly normal...
Finally, I dedicate this thesis to my children, Isak and Inez. They have constantly reminded me of the fact that there are more important things in life than pricing kernels and financial frictions. In the end, these two little rays of sunshine made all the difference. This is to you.
Stockholm, November 10, 2018 Tommy von Br¨omsen
Introduction
In order to earn excess returns, investors must take on risk. For an inter- national investor that risk can be characterized as either country-specific or global. While the former risks are contained within a given country, global risks are spread across economies and can potentially affect investors around the world. Hence, global risks are shared among international in- vestors. Understanding the origin of these risks, how they behave over time, and how they are transmitted to global asset prices (and possibly to the real economy), is important to researchers, policy makers and other market participants. Moreover, as the globalization continues to evolve, economies become more integrated. As a result, disturbances that were previously contained within a given country now have a stronger tendency to spread to other economies. Hence, there are good reasons to believe that global risks will be more dominant in the future, and therefore also more imperative in determining asset prices.
This doctoral thesis - consisting of three independent research papers - is aimed at improving our understanding of international risk sharing. In the first paper I find that global interest rates contain predictive informa- tion about the US dollar. In the second paper I show that the correlation between two risk factors in currency markets captures a time-varying risk premium. In the final paper I study firms’ cash holdings through the lens of a real business cycle model with a financial friction. This model belongs to a group of models that has received a lot of attention in the aftermath of the global financial crisis, in which disturbances in the US were spread to the rest of the global economy.
More specifically, the first paper, The Global Interest Rate Differential and the Dollar Carry Trade, studies the global interest rate differential (GID), defined as the difference between the highest and the lowest interest rate among the ten most liquid currencies. I document that changes in
x ESSAYS IN INTERNATIONAL RISK SHARING
GID contain predictive information about the excess returns of a popular currency strategy referred to as the dollar carry trade. I then present a novel trading strategy which exploits the predictability and delivers a Sharpe ratio superior to the dollar carry trade, both in- and out-of-sample. These empirical findings can be understood through the lens of a reduced-form, no-arbitrage model, where the main innovation is a systematic increase in the volatility of pricing kernels that amplifies expected currency returns.
This amplification effect - which is associated with periods of a decreasing global interest rate differential - sheds crucial light on the profitability of the dollar carry trade and, as a result, on the dynamics of the US dollar itself.
The second paper, Correlated Risk Factors in Currency Markets, stud- ies the correlation between risk factors. In a factor model, consider an asset with positive loadings on two, uncorrelated risk factors. Since the factors are uncorrelated, the asset returns are benefiting from diversifying across the factors. A change in the correlation between the factors has effects on the diversification, which directly affects the volatility of returns and - via volatility timing - future Sharpe ratios. I find strong evidence of these economic links in currency markets. First, I document that the correlation (CORR) between the dollar factor and the carry factor is highly time- varying, across almost the entire [−1, 1] interval. Second, the correlation between CORR and the difference between the volatilities of low and high interest rate currencies, σL − σH, is 87%. This is consistent with CORR affecting the benefits of diversification. Third, for high (low) interest rate currency returns, a positive CORR (above 0.25) is associated with low (high) volatility and a high (low) future Sharpe ratio. The reverse holds for negative CORR (below −0.25). These results extend to the standard, high-minus-low carry trade strategy (HML). Comparing negative to posi- tive CORR, the average next-month Sharpe ratio of the HML carry trade is 0.20 and 0.99, respectively. Fourth, I show that CORR is procyclical and positively related to both the US interest rate as well as global interest rate differentials. CORR is also linked to the time-varying characteristic of the US dollar as a safe haven. Finally, I provide evidence that CORR contains information about the predictability of carry trade crashes.
The third paper, Operational Cash Holdings in a Real Business Cycle Model with Financial Frictions, is motivated by the observation that cash
xi holdings of US firms behave in a procyclical manner. To better understand why firms tend to hold more cash in good economic times I first distinguish between operational cash (used to run the day-to-day business operations), and non-operational cash (used for precautionary motives). In a real busi- ness cycle model I find that the procyclical behavior of operational cash holdings can be explained by both a standard productivity shock and a financial shock (affecting firms’ ability to borrow). The productivity shock is the main driver of output in the model and is therefore tightly connected to the firms’ need for operational cash to finance wages and investments.
When extracting the financial shock series from the data, we see a pro- cyclical pattern. As a result, in good economic times firms tend to shift their financing from equity to debt. To facilitate this shift, firms engage in share repurchases and dividend payments, which require operational cash.
I also find that the model-implied volatility of non-operational cash hold- ings is significantly larger than that of operational cash holdings, for which I provide economic intuition.
The remainder of this doctoral thesis consists of the three papers sum- marized above.
Contents
1 The Global Interest Rate Differential and the Dollar Carry
Trade 1
1.1 Introduction . . . 2
1.2 Empirical setting . . . 8
1.2.1 Currency returns . . . 8
1.2.2 Carry trade returns . . . 9
1.2.3 The global interest rate differential (GID) . . . . 10
1.2.4 Data . . . 11
1.3 Empirical findings . . . 12
1.3.1 Predicting the dollar carry trade . . . 12
1.3.2 A superior trading strategy . . . 19
1.4 A reduced-form model . . . 24
1.4.1 Pricing kernels and expected returns . . . 24
1.4.2 The amplification effect . . . 26
1.4.3 Linking GID to the amplification effect . . . . 28
1.4.4 Relating GID to expected currency returns in general 32 1.4.5 The profitability of the dollar carry trade . . . 39
1.5 Conclusions . . . 40
1.6 Bibliography . . . 43
1.A Appendix . . . 47
1.A.1 Cleaning the data . . . 47
1.A.2 Robustness checks . . . 49
1.A.3 Derivation of the model equations . . . 52
2 Correlated Risk Factors in Currency Markets 55 2.1 Introduction . . . 56
2.2 Empirical setting . . . 62
xiii
xiv ESSAYS IN INTERNATIONAL RISK SHARING
2.2.1 The carry trade strategy . . . 62
2.2.2 The dollar factor . . . 64
2.2.3 Data . . . 64
2.2.4 Excess carry trade returns . . . 65
2.2.5 Constructing CORR . . . 67
2.3 Linking CORR to currency returns . . . 69
2.3.1 CORR and the cross-section of currency returns . . . 70
2.3.2 CORR and carry trade returns . . . 74
2.3.3 Understanding CORR . . . 75
2.4 Regression analysis . . . 80
2.4.1 The interest rate differential . . . 80
2.4.2 A simple look at the data . . . 81
2.4.3 Connections to the US business cycle . . . 82
2.4.4 Regression results . . . 85
2.5 Predicting currency crashes . . . 90
2.6 Conclusions . . . 94
2.7 Bibliography . . . 96
2.A Appendix . . . 101
2.A.1 Constructing daily HML returns . . . 101
3 Operational Cash Holdings in a Real Business Cycle Model with Financial Frictions 103 3.1 Introduction . . . 104
3.2 Literature review . . . 111
3.3 Model . . . 115
3.3.1 Firms . . . 115
3.3.2 Households . . . 122
3.3.3 Government . . . 123
3.3.4 Equilibrium . . . 124
3.4 Simulation . . . 126
3.4.1 Data . . . 127
3.4.2 Calibration . . . 129
3.4.3 Results . . . 130
3.5 Conclusions . . . 139
3.6 Bibliography . . . 141
3.A Appendix . . . 146
CONTENTS xv
3.A.1 Interest expenses . . . 146
3.A.2 First-order conditions for firms . . . 146
3.A.3 First-order conditions for households . . . 147
3.A.4 Full set of equilibrium conditions . . . 149
3.A.5 Data sources . . . 150
