Lustig and Verdelhan: Carry Trade

Evaluating the Carry Trade as a Trading and Investment Strategy

Hanno Lustig (UCLA) and Adrien Verdelhan (Boston University)

Introduction

Low-yield currencies consistently produce low returns and high yield currencies
consistently produce high returns for US investors. This is true for a wide range of
currencies in both developed and emerging markets, and it holds true over long periods
of time.1 As a result, the carry trade in currency markets, borrowing in low-yield
currencies to go long in high-yield currencies, has been a profitable investment strategy
over the last three and a half decades. In fact, a levered position in currency markets has
consistently outperformed the US stock market. However, carry trade investors load up
on systematic risk. Cross-country differences in interest rates convey information about
differences in the risk characteristics of these currencies, much in the same way that
differences in book-to-market ratios convey information about stocks.

Returns on the Carry Trade

We use st to denote the exchange rate in units of foreign currency per dollar2. The
monthly excess return in dollars on a long position in foreign currency ( rxt +1 ) is the one
–month forward discount ( f t ,1 − st ) less the rate of depreciation ∆st +1 of the foreign
currency (also in logs):

rxt +1 = ( f t ,1 − st ) − ∆st +1 (0.1)

1 There is one exception: currencies with very high yields in countries that experience high rates
of inflation

2 The following variables are denoted in logs: st, ft,1, ∆st+1, r*t, and rt.

1
Lustig and Verdelhan: Carry Trade

The forward discount is simply the spread between the foreign and domestic spot
interest rate: f t ,1 = r t − rt .3 Hence, a US investor who is long in foreign currency and
*

short in dollars simply earns the interest rate spread minus the rate of depreciation.

It is common practice to sort stocks into portfolios based on characteristics like size and
book-to-market as these portfolios deliver a sharper picture of the risk-return trade-off
in stock markets. We apply the same approach in currency markets. We evaluate the
historical performance of a carry-trade investment strategy in currency markets. To do
so, we construct a sample of 22 developed country currencies between 1971 and 2005i.
At the start of each month, we sort these foreign currencies into six portfolios based on
the forward discount or the interest rate difference with the US. The first portfolio has
the lowest interest rate currencies. Current examples include the Japanese Yen and
Swiss Franc. The last portfolio has the highest interest rate currencies, such as the
Australian and New Zealand dollar. This sorting averages out currency-specific
variation in foreign currency returns, and it allows us to focus on the variation in
average returns across different currencies that is driven by the interest rate gap, not by
other factors.

Low interest rate currencies -----------------------------Æ High interest rate currencies

3 For now, we abstract from sovereign default risk, but we will check our results on a shorter
sample of returns computed from forward rates that are not subject to this deficiency.

2
 Lustig and Verdelhan: Carry Trade

Log Excess Returns 1 2 3 4 5 6 HML

Average -1.06 1.44 1.07 2.47 2.42 3.29 4.35

standard deviation 9.54 9.88 9.78 8.81 8.97 8.86 6.77

SR -0.11 0.15 0.11 0.28 0.27 0.37 0.64

Log Forward Discount 1 2 3 4 5 6

Average -2.35 -0.65 0.88 2.80 5.28 17.24 19.58

standard deviation 0.66 0.66 0.60 0.68 1.07 3.17 3.02

Log Changes in Spot Rate 1 2 3 4 5 6

average -1.29 -2.09 -0.18 0.33 2.86 13.95 15.24

standard deviation 9.49 9.82 9.73 8.79 8.91 9.38 7.41

Inflation 1 2 3 4 5 6

4.62 5.01 5.56 7.30 9.57 21.88

Table 1: Developed Country Currencies. Monthly Returns (annualized, in percentage

points), computed using T-Bill interest rates. The sample is from January 1971 to
December 2005.

We compute the monthly returns for each portfolio by taking an equally weighted
average of the log returns for all of the currencies in each portfolio, and we use 3-month
T-Bill rates for all these currencies to compute the forward discount. The first panel in
Table 1 reports the average excess returns that a US investor would have earned
between 1971 and 2005 by taking long positions in these foreign currency portfolios.
The average return on the first portfolio is -1 %, while the average excess return on the
last portfolio is 3.29 %. The average excess returns increase almost monotonically from
portfolio 1 to portfolio 6. The annualized return on a carry-trade strategy that goes long
in the last portfolio and short in the first portfolio is 4.35 %, and the standard deviation
of the return on this currency high-minus-low (HML) strategy is only 6.77 %, because
the long-short strategy eliminates dollar risk.

To create a good benchmark, we lever up the currency HML strategy to match the
volatility of the returns on the S&P 500 over the same sample (15.4 %). A US investor
who invested $1 in the levered currency HML strategy in 1971 would have ended up

3
Lustig and Verdelhan: Carry Trade

with $135 in 2005, but only $ 40 in the stock market. Over this sample, the carry trade
consistently outperforms the market.

Currency HML Cumulative Returns

160

140
Levered HML
Market
120

100

80

60

40

20

0
Feb-71 Feb-76 Feb-81 Feb-86 Feb-91 Feb-96 Feb-01

Figure 1: Cumulative Returns on the levered currency HML strategy and the S&P 500.

Interest rates, Inflation and Risk

Why do high interest rate currencies earn higher returns for US investors? On average,
the low interest rate currencies do not appreciate enough to offset the lower interest rate,
while high interest rate currencies do not depreciate enough to offset the higher interest
rates. The second panel in Table 1 reports average portfolio-by-portfolio forward
discounts, while the third panel reports rates of depreciation. The log excess return is the
difference between the log forward discount and the rate of depreciation of the foreign
currency. The currencies in the first portfolio appreciate by 1.29 %, while the interest rate
gap is larger: -2.35% on average. On the other hand, the highest interest rate currencies
depreciate by less than 14%, while the average interest rate gap is 17.3%

It’s a risk premium!

Why then, are low interest rate currencies less risky than high interest rate currencies? It
turns out that currencies behave very differently in bad times when (i) the volatility in

4
Lustig and Verdelhan: Carry Trade

currency markets rises, (ii) when the volatility in stock markets rises and (iii) when the
US economy takes a turn for the worse. In all three cases, the returns on the currency
HML strategy tend to be lowii. The beta of the unlevered HML return with respect to the
volatility of currency markets (measured as the realized average standard deviation of
monthly changes in the log spot rates over the last 12 months) is 1.95. This means that a
100 basis points increase in the monthly standard deviation lowers the annualized
return by 195 basis points on average. The first portfolio’s return increases on average by
50 basis points, while the last portfolio’s return declines on average by 145 basis points.
In addition, during these episodes of high volatility, the HML return becomes highly
correlated with the US stock market return.

Rolling correlations of daily exchange rate changes with return on S&P 500

0.8

0.6

0.4

0.2
AUSD
NZD
0
CHF
07

07

07
07

07

07

07

07

07

08
07

07

07

JPY
20

20

20
20

20

20

20

20

20

20

20

20

20
-0.2
8/

7/

7/
/

/
2/

1/

1/
31

31

30

30

29

28

26
/2

/2

/2
3/

4/

5/
1/

5/

6/

7/

8/

9/

1/
10

11

12

-0.4

-0.6

-0.8

-1

Figure 2: One-Month Rolling Correlation of Daily Exchange Rate changes with S&P
500 Returns

To illustrate this point, we take a closer look at the last 12 months in currency
markets. In February and in July of 2007, US stock markets were subject to large and
sudden increases in volatility because of developments in US mortgage markets.

Figure 2 plots the one-month correlation of daily changes in the exchange rate for
two low interest rate currencies, the Japanese Yen and the Swiss franc, and two high
interest rate currencies, the NZD and the AUSD, with the returns on the S&P 500.
None of these currencies were highly correlated with stock market returns until

5
Lustig and Verdelhan: Carry Trade

February 2007, the start of the sub-prime crisis. The sharp increase in the VIX index
coincided with a sharp decrease in the correlation for the low interest rate currencies
and a sharp increase in the correlation for the high interest rate currencies. The same
scenario unfolds at the start of the summer in 2007. Figure 3 shows exactly how
much market risk carry trade investors have exposed themselves to. It plots the VIX
index against the market beta of a levered currency HML strategy in these four
currencies (long in AUD and NZD, short in Yen and Swiss Franc). The leverage is
2.2. In the first instance, the market beta of levered HML dramatically increases from
0 to 2.25 in a matter of days. At the start of the summer, the increase is from .5 to 2.

Conclusion

A carry trade investment strategy with the same volatility as the US stock market
consistently outperforms a buy-and-hold in the US stock market over the last three
and a half decades. However, a currency’s interest rate conveys information about its
riskiness, just like a book-to-market ratio does for stocks. This becomes most
apparent in bad times for US investors.

Rolling market beta of currency HML

3.5 35.00

3
30.00
Levered HML Beta

VIX
2.5
25.00

2
20.00

1.5

15.00
1

10.00
0.5

5.00
0

-0.5 0.00
07

07

07

07

07

07

07

07

07

7
00

00

00
20

20

20

20

20

20

20

20

20

/2

/2

/2
3/

3/

3/

3/

3/

3/

3/

3/

3/

/3

/3

/3
1/

2/

3/

4/

5/

6/

7/

8/

9/

10

11

12

Figure 3: Rolling market beta of currency HML with the VIX index

6
Lustig and Verdelhan: Carry Trade

i Hanno Lustig and Adrien Verdelhan, The Cross-section of Foreign Currency Risk Premia and
US Consumption Growth Risk, American Economic Review, March 2007, Pages 89-117.

ii Hanno Lustig, Nick Roussanov and Adrien Verdelhan, Common Risk Factors in Currency
Markets, Working Paper UCLA, 2007.