The launch of the Rydex CurrencyShares Euro Trust (NYSE Arca: FXE) in December 2005 opened up the currency market to mainstream investors for the first time ever. The exchange-traded trust, which provides exposure to the euro and generates interest payments based on local (EU) bank rates, was an instant hit, rapidly gathering over $1 billion in assets. Noting the demand, Rydex quickly added to the CurrencyShares family, rolling out an additional seven trusts over the next 15 months.
In May 2007, Barclays PLC joined the party, issuing three iPath currency exchange-traded notes (ETNs) designed to track the value of individual currencies.
(PowerShares and other ETF providers have also launched currency-focused products recently; those funds, however, include multi-currency and currency-strategy ETFs, and are beyond the scope of this article.)
The success of these new currency exchange-traded vehicles (ETVs) is no surprise. Prior to the CurrencyShares trusts and iPath currency ETNs, it was very difficult for individual investors to gain exposure to the foreign exchange market—ironic, considering that foreign exchange is the most liquid asset market in the world, with daily turnover of $3.2 trillion.1
Of course, it was possible for investors to tap into these markets before the launch of these new vehicles. Investors could open foreign currency bank accounts or tap in to the highly liquid currency futures markets. But each approach had its flaw. Few banks offer foreign currency accounts to individuals, and the Patriot Act requires U.S. citizens to disclose all foreign bank accounts with deposits above $10,000 to the Department of the Treasury. Meanwhile, foreign currency futures, though wildly successful with institutional investors, require margin accounts and demand that investors roll into new contracts as the near-month contract expires, something many investors are uncomfortable doing.
By contrast, these new ETVs offer clean, crisp exposure, all from the comfort of a regular equity brokerage account.
While some investors have used these funds to speculate on currency movements, the most popular uses seem to be as a hedge against existing currency exposure or as a noncorrelated asset to add to a portfolio.
Correlations Among Currencies
Individual-currency ETVs can be hedged against other individual- currency ETVs and other asset classes to help reduce market risk.
Within the currency market, correlations exist between currencies within the same geographic region and between currencies from countries that depend on similar industries. Figures 3 and 4 show the correlation matrix for the CurrencyShares ETFs and iPath currency ETNs. They highlight the high level of correlation between the currencies in Europe (FXE, FXB, FXF and FXS), as well as the strong correlation (52 percent) between the CurrencyShares Australian Dollar Trust (FXA) and the CurrencyShares British Pound Trust (FXB). The latter can be attributed to the fact that the British and Australians both annually invest $35 billion (USD) in each other’s economy.2
The Japanese yen has a low or even inverse correlation with many other currencies. This could partially be attributed to the very low 50 basis point (0.50 percent) interest rate in Japan, as well as the fact that the yen is the only Asian currency in the matrix.
Correlations With Equities
It is interesting to analyze the positive correlation relationship between certain international equity ETFs and individual-currency ETVs, as it helps us understand the degree to which ETVs can serve as a hedging tool for institutional traders exposed to currency risk as a result of owning foreign equities.
Figures 5 and 6 group the CurrencyShares and iPath Currency ETNs with an appropriate iShares international equity ETF and showcase the correlations between each. A one-factor regression was generated where the one-day return of each CurrencyShare or iPath ETN was the dependent variable, and the one-day return of the related iShares international ETF was the independent variable. The results show that all of the iShares were positively correlated to their corresponding currency vehicles, and that they were statistically significant with a 99 percent degree of confidence.
Interestingly, the results also show that the iShares equity index funds and the currency vehicles are not a perfect proxy for one another, since there is equity market risk with the index funds. The betas in Figure 5, for instance, suggest that for every 1 percent return in the iShares equity ETFs, the related CurrencyShares would return anywhere from 7 to 29 basis points (0.07 percent to 0.29 percent) in the same direction. The betas are much lower for the iPath currency ETNs, most likely due to limited data.
The in-sample single-factor regression model of the CurrencyShares Australian Dollar Trust using the iShares MSCI Australia Index Fund (Figure 7) provides a visual image showing that the foreign currency ETF is not a perfect hedge for the respective equity index, even though there is positive correlation. The relationship diverged, for instance, in August 2007, during the global credit crunch.3
Correlations With Gold
It has been over 35 years since the collapse of the Bretton Woods Agreement and the end of the gold standard, but even in a free-float currency economy, there still remains a correlation between the foreign exchange market and the gold market.
Figures 8 and 9 display the results of regressions for the one-day returns of the CurrencyShares and iPath Currency ETNs using the one-day returns of the streetTRACKS Gold Shares ETF (NYSE: GLD) as the independent variable. GLD has a statistically significant positive correlation to all the CurrencyShares except the CurrencyShares Japanese Yen Trust; however, the results also make clear that the correlations are not high enough to justify the idea that a GLD position creates exposure to currency markets.
Combining Asset Classes
Multifactor regressions provide a tool for analyzing methods of combining asset classes for hedging exposure. The multifactor regression in Figure 10 models the CurrencyShares Swedish Krona Trust against the CurrencyShares Euro Trust and the iShares MSCI Sweden Index Fund. The results show that the regression has an adjusted R-squared of 71 percent, and that both variables were statistically significant with a 99 percent degree of confidence.
Figure 11 plots the return of the in-sample model against the return of the CurrencyShares Swedish Trust, and illustrates the view that the European currency markets have been highly correlated. Investors with a bullish view on the Swedish krona should consider their current portfolio exposure to European equities and other European currencies before making a final allocation.
The second in-sample multifactor regression (shown in Figure 12) analyzes the CurrencyShares Euro Trust daily returns against the streetTRACKS Gold Shares, iShares S&P 350 Euro Index Fund (IEV) and the iShares Lehman 1-3 Year Bond Fund (SHY) daily returns. The purpose for including the iShares Lehman 1-3 Year Bond Fund was based upon the theory of interest rate parity, which states that the differential of short-term interest rates between two foreign countries should have a positive relationship on the valuation of their currency rate. The iShares S&P Lehman 1-3 Year Bond Fund is the closest ETV proxy for short-term interest rates.
As interest rates and bond prices have an inverse relationship, the regression supports the broad hypothesis and we observe that the return of the iShares Lehman Bond fund had a positive beta against the CurrencyShares Euro Trust return, holding all else equal. The three-factor model had an adjusted R-squared (or goodness of fit) of 30 percent. This relationship could weaken for many reasons, such as if the Euro Zone’s interest rate volatility increased, which is not a factor incorporated in the model.
Figure 13 plots the return of the in-sample model against the return of the Euro Currency Trust.
The results of this analysis show that single-currency exposure can be a useful hedging tool in a variety of situations. Global asset correlations are currently at their highest levels in years, putting a heightened value on unique asset classes that can be incorporated into portfolios. The launch of new ETVs like the CurrencyShares Trusts and iPath Currency ETNs has significantly opened up the foreign exchange market to all investors, who can now tap in to these assets as they build more sophisticated, institutional-caliber portfolios.
R2: R-Squared is a statistical measure of how well a regression line approximates real data points. R-Squared is a descriptive measure between zero and one, an indication of how well the independent variable is predicting the dependent variable. An R-Squared of zero is low and an R2 of one is high.
T-Statistic: After an estimation of a coefficient, the t-statistic for the coefficient is the ratio of coefficient to its standard error. That can be tested against a t distribution to determine how probable it is the true value of the coefficient is really zero.
Beta: A mathematical measure of the sensitivity of rates of return on a portfolio or given stock compared with rates of return on the market as a whole. A beta of 1.0 indicates that an asset closely follows the market. A beta greater than 1.0 indicates greater volatility than the market.
1www.bis.org 2007 Triennial Survey
2Australian Government Department of Foreign Affairs and Trade