How do different asset classes interact with the Japanese yen-U.S. dollar exchange rate? To investigate, we consider relationships between Invesco CurrencyShares Japanese Yen (FXY) and the exchange-traded fund (ETF) asset class proxies used in “Simple Asset Class ETF Momentum Strategy” (SACEMS) at a monthly measurement frequency. Using monthly dividend-adjusted closing prices for FXY and the asset class proxies since March 2007 as available through July 2019, we find that:
The ETF asset class proxies are:
- PowerShares DB Commodity Index Tracking (DBC)
- iShares MSCI Emerging Markets Index (EEM)
- iShares JPMorgan Emerging Markets Bond Fund (EMB), available only since December 2007
- iShares MSCI EAFE Index (EFA)
- SPDR Gold Shares (GLD)
- iShares Russell 2000 Index (IWM)
- SPDR S&P 500 (SPY)
- iShares Barclays 20+ Year Treasury Bond (TLT)
- Vanguard REIT ETF (VNQ)
The following chart compares lead-lag correlations between FXY monthly return and asset class ETF monthly returns, ranging from asset class ETF returns lead FXY return by 12 months (-12) to FXY return leads asset class ETF returns by 12 months (12). Results suggest that:
- Except for some contemporaneous relationships (0), correlations across lags are modest and very noisy. In other words, relationships are unreliable.
- TLT and GLD have notable positive contemporaneous correlations with FXY return, while SPY and IWM have negative (but weaker) correlations. In other words, appreciation (depreciation) of the yen relative to the U.S. dollar tends to be good (bad) for U.S. government bonds and gold but bad (good) for U.S. stocks.
To check for non-linear effects, we look at average asset class returns by ranked fifth (quintile) of FXY returns.
The next chart summarizes average same-month returns of asset class ETFs by quintile of monthly FXY returns. There are 28-29 observations per quintile. Results indicate that the negative contemporaneous correlations above come from the quintiles of extreme monthly declines (quintile 1) and extreme monthly advances (quintile 5) in FXY.
In case there is some non-linear predictive power of FXY return, we look at next-month asset class ETF returns.
The final chart summarizes average next-month returns of asset class ETFs by quintile of monthly FXY returns, using the same scale as the prior chart. As suggested by the top chart above, results are noisy. Change in FXY has no notable power to predict next-month asset class ETF returns. There is perhaps an indication that a relatively strong yen is bad for equities and good for U.S. government bonds and gold, but correlations are weak.
In summary, evidence suggests that a strong (weak) yen relative to U.S. dollar is contemporaneously bad (good) for equities, but the opposite for U.S . government bonds and gold. There may be weak continuation of the strong yen effect for a month.
Cautions regarding findings include:
- Sample size is small, especially in terms of quintile breakdowns and variety of market conditions.
- All analyses are in-sample. An investor operating in real time could not have known the findings.
- Results may differ for other measurement frequencies.
