In a detailed post last week (see post Why Currency Hedging is Necessary), reader Avon explained how foreign stocks are much riskier when one considers the effect of exchange rate fluctuations. He argued that even when exchange rate fluctuations wash out over the long term, the added volatility of these fluctuations increase the risk of holding foreign stocks. Therefore, he suggested that currency hedging is worthwhile even if the costs of hedging run close to 2% per year.

I turned to Triumph of the Optimists, an excellent book on 101 years of investment returns from sixteen countries around the world by Elroy Dimson, Paul Marsh and Mike Staunton to find out how currency fluctuations have affected investment returns in the past. The first point that should be made is that currency values have fluctuated considerably in the past. Most currencies depreciated against the U.S. dollar and the volatilities were quite large. The Italian Lira, for instance, depreciated at an annual rate of 5.7% against the U.S. dollar with a standard deviation of 17.4. As an aside, if the Canadian dollar depreciated at a similar rate, $100 will be worth just 37 cents after 101 years. Fortunately, compared to most other currencies, the Canadian dollar was very stable (depreciation of just 0.4%) and exhibited the lowest volatility — a standard deviation of just 4.3. Even in the era of floating exchange rates, the Canadian dollar exhibited low volatility: a standard deviation of just 4.7% in the 1972-2000 period.

Despite the volatility in currency rates, currency risk did not add greatly to investor risk. Dimson et al. note: “The total dollar risk was generally less than the sum of the local market and currency risks because the correlation between the two returns was such that they often offset one another”. For instance, an U.S. investor in Canadian stocks would have experienced real returns with a standard deviation of 16.8% in local currency, 4.6% in exchange rate and 18.4% in U.S. dollar terms. In some markets such as Netherlands, Denmark and Sweden, the exchange risk offset local market risk so much that dollar risk was lower than local market risk.

The authors conclude:

International investors, can, of course, choose to hedge currency risk. We have seen, however, that even for single country investments, exchange risk does not greatly increase portfolio risk, and its impact is even smaller in the context of internationally diversified portfolios. Indeed, unhedged international portfolios might even have lower long-term variability than their hedged equivalents when measured in real terms.

This article has 19 comments

  1. On your last point, you’ll find this paper interesting:

    It argues that for Canadian investors, unhedged portfolios will reduce risk because our currrency tends to perform poorly in difficult periods (look at lask week). Therefore it smooths out the negative performance of global equity portfolios.

  2. Pingback: Tweets that mention Currency Fluctuations and Stock Market Returns | Canadian Capitalist --

  3. I dropped many of my American stocks last year because I couldn’t handle the currency fluctuations. As much as I thought I had solid (US) companies, I couldn’t guarantee myself a conservative 10-15% return on them because of the dollar. Kept Apple though 🙂

    I’m sure I’ll revisit foreign investments after I gain more knowledge.

  4. I think the main point is that Canadian investors should ratchet down their USD exposure as they near retirement. Right now I am 33 and 50% of my total portfolio is in USD. I wouldn’t feel comfortable with this exposure at 65! According to my IPS, i have to ratchet it down so that I have max 15% in USD at 65.

  5. I must say I have disagreed with your previous posts on currency hedging.

    My issue has been that you looked at the decision to hedge from an individual investment point of view and strictly the cost of hedging. ALL, I repeat ALL investment decisions should be made with the entire portfolio in mind. We need to consider how each asset class we add to our portfolio adds the risk profile of our portfolio.

    In this case not only is the volatility of the currency important but the correlation with the other assets in the portfolio.

  6. Thanks for doing the reading on this one. My spider sense always makes me suspicious of consistent predictable drags on returns such as MERs and currency-hedging costs. Even if they look small, their consistently negative effect accumulates over time.

  7. Canadian Capitalist

    @John: Thanks for the paper. I haven’t read it fully yet but it seems to support the view that foreign investors should not hedge their currency exposure in US stocks due to negative correlation between US currency fluctuation and stock market returns. That’s consistent with the findings of Dimson et al.

    @Ryan Martin: Expected returns of 10% to 15% means an aggressive portfolio IMO because the earnings yield on the market is only about 6.5% or so.

    @DM: Agree. As investors get older, they should ratchet down their equity exposure, which automatically implies ratcheting down US dollar exposure. Our personal portfolios resemble the Sleepy Portfolio’s allocation. We have 22.5% exposure to US dollar and another 27.5% exposure to other foreign currencies (Euro, Yen, Pound and emerging market currencies).

    @Stephen: I plead guilty that I looked at currency mainly from a cost point of view. But cost is very important, IMO. If hedging currency costs 1.5% to 2% every year, the drag is so large that a currency has to depreciate a lot just to break even. There simply isn’t much of a track record for currency hedging, so my default position is still unhedged exposure to foreign stocks.

    I agree with your comment that it is important to look at how adding an asset class impacts the portfolio as a whole. It’s a topic covered in Triumph of the Optimists and I’ll write up a post in the future.

    @Michael: Triumph of the Optimists is just pure gold. I’m thinking that it is worth every penny of its $135 cover price especially if they put out a new edition with 10 more years worth of data.

  8. I think that investing in foreign markets is just a ruse for financial services companies to squeeze more money from retail and institutional investors. Foreign mutual funds all charge a higher MER than domestic funds. Stock pickers like me really only have access to non-North American stocks through ADRs listed in New York.

    In reality, in this new globalized world, we are exposed to the global markets anyways. Scotiabank is famously deriving a lot of their income from Latin America & the Caribbean. And who do you think is buying all of the commodities that our natural resource companies are exporting? It’s certainly not being consumed here in Canada. So, in my opinion, even if every stock we have is listed on the TSX, you’re only kidding yourself if you think you have no foreign exposure. The perfect example is these past few days where markets around the world are plummeting and recovering in lock-step, in complete unison.

    Gee, I wonder why the financial services industry isn’t pointing that these facts? Possibly because they earn much higher MERs on foreign funds, maybe because they also don’t always have to disclose what kinds of currency hedges and such are in place or maybe because they can pretty much invest in anything they want in between quarterly reports, or maybe so that they can travel to great places paid for by the money they manage? After all, how else are these money managers going to have an all-expenses paid vacation to the Bahamas every winter and still be able to call it “work”, because they have to do their “research” to find the best golf courses in which they think you should invest?

    I’ve seen how all these slick haired, smooth talking, luxury sports car driving crowd operate… ;0)

    • Canadian Capitalist

      @Phil: I agree that a number of companies listed on the TSX have global sales, not just in Canada. But the TSX is missing some major sectors such as Technology, Consumer staples, Pharmaceuticals etc. to name a few. It may well be that mutual fund companies like to tout foreign investing due to higher fees. Personally, I just avoid that route preferring to get foreign stocks through broad-market, ultra-cheap ETFs.

  9. @CC and @Phil S: My approach is somewhere in between. I own US and Canadian stock ETFs to get braod coverage of sectors and trust that the international operations of these companies give me sufficient international exposure.

  10. I do not hedge for two reasons:
    1) I am yet to see any evidence that it lowers the overall returns over the long run and I am 30 years from retirement.

    2) To protect myself from a complete demise of the Canadian dollar due to hyper-inflation or another external event in the Canadian economy. By keeping my currency exposure to both the US and the Eurozone I at least have some protection (for some of my assets) in the worlds two biggest economies plus to a lesser extend other developed/emerging markets. Although this seems like a long shot, my investment horizon is 30 years+ so who knows what happens and besides it helps me sleep at night.

  11. Avon Barksdale

    As I pointed out in my previous post, the currency hedge issue for DIY middle class investors is not trivial or obvious. To simply state that buying the lower fee US equivalents on the NYSE is the obvious right thing to do oversimplifies a complicated optimization problem.

    First, you are paying extra fees to buy the US securities through FX bid-ask spreads and possibly additional ECN fees. Some discount brokerages offer competitive spreads but you should be thinking that it will cost you anywhere between 1% to 3% every time you buy US currency. That kind of hurdle is large for people who do not have a lot of assets saved up yet.

    Anyway, despite the FX fees, you are exposed to additional volatility with the US dollar. You can do the calculations yourself to see it, or you can look at S&P’s fact sheets regarding their currency hedged indexes. I want to stress that you will see a gain or a loss with certainty from currency fluctuations (in mathematical parlance, the probability that you see no gain or loss is associated with a set of measure zero – mathematicians would indicate this by saying that you will see a gain or loss almost surely). So, to understand what the risks are, we need to understand how big the spread is in that gain or loss. This is why risk management places such importance on the variance of the portfolio and not just expected return. I have demonstrated in my previous post that the standard deviation for the unhedged portfolio is almost 2 times larger (if we assume no long term correlation between stock returns and the CAD-UAD exchange rate, which is close enough for demonstration purposes). Simply put, the poor returns in Canadian dollars of an unhedged portfoio over the last ten years is not a one-off thing – it can be quite typical.

    To give you an idea why variance matters, supposed you were offered a lottery which has two outcomes based on the flip of fair coin: Heads- $1 million, Tails- you lose your house and all your possessions. Now, for most young DIY investors this lottery has a positive expectation, but no rightly sane person would take the lottery. In fact, if the lottery was forced on them, most people would try to buy their way out of it. This is why rational risk averse investors care very much about the variance of their portfolios. If you understand the downside risk well, the upside will take care of itself.

    Now let’s return to the currency hedging question. You have a lower expectation with the hedged ETFs but you have a lower variance as well. In an efficient market, the level of the reduction in the expectation will be matched to the reduction of the variance in your portfolio. You will get the fair price – the market price of risk. Each investor here will have to determine her own price for risk on the margin (i.e. her reservation price for risk reduction). If the hedged ETFs are too much for you, don’t buy them.

    You must remember that the currency risk has no expectation of reward. People invest in the S&P500 because the stock market drifts up (a submartingale process), but exchange rates are nearly perfect martingales (no drift up or down over the long run). This is why no one just uses foreign currency as a buy-and-hold investment -it’s just pure volatility. This is also the reason why the FX derivatives market is so huge.

    And while everyone here seems to worry about the implosion of the Canadian dollar, there are real economic risks with the US dollar. One day the US is going to have to get serious about taxes. If Americans do not find a way to increase federal revenue, they will pay one way or another. It is possible that the US dollar could enter a prolonged weakening against other G7 currencies – especially if energy prices rise. I personally don’t have a view which way the dollar will go, but the point is, there is risk with a completely unhedged strategy if you have the majority of your financial obligations in Canada on retirement.

    The currency hedging issue is not trivial or a no-brainer and investors here would be wise to consider and understand all risks in their investment choices.

  12. Canadian Capitalist

    @Avon: I looked up the market history posted here because my sense was that your model overstates the risk of CAD-USD fluctuations (at least in the past). Dimson et. al found that the annual SD of CAD-USD fluctuation was 4.2 (not 15 as assumed in your post) and it is not correct to assume that currency fluctuations are independent of stock market returns. In fact, Dimson found negative correlation between stock market returns in local currency and USD-foreign currency changes.

    The past century was really a US century. The USD was strong — only the Swiss Franc appreciated against it over the entire century. Even then for many markets, an US investor experienced less volatility than a local investor due to negative correlation. In fact, in all markets, the SD of USD returns was much less than the SD of local currency and SD of local markets.

    The point is even if you ignore costs on all sides, you can make a very strong case that having an unhedged exposure to foreign markets is a very good thing for a portfolio. The costs involved in hedging (which I think we can agree will run at least 0.5% per year) simply makes the odds of outperformance in favour of unhedged holding even better.

  13. Avon Barksdale

    Thanks for the post CC.

    First the last century of currency data is not a good measure going forward. The Bretton Woods system was in place from the end of WWII until 1971 when inflationary pressures in the US economy forced Nixon to abandon the gold standard. Thus, we should only look at currency volatility from the mid 70s and probably the early 80s onward – i.e. once we entered a free floating currency regime. The annual volatility of the CAD-USD typically sits in the10+% per annum range. I took 15% for round figures. My results don’t change much for a volatility of about 12% – which is quite typical. I encourage you to verify this all yourself. Annual volatility of 4.3% is very low.. To get an estimate of the annual volatility, use the daily fluctuations and then scale up. Don’t use annual returns for the last 30 years – 30 data points is no where near enough data to make a conclusion on the volatility risk. Right now, implied volatility on European options (nothing to do with Europe, but having to do with exercise privileges) on the USD out to December 2010 all sit in the 11% to 13% range. Again, this is pretty typical. BTW, if you believe that annual volatility is 4.3%, you are saying that if the dollar is at parity at the beginning of the year, there is only about one chance in 50 that it would trade at or below 91 cents at the end of the year. I think the CAD-USD is a tad more volatile than that (and so would most option traders!).

    About the only time there is a strong correlation between the S&P500 and the CAD-USD exchange rate is during a flight to quality event. In this case the S&P500 declines sharply and investors flock to US treasury instruments as a safe haven. This drives the USD up against all world currencies. But this is bad news for unhedged investors. After a flight to quality event, your portfolio will become unbalanced – your stocks declined and you are now overweight in bonds. This is the time that you need to rebalance by putting more money into stocks during your next investment installment (or from selling some of your bonds). However, once the flight to quality event passes, the USD starts to weaken against world currencies which erodes the unhedged portfolio’s performance. Rebalancing causes you to buy high (due to foreign currency effects) – having precisely the opposite effect rebalancing is supposes to provide. This is EXACTLY what happened in 2008. All unhedged investors who dumped money into the market in late 2008 and early 2009 have seen most of their gains wiped out by the appreciation of the Canadian dollar. This is where correlation just kills you – you have great difficulty taking advantage of a sharp decline in the S&P500. Just one flight to quality event in a decade can more than pay for the hedged portfolio.

    Again, the hedging issue is not trivial or obvious. In fact, I do not advocate either hedged or unhedged, just that people need to understand that there is no free lunch – don’t expect radical outperformance from the unhedged portfolio on a risk adjusted basis. I encourage investors to read as much as possible so that they understand all risks – whatever strategy they use.

  14. To me expected return is the least important reason not to hedge. For me personally there are two very good reasons NOT to hedge that I don’t think have been mentioned.

    I don’t know where I’m going to live in 10, 20 or 30 years. I don’t know where I’ll work, and I don’t know where I’ll retire. Probably here in Canada, but I’m not sure. Could be the US. Could be Australia. Could be France. Who knows? So I don’t want to tie my investing fortunes so closely to my home country.


    Investing overseas is insurance against unexpected events here at home. We could have a terrible government that turns Canada into Greece some time in the next few decades. Quebec could split and Canada could do a bad job of managing it and experience a financial and currency crisis. Someone could discover cheap fusion and make the oil sands into a valueless albatross overnight. Not hedging broadens my exposure and protects me from that.

    I’ll still keep most of my assets in CAD, but part of the reason for owning overseas assets is to be exposed to the overseas currency.

  15. Pingback: Looking To Make Some Extra Cash? Forex Trading May Be For You! | IPAD AND IPAQ

  16. I believed the correlation between the USD and the CAD is quite close, ie their effects will not offset against one another. I would prefer to have investments in an emerging market currency, as a better insurance against a failure in CAD, or in Canada. After all, emerging markets are the future, as they all say, and developed economies are at the most, stagnant. So, why not emerging market currencies ?

    • @Rox: I don’t buy the argument that foreign currencies are an asset class. In any case, Canadian investors owning currency unhedged foreign equities will already have plenty of foreign currency exposure.

      I haven’t run the numbers but it is hard to see how CAD and USD are correlated. When the CAD is strong, USD is weak and vice-versa. That’s how it has been as long as I can remember.

  17. CC, when I said emerging market currency, I do not mean purely the currency itself, as in forex per se, I meant as a whole, the many asset classes which are denominated and valued in a currency used by an emerging market. And yes, foreign equities would be an investment in a foreign currency too.

    Anyway, since we are on this subject, unlike yourself, I regard holding foreign currencies in Term Deposit Accounts as an asset class too, especially when they are in high yield currencies such as the AUD.

    Well, I as for the pair, I don’t look at things as the USD-CAD pair, I look at things as in CAD-SGD, but I guessed I am more biased in this sense because of my natural need for the SGD.