Some investors propose hedging all currency exposure while others see no benefit to hedging at all, so is currency exposure a risk?
This is an easy question for global bond investors. Currency volatility is typically high in comparison to bond volatility, so currency hedging almost always reduces risk significantly. However, results are no longer straightforward when we look at equities. Let’s consider the historic data and what the effect of currency hedging has been in the past (the relevant data starts in the 1970s when currencies became free-floating).
From the early 1970s to 1995, currency hedging reduced risk for investors in all of the main developed market currencies (local currency/'hedged' volatility is lower than any unhedged base currency). However, this was no longer true over the past 20 years – currency hedging in the more recent period increased risks for some base currencies, and reduced it for others.
What is going on here? How is it possible that we hedge a risk and total volatility isn't going down? Well first of all, we may be hedging the wrong exposures. Many underlying corporates are now multi-nationals so hedging based on the currency of their stock market listing may be wrong (just look at the FTSE 100 with its many British multinationals but also non-domestic/foreign companies).
Currency hedging over the last 20 years increased risks for some base currencies, and reduced it for others
Second, we may overestimate the importance of the currency risk as a large part of it 'disappears' in the noisy equity returns. For example, if (hedged/local currency) equity volatility was 15% and currency volatility 8% and, assuming both risks are uncorrelated, then statistics predicts the combination of the two would have a volatility of just 17%, so three-quarters of the currency volatility is diversified away. Incidentally, the 2% increase is similar to the gap observed up to around 1995.
Third, and I think most importantly, we cannot treat currency risk as an 'independent' risk factor. If a currency is a safe haven and appreciates in a crisis then it is negatively correlated with market risk and the currency exposure may actually reduce portfolio risk. On the other end, exposure to risky currencies that tend to sell off in a crisis will add further to portfolio risk. If the market risk in currencies is a more recent feature, then this would explain the wider spread of volatilities and the apparent risk increase of hedging for some base currencies.
We cannot treat currency risk as an 'independent' risk factor
We can look at the historic market risk of currencies using our correlation galaxies where I have combined developed and emerging currencies with asset classes ranging from safe bonds to risky equities.
The data lines up along two axes: a market risk dimension running from equities towards global bonds. Separate to this, the currency dimension plays out between the two blocks: euro and US dollar (USD). Most emerging market (EM) currencies are 'risky' and are more closely linked to the USD (Polish zloty being the notable exception as it clings on to the euro).
Developed market currencies are typically 'less risky' than EM, but there are still significant differences. The commodity-linked Canadian and Australian dollar show some market risk while the yen and Swiss franc are the most defensive 'safe haven' currencies.
Based on risk considerations, an investor in a base currency that is more risky should have a lower level of hedging – as the foreign exposures tend to appreciate in a crisis and balance some of the equity sell-off. The target hedge ratio will then vary across base currencies, Japanese investors should hedge the most and EM investors the least. See Campbell/Viceira (2010) and Lustig/Verdelhan (2007) for academic research showing similar conclusions.
So why is this pattern only showing in the more recent data since 1995? I believe this is due to financial globalisation and the rapid rise in international portfolio investment flows. The risky currencies in the above analysis tend to be higher yielding ones and those that 'import' foreign capital and attract global portfolio investments in which flows have massively increased.
What we observe in markets – even in the higher frequency data – is that these global capital flows can be volatile and move in line with global risk appetite. If things go wrong then risk averse investors pull back their portfolio investments often indiscriminately, and so cause a sell-off in these risky 'investment' currencies and appreciation in risk-off 'funding' currencies; the effects of which have been examined in multiple papers, for instance Brennan/Cao (1997), Hau/Rey (2006), Brunnermeier/Nagel/Pederson (2008) and de Bock/de Carvalho Filho (2015).
Based on risk considerations, investors in more risky base currencies should have a lower level of hedging
My conclusions from the above research are that currency hedging is important for portfolio risk management. However, we need to (a) be careful what data we use for historic analysis (and I believe the more recent data from around 1995 onwards is more relevant) and (b) need to derive results specifically for the base currency we look at (we can’t apply USD-based research to GBP portfolios). Currency hedge ratios should be high for safe-haven currencies and low for risky currencies. For sterling investors, the results of currency hedging have been mixed and a balanced approach (neither fully hedged nor unhedged) seems most appropriate.
How much ‘juice’ is left in factors and what could understanding the role of luck and skill in poker have to do with this?
Physical gilts have generally offered a yield pick-up over equivalent swap-based exposure. Yet this premium has declined over the past two years as pension funds have continued to invest heavily in gilts to match liabilities. Let's consider how much this premium might have to fall for investors to swap bonds for swaps.