Deafening noise has surrounded the recent episode in emerging markets. Can it really be the case that the secular growth story was a myth, and the BRICs are now broken? Economies which spent the last 15 years building foreign exchange reserves to insure against balance of payments crises are nonetheless keeling over because the Federal Reserve is not even raising interest rates, but is reducing the rate at which it purchases US bonds? Is it really true that a category of “asset class” poorly described as “emerging markets” somehow renders economies as diverse as Indonesia, Turkey, Thailand and Brazil, permanently correlated?
Investors trying to make sense of this analytical chaos are likely to struggle. Fortunately, most of what we know about FX predictability renders all of the above virtually irrelevant.
The finance literature makes two empirical claims about FX returns. There appears to be a forward rate bias, interest rate differentials predict returns and not spot rate changes. This runs contrary to the original assumptions of economic theory that spot rate moves should be expected to offset differences in interest rates between countries. The second claim is that real exchange rates tend to mean revert over very long time horizons: so value has some power in predicting spot returns.
Both of these claims have been refined in more recent literature, but broadly confirmed. More sophisticated models of real exchange rate predictability suggest that the further an exchange rate is from its long run fair value the more rapid and likely it is to reverse course. This makes intuitive sense. Extremely over- or under-valued exchange rates are likely to have economic consequences that encourage self-correction. Extremes in asset pricing also tend to coincide with exaggerated optimism or pessimism.
The forward rate bias literature was originally focused on developed market exchange rates. More recent works has found similar evidence in emerging FX, and higher returns. A strategy of buying the highest yielding emerging exchange rates and shorting the lowest yielding, has delivered equity-style returns, with better sharpe ratios.
Despite this evidence, systematic allocations of capital to FX carry strategies can be very unattractive: if levels of carry are low and spot rates have been relatively stable and elevated. Arguing that one should always invest in FX carry strategies is a bit like saying that one should always be long of equities, because there is an equity risk premium. But there is a huge literature and permanent debate over what P/E multiples offer sufficient compensation, when are equities cheap, or expensive? Similarly in bond markets. But how much carry constitutes good compensation for the risk of buying Brazilian real, Turkish lira, or Indian rupee? When is carry “cheap”, or “expensive”?
The finance literature has little insight on this point. Intuition helps. Getting paid an additional 2% to own the BRL over the US dollar is unlikely to dwarf spot exchange rate moves, even over 3-5yr time horizons. The real has moved by an annual average of 15% over the last 10 years. Earning 10-15% annually in carry dramatically improves the odds.
In part, this intuition amounts to adjusting the carry for volatility. Quantifying this formally is fraught with difficulties, however. FX volatility is highly unstable and conventional risk analysis encourages myopia. Short sample periods, say one to the 3-months, may provide best estimates of where FX volatility will be next week, but tell us very little about where it will be in 18 months time. Sensible carry strategies have to take a medium-term perspective. It also makes perfect sense that opportunities will occur to earn risk premia when volatility is high. This should be exploited.
So where does this leave us after the recent EM FX episode? The economics looks terribly complicated. No one really knows what trend growth in countries like India is: Why has inflation been trending higher despite a slowdown in growth? We can make some sensible guesses. At best, this is speculation. How significant are the macroeconomic consequences of the negative terms of trade shock for Brazil and Indonesia as a result of the changing mix of Chinese growth? What will happen to credit quality in both these economies, with slower growth?
Even if we knew the answers to these questions, it is by no means clear what the currencies are likely to do. Fortunately, the investment case is far simpler than this. The annualised carry on an equally-weighted basket of three of the highest yielding EM FX rates – the Brazilian real, Indonesian rupiah and Indian rupee – is around 15%. Six months ago this was 5%. This huge increase in carry totally changes the return properties for an investor with a medium-term horizon. Compounding away with 15% interest rates over three years makes it is highly probable that returns generated by carry will dominate spot moves. Particularly if diversified over very different economies on the long and short sides.
The return prospects are further enhanced by recent collapses in spot prices. The Brazilian real fell 60% from its peak in 2011 to its recent low in August. That is an extreme shift in Brazil’s competitiveness. The same is true of India. Below is a chart of the real (inflation-adjusted) trade-weighted Indian Rupee, since 1992 (Fig 1). The post-tapering fall looks like a very plausible overshoot. To the extent that we know anything about prospective moves in spot price some reversion to fair value is most probable.
Risk and diversifcation
Finally, a few thoughts on the risk and portfolio construction implications of forward rate FX strategies. Most of the finance literature is silent on this issue. Some recent work has tried to enhance FX carry return by improving the timing of capital allocation, but this misses the point. The real issue in risk management is does the strategy offer genuine diversification and what are its return properties in the scenarios we care most about.
Carry strategies can have terrible portfolio properties. They are often positively correlated with other risk assets, such as equities. And worse still, the return distribution is skewed towards losses. In other words, when you care most you are likely to suffer the greatest losses – for example in 2008.
However, these risk characteristics are not static. If interest rate differentials are elevated in relatively low inflation environments, as is currently the case, there is a very high probability that carry will dominate spot price moves. Also, these returns should be independent of factors affecting other risk assets, say equities. This recent EM currency episode has occurred at the same time as many other risk assets have performed extremely well. The reverse could also happen. But perhaps more importantly the factors influencing, say, the profitability of Italian equities are entire different to those like to affect the interest rate component of the FX return of the Indonesian rupiah. That is genuine diversification.