Macroeconomic considerations, like demographics and GDP growth, make emerging markets an obvious place to look for potential investments. Over the long-term, emerging markets are likely to enjoy higher demographic and economic growth than that in developed markets. However, currency risk is a key uncertainty. 

At Equity International, a crucial question we ask ourselves is, “Are we getting paid for currency risk?” Additionally, do we have the staying power to withstand currency volatility over the short term and benefit from potential long-term growth? One basic way to think about mitigating currency risk is to divide it into two categories: currency forecasting and currency hedging. 

In this short essay, we will focus on currency forecasting and will share some of Equity International’s thoughts and processes on long-term currency forecasting of the Mexican peso (and more specifically, the bilateral USDMXN), from both a qualitative and quantitative perspective. Currency hedging is a vital topic, but for purposes of this article, we will assume that an investor is unhedged. Although Equity International invests in many emerging markets, the concentration on Mexico helps center this analysis on one country in which our company has had extensive experience. 

To accurately forecast the future, it’s important to understand the past. The free-floating Mexican peso known today had its origin in the Tequila Crisis of 1994. NAFTA’s implementation and the corresponding uprising in the Chiapas region were vital contributors to the crisis, but the former fixed foreign exchange regime was in part responsible for Mexico’s economic imbalances. Mexico’s move away from its fixed exchange rate regime was not without its problems. After the peso started floating in December 1994, the currency continued to depreciate and declined by 34 percent relative to the dollar in 1995. Nevertheless, the severe currency depreciation in 1994-95 has not been repeated to the same extent in subsequent years.

Equity International considers both qualitative and quantitative factors when forecasting USDMXN. 

Qualitative
On the qualitative side of the equation, we pay close attention to elections, NAFTA renegotiations and other influences. Obviously, the election landscape has changed both in the U.S. and in Mexico since November 2016. Donald Trump’s election initially had a significant negative impact on the peso, but the currency has since mounted a strong recovery. More short-term volatility could arise if Andrés Manuel López Obrador is elected President of Mexico in July 2018 given his populist rhetoric. If this does occur, then there likely will be a selloff in the peso, but without a majority in Congress, López Obrador could find it difficult to significantly change Mexico’s economic policies. Those economic policies and economic fundamentals ultimately will drive the exchange rate in the long-run.

NAFTA renegotiations pose some risks for Mexico, but the downside scenarios look manageable. Changes to the rules of origin, a reversion to Most Favored Nation (MFN) tariffs and other adjustments would be far from optimal for Mexico, but it is important to look past the headlines when determining how these alterations could impact USDMXN. 

For example, if MFN tariff rates applied, then the simple average tariff for the United States would be about 3.5 percent, and the tariff for Mexico about 7 percent. However, if the U.S. withdraws from NAFTA, the existing NAFTA rates may be continued for one year under Section 125 of the Trade Act of 1974, which could allow time for adjustments and reversals. In addition, the U.S. would still have a number of free trade agreements with Latin American countries in proximity to Mexico. So, the economic cost may be less than that implied by the tariff rates. 

In the end, both parties have long-term economic incentives to maintain many NAFTA provisions, which may make smaller, rather than larger, policy changes more likely over time.

Quantitative
On the quantitative side, Equity International takes a forecast combination approach with respect to currency forecasting. We do not look at any one model using historical data over a specific time period, but consider many different types of models using historical data over varied horizons. This seems to help with the parameter instability that plagues so many regressions. With respect to the forecast horizon, Equity International considers predictions over several years (e.g., five years) and combines models with explicit horizons with those that identify disequilibria without a specific horizon. 

The next paragraphs highlight a sample of the simple quantitative equilibrium methods and models that we utilize. We found that simple models frequently outperform complex models. So, their empirical promise and the short nature of this essay warrants a focus on these methods.

Method 1
The first method is to observe large deviations of the natural logarithm of nominal USDMXN from its own trend. The historical time period used and the method for determining the trend both are key assumptions, but large deviations from the trend are often visible regardless of the method employed. 

For example, one would have noticed a large deviation from trend in 2008 in conjunction with the Global Financial Crisis. An investor who was long MXN after 2008 would have enjoyed a five-year period with a roughly 5% appreciation in MXN relative to USD (a good outcome since MXN should depreciate against USD on average over long periods of time). 

We are mindful of structural breaks and consider other more complex models that allow for these breaks, which coincide with political changes, sustainable shifts in inflation policy, etc. But part of the appeal of a simple model with a static trend line or a smoothed trend line is that it does not overreact to short-term changes. 

Method 2
The second method is to note large deviations from trend in a currency’s real exchange rate or real effective exchange rate (REER).

Similar to the previous method, the model identifies hypothesized disequilibria, instead of forecasting over a specific period of time. A bilateral real exchange rate adjusts for inflation differentials between two countries and the REER adjusts both for inflation and trade level among a country’s trading partners. 

Using these methods, one would have noticed a large deviation from trend in 2016 after the significant depreciation in MXN during 2014-16. So far, the appreciation of MXN in 2017 seems to lend credibility to this type of model, but more time is needed to determine whether a purchase in 2016 will be an auspicious entry point for a long-term investment. 

We are cognizant of critiques that productivity adjustments and other modifications may be appropriate for real exchange rates, and we do consider some behavioral equilibrium exchange rate models that allow for these nuances. But, productivity adjustments in their standard form do not necessarily add much to forecast accuracy.

Method 3
The final method uses cointegration to detect equilibrium levels between a currency and another variable or variables. 

Cointegration indicates that the variables in question share a stochastic trend, but a linear combination of the variables exhibits no trend. In other words, we look for a variable or set of variables that are related to MXN over a long period of time and observe whether their movements deviate from long-term patterns. 

Long-term and short-term sovereign interest rates in the U.S. represent a classic example of cointegrated variables. The slope and shape of the yield curve shifts frequently, but the level of short-term interest rates and long-term interest rates move together over long periods of time. 

With respect to Mexico, crude oil is a reasonable variable to use, although the Mexican economy is increasingly less reliant on oil. While a comparison of the natural logarithm of the real USDMXN exchange rate and the logarithm of real crude oil prices does not always show cointegration according to formal tests, the measure can be a helpful supplemental calculation. For example, this type of analysis would have suggested than MXN was undervalued both in 2008 and in 2016 as the above methods demonstrated as well.

Equity International’s processes on long-term currency forecasting are multifaceted and encompass both qualitative and quantitative considerations. In Mexico, there may be some short-term volatility with upcoming political elections and NAFTA renegotiations. But, for those investors with a sufficiently long time horizon, the country can be one of the more appealing emerging economies.