When FX forward positions express views on country-specific developments one can shape the trade to its rationale by hedging against significant unrelated global influences. Almost all major exchange rates are sensitive to directional global market moves and USD-based exchange rates are typically also exposed to EURUSD changes. A simple empirical analysis for 29 currencies for 1999-2017 suggests that the largest part of these influences has been predictable out-of-sample and hence “hedgeable”. Even volatility-adjusted relative positions across EM or FX carry currencies may sometimes be hedged against market directional influences.

This is an original SRSV post.
It ties in with this sites’ summary on systematic value through macro trends, particularly the section on how to align macroeconomic trends with market positions

Hedging FX forwards: purpose and a simple method

If FX forward positions versus their natural funding currency (EUR or USD) are meant to express a view on developments that are specific to a single currency or a group of currencies they should be hedged against the estimated influence of unrelated global factors, provided that this influence is significant and predictable. If FX positions are meant to express currency-specific developments relative to one or more similar currencies global influences are naturally weaker, but their strength and case for hedging should still be considered.

In the below sections hedges against global directional risk and EURUSD exchange rate changes are considered. Hedge ratios are re-estimated once per month at the beginning of a month and applied to trading positions for the subsequent month. Hedge ratios are estimated as beta (sensitivity) to the relevant benchmark using weighted least-squares regressions based on rolling windows of daily and weekly data. In these regressions past observations carry exponentially decreasing weights backward in time. While daily data provide more observations, they also often understate the correlation of FX forwards due to time zone effects, i.e. incomplete overlap of trading hours during the day. Therefore, we use a weighted average of weekly data estimate (26 weeks half-life of lookback window) and a daily data estimate (63 days half-life of lookback window). This has been set a-priori based on expert judgment without any attempt of statistical optimization.

The impact of hedging is then studied in three steps:

  • First, we check the rolling betas for stability and theoretical plausibility. Stable betas should mostly have the same sign and limited fluctuations. As to plausibility, estimated betas of high-carry and EM currencies to the global directional risk basket should positive. Likewise, the estimated betas” of currencies that trade against the USD should mostly have positive beta with respect to EURUSD changes.
  • Second, we assess how successful rolling hedges have been in removing global influences. This success depends on their ability to predict their betas out of sample. As basis for testing the out-of-sample efficiency of hedging we look at the residual betas, after hedging. We calculate them on a 1-year rolling basis (using daily and weekly frequency), which seems to be the shortest meaningful estimation period. Successful hedging would require that the absolute size of betas of the hedged positions declines visibly and significantly. Unsuccessful adjustment would result in underhedging or overhedging and no meaningful reduction in absolute betas.
  • Third, in some cases we check how the rolling hedges have affected the long-term performance and correlations of FX forwards to get an idea how positions have changed their “character”.

For an explanation of the data, FX forward returns and currency acronyms see annex at the bottom of this post.

Directional risk hedging

Most obviously FX positions may need to be hedged against broad global directional market risk. This market risk is approximated by a cross-asset basket, consisting of 1/3 global equity index futures, 1/3 credit default swaps and 1/3 carry and EM currencies. Hedge ratios are estimated as betas of 1-month FX forward returns with respect to returns on that basket.

Estimations look broadly sensible and stable. Within developed markets the “risk betas” of AUD, CAD, NOK, NZD and SEK have consistently been positive. Also, the estimated betas of EM currencies with respect to the global directional risk basket have almost always been positive.The estimated hedge ratio of CHF has consistently been negative. EUR, JPY and EUR display no such stable relations.

There has been a robustly positive link between FX “risk beta” and FX carry. In each year since 2000 the currencies with higher carry have displayed greater sensitivity to global directional market risk.

Rolling monthly hedging drastically reduces subsequent sensitivity of FX positions to directional market moves. The mean absolute beta with respect to the global directional risk basket drops by three quarters, from 0.8 to 0.2 after hedging The below panel shows almost all the blue lines (rolling annual directional beta after hedging) closer to the zero line than the black lines (directional beta before hedging). This holds true also for CHF, which had consistently a negative beta, and EUR, GBP and JPY, which posted changing betas. Based on these charts there is no sign of systematic overhedging or underhedging. However, hedge ratios based on regression estimates can fail to capture sudden increases or declines in beta, such as in the case of Turkey and South Africa in the 2000s and Malaysia and Russia in the 2010s.

Hedging directional risk of currencies has a profound impact on their long-term performance. In developed markets only the NZD and, by a small measure, the CHF posted positive long-term returns. Other apparently “profitable” long-term currency positions, such as AUD, CAD and NOK, pay back all their profit or more in form of hedging costs. SEK and GBP longs incurred sizable long-term drawdowns. Also in the EMFX space hedging would have cut deeply into long-term returns, with only half of the currencies left in positive territory and mostly negative performances since 2010.

EURUSD hedging

FX positions of smaller countries’ currencies can also be adjusted for exchange rate shifts between the largest ones. A particularly strong case can be made for hedging against EURUSD changes, which confound the idiosyncratic returns on currency positions in smaller countries in dependence upon their natural trading benchmark. FX forwards of currencies that trade against USD should plausibly be correlated positively with EURUSD and the more so the stronger their economic ties with the euro area and the greater their share of USD debt.

Empirically, the USD-based FX forward returns almost always displayed positive euro betas, as should have been expected.

By contrast, currencies that trade against the euro or a basket show small size and changing signs of their beta. Indeed, since most EUR-based currencies are closely integrated with the euro area there is no strong fundamental case for hedging them against EURUSD risk.

Altogether, regression-based rolling monthly hedging reduces the mean absolute empirical beta to EURUSD by 60% from 0.25 to 0.1. The below panel shows that for USD-based currencies the hedged positions (blue lines) display considerably less annualized rolling euro sensitivity than the unhedged positions (black lines).

However, for euro- or basket-based currencies hedging has not delivered a consistent improvement. The exception has been the central European currencies in the early 2000s, as these were often traded against USD despite the economic dominance of the euro: as a consequence EURUSD beta was negative and there was a legitimate case for hedging.

Double hedging

Since the influence of global directional market moves and EURUSD moves are conceptually different, FX forward positions may best be “double hedged”. Here we present the empirical properties of two-stage hedging, where the first stage hedges against directional risk and the second against EURUSD risk.

The empirical findings on the primary betas for directional hedging have already been discussed above. Estimated incremental euro betas for almost all USD-based currencies have been positive almost all of the time. However, they have mostly become smaller compared to their size before the positions had been hedged for the global directional risk basket. In particular, the high euro beta on non-risk-adjusted returns after the great financial crisis seems to reflect an “omitted variable bias” resulting from the positive correlation of EURUSD with global risk during periods of USD funding pressure. This suggests that EURUSD is better suited as a secondary hedge.

For the euro- and basket-based based currencies the incremental EURUSD sensitivity was generally lower than the outright EURUSD sensitivity, particularly after the great financial crisis. The only exception was CHF. These changes are likewise explained by the de-facto removal of the accidental risk correlation of EURUSD from the estimated hedge ratio.

Regression-based rolling hedging against EURUSD reduced the mean absolute beta of positions roughly by about 63%, from 0.19 to 0.07. In the below panel for USD-based currencies’ euro betas after euro hedging (blue lines) were clearly shifted closer to the zero line when compared to the pre-hedging betas (black lines). Interestingly, even for euro-based currencies there has been a tendency of smaller and more stationary betas.

Correlations across global FX forwards naturally decrease markedly as a consequence of double hedging.

Hedging relative positions

FX forwards of EM or carry currencies form more homogeneous groups of risk than global currencies in general. Normalized relative positions adjust for recent differences in return variability and then trade one currency against the basket. They are a sensible way to express views on relative developments within the group. It is plausible that all currencies of these groups are positively affected by directional global market moves and that the main differences of market effects are captured by the differences in FX return variances. To check this the below sections investigate if normalized relative positions still have had predictable and stable residual beta with respect to global directional market moves.

Indeed, the estimated incremental risk betas of normalized relative positions (blue lines) are much smaller than those for outright positions (black lines) and not very stable across time. There is a tendency for some low-vol currencies (CZK, PEN and RON) to record negative incremental beta and for some high-vol and classic EM carry currencies (BRL, MXN, ZAR) to record positive incremental beta. This may indicate that directional volatility plays a greater role in liquid carry currencies when compared to idiosyncratic volatility.

Despite unstable and small betas, there has been some modest benefit of hedging relative normalized positions. The overall reduction in the mean asolute beta has been 45% from 0.46 to 0.26. A modest improvement can be seen in the below line charts of pre- and post-hedging directional betas (black and blue lines respectively). On balance, the risk betas of hedged positions have been closer to the zero line in about half of the countries. Also in the cases of CZK, PEN and RON a persistent positive beta of the relative position has been removed.

Relative and hedged long-term performances of EM currencies (blue lines below) naturally tell very different tales compared to their outright performances. Some currencies with low volatility and low beta, such as PEN, PHP and RON posted superb long-term returns, whereas some popular carry currencies, such as MXN and ZAR posted negative performance. The worst relative risk-adjusted position since 2000 would have been a long in TWD, which would have lost more than 150% of its notional contract value.

Annex: Data

The empirical analysis of forward returns is based on a panel that comprises 9 developed market currencies and 20 emerging market currencies from 1999 to 2017. The starting date has been chosen in accordance with the replacement of most European currencies with the euro and the expansion of floating exchange rate regimes in emerging markets. To be eligible for this analysis currencies had to be largely convertible, floating and sufficiently liquid, at least for a part of the sample period.

  • The developed markets group includes the Australian dollar (AUD), Canadian dollar (CAD), Swiss franc (CHF), euro (EUR), British pound (GBP), Japanese yen (JPY), Norwegian krone (NOK), New Zealand dollar (NZD) and the Swedish krona (SEK).
  • The emerging markets group includes the Brazilian real (BRL), Chilean peso (CLP), Colombian peso (COP), Czech koruna (CZK), Hungarian forint (HUF), Indonesian rupiah (IDR), Israeli shekel (ILS), Indian rupee (INR), Korean won (KRW), Mexican peso (MXN), Malaysian ringgit (MYR), Peruvian sol (PEN), Philippine peso (PHP), Polish zloty (PLN), Romanian leu (RON), Russian ruble (RUB), Thai baht (THB), Turkish lira (TRY), Taiwanese dollar (TWD) and the South African rand (ZAR).
  • When currencies are temporarily pegged or their markets disrupted, particularly due to convertibility restrictions, they are excluded from the analysis for the length of the disruption period.

Generic FX forward returns are calculated based on regular forward contracts or, where appropriate, non-deliverable forwards with initial maturity of just over one month. The contracts are “rolled” at the beginning of each month of the approximate initial maturity. The contracts are based on the exchange rate of the specified local currency versus the U.S. dollar for most cases, except for some European currencies that naturally trade against the euro (i.e. CHF, HUF, NOK, PLN, RON, SEK) and some currencies of the European periphery that are benchmarked against an equal basket of U.S. dollar and euro (i.e. GBP, RUB, TRY).