Risk control and plan sponsors
As investors pursue larger allocations to international assets, they are forced to consider what they must do with the ensuing currency risk. There appears to be a general consensus that:
q holding currencies passively is not likely to provide return, but instead engender volatility, and
q active currency management can exploit inefficiencies in the currency markets.
Further, it appears that one can be fairly confident that the excess return from active currency overlay is generated from skill.
At a strategic level, plan sponsors select currency hedging benchmarks to reflect optimal risk-return trade-offs to reflect either asset-liability considerations or asset-only considerations. In essence, it has been argued that as an investor increases its international allocation, they must make an optimal trade-off between the fact that passive portfolios of currencies provide diversification benefits as they are not correlated with other assets, but these allocations also provide uncompensated volatility. Hence the strategic hedging decision – whether to be unhedged, partially hedged or fully hedged – is made largely on considerations of risk, where risk is defined as the volatility of returns, funded ratios, contribution rates etc.
In most asset classes the next decision is whether to manage the investment class passively or actively. In currency overlay, quite unlike any other asset class, plan sponsors are faced with three types of managers:
q Passive managers who implement hedges equivalent to benchmark levels with no deviation.
q Risk-control and options replications managers who argue that it is difficult to create alpha trading fundamentals and technicals in currency markets and that currency returns can occasionally be substantially negative. These strategies will deviate from the benchmark to prevent large negative returns.
q Active managers who normally deviate from benchmarks and trade on fundamental and/or technical analysis of currencies.
Therefore, passive managers could potentially be exposed to large negative and positive return outcomes (the fat tail problem) and cause problems for sponsors trying to achieve objectives issues. However, no research has been conducted to date on the impact of active management for such return outcomes. In this article, we will demonstrate that active management not only produces positive alpha with a generally lower standard deviation, but also improves the predictability of returns for plan sponsors (ie, the distribution becomes more normal). The beneficial effects of currency overlay, from a risk perspective, are shown both at the level of international returns as well as overall portfolio returns.
The question that this article will explore is how plan sponsors should define risk control in currencies from an overall plan perspective, if this is their objective. Further, it demonstrates how they should monitor such programmes while achieving the highest alpha. We are restricted to data from JP Morgan Investment Management’s (JPMIM’s) existing mandates with currency overlay to demonstrate some of these results.
Any discussion of risk is complicated by the fact that there are different types of risk (eg, strategic and tactical) as well as different parameters by which they can be measured (eg, volatility, distribution of returns). In this section, we focus the discussion on how currency overlay managers can assist plan sponsors in managing the strategic risk engendered by currencies – ie, at the overall portfolio level. We do not consider issues relating to tracking error as they are covered elsewhere.
We first define three simple measures by which the risk of the currency component can be defined and monitored with respect to the impact of currencies on overall portfolio returns. They are:
q the volatility of currency returns;
q the distribution of currency returns, and
q the correlation of the excess currency returns to other asset classes.
We briefly describe how each of these needs to managed by overlay managers to control risk for a plan sponsor.
At the most basic level, unhedged and partially hedged benchmarks expose plan sponsors to currency volatility. In the case of fully hedged benchmarks, all exposures are converted back to the base currencies and hence there is no residual currency volatility. Therefore, on a strategic basis, if currency volatility (from the benchmark) is largely uncompensated on a passive basis, at a minimum, an active currency manager should lower the volatility of the currency returns relative to benchmark for unhedged and partially hedged mandates. Table 1 provides data on seven of our clients, covering different benchmark choices. For each client, we have provided mandates, benchmark return and volatility and actively managed return (gross of fees) and volatility as of June 1999.
As can be seen from Table 1, for all unhedged and partially hedged mandates (Clients 1–5), excess returns have been generated with volatility lower than that of the benchmark. This result applies across other accounts not reported here. As expected, in the case of fully hedged mandates, including Clients 7 and 6 (who initially had a fully-hedged mandate), active management implies greater volatility than the benchmark volatility. Hence under this definition of risk, even fundamental/technical managers can provide risk control, while adding alpha.
In case currency returns are not normally distributed and sponsors are averse to frequent occurrences of significant negative returns, then currency overlay management should change the distribution of returns. Essentially, the active management of currency exposure should change the distribution of returns to be either more concentrated around the 0% return or shifted to the right of the benchmark distribution so that the mean expected return is higher than that of the benchmark, while minimising the frequency of large negative return outcomes. The charts on page 70 show how this can be monitored and confirms an earlier claim that the distinction between the nomenclature applied to risk control and other active strategy is largely one of perception rather than results. We have taken the frequency distribution of returns along with the cumulative distribution of portfolios managed against different benchmarks for a set of clients highlighted in Table 1. What is obvious from these charts is that active currency management using fundamentals and technicals alters the distribution of currency returns in a fashion consistent with pure risk-control strategies. Three key results can be discerned:
q the mean is shifted to the right with active management;
q the distribution is more concentrated around the actual active portfolio mean than the benchmark; and
q the large negative returns are largely absent in the actual active portfolios. Obviously, the distribution becomes more normal and predictable for asset-liability planning.
In all cases except Client 7, the columns representing the frequency of active management returns are shorter than those for the benchmark for returns less than zero. Chart 7 which represents the returns for a fully hedged benchmark does not have this result, but produces a more normal distribution of returns vis-à-vis the benchmark with positive alpha. Once again, active management produces alpha and provides risk control.
The third way in which active management can potentially lower the overall risk of the portfolio for a given return target is if the excess returns of a currency strategy are less than perfectly or even uncorrelated with the returns of basic bond and equity investments. Earlier research has pointed out that this appears to be the case and we reproduce a table demonstrating these results. Table 2 highlights the data of specific JPMIM currency overlay clients under this definition of risk control. The lack of significant correlation of the excess returns with the S&P 500 benchmark, the Salomon Bond Index and MSCI EAFE (unhedged) indicates that currency management with positive alpha can be risk reducing at the overall portfolio level. Essentially, active management shifts the efficient frontier up so that a target return can be achieved with lower volatility.
Within the currency overlay business, a dichotomy has historically been made within the active currency overlay community among those that are focused on fundamentals and technicals and those managers who are called risk-control managers. This paper has sought to demonstrate that if plan sponsors define strategic risk control as managing the risk of the overall portfolio, then there are three ways (or some mix of the three) in which risk control can be achieved and monitored through active currency management. First, the manager can lower the absolute volatility of currency returns; second, by altering the distribution of returns to be less biased towards negative returns; and third, by the benefit provided by the lack of correlation between currency alpha and other asset class returns. However, a counter-intuitive result is demonstrated, namely that fundamental and/or technical manager (or a mixed manager) could also be providing strategic risk control to plan sponsors while targeting positive alpha. Since this is the case, the more appropriate comparison among managers (and the selection thereof) should be based on the risk-adjusted alphas that they produce over long horizons and the degree of skill inherent in such performance rather than generic titles that mask the true perspective.
Arun Muralidhar is head of currency research and Khosrow Mehrzad is vice president in charge of the direct hedge fund at JP Morgan Investment Management in London