Increased foreign exposure and the search for new sources of alpha have heightened interest in active currency overlay and according to empirical evidence from consultant Frank Russell, the attention is sensible – active overlay can add value. In the most comprehensive review to date, Capturing alpha through active currency overlay, Russell analyses over $85bn (e93.7bn) in accounts belonging to 18 of the top managers. Critics of active overlay have used the size and liquidity of international currency markets as evidence of efficiency. Not so, says Frank Russell. Often market participants aren’t trading for profit – central banks lose money supporting depreciating currencies, for example – and opportunistic investment is available and preferable to the passive alternative.
The report builds on an endorsement of active overlay published in 1998 by Brian Strange (see IPE September 1998), now at JP Morgan in London. Using a larger sample, the report takes the debate a stage further but the conclusions mirror Strange’s findings – active currency overlay can produce positive excess returns. Frank Russell starts with the basics, measuring the results of accounts in aggregate. Initially the report analyses all available data, 241 accounts in total, irrespective of age. It then does the same for those accounts with three and five years of trailing data, as of June 1999. Of the 241 accounts, 180 or 75% have added value since inception. 81% of those over three and five years, added value. For US dollar accounts, of which there are158, 135, or 85%, added value. Those only three years old had a corresponding figure of 86%. According to the report these accounts have a better record due to a strengthening dollar over the sample period.
When the data is broken down into months, the results are less impressive. There are over 10,000 months of available data, of which 54% registered positive returns. For the equivalent dollar accounts, 56% of monthly returns were positive. Accounts posted better results in their entirety because managers had larger gains than losses on average, thereby providing positive excess returns in the long run.
The report says the results are interesting as a starting point but are unrepresentative of a client’s opportunity set. To avoid biasing the numbers by managers with numerous accounts, Frank Russell calculates an equally weighted composite for each manager. Composites are designed by sorting accounts by base currencies and benchmark hedge ratios, by weighting equally excess return for each account and by counting managers once a month. Sufficient data enabled Russell to create further composites for 0%, 50% and 100% hedged benchmarks for US dollar investors. These were formed by taking the mean of a manager’s account, then averaging them.
Twelve managers had accounts for a US dollar benchmark hedge ratio of 0% and 50%, six for a 100% dollar hedged benchmark. Of the 0% and 50% composites all 12 registered positive returns since inception. Four of the total six accounts with a 100% US$ composite registered success, whereas only one out of four with three-year trailing data had a positive return.
Bolstering these encouraging conclusions is a section dealing with the magnitude and volatility of excess returns. The report looks at average excess return, tracking error and information ratios and begins by returning to the summary statistics. Taken in aggregate, the 241 accounts produced an average excess return of 1.06% per year with an average tracking error of 2.25%. Accounts in existence as of June 1999 and with three years relevant data added 1.17% with a tracking error of 2.10%, those with five years’ returned 1.39%, tracking error was 2.56%.
Composite accounts produced higher excess returns than separate accounts albeit at a price (Table 2). Average excess return for all composites was 1.48% with a corresponding tracking error of 2.63% (opposed to 2.25% for all 241 accounts). Tracking error for those composites with three years was 2.69% as opposed to 2.10% and for those with five years, 3.04% rather than 2.56%. When the data is deconstructed one step further, into dollar composites, results are startling. Although the 0% US$ has a higher average excess return than all composites, tracking error is extremely high. The 50% US$ composite, that the report says is a proxy for a symmetrical mandate and therefore the most indicative of performance, slightly smaller excess returns than all the composites but tracking error is significantly lower.
As a final endorsement, the report compares volatility, or tracking error, of overlay managers with fixed income, US equity and non-US equity portfolios. Overlay managers using a 50% US$ hedged benchmark find tracking error falls between US fixed income and US equity and significantly below the tracking error of a non-US equity portfolio, placing it in what Frank Russell calls a lower risk category. With positive success ratios, consistent excess returns and low tracking errors, the report is a resounding endorsement of active currency overlay.
In the last quarter, the report sensibly lists a number of caveats and potential complications associated with overlay. Symmetry, or the ability to hedge both sides of the benchmark, is unique to currency overlay. Managers are allowed the flexibility to win in all currency environments, increasing potential for more consistent alpha and lower tracking error. Asymmetrical mandates prevent the manager adding value so easily in all currency environments.
Selecting the benchmark can lead to a conflict. The report cites a situation in which an unhedged benchmark may better reflect the needs of the plan long-term but the investor may prefer a more practical 50% hedged benchmark. Investors must reconcile this dilemma if the two are at odds. Hedging around the benchmark is also preferable according to the report. Unhedged US dollar composites exhibited the largest excess returns and tracking error for the overall period. Fully hedged US dollar portfolios had the lowest returns and a lower tracking error. Portfolios with 50% hedged dollar benchmarks, for example, had lower returns than the unhedged dollar composite but a far smaller tracking error, 1.8% as opposed to 2.76%. In other words, the benchmark affects return and volatility, something the investor needs to be aware of.
Investors should be aware they can manage risk by hedging between non-base currencies, by limiting concentration of currency pairs and by defining acceptable investment vehicles (futures, options etc). As with other investment strategies, the report warns previous performance is no guarantee of future returns. Although there are few overlay managers, each has a different approach and investors must scrutinise potential managers as they should other active strategies. Overlay strategies also have unique reporting procedure. Overlay managers are often the only source of information thanks to the technicalities of calculating performance and the apparent lack of custodians’ reporting expertise. In other words, the investor is unable to obtain independent, third party analysis.
Despite the reservations, the report backs currency overlay and reiterates Brian Strange’s conclusions. “Excess returns, in combination with reasonable tracking errors, support Russell’s qualitative assessment that positive excess returns from active currency overlay are likely to be sustainable and transportable across various base currencies and hedge ratios,” it says, adding the caveat: “while excess return potential is attractive, implementation and operational considerations may be formidable to some investors and may even diminish potential investment benefits. Currency overlay strategies have complexities that investors must understand and be prepared to tackle.” IPE