Over the last twenty years, institutional investors have been allocating resources less toward traditional assets like equities and bonds, and more towards alternative investments like hedge funds, private equities and commodities. A survey conducted by the National Association of College and University Business Officers (2008) found that US university endowments larger than $1 billion allocated more than 20% of their assets to hedge funds. This increased interest in alternative assets naturally puts currencies on the radar of institutional investors.
Research has shown that currencies can be an excellent source for additional return. Various trading strategies like trend following and the carry trade have been profitable in the past (see Irvin and Park (2007) for a recent survey of technical trading and Froot and Thaler (1990) for a survey of the carry trade). Moreover, currency returns are generally uncorrelated with returns from other asset classes (see Burnside, et al. 2006). Thus, the improvement in return per unit risk on a portfolio level can be substantial.
However, the decision to allocate assets to a currency fund raises some challenging questions, such as:
- How to set an appropriate benchmark for performance?
- Are past and future performance related?
- Do managers maintain a consistent style that would allow investors to diversify across different currency funds?
These questions are difficult to address and also suffer from survivorship bias. Available data are biased toward managers who perform well and survive, while managers who perform poorly often close their funds and are unavailable for research studies. In our recent research, we have begun to tackle these issues.
In a prior study (Pojarliev and Levich, 2008a), we built on several well-known currency trading strategies and investigated four potential factors that could explain currency returns earned by professional managers. These four factors represent transparent, easily replicated trading strategies (carry, trend, and value) and the volatility of spot foreign exchange rates. We found that a significant portion of currency trading returns were related to the four specific trading styles or risk factors. These “beta returns” reflect exposure to a particular style or risk. In addition, some funds were able to generate significant additional returns (often referred to as alpha) not related to these factors.
In a recent study (Pojarliev and Levich, 2008b), we make use of a new database on daily currency fund manager returns over a three-year period, 2005-08. With data for all managers since inception, even those who eventually exit the sample, we control for survivorship and backfill bias. The new data set allows us to validate the results from our previous study, namely that the four factors explain a significant part of the returns of professional currency managers. More importantly, this higher frequency data allows us to estimate both alpha measures of performance and beta investment style factors on a yearly basis, which in turn allows us to test for persistence.
Are alpha and beta persistent?
In our research, we find no indication of performance persistence – funds that do well in one year are no more likely to do well in the following year. But we do find statistically significant evidence of style persistence; funds with returns linked to the carry, trend, value or volatility factor in one year are likely to remain so in the following year. These results have important implications for the investment management industry. Plan sponsors should be careful when selecting currency managers based only on past performance data. As in other venues of investment management, in this sample of currency managers it appears that past performance is no indication for future performance. On the other hand, style persistence indicates that choosing currency managers with different styles makes sense as they are likely to maintain their style, and thus continue to offer a diversification benefit. Plan sponsors usually seek to hire more than one currency manager representing different styles to obtain diversification benefits. From this perspective, our evidence for style persistence is welcome. However, style persistence may imply that individual managers are less able to exploit market timing ability.
Differences between survivors and dead funds
Our research also uncovered some significant differences between managers who did not survive our three-year sample and managers who were still active in April 2008. Not surprisingly, we find that surviving managers produced an alpha that was higher than exiting managers, and with a difference of about 9.5 bps per week (or nearly 5% per year), the difference is both statistically and economically significant (see Figure 1). Our analysis suggests that living mangers generally tracked our four factors more closely (with a higher R-squared) than managers that would eventually drop out. Sticking closer to the “benchmarks” has helped some managers to stay in business.
Figure 1: Cumulative performance of the “live” and “dead” portfolios
Notes: Cumulative performance on two portfolios of funds listed on the Deutsche Bank FXSelect platform. One portfolio consists only of “dead” funds, i.e. the funds which were no longer on the platform in April 2008, and a second portfolio consists of “live” funds, i.e. those that were still active as of April 2008.
Source: Deutsche Bank and authors calculations.
Components of their investment strategies, as evidenced by the sign of factor coefficients, varied between surviving and deceased managers as well. Contrary to the presumption in the market that the recent underperformance of trend-following strategies was the main reason for the lacklustre performance of currency managers over the last three years, we found that “betting for liquidation of carry strategies” caused more damage for some managers. Given the recent liquidation of carry trades (AUD/JPY declined 27% in the first 8 days of October 2008) this is quite ironic. It looks like “betting for liquidation of carry trades” is now a smart strategy, but many of the funds which applied it during our investigation period did not survive to bear the fruits. Market timing remains crucial. Additional tests support this notion – live managers owe some of their success to timing skills in the trend following strategy, whereas dead managers demonstrated negative timing with respect to volatility.
Overall, our results lend further support to the notion that style factors explain a substantial part of returns of professional currency fund managers. While the track records of individual managers – their strategies and performance, as well as their longevity – vary considerably, managers appear to have two things in common – a lack of performance persistence and a tendency for style persistence. Market timing and not deviating from well established trading strategies proved to be important factors for survival.
Burnside, C., M. Eichenbaum, I. Kleschelski, and S. Rebelo, “The Returns to Currency Speculation,” NBER working paper 12916, August 2006.
Froot, K. and R. Thaler. 1990. “Anomalies: Foreign Exchange,” Journal of Economic Perspectives, vol. 4, no. 3 (Summer): 179-92.
Irvin, S. and C. Park. 2007. "What Do We Know About the Profitability of Technical Analysis?" Journal of Economic Surveys, vol. 21, no. 4, pp. 786-826.
National Association of College and University Business Officers (2008), 2007 NACUBO Endowment Study, Washington, D.C.
Pojarliev, M. and R.M. Levich. 2008a. “Do Professional Currency Managers Beat the Benchmark?” Financial Analysts Journal, (Sep/Oct): vol. 64, no. 5, pp. 18-32.
Pojarliev, M. and R.M. Levich. 2008b. “Trades of the Living Dead: Style Differences, Style Persistence and Performance of Currency Fund Managers?” NBER working paper 14355, September 2008.