The NoLoad FundX Mutual Fund Momentum Approach
Posted in Momentum Investing, Mutual/Hedge Funds
February 10, 2009
A reader offered:
“I suggest you test the performance of NoLoad FundX, a newsletter often cited as a violation of the Efficient Market Hypothesis. They offer decent data for an analysis. You can also use their mutual fund, FundX Upgrader (FUNDX), for 2008 performance.”
The NoLoad FundX approach is essentially momentum-based (with funds rather than stocks), as follows: “Upgrading involves measuring near-term performance of mutual funds (twelve months and less) and comparing them to returns of other funds with similar risk. We invest in funds with the best recent returns, and hold them as long as they continue to outperform. When a fund drops in our ranks, we ‘Upgrade’ to the new market leaders.” Using results of relevant research, the performance data presented by NoLoad FundX and actual performance data for the FUNDX mutual fund, we conclude that:
There is evidence to support a belief in the efficacy of applying a medium-term momentum strategy to funds. In aggregate, this evidence suggests that exploiting fund-based momentum may depend critically on: (1) limiting the universe of funds considered to those focused on specific asset classes and styles; (2) concentrating investment in a very few of the best-performing of these focused funds at any one time; and, (3) fairly frequent reallocation as old winners fade and new ones emerge.
There is also research that measures mutual fund performance persistence for fixed holding periods based on past performance. This research, somewhat less connected to the NoLoad FundX approach than that listed above, generally concludes that mutual fund performance persistence over periods of one to a few years exists more reliably for past losers than past winners.
The performance results presented at “Performance of Upgrading Monthly to the Top Five Funds” indicates a fund momentum effect, outperforming the S&P 500 index in 20 of 28 years (underperforming during five-year and three-year losing streaks). However, these results may not realistically incorporate all the frictions of actual fund-switching, whether for traditional mutual funds or for exchange-traded funds (ETF). In fact, NoLoad FundX amends these results with the statement: “Although fund-imposed redemption fees are factored into the results, brokerage fees, taxes and any outside management fees…are not. If applicable, these additional costs would have a negative impact on one’s actual returns.”
For a current view of the performance potential of the NoLoad FundX strategy, with trading frictions, we compare the total return performance of the FUNDX mutual fund over its entire available history (since November 2001) with those of five ETFs:
- DIAMONDS Trust, Series 1 (DIA) – very large capitalization U.S. equities.
- iShares Russell Midcap Index (IWR) – mid-capitalization U.S. equities
- iShares Russell 2000 Index (IWM) – small capitalization U.S. equities.
- iShares Barclays 7-10 Year Treasury (IEF) – mid-term U.S. Treasury notes (available only since July 2002).
- SPDR Gold Shares (GLD) – gold (available only since November 2004).
As of 1/31/09, FUNDX has positions in DIA, IWM and GLD (among a total of 41 positions).
The following table summarizes total returns (using adjusted closing prices for the ETFs) for seven investing horizons since the inception of FUNDX at the beginning of November 2001. Some observations are:
FUNDX matches or beats returns of all three equity ETFs for three of the seven investment horizons.
FUNDX substantially underperforms the bond and gold ETFs for all horizons available for comparison. The point is not that investors should benchmark FUNDX against bond or gold returns (although, as noted, FUNDX holds GLD as of 1/31/09, and gold probably should be factored into any bencmark for the fund), but that a momentum strategy that excludes or arbitrarily segments asset classes may not be optimal.
For taxable accounts, taxes would likely degrade the FUNDX performance relative to buy-and-hold returns of the ETFs substantially due to FUNDX trading of funds (“upgrading”).
These real results, in the context of the strong hypothetical intermediate-term momentum results presented for the NoLoad FundX approach, indicate that the FUNDX mutual fund implementation of the approach may not substantially exploit available momentum. Possible reasons for this ostensible underachievement are:
- The fees, expenses and other frictions involved in implementing FUNDX substantially offset momentum returns (the market is mostly efficient after frictions).
- FUNDX does not sufficiently concentrate positions among just a very few funds with the highest momentum.
- FUNDX does not reallocate frequently enough to those funds with the very highest levels of momentum.
In addition, the NoLoad FundX approach might benefit from consolidating a wide enough range of asset classes (for example, to include bond funds) within a single analysis universe.
In summary, while research generally supports belief in an intermediate-term momentum effect for equities, it is not obvious that the FUNDX mutual fund substantially exploits the effect. Investors may be able to capture more of the effect by applying the NoLoad FundX approach more purely themselves.
For another perspective on NoLoad FundX, see Mark Hulbert’s 2006 article entitled “What Constitutes the Long Term?”, which addresses the streakiness of the performance of this strategy.
