I was flipping though Mike Carr’s book again the other day, and decided to download the monthly data for the Fidelity Select Sector funds from Yahoo. (Part of the reason was due to my frustration with the state of asset class testing software that I did a post on the other day).
I reran my simple rotation system on the universe of 23 funds that had data back to 1988. Results were broadly similar as those in my book, albeit with higher drawdowns. (These all are equity funds, so they lose the diversification benefit of including bonds, etc). This test was meant to be instructive rather than exhaustive.
Simply taking the top fund, updated monthly, based on the average rolling 3/6/12 month performance resulted in compound returns of 18% per year, vs. 10% for the average fund.
Simply taking the top 3 funds , updated monthly, based on the average rolling 3/6/12 month performance resulted in compound returns of 16% per year, vs. 10% for the average fund.
Both systems had similar volatility and drawdowns (54%, occurring in 08/09) as the underlying funds (average of 61%).
I think Fido has some screwy rules for holding their funds, but I see no reason why this wouldn’t work splendidly with a diverse portoflio of ETFs or mutual funds. The sample 5, 10, 20 fund allocations from my book would be a good start.
Didn't mention it in the other thread, but I use Amibroker for testing systems like this because it has a rotational trader. Might be worth a look.
I think I own Amibroker but never got into it. . .
It's good – can automatically download data from Yahoo, and then, from there, you can program a simple rotation pretty simply – email me if you want some sample code to get you started.
no thank you, I want to be programming less, not more! sounds like a good
intern project though .. .
LOL – it's all of about 10 lines of code – I think you could handle it.
Also Portfolio123 just added the same capability.
only back to 00 though right?
Don't know – I don't have enough time to look at P123 to see the capabilities – if they had added it 2 years ago I would be using it. Maybe it's a job for my intern.
What would the results be if you add the 10-mo-MA filter to the entry (ie no tinvest if the selected fund is below 10-mo-MA)?
I would imagine it would be very easy to test if you already have all the data downloaded in Excel.
Mebane, not a statistician here… when you say average rolling are you simply ranking the 23 funds from 1 to 23 based on 3 mo, 6 mo, and 12 mo and whomever has the lowest number (i.e. #1 in 12 mo, #3 in 3 mo, #4 in 6 mo =
are the 3 to go with? then rebalance monthly?
the drawdown of 54% is a bit too much for my taste though lol
To build on what Andrey wrote: maybe the drawdowns could be reduced by going to cash (or equivalent) when the market goes under the 10-month average at month's end?
check back tomorrow.
check back soon
Mebane,
This is similar to what Janet Brown has been doing successfully for years with her FundX strategy.
“Janet M. Brown, President of DAL and Managing Editor of NoLoad Fund*X, joined DAL in 1978. Janet has been researching funds and developing fund investment strategies for many years. Prior to joining DAL, she worked in Brussels with a major financial services company where she specialized in mutual funds. Janet is frequently interviewed by the media on investment and mutual fund issues. “
Yes and they got killed with it during the downturn. The strategy does not do well in bear markets in general – too much quick-shifting momentum without longer term trends. Combine with a 200dma and it's more interesting.
I haven't checked their track record lately, but makes sense. Thanks for correcting my observation.
idea… When using a mutual fund allocation with restrictions: Why not allocate a minimal % to all funds in an effort to avoid liquidation penalties (would it work?). Rebalance weighting to the Top X funds.
Have you considered weighting the rotational strategy?
ex: 5 Funds
1: 35%
2: 25%
3: 20%
4: 10%
5: 10%
idea… When using a mutual fund allocation with restrictions: Why not allocate a minimal % to all funds in an effort to avoid liquidation penalties (would it work?). Rebalance weighting to the Top X funds.
Have you considered weighting the rotational strategy?
ex: 5 Funds
1: 35%
2: 25%
3: 20%
4: 10%
5: 10%