As a follow up to yesterdays post titled "How to beat the best market timing strategy" . A reader commented on the difference in annual returns, Hulbert states an 11.4% annualized return and I cited a 5.4% annualized return. The difference is 4 fold. ! 1) I tested the strategy on the S&P while Hulbert used the Wilshire 5000, 2) I didn't account for growth while not invested 3) I didn't get the exception rule coded and 4) we started on different dates.
1) The strategy is only invested 30% of all days. I didn't factor in how the account would grow while it was sitting in cash for the remaining 70% of days. From 1980 to present 7/21/11 the average 90 day T bill yield was 5.17% if we assumed earning 70% of that yield while this model is in cash there would be a 3.6% improvement to the 5.4% annual return, bringing it up to 9% annual return.
2) I dont have Wilshire 5000 data to run this strategy, I do have Russell 3000 data back to September 1987. Perhaps the Russell 3000 is more representative of the Wilshire 5000 versus the S&P 500. Below is a chart of using the original timing strategy on the Russell 3000 ( $RUA) and as a comparison the timing strategy on the S&P ( $SPX) . Using the Russell 3000 boost the annual return to 6.26% if we add the 3.6% improvement from the money market the model is at 9.86% annual return.
3) I still cant seem to code the exception rules, its making smoke come out of my ears, but I haven't given up.
4) I don't have the data to start in the early 1980's which Hulbert quoted the 11.4% annual performance from.
The difference between Hulberts quoted 11.4% annualized return and the 9.86% annualized return that I came up with is a combination of not having Wilshire 5000 data, not starting on the same date and not having the exception rules coded.
The title of yesterdays post was "HOW TO BEAT the best market timing strategy ever" . If we test the "How to beat " rules on the Russell 3000 since 1987 the account would have grown 10.6% annualized then add in the 3.6% per year in money market growth brings the return to 14.2% then tack on any additional boost in returns from the other variables we weren't able to account for, perhaps another 1-2% = ~15.2 to 16.2% return per year. Below you will find the equity curve using the "How to Beat " rules and original rules as a comparison both tested using the Russell 3000 and no account for money market growth.
FYI there is an ETF for the Wilshire 5000 ticker WFVK, it trades by appointment only 1,000 shares if you are lucky and has only been around for 1 year.
Check out some of our market timing models that are even better. We don' just beat the market we leave bruises! Join our team and become a champion, sign up for a free 14 day trial, here.
SPY, XLF, QQQ, EEM, EWJ, TNA, TZA,IWM, DIA, QID, IWB, IWO
4 comments:
I have an Excel spreadsheet with daily Wilshire 5000 index price data for the period 12/3/1979 - 8/1/2008. I never updated it since Aug 2008, though. If you're interested, I can send it to RipeTrade@gmail.com.
No , thank you. I do appreciate the offer though. The problem is my software platform isn't capable of pulling data from an outside source other than the 1 data provider that I pay. I think I probably got close enough with the Russell 3000. Besides if there isn't any efficient way to trade the Wilshire with an ETF or a no load fund, then whats the point of testing a strategy on it.
"If there isn't any efficient way to trade the Wilshire with an ETF or a no load fund, then whats the point of testing a strategy on it."
As my British friends say, spot-on.
"The strategy is only invested 30% of all days. I didn't factor in how the account would grow while it was sitting in cash for the remaining 70% of days. From 1980 to present 7/21/11 the average 90 day T bill yield was 5.17%..."
This brings up a great point. If someone gives performance figures for a model that spends a lot of time on the sidelines and the performance figures include T-bill interest, the model can be expected to seriously lag historical returns in these days of zero interest rates.
Post a Comment