Behind the scenes, I’m working on an Asset Class ETF Rotational Model which rotates weekly. This post will explore one aspect of building this model: Which day of the week is the best day on which to rotate?
Using Faber’s TAA as a rough guide, I’ve selected 5 ETFs with which to use for testing a weekly rotational model.
- VTI Vanguard Total Stock Market
- VEU Vanguard FTSE All-World Ex- U.S.
- DBC Power Shares DB Commodity Index Tracking
- IYR iShares Dow Jones U.S. Real Estate Index
- IEF iShares Barclays 7-10 Year Treasury Bond
One of my earliest thoughts about this system was wondering whether there was a particular day of the week on which to rotate that would be better than the others. It appears there is.
The Rules:
- Buy the ETF on Monday’s close and sell at the close 5 trading days (1 week) later.
- Repeat for each day of the week.
No commissions or slippage were included. All available history for the ETFs was used.
The Results:
Thoughts and Caveats:
If you had to rotate on a particular day of the week and may only hold the trade for a week, on which day(s) would you rotate? On which day(s) would you definitely not want to rotate?
There are no guarantees that this pattern will persist, but IYR and VTI have more than a decade’s worth of history.
If the pattern does persist, rotating on the best day could add enough of an edge to cover commissions or even the expense ratio of the ETFs.
I’m curious what the results would be if the hold time was extended to 2, 3, or even more weeks. Would the anomaly still persist over a longer time frame?
Wow. I should have chose a more exciting title for this piece.
Nice way to look at things Woodshedder. I wonder if using day of month in combination with the system you are using for your wife’s Fidelity account would add any incremental return.
I’ve had to put my test of Faber’s models on hold as my little duckling now takes up all of my free time…
Duck, there is really no way to institute a day of the month rotation in Fidelity since the funds have to be held for 30 days and then any day thereafter they may be rotated. In short, the days will be constantly changing.
Can this rotation be applied to stocks in general? I’ve always wondered if there are better days to trade than others for stocks.
I think this is a very useful discovery for the EFT asset classe. Thank you!
For the rotation strategy to work, the set of stocks or etfs or what have you must have some subset whose returns are negatively correlated with the rest. As a practical matter, you must have some fixed income assets (bond funds for example) in your basket. Here it is IEF. You can replace IEF by AGG or TIP or TLT and still get good results. Quarterly rotation with any of these will get you 12% or above annualized for 2003-2012.
I’ve been working on a similar approach to goosing rotation model returns using a different blend of ETFs. I’ve found that adding a conditional filter to the day of week trade(such as Close > “x” day MA) will produce a trading day target that may be offset from your study by a day or 2. Based on many runs of rather simple long/short, stocks/bond rotation models it can be expected that there may in fact be no weekly changes in the momentum ranking of portfolio components for many weeks, meaning no trading is required unless the model rebalances position sizing to maintain target % capital allocation proportions. That being said, I’ve concluded that absolute returns on a market neutral portfolio can be increased by about 2-4% APR using an optimum day of week transaction strategy.
Thanks BZB. Have you blogged anything about the conditional filter. I’d like to read more about it.
Yes, I’ve seen that the system may go for many weeks without rotating. I suspect that is ideal, or the system may be getting whipsawed.
Your estimate of 2-4% APR is nothing to sneeze at. I’m glad I’m heading in the right direction with this research.
Try this:
1. Start on the close of the first full week of the year.
2. Buy the one that did the best for the the period of twelve weeks ending on the close of the prior week.
3. Hold for sixteen weeks.
4. Repeat step 2 and 3 every sixteen weeks except that close the position at the end of the year. (So the last period will be only three to four weeks.)
5. Repeat Steps 1-4 for the years 2003-2012.
If you do it right, you will be amazed at the results. Like really amazed.
To ensure that this is a robust strategy, try evaluation periods of 10,11,12,13,14 weeks and holding periods of 14,15,16,17, and 18 weeks.
gregor, I started working on something similar to that some time ago, and then quit, as the whole thing felt curve fit. I saw the idea here: http://seekingalpha.com/instablog/709762-varan/251242-a-low-drawdown-strategy-for-sector-rotation-for-fidelity-select-funds
Anyway, it must be serendipity that you’ve also mentioned it, therefore, I must test it! Thanks!
That’s why you make sure that changing the parameters (the holding period and the evaluation period) still gives you good returns. That is the only way to assess the robustness of any strategy which is designed on the basis of historical data. Even if you optimize on the basis of a part of the data and do cross validation by testing with the rest, and select the parameters on the basis of cross-validation performance you are still ‘curve-fitting’.