Trendfollowing applied to asset allocation

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Post Reply
budonk
Roundtable Fellow
Roundtable Fellow
Posts: 62
Joined: Tue Dec 30, 2003 7:44 pm
Location: Lake Tahoe, NV

Trendfollowing applied to asset allocation

Post by budonk »

or, titled " A Quantitative Approach to Tactical Asset Allocation" in this months Journal of Wealth Management.

http://ssrn.com/abstract=962461
sluggo
Roundtable Knight
Roundtable Knight
Posts: 2987
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo »

Could you share a little bit about the review process used by the Journal of Wealth Management? For example, how much time elapses between submission of manuscript and receipt of reviewer's first round of commentary? How many different reviewers provided feedback to the authors? Are the reviews and reviewers anonymous? Does the Associate Editor take an active role in the review process, or is he mostly a facilitator and checklist-keeper? Thanks much.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

Palmer Wright, familiar to those of us who used to receive the TR list, pointed out the following site a while ago.

http://www.fundx.com/

This mutual fund advisory service probably achieves rather a similar purpose using a mostly mechanical system to switch between mutual funds. If the Far East equity market is hot, the system has you in the relevant iShares ETF etc.

I have not yet read the paper pointed out by budonk but bearing in mind the backtesting goes back to 1972, I am not entirely clear why Sluggo is so concerned as to the timeliness of the decision making process, nor indeed as to the process itself. It seems to me that a mechanical method of allocating accross assets classes has a great deal of merit. And as to timeliness and being bogged down by red tape, if you want to apply such timing to your own investments - well you will not face any red tape and can act whenever your system tells you to.
budonk
Roundtable Fellow
Roundtable Fellow
Posts: 62
Joined: Tue Dec 30, 2003 7:44 pm
Location: Lake Tahoe, NV

Post by budonk »

AFJ - I think Sluggo is referring to the review process of getting a paper published at an academic journal...correct me if am wrong Sluggo.

The #1 comment I would make is that the author must make sure he has addressed the target audience for the publication. The readers of the FAJ, JWM, JPM, or JTA are all different. Likewise, the submission process, submission rules and formats are all different. When I first submitted my paper to the FAJ, is was not ready for publication. It was a great research topic, but was not well written and a bit draft-y.

If they like the paper, they will work to turn a "B" paper into an "A" paper. But they will outright decline a poorly written paper (including papers that do not properly have knowledge of the academic literature).

Most journals have a near 100% rejection rate (the JPortfolio Managment had something like a 1-2 year queue for papers it already wanted to publish). As far as my experience with the JWM, it was smooth. All anonymous reviewers, the editor was involved but mostly a gatekeeper. (Note: I had already had the paper commented on by dozens of academics, etc). The FAJ is more computerized.

Lemme know if you have any more questions. Like to hear your thoughts on the paper as well. Feel free to contact me off-forum as well.
sluggo
Roundtable Knight
Roundtable Knight
Posts: 2987
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo »

budonk wrote:AFJ - I think Sluggo is referring to the review process of getting a paper published in this particular academic journal

Yes exactly. By "submission of manuscript" I meant, an author sends ("submits") a first draft of an article ("manuscript") to the Editorial Board of a journal, hoping they will accept it for publication. Sorry if the lingo was confusing to Roundtable readers. And thanks for letting us know how it works at JWM. I've been an author and an anonymous referree (though thankfully not an Associate Editor!) but not for this particular journal.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

The fact of the matter is that using a very long term trend following method on a widely diversified basket of world futures IS asset allocation.

And it is really investing, not trading, when a winning trade lasts an average of 3 years or some such period.

Think of it: you have stock markets and these days you can go almost as "emerging" as you like. Though for some, the Hang Seng may be quite exotic enough. And you have allocations to metals, agricultural products, bonds/interest rates, energies and so on.

And think of the ease and convenience of futures compared to cash stocks or even mutual funds. Complete coverage of most markets with a mere 50 or 100 instruments - CBOT even have a real estate contract now (let's see how it develops). Built in leverage which you can ratchet up or down. Instant and cheap dealing. No stock splits, dividends, rights issues to worry about. Instant access to your money - no lock in periods of up to 180 days on mutual funds and 2 years on hedge funds.
Daily statements and NAV.

I found the paper interesting although a much fairer comparison would include com and slippage. I also found the fundx website interesting - a different take on very long term trend following, using a different methodology and different instruments.

But when it boils down to which instruments to use, it may well be that futures can produce a very similar risk reward ratio for a lot less effort than other avenues.

And as they used to say on Blue Peter (the charming childrens' television programme) "this is something you can do at home"!
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

By the way, one silly little point: the NAREIT total return indices are "gross": did you adjust down to net terms for your testing? Helpfully, MSCI, as you will have noticed, gives you a choice of gross or net (IE after deducting withholding tax on dividends). Not sure on the bond index you used - was that net or gross? I assume net.

I suspect none of this would materially affect your findings.

I tested the MSCI World Net USD Standard Index using buy and hold and then using a simple 12 month Bollinger Band breakout with a 0.5 std dev band. Using monthly data. I then split the index into its many component parts and retested. I also tested a mix of Value and Growth for each country in one bag. Splitting the index up produced much better risk adjusted returns. The simple momentum strategy beat buy and hold hands down on a risk adjusted basis. No doubt I would have been even more impressed if I had added a liberal portion of a bond index, a REIT Index, an alternative asset index and so forth - but my specific interest was the stock markets.

No re-balancing for the buy and hold and taking no account of the initial market cap weightings back in 1969 or thereafter. I simply divided the trading equity by the number of trading instruments. For both systems.

Put simply, "diversification pays off handsomely" and "a market timing strategy makes an investment a hell of a lot easier to live with".

Again, a most enjoyable paper and I am glad to have been stimulated into testing a trend following strategy on different instruments.

At the end of the day a very long term trend following strategy on a very widely diversified and very large portfolio of futures can produce results as good or better than a system using cash instruments. And as for leverage – that is optional with futures and a cheap option at that. Ratchet it up or down as you choose.

The fact remains however that the majority of the investing public would find it a great deal easier to use an LTTF strategy with ETFs or no load mutual funds.

And the investing public would be very well advised to use a strategy such as that you outline and to spurn the host of largely useless active fund managers and advisors who charge so handsomely for such sub sub-optimal results.

It may amuse US readers to learn that the concept of no Load Mutual funds does not seem to exist in Europe. A typical actively managed open ended fund over here charges an entry fee of 2 to 5%, annual management fees of anything up to 3% and a bid offer spread of anything up to 7%.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

In the unlikely event of anyone having interest, I attach a spreadsheet showing some results of tests I ran on the MSCI World Index and it constituents using:

1) buy and hold and 2) a Bollinger band breakout system.

I have given little thought to money management as yet and no doubt I will come up with many further questions to test and will discover errors I have made.

None the less, the results show prima facie support for the supremacy of a simple momentum based system over buy and hold in the stock markets. At least when traded in a tax free account.

Personally, I would find it much easier to live with the shallower and shorter drawdowns of a market timing system and to have my money earning some interest in the bank during the miserable bear years. Note I have only assumed 3% as the rate of interest earned. Probably 6% is nearer the mark in the long term.

Of course, inflation would have taken quite a chunk out of these returns: but a modicum of gearing may perhaps have circumvented this problem , even at the awful borrowing rates prevalent at times over the past 30 years. A topic for some further research perhaps.
Attachments
MSCI Tests.xls
(18 KiB) Downloaded 600 times
mc
Roundtable Knight
Roundtable Knight
Posts: 167
Joined: Fri Nov 10, 2006 3:32 am

Post by mc »

Hi, I have followed your updates in this thread and studied your spreadsheet. Good work and thank you for sharing it. I also enjoyed the finance article - easier maths than usual.

This shows again that momentum is a good performing indicator and in particular in curtailing the worst drawdown during those sharp bear markets affecting buy-and-hold. If you lose as much as 45% you have to make 82% to get to back where you started! Avoiding getting into that predicament is rule #1, right?

ETFs are an attractive recent option to many (European) mutual funds/unit trusts and keeping management fees and trx costs down is essential in the long term. Taxes are another real-world consideration usually left out of simulations (including academic ones) and the tax implications would probably be very different for the two strategies, ignoring tax havens. A question is how buy-and-hold would have performed with annual equally-weighted rebalancing.
Post Reply