Who can beat the magic formula?

[frog03]

 

Very long term, the best stock mutual fund managers do about 5 points a year better than the market (see FPA Capital, First Eagle, Third Avenue...).

Berkowitz (on his managed accounts which started a bit before the mutual fund) and the Kinetics Paradigm fund do about 9 point better than the market.

Now, some hedge funds do even better (see recent post about Einhorn, Sprott, some others) but often do not accept new investors or have hefty requirements.

 

Seems to me that for a relativeley small investor (25K or more) that does not want to/cannot do the work himself, it is virtually impossible to find a manager that is likely to beat the magic formula especially on a pre-tax basis.  Do you agree?

 

[beerbaron]

Pure quantitative selection does not make any sense. If more investors go to magic formula then the amount of companies falling into this category will be reduced, the diversification falling we can also expect the risk to increase as the remaining companies are probably lower quality. I think it would be wise to apply the magic formula only when it is out of favour.

 

If I were using the magic formula I would tweak the timing. I would only buy it when the amount of companies in the screening is 15% above it's historical average. That way you buy when value investing is out of favor.

 

Small investors without knowledge should just buy the lowest fees Equity Index ETF coupled with bonds ETF. The ratio of the two should be mixed depending on the investor's tolerance to volatility.

 

BeerBaron

[twacowfca]

 

Very long term, the best stock mutual fund managers do about 5 points a year better than the market (see FPA Capital, First Eagle, Third Avenue...).

Berkowitz (on his managed accounts which started a bit before the mutual fund) and the Kinetics Paradigm fund do about 9 point better than the market.

Now, some hedge funds do even better (see recent post about Einhorn, Sprott, some others) but often do not accept new investors or have hefty requirements.

 

Seems to me that for a relativeley small investor (25K or more) that does not want to/cannot do the work himself, it is virtually impossible to find a manager that is likely to beat the magic formula especially on a pre-tax basis.  Do you agree?

 

 

 

 

The MF is a good screen.  The problem with backtested screens is that these rarely identify durable criteria.  The hypothetical returns from models based on such screens usually involve switching frequently from stocks that no longer meet the criteria to other stocks that do.

 

In the real world, flipping stocks generates frictional transaction costs that usually cancel the hypothetical gains.  Nevertheless, the MF model uses a holding period that is a little longer than similar value oriented models.  They have recently launched a fund based on the model which is currently outperforming.  It will be interesting to see how well it performs through the entire cycle, including the less frequent periods when the MF model tanks.  If you forced me to make a bet, I would bet that their model in actual trading will outperform most other funds, including index funds.

[ericd1]

The Ivy Portfolio by Faber and Richardson offers an interesting approach to active style management using ETF's. You might want to pick up a copy. You don't invest in any one manager, but rather in the top performing equity styles, or asset classes. I've used a similar strategy for over 11 years using actively managed funds, which can outperform index funds, or ETF's tracking indices and have outperformed the market by 7.8% annually. Not record breaking, but I'm satisfied.

 

 

[twacowfca]

We took more than a cursory look ( but less than definitive ) at MFI a couple of years ago. Our tentative conclusions were as follows:

 

It works best when value investing in general outperforms in the market.  Don't expect good performance in a bubble or when the market rolls over on top of the bear.

 

It seems to work better for some industries than others.  Tech especially seems to be problematical; it may be that when a tech co shows up on the MF list, this sometimes signals an impending loss of their fleeting advantage.

 

Companies and industries that don't have subpar economics seem to do better than others when they show up on the MFI list at a time that the MF model is outperforming.

 

Does anyone have more info about which industries work especially well?  When the economy was in good shape, specialty retailing and consumer cos with brand names seemed to do fairly well, but poorly when the model wasn't working.

 

 

[manualofideas]

We did a report on magic formula investing a while back that some members may find useful. We discussed the pros of the magic formula approach on pages 4-6 and included some suggestions for narrowing down the list of magic formula screen results.

 

http://manualofideas.com/files/moi200811.pdf

 

[Parsad]

Seems to me that for a relativeley small investor (25K or more) that does not want to/cannot do the work himself, it is virtually impossible to find a manager that is likely to beat the magic formula especially on a pre-tax basis.  Do you agree?

 

I'm sorry, I'm one of the handful of people who don't buy this.  Magic Formula is just one of the hottest new market tactics available.  Please find me anyone who actually practices Magic Formula and has beaten the markets regularly...and not back-tested data, but actual published results somewhere, be it an audited fund manager, or private investor who actually is on the record with their results and we can compare. 

 

I've seen so many Magic Formula blogs and I know some young managers are now embracing Magic Formula, but let's see the actual results that can be verified.  I know of many investors who have killed the markets over the last decade (Buffett, Simpson, Klarman, Watsa, Pabrai (PIF 1&2), Biglari, etc.), but I would like to see as many managers who have done so using Magic Formula.  Cheers!

 

[Guest HarryLong]

For what it's worth, the main criteria should be that for any system, like MFI, any proposed improvement should be objectively systematic.

 

The Manual of Ideas suggestions don't quite make the cut. They're a checklist, but leave room for discretion, rather than being fully systematic.

 

MFI can be improved upon with additional systematic overlays.

[twacowfca]

For what it's worth, the main criteria should be that for any system, like MFI, any proposed improvement should be objectively systematic.

 

The Manual of Ideas suggestions don't quite make the cut. They're a checklist, but leave room for discretion, rather than being fully systematic.

 

MFI can be improved upon with additional systematic overlays.

 

 

We tried some systematic overlays.  Only two of these that we could identify seemed to be promising.

 

 

Putting MFI in mothballs when the economic cycle ( as measured by objective criteria such as standard deviations for unemployment) is peaking , rolling over, and approaching the bottom.  However, this would improve the returns of virtually all long models.  

 

Secondly,  the quality of the business, measured by the industry code broadly defined, not too specific.

 

I asked Joel once in a Q&A session about MFI which industries performed better in his model.  He answered,  "I can't talk about that".  But he was willing to answer all other questions about the model.  I posed the same question on another occasion to one of his close younger associates, and received the same answer.

 

 

[Guest Broxburnboy]

 

Seems to me that for a relativeley small investor (25K or more) that does not want to/cannot do the work himself, it is virtually impossible to find a manager that is likely to beat the magic formula especially on a pre-tax basis.  Do you agree?

 

 

The small investor has several advantages over mutual fund managers, if s/he does their own due dilligence. The whole micro and mini cap universe is open to them, whereas most fund manages can't or won't play in that space.

My own personal experience is that the individual investor should only buy into funds that lever that fund's advantages: hedge strategies, access to M&A arbitrage, superior independent research and experience, sector expertise.

 

In my opinion ETFs and balanced funds are for those who do not wish to learn the investing game and are willing to settle for average performance. Even the most cursory DD in the banking sector for example will reveal an investment that will outperform that sector and consequently any sector balanced mutual fund.

The funds which I have retained over time are "hedge funds" with very generous incentive bonuses for the fund mangers but consistently outperform their rivals. Past performance during bear markets seems to me to be the best predictor of future performance during any market.

 

Cheers BBB

[twacowfca]

For what it's worth, the main criteria should be that for any system, like MFI, any proposed improvement should be objectively systematic.

 

The Manual of Ideas suggestions don't quite make the cut. They're a checklist, but leave room for discretion, rather than being fully systematic.

 

MFI can be improved upon with additional systematic overlays.

 

 

I agree with your assessment, Harry.  The MOI checklist has a lot of generally good value investing criteria, but some of these checks may not resonate with the MF model.  For example, excluding a co if there is substantial insider selling and excluding faddish cos.  Our biggest winner from the MFI list was a co that should have been excluded by the checklist: Deckers Outdoors.  Scuttlebutt research with the teenage daughters of friends suggested that their boots were still very popular, although many suspected they were a fad in the fall of 05.  Insider selling was massive.  Normally a big red flag, but a little digging revealed that the selling was by the founder who was mad at the CEO who had an awesome track record elsewhere and had transformed their business for the better.  Finally, the huge late summer and early fall build up in their inventory was not a red flag, as it is on screens for such cos because their boots were still hot and most sales were in the fall and winter, necessitating a huge buildup before their prime, cold weather season.

 

In summary, the MFI screen is a great starting point for digging deeper to see if you can find a gem.

[SharperDingaan]

 

Very sexy, but please explain why I shouldn’t just look up which are the current worst performing sectors on ‘XYZ’ exchange, & buy that sector (single share, index, or mutual fund).

 

Its highly unlikely that 2 years from now the current sector ‘dog’ will still be on the bottom, & very likely that the next 2 years are not going to economically look like the back-tested data set.

 

SD 

 

[netnet]

We have found MF to be a highly lucrative area in which to hunt.  The three turnoffs are 1) unexplained insider selling 2) tech companies 3) unacceptable business risk--such as single  troubled customer, etc.  We got hammered in the downturn, who didn't?

 

Very sexy, but please explain why I shouldn’t just look up which are the current worst performing sectors on ‘XYZ’ exchange, & buy that sector (single share, index, or mutual fund).

 

Its highly unlikely that 2 years from now the current sector ‘dog’ will still be on the bottom, & very likely that the next 2 years are not going to economically look like the back-tested data set.

 

Sharper, we are generally playing a mean reversion game, but MF should get you better results because the companies are sorted and ranked.

 

 

netnet

[beerbaron]

Did anybody do a backtracking for a stock with:

 

1- Low PE ratio (or high yield)

2- High dividend distribution

 

And apply the same principles then the magic formula (lowest score is the best)

 

The way I see it the MF gives Fisher style and the other would give Graham style. In our kind of environment I would prefer Graham as past return on investment might have nothing to do with future.

 

BeerBaron

[twacowfca]

Did anybody do a backtracking for a stock with:

 

1- Low PE ratio (or high yield)

2- High dividend distribution

 

And apply the same principles then the magic formula (lowest score is the best)

 

The way I see it the MF gives Fisher style and the other would give Graham style. In our kind of environment I would prefer Graham as past return on investment might have nothing to do with future.

 

BeerBaron

 

 

Low p/e is somewhat redundant because their model measures this aspect of value in a somewhat different way.  High dividend yield is rarely found in MFI stocks, and is a different aspect of value that doesn't improve their model results, if I'm not mistaken, in our attempts to tweak their model a few years ago.