As my readers know, I believe the best formula to magic formula stocks is to go with ones that pay dividends. Even more specifically, go with stocks with larger market caps that pay dividends.
Of course nothing in life is guaranteed or certain and this is certainly the case here. This is illustrated by the fact that dividend stocks are doing well (for tranches started in 2016 they are up 15.6%, while all MFI is 14.3% and R3K is 9.2%), but ironically (hah, hah) the smaller market caps are driving the outperformance. By. A. Lot. This has been happening for the last couple of years. But it seems to be exacerbated by the Trump rally. For those not living under a rock, the Trump rally has been more beneficial to smaller cap companies as they tend to to have less international sales (which may be hurt by stronger dollar). Check out this table:
| Year | big | small |
| 2006 | 51% | 8% |
| 2007 | 6% | -26% |
| 2008 | -2% | -7% |
| 2009 | 46% | 57% |
| 2010 | 24% | 11% |
| 2011 | 21% | -2% |
| 2012 | 50% | 21% |
| 2013 | 41% | 7% |
| 2014 | 10% | 7% |
| 2015 | 0% | 2% |
| 2016 | 7% | 32% |
| Grand Total | 25% | 8% |
Big and Small here are simply a quantile splitting of dividend stocks each tracking month. So Big are stocks amongst 25 largest in the screen and small are in 25 smallest. You can quickly see (aha) that over time big has been better than small (again these also are stocks that have dividends of 2.6% or greater). But that has flipped in 2015 and 2016. A wee bit frustrating.
Though even with this, my Formula tranches are hanging in there:
| 12/31/15 Stocks | Start | Current | Dividend | Pct Gain | R3K Gain |
| BBY | $30.31 | $48.44 | $1.57 | 65.0% | 12.1% |
| GME | $28.34 | $25.97 | $1.48 | -3.1% | 12.1% |
| ILG | $15.89 | $18.46 | $0.48 | 19.2% | 12.1% |
| PPC | $22.22 | $18.36 | $2.75 | -5.0% | 12.1% |
| VIAB | $40.93 | $38.12 | $1.00 | -4.4% | 12.1% |
| Totals | 14.3% | 12.1% | |||
| 4/1/16 Stocks | Start | Current | Dividend | Pct Gain | R3K Gain |
| CALM | $51.27 | $40.40 | $0.44 | -20.4% | 11.8% |
| HPQ | $12.10 | $16.10 | $0.25 | 35.0% | 11.8% |
| TIME | $15.24 | $16.50 | $0.57 | 12.0% | 11.8% |
| ILG | $13.85 | $18.46 | $0.36 | 35.8% | 11.8% |
| TSRA | $30.79 | $40.60 | $0.60 | 33.8% | 11.8% |
| Totals | 19.3% | 11.8% | |||
| 7/1/2016 | Start | Current | Dividend | Pct Gain | R3K Gain |
| ILG | $16.76 | $18.46 | $0.24 | 11.5% | 7.5% |
| VIAB | $44.00 | $38.12 | $0.20 | -12.9% | 7.5% |
| CPLA | $53.22 | $89.30 | $0.78 | 69.3% | 7.5% |
| HRB | $23.62 | $23.35 | $0.44 | 0.7% | 7.5% |
| PBI | $17.69 | $15.59 | $0.38 | -9.8% | 7.5% |
| Totals | 11.8% | 7.5% | |||
| 10/3/2016 | Start | Current | Dividend | Pct Gain | R3K Gain |
| CPLA | $58.15 | $89.30 | $0.39 | 54.2% | 4.6% |
| LDOS | $43.01 | $52.09 | $0.00 | 21.1% | 4.6% |
| GME | $27.70 | $25.97 | $0.37 | -4.9% | 4.6% |
| PBI | $18.15 | $15.59 | $0.19 | -13.1% | 4.6% |
| CSCO | $31.72 | $29.95 | $0.26 | -4.8% | 4.6% |
| Totals | 10.5% | 4.6% |
I think the reason is that my $600m market cap threshhold gets many stocks that would not be in 25 largest (CPLA for example). Looking back, the typical break point is about $2.5 billion. So if you take the 50 stock screen with $100m market cap, 1/2 stocks will be greater than $2.5b and 1/2 smaller.
If you take the table (Big vs Small) that I published just now and made it big, medium and small (with medium going down to $600m), you can see I have benefited from this in 2016 (though not 2015).
| Year | big | Medium | small2 |
| 2006 | 47% | 44% | 7% |
| 2007 | 15% | -10% | -29% |
| 2008 | 22% | -8% | -9% |
| 2009 | 32% | 53% | 60% |
| 2010 | 31% | 13% | 11% |
| 2011 | 22% | 16% | -8% |
| 2012 | 53% | 30% | 25% |
| 2013 | 41% | 19% | 5% |
| 2014 | 12% | -2% | 8% |
| 2015 | 3% | -10% | 3% |
| 2016 | 6% | 28% | 31% |
| Grand Total | 26% | 16% | 8% |
By Dividend Yield
I also looked at data by dividend yield. Now this is a bit circular. If GME has a constant dividend, you are obviously getting a better "deal" if you buy it when the dividend is 5% versus 4%. So to some extent, you clearly get better returns when yields are higher as there is some implication the market is discounted at that point (such as 2009).
| Yield Range | Big | Small | Total |
| no dividend | 3.8% | 4.7% | 4.4% |
| Under 1% | 4.3% | 27.0% | 11.1% |
| 1.0% to 2.6% | 8.9% | 8.2% | 8.7% |
| 2.6% to 4% | 19.9% | 11.9% | 17.3% |
| 4% to 6% | 28.0% | 14.2% | 23.0% |
| 6% to 10% | 15.1% | 1.8% | 7.7% |
| over 10% | 36.7% | 4.8% | 16.3% |
So for "Big" stocks (again, here Big is top 50th percentiles), you almost get a linear improvement. Higher yield range => higher return. Small stocks are a bit more choppy. If you split this by year, you'd see more stocks in the higher yield ranges when the market was discounted. I don't do that by year here as the data starts getting to crumbly.
Summary
Not sure where I am going with this. But here are some high level takeaways:
- There definitely seems to be a correlation to dividend yield and returns.
- There has also been a correlation between market cap and returns, though that seems to be reversing (perhaps as dollar strengthening hurts larger firms internationally).
- If you really wanted to optimize - you might want to decrease bet based on number of firms in each dividend yield bucket. To some extent, this is like taking chips off the table when market is extended.
- History is just an indicator. The problem with any "formula" is it may have worked because other factors were in play that no longer exist. So perhaps certain things worked better as treasuries were so low and yields were decreasing. Perhaps they don't work when treasury yields are increasing.
- Really the only thing I know for certain is the answer to the ultimate question of life is (wait for it) 42!
Posts
No comments:
Post a Comment