1. Net-Net Screen. "Net-Net" tells us to consider buying stocks priced below a stripped-down net worth--which certainly makes sense. In fact, the only logical reason we generally pay 10 to 100 times more than a stock is worth is because we believe some even more gullible person is someday going to pay more. We then listen endlessly to experts and politicians arguing over why there are "outrageous" stock market crashes. Meanwhile the only real mystery to me is why we are so mystified. So... Net-Net sounds good... but let's see if it works...
I found 2001-2010 backtests for net value investing at OldSchoolValue.com. I am uncertain about the correct names for these tests. Old School Value has what they call the NNWC screen, or Net-Net Working Capital, and a separate screen they call NCAV, or Net Current Asset Value, using quite different formulas. At any rate, the NNWC screen at Old School Value uses the same formula as the Net-Net screen at Guru Focus, even though Guru Focus also calls this the "NCAV Bargain Screen."
Guru Focus mentions a test done for 1970-1983 resulting in almost a 30% annual average gain, but gives no year-by-year results. The Old School Value tests show an average gain more in the neighborhood of 40%. I am posting year-by-year results below. These figures are very approximate, taken from eyeballing a linear chart, and taking the average between NCAV and NNWC results. Also, there are two different pages showing different results for 2003, so I took the lowest of the two. Additional Old School Value criteria are: Volume always over 30k, NCAV always bought at 2/3 base value, 1% deducted for slippage, rebalanced every 6 months.
|2001||2002||2003||2004||2005||2006||2007||2008||2009||2010||5-yr avg ||10-yr avg|
- All Net-Net backtests are necessarily based on only about 10-20 qualifying equities per year, so results will certainly tend to be variable. Interestingly however, a 3-year backtest by Old School Value for NCAV at 1/1, instead of at 2/3 value, showed almost no gain at all. This along with all other testing seems to support the hypothesis that, if buying stocks of companies with zero popular appeal at 2/3 their net value, then on average you can expect to gain perhaps 1/3 their net value within one year. Also, extraordinary losses seem to produce even more extraordinary rebounds. Therefore, 2001-2010 results are probably unusually high. Something around 30% average annual gain seems a reasonable long-term supposition.
- Unfortunately, the idea that using Net-Net will be more stable than other forms of investing seems disproven. Nonetheless, the 2008 downturn of -46% is not severe relative to the average annual returns. During 2008 many mutual funds, newsletters and even precious metal funds saw downturns of -40% and sometimes even -60%, often without a swift rebound.
- Guru Focus probably improves results by listing only companies with a positive 12-month cashflow and negligible debt. Other Net-Net screens are also available such as at GrahamInvestor.com. It might be feasible to cross-reference all Net-Net and NCAV screens, as well as other relevant screens.
- Suggestion. I like the idea of comparing companies based on their most basic values. I wish that Guru Focus would list every available Net-Net value, not only the "best" results. Then I could use Net-Net in place of P/E to reinforce all my buying decisions.
- In backtesting for 1998-2008, Predictability by itself showed an amazingly consistent correlation to gain, with peak results at about 10-12% average annually for stocks rated 4-5 stars on "predictability" alone. This by itself is good and, by adding other value-based parameters, as GF has done with the UPC and BMC screens, we can certainly expect better.
- However we are only given the overall average for 1998-2008. What were the year-by-year scores? Then we can have much clearer verification of the theory that Predictable stocks remain more stable as well as better performing than the general market.
- The only other complaint against Predictability is that it does not include companies that may have qualified but went bankrupt. However the author believes this does not significantly alter results. Further research could possibly settle this issue.
- It seems likely to me that investors can do well simply by adding Predictability to any personal favorite screening parameters. "Predictability investing" seems to me similar to "index investing" or "blue chip investing," only basically more sensible, and also, always "picking your own index" directly from the total
range of stocks.
- Suggestions. (a) Let's please have Predictability
backtests for our last 4 unique decades? The 1970's, 1980's, 1990's and
2000's? (b) Also for each decade's backtest, please include year-by-year
results? (c) I think "Dependability" would be a more accurate screen title than "Predictability."
- BMC = Buffet-Munger Companies. This screen combines the Predictability screen with Buffet and Munger methods of finding "good companies at fair or undervalued prices." Standards also include high competitiveness and low debt. PEPG is used to determine low prices (P/E ratio divided by the average growth rate of EBITDA over the past 5 years).
- Backtesting of BMC for 1998-2008 resulted in an annualized gain of 20%, vs. 2.7% for the general market. This is excellent although again, I don't understand why we are not given a year-by-year breakdown, as we are for the performance of every guru.
- UPC = Undervalued Predictable Companies = Predictability + DCF. DCF (Discounted Free Cash Flow) is a somewhat intricate method, inspired by Warren Buffett, to determine whether a stock is under-priced.
- UPC and BMC therefore seem similar, except that we only have backtesting results for BMC.
- Presumably both may be combined by cross-referencing BMC with DCF.
- LPS screens for historically low Price/Sales ratios.
- LPB screens for historically low Price/Book Value ratios.
- Arnold Van Den Berg and Ken Fisher are cited as two gurus who extensively use Price/Sales. However both of these gurus show scores consistently echoing one another and the S&P 500 and with total returns that are actually lower than the general market for 2001-2010.
- Therefore it seems doubtful that P/S or P/B add anything of value. In essence, by comparing the price ratio to its past performance, these seem equivalent to technical analysis. P/S or P/B may indicate potential short-term price swings, but seem not likely to indicate where we are in long-term fundamental value.
- Incidentally, if we don't know what the "high" value was, then it seems to me, we can not begin to judge how valuable the "low" might be.
- At any rate, unless I am mistaken in my review of the available performance data, it seems to me that LPS and LPB should be ignored.
- Robert Rodriguez is cited as using 52WL as "one of" his strategies to average 15% annually over the past 20 years. However he must have earned most of this in the early 90's. As of June 2011, according to Guru Focus records, Rodriguez only seems to have averaged about 9% over 15 years, 11% over 10 years and 3% over 5 years.
- Peter Lynch is quoted as saying, "Insiders might sell their shares for any number of reasons, but they buy them for only one: they think the price will rise." I think not. They also buy stocks because of blind optimism, which is normally a helpful trait for dealing with the stress of being a CEO. Or they feel it is necessary to convince someone their company will recover, partly by going on record as buying. Whereas when a CEO or CFO sells at a peak, just as when rats leave a ship in a calm ocean, arguably that is more of a clear indication of something. However what I think or Lynch thinks hardly matters. How about some clear backtesting numbers of insider selling vs. insider buying? That shouldn't be so difficult.
- It is cited that insiders average 6% better than the general market. This does persuade me that it might be effective to follow insiders. However I think it's even better to follow gurus, who also can be found doing at least 6% better than the market. In addition, they also know what the insiders are doing.
- The guru-based 52WL screen seems likely to be handy to flag "potential" bargains. However other homework is need to determine the relative immediate upward potential of each such stock.