|
Factoids Yahoo Banks Excite Banks Banking News Bankstocks.com 2006 Updates Dec Nov Oct Sept Aug July Jun6 May April Mar Feb Jan 2005 Updates Dec Nov Oct Sept Aug July Jun May Aprl Mar Feb Jan 2004 Updates Dec Nov Oct Sept Aug July Jun May Aprl Mar Feb Jan 2003 Updates Dec Nov Oct |
Where the Forecaster Failed & Why In 2006, geography was destiny. It was bad to be in California and Oregon - where the real estate market was bad and the net interest margins were falling. It was good to be in Texas and Oklahoma - where the energy sector helped job growth and the real estate markets. And perhaps three southern banks may have received a post-Katrina bounce. And there was another theme - the yield hogs supported two stocks that should have fallen. There were too many 'exceptions' to feel good about the Predictor. I would like it to be accurate at least 70% of the time. But maybe 2006 was an exceptional year. Houston based PRSP rose 20% while CBSS - with its Dallas exposure - rose 23% and Oklahoma based BOKF rose 21% while the metrics for all three projected close to flat years. It is surprising that the amount of error was so consistent. California based CNY and UB both fell, with UB falling much more than projected. Since I did not have beginning of the year FFO estimates for most of these banks, I did not include those EPS' where I had them - so I could compare apples to apples. If I had included the falling 2006 EPS estimates for these two banks, then the projections would have been significantly more in error. Oregon's UMPQ was flat when the metrics suggested a huge gain. ALAB, BXS, HBHC and TRMK all rose much faster than their metrics suggested - thus I am guessing on the post-Katrina bounce. I believe I read that BXS had substantial deposit increases from insurance checks related to Katrina. With the beginning sector average yield in the low three's, FHN yielding its 4.68%, and FNB at 5.41%, stuck out as bait for the yield hogs - who I suspect supported their prices even thought the metrics suggested that they should fall around 10%. Both begin 2007 looking over-valued. I have yet to generate a theory why UMBF's P/E is so far above sector average - which is why the projection was so far off on that stock. And that makes 13 exceptions in a sector of 21. In 9 of the 21 cases, the estimates were within 5%. Getting that close that often should mean that I am on to something. But being so wrong so often means being uncertain all the time - even when you are right. Where the Forecaster Was Right & Why While the forecaster failed to consistently forecast individual stock returns, it did a very good job at selecting 'groups' of stocks. The group of stocks that were under-valued at price at a logical P/E ratio outperformed those that were over-valued [18.49% to 3.3%]. The group of stocks that were under-valued at price at a logical Price/Book ratio outperformed those that were over-valued [17.63% to 4.69%]. The differences in those returns are substantial. In 2006, the results were slightly skewed so that only eight of the 21 [38%] stocks beat sector average. But if one chose from the group of the under-valued, your odds were seven in 13 [54%] in choosing a stock that beat sector average. That 16 point difference is a 42% improvement in the odds. Add more stocks to your portfolio and your odds will rise further. And it is just as important to avoid the losers as it is to pick the winners. If one chose from they eight stocks that were over-valued - your odds of beating the sector average was only 12.5%. Going by the forecaster and buying 'best' three stocks at the beginning of the year, one would have bought RF, WTNY and ZION. Only ZION underperformed the sector - and only by 177 basis points. This portfolio would have had an average return of 15.62% vs. the sector average total return of 12.78%. Ironically, I would have only bought ZION, not liking RF or WTNY for their relatively poor histories of dividend growth. But I did not own any of the three. I owned CBSS - which beat the sector substantially. Some of that choice was skill - some of that choice was luck. Why should one expect this level of predictivity in the future? Because as long as humans are setting the prices [or as long as humans are writing the programs for those 'programmed' trades], some prices will be illogical. Some prices will be overly influenced by yield hogs. Some prices will be overly influenced by momentum players. Some prices will be set by those overly influenced by the results of the last quarter. And prices will always be set predominantly by those who have a short-term focus - and I intend on playing a long-term game and will measure my success with long-term metrics. Home Page Factoids Previous Update |