Kim Eng's Long Dec Short Jan Strategy for SMT 1 comments

(P.S: Sorry for any disturbances the advertisements above may have caused you)
Occasionally I come across stock recommendations based on premises so absurd that I just have to write about it. Today's analyst report by Kim Eng's Range Trader on Surface Mount (SMT) recommending a buy in December 05 and a sell in January 06 is a case in point.

Let me clarify that I have absolutely nothing against analyst reports in general. In fact, I believe a lot of the time they are useful for the small investor, as outlined in an earlier article The Case for Analyst Reports. I also have nothing against SMT; in fact I believe it is one of the better price-for-value tech stocks on the SGX, given its steady fundamentals and creditable dividend policy.

As for charting, I myself am not skilled at reading charts beyond the price-volume and the moving average; things like Bollinger charts and MACD indicators are arcane to me. I do, however, appreciate that technicals analysis, if done properly, can provide hints/indications of good (or bad) news or fundamentals which have not yet been made public, and some have managed (or at least claimed to have managed) to integrate price-volume stock analysis with fundamentals analysis of the company's business to pick the best stocks to inject their limited liquidity into. So, it's an approach which attempts to combine publicly available information (fundamentals) with non-publicly available information (through price-volume hints interpreted through charting) to provide a more comprehensive picture. A logical approach.

But when someone attempts to recommend a trading strategy based on historical trading patterns of the particular stock without caring about explaining the possible underlying reasons why it should behave so, it is rather irritating. There is of course the well-known Capricorn effect where stocks tend to appreciate at year-end for window-dressing (ie. boosting reported portfolio gains) but there is no mention of this in the analyst report. Instead, Kim Eng's technical analyst observes that since the stock has attained temporal peaks in Jan 2004, July 2004, Jan 2005, July 2005, it is therefore likely to peak in Jan 2006, hence the proposed "Long December Short January" strategy.

The analyst recognises the undervaluation of SMT given its prospects, and therefore recommends a long on it in December 2005. However, based on historical price pattern he recommends a sell in January 2006. Clearly the fundamentals won't have changed much within this one month; they don't enter into his consideration for the "sell in January" call anyway. This is the crux of the problem: my belief, as stated above, is that technical analysis is built on the premise that non-public, or even insider, information may be embedded in the price-volume trend/pattern. If no regard is paid to the fundamentals that belie the charts then one is not reading them intelligently. From what I remember, SMT peaked in Jan 04 because it had risen above $1 in late 2003 during the bull run ie. it became overvalued, hence the correction; it peaked in Jul 04 because of oil concerns; it peaked in Jan 05 because of poorer than expected growth; and it peaked in Jul 05 because the general tech rally ran out of steam. To me these are timing coincidences that could just as well have happened in other months; reading too much into them smacks of data-mining and reading patterns where there are none.

References:
(1) Kim Eng analyst report Dec 13: Surface Mount Technology- Long December Short January