The great thing about being gainfully unemployed is that I can react better to a major event such as Investors Conference 2018. I have spent the better part of yesterday with a Bloomberg terminal and can now write something on what the other speakers are saying.
Note that it is not my intention to diss anyone's investment strategy. Backtesting itself is highly subjective and results can vary when the backtesting period is varied.
a) My own STI ETF presentation
The BIGScribe directors are holding all speakers accountable for their claims. In this case, I am lucky because I did not make a stock recommendation but merely three broad strategies involving Singapore blue chip stock components of the STI. So yesterday, I tried to expand my backtesting period to 15 years to see if my results can hold.
Holding an equal-weighted blue chip portfolio returned only 4.56% over 15 years. If you pick the top 15 dividend stocks, performance actually gets worse and the portfolio only returned 4.19%. Fortunately, the low PE strategy retained some of its mojo and continued to return 6.75%.
This gives me some confidence that at least some of my ideas can withstand the test of time.
Now that I have subjected my own investment ideas against my own standards, let's move to the ideas shared by other speakers.
b) Starhill REIT
This is the third time Brian and participated on the same event. In every seminar, I listen intently to Brian because I like his candid sharing of his personal portfolio and finely-honed gambling instincts ( As in the gambling instincts that make someone win BIG ).
When Brian listed Starhill REIT in his portfolio, I perked up. You see, I tend to skip retail REITs because I have no idea how the retail sector is responding the disruption from online vendors. My idea is that solid management would eventually be beaten by industry fundamentals.
I did a double-take on Starhill and backtested one of my favourite strategies, selecting the highest dividends yields then choosing those with the lowest gearing. Lo and behold, Starhill REIT magically appeared when previous backtests showed that it hadn't.
This gives me confidence that Brian knows what he is doing and I will not hesitate to add Starhill into my portfolio moving forward.
After all, the best investors steal ideas from others.
c) Sustainable Competitive Advantage
I did not warm up to quite a few ideas from other speakers.
One difficult concept is that Old Chang Kee has some kind of sustainable competitive advantage over their business. I think that any Mak Chik with a basket of Sardine Epoks can potentially change the game for Old Chang Kee so I don't really buy that argument. Maybe Old Chang Kee might have an Investment Moat based on logistics and their internal SOP but F&B is a tough business in Singapore and we are fickle eaters.
As a consequence of this, I do not really speak on moats or margin of safety. I prefer things that I can measure. Anyway, I don't aspire to become Warren Buffett and I think that these concepts are too fluffy to be useful to serious investor.
But there is no harm reading what analysts have to say.
d) Growth at a Reasonable Price
The bulk of my backtesting was focused on GARP with the PEG ratio being the focus on new strategies.
We can say generally speaking that GARP works. Strategies I employed on local stocks returned 12.44% with a semi-variance of 12.38%. A Sharpe ratio of 0.58 is not bad, but does not belong to the set of strategies which have to cross the 0.80 bar to be put into action.
However,employing GARP on local blue chip stocks disappointed me. Lowest PEG stocks in the STI underperformed the equal weighted strategy at 3.93%. So if you wish to apply GARP strategies, make sure you employ this on the larger local stock universe and listen intently to the speaker that day.
It is meant to be combined with other screens.
Over the next few weeks, I will be promoting my next event in which I fly solo, so do keep a lookout on this space.
( Hats off if anyone can relate the pic to this post. )
Problem with that Dune test is that it focused on only a single factor, at a single moment in time, with a binary live or die result.
ReplyDeleteMany such tests can be designed to always fail, in entertaining ways. And usually a favorite pastime of dictators & medieval rulers.
Subjective backtesting is another term for data-mining. 15, or even 30 years is basically too short to provide robust conclusions. But sometimes data limitation.
Teh Hooi Ling used to run paper portfolios in the BT using some of those factors. I wasn't really into investing then. I vaguely recall the low PB one performing the best? But then you have academic research now saying low PB has lost its mojo in the last 20 years when using US or international stocks.
¯\_(ツ)_/¯
Wow ! Glad you caught the Dune reference !
ReplyDeleteI am not naive about back-testing.
ReplyDeleteIf it is that magical, the only successful professionals out there will be quants. If you do too much data snooping, you will end up with a number of factors that cannot explain how the returns come about. At best, it is a good way to compare strategies over time.
So far I find that it's a better way to explain the use of an investment concept compared to just pointing out that a company allegedly has a competitive advantage.
I wish I had a better tool.
Hi Christopher
ReplyDeleteI attended the Investors Exchange and wanted to ask you a few questions on your presentation. One strategy that you shared is to buy 15 STI stocks with the lowest PE ratio. At rebalancing time, we need to keep the 15 stocks with the lowest PE ratio. Do we need to ensure the $ allocation is the same across all the stocks? Is the strategy considered growth strategy? Thanks.
Steven,
ReplyDeleteIf some of the stocks get ahead of themselves, you will need to sell them to maintain the consistent size of these stocks. I would exercise some caution as this can incur brokerage expenses, so you may wish to only rebalance if the stock is now 25% larger than it should ideally be.
Since P/E is a value metric, this is a value strategy rather than a growth strategy.
Regards