Saturday, December 22, 2007

Of estimates and consensus thinking

I attended an investment session where the instructor asked the class (of around 20 pple) to estimate the size of Thailand vs Singapore. Was it 50x bigger? Or 100x bigger? Or 500x or what?

He wanted to prove a point. The true answer will lie in the range of everybody's estimate. Bcos someone was bound to get it right. Well his point was quite valid, in the end, the answer did lie within the range of everyone's estimate.

But what was more striking to me was that most estimates are wrong and some VERY WRONG. For those dying to know how big is Thailand vs Singapore, well it's actually 73x. The closest estimate was 50x. And only one guy got that close. Some had it 10,000x. My estimate was 400x.

This made me think very deeply about the nature of estimates. And more specifically, estimates of future earnings of listed companies. We know that the sell-side or brokers have their army of analysts to forecast listed companies' earnings for next yr, or 2 yrs out. Maybe, just maybe the analysts' estimates on a listed company's EPS that we, and most investors rely on, might usually be wrong as well. And it's logical that they should be wrong. Bcos estimates, by virtue that they are estimates, are usually wrong!

Of course, you may argue that analysts have access to information since they get to talk to industry people, competitors, company management etc. Well the analogy with Thailand vs Singapore may not be quite right today, since we have Google and Wikipedia.

But imagine if it were the Stone Age and the class was given 1 yr to walk Thailand and Singapore and come up with an estimate, how likely is it for the class to get it right? Probably as likely as the analysts to get next yr's EPS right, right? Which implies that estimates based on some info but INCOMPLETE info is not much help and that's the way it should be.

So consensus thinking and crowd thinking, by logically extending the argument, can actually be usually wrong. This can be quite scary bcos most of us (well some of us) usually follow others' action thinking that they did their homework so we are safe. E.g. I will go for a stall with a respectable queue in front of the shop at an unfamiliar hawker centre. As for financial markets, there is this thinking that even if we are wrong, so would most others and so it shouldn't be that bad.

Now based on the recent poll, I guess most pple would agree that Singtel is a bad investment since most pple thought that Singtel gave back 0% return since IPO. But guess what, the actual answer is more than 44% return since IPO, which is at least 3%pa based on the price of Singtel when the poll started (around S$3.60). Bcos Singtel gave back lots of dividend and capital back to shareholders during the 15 years it was listed. Since then, Singtel reached a new high of S$4.00 or so. That's another 10%. So again, most people are wrong. Ok you may argue 3%pa is not very attractive, esp after putting your money there for 15 yrs. Well its better than fixed D, and the point here is actually estimates are usually wrong, just a reminder.

Also it's a mere 15 years since Singtel IPOed. Statistically, it's not really that significant yet. Yes in order to be of statistically significant, the track record has to be 18 yrs or more! If you hold on to Singtel for the next 3 yrs or more, maybe the annual return will converge 8%pa or something.

So I guess the moral of the story here is this: Don't trust what most people do, they are usually wrong. Do your own homework and come up with the logical conclusion. Or you can visit this blog (which tries to post accurate logical conclusion on most stuff) more often.


  1. Hi Jay,

    I agree with you that most analyst get their estimates wrong else we would be seeing a lot of very rich Analysts by now.

    However, it is also very difficult for the average investor like me to verify the Analysts' reports. They are the experts with all the tools and access but if their information is 'incomplete, where else can we verify their claims? Of course, certain information like past history can be verified because it is recorded but future plans in particular is difficult to verify.

    Hence, like your analogy on people buying from the stall with the longest queue in an unfamiliar hawker centre, the same goes to unfamiliar stocks. Most people will usually go for stocks with the most 'buy' call.


  2. Hi Derek, it is true that most retail investors are not in a position to 2nd guess sell-side analysts, given their limited access to a lot of information about co.s.

    I guess that message here is that one should try to exercise independent thinking and not take things at face value.

    This is actually a lesson for myself as well with my journey to learn about insurance.

    For years, I did not want to learn about it, and left it to friends and relatives who were agents. I thought they would help do a good job for me.

    But in the end, it has to be my own independent thinking that does the job. To do what`s best for me.