Interdum stultus opportuna loquitur...

Wednesday, June 29, 2005

BozoRant: My Error, My Fault, Mea Culpa

Note - from June 24th 2009, this blog has migrated from Blogger to a self-hosted version. Click here to go straight there.

There are three statistics-related aphorisms that I particularly love - but my favourite is an anonymous quotation referred to by Sir Josiah Stamp (who also generated some other aphorisms that I like).

The Stats-related one is as follows:

"The government is very keen on amassing statistics. They collect them, add them, raise them to the nth power, take the cube root and prepare wonderful diagrams. But you must never forget that every one of these figures comes in the first instance from the village watchman, who just puts down what he damn well pleases."

The flummery of statistical compilation aside, there is also a decent amount of blockheadedness applied to the numbers once they are compiled. I performed one such act of blithering blockheadedness the other day, when I allowed the following to slip through:

Speaking of dividends, the yield on the ASX200 is now below 2% - and has been for several months. I don't know about you, but paying $1 for a 2c dividend (franked, on average, 76%) strikes me as idiotic. You get better cash extraction from a kiddie's Dollar-mite account (do they still have those?).

I note with mirth that the HTML tag used to identify indented quotations is <blockquote>; <blockheadquote> would be more appropriate in this instance.

Firstly, there's a typo (as I alluded to in a response to a comment posted yesterday by Sam Jones); it should have read ASX20 rather than ASX200. I wrote a long-winded explanation of my methodology last night, and I was immediately struck by what looked like a cockup (which I pointed out at the time)... namely, to reconcile the numbers I had for the ASX100 yield and the ASX200 yield, the 'next 100' must have yielded over 20%... patently ludicrous.

Now, I've never claimed infallibility. I do claim to be right a lot more often than I'm wrong; the bigger the issue, the more right I think I am - simply because I think about it harder... call it conceit, but I think that when I think hard about things, I tend to be able to get them pretty right. Rest assured though - if I am ever wrong I will always say so.

Anyhow - in my response to Sam, you can see that I had a suspicion that I had 'laid a grossie'.

So then I started checking calculations, and bugger me if I didn't find the problem. And bugger me twice if it wasn't all what we used to call a FAH problem (hover your mouse over the bold text to see what that means).

it starts with a firmly-held principle of calculation, then descends into farcical (short--lived) incompetence, for which I apologise.

First - the firmly-held principle of calculation.

I am not a believer in using 'composites' to generate index-wide numbers. So for example I tend to use 'source' numbers (EPS, DPS, shares on issue and last price) to generate numbers like Market Cap and aggregate PE for an index, rather than just using a market-cap weighting of individual stock yields and/or PE ratios. My justification for this is a nervousness about accounting for the price that appears in the PER and Yield calculations for individual stocks; if the stock's share price has move substantially and the PER calculation doesn't capture that properly, the aggregate PE will be biased.

In this case, the belts-and-braces approach turned and bit me on my flabby white arse.

In my haste, I had based all of my calculations except the one for the ASX200, on the Final dividend payments for each stock, rather than the Total dividends paid. In other words, I had left out all the Interim dividends. Thankfully I had not performed the calculation for the ASX200 at that stage (because I had not checked all of the data for the 'next 100), and when I did it I discovered that the numbers simply didn't 'hang together'.

So I withdraw the claim that the ASX200 (or oven the ASX20) yield is below 2%. It is, in fact, a hair under 3.3% (which is still WAY too low, when bonds are yielding 4.5% and are all overpriced). I have fixed the algorithm that does the index PER/Yield calculation, and this stuffup will never happen again.

Everything else I said in the reply to Sam was valid though - the presence of mobs of Property Trusts in the 'next 100' is what raises the yield on that vehicle relative to the AX20 (which includes non-yielders like New Corpse).

Thanks again to Sam for the questions that pointed me on the road to fixing the stuffup. As I said above, I never claim to be infallible, and so if EVER anything strikes you as probably wrong, for Odin's sake tell me so that I can set the record straight if needs be.