moorfield wrote:Example3: I see dividenddata already has marked down GSK's dividend to 55p, however I have it on a yield of 4.9% today which recognizes that one can still buy the last two dividends due to be paid next January and April for its current financial year before the corporate split happens: ie. 19p + 23p + (19p + 19p)*55/80 = 68.125p (we do not know the new dividend timetables of course but for now one might apportion the aggregate amount quarterly thus). That figure will continue to change again, of course.
I agree that that's a reasonable estimate of what the next 12 months income will be, but as you say it will continue to change - and it seems highly likely that the change will be downwards as quarterly payments made at the old level of dividends disappear from your sum, to be replaced by quarterly payments at the new, lower level. The way I would suggest treating yields in cases like GSK when there's a pretty-certain level of dividend for the next few payments and a significantly different, lower starting level for payments after that is to regard the excess of what you're expecting over the next 12 months over what it would be if the new level of dividends were starting immediately as a 'cashback discount' on the share price and calculate the yield as the new starting level of the dividend divided by the share price minus that discount.
So for GSK at its current price of 1388.2p, if the new starting level of dividends of 55p indicated by the company were starting immediately, you would be expecting 55p per share over the next 12 months, but you're actually expecting 68.125p. So you're expecting to be 68.125p-55p = 13.125p per share better off than you would be if the new starting level were starting immediately, with that 13.125p 'bonus' being one-off and expected in the next six months. So the
effective price for GSK shares paying at that new level starting now is the actual market price minus that 'bonus' that you can pretty confidently expect to get in the short term, i.e. 1388.2p-13.125p = 1375.075p. And the yield of those shares would clearly be 55p/1375.075p = 4.00%.
Basically, calculating it as the next 12 months income divided by the current share price (68.125p/1388.2p = 4.91%) is looking at the future income through distinctly rose-tinted glasses, because that level of 'next 12 months' income is highly unlikely to be sustained beyond the short term - it's pretty certain to drop to about 23p + (19p + 19p + 19p)*55/80 = 62.1875p in the next three months and about 55p in the three months that follow them. And more generally, if you have some sort of general idea of the pattern of dividends in the future, of the "will start at 55p and no current strong reason to believe they'll subsequently behave very differently to dividends in general" type, but some very-likely-known short-term deviations from them, treat those deviations as discounts or excesses on the share price used in the yield calculation according to whether they're positive or negative deviations.
Or alternatively, just look at whether the deviations from the expected long-term pattern are big enough to be worth taking into account at all. In the case of GSK, the short-term, one-off deviations totalling 13.125p from expecting dividends to go to their new starting level of 55p immediately are very roughly 1% of the share price, which is reflected in the fact that if I use the new starting level and don't discount the share price by that deviation, my yield calculation will be 55p/1388.2p = 3.96%, about 1% lower than the 4.00% calculated above taking the deviation into account. Does that sort of difference in yield actually matter to you? I can't judge how individual HYPers will answer that question - but those for whom the answer is "No" might as well just calculate the yield using the new starting level immediately, ignoring the short-term deviation from that caused by the new level not starting immediately.
Put another way, one should expect the bulk of the dividends produced by a HYP share to be produced in the long term and only a small fraction of them in the short term - so oddities in their pattern that
only affect them in the short term need to be pretty large to have a significant effect on the desirability of the share.
Gengulphus