Smautf wrote:I confess, various assessments of BHP's yield have been puzzling me lately.
BHP Group paid a special dividend of almost 80p at the end of January, arising from the sale of their US shale business. Last year's final dividend (paid 25.9.18) was 48.7p, and the current year's interim dividend (payable on 26.3.19) will be 42.25p.
That's a total of (roughly) £1.70 in the year to 30.6.19 and a yield of 9.67% on today's price of £17.58. Strip the special out and the yield is closer to 5.1%.
So I have no idea where either Pyad or ShareCast are getting their numbers from ! The same, by the way, goes for HYPTUSS which - the last time I ran it - suggested a forward yield of 8.1%.
... Does anyone have an explanation ?
Thanks - your reminder of that special and wondering where ShareCast are getting their numbers from has suggested an explanation to me. I can't prove my explanation is correct, but based on what I've seen about forecast dividends and yields in the past, I think it quite likely - it is:
I know that data providers generally give consensus forecasts - i.e. some sort of average they've formed from the forecasts made by individual analysts' forecasts. Which forecasts they use may vary between data providers - for instance, they might have different lists of brokers/analysts that they pay attention to, and they might ignore a forecast for being out of date if too long has elapsed since it was produced (e.g.
ShareCast's list of brokers they take into account for their broker view statistics for BHP might not be exactly the same as they use for forecast figures, but it's probably at least similar). And what sort of averaging they use may also vary - for instance, they might use the normal average (= arithmetic mean in mathematical terms), or they might use the median to try to avoid the figure being influenced unduly by an outlier forecast produced by an analyst who is much more optimistic (or much more pessimistic) than the rest about the company's prospects, or they might use other techniques to achieve the same aim - e.g. one I saw described in detail many years ago was IIRC "if there are three or more forecasts, discard the highest and the lowest, otherwise keep them all; then take the arithmetic mean of the forecasts that have been kept".
Such averaging systems are basically always capable of producing a 'consensus' figure that no individual forecaster remotely agrees with. As an extreme example, if a company looks as though it's in trouble and exactly half the forecasters think it will bite the bullet and cancel its dividend entirely, while the other half think it will tough it out and maintain its dividend for at least one more year, then all three of the systems I've described will reckon the 'consensus' is that it will cut its dividend by 50% - when not a single forecaster is forecasting anything of the sort! And both differences in the forecasts used and in the averaging method can make quite significant differences to the figure obtained - for instances, if there are ten forecasts and seven forecast cancellation while three forecast toughing it out, then the normal average gives a 70% cut, the median a 100% cut and the discard-highest-and-lowest-then-average method a 75% cut, but another provider who reckons that five of the cancellation forecasts and none of the toughing-out forecasts are out of date gets 'consensus' forecasts of 40%, 0% or 33% cuts respectively...
What I suspect might well be going on here is that some of the forecasters are counting the special and so getting a dividend of about 170p for the current financial year, while others are not and so getting a dividend of about 90p for the current financial year. As long as there are significant numbers of forecasters in both of those categories, the same sort of effect as I've described would end up producing 'consensus' figures between the two, such as ShareCast's 143.5p or pyad's data provider's 158.8p
and it would make the fact that they differ markedly very understandable if they're different data providers.
Smautf wrote:I don't recall seeing a commitment to hugely increased payouts from the company although I might easily have missed something. ...
...
NB Next year, I assume there will be no special and that the payments will be closer to the 90p total of 2018 final and 2019 interim - which means ShareCast's forecast of 96p for the year to 30.6.20 feels more reliable than their forecast for the year we're actually in !
That describes my position as well.
Gengulphus