Luniversal wrote:Not so sure about 'plenty of choice' these days. ...
...
For illustration, here is a 'dividend aristocrats' HYP which I posted on TMF in Nov. 2015. It is made up of 15 shares whose divis had each grown by 10% pa or more compound (i.e. c. 7% real) ever since 2000. None had ever reduced income in nominal amount between financial years. They were well prized by punters, but were yielding enough for a portfolio average yield of 3.3%: one-tenth below the All-Share.
The idea was to play fast catch-up: the average payout growth was 14.9% pa (which already sounds nostalgic when Imps is regarded as remarkable for committing to 10%). An eternity investor ought to bear the first few years' subpar payout for the sake of the juiciness to come.
...
... [Portfolio list of shares omitted for brevity - it's available in my table below] ...
...
I wrote then: "Nobody can say whether or when any income portfolio would begin to act up. But if these did, the holder would be entitled to grumble loudly."
Well, grumble away... because less than four years later, half this lot are in the doldrums or worse.
Vodafone and Mitie are cutters, Capita and Cobham are off the list and SSE's payout looks threatened for all those protestations of eternal fidelity and priority for income. Growth has declined to a trickle at BATS and RPS. WPP is on a freeze, maybe with worse to follow. Others are doing okay, but none is exactly bounding away to cover for the defaulters.
I wouldn't rate this HYP's chances of good future dividends from RPC highly either ;
-), for reasons that will become apparent below...
More seriously: this says essentially nothing about whether there is plenty of choice these days. It's quite easy for someone to claim that passing a particular test makes a share a 'dividend aristocrat' (or whatever other phrase they feel puts a good spin on it and/or have picked up from someone else) and select a HYP on the basis of that criterion. That's what you did when you selected this HYP (*) in 2015, and what the grumbling should be about is that the criterion used turned out to be a poor one for HYP share selection. Maybe the amount of choice for a HYP constructed along those lines has dropped since 2015, but to find out whether that's the case requires one to look at the shares that meet that criterion now, not the ones that were available back then. And even if it has dropped, that just means that as a HYP share selection criterion that allows plenty of choice, it has the weakness of sometimes being too restrictive.
But
why has it turned out to be a poor one? Just 15 companies selected on just one investment date is a decidedly small sample, so don't expect any firm conclusions from what follows - but examining the companies' and portfolio's dividend histories for weaknesses might suggest some possible answers for further investigation. So I've put the company dividend histories together - first the histories from 2000 to 2015, which should be pretty similar to the histories on which you selected the portfolio. Almost certainly not identical to them, as I don't know the exact date that you selected the portfolio, and even if I did, there are various ways that the dividend histories can be assigned to years: by company accounting year, by calendar year, by tax year and by portfolio year being the main possibilities, and the last three have the options of taking a dividend into account on its announcement date, its ex-dividend date or its payment date - see (**) below for the exact choices I've made for this table if interested. Also, when available, I've taken the figures from the
http://www.dividenddata.co.uk dividend history pages for the companies, since they not only bring the dividend histories together in a single place, but also do the adjustments for share splits/consolidations and rights issues for me - a big saving in effort for me and one without which I probably wouldn't have taken the task on. Unfortunately, they don't cover James Halstead or RPS, and neither of those companies' investor relations sites have dividend history pages that I can find, so I've had to assemble them and do the adjustments for myself - each of their columns took more effort than the other 13 companies combined! I've italicised those companies' figures to indicate the different source.
The fact that the average dividend CAGR between 2000 and 2015 has come out as 14.5% compared with your 14.9% figure is not a problem - it's a small difference, and well within the scope of differences in the way we've assembled the dividend histories. Rather, it's reassurance that I'm seeing basically the same picture as you did in 2015, just viewed from not quite the same place.
If however I look at the details of that picture, it's very noticeable that the average CAGR of the shares in 2015 drops off quite a bit as one goes from looking at the previous 15 years to the previous 10 years and then the previous 5 years, from 14.5% to 12.9% to 10.5%. And looking at the individual shares, Vodafone shows a particularly big drop, from 15.0% to 6.6% to 5.2% - and the reason is not hard to find in its detailed history: it nearly tripled its dividend between 2003 and 2005, and had mostly been a decent-to-good but not spectacular dividend grower otherwise. In short, it basically earned its 'dividend aristocrat' status on 10-year-old past glories and not on anything more current. It's the most extreme example of that in terms of how long ago the past glories occurred, but all of Babcock, British American Tobacco, Capita and Mitie show similar evidence of earning their 'dividend aristocrat' status on 5+-year-old past glories. And Diageo and SSE are complete pretenders to the status as far as I can see - your sentence "
It is made up of 15 shares whose divis had each grown by 10% pa or more compound (i.e. c. 7% real) ever since 2000." should really be "
It is made up of 13 shares whose divis had each grown by 10% pa or more compound (i.e. c. 7% real) ever since 2000, plus two others I felt belonged in it."
In short, 'slowing dividend growth' rot was already in this HYP when you selected it! And as you've observed, it's become worse since:
Which reveals why I hold little hope of future dividends from RPC - the company has just been taken over... It was purely a cash takeover, with proceeds of 793p per share, was done by a scheme of arrangement that became effective on Monday and the shares were delisted on the same day. So holders should get the proceeds of 793p per share - they should arrive by Monday 15th according to the expected timetable. So if you want to keep this portfolio going as a 15-share HYP, you probably ought to be working out about now what replacement(s) you think it should buy for RPC.
I should note that:
* I've adjusted the italicised dividend figures for British American Tobacco from those in
http://www.dividenddata.com due to its switch to paying out quarterly, which resulted in it technically only paying an interim and a quarterly for 2017. I've adjusted by attributing Q1 dividends for 2018 and 2019 to 2017 and 2018 - which still makes the 2017 figure unreasonably low, but not as badly.
* I've similarly adjusted the italicised figures for Vodafone due to its switch to declaring dividends in euros and determining the exchange rate late and without announcement (I've adjusted by finding the actual sterling dividend figures from my own records, except for the "2018" (i.e. 2018-2019) final, which I've done at the current exchange rate because the actual rate isn't yet known - I've marked the figure as approximate as a result).
* Clearly the rot has become worse - though I should say that the average CAGR of a portfolio's constituent shares is not a precise guide to the CAGR of the portfolio as a whole, except in the case of a portfolio that is equally weighted at the start of a 1-year period, and the inaccuracy can become major if the individual share CAGRs vary greatly, the period is many years and/or the portfolio does not start the period equally weighted. As a simple example to illustrate the point, consider a hypothetical portfolio with two holdings, equally weighted (by dividend income since we're looking at dividend CAGRs) to produce £1000 income each year. One holding does very poorly, consistently cutting its dividend by 40% every year - it clearly has a dividend CAGR of -40%. The other holding does very well, consistently increasing its dividend by 20% each year - so it equally clearly has a dividend CAGR of +20%, and the average CAGR of the shares is (-40%+20%)/2 = -10%. So the portfolio initially has dividend income of £1000+£1000 = £2000, and one year later its dividend income is £600+£1200 = £1800, a 10% fall exactly as you might expect from a -10% CAGR. But two years later, the income is £360+£1440 = £1800, i.e. it hasn't fallen any further, and a 10% income drop over two years annualises to a CAGR of -5.1%; three years later, the income has grown from its low of £1800 in the previous two years to £216+£1728 = £1944, a portfolio dividend CAGR of -0.9%; four years later, the income becomes £2073.60+£129.60 = £2203.20, a portfolio dividend CAGR of 2.4%; and so it goes on - in the very long term, the portfolio dividend CAGR approaches 20%, the maximum of the individual shares' dividend CAGRs rather than their average.
* In this case, the CAGRs of the 14 individual holding CAGRs (excluding RPC because I simply don't have a sensible final dividend figure for it due to the takeover) vary very widely from -100.0% up to 10.3%. and that means that the incuuracy in treating their average CAGR as an indication of portfolio income performance is quite big, making the -12.8% average given in the table pretty big. Details depend on how the portfolio was weighted, but if it was equally weighted by income with each original holding having produced £1000 income in 2015, the portfolio income in 2018 would have been £13065, which implies a portfolio dividend CAGR of -2.3%. Not good, but a far cry from being as bad as -12.8%. So the -12.8% figure is pretty badly misleading... (And I did consider not calculating it at all, but decided that if I did, someone would - so I might as well and get the explanation why averaging CAGRs can be misleading in from the start!)
A very quick summary of the above is firstly, don't chase dividend growth: uncritical reliance on high dividend growth measures without looking carefully at how sustainable the growth is can be just as misleading as similar uncritical reliance of high dividend yields, even when the dividend growth measures are long-term ones like 15-year dividend CAGRs. And secondly, be careful about averaging: for some measures like CAGRs, you need to work out the measure for the portfolio as a whole if you want a reasonably correct answer for the portfolio, not do it for the individual shares and average the results (not that you have necessarily done that - but I think a health warning to others about your mention of the average CAGR of the 15 shares in this HYP is worth giving).
(*) Note that it definitely isn't a HYP according to
this board's guidance, not even when it was selected in 2015: James Halstead isn't and wasn't then in the FTSE 350, and containing just one non-qualifying share means that it breaks that guidance's "comprised exclusively" requirement. However, 2015 is before TLF and this board came into existence, let alone this board's guidance, and I'll presume that it was a HYP by the
TMF guidance - the only guidance around when it was selected. In particular, that guidance didn't require FTSE 350 membership, and it said that a small proportion of a HYP not qualifying didn't prevent it being a HYP. So just to be clear: this footnote is
not to say that your 'dividend aristocrats' HYP is off-topic for the board - it is to say that I think it's
on-topic and why.
(**) I've chosen to account for dividends in the table by company accounting year, but that raises the issue of which year to put that against when the company accounting year doesn't end on December 31st (which it does for 6 of the 15 companies). For the other year ends which occur, I've put the dividend against the starting year for the 5 companies with March accounting year ends (e.g. SSE's "2015" entry is for its accounting year from April 1st 2015 to March 31st 2016) and the ending year for the 2 companies with September accounting year ends (e.g. Sage's "2015" entry is for its accounting year from October 1st 2014 to September 30th 2015), basically to make the periods have 9 months in common with the calendar year rather than 3. For the 2 companies with June accounting year ends, that doesn't really resolve the issue, but I've chosen to put the dividend against the ending year to ensure I have figures for their "2018" entries.
Gengulphus