Remove ads

Working out the Yield

General discussions about equity high-yield income strategies
pds2008
Posts: 33
Joined: May 19th, 2017, 2:24 pm
Has thanked: 15 times
Been thanked: 13 times

Working out the Yield

#79897

Postby pds2008 » September 8th, 2017, 12:24 pm

Hello

I am having difficulties establishing a common dividend yield figure when looking at various indices.

Looking it BHP Billiton just now, Morningstar has the yield figure at 3.03%, Hargreaves Lansdown has 4.6% and DigitalLook has 4.36%.

Is there a simple explanation and can some kind soul advise on a foolproof (oops) approach to establishing an accurate yield

Many thanks

P

Itsallaguess
Lemon Quarter
Posts: 1427
Joined: November 4th, 2016, 1:16 pm
Has thanked: 166 times
Been thanked: 562 times

Re: Working out the Yield

#79910

Postby Itsallaguess » September 8th, 2017, 1:00 pm

pds2008 wrote:
I am having difficulties establishing a common dividend yield figure when looking at various indices.

Looking it BHP Billiton just now, Morningstar has the yield figure at 3.03%, Hargreaves Lansdown has 4.6% and DigitalLook has 4.36%.

Is there a simple explanation and can some kind soul advise on a foolproof (oops) approach to establishing an accurate yield


It depends.....

In terms of a 'simple explanation', I would say that if you agree with the view that the company will at least continue to pay out dividends at the same rate as it's last yearly distribution, then the 'cleanest' way to manually work out a useful 'Yield' figure would be to find the last 12-months dividend-payments (either paid or declared in recent RNS result releases....), and divide that cash distribution total into the current share-price, to give what's often called a 'running-yield'.

The advantage of using a 'running yield' is that you're personally making no major company-related forecasts over and above the simple view that you expect the next 12-months dividend amounts to be at least the same as the previous 12-months distribution.

In the absence of any news that might cause that view to be overly-optimistic, then generally, and certainly at a portfolio-level, this would be considered a fairly conservative position to take regarding a view on yields.

For a worked-example, let's take Company A, that has either paid or declared (via RNS result statements) the following quarterly dividends over the past 12-months -

Company A - Past 12-months dividend distributions -

Quarterly Dividend 1 - Paid - 25p
Quarterly Dividend 2 - Paid - 25p
Quarterly Dividend 3 - Paid - 25p
Quarterly Dividend 4 - (Declared in last results RNS, will be paid in future) - 25p

Total dividend distribution (paid or declared) over the past 12-months = 100p

Current share price = 1850p

'Running Yield' = 100p / 1850p = 5.4%


Cheers,

Itsallaguess

moorfield
Lemon Slice
Posts: 780
Joined: November 7th, 2016, 1:56 pm
Has thanked: 52 times
Been thanked: 126 times

Re: Working out the Yield

#79925

Postby moorfield » September 8th, 2017, 1:28 pm

I use either (i) last (interim + final) dividends paid or (ii) last dividend paid and next dividend announced. (In some years that may differ slightly from the dividend received in my account due to the timing of top ups, however I use this method to be consistent with how running yields are often reported. Special dividends are excluded.)

BrummieDave
2 Lemon pips
Posts: 204
Joined: November 6th, 2016, 7:29 pm
Has thanked: 7 times
Been thanked: 39 times

Re: Working out the Yield

#79950

Postby BrummieDave » September 8th, 2017, 2:28 pm

I find this area confusing too.

If you take something that is specifically aimed at income seekers such as SPDR S&P Global Dividend Aristocrats UCITS ETF, the SPDR website shows two figures for different yields:

A distribution yield of 3.32% (defined as the 12 month historical dividend pay-out per share divided by the NAV)

An Index Dividend Yield of 4.55% (defined as the weighted average of gross dividend Yield of stocks, in the index)

HL's website just shows the latter of the two.

If £1,000 was invested today, what income would an investor expect?

SalvorHardin
2 Lemon pips
Posts: 212
Joined: November 4th, 2016, 10:32 am
Has thanked: 352 times
Been thanked: 89 times

Re: Working out the Yield

#79967

Postby SalvorHardin » September 8th, 2017, 3:26 pm

A major problem arises when you have companies, like BHP Billiton, who declare dividends in a currency other than sterling but then pay UK investors in sterling. When there has been a substantial movement in the foreign exchange rate between the payment date of that dividend and the present day then the historical yield figure can look out of very different to the published yield at most sources.

For example, an annual dividend of $1.00 on a share price of 3,500p at an exchange rate of $1.50 per £1 is a yield of 1.9%. Now if sterling falls to $1.30 (and everything else remains unchanged) the yield rises to 2.2%.

Some data sites calculate dividends differently, which introduces more differences.
The most common seems to be to use the most recent year's worth of dividends (historical yield). But some take account of forecast changes to the dividend (forward yield). And some treat special one-off dividends as regular dividends because their software cannot discriminate between the two types of dividend.

I go to the source (company dividend records) and use the historical dividend data. For foreign dividends I calculate the yield using the current exchange rate. My one concession to using the forward yield is where the company has announced a change in future dividends (e.g. American companies which increase a quarter's dividend will apply the same increase to all four quarterly dividends).

My spreadsheet for companies paying foreign currency dividends which are then converted to sterling shows two yields, one using the historical payments in sterling and another assuming that it is paid in the foreign currency. When sterling has been volatile, as over the past fifteen months, the two figures can be quite different.

DiamondEcho
Lemon Quarter
Posts: 1940
Joined: November 4th, 2016, 3:39 pm
Has thanked: 1239 times
Been thanked: 232 times

Re: Working out the Yield

#80015

Postby DiamondEcho » September 8th, 2017, 6:48 pm

Anything x-currency I assume most sources might miscalculate it, and so check and if necessary correct it myself in spreadsheets. Many sources of div% info simply cannot deal with the likes of Special-divs etc., so be aware of that too.
I also work on a '12month trailing yield basis' as an assumption going forward. PLCs are loath to cut, so that seems a reasonable base-case for the future. I prefer unexpected upside than downside - this conservative approach seeks to tie in with that
If FX rates on x-currency divs have really swung in the interim [3-6-9 months etc] I might also go and see how that impacts things.
I plan for the year ahead on the past year actuals [base-line], and major shifts in any FX rates. Only once or twice have I chopped my own personal forecasts where I'm additionally sure further trouble lies ahead, and I want to rein in my expectations prior to the actualitee.

staffordian
Lemon Slice
Posts: 594
Joined: November 4th, 2016, 4:20 pm
Has thanked: 232 times
Been thanked: 75 times

Re: Working out the Yield

#80016

Postby staffordian » September 8th, 2017, 6:58 pm

Also worth mentioning that some sources quote historical yields, such as itsallaguess detailed in the first reply, and some quote forecast yields, which are, of course, largely speculation.

Itsallaguess
Lemon Quarter
Posts: 1427
Joined: November 4th, 2016, 1:16 pm
Has thanked: 166 times
Been thanked: 562 times

Re: Working out the Yield

#80019

Postby Itsallaguess » September 8th, 2017, 7:45 pm

staffordian wrote:
Also worth mentioning that some sources quote historical yields, such as itsallaguess detailed in the first reply, and some quote forecast yields, which are, of course, largely speculation.


I don't think they are 'largely speculation' at all, on the whole.

I think the vast majority of larger-cap companies provide forward-guidance in terms of earnings, profit, and dividends, and it's that forward-guidance that informs most of the forecast-yields that we see mentioned on these boards.

Of course, companies stumble sometimes, and we've certainly seen examples of this in recent months, and the market can be quite unforgiving in those circumstances, when forward-guidance turns out to have been undelivered for one reason or another, but I think on the whole, 'forecast yields' have a place in future-dividend estimates, especially at an 'overall portfolio' level.

So I think 'speculation' is too strong a term, but it's right that we should understand that they are based on future-guidance, which isn't written in stone by any means, but is at least originating from the dividend-paying companies, and is fairly regularly updated into the brokers providing these 'forecast yields'...

Let's not forget that even using 'running-yield' is making an assumption that the previous 12-month dividend-distribution will at least be maintained....

That's not much further from 'speculation' than taking heed of guidance coming from the companies via the brokers really, if you think about it....

Cheers,

Itsallaguess

Gengulphus
Lemon Quarter
Posts: 1573
Joined: November 4th, 2016, 1:17 am
Been thanked: 552 times

Re: Working out the Yield

#80070

Postby Gengulphus » September 9th, 2017, 9:53 am

pds2008 wrote:I am having difficulties establishing a common dividend yield figure when looking at various indices.

Looking it BHP Billiton just now, Morningstar has the yield figure at 3.03%, Hargreaves Lansdown has 4.6% and DigitalLook has 4.36%.

Is there a simple explanation ...

Only the rather useless "it's because different people use different methods of calculating the yield". Explanations of their different methods depend on various subtleties and so cannot be entirely simple...

pds2008 wrote:... and can some kind soul advise on a foolproof (oops) approach to establishing an accurate yield

Depends what you mean by "accurate". If you mean a reasonably precise figure that everyone would agree is the right figure to use, no. Sources won't always agree with themselves - if I look at the figures in http://www.digitallook.com/equity/BHP_Billiton right now, I see 4.5% and 4.4% (across the top row of the table under "Key Financials"), 4.5%, 3.7%, 6.3%, 2.4% and 5.4% (in the right column of the table under "Key Fundamentals") and 3.9% (in the right column of the table under "forecast")!

If you mean a reasonably precise figure of the type you want to use, then yes - but you have to decide what type you want to use! The basic method is simply to get the relevant dividend figure and share price, divide the dividend by the share price and multiply by 100 to convert to a percentage. Alternatively, you can try to find a site that calculates the type you want, but it can be difficult to know definitely and precisely what calculations a site does. So I would tend to calculate the yield myself and compare with a site's quoted yields for a good variety of shares before deciding that they're what I want. And even then, I'll generally only trust a site's yields for filtering purposes and subsequent narrowing down of the filter results to a purchase candidate - I won't actually buy that candidate until I've checked its yield with my own calculation.

The problems about deciding what type of yield you want to use mainly lie in what the relevant dividend figure is, specifically what year's worth of dividends you use and whether you exclude any dividends. The main choices for what year's worth of dividends to use are:

* The historical dividend, which is the total dividends declared by the company for its last reported-on financial year. Strong points are that the figures are almost completely certain and (IMHO) the best possible indication of what the directors intended to do with the dividend. The main problem is that it can be seriously out-of-date - typically by up to a year plus however many months it takes the company after a year end to report on that year. Other problems include occasional changes of financial year end by a company, resulting in e.g. a 9-month or 15-month "year" (which needs the dividend to be adjusted to be 12 months' worth), corporate actions such as rights issues and share splits/consolidations (which need adjustments for post-action shares being different from pre-action shares), and companies that declare their dividends in a foreign currency and only produce a sterling equivalent at a later date (which need you to translate the declared dividend in the meantime using a best guess at the applicable exchange rate - usually just the current exchange rate).

* The rolling historical dividend, which is the total dividends declared by the company for the last year, whether or not they're for the same company financial year - so they are typically either the last interim and the last final dividend declared, or the last four quarterly dividends declared, in either case regardless of whether they're for the same company financial year. It has much the same strengths and problems as the historical dividend, though the out-of-dateness is less - typically only up to 6 months plus the reporting delay - and there are a few more potential problems caused by companies changing between quarterly and twice-yearly dividends, by companies changing their interim:final dividend split, and by companies deciding late in a financial year that they need to cut their dividend (see my recent post viewtopic.php?f=15&t=7253&p=79331#p79331 for an example).

* Various forecast dividends - various of them because there's one for each analyst producing forecasts for the company, plus one for each provider producing consensus forecasts (typically averages of analyst forecasts, possibly after some adjustments such as excluding analysts who haven't produced a forecast since the last results, or excluding the highest and lowest forecasts to try to prevent 'outlier' forecasts having an undue influence on the final figure). Their strength is that they are attempts to produce what one is really interested in (*), namely the dividends one will get in the future; their main problem is the fact that they are attempts - individual analysts most certainly don't always get it even approximately right, and while consensus figures are better, they're still not all that good. Other weaknesses include a lack of availability of many forecasts for regulatory reasons and/or the fact that their providers want to charge more for them than one wants to pay, and (especially for the free forecasts such as DigitalLook's consensus forecasts) a certain amount of out-of-dateness due to the forecasts not being updated promptly for recent news - it's quite noticeable that they tend to take weeks-to-a-few-months to adjust fully to major news.

* As a special case of the last, dividends one forecasts oneself. Hard work in general, and prohibitively so when done for more than a handful of shares - but when a company announces a change of dividend policy, it's often the only way to get a realistic figure while the forecasts catch up, and sometimes really quite easy. For example, I've had no difficulty at all in forecasting Carillion's 2017 dividend as 0p since their trading statement a couple of months ago! (And while DigitalLook's consensus forecast has now largely caught up with that news, it's still 1.05p rather than 0p...)

The other issue is which dividends to include. The easiest case is ordinary dividends, which should be included. The next easiest case is special dividends that are accompanied by share consolidations: they should IMHO almost never be included, the reason being that (apart from tax considerations for non-tax-sheltered holdings) they are almost always equivalent to a sale. I.e. look at the cash produced by the special dividend (possibly plus a 'fractional entitlement' payment from the consolidation) and divide it by the number of shares lost to the consolidation, and you'll almost always get something reasonably close to the current market price - so in effect (again, apart from tax if relevant), the company has bought those shares off you rather than paid you a dividend.

That leaves special dividends that are not accompanied by share consolidations. For those, I reckon a judgement call is needed: is the special dividend one you expect the company to produce regularly in the future? If so, it's relevant to what you're really interested in (*), i.e. the future dividend income you expect from your investment, and so can reasonably be included; if not, then it's not very relevant to that and so is IMHO best not included. I should add that in my experience, it used to be that companies called dividends "ordinary" if they did expect to repeat them regularly and "special" if they didn't, making for a very simple "include ordinaries, don't include specials" guideline. But that has been messed up in recent years by an increasing number of companies calling dividends that they do expect to repeat "special" (I suspect for 'spin' purposes - they want to be able to cut their to-be-expected dividend payments while still being able to claim that they didn't cut their ordinary dividend...). Annoying, as one can no longer take companies' descriptions of their dividends at face value, and I've taken to calling such dividends "not-so-special" dividends...

One other complication is that one very occasionally comes across payments that are clearly intended to be equivalent to ordinary dividends, but aren't actually dividends. An example is Rolls-Royce's regular payments via issuing redeemable C shares, that shareholders can redeem for 0.1p each (including being able to set up a standing instruction to redeem all such shares they get, so that all the shareholder really sees is the cash arriving, as for a dividend). They do that for Corporation Tax reasons, and the only real consequence for shareholders is that the payments are taxed by CGT rather than Income Tax (which makes no difference at all if the shares are held in tax shelters). I would include such payments as though they were dividends, even though technically they are not.

Anyway, you don't necessarily have to follow my recommendations in the above, and even if you do, there is room for differences of opinion about which special dividends can be expected to be repeated regularly. So you can decide between various somewhat-differing versions of historical and rolling historical dividends, and also of dividends you forecast for yourself. And as regards dividends forecast by others, you're dependent on which decisions they make about what to include...

The other figure to decide upon is the share price to use when calculating the yield. If you're going to use the result to make a buying or selling decision, the answer is simple: the price at which you can buy or sell. So for a buy or sell you're intending to do immediately once you've made the decision, the current market price (or to be precise, the market price when you actually do the buy or sell, but the best you can really do in practice is use the current market price for the decision, then quickly check it hasn't moved significantly against you at the point of actually trading).

But if you're planning to use a limit order, then your intended limit price. E.g. if the dividend figure for a share is 10p, you could reasonably enter a limit order at 200p and say that you intend to buy at a yield of 10p/200p = 5%. And other prices may be relevant for other purposes - e.g. the price at which you bought if you're trying to analyse your past buying decisions to see how well you've been making them. But in general, be clear with yourself about what your purpose is for using a price other than the market price and that the resulting figure is what the yield was or hopefully will be, not what it is now. (And if you post such figures, be very clear about those points in your post - both to avoid unintentionally misleading others and to avoid provoking disputes...)

To sum this up, basically your question isn't a simple one - there are quite a lot of choices about exactly what you want, and they're all less than ideal in various different ways... Sorry!

(*) These two statements about what you're really interested in do assume that your purpose is to make a buying or selling decision, so they aren't necessarily applicable if you have some other purpose.

Gengulphus


Return to “High Yield Shares & Strategies - general”

Who is online

Users browsing this forum: No registered users and 1 guest