Building a Portfolio – Dividend Portfolio 3
Today’s post is about the qualitative factors that go into building a portfolio of dividend stocks. We’ll also take another look at screening for stocks.
Building a Portfolio
Last week we had our first look at the quantitative factors which underpin the selection of dividend stocks, and we tried – none too successfully – to use free stock screeners to identify good candidates.
Today we’ll have another go at screening, but first I want to talk about the quantitative factors that might help us to build a dividend portfolio.
As always, the key consideration in constructing a portfolio is diversification. Indeed, diversification is the reason portfolios exist.
- In the sense that all our dividends will come from stocks listed on the UK market, we will not be diversifying across asset classes or geographies.
- But in the sense that the portfolio will consist of around thirty stocks, we are diversifying away stock-specific risk.
We will diversify across industry sectors, since companies in the same industry are often hit by the same issues at the same time (look at energy and mining stocks at the moment).
There is also some scope for diversity of geography across the portfolio’s earnings base.
- many of the larger UK stocks have a good proportion of international earnings.
- the thinking here is the same – a local economic slowdown (or even a natural catastrophe) will not affect the whole portfolio if they earn their money across the globe.
We will try to start the portfolio with companies that have a market cap above £500M, so there is limited scope for diversifying across company size.
- this is because larger companies are safer – they are more likely to continue to pay a dividend after a setback
- if we struggle to put the portfolio together from companies of this size, we may have to drop down to £250M, or even to £100M.
The basic qualitative measure of a company is indeed its “quality”. Quality means different things to different people, but it usually revolves around the amount of earnings and the reliability of earnings.
- Within the context of a dividend portfolio, we can optional substitute dividends for earnings, so we might look for regular, growing dividends from high and stable earnings.
To remain entirely qualitative, we need to restrict ourselves to company quality ratings, such as whether they appear on lists of “dividend aristocrats”. ((Note that inclusion on these lists is usually based on analysis of underlying quantitative factors – the numbers ))
Such lists are more popular in the US than the UK, but there is a UK index (from S&P), which measures “the performance of the 30 highest dividend-yielding UK companies … that have followed a managed dividends policy of increasing or stable dividends for at least 10 consecutive years”.
For those dividend investors who insist on an unbroken track-record of dividends, such an index can be very useful.
SPDR run an ETF (LON:UKDV) that tracks the index and yields around 4% pa for an annual fee of 0.3%.
At the time of writing, the ten largest holdings were as shown below:
And the sector breakdown was as follows:
This is not strictly a passive ETF, since it selects the 30 highest-yielders with long unbroken track records – it’s a kind of smart beta.
Since it yields more than 3.5%, even after the 0.3% annual charge, it’s eligible for the dividend portfolio.
There’s also an ETF that tracks the European Dividend Aristocrats (LON:EUDI), but this only yields 3.16% (2.86% after charges), so that doesn’t qualify.
And SPDR also have an ETF that tracks emerging market dividend payers (LON:EMDV). This currently yields 6.19% (or 5.66% after charges) from 102 underlying stocks, so it also qualifies for the portfolio.
iShares also have a range of dividend ETFs: (IAPD, SEDY)
- LON:IUKD tracks 50 UK dividend stocks, and yields 5.69% (5.29% after charges) – eligible
- LON:IDVY tracks 30 Euro stocks, and yields 4.2% (3.8% after charges) – eligible
- LON:IAPD tracks 35 APAC stocks, and yields 5.96% (5.37% after charges – eligible
LON:SEDY tracks emerging markets but only yields 3.33% after charges. There’s also a US dividend tracker but its yield is too low to qualify.
So there are five dividend ETFs that we can look at – 2 UK and one each for Europe, emerging markets and APAC.
- Next time we’ll take a quick look at the investment trusts that might be suitable for our dividend / income portfolio.
After that aside, back to quality. A halfway house / gray area between qualitative and quantitative analysis would be counting.
- We could get this information directly from the company’s accounts, though we will look for better (more machine-readable) sources.
Other things we might count include:
- increase in sales vs previous year
- increase in profits
- increase in dividends
We have already compared our stocks to the market by including the long-term average yield of the UK market (3.5%) as one of our base criteria.
- We can take this idea further by comparing other company characteristics to the market.
Since we are comparing two numbers, this is not strictly qualitative, but the result it produces – “better” or “worse” – can be thought of that way. It’s another grey area.
Things we might compare (apart from yield) include:
Of course this then sets us the challenge of sourcing this market-average – or at least index-average -data.
This is intended to be a long-term growth portfolio rather than an actively traded one. So we will need to consider at least the medium term prospects of any stocks we plan to buy.
- Industries that have the potential to be disrupted within 5 to 10 years are not ideal candidates.
This is one of Warren Buffett’s staples – if you can’t understand a business (you can’t understand how it makes money) then you can’t invest in it.
- I think this applies much more to the Dividend Portfolio than say to our AIM portfolio, which is at least in part a trading account.
As well as the source of profits, you need to be able to understand the forces acting on the company’s industry sector, and how its strategy will protect it in the future.
- If you don’t feel comfortable with any of this, then it’s probably better to be safe than sorry.
Once we have our thirty companies / funds, we can mostly ignore them. The dividend will be announced twice a year with the interim and final results, and we can check that a company is still suitable.
But in addition to these checks, it’s probably a good idea to run the original stock screen (and qualitative tests) around once a year, to check that even better candidates haven’t emerged.
- It’s also possible that some stocks will have revalued (gone up in price) to the point where they form too large a proportion of the portfolio, or are now valued too highly to qualify.
- It’s usually best to sell at least some of the highly valued company and re-invest in something cheaper.
You might aim for a portfolio turnover of 20% – six stocks out of thirty.
- You could make all these changes at one point each year (say January), or if you are prepared to put in a bit more effort, you might stagger them through the year.
Portfolio turnover of 20% gives an average holding period for each stock of five years.
- With a long-term dividend portfolio you probably don’t want to drop too far below this.
Last week we looked at the free screeners from Google and the FT. The Morningstar screener wouldn’t play ball, and that’s still the case today.
- So I’ve dropped it from my list of free screeners, and replaced it with the Hargreaves Lansdown Screener.
- Screening uses a visual histogram of the stock universe (as per the Google screener) and screens can be saved to be used again in the future.
The HL screener has 7,778 stocks in its database, but that includes international stocks. When I select for UK only, this drops to 1,702 stocks.
- Setting Market cap to minimum £500M, yield to minimum 3.5% and PE to maximum 19 gets us down to 87 stocks, which is in line with the results from the FT and Google.
Here’s the basic screen:
To drill down any further, we need to switch to the Advanced screener.
- Price to Sales and Price to Book were both there, and setting each of them to a maximum of 3 got us down to 57 stocks.
Of the quality screens, Profit Growth was there – I set this to minimum 5% pa.
- I had to use Operating Margin as a proxy for ROCE – I set this to minimum 10%.
None of the other quality screens were present, but it was a better story on dividends:
- I set Dividend per share growth to minimum 5% pa
- Dividend cover was set to 1.5
- Forecast dividend growth and years of dividends were missing
Here’s how the HL screener matches up to the other two free screeners:
- The HL screener is slightly worse overall, but slightly better as a dividend stocks screener.
Here are the final settings for the HL advanced screener:
And here are the seven results that it produced:
Screening Plan B
One way to get around the problems we’ve had so far with using a single filter to find the perfect dividend stock, would be to use more than one filter.
- Let’s use a base filter to avoid completely unsuitable stocks
- and tweak it in a variety of directions in order to come up with few different kinds of dividend stocks
- we might be able to build a balanced and diversified portfolio from these sub-sets
The problem with this idea is that the free screening tools we’ve looked at so far don’t include all the types of extra data we’d like to include.
Let’s say that we have a base filter of market cap £500M, yield of 3.5% and PE of 19 or less.
- Last week that got us down to between 70 and 80 stocks, and today it produced 87 stocks.
The extra fields that we’d like to add in turn would be perhaps:
- ROCE or operating margin
- debt / equity ratio
- interest cover
- Price to Book
- dividend cover
- years of dividends
- historic or forecast dividend growth rate
The first four were in last week’s screeners, the last three weren’t.
- For the HL screener, we have three that we can look at: operating margin, price to book and a combination of dividend cover and historic dividend growth rate.
Operating margin of minimum 10% produces 44 stocks, of which this is the top few:
Price to book
Price to book of maximum 3 produces 65 stocks:
Adding dividend cover of 1.5 times and historical dividend growth of 5% pa produced 30 stocks:
I think this combined approach has some merit, and the last screen is probably the best we have seen so far.
- At the moment I am leaning towards a portfolio made from a combination of ETFs, investment trusts and the best companies from this list.
Here’s a summary of what we have covered today.
- We need to build a diversified portfolio of 30 or so dividend stocks
- we will diversify across sectors and geography
- we will focus on larger companies, since their dividends are safer
- “Quality” is focused around dividend reliability and growth, but also earnings, profits and return on capital
- Lists of “dividend aristocrats” can be useful, as can the ETFs which track them
- Counting the reliability and growth of dividends, sales and profits can help us to identify the safest dividend stocks
- Comparisons to market- or index-average numbers (yield, valuations, growth rates) can be useful
- we need to find a data source for these averages
- This is a long-term portfolio, and industries facing potential disruption should be avoided
- We need to understand how each company makes money, and how its strategy will protect it from the forces acting on its industry
- We will review each company’s place when it’s interim and final results are announced, and rebalance once a year, aiming for less than 20% annual turnover.
- The HL free screener has replaced the non-functioning Morningstar screener on our list
- The HL screener has one fewer of our 14 ideal filters than the Google and FT screeners, but is more focused on dividends
- Using a base screen (market cap, yield, PE) and adding extra conditions one by one seems to be a useful approach to produce a useful number of candidate stocks
- adding dividend cover and historic dividend growth is especially promising, with 30 matching stocks
In the next post we’ll look at a practical example of somebody putting together a dividend portfolio.
Until next time.