This is just a quick update on recent events in the UK Value Investor model portfolio, including buys, sells and performance relative to a FTSE All-Share tracker.
Before I get into the nitty-gritty, here’s a recap of the model portfolio’s goals:
- High Growth: Generate higher capital gains than a FTSE All-Share tracker
- High Yield: Have a higher dividend yield than a FTSE All-Share tracker
- Low Risk: Have lower volatility and smaller peak-to-trough declines than a FTSE All-Share tracker
- Consistent long-term results: Grow from £50,000 to £1,000,000 in less than 30 years
I’m currently competing against a virtual portfolio which is fully invested in a FTSE All-Share tracker from Vanguard. Vanguard’s funds are some of the cheapest available and should therefore be the best performing and hardest to beat.
Here’s a snapshot of the results of these two portfolios so far:
Higher capital gains? Yes
I know that weekly, monthly and even yearly ups and downs are not important, so I measure capital gains over periods of five years or longer:
| Results to 01/07/17 | Model portfolio (A) | FTSE All-Share (B) | Difference (A – B) |
|---|---|---|---|
| Value at inception (01/03/11) | £50,000 | £50,000 | £0 |
| Value Today | £94,213 | £82,318 | £11,895 |
| 5-Yr total return | 88.3% | 64.8% | + 23.5% |
| 5-Yr annualised return | 13.5% | 10.5% | + 3.0% |
| Total return from inception (01/03/11) | 88.8% | 64.6% | + 24.2% |
| Annualised return from inception | 10.6% | 8.2% | + 2.4% |
As long as the portfolio beats the FTSE All-Share then I’m happy, but what I really want is to get at least 10% per year (on average) which will keep the portfolio on target to hit £1 million within 30 years.
If you’re interested in short-term returns (which you shouldn’t be) then both the model portfolio and the All-Share are up by around 7% so far this year, which is well above average for a six-month period.
However, those gains could easily be wiped out by a minor market “correction”, which is precisely why I don’t like to think about short-term returns.
Higher dividend yield? Yes
The model portfolio has always had a higher dividend yield than the FTSE All-Share and that continues to be the case:
- Model portfolio yield: 4.1%
- FTSE All-Share tracker yield: 3.15%
If (like me) you’d like to live off your portfolio’s dividend income one day, then the yield is obviously very important.
For example, to get a £20,000 dividend income from a portfolio yielding 4.1% (like the model portfolio) you’d need almost £490,000 invested.
But if your portfolio only yields 3.15% (like the All-Share tracker) you’d need almost £635,000 invested.
So you would need to save up another £145,000 just to make up for that 1% shortfall in yield.
As they say at Tesco, every little does indeed help.
Lower risk? Yes
The chart above shows peak-to-trough declines for the model portfolio and its All-Share tracker benchmark portfolio.
Hopefully, it’s clear that both portfolios are about 2% below their all-time highs today, and also that the model portfolio has seen significantly smaller declines than the FTSE All-Share since 2011:
| Peak-to-trough declines to 01/07/17 | Model portfolio (A) | FTSE All-Share (B) | Difference (A – B) |
|---|---|---|---|
| Maximum decline over 5 years | – 3.9% | – 11.4% | + 7.5% |
| Maximum decline from inception | – 8.1% | – 13.5% | + 5.5% |
Keeping volatility and declines to an acceptable level is something that many investors ignore.
They chase high returns from high-risk stocks, which is fine when the market’s going up. But when a bull market turns into a bear market, many of these investors find out that their tolerance for losing money is much lower than they thought.
Many of them end up panic-selling at the bottom of a bear market, locking in losses which would otherwise only be temporary.
So although focusing on controlling risk isn’t exactly “sexy”, I think it’s hugely important. And when in doubt, you should probably underestimate your risk tolerance, just to be on the safe side.
Growing to £1m within 30 years? Perhaps
I wrote recently about my new goal to grow the model portfolio from £50,000 to £1,000,000 pounds before its 30th anniversary in 2041.
For this to happen, the portfolio will need to double in value four and a quarter times in 30 years, which means doubling in value every seven years or thereabouts.
This will require a long-term average growth rate of at least 10% per year.
Today the model portfolio has a virtual value of £94,213, almost £12,000 ahead of its FTSE All-Share tracker benchmark portfolio, which sits at £82,318.
So the first of those four and a quarter doublings is almost complete, and it has taken just over six years from inception in 2011 to (almost) get there.
And that means the portfolio is still on schedule to achieve its 30-year one million pound goal:
Recent purchases, sales and the benefits of Stoic equanimity
If you’ve been reading this blog for more than a few months, you’ll know that I have a long-established policy of buying or selling exactly one company each month.
This steady approach allows me to take profits on winners, weed out losers and replace both with even better companies at even lower prices (hopefully).
Here are the three trades I’ve made since the last portfolio review in April:
- May: Sold BAE Systems for a 142% gain over six years
- June: Bought a FTSE 250-listed UK-focused company operating in the Support Services sector
- July: Sold Morrisons for a 2% loss over four years
As you can see, some investments perform well while others don’t.
It’s important to remember that if an investment goes badly (like Morrisons) it does not mean you’re a bad investor, or that your strategy is bad. Investing is not as black and white as that.
Instead, investing is more like a game of chance. For example:
- Imagine you have a single dice (or “die”, if you’re pedantic) with four green sides and two red sides
- Now roll the dice
- If the dice lands green side up, you win £1,000; red side up, you lose £1,000
- Repeat 60 times
On average, you should expect to get green twice as often as you get red, so after 60 rolls you should expect to get green 40 times and red 20 times. That’s £40,000 won and £20,000 lost, giving an expected profit from this game of £20,000.
So despite this being a good game from your point of view, because you’re very likely to make an easy few thousand pounds, you can expect to lose money about a third of the time.
The stock market is exactly the same. Even if you’re playing by a set of rules which you know will lead to a positive outcome in the long term, you can still expect to lose money on a regular basis.
Of course, you have to have more winners than losers, but you also have to be able to accept the fact that you will lose money on some investments.
Some investors are psychologically crushed whenever they lose money, and I think that’s a serious mistake.
A better approach is to accept that you’ve lost money with stoic equanimity and that occasionally losing money is an inescapable part of the investment game.
When you lose money you should review the original purchase decision and your decision-making process. Work out what went wrong and decide if it was something that was avoidable or if it really was just bad luck.
If it was avoidable then make the necessary changes to your investment process. If it was just bad luck, take the loss on the chin and keep rolling the dice.
Leaning towards mid-cap cyclical UK-focused companies
As the level of uncertainty around the UK economy has increased, the valuations of many UK-focused companies have declined.
However, as a value investor, I am drawn towards lower valuations, so over the past three months the portfolio had nudged its way towards smaller, more cyclical, less international companies.
And I’m not alone in thinking that UK-focused cyclical companies are attractively priced, as Neil Woodford’s blog has said much the same thing.
In concrete terms, here’s how the model portfolio has changed:
- More smaller companies: FTSE 250 stocks now make up 46% of the portfolio, compared to 42% in April
- More cyclical companies: Cyclical sector stocks now make up 57% of the portfolio, compared to 50% in April
- More UK-focused companies: 46% of the portfolio’s total revenues now come from the UK, compared to 41% in April
I don’t want to go overboard with this idea though, so I have rules which limit how cyclical and how UK-centric the portfolio can be:
- Don’t have more than 66% of the portfolio invested in cyclical sector stocks
- Don’t have more than 50% of the portfolio’s total revenues coming from the UK
If I’m wrong about the attractiveness of UK cyclical stocks then these rules will help to protect the portfolio from my erroneous judgement.
To one million pounds, and beyond
So that’s one more quarter down in the portfolio’s long journey to a million pounds.
I’m reasonably satisfied as the portfolio has continued to hit all its performance targets, with more income, more growth and less risk than the FTSE All-Share.
It’s also on track to reach the million-pound mark before 2040, and perhaps even before I reach 65 in 2037.
Hopefully, we’ll all still be around at that point to see how things panned out.
Hello John,
Wise words indeed. After reading ‘The Intelligent Investor’ (the best book on investing ever written according to Warren Buffet), I struggled to find a practical UK-based newsletter/blog along similar ideas.
Since finding your website and subscribing, I am extremely happy with my decision. Your thoughts and writings are carefully considered, and are open and honest about taking the rough with the smooth.
My biggest weekly loss was £16k (I know I shouldn’t, but I like to keep a wiggly graph), but my investments are very well up overall and I’m very very happy about how things are going.
Keep up the good work John, and I’m looking forward to your millionaire announcement in due course!
Best Wishes,
Matthew
Hi Matthew, I had exactly the same problem finding UK-focused value investing blogs back in 2007/8 (also after reading the Intelligent Investor), so I decided to start this one. Amazingly enough, almost ten years later it’s still going!
As for your wiggly graph, at least it’s going in the right direction. And with a bit of luck – and some of that careful consideration you mentioned – both our wiggly graphs will continue going (mostly) in the right direction for at least the next ten years, and hopefully a lot more.
John hi
I’d be interested to see a bit more use of simulation. specifically:
– would it be possible to use previous years’ data, and apply the current picking/stock churning strategy ? (probably needs to be a more automated version – less use of the judgement calls re what to buy and what to sell)
– do you ever look at the stocks sold, 1-2-3 years later, and assess the returns they would have given, if they had been retained ? … how would they have done relative to the relevant index (FTSE all share, divs reinvested) and how would they have done relative to their replacements bought in the following month?
I see the first of these as a really useful test of how the strategy would do in a serious downturn (2007-09 and also the dotcom bust – and others too, if you can go back far enough)
I see the second as a useful resource for measuring the impact of the strategy and possibly refining it – also allowing the possibility of testing whether the ‘rules’ bit and the ‘judgement’ bit are both contributing positively,
Happy to say more if it would be useful (email me is probably easier than responding via this dialogue?)
Hi David
Interesting points. I’ll take each one in turn (this is a long comment; probably should have sent by email!):
AUTOMATED BACK-TESTING – Yes, in principle I could back-test the strategy using historic data, assuming a) the data is available and b) the judgement element is removed or randomised, or otherwise automated.
This might be interesting, but I’m not sure how useful it would be because:
a) I’m very wary of back-testing. Long-Term Capital Management used extensive back-testing. They had an army of PhD’s and they still almost broke the global financial system.
b) My strategy evolves over time as new lessons are learned, so it’s not an entirely static system.
c) I’m not sure what I would do with the outcome. If it showed some massive problems then that might be useful, but what if it showed a slight but inconsistent underperformance over the last few decades? Am I to put more weight on that than the outperformance it’s achieved since 2011?
As I said, it might be interesting, but I would be very wary of making any changes based on back-testing. Companies and the world in which they operate are complex systems, so their past is not their future and their future is almost entirely unknowable.
Other than a few very basic factors like CAPE (cyclically adjusted PE) mean reversion, I tend not to put too much faith in the minute details of the past.
However… it would be interesting to do some back-testing, if I had the time, the data and the inclination!
As for whether back-testing is a useful way to see how the strategy would have performed in past downturns, I think it might be, but it might not. I think how a portfolio responds to a downturn has three main factors: 1) The average valuation – high valuations have further to fall, 2) the average level of debt – more indebted companies have bigger problems in economic downturns, and 3) cyclicality – cyclical companies typically have more problems during downturns.
So as a guess I’d say that if a portfolio was trading at lower average valuations than the market, with lower average levels of debt and a higher weighting to defensives over cyclicals than the market, I would expect that in most cases such a portfolio would be more defensive, i.e. fall less than the market during a downturn.
Back-testing might reinforce this belief, but for now I’m happy with a gut-feel approach to this. I’m sure we’ll have a bear market at some point, and then we’ll have a better idea.
But, back-testing the 2008/9 bear market might be doable. I think I have data for that period for quite a few stocks, so I feed that data into my spreadsheet it might produce a portfolio which I can then track through the 2008/9 period.
I’ll put that on my to-do list, although I can’t say when it might get done (if at all, depending on what else is going on).
If I don’t write about this within a month or so, give me a nudge.
POST-SALE ANALYSIS OF STOCKS SOLD – Again, this might be interesting, but I think it would be very hard to extract some actionable principles from it. Over what time period should sold stocks be analysed? 1yr, 5yrs, 10yrs, 20yrs? If they outperform over say 3 yrs, what do I do about it? Should I hold onto stocks for 3yrs longer? What if they subsequently underperform after 3yrs? That sounds like a recipe hindsight cherry picking of past data. Ultimately, I think this could be a lot of work for very little return (although I’m happy to be proved wrong if someone else can do the analysis and show the value in it!)
As for contribution of rules versus judgement, I think the judgement element adds nothing to performance. I only use judgement to decide which of my bottom five-ranked holdings to sell, but in the end they almost always get sold at some point. I would expect that the judgement element nets out to zero because judgement-based winners will probably be offset by judgement-based losers.
Rather than boosting performance, the judgement element is there mostly for psychological reasons, so that I’m interested in the process and have some feeling of control over it. If I programmed my investment strategy into a trading computer program and ran it every month, it might work just as well but it would be mind-numbingly boring! And I invest as much for enjoyment as I do for financial returns.
(apologies for any typos, and any response is probably best sent by email!)
John
Hope that this is the most appropriate place to post comments / questions
You will be of course familiar with Siegels constant and the debate that took place between Gross and Siegel a few years ago. I guess it’s the nature of economic theory but I haven’t seen any further conclusive research either proving or disproving the existence of the constant. Have you found anything or have any strong views? Your half yearly calculation of fair value of the Ftse has some parallels.
I think Siegel’s constant is not a true constant but is probably a combination of one “constant” and one variable.
The “constant” (obviously not a true constant like gravity) is that US stocks have consistently returned something like 3% per year above GDP growth, but with quite a bit of volatility.
The variable is DGP growth.
So Siegel’s constant of 6.5% (although it actually changes over time) is roughly 3% from equities (essentially dividends and buybacks) plus 3.5% of GDP growth.
Will Siegel’s constant continue to be around 6.5% over the next century? Personally I doubt it, but you never know. The key problem I see is population growth, which as been a big tailwind for US economic growth over the last few centuries. But historic high growth rates can’t go on forever, although the US is a big country and thinly populated, so perhaps they can for another century (I haven’t looked at data like falling birth rates etc. for the US).
Here’s an interesting link on the GDP + 3% argument:
https://www.pragcap.com/u-s-equities-long-term-real-returns/
Thanks John makes sense