This post on the equity risk premium is from former hedge fund manager turned author Lars Kroijer, an occasional contributor to Monevator. He also wrote Investing Demystified.
We’ve previously seen how the equity risk premium was 4.3% during the period 1900-2014.
All very interesting – but what’s it got to do with you?
Well, I’d argue we have good reason to expect global equities to outperform UK and US government bonds by a similar magnitude of 4-5% in the future, too.
This means you can plug this sort of real return figure – say 4% – into your compound interest calculators and other such tools when working out how equities fit into your investment strategy.
It’s a practical way of dealing with the uncertainty of shares.
A real return, but not a certain one
Now, some people do criticise this approach.
They argue that using historical returns to predict future returns will lead you to expect higher returns at peak markets, and lower returns at market lows.
And to be sure, historical returns from UK equities looked a lot better on 1 June 2007 – before shares crashed in the financial crisis – than on 1 March 2009, in the depths of the aftermath.
Even worse, perhaps it was because you were attracted by the high historical returns you saw in the data in mid-2007 that you decided to invest in equities – retrospectively right at a market peak.
Combining high historical returns with low expected risk made equity markets look most attractive, just when in hindsight they weren’t.
I understand this criticism but – aside from the general unavailability of crystal balls to predict short-term stock market crashes – I think the length of data mitigates it, at least when we’re involved in long-term planning, not market timing.
With hundreds of years of data across many geographies – incorporating spectacular rises, huge falls, and everything in between – I think historical data is the best guide to the kind of risk and return we can expect from the equity markets going forward.
A more practical complaint is that in the past investors couldn’t actually buy the whole world of equities.
One of the leading index providers, MSCI, only started tracking a ‘world index’ in the late 1960s.
Easily investable products that followed this or similar indices did not arrive for decades after.
So perhaps investor expectations have changed – and the equity risk premium shifted – due it being easier and cheaper to invest in globally diversified equities now than was previously the case?
Time will tell.
Alternative approaches
It’s worth noting there are other ways to derive a figure for your expected returns from equities.
For example, you might look at the dividend yield of the stock markets, or the average P/E ratio.
Combining either of these measures with longer-term earnings growth estimates can yield you an estimate of projected stock market returns.
The problem with these measures is they use quite short-term financial data, and combine it with a highly unpredictable long-term growth rate in order to extrapolate something as uncertain as future stock market returns.
To be honest, I don’t see that as an improvement on looking at historical data.
Other people suggest conducting surveys asking investors what their projections are for the markets, to try to gauge what returns they’re demanding.
While an interesting idea, these surveys are criticized for being heavily sentiment-driven.
They might also tell you more about investors’ desired returns than what they actually expect to earn.
Lars’ predictions
On average I expect to make a 4-5% real return per year above the minimal risk rate (that is, the return I can get from short-term US government bonds) from a broad based world equity portfolio.
This is the figure I’d use in my financial planning when figuring out what I hope to achieve over the long-term.
Of course I do not expect this return to materialize every year.
However if I had to make a guess on the compounding annual rate from global equities going forward, I would plump for this sort of range.
Expected future real returns
| World equities | 4.5-5.5% |
| Minimal risk asset | 0.5% |
| —- | |
| Equity risk premium | 4-5% |
Note that while the equity premium here is compared to short-term US bonds, I would expect the same premium to other minimal risk currency government bonds.
This is because the real return expectation of short-term US government bonds is roughly similar to that of other highly rated countries such as the UK, Germany, and Japan.
Apologies in advance
By the way, if you consider these expected returns to be disappointing, I’m sorry.
Writing higher numbers in this article or putting them into a spreadsheet won’t make them true.
Anyway a 4-5% annual return premium to the minimal risk asset if achieved will quickly add up to a lot. You could expect to double your money in real terms roughly every 15 years.
Indeed, some people would suggest that by expecting equity markets to be as favourable in the future as in the past, I’m indulging in wishful thinking!
Why the risk premium exists
It may sit wrong withyou to have something as important as what you can expect to make in the stock market be based on something as unscientific as historical returns – or for that matter my ‘guesstimate’.
Perhaps so, but until someone comes up with a reliably better method of predicting stock market returns it’s the best we have, and in my view a very decent guide.
Also, it’s not superstition we’re dealing with here.
We know that the equity premium should be something – because if there were no expected higher rewards from investing in riskier equities, then we would all simply keep our money in low risk bonds.
Another argument with simplistically predicting a stable risk premium is that we don’t change it much with the world around us.
It doesn’t seem right that the expected returns going forward should be the same in the relatively stable period preceding the deep stock market crash of 2008, say, as during the height of panic in October of that year.
Did someone who contemplated investing in the market in the calm of 2006 really expect to be rewarded with the same return as someone who stepped in during the despair and frenzy of October 2008?
Very probably not; someone willing to invest at a moment of high panic would likely expect to be compensated for taking on that extra risk.
This suggests the equity risk premium is not a constant number, but is somehow dependent on the risk of the market.
At a time of higher expected long-term risk, equity investors will likely be expecting higher long-term returns.
We might therefore think of the 4-5% equity premium I’ve outlined as an expected average based on an average level of risk.
Putting the risk premium into practice
In the interest of trying to make something as complicated as the global financial markets into something almost provocatively simple, here is a chart showing what we can expect in terms of returns after inflation.
The equity risk premium in graphical terms.
The graph shows how as risk increases, so does our expected return.
So if you’re an investor who wants to achieve returns in excess of the minimal risk return, you can invest in a broad portfolio of world equities.
In my opinion you can reasonably expect over the long-term to make a real return of 4-5% per year above the rate of minimal risk government bonds, which I expect to be about 0.5% per year.
However you can also expect your annual return to vary significantly, with a standard deviation of about 20% per year.
If that sounds too risky to you, you can combine an investment in shares with an investment in minimal risk government bonds to find your preferred level of risk (that is, volatility).
| Minimal risk | Low risk | Medium risk | High risk |
| 100% Bonds | 75% Bonds | 50% Bonds | 0% Bonds |
| 0% Equities | 25% Equities | 50% Equities | 100% Equities |
By varying the ratio of equities to bonds, you can blend your portfolio to suit your best guess of your individual risk tolerance.
Adding a greater percentage of bonds reduces the volatility in your portfolio, at the expense of reducing your returns.
Simple is best
I believe by following such a strategy you will do better over the long term than the vast majority of investors who pay large fees needlessly to earn consequently poorer investment returns.
Such a strategy can be enabled by combining just two index tracking securities – one tracking your minimal risk asset, and one tracking the world equity markets.
If this seems just too simple then remember your simple portfolio hides a lot of complexity below the surface.
Your world equity tracker is giving you exposure to a large number of often well-known companies, operating in many currencies all over the world. From just your two securities you are therefore getting amazing diversification, along with a minimal risk security in the shape of government bonds that give you the greatest amount of security possible.
An excellent portfolio made up of just two securities… Who said investing had to be complicated?
Lars Kroijer’s book Investing Demystified is available from Amazon. He is donating all his profits from his book to medical research. He also wrote Confessions of a Hedge Fund Manager.
Comments on this entry are closed.
There is something to be said for this super simple approach, perhaps it’s best for a SIPP, where you want to be more sure of not losing all your money and having zilch to retire on? Obviously it depends not just upon your risk tolerance, but also on your age when you begin…
Nice follow-up to the earlier article!
Cheers
There are a range of methods to estimate the real return on equities. I don’t like the terminology equity risk premium, because that is the term for the return over the risk free asset, and there is no agreed definition of what the risk free asset is.
The greatest understanding of the future real returns that might be achieved is achieved by considering all the different methods. If I had to choose a single method to estimate future real returns on equities then I would go for the supply side models, which calculate what the market will provide by way of the supply of those dividends, that is it asks the question what are those supplied dividends worth? At its simplest level these models work based on something like dividend yields are say 3% and dividends increase in line (say) with prices or just below, hence your expected real return on equities is mathematically on those assumptions perhaps 3%pa (that’s not my estimate just an explanation of the method). At a more detailed level you look at what dividends are being paid, possibly averaged over some years based on some Shiller type ratio and allowing for share buybacks and sustainable payout ratios. You calculate the growth in dividends noting that the growth of dividends is related to Gross Domestic Product, net of a dilution affect from the growth of new companies before they issue equity capital. Using supply side methods can still give you a range of estimates, the range is telling you there is uncertainty involved, and that’s part of the risk of equity investment.
While it is useful to look at all methods the suggestion of historical returns as the best method is a surprising choice. Stock markets are complex adaptive systems and using the past to predict the future is too simplistic in my view. If you look what has actually caused those returns in the past, allowing for factors such as increases in price earnings multiples, you at least gain more insight. And if you break down those past returns you are already in the process of moving to more prospective methods.
An absolutely wonderful book on the subject is Rethinking the Equity Risk Premium by Laurence B. Siegel (Editor), Martin L. Leibowitz (Editor), P. Brett Hammond (Editor). This is a set of papers by academics and practitioners on the subject. You can download it for 99p on your kindle. It is not an easy read, but it’s a truly wonderful book. It’s all the better for the fact that you are reading the views of a range of different experts on the subject. The conclusion I reached is how difficult it is to predict what real returns might be received in the future, even over 30 or 40 year time periods, and the variety of estimates that are out there. However it also gave me confidence that having a good chunk of equity investment is a sensible choice albeit there are no guarantees.
A good article, very clear & well written – ironically, it goes to show how much effort to confuse goes into the standard commercial info put out there by companies trying to sucker you into a bad deal. 🙂
As you allude to in your article, some variables in modern times are new, so historical comparisons are even more difficult, but equally some fundamentals [like human psychology] will always be unchanged.
One new factor is the increasing AI/computerisation of equity dealing, think high-frequency trading, & my guess is that this will lead to more volatility. While that will be relevant to the short-term, it will probably increase the general perception of overall risk (negatively) for the equity asset class, at least amongst relatively amateur investors.
So if that were true, do you think a significant disparity between real & perceived risk would then change that real risk? [a lot of investors dropping it as a part of their portfolio for example ….. perhaps to be replaced by day traders attracted to the opportunities afforded by the volatility]
…..& if so, what effect would it have on the returns for those equity investors who persevere long-term under the new circumstances?
excellent well written, sensible article.
for tracking the minimal risk asset , the short term gov bonds , does Lars think go global here or remain uk only?
Lars Kroijer is the global version of Jack Bogle. He is right too.
Real returns:
Charles Ellis: 4,5%
Jeremy Siegel: 5%
Simplicity, for the majority of your portfolio, is key.
I did wonder about the increasing ease of investing, alongside the availability of cheaper platforms, pushing up the demand for equity and so reducing returns…any evidence of this?
Mr Z
What about the larger risk premium for investing in higher risk equities, such as small caps or emerging markets? Often people say there are bonds and equities and the way to vary your return/risk is to vary the mix depending on your age. But if you’re a long way from retirement with a bigger risk appetite, surely you should be mixing in a much larger proportion of small caps and emerging markets which might well have risk premiums of 5-7%?
great article
on a slight tangent – if you’re trying to work out the dividends paid within an acc type fund does the following sound right?
no. of units held on the ex dividend date x dividend amount for that ex dividend date
i can get the ex dividend dates and associated dividend amounts from trustnet no problem
you then look at the associated payment dates to figure out which tax return year the dividend payments fall into..
man – ISAs are so much better than all this bollocks 😉
Your analysis may well work out over (say) a 50 year horizon.
However, over say the next 10 years, let’s consider:
1 The US makes up 58.8% of the MSCI world index.
2 Based on the US CAPE, the predicted US returns over the next 10 years are rather low.
So I wouldn’t want 58.8% of my portfolio in the US. Currently I’m at 1.97% and considering reducing it to zero.
Don’t forget that prior to the the dotcom bubble there was another bubble. At the end of the 1980s the share of market capitalization that Japan represented in global equity markets (MSCI WORLD) was
as high as 45%.
Therhino, just switch to income units and reinvest in neat chunks like 5-10k.
Makes it much simpler to keep track
@Marco – thanks for the tip. I’m starting to think its not that bad though to work out the dividends paid within an acc type fund. The critical bit of book keeping is recording the no. of units you have on the ex dividend date.
I’ve left a comment at http://monevator.com/accumulation-funds-dividends/ with a suggestion on how that article could possibly be made a tiny bit clearer in relation to ex dividend and payment dates.
I also wonder whether HMRC are even bothered about you filling in the dividend sections if you are a basic rate tax payer? My understanding is that you will have already paid the tax automatically if this is the case, similar to the default case of paying basic rate on interest in a bank account if you do nothing, i.e. don’t fill in an R85.
Curiosity strikes: how much of the “equity premium” do we assign to merely keeping even with expected inflation? Why an excess, if any? My own thought: equity premium chosen to keep ahead of inflation, plus excess sufficient to avoid wipeout across the period one absorbs volatility. Equity premium thus should be (and is) higher in cyclicals such as oil.
Quibble on using “risk” when “volatility” is meant, eg on the third line of the legend for the upper-right point on your big chart. However, that’s coming from inaccurate terminology in academia and what’s in use in the current literature, so can’t lay blame on you. “Stocks are risky (because they’re volatile)” seems to me to confuse cause and effect — stocks have risks, actual and perceived, thus their prices are far more volatile than those of the baseline bond.
@JB
“Quibble on using “risk” when “volatility” is meant, eg on the third line of the legend for the upper-right point on your big chart. However, that’s coming from inaccurate terminology in academia and what’s in use in the current literature, so can’t lay blame on you.”
Academia are perhaps using volatility as a risk measure because it makes life simple in calculations?
Warren Buffet has famously had his two cents worth on this presumption.
‘Downside Risk’ seems a more sensible measure for the risk aware investor.
With a ‘Well Balanced Portfolio’ volatility equals opportunity (rebalancing opportunity to add low and reduce high). If markets did not fluctuate this opportunity would not exist.
@ Joseph Beckenbach – “Equity premium thus should be (and is) higher in cyclicals such as oil.”
Is it really higher? I looked on Morningstar for long-term returns of the Vanguard Energy fund (VGENX) and its returns look slightly lower than S&P500 total return since 05/23/1984, when VGENX started. The US energy equity index that Morningstar uses as a benchmark has returned much less than S&P500 total return since the same date. In the last ten years VGENX and the energy index have returned less than total US stock index and S&P 500. You could start from the spring of 2000 to see a short-term higher return for oil stocks thanks to the post-dotcom-bubble oil boom, but that’s cherry-picking. 2007 to summer ’08 would also be a bad place to start because of the oil-price peak.
Ok you said “oil” on its own as a sector but many oil companies also do natural gas production and electricity too so energy funds are the only sensible proxy to look at to me.
Sectors that start small and have huge productivity growth will probably have the biggest premium in future. Healthcare is currently a sector that has had a genuine higher premium I believe – Vanguard’s healthcare fund has returned 17% a year on average since 1984. I think healthcare has been a fast-growth sector because of the aging of society plus new biotech developments, while fossil fuel energy is a highly-saturated sector that had its excess growth a long time ago.
The future is unknown which is why a total world stock index will probably outperform most people’s guesses over the next 30 years or so. If I was forced to pick one sector for outperformance in future I would go for green energy. The iShares Clean Energy ETF has been a disastrous performer since its launch in 2008 (minus 19% a year), making it a contrarian play I suppose:
http://www.ishares.com/us/products/239738/ishares-global-clean-energy-etf
Personally I don’t follow his advice (yet) but Investing Demystified is one of my favourite books. This is so clear and deep! So Lars is one of my DIY gurus. Watch him:
https://www.youtube.com/watch?v=xLyccwXd0hY
i don’t think there’s any evidence for different equity risk premiums for different market sectors. sectors can do a lot better or worse than the market over a period of decades, but is it at all predictable in advance which will do better?
if anything, there is some evidence for a *low* volatility premium – see http://monevator.com/low-volatility-premium/ – but that is a murky area. and it doesn’t necessarily translate into premiums for market sectors, anyway.
clean energy will be of vital importance for the continued success of the broader economy. but it doesn’t necessarily follow that investors in that sector will do especially well. that is not always the case with new technology. a lot of the research for new technologies takes so long to pay off that only the State is prepared to put the money in. and there is a danger that the clean energy businesses that it’s possible to invest in are ones that VCs were able to package up and float quickly, so that *they* get good returns, regardless of whether those businesses’ longer term prospects are good.
straying blatantly into the world of active investing … if the rise in world temperatures is to be limited to 2 degrees, only about 1/5 to 1/3 of known reserves can be burnt. it follows either that oil, gas and coal companies are currently massively overvalued – a “carbon bubble”; or that we will be facing the economic consequences of more extreme temperature rises (perhaps 5 or 6 degrees), which will probably lead to an even greater loss of value in other market sectors (and that’s just the financial costs). even mark carney has started talking about this …
thanks a lot for the comments. There is a lot of academic literature on both the size and methodology of finding the equity risk premium (I also don’t particularly like the term as it confuses non-finance people) and I have tried to summarize them in a non-academic way in the book, including dividend discount / surveys / PE growth/ etc. I think most people are just uncomfortable with not having something more firm/tangible to cling on to when trying to figure out something as important as what they should expect to make from their equity portfolio.
On another note, I am starting the process with the publisher of doing the 2nd edition of Investing Demystified. If any of you have read the book and don’t mind I would greatly appreciate if the publisher (Financial Times Publishing) could email you a short Q&A. Obviously anonymously. If you don’t mind, you can email me on lars@kroijer.com or directly to publisher if you prefer on lisa.robinson@pearson.com – many thanks!
@ Lars – That’s great news!