‘Please, sir, I want some more.’- Oliver Twist (Charles Dickens)
That we are living in a low return world is now so widely accepted as to be bordering on a cliche- but HOW low will long term returns likely be? With the caveat that no one really knows, it is possible to come up with a reasonable set of assumptions that can provide us with a range of potential outcomes. We looked at this issue around a year ago, so now might be a good idea to revisit the prospects from a slightly different angle, as there are many ways to arrive at a conclusion, some of which will be similar.
One such method that I have been aware of for some time is that pioneered by John Hussman, a US fund manager. His starting position is that, as a security is a claim on a long term stream of cash flows it is the valuation of that security that determines the expected future return thereon. As the price rises, so the expected rate of return drops. As markets are at or very near all time highs, this suggests that returns will be at all time lows. Not so, as we shall see below (though they obviously do not compare as favourably as those prospective returns on offer at previous lows). [See below for the maths behind the calculations].
It is also important to note that although the returns achieved at the end of the period are estimable, how the returns get to that point are not; a 10 year return of 3.48% per annum could be simply an annual gain of 3.48% per year, (a rather dull decade) OR a sharp fall followed by an equally strong rally (much more “interesting”, rather like what has transpired post 2007). There is no way of determining how this return occurs, only the final destination and even that is not certain.
As stated above, opinion is divided on the subject. GMO, the US asset manager believes return will be even lower than this for the S&P 500 (see page 3 of the pdf). though this is over a 7 year horizon. Using Hussman’s method, 7 year returns are expected to be +3.69% p.a.
The crucial determining factor in these calculations is the terminal P/E ratio. One of the main reasons that 10 year returns have been much better than expected is that the P/E ratio has risen so strongly in the last 10 years- investors are paying ever higher prices for each Dollar/Pound of earnings.
The higher the terminal P/E is, the more that the actual returns will be better than this model predicts; using the US example, IF the market P/E is 20 x in 10 years, the annualised return for the Index will actually be 5.95% per annum (and thus overall returns for the 60/40 portfolio will be higher than this model “predicts”). Of course if it is lower, then the opposite applies.
The current low returns environment is clearly a function of the policies enacted by Global Central Banks, (QE) which have effectively equalised returns across asset classes, such that risk -adjusted returns are now more or less the same for all. Whether these policies can be sustained for another 10 years remains to be seen.
The expected return is given by the following formula, broken down into the capital gain and the income.
Long term total return = (1+g) (future PE / current PE)^(1/T) – 1 [Cap Gain]
+ dividend yield(current PE / future PE + 1) / 2 [Income] -This an approximation of the average yield during the time horizon.
Currently, (13/9/17) the S&P 500 is at 2498.4, with EPS (Earnings per Share) of $100.3, implying a P/E ratio of 24.91 x last years earnings. Long term EPS growth for the Index has remained in a well-defined range of 6% growth (on a peak-to-peak basis) for the last 60 years or so, so we shall use that as the growth term. According to Advisor Perspectives, the Long Term Average P/E ratio (based on trailing earnings) is 16.7 x . The Dividend Yield is now 2%. Plugging those numbers into the formula gives us a 10 year expected return of:
(1.06) x (16.7/24.9)^0.1-1 = 0.0185
+ 0.02 (24.9/16.7+1)/ 2 = 0.0249. Adding these together gives an expected 10 year return of 4.34% p.a.
[ Using the FTSE 100 as a rough proxy for the UK Market, with a current P/E of 30- trailing EPS is thus 238.8p, a long term growth rate of 5%, an historic average P/E of 15 x and a Dividend Yield of 3.1%, one arrives at an expected return of:
(1.05) x (15/30)^0.1= -2.03 +
0.031 (30/15+1)/2 = 4.65, giving a total return of +2.62% p.a. ]
For Bonds, the expected return is simply the Yield to Maturity (assuming no default) which are currently 2.19% (US) and 1.15% (UK)
Thus, the typical 60:40 portfolio [2] could expect 10-year returns in the region of 4.34% x 0.6 + 2.19 x 0.4 = 3.48% for the US portfolio and 2.62 x 0.6 + 1.15 x 0.4= 2.03% for the UK version. One can do the same for other markets (or indeed the World Indices) and get similar results.
Of course, one can arrive at a number of valid outputs depending on the time horizon used in this formula. If one changes the time horizon, (to say 12 years, rather than 10) you get a slightly higher rate of expected return. In the case of the US, the 12 year annualised return is 5.02% p.a.
[2] For some strange reason, financial analysts and strategists insist on including a 10% cash weighting in portfolios. In the current environment that appears nonsensical, so we, shall as we do in practice for our portfolios, ignore it these calculations.