Often we get asked what our view is of expected returns and, although our investment process is not conditional upon a set of expected asset class returns, we thought it might be interesting to see what a passive balanced portfolio might be expected to return over the course of the next ten years.

We have set an investment horizon of ten years simply because over this time period there is a clear and statistically significant relationship between valuation and returns that is not seen at short horizon dates, certainly not at a one year horizon.

Our analysis is just to give a baseline for what the expected passive return on a balanced portfolio might look like over the next decade.

Starting with Equities: the total return you earn over any period is comprised of capital gain (growth in earnings and the change in the Price/Earnings ratio) and income (dividends).

Without showing some lengthy arithmetic, take it from us (for those of a more nerdy nature we can provide details of the equations and a spreadsheet should you wish) that the Expected Equity index Return (ER) can be closely approximated (from The Gordon Growth Model) as,

ER = (1 + g) * (End PE ratio/Start PE ratio) ^ (1/t) – 1 + Dividend Yield * (Start PE ratio/ End PE ratio + 1)/2

Capital gain and Income are separated.


g = the growth rate of earnings per share over our time horizon, in our case ten years

End PE ratio = the Price/Earnings ratio at the 10 year horizon date, we use the Price/Trailing 10 year average reported earnings

Start PE ratio = the Price/Earnings ratio today, we use the Price/Trailing 10 year average reported earnings

t = the time horizon, in our case this is ten years

Dividend Yield = today’s dividend yield

Now, you might say “if I knew the PE ratio in ten years’ time and I knew how much earnings per share would grow over the next ten years I would be a genius with perfect foresight and life would be easy!”. In fact, you would be correct in thinking that, so we obviously need to make some assumptions/forecasts.

As anyone who has spent time in our company knows, we don’t like forecasts and we have next to no time for forecasters.

(For a fantastic demolition of the futility of forecasting and the witchcraft of forecasters see the following wonderful papers by Dr. Kesten Green and Professor J Scott Armstrong, University of South Australia: “Simple versus Complex Forecasting: The Evidence” and “The Seer-sucker Theory: The Value of Experts in Forecasting

So, rather than forecast the inputs, let’s set them at “normalized” or long run trend levels,

We assume earnings per share growth is at its long run trailing rate over the next ten years, we assume the PE ratio in ten years simply reverts back to its long run historical median level. No forecasts to manipulate, no exaggerated claims, just more of the same we saw over the last 5 decades and more…

Moreover, we proxy expected inflation, needed to convert nominal into real returns, with trailing 3 year average inflation.

So, taking those inputs, we calculate the implied ten year annualized return for the equity indices in the US, UK, Germany and Japan and plot them with real cash yields and real ten year bond yields in each case to give you an “implied return curve”:

In all cases the real implied cash and bond returns are pretty much zero – that is what QE does for you (it drives them to zero to force you out along the curve seeking returns in equities, it drives the curve to the horizontal axis as it has done in the US).

The implied real equity index return in the USA is a measly 2.5%. It is far more attractive elsewhere but we should never forget the down-market sensitivity of non-US markets to the US.


What of a balanced portfolio? Well if we look at a 50/50 Equity/Bonds mix in each case, and a weighted average for a portfolio of all four markets, this is what we get:

A US balanced portfolio has an implied return of 2% – the standard annualized return assumption for most big US pension funds is 7-8% – with the implied real return below 5% in all cases. Interestingly, a weighted average (using equity index weights) is just 2.5%


Of course, it is possible to get more, but that implies that earnings per share growth  (from the starting point of record profit margins) exceeds its long run, and highly stable, trend growth rate and/or valuation multiples go significantly higher.

That would be the best case scenario and, like everyone we like to hope for the best, but a prudent investor always plans for the worst (the event that earnings per share growth is slower than trend or the valuation multiple falls below the long run median level at any point over the next ten years) and that is exactly how we approach portfolio construction.

In the best case our fund should produce solid returns and in the worst case it should substantially lessen portfolio drawdown.