Dan,
I would have to sack myself as a Financial Planner if I was to agree that I thought you are asking the right question here. Hence all my observations around this being more complex than the question you are asking.
However, as I'm giving a lecture on this tonight at DBS let me put on my asset management hat for moment.
Here is a graphic from Asset Management Company Vanguard which I think gets to the crux of your question.
This is simply saying that the average return of a portfolio between 2000 and 2012 has been lower than the average return between 1926 and 2012. In other words investors should prepare themselves for lower returns in the future than they have received in the past. I think it is helpful to put some real evidence in to support statements that are fundamental to the debate.
The real question is surely this: "well, how much lower and how do we estimated it?"
So, again I produced a guide which I have already referenced called "realistic expectations" and in this I turned the question on its head and asked these questions:
Do I know what the average return from "cash" (I'm using the German 3 Month Money market rate as a proxy for the risk free rate) has been over some long time horizon? - answer yes
Do I know what the equity risk premium has averaged? - answer yes
Do I know what cash is paying today? - answer yes
So, if I am willing to assume that the equity risk premium will stay reasonably constant over time (happy to debate this), then all I need to do is assume a lower risk free rate going forward and solve for the portfolio that matches my expected return given a lower risk free rate.
Now in reality these assumptions are fed into the Black Litterman model to generate the actual asset allocation, but in order to simplify the problem a lot, let's work with historical data to calculate the realized excess return of a given portfolio and simply work out what part of that can be considered risk premium and what part is risk free rate and we are in business.
Assumptions:
We are willing to use some proxy for the Euro prior to 1999. I am using the German Mark in order to extend my analysis back to 1970.
We know our costs from my earlier post are around 1.25%. (Happy to debate paying extra in fund management fees in order to buy Ferraris for fund managers. )
Your target is 3% nominal (I see no reference to inflation in your question - another huge concern)
So your target gross nominal return needs to be 4.25%pa and gross real needs to be 6.25% assuming ECB inflation target of 2%pa.
The following table takes a range of portfolios from 80% Bonds/20% Equities to 98% Equities for the period 1970 to 2014
Source: Dimensional Fund Advisers
How to read the chart
1) Ignore everything except the first column
2) First Column has the Euro short term rate at the top. This is the return you would have earned from "cash" between 1970 and 2014. Over this period the return was 5.18%pa (source Deutsche Bundesbank) So, this is the "risk free rate" you could have had this for "free". Note that over this period inflation averaged 2.73%pa
3) What we are interested in is this. Cash isn't paying 5% today - more like zero. So, if I start with a lower interest rate, what should I expect from an investment portfolio.
4) The number shown in bold in the table is the risk premium for the portfolio above this "base" number. So top left hand corner, the number is 1.889.
That means that the conservative portfolio (80/20) averaged 1.889% above the Euro Short term rate of 5.18%pa over the period.
Note that the period include high inflation in the 70s, poor market returns, good market returns and both falling and rising interest rates.
5) The second number (T Stat) is 2.137 this is a measure of statistical significance and measure how confident we should be in the average premium (it was developed in Dublin incidentally) the test looks at the mean, the standard deviation of the mean and the number of observations. A T stat of 2 or more is generally taken to be statistically significant.
The table gives me the actual risk premium of each portfolio between 1970 and 2014 above the German Short Term Money Market Rate and my base currency is DMY/Euro. Note for the avoidance of doubt these are my research portfolios and not the off-the-shelf DFA portfolios referenced by Stephen earlier.
For forward planning purposes, if we assume a Euro short term rate of say 1% over the medium term then our Expected portfolio return is the Euro short term rate plus the average premium. So, putting it all together:
For a nominal target of 4.25%pa we need around 45%- 50% in Equities
For a real (inflation protected) return of 6.25%pa we need around 80% in equities.
Note that we also have T Statistics of more than 2 for each portfolio so we can be 95% confident that this is the real risk premium for the portfolio and not just background noise.
Caveats
Each portfolio is more complex than a simple MSCI or FTSE All World Equity portfolio and a much lower equity risk premium would be obtained running the same analysis using just a developed equity index. In other words, some of the heavy lifting is being done here by taking more equity risk and less fixed interest risk. Overall this gives the portfolios a better sharpe ratio and better downside protection.