We now reach the end of this series of articles relating to investment portfolio planning, with particular attention to risk and return, volatility, diversification, correlation and, finally, the “efficient frontier”.
First of all, we need to conclude our cross-correlation chart by including UK and overseas equities.
As we noted last week, these two equity classes have been quite highly correlated but with sufficiently wide and frequent variations to indicate that diversification between the two will reduce overall volatility of the portfolio. But what about their correlation with the other major asset classes? The complete correlation chart is shown here.
These last two lines show UK equities have proved to be very lowly correlated with all the other major asset classes over recent years. Indeed, the negative correlation with long-dated fixed-interest gilts confirms the increasing trend for these two asset classes to be viewed as the prime alternative asset classes.
In other words, when institutional investors withdraw money in significant amounts from equities – thereby causing prices to fall – a natural home for that money has been fixed-interest UK Government bonds – thereby causing prices of those gilts to rise.
Negative correlation generally means that the rise in value of one asset has some degree of opposite effect on the other asset. Here, diversification leads to a very significant reduction in portfolio volatility.
The correlation factors involving overseas equities show an even more marked lack of correlation with the other asset classes, with the exception of UK equities. Here, this low correlation not only indicates that diversification reduces portfolio volatility, but also it does not do so at the expense of significantly lower investment returns.
Now we need to identify the make-up of a model portfolio. What combination of which assets will produce the highest possible returns while not exceeding a stated overall portfolio volatility? The answer is the “efficient frontier”.
It may be derived from looking back at years in which equity returns have been poor (which, for the purposes of the chart above, we stipulate as being below 13 per cent overall return). Note just how few times the return on UK equities has fallen below the 13 per cent and, when it has, in around half of those years the return on overseas equities would have given a significantly higher return.
More important might be to note that in five of the seven “bad equity years” property has yielded returns higher – usually significantly higher – than 13 per cent.
Fixed-interest gilts have done the same on two occasions, index-linked gilts on one occasion and cash also on one occasion (well, almost – 12.6 per cent in 1990).
The property statistic in particular helps to indicate its low historical correlation with equities – as indeed does property's very strong performance in 2000 and the first part of 2001, during disappointing times for UK equities.
Now, if we were to speculate that this past 20 years or so might possibly give some kind of guide to the possible future performance of the different asset classes – at least as far as correlation is concerned – then we start to head towards a basis for a model portfolio. This would be especially the case for a portfolio from which assets are expected to be encashed on a regular basis to meet a required flow of income – for example, a portfolio to be used with income drawdown.
Of course, we must acknowledge the fact (or at least the belief by some authorities in high office) that past performance may be no guide to the future but there is little doubt it is not a bad starting point (see chart).
It is also interesting to note that the major constituent assets in this model portfolio are also lowly correlated with each other, with the exception of UK equities and overseas equities which, in any case, have demonstrated frequent wide variations.
Thus our portfolio would, over the last couple of decades or so, have given high overall returns with very low combined volatility. Moreover, and I would suggest this could be the most important issue for many investors, there is not a single year when the need to withdraw money from the portfolio by encashing assets could not have been met from the encashment of an asset which had performed very well (that is, at least 13 per cent) over the preceding year.
Of course, one model portfolio – even if it is accepted as being based on sound fundamental principles – will not answer every client's needs.
In particular, clients with a higher tolerance to risk may be better served by a higher proportion of higher volatile investments, though still keeping at least one eye on low correlation portfolio structure.
The range below is therefore suggested (not by me, I stress, but by the high return/low correlation strategy) as being close enough to an efficient frontier while still allowing sufficient flexibility for client needs.
One last point. If these combinations look remarkably like the make-up of many with-profits or managed funds, then perhaps it is no coincidence. If, this series of articles has only led to a suggested asset mix which has been followed by major investment life offices for many decades, I have done no more than confirm that the professionals have never been getting it far wrong.
Next week, on to some incredibly important court cases – very recent and future – which will have a massive impact on the pensions world.