"What we learn from history is that we do not learn from history." - Benjamin Disraeli

To see if there was a monotonic relationship between last year's rankings and this years rankings. I used the figures for the return of each asset class provided by Vanguard in "Realistic Expectations for Share Market Returns".

It turns out that chasing last year's top returns was not necessarily the worst strategy to use, at least over the last 20 years. But it did come close.

The rankings are the returns experienced in the previous 12 months. A "1" ranking is given to the asset class that performed the best over the previous 12 month period. A ranking of "5" is given to the worst performer.

The five asset classes considered are International Shares, Australian Shares, Property, Australian Fixed Interest (bonds) and Australian cash (bills).

It looks like the relationship is not monotonic (which is a mathematical way of saying each ranking gets better than the one before) but clearly the only class that really has tremendously high returns is the strict contrarian one. If I had 100 years of data to play with it would be interesting to see if the relationships are exactly the same or if the relationship really is monotonic or whether the median rankings are always the worst. What is clear though is that the worst asset class from last year seems on average to be the best the following year. If you are going to use contrarian asset allocation there is little point in diversifying asset classes unless volatility is a very major concern. I found that on the few occasions when there was only a point or two between the 5th and the 4th asset class that if you allocated 50% to each it made only a small impact on the overall performance, but it diminished volatility by a percentage point.

Click here to see the Excel spreadsheet where I tested the strategies.