Last week’s post “Risk and Return in Pictures” received some requests to chop up the data in different ways, so this follow up post will try to do just that. I suggest taking a look at the original post by clicking here if you want a refresher.
Request 1: Use Different Start Points
One commenter posited that starting at different points in time might make the results look quite different, namely starting the portfolios in 1999 would make the results less dramatic. I think what he was getting at was to look at the peak of the All Equity portfolio and start the charts from there. The peak occured during September of 2000. When we plot the charts from here and again until the end of August 2008 we get this:
(You can click the graphs to enlarge them)
If we had data to the end of September 2008, the outperformance of the 10% equity portfolio would certainly be more dramatic still since September was a horrible month for equities. However, getting back to the task at hand and since we have the luxury of cherry picking the data, let’s now look at the trough of the All Equity portfolio when the tech bubble burst as a starting point. This will probably have more significance today as many people are pondering a switch from equities to fixed income due to the recent market declines.
(Click to enlarge)
As we see from the above graph, switching to a safer portfolio when the market last crashed would’ve been quite detrimental to your bottom line. The All Equity portfolio had an annualized average return of 17.03% during this 6 years, versus the 10% Equity porfolio’s annualized return of 6.94%.
Request 2: How Did Rebalancing Affect the Volatility?
Going back to the original time span of January 1988 to August 2008 here are the annualized Standard Deviations (higher numbers mean more volatility) for each portfolio WITHOUT ever rebalancing versus WITH annual rebalancing:
Portfolio No Rebalancing Annual Rebalancing
10% Equity 5.10% 5.03%
20% Equity 5.63% 5.37%
30% Equity 6.46% 5.98%
40% Equity 7.46% 6.80%
50% Equity 8.54% 7.77%
60% Equity 9.67% 8.83%
70% Equity 10.81% 9.96%
80% Equity 11.96% 11.14%
90% Equity 13.11% 12.36%
100% Equity 14.25% 13.61%
Here we see that rebalancing had the effect of slightly lowering the portoflios’ volatilies, however as mentioned before, I prefer to use a dynamic rebalancing schedule which is based on reaching a set allocation deviation as opposed to just picking a periodic rebalancing schedule. Also, I’ll note that the performance differences were minimal – but this is a separate discussion altogether (one which involves looking at how much the allocations drift over time, and also the effect of pretty much 20 years of falling interest rates until now – my guess is that the next 20 years will not be the same!)
Request 3: Log Plot to Highlight Earlier Changes
Capping off today’s post is the original graph using a logarithmic Y axis. This serves to highlight the changes in the earlier periods of the graph most notably. Whereas in the non-log graph, all the portfolios seemed to move in lock-step for the most part, now we see some more dramatic swings in line with the latter part of the chart.
The charts in this post and the previous part in the series leave us lots to talk about, and quite frankly I’d love to start writing about it here but Fiona will have my head on a plate if I don’t call it a night – so I’ll leave it to you guys to start with the discussions and I’ll continue my own thoughts later this week. (And yes, I know the next part in the DFA series is long overdue – also on the to-do list!) :)