This is a summary of the hypothetical results that could have been achieved between the period of July 31, 2009 and June 1, 2018 had one followed the VT Model’s signals during that time. Results are completely hypothetical, do not include slippage or commissions and may contain errors. The hypothetical results do not reflect what one will achieve in the future and are being provided for informational purposes only. Future results will be different.
At a high level, the VT Model underperformed VT in terms of CAGR and total return. However, it had a smaller maximum drawdown. The underperformance in terms of CAGR and total return is to be expected during the bullish part of the full cycle.
The comparative performance can be seen in the following chart. Of note, the smaller drawdowns occurred during the 2011 and 2015/16 market corrections.
As can be seen from the following chart, the VT Model shifted its allocation during those market corrections to IEF as many of the filters went negative.
Next, we will look at the performance of each individual filter to see how each one did individually over the period.
A few things jump out from the table. First, there was a wide variance in returns amongst the filters. This is not too surprising given that the sample period is relatively short, the trades were few and there is a considerable amount of timing luck with such strategies. Second, each of the individual maximum drawdown numbers was greater than the maximum drawdown for the overall strategy. This may seem odd, but it is a function of the independent movement of each filter.
We can see this when we chart the performance of all the filters on one chart. The combination of all the filters (black line) appears to be the smoothest line.
Finally, let’s combine the VT model with our version of the Permanent Portfolio to see how the combined performance would have looked during the period.
A portfolio comprised of 50% allocated to the VT Model and 50% allocated to our version of the Permanent Portfolio would have hypothetically earned a CAGR of 7.41% over the period with a maximum drawdown of 7.88%. The CAGR almost equals the max drawdown, which is favorable. Also, the percentage of positive rolling 52-week returns is higher than either VT Model or the Permanent Portfolio on its own.
The following chart shows the performance of the VT Model, the Permanent Portfolio and the combined portfolio (50/50):
In summary, the VT Model would have performed reasonably well during the period in that only a marginal amount of total return was sacrificed for the potential of a lower maximum drawdown, which is the point of employing a risk management system such as this. However, whether it will perform well in the future, especially during a recessionary bear market, is unknown. I believe that the logic of the filters is sound, but whether that will translate into better risk adjusted returns is completely unknown. Market dynamics may be entirely different going forward.