My investment strategy consists of 4 parts:
1. A long-only systematic strategy that invests in either global equities (VT or VTI/VXUS) or 7-10 year US treasuries (IEF) (“VT Model”). 50% of capital is allocated to the VT Model;
2. A version of the “Permanent Portfolio”, which is a buy and hold portfolio made up of 25% global equities (ACWV), 25% long-term US treasuries (TLT), 25% gold (SGOL or GLD) and 25% short-term US treasuries (VGSH or SHY). Holdings are rebalanced annually. 50% of capital is allocated to the Permanent Portfolio;
3. A mean reversion strategy that uses margin of up to 80-100% of the account (“MR Strategy”); and
4. A discretionary trading component that uses margin of up to 20% of the account (“Discretionary Strategy”).
VT Model – 50% of Capital
The goal of the VT Model is to create a simple, low-cost, adaptive asset allocation strategy, which attempts to improve the risk-adjusted returns of an all equity portfolio.
Historically, equities have provided the greatest portfolio returns. Thus, to generate strong returns over the long term, it is desirable to hold a portfolio with a 100% allocation to equities. However, the returns associated with an all-equity portfolio come with a great deal of pain in the form of large drawdowns.
Large drawdowns must be avoided for two reasons: (1) for anyone in, or approaching, the withdrawal stage of his or her life, they can have a very damaging effect on the safe withdrawal rate and a portfolio’s survival; and (2) large drawdowns are psychologically very difficult to handle and often result in bad behavior like selling at market bottoms. Avoiding very large drawdowns limits the risk of doing anything irrational. Even if absolute compound returns are lower from employing risk mitigation strategies such as those included in the VT Model, the actual returns experienced by the individual may be higher due to the avoidance of poor behavior.
The VT Model attempts to reduce such drawdowns by using a variety of technical indicators and economic signals to adjust the allocation to VT on a weekly basis to a level between 0% and 100%. These are mechanical adjustments and not discretionary.
When all signals are positive, allocation to VT will be 100%. When all signals are negative, the allocation to VT will be zero. Usually, the allocation will be somewhere in between 0% and 100%. Amounts not allocated to VT will be allocated to IEF. If the VT Model performs well, returns will be similar (though not necessarily greater) to VT over the long term, but with lower maximum drawdowns. The goal is to reduce maximum drawdowns during severe bear markets by at least 50%.
Equity allocations are made to global equities, rather than equities of any country or region. This eliminates any home country bias. VT is a global equities ETF that is roughly divided 50/50 between US and non-US equities. US equities will typically outperform non-US equities during periods of US dollar strength, whereas non-US equities typically outperform US equities during periods of US dollar weakness. By having roughly equal allocation to US and non-US equities, the hope is to smooth out the impact of US dollar fluctuations. Obviously, when US markets are outperforming global markets, VT will lag US markets but will outperform non-US markets. Similarly, when non-US markets are outperforming US markets, VT will lag non-US markets, but outperform US markets.
Permanent Portfolio – 50% of Capital
The Permanent Portfolio, invented by Harry Browne, has proven to be remarkably stable over time, due to the uncorrelated returns of each asset class. Here is the performance of the Permanent Portfolio using Portfolio Visualizer since 1986:
Obviously, it underperformed US equities, but the worst year was -3% versus -37%, and the maximum drawdown was 13% versus 51%. Clearly much easier to tolerate than an all equity portfolio.
By allocating 50% of the overall portfolio to the Permanent Portfolio, I am hoping to add a core, stable strategy to the overall portfolio and one that will only be moderately correlated with the VT Model. The low to mid correlation between the VT Model and the Permanent Portfolio should help smooth out returns of the overall portfolio. My only change to the originally created Permanent Portfolio is to include international equities as well as US equities. Again, this removes any home country bias.
MR Strategy – Traded with Margin of up to 80-100% of Capital
The MR Strategy trades ETFs and low beta large cap stocks on a short term basis (a few days). This strategy is complementary to the VT Model as it is intended to do better during moderately volatile and choppy markets, when the VT Model will be subject to whipsaws. Conversely, the MR Strategy does not trade when market volatility is low, when the VT Model will normally be fully long. Because it will not trade during low volatility periods, there will be extended periods when the MR Strategy will not be active. As a result, although total allocation to the MR Strategy may be as high as 100% of the account, its average exposure and margin utilization will be about 30%.
Discretionary Strategy – Traded with Margin of up to 20% of Capital
Margin of up to 20% of the total account may be traded on a discretionary basis. Discretionary trades will focus on allocating capital (long/short) to asset classes that are trading at extremes from a sentiment and trend perspective and to trades that would not be correlated with the overall portfolio. Occasionally this may include positions contrary to the VT Model when sentiment and trend reach extremes, typically after corrections. Discretionary trades seek to add positive absolute returns to the overall portfolio, but not meaningfully increase the risk of the overall portfolio. These trades will typically be short-term (days to weeks).