The Kahneman-Tversky (K-T) Portfolio is a simple portfolio managed using the Dual Momentum (DM) investment approach. Since at least one new member (@Ken Dorge) has indicated an interest in adopting this simple model I will spend a little more time than usual with some explanations/ideas.
The model/approach was made popular by Gary Antonacci (https://dualmomentum.net/ in his book “Dual Momentum Investing”. In it’s original simplest form it requires the selection of only one asset/ETF from a group of three asset classes (US Equities, International Equities and Bonds) based on their relative “momentum” as measured by the 12-month Rate-Of-Change of prices.
The following screenshot shows this portfolio set up in the Kipling workbook (the Kahneman portion of the portfolio):
Note the selection of DM as the model to be used and the 100% weighting to the single 252 trading-day lookback period. VTI represents US Equities, VEU International Equities and TLT Bonds. Antonacci suggests using something like AGG (Aggregate Treasury/Corporate Bond ETF) to represent bonds but I have chosen to use TLT because returns are likely to be higher when Bonds are performing well – the expectation being that this will occur when equities are out of favor, although this inverse correlation is not always observed, especially in times of severe economic downturns.
The Momentum/Relative Strength of the assets is shown in the 4th Column from the left, where SHY (short-term bond ETF) is used as the baseline for “absolute” momentum. As can be seen in the above figure, none of the assets are showing positive momentum and Antonacci would allocate to a Bond ETF (e.g. TLT or AGG). Because TLT is a more aggressive Bond fund, I choose to allocate to SHY rather than TLT but, as we’ll see below, even this is not necessarily the best option at certain times – we would have been better off in Cash over the past few months.
Although not part of Antonacci’s “system” I also like to look at rotation graphs that show the relative strength and direction of “momentum” for each asset relative to a benchmark (in this case SHY):
As we can see in the above figure all assets are showing negative (12-month) momentum relative to SHY (top right corner at the (0,0) origin) in both the longer term (horizontal axis) and the shorter term (vertical axis). However, we do see that, in the shorter term, all assets are showing a little more strength (moving upwards along the y-axis).
The above portion of the K-T portfolio follows Antonacci’s original suggestion – and long term back-tests (~40 years) show this system to perform very well. However, the system is slow to react to changes in market behavior – that is good in that it generates very few adjustments or the need to trade – but may be bad for investors with a shorter term time horizon (i.e. some of us “old fogies”). In this case, maybe we want to shorten the look-back period over which we measure “momentum”.
The Tversky portion of the K-T portolio uses a weighted combination of 60- and 100-day look-backs:
At the present time there is no difference in the rankings – but this portion of the portfolio will react faster to changes in market behavior than the Kahneman portion.
Let’s check the rotation graphs to see this:
Obviously a lot more “action” here – although everything is still in the bottom left (negative-negative, long/short-term momentum/relative strength).
How does all this translate to portfolio performance? :
As we can see, the portfolio has not been able to keep up with the benchmark AOR fund over the past 22 months – although it did outperform over the June-Sep 2021 period. Another reason that the portfolio performance has not been great is due to the use of trailing stop-loss orders (TSLOs) in the bullish 2021 markets. I have since discontinued using TSLOs (never was a big fan in the first place but was trying to “play-it-safe”).
As mentioned above, holding SHY over the past 3 months has not been all that good – better than holding equities but not better than moving to Cash (not an “official” option for the DM system).
Is using the original 252-day (12-month) look-back better than using the shorter-term look-backs? Maybe, maybe not – through the Sep 2021-Feb 2022 period the Tversky (fast) portion performed better than the Kahneman (slow) portion of the portfolio. This is why we encourage diversification even if it requires a little more work.
I have no plans to get back into the equity or bond (TLT) markets as of this review date – but don’t forget that review date luck is probably the greatest factor in determining performance.
Hope this extra level of explanation helps.