This section documents my research related to algorithmic investing, much of which was carried out in collaboration with members of the Silicon Valley Computerized Investing (CI) group, formally affiliated with AAII.
The research relies on
back-testing. The SEC calls my
observations “hypothetical” because they did not result from actual
security transactions.
It was necessary to
synthesize fund price histories for back-testing because few funds have
twenty-year histories, and no funds have fifty-year histories.
Price histories were
developed from actual securities to the extent possible. Actual price histories include a measure
of investment expenses. Histories
beyond about twenty years generally reflect investment indices. Index price histories do not include
investment expenses, which tends to overstate performance.
Backtesting is clairvoyant
in the sense of knowing the closing prices on the last day of the month and
being able to execute trades at these prices. There is often a significant difference
between the closing prices of the securities comprising the NASDAQ 100 and the
inter-day prices at which most retail trades are executed. There is generally a difference between closing prices of ETFs and mutual funds and the actual
prices received when transitioning a portfolio between ETFs and mutual funds.
This research is
unaudited. It is possible that the
results are wrong or even that the results have been fabricated.
No one knows how
historical results, real or hypothetical, apply to the future.
Even if there are no
errors and even if the differences between hypothetical and actual results are
unimportant, my conclusions may not apply to your specific circumstances and
therefore cannot be considered recommendations.
For the record, I pay for
my SectorSurfer® subscription. I have no financial relationship with
Gary Antonacci, SumGrowth Strategies, LLC, Silicon
Cloud Technologies, LLC, Julex Capital Management, LLC or any other individual
or organization mentioned in this research.
Investors need to consider
these issues carefully. If you do
not understand any part of this disclaimer, seek competent advice.
SectorSurfer® is a trend following (momentum) investment strategy which determines market trends as a double exponential moving average.
· SectorSurfer Forward Walk Progressive Tuning, an introduction. This article illustrates the double exponential moving average trend calculation and how SectorSurfer can be evaluated in terms of both tuning and return. Start here if you are new to SectorSurfer®.
· Presentation handouts.
Easily Implemented Momentum Strategies to Increase Return and Reduce Risk – Presentation to the AAII Silicon Valley Chapter, February 10, 2017.
Risk adverse investors can reap performance gains using Antonacci’s Dual Momentum investment strategy. Peter mentions how aggressive investors can obtain higher returns without the risk of selection bias and that all investors need to be prepared for disappointing short term performance due to the imperfect nature of momentum algorithms.
Three Momentum Algorithms, the white paper from which the February
2017 AAII presentation was drawn.
Taming Drawdowns and Increasing Risk Adjusted Returns with Don Maurer and Allan Zmyslowski, July 2017. The conclusions are
·
Small
value and small momentum portfolios have provided higher returns than US large
cap portfolios over long intervals.
However, these portfolios can sometimes provide no benefit for decades
and they have occasionally lost a third of their relative value over a few
years. Small value and small
momentum investment strategies are more appropriate for patient investors with
long time horizons or who are otherwise able to weather periods of
underperformance.
·
Small
value, small momentum and Cloonan’s Level3-type portfolios exhibit larger
drawdowns than conventional benchmarks.
Level3 portfolios with risk control have drawdowns which are comparable
to the drawdowns of the conservative Wellesley Income fund
and which are smaller than the drawdown of the benchmarks.
·
Risk
control generally reduces the return but the reductions in return are smaller
than the reductions caused by a permanent bond allocation.
·
History
suggests that the SIMPLE, Pinkerton and Level3 portfolios with risk control are
safer than conventional benchmarks because they provide smaller drawdowns and
thereby reduce the risk of panic selling.
They are also safer because they improve the chances of accumulating
enough for retirement and reduce the risk of running out of money.
·
High
return factor investing is a do-it-yourself strategy since individual investors
are less affected by the liquidity issues which plague fund managers.
The SIMPLE Investment Strategy with Don Maurer and Allan Zmyslowski, September 2017.
This article begins with a discussion of the constraints which challenge factor investing. We then address downside risk mitigation and argue that tactical changes to the bond allocation (otherwise known as market timing) is cost effective compared to the usual approach of including a static bond allocation. We trace the evolution of the SIMPLE strategy from its origin as Antonacci’s Dual Momentum strategy, describe backtested performance since 1974 and illustrate the effect of the SIMPLE strategy on savings accumulation and portfolio longevity. We conclude that
· Factor investing offers the possibility of considerably higher returns than traditional benchmarks but practical issues limit the return potential for funds and large portfolios.
· The SIMPLE strategy has, based on backtesting, provided higher returns, higher risk adjusted returns, smaller drawdowns and a lower underperformance frequency as compared to traditional benchmarks.
· The SIMPLE portfolio reduces the return risk associated with saving for retirement and the risk of running out of money during retirement.
· The potential improvements with the SIMPLE strategy are so large that planners should consider revising the guidelines for pre-retirement savings rates and for post-retirement withdrawal rates.
The extended version
of the SIMPLE Investment Strategy contains additional detail.
Reliable Timing Algorithms. The postings in this group discuss the choice of the benchmark portfolio, what
can be learned from the managed performance of large cap US stocks since 1927
and the results of managing large cap US stocks and more complex portfolios
using three dozen timing algorithms. This section also documents the
definition of many timing and allocation algorithms and discusses common
investment mathematics and how historical data can be corrected.
Definition of
Timing and Allocation Algorithms.
Investment Math and Correcting Yahoo Equity Curves. Common investment calculations and how the curated equity curves were developed.
Curated Data. This EXCEL file illustrates common investment calculations and documents the curated equity curves for large, mid and small-cap US stocks, foreign stocks, US real estate, CASH and 4-week Treasury bills. The spreadsheet also illustrates the calculation of the SPVOL timer and the AbsMom5_1+ DR*VOL+IUC composite timer.
Reliable Timing Algorithms. A “reliable” algorithm has high WINS, a high reliability index and good traditional statistics. An equally weighted composite of the AbsMom5_1, DR*VOL and IUC timing algorithms meets these criteria.
Handouts of October 3, 2019 presentation to the CIMI group.
Conservative Investment Strategies, Silicon Valley CI Group, January 7, 2021.
Abstract. The goal was a strategy with a volatility
comparable to that of the conservative 20:80 benchmark, but with a lower
drawdown and consistently higher return.
Maurer’s S121 tactical strategy and variations thereupon did not
achieve this goal.
The revised goal is a
strategy with a volatility comparable to that of the
moderate 60:40 benchmark. Tactical
strategies and the Swan active option strategy provide lower drawdowns than the
benchmark and some provide consistently higher returns. Passive option strategies were
disappointing.
Tactical strategies
control volatility by adjusting bond allocations. The presentation speculates about
performance in a low interest rate environment.
· Slides
· Text
Assessing Tactical Investment Strategies Using Short Histories, Silicon Valley Computerized Investing Group, April 1, 2021.
The primary investment
risks are volatility, the day-to-day changes in portfolio value; the risk pf
large price drawdowns during market corrections and the longevity risk of
running out of money before death.
The presentation will Illustrate how volatility, drawdown and longevity
risks can be used to rank investment strategies. Tactical strategies are
identified which, historically, have had substantially better drawdown and
longevity characteristics than the traditional 60:40 strategy while exhibiting
similar moderate volatilities.
· Slides
· Text
· Appendix C. Spreadsheet of Strategy Statistics
What Tactical Strategies Offer Dividend Investors. AAII Portland Income Special Interest Group, July 11, 2021.
Understand the differences between dividend and tactical
strategies with respect to initial spend, volatility, drawdown
and longevity. Learn how the AAII
dividend portfolio has performed relative to dividend funds. Understand how tactical strategies
mitigate volatility and discover two tactical strategies which have attractive
backtests over the past thirty-one years.
· Slides
· Text
Why More Investors
Should Be Using Tactical Allocation As Their Core
Investment Strategy, Silicon Valley Computerized Investing Group, October
6, 2022. With the exceptions of
near-term saving goals and tax intensive environments, tactical strategies are
well suited to the core investment portfolio. Not as a sliver but as a major portfolio
commitment.
Backtests using investable securities over the past fifty
years show that tactical strategies which are suitable for the core portfolio
would have nearly doubled returns net of inflation as compared to return of the
traditional 60:40 portfolio.
Volatilities would have been similar, drawdowns
would have been reduced and the probabilities of achieving financial goals
would have increased. For example,
the 4% rule for spending in retirement would have become the 6% rule.
Some tactical strategies do not require market timing.
Tactical strategies which are suitable for the core
portfolio are not difficult to implement.
For those without the time or inclination, comparable tactical
strategies are available in separately managed accounts.
Relative strength shows that individual tactical
strategies do not consistently outperform over time and that the investment
horizon must extend for a decade for the outperformance to become
pronounced. Inconsistent
performance implies that the investor who extrapolates from only a decade or
two of historical backtesting does so with
considerable risk.
· Slides
· Text
Lessons from a Hundred Year Tactical Allocation Strategy, Silicon Valley Computerized Investing Group, December 7, 2023.
· Slides
· Description of Investment Universe
I would appreciate learning of errors and being advised about other tactical approaches. Please remember that future allocation signals may not be as effective as past signals.
Peter James Lingane
peter@lingane.com
Updated December 7, 2023