Hi Guys,
We’ll never know when to hold’em; We’ll never know when to fold’em, if we don’t know the odds. For completeness, you also need to know the Expected Payout, but that’s a subject of another story.
When in Las Vegas, it is never a smart or profitable decision to play Keno; the odds are terrible for the player. The only consistent winners are the casinos. Not surprisingly, the luxury casino/hotels have the poorest payout odds, but they do employ the best looking Keno runners. There are always hard tradeoffs.
From a fair odds perspective, Blackjack should be the game of choice. With just a little study and a persistent discipline, the odds at “21” can be effectively equalized without defaulting to a dangerous card counting system.
As Louis Pasteur remarked; “When observation is concerned, chance favors only the prepared mind”. And the prepared gambler.
That’s, of course, only true sometimes since luck is an uncontrollable factor. Emotions and hope often prevail over logic in decision making. The winning odds in lotteries are lousy, and most everyone knows that the odds are lousy. Yet lotteries attract huge monies. Hope for riches is an incentive that folks can not disregard. Gambling on an uncertain investment return has sport-like characteristics.
Most folks would likely prefer the clockwork precision world of Isaac Newton. When he discovered the classic mechanical Laws of Motion and Gravity, many of the 17th century hoi polloi believed that happenings were preordained and totally predictable. That faulty interpretation ended when Werner Heisenberg introduced his statistically based
Uncertainty Principles to explain some of physics more complex and perplexing issues. Life is fuzzy.
Like it or not, the investment world is more like Heisenberg’s model than like the Newtonian perspective. Uncertainty rules the investment day so statistical expectation is a superior framework over a clean, single point forecast for planning purposes .
Consequently, to tilt the odds in the investment universe, an understanding of statistics, their merits and their shortcomings, is essential. You must know the odds, estimate the expected payoffs, and have a plan to exploit them in your favor. An additional dimension of complexity is introduced because these odds are not constant over time.
Based on long-term statistical data tables, we know that the US equity market has a modest 0.03 % return daily upward bias. That’s the financial incentive to invest despite its roughly 1.00 % daily price volatility (standard deviation).
These same statistical pricing data sets amply illustrate that time is an investors best friend. On a daily basis, the S&P 500 Index (serving as an equity market proxy) has delivered positive outcomes 53 % of the time. As the time scale progressively stretches to weekly, monthly, quarterly, annual, and 5-year increments, the likelihood of positive outcomes increases to 56 %, 59 %, 64 %, 71 %, and 81 %, respectively. Indeed, time is a powerful ally.
The sell-in-May axiom reminds us that the progress is not uniform. The long-term monthly data sets show two negative months (February and September) with weak positive returns recorded in the summer months. Even this weaker period generates likely rewards that exceed the current low yields offered by alternate fixed income candidate vehicles.
Overall, because of their perceived risk (loosely measured by return volatility), equities have awarded investors with a risk premium of 4 to 6 % over a various assortment of zero risk options like short-term or 10-year government bonds. That’s likely to remain intact going forward.
However, as a general observation, predicting future market returns is a Loser’s game, especially for the short-term annual prediction drill. Numerous Guru attempts are made producing an equal number of failures. Longer term forecasts (like 10-year periods) are imperfect, but have proven to be much more reliable.
Recent academic and industry studies suggest that some forms of P/E ratio can explain about 40 % of the equity market movements. That’s surely a step in the right direction, but it is a double-edged sword. That same finding also means that 60 % of the action remains clouded in mystery. In the end, forecasts are not perfect so portfolios should be assembled to reflect this uncertainty.
I interpret that bottomline conclusion in terms of my portfolio’s asset allocation. To make the portfolio robust a
gainst this uncertainty, the portfolio must retain its broad diversification characteristics. It must protect a
gainst an array of future scenarios that include downside probabilities.
Here is a reference to a nice Vanguard study prepared late last year that catalogues candidate forecasting signals and their shortfalls. The title of the work is “Forecasting Stock Returns: What Signals Matter, and What do They Say Now?” Here is the Link to the document:
https://personal.vanguard.com/pdf/s338.pdfThe study explores over a dozen candidate market direction signals to project future equity returns. Various P/E ratio formulations did the best job. Most others failed miserably. The Vanguard researchers conclude that attempts to predict near-term equity performance (like next year’s returns) is a lost cause; one-year estimates are worthless.
The study further concludes that there is a modest ability to capture reasonably reliable 10-year annualized returns. The P/E 40 % explanatory power observation that I mentioned earlier was gleaned from this study.
For forecasting purposes, Vanguard uses a Monte Carlo simulation code that explores market return as a function of leading signal correlations. This proprietary tool is called the Vanguard
Capital Markets Model (VCMM). A primary output of the code is a multiyear market returns projection map. One dimension of that map is an event probability distribution.
When applied to the current economic and financial environments, the simulations are more positive than negative about the upcoming 10-year equity market prospects. Vanguard sees a higher likelihood of positive returns, slightly muted when contrasted a
gainst historical equity rewards.
Of particular interest are the final charts in the referenced report. The plots show the predicted return’s probability distributions for this year, and the next 10-year period.
Note that the VCMM code has the rare capability to forecast low percentage (probability), infrequent fat tail outcomes, essentially Black Swans. That’s an exciting feature. However, I seriously doubt the tools have sufficient accuracy at these extremes given how far removed these events are from the data rich historic average returns.
I recommend that you visit the article. Sorry that I didn’t post this Link earlier, but I neglected to read the report for several months as it languished on my to-do list.
Best Regards.
