November 2, 2020: Polls Are Useless vs. This Investment Strategy Doesn’t Work!

Re-balance Cycle Reminder All MyPlanIQ’s newsletters are archived here.

Regular AAC (Asset Allocation Composite), SAA and TAA portfolios are always rebalanced on the first trading day of a month. the next re-balance will be on Tuesday December 1, 2020.

Please note: As of March 1, 2020, we officially phased out our old rebalance calendar for both SAA and TAA. They are now always rebalanced on the first trading day of a month.

As a reminder to expert users: advanced portfolios are still re-balanced based on their original re-balance schedules and they are not the same as those used in Strategic and Tactical Asset Allocation (SAA and TAA) portfolios of a plan.

Polls Are Useless vs. This Investment Strategy Doesn’t Work!

By the time you are reading this newsletter, it’s probably a nail biting time (less than 24 hours to go) to wait for the US president election result. The voter turnout of this election might be the highest presidential-year turnout since 1908. To make the matter more serious, it’s hotly debated and highly polarized.

What’s interesting is that we can see/learn something from election polling and apply this to investing in financial markets.

Polls are useless! Are they really?

Many people are dismissing election polls and think they are useless. This is because most polls faltered in the 2016 presidential election: they were predicting Hillary Clinton’s victory all along leading to the election night and then people were shocked to see Donald Trump won the election.

Of course, some pollsters were quick to point out that they were actually not that far off in 2016: in fact, Clinton did win the popular vote in 2016. So technically, the polls were not completely wrong at that time.

Now that we are on the eve of another election night, will we see the same thing happen again as at the moment, many polls are again forecasting the Democratic candidate Joe Biden will win the election? For example, the popular FiveThirtyEight.com tracks and synthesizes all major polls. Based on their statistical simulation, at the moment, it has the following forecast:

Before you quickly celebrate or dismiss the forecast, you are reminded that this newsletter does not intend to endorse or predict the election result. What we would like to do is to objectively understand and analyze this a bit deeper as how we approach to this type of forecast or methodology is very relevant/similar to what we do in investing.

What the above figures say is that among 40 thousand simulations, it observes 89 out 100 times when Biden wins while Trump wins 10 in 100 times( the other 1 left is hung or tie election). Put it another way, the likelihood or probability for Biden to win is about 89% and for Trump, it’s about 10%.

If asked to interpret the above prediction, virtually everyone (even Biden’s supporters) would agree that Trump still has a chance to win. In fact, 538 does put out an article right now as I’m Here To Remind You That Trump Can Still Win.

But come tomorrow or after the election, if Trump indeed wins, people will again quickly dismiss these polls and say they are useless.

This reveals a popular human tendency: though we accept statistical probability at its face value (in 2016, 538 predicted Trump had 28.6% chance to win), when an event with low odd does happen, people will forget about statistics and just dismiss the whole process.

But having a low probability, even at 1% or less, does mean such an event indeed can happen. Technically what the Trump’s 28.6% winning probability, for example, of the 538’s prediction for the 2016 election means is that Trump will win on average 28.6 number of times if the election is held in the ‘exact’ same setting for 100 times . Or to be more rigorous, Trump will win on average 286 times out of 1000 ‘elections’ or the same ratio when more ‘elections’ or samples are carried out.

So a single actual occurrence of Trump’s win in 2016 does not really invalidate the polls. However, it certainly doesn’t give strong support to the polls either. The difficulty for many people to comprehend lies in that we can’t afford to repetitively ‘test’ out this prediction in real world: for example, how in the world can we repeat the election in the same exact settings for 1000 times?

This problem is further exacerbated as it comes with some real substantial political, social and economical impacts. People tend to get very emotional when the results are more materially relevant to their own well-being.

On the other hand, there are indeed some problems (some of them might be serious) in the methodology of a poll. For example, the bias of samples, the ‘shy’ Trump supporters etc. So, it’s actually more important to first vet and validate the methodology of a polling analytically, instead of simply waiting for ‘real time’ results.

The above discussion by no means supports either candidate or even the polls (as indeed, there are quite some methodological deficiencies in most polling methods). What we want to point out is that the process of understanding whether a method works or not is also applicable to financial investing.

Investment strategies and mistakes

Long time readers might have sensed where we are leading to on investment strategies: over time, we have published many newsletters to touch on the topic of investment mistakes vs. temporary losses or miscalls:

We strongly suggest readers who are interested in this important subject to read the above newsletters.

Similar to election polls, investment strategies materially affect people’s well-being immediately (you can see your real time account balance statements these days). They are time related (so called time series): in investments, it requires a long time to observe large enough samples to get expected average results (returns). For example, we stated in August 7, 2017 newsletter:

A strategic buy and hold strategy would require 15 to 20 years minimum to get a reasonable return
A tactical asset allocation strategy like ours would require 10 to 15 years minimum to get a reasonable return
A total return bond fixed income portfolio would require 2 years minimum to get a reasonable return (see July 24, 2017: Total Return Bond Fund Portfolios And Cash)

So people don’t have patience or just simply can’t afford to observe this in such a long time to validate or invalidate a strategy.

Similar to polling, we believe investors should not simply rely on real time results to validate a strategy (in fact, just like what’s discussed above, this is often not effective). What’s more important is to see whether an investment strategy has strong intuitive and fundamental backings such as long term stock or equity risk premium or inherent price momentum. These should always serve as the first line of filtration to weed out bad strategies that are often produced by data snooping or other statistic fluke (purely by luck). Only after that, one should look at historical back testing/simulation results to confirm/invalidate it.

After carefully vetting and selecting a strategy, the next step is to commit to it for a long enough of time: the reason is that for any strategy to converge/produce an expected or its average outcome (such as average annual returns), it just simply requires large enough of samples. During this period, one has to rigorously stick to the strategy and implement it as it is.

The second step is the one that can really challenge investors: when a strategy (or a portfolio) loses a good amount of money or underperform a market index at a time, is it simply just an inherent part of the samples or it signals its ‘real’ failure. This again leads us back to the first step: understanding thoroughly on the behavior of a strategy, even before you are committed to it is extremely important to help you navigate through real time turbulence. For example, buying and holding a stock index fund such as SPY can result in large drawdown as big as one half to two third at some point. On the other hand, buying and holding such a stock market aggregate will most likely make money in a long enough (at least 20 years or even longer) period. With this understanding, investors should be able to ride out some violent bear markets.

To summarize, we shouldn’t be too quick to dismiss or accept an investment strategy, just like an election poll. One should first thoroughly understand its methodology and behavior (and possible extreme outcomes) and then just let large enough real time samples play out in real life to gain its positive average outcome.

Market overview

It’s everyone’s guess how the short term market reaction will be for the US election. However, we are clear on some noticeable conditions including extremely high stock valuation and the rising pandemic wave. On the other hand, it’s somewhat encouraging to see that S&P 500 dramatically outperformed the beaten down analysts’ earnings expectation for Q3 2020. For SP 500 companies, based on Factset:

86% of S&P 500 companies have reported a positive EPS surprise and 81% have reported a positive revenue surprise. This is the highest positive earnings surprise since 2008.
The blended earnings decline is -9.8%, much better than -21.1% expected on September 30.

In a word, US companies have done way much better than what were expected. Looking ahead, S&P 500 companies EPS (Earnings Per Share) are:

So there won’t be much earnings growth in 2021 for S&P 500 ‘conglomerate’. The most stock price growth comes from price earnings (PE) multiple expansion.

Investors shouldn’t be complacent on the markets. As always, we should follow our strategies to navigate through this period:

For strategic allocation (buy and hold) investors, ignore the current market behavior. Remember, as what we have emphasized numerous times, when you choose and commit to a strategic portfolio, you essentially know and commit that your investment horizon (or the time you need to utilize this capital) is 20 years or longer. As we pointed out, if your investments are those diversified (index) funds such as an S&P 500 index fund (VFINX, for example), you know your money is in some solid ‘business’ that eventually (20 years later) will deliver some reasonable returns. As long as you are comfortable with this thesis, you should sit tight and forget about the current gyration.
For tactical investors, again, you have to ignore the current market noise. Furthermore, you should follow your strategy rigorously, especially in a time like this. Human emotion, both optimistic and pessimistic, and human desire, both greedy and fearful, are your worst enemies. This has been shown to be true time and time again.

Stock valuation now reached another high. For the moment, we believe it’s prudent to be cautious while riding on market uptrend. However how serious a correction might be, we have confidence in the US economy in the long term and thus in the stocks in aggregate. We just need to manage through interim losses carefully.

We again would like to emphasize that for any new investor and new money, the best way to step into this kind of markets is through dollar cost average (DCA), i.e. invest and/or follow a model portfolio in several phases (such as 2 or 3 months) instead of the whole sum at one shot.

Enjoy Newsletter

How can we improve this newsletter? Please take our survey

–Thanks to those who have already contributed — we appreciate it.

Latest Articles

Disclaimer:

Any investment in securities including mutual funds, ETFs, closed end funds, stocks and any other securities could lose money over any period of time. All investments involve risk. Losses may exceed the principal invested. Past performance is not an indicator of future performance. There is no guarantee for future results in your investment and any other actions based on the information provided on the website including, but not limited to, strategies, portfolios, articles, performance data and results of any tools. All rights are reserved and enforced. By accessing the website, you agree not to copy and redistribute the information provided herein without the explicit consent from MyPlanIQ.