Press "Enter" to skip to content

Technical Analysis: The Problem with Trends

Some individuals use the trends visible in stock price movements to make investment decisions.  This is otherwise known as Technical Analysis.  But does this type of stock analysis work, or is it just searching to see patterns and trends where there are none.

How do Analysts Evaluate Stocks?

So let me start by explaining, most stock analysts lie in one of two camps.  

Fundamental Analysis

The first category of analysts conducts what is known as fundamental analysis.  These folks analyze the underlying health of companies in the stock market.    They then utilize this analysis of things like debt levels, macroeconomic situation, profits, revenues, and other indicators to make a decision on whether to invest.  I previously wrote about the difficulties involved in these techniques here.

Technical Analysis 

The second category of analysts utilizes what is known as Technical Analysis.  The topic of this post, technical analysis is basically looking for patterns in the prices of stocks.  The most common forms of technical analysis utilize some sort of moving average.    From these moving averages, the practitioner attempts to determine whether a trend, support, or momentum of a stock indicates the future stock direction.

Rationale for Technical Analysis

Proponents of Technical Analysis tend to believe that fundamental analysis won’t work since all known information is already priced in.    For example,  things like the US’s large amount of corporate or sovereign debt are already priced into a stock.  As such, to someone who practices technical analysis it is not worth analyzing the fundamentals.  Instead, it is better to analyze stock movement trends.

While technical analysis does line up with one of the main ideas of the efficient market hypothesis, that stocks price in all available information, the similarities end there.    Technical analysis is predicated on the belief that past stock movements can allow one to determine future movements.

Past Performance Does Not Equal Future Results

If I were to stop this post right here I would with the obvious quote “past performance does not equal future results”.  There is little to no evidence that past performance can be a guide to future stock results via charting or other technical analysis techniques.    But we strive to give a deeper analysis of the issue so I will.

Techniques of Technical Analysis

The techniques of technical analysis are fairly sound when analyzing data.    My mind immediately slips to a control chart in six sigma when I discuss Technical analysis.  Basically, a control chart maps out the results of some sort of test. 

Control Chart

Anyway, on this chart you record each recordable result of an experiment.    Also to this chart you add the mean and standard deviations of the data.    As you make changes you observe where the data lies in relation to those standard deviations.  If it begins to hit beyond a certain level of standard deviations repetitively you note the process is not in control and may be trending a certain direction.

Control Chart 

Sound Statistical Analysis as a Foundation

Control Charts are a simple way to see if a process is on the move and portends future results.  Similar analysis exists all over-engineering with variations including moving averages, control points, and basically any other data analysis technic. Technical analysts utilize all of them.    Again a sound general technique for analyzing data.

Lacks Control of Variables

But…  It overlooks one important thing.  Control for variables.  When you analyze something as an engineer or scientist you tweak one lever at a time in a  controlled manner.  You then observe the results of those changes over time.  But our economy is not one controlled input.  Quite the contrary, it is an ever-changing system.

Let’s take a hypothetical stock X.  Stock X shows via technical analysis that it is trending in an upwards price direction and has been for some time.    This might be because of great management, the sweet spot for product placement, growing demand, investor sentiment, or anything else we want to consider.    Now if we assume all those things will remain constant, or continue to improve, then tomorrow stock X should follow today’s trend.  A great argument for technical analysis.

Tomorrow May Not Be Like Today

The thing is, tomorrow could be completely different.  And not in an I changed one thing way.    A new competitor could enter the field.  The government could take action against the company.  A recession could start.  Company management could even be accused of maleficience.     It could just be as simple as an article appeared claiming the stock was about to fall.   Tomorrow could hold any one of those and many other events, sometimes all of the above and sometimes none.  There is no way to know.

Berkshire Hathaway as a Recent Eample

Want an example? Take Berkshire Hathaway’s recent issues.  Their stock was plugging away on the way up when out of the blue comes the Kraft Heinz debacle.  There is nothing in the stocks prior price movement that could have predicted that price change.  And that my friends show the glaring reason why I don’t do such analysis…

Market Trends Versus Individual Stock Analysis

But wait, what if we are not analyzing for individual stock movements but instead market trends.  Perhaps you come up with a hypothesis that the market will always move up on the last day of each month as 401ks invest.  Or maybe the end of December should always go up since it is after bonus period.    These are alternative forms of technical analysis.

Trending of Whole Markets has Value 

This might be where things get interesting.    This type of trend analysis can work.  Looking for patterns in trading patterns as a whole has value.  Watching the flow of money.  Those things can and have worked for some in the past.  At least they can until everyone knows about them.

Trends are Only Stable so Long as The Public Doesn’t Notice Them

You see, if everyone knew that the stock market would go up on the last day of the month as 401ks invest, then everyone would buy stocks the day before that last day. This would drive up the price on that day, decreasing the amount of possible increase on the last day.    Furthermore, those same people might sell in mass on the last day to lock in their gains.    The result, after the cats out of the bag, would mean the last day of the month becomes the day that loses.  

Only The First Mover Need Apply

In other words, this type of technical analysis, analyzing flows, only works if you are the first person to discover it and you keep quiet.  Now we’ve talked before at length of how you are one person analyzing stocks in your spare time to invest.  A hedge fund meanwhile is many super smart people analyzing stocks to invest as a full time job.    So what do you think are the odds that you would be the first person to discover such flows?    Or for your chosen hedge fund investment what are the odds no one would leak that theory for their own gain?    That’s right, near 0.

Granularity, One More Critique of Technical Analysis

Now I do have to level one more criticism at technical analysis before we close.     That is the issue of granularity.  Let’s assume for a moment, despite my entire post to the contrary,  that past performance can predict future results. 

Now let’s take another example stock.  Let’s say Amazon.  If I were to look at Amazon on a yearly basis of granularity what would I see?  A stock with nearly straight up increases.    But let’s drill deep in, what if you looked at Amazon on a  daily basis for the last month?  The movements are all over the place but the average appears flat line.  Depending on length of view you’d see no movement up or down for the stock.    Now finally, take that same stock and zoom into a mid view of say monthly over the life of the stock.  Suddenly a very recent downward trend shows up.  So what is the trend for Amazon?  Is it up, down or sideways? *Note This was written in January, Amazon stock has since recovered somewhat with the market.

Attempts to Control for Granularity

Technical analysts try to utilize various methods to control for these granularity issues.  Moving averages, comparing short term and long term moving averages, and other combinations of analysis come to mind.  All legitimate methods for controlling for granularity, but again only when you control the variables.  They only work when you know what to expect going forward and can thus determine how to look at the data to control for the variable of time.    IE you need to know something about the expected future inputs to decide how best to look at the data.  No one can legitimately say that for stocks.   

Technical Analysis in the Main Stream

So needless to say I do not favor technical analysis.  Every time I read in the paper about something like the death cross I just shake my head.  Even if it is correct, If it’s in the paper then don’t expect it to make you wealthy since it is not a secret.  I guess what I am saying is I see about as much value in the mainstream practice of technical analysis as I would from buying stocks based on a horoscope.  It might be just as accurate. 

Is anyone reading this a proponent of technical analysis of stocks?  If you disagree with my analysis (or agree for that matter) I’d love to hear from you in the comments.

4 Comments

  1. Matt
    Matt May 15, 2019

    I pay attention to technical analysis, but only in a psychological sense: that is, a material number of people follow this methodology, so if I am planning an action (buy or sell) or trying to understand the current market price of a stock, it might provide some insight, or help me to maximize my own strategy.

    Otherwise, I agree with all your thoughts. Humans evolved into pattern-seeking animals, which has helped us greatly to survive to this point. But we also fall prey to seeing patterns (and inferring causality) where they are not really present. This was explored very well in Nicholas Nassim Taleb’s Fooled by Randomness.

    My other favorite thought along this line is George Box’s quote, “All models are wrong, but some are useful.” Whenever you are using a tool, you need to understand the limits of that tool and, in particular, where it is and is not useful. Ignoring that is how people can go off the rails.

    • FullTimeFinance
      FullTimeFinance May 15, 2019

      Love the quote, haven’t heard that one before but it fits.

  2. Nate Matherson
    Nate Matherson June 14, 2019

    Really in-depth post, thanks!

    I am interested in the idea of technical analysis, but like you I find myself shaking my head a lot. I am in it for the long haul, trying to buy companies that will last as our world changes. I guess TA could be great for timing entry, but if you dollar cost average it will probably average out about the same.

    • FullTimeFinance
      FullTimeFinance June 15, 2019

      I’m not even sure it works for timing entry. Most studies show time in the market trumps all.

Leave a Reply

Your email address will not be published.