More on Technical Analysis

by: Felix Salmon

It seems I hit a nerve yesterday: 26 comments and counting here, as well as 18 over at Seeking Alpha, and quite a few more chez Ritholtz both yesterday and today; Barry Ritholtz also has a more formal response here. Adding it all up, it seems that the majority of the comments disagree with me, but they all seem to disagree for different reasons.

Some say that technical analysis is a trading tool which tells you how and when to enter or exit positions which you're getting in or out of for fundamental reasons. Some say that it's a way of quantifying market psychology, and that there are real causal reasons for what it says. Some say that it's a fine art, which only works when used in the hands of the adept. Some say that it's all one big self-fulfilling prophecy, that it works because people believe that it works, creating a herd mentality. And then there's my favorite comment of all:

 TA showed me a seven-year double top last year (2000-2007) and I sold all my stocks on the day in July that the DOW closed at 14,002 and have been in a money market fund ever since...

Evidently for some people technical analysis is so powerful that recognizing a "seven-year double top" can and should be a reason to sell every single stock you own.

And a lot of people decided to take the ad hominem route, asking whether I was a trader, how much money I've made in the market, etc etc. For the record: I'm not a trader, I've made no money in the market, and I would never, ever recommend that anybody ever come to me for investment and/or trading advice. The full extent of my advice on both fronts is simple: don't do it. Don't try to be clever, don't try to beat the market, don't think you're smarter than anybody else. If you're asking me for advice, that's a pretty good indication right there that you're not going to be any kind of outperformer. On the other hand, don't think you're going to outperform through use of technical analysis, either.

The most basic argument against technical analysis is the same as that against any kind of active investment strategy. There are technical analysts, and then there's everybody else. If the technical analysts outperform, then everybody else, in aggregate, must be underperforming. Put aside the passive index performers: they simply track the market as a whole. So who are the investors who consistently underperform the market? Do they happen to be exactly or even mostly the same as the group of people who don't use technical analysis? It seems improbable, to say the least, especially given the fact that celebrated investors like Warren Buffett and Peter Lynch loudly and frequently declare that they think technical analysis is bullshit.

I think the silliest arguments in favor of technical analysis are the ones which basically assume their conclusions:

 Technical analysis is a tool. Like any tool, it needs to be used correctly, or the tool user is likely to sustain injuries... TA expertise takes time and effort... The reading part of a TA education would begin with Technical Analysis of Stock Trends (Edwards and Magee)...that's about 700 pages. Then add more reading, and a few years of daily market observation and price behavior, and then, maybe, you're qualified to make pronouncements on the usefulness of this tool.

The problem here is that the argument is unfalsifiable: anybody who has anything bad to say about technical analysis, and anybody who's lost money using technical analysis, is by definition unqualified to make such a determination. It's worth remembering that given the sheer number of people who use technical analysis on a daily basis, statistically speaking a lot of them will make money quite consistently on say an annual basis for many years. But the existence of such people is no indication that technical analysis is any better than picking stocks based on Super Bowl results.

As for the idea that technical analysis is one big self-fulfilling prophecy - well, I think that's easily disproven by the sheer quantity of disagreement that there is in the field over pretty much any chart.

Barry's take is more sophisticated - but then again he says right off the bat that technicals "are not a way to forecast the future", which does kinda make you wonder what they're good for. He asks what is for me a very easy-to-answer question:

 If you could look at one and only one source before buying your next stock or fund, which would you choose: a fundamental analyst's report (with no charts in it), or any chart of your choosing?

I'm no fan of fundamental analysts, but largely because they spend altogether too much time looking at the share price. An analyst's report which was written with no concept of the share price could, I think, be hugely valuable. The best investments on both the long and short side are where you buy a company trading at a fraction of its value, or where you sell a company trading at a large multiple of its value.

Let's say you have a good fundamental analyst's report which came to a compelling conclusion that a company is worth somewhere in the region of $35 per share. Then if it's trading at $7, you buy. If it's trading at $150, you sell. No chart could ever be clearer than that. The problem is that fundamental analysts tend to be shy, and generally base their valuations on wherever the stock is trading right now. And that's not very useful at all.

Insofar as technical analysis is any better than pure guesswork or astrology, I think it works as a way of gauging market sentiment and psychology. But here's the problem: everything in technical analysis seems to work backwards. No one ever seems to start from first principles and make predictions of what would happen to price-and-volume charts when sentiment does this or does that.

Instead, people start with historical charts, and look at the way prices and volumes have historically behaved during periods of given market sentiment. It's a way of guaranteeing that you're always right about the past, but I'm far from convinced it's at all useful when it comes to being right about the present or the future. As any quant fund will tell you, those kind of models work until they don't. Meanwhile, the technical-analysis crew seems to think that they're dealing with immutable, permanent, semi-scientific probabilistic laws. In the markets, there's no such thing.