Are you really a chart expert?

When it comes to charts, everyone is an expert -- or so they think!

It can be expensive to be overconfident.  In this post and another on my agenda I will illustrate the problem.  If you get both problems right, you can have confidence in your chart-reading skill.

Tracing a Dangerous Path

One of the most common charts we see compares current circumstances to something that happened in the past.  It often comes as a warning.

My email today provided this example, suggested for consideration by one of my most astute friends.  He is also a very successful investor and market observer.

 

 

So we are tracing the same steps that we did last year, with economic disaster to follow.

But wait --- there's more!

We are following the same path as in the Summer of 2011 and the debt ceiling debate.  Oh my!

My friend did not reveal the source of the charts, but I did not need a hint.  I get similar emails every week.  The charts are all from the same source.  The perpetrators have a wonderful and profitable business model.  They have identified a market of people who want to be Scared Witless (TM euphemism, OldProf).  They love to hear about conspiracies.  They want to have their worst fears confirmed.

My own market is much smaller, but I am proud of the readership.  It consists of people who employ critical thinking when evaluating evidence.  They want to profit from their investments.

Problems with the Chart

Regular readers will have already pounced on the main problem with this chart -- the twisting of scales.  Modern computers and software have created enough charting power to overwhelm the (lack of) skill of the user.  You can now look back in history, adjust the scales from two different time periods, find a brief period that seems similar, and then show a prediction.  Bravo!

You could just as easily find a time period that showed the opposite.

When you read something like this, it is time to turn the page.  I recently awarded The Silver Bullet to Tom Brakke for exposing this kind of deception.

Problems with the Logic

Moving beyond the chart itself, let us suppose that you paused to consider the actual reasoning.  Try to forget the chart and use words instead.  Here is my best effort:

The Citi Economic Surprise Index has moved higher, just as it did last year.  It is tracking last year's move almost tick-for-tick.  We can see the collapse last year.  Ergo, we should expect a series of economic disappointments to start 2013.

It seems pretty foolish when stated that way, but it is the story they are selling.  Please note what the chart does not say:

• This is something that happens every year (not just one time);
• This is something that happens whenever there is a similar negotiation (not just one time).

There are many differences between last year and this year.  Just for starters:

• Election versus non-election year
• Debt ceiling for a temporary change versus long-term policy
• Public preamble and preparation.

But those are only starters.  The basic problem is the following: 

You cannot make valid inferences from one prior case.

Anyone cherry-picking a prior history to do this is selling snakewater.  Beware!

An Alternative Viewpoint

Let us pretend that we did not start with the original chart.

In fact, I did not.  I have often looked at the Citi Surprise Index in the past, but I have not found much predictive value.  If the economy turns positive, the expectations increase.  I have tried but cannot sync it with any meaningful prediction.

Here is one take, from Easynomics:

Keep in mind what is basically happening, as it is usually a cycle.  Expectations rise as a result of improving data, and then it becomes more likely that data will disappoint.  It doesn’t actually mean the data get worse, only that they disappoint versus expectations.  The reverse then happens in order to complete the cycle.  We hope that the downturns don’t go as deep as the upturns go high.

Dr. Ed also follows this index.

 

Conclusion

Good luck in tracking stock prices versus the surprise index.  I do not see a fit, but suggestions are welcome.

There is a lot of irony in the original presentation.  Step back from the charts and think about it.

If the Citi index showed that there were disappointments, that would be the end of the story.  The perma-bear site would simply report that the macro conditions were bad.

Since results have been beating expectations, and you are selling snakewater, there is a problem.  Solution?  If you do enough data mining you can find a time period where beating expectations was actually bearish...  Black is white.  Bad is good.

Let's see -- last year - - wonderful.  One case is enough.

If it does not fit last year, let's go back in time......

Comments

[oldprof]

Dan -- Thanks for highlighting this. Barry's post is here: http://www.ritholtz.com/blog/2012/12/10-midweek-pm-reads-5/

It is nice to see Dwaine getting some well-deserved attention.

As you (and other regular readers) know, I have been highlighting this work for over a year!

Thanks again,

Jeff

[Proteus]

Torture the data until they confess.

I've played with the Citi index before, and here's what I see...
The Citi index goes up and down, sometimes a lot. The peaks can be more extreme than one might imagine. If you see a peak at an extreme value that is lower (or higher) than any others in the past 10 years, think about buying (or selling). Otherwise, it's just a moderately interesting chart that sometimes leads and sometimes lags market turns, and sometimes has the wrong sign.

[Eric]

I think the other problem with this chart comparison is that it does not acknowledge what really precipitated the correction of 2011: the S&P downgrade of US debt.

[wkevinw]

Great stuff as always.
Back to the "scared" internet.
I was wondering what you think of Kyle Bass' prediction of a Japan bond/inflation/currency crisis.

Thanks for your hard work.

Kevin