My Photo
Note: Jeff does not accept guest blog posts on A Dash of Insight.

For inquiries regarding advertising and republication, contact [email protected]

Follow Jeff on Twitter!

Enter your email address:

Delivered by FeedBurner


  • Seeking Alpha
    Seeking Alpha Certified
  • AllTopSites
    Alltop, all the top stories
  • iStockAnalyst
Talk Markets
Forexpros Contributor
Copyright 2005-2014
All Rights Reserved

« Weighing the Week Ahead: Approaching Dangerous Curves? | Main | Weighing the Week Ahead: Will US Leaders Become Cliff Divers? »

November 07, 2012



From an investor's perspective, I see two games taking place with respect to the business cycle. First, there is the actual business cycle. Investors should define the cycle in a way that best suits their needs. This doesn't mean we all have to create our own models; but it does mean there isn't one "true" model we all must use.

Second is the public perception of the business cycle. This is where work like that of Chauvet and Hamilton comes into play. At first glance, my guesstimation is that the lag between the business cycle dates and when the NBER announces them serves to flatten the cycle. A lag between the start of recession and the announcement of the start date can serve to mute the rate of downturn, as some people maintain the economy is slowing but not in recession. Likewise coming out of recession those who still believe the economy to be in recession will dampen the rate of recovery. Shortening the time between actual date and the announcement of that date could steepen the cycle on both sides of the peak/trough.


The recession probability indicator doesn't look all that useful. From the charts it certainly doesn't look predictive at all, and while it is highly correlated with the NBER dates, I'm skeptical that they really provide any extra information than just looking at the quarterly GDP estimates and monthly employment. In fact, it really just confirms my pre-existing conclusion that the only way to improve our real-time ability to monitor the economy is by producing monthly instead of quarterly GDP estimates.

Meteor Blades

A thoughtful presentation and a great reply to an unthoughtful comment on your presentation.


If it is the case that the margin between the reliability of prediction 18 months out when compared to the actual situation that eventuates is large, with known methodology, isn't is therefore the case that new methodology is more liklely to be able to develop tighter predictions than current methodolgy than would otherwise be the case?

And therefore doesn't that diminish the importance of a proven track record?

As long as you diversified your risk you would in fact be better off giving a larger weighting to the predictions of the new approach as a result - wouldn't you?

I mean, that the future is identical to the past, and thus what has been done before, as long as it has been done a lot, is the best we can get would seem to not be consistent with an acceptance that technology drives the capacity to produce a particular good - wouldn't it?


If you score more points you are likely to win more games.

Therefore if you raise the cutoff point the likelihood of your prediction being right is logically more likely.

In the graph the higher the redline rises the greater is the association with a recession.

The case is proven - isn't it?


It would be interesting to have an estimate of alpha and/or beta risks - that is rejecting a true test or accepting a false positive. It is really rare in any hypothesis testing process to see both low alpha risk and low beta risk - especially so in 'social sciences' like Econ. Knowing an estimate of that would be helpful in knowing what to make of the various recession calls.

Now having a high alpha or beta isn't bad - sometimes you are willing to accept a few false positives (say throw out good parts) if one missed bad result is a disaster (say a plane falls from the sky). The point though is you should know the relative alpha beta and report it and if you don't know them report that too.


Steven -- It is on FRED -- the fine data source maintained by the St. Louis Fed. They provide 61,000 different data series, an excellent resource for serious researchers. There is also an excellent charting facility and references to check out the data source. They clearly provide a link to the original study, which I used as the basis for this post. It was available to anyone, as was the data on the authors' web sites.

Complaining that "the Fed" has it on their site is like objecting about some book in the library. The presence on the site is a resource, not an endorsement.

This is really an important lesson, and your comments are helping to emphasize that.



Steven Kopits

And if so, what's it doing on the Fed's website?

Steven Kopits

OK. So there are three criticisms.

One is that this tool is subject to false positives, and therefore is not reliable. Jeff notes: "Look at the chart above and pretend that we were going back to the future in 2005. You could have drawn the line at 15% and said everything that Lance says now. You would have been completely wrong!"

This, of course, is incorrect. You would have been wrong 50% of the time--and right 50% of the time.

Next, you need 80% for three months, but apparently that's not visible on the chart.

And finally, data appears to have been smoothed away (and this appears to be some on-going thing) which means the graph omits certain data points which would be important in our interpretation.

Now, if you have to make this many apologies for a simple visual tool, then it's just not a good tool. Is that what you're trying to say?

Don Ake

You need to have a significant recovery before you can have a significant recession! -

Colorado Investment Manager

Great post and Ritholz approved as I was directed here from his Big Picture blog.


Im not an economist, just a network engineer who is trying to teach himself a lot about markets and money.Yes, I plan for you all to belittle and trash me but while I still have the right to free speech ill say what Im feeling. My gut tells me the country will see a DEEP recession and we are on the cusp of going into one. Look at dividend payments on a whole over the last year, the trend is NOT good and very steep on the downside as well because companies have been hoarding cash so they can survive the next downturn. Company earnings have been really bad, just take a look at UPS’s last earning report, even McDonalds took a hit. That right there is cause for concern. The tanking markets are not a knee jerk reaction to the re-election of an anti-business big government socialist, although his re-election hasn’t helped people feel confident in the business world, unless you run an abortion clinic or union based shop. The chatter all over Wall St. is about the next recession…and its on its way regardless..

Scott Rothbort

Jeff - as always excellent work. I would like to refresh people's attention to the Hindenburg Omen scare in 2010. That was also faulty data mining research. Unfortunately since the tech bust a decade ago, negative commentary gets more attention without having to pass the legitimacy test than good old fashioned quality research - Scott


Swiss Maven -- There are many roads to investment success. The Warren Buffett school basically ignores smarty pants guys like you and takes advantage of stocks that represent good values. If you subtract Apple's cash of $125 per share, it is trading at a single-digit P/E multiple. For most of us, that is attractive.

Most people who have tried to trade Apple from the technicals find themselves chasing rallies from the sidelines.

When you have more experience, you will learn to have respect for other approaches and be less dogmatic in your opinions.

Check back with me in a year or so:)

And especially if you want to drop the cloak of anonymity and reveal some long-term performance!


Swiss Maven

This is a "dash of insight?" Holy Crow. AAPL blew its top at 700, is already down 23% to 540 in just 6 weeks, and you are BUYING it? Can't you see that the macro-trend support has been broken? Any rallies from here out will no longer be “bull market rallies” but rallies in what is now a new technical bear market for AAPL. It will not again see 700 for many a year. My insight is that your insight is worthless.


scm0330 -- I have often noted that the stock market is not a futures contract on GDP. People form unsound opinions about the economy all of the time.

I can't predict the zaniness of others, but I can do something about understanding the economy. Stocks respond (ultimately) to corporate earnings. Earnings get whacked during recessions. I therefore look at the trend in earnings and then adjust for financial and recession risk.

If you correctly view the recession forecasting methods as identifying where we are in the business cycle, it is helpful in identifying the right stocks and sectors. This may require accepting a little pain until the data finally convinces the skeptics.

Great question, which I should discuss more extensively.



RB - Here is a good way to think about it. I am not a recession forecaster. I do not have a method or any independent research, and therefore I have no track record.

Because of my background, I am very good at reviewing the research methods and conclusions of experts in several different fields. Think of me as a finder of experts.

When someone offers an opinion about recessions (or any similar topic) they should either have a proven record using their own research, or good evidence that they have found others who have strong evidence.

So Josh and Barry, who both are willing to change their opinions with the evidence, are free to draw conclusions. I don't know of any methods that have proven effectiveness for an 18-month prediction!


paul t

Thank you very much indeed for this and similar posts. You and Krugman have the delightful habit of naming names!



Thanks for your perspective here. I had a related question.

You spend a lot of time discussing whether or not a recession is in the offing, based on the best forecasting methods you're able to identify. How, if at all, do you relate recession calls to owning stocks? There are periods in time that are not recessionary, but are not rewarding to equity investors. Conversely, some great times to own/increase equity exposure is in the teeth of a recession. (Obviously, identifying inflection points in the business cycle is very important to success here.)

I guess what I'm trying to understand is how to relate your extensive discussions surrounding recession forecasting with the decision to own stocks.


Scratch the last line !

The comments to this entry are closed.