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

For inquiries regarding advertising and republication, contact

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

« Starting with the Result: The Blowout of the Week | Main | The SEC in Action... err, Partial Action »

December 16, 2008


Mike C

"What time frame do you like? At least you are asking the right question!"

Well, that is something I've been thinking alot about lately. What time period of the U.S. economy and stock market is the most relevant analogue to today? On some level at some point there is a qualitative subjective judgement call here, right? This isn't physics or engineering with constant rules that always apply.

If you'll recall, I asked a question about the Kondratieff cycle. Briefly put, the 4 seasons of the Kondratieff cycle last about 60-70 years, and the hypothesis is that this deflationary, deleveraging cycle is the onset of the Kondratieff winter, and one must go back to the last Kondratieff winter for comparison which takes you back to the 1930s/1940s.

To oversimplify, there has been a 50-year buildup in debt that MUST unwind and run its course. I think David Merkel over at Aleph Blog has been stressing this point recently repeatedly. Does this have implications that run a number of years for the overall economy, corporate profits, and what are considered "fair value" stock valuations? Do we actually need to throw out everything from 1982+ or post 1987 crash in that it will actually mislead rather then inform about the next 10-15 years?

Bill Gross has remarked that what we are seeing isn't cyclical or even secular, but trans-generational. Here is another interesting piece on that theme:

Is something different this time? Compared to what time?


Mike C -- As usual, a very good question!

Most people developing a model use a simple rule: Get as much data as possible and use it. They do not consider relevance. This applies to some of the big name types.

I try to use only relevant data. It depends upon the problem. For many purposes I use post 1987 because of the effect of the Crash. For the performance of our methods, we use the entire history, since it spans a couple of market cycles.

When it comes to questions of the Fed, economic policy, and forecasting, it is pretty obvious that information from an era where there were computer models were in a formative state and the Fed had a different approach -- these data are not relevant.

Anyone looking at a chart of the frequency and duration of recessions can see this at a glance, although most do not bother to do it.

Briefly put, I do not care what happened under Taft, Hoover, or even Ike. Nixon and Ford were pretty unusual times with strange policies that we will never try again.

What time frame do you like? At least you are asking the right question!


Mike C

"There are so many who think they know so much. In fact, investors should look not to a single market call, but to long-term history."


In terms of "long-term history", especially in the context of building and using models do you still think it is only valid to go back to 1980+ or does one need to go back much, much further perhaps even to 19th century data to really get a full "long-term" picture? I have an opinion, but I can usually count on you to provide a very well-reasoned counterargument which is useful in avoiding confirmation bias. :)

The comments to this entry are closed.