If you would like an email sent to you when I update the blog please send an email with "subscribe" in the subject line to traderbob58@gmail.com. To be removed use "unsubscribe".

Search This Blog or Web

Monday, January 14, 2013

Conspiracy theory?

I am not much for conspiracy theories.  Most either provide no real evidence or what they do provide is so flimsy as to be meaningless.  However, there is something that is making me a bit curious.  Here is a chart I showed in Welcome to the recession Mr. President.

That is what the chart looked like in early Nov. with data for Oct.  At that time it was clearly looking like we were in a recession.  I think the chart made the rounds on the blog circuit.  I even saw a post claiming it was not indicating a recession.  Here is what the chart looks like now.

The data was clearly revised lower.  That in itself is not all that unusual in the world today.  However, what I find odd is that we are still only showing Oct. data.

From the FRED web site.
2012-10: 7.34 Percent   Last 5 Observations

Monthly, Not Seasonally Adjusted, Updated: 2013-01-02 12:16 PM CST

So the data is supposedly updated this month, but we are still only showing Oct. data.  We had Oct. data when I showed the chart on Nov. 7.  What has happened since?  If you click on the last 5 observations link you get this.

2012-10: 7.34 Percent   Hide Last 5 Observations

2012-09:  4.80
2012-08:  4.46
2012-07:  0.60
2012-06:  0.44

The data series is monthly and Oct. data was there in early Nov.  Did the entire FED staff go on vacation since the election or something, LOL.  It seems pretty odd to me.  My conspiracy theory is the data has continued to get worse, but they don't want anybody to know we are likely in a recession.  I am having trouble coming up with a reason for the data being updated on 2013-01-02, but the chart still only shows Oct. when I know it already showed Oct. data in Nov.  I can think of lots of reasons why the FED might not want people to think we were already in a recession though, LOL.  It will be interesting to see what it says when it finally gets updated again.



Anonymous said...

If you click on "Vintage Series in ALFRED" on the left hand side, you get directed to http://alfred.stlouisfed.org/series?seid=RECPROUSM156N. You can then select "Download Data" near the upper left of the chart. Then, select your date range and include all revisions. Sure enough, August was revised down from 19.56 to 3.8 and up again to 4.46 (at least I think this is the progression based on the time stamps). July is when ECRI initiated its recession call I believe. What I don't understand is why there are revisions going back years in this data set. For example, why is there a revision to 1990 data being made in 2012/13? After reading the FAQs about how this data is used, there is some smoothing that is done. Regardless of the revisions and continuous smoothing, I noticed that once the data point moves above 2, the progression continues to above 20. The exceptions are 01/01/78, 04/01/79, 07/01/89, 08/01/05. And even with the downward revision, we have a reading of 7.34. Only 3 times in the past 40 years have we had a reading as high and it not progress to a higher data point above 20 (indicating a recession): 01/01/78, 04/01/79, 08/01/05 (78 and 79 were anomolies as they each had readings of less than 1.27 before or after. 05 had one reading above 2 before and nothing after) . At minimum, based on aforementioned, we are looking at a 70% liklihood of recession. And, if we look at any three month reading, we see that 100% of the time that there are three consecutive months above 4, it leads to readings above 20 and a recession signal. So no matter how you look at it, I don't see how were are not likely in a recession or heading into one... unless this data can have huge downward revisions to make it fit actual recessions, which would then make it a pointless indicator if it always subject to huge downward revisions.

Anonymous said...

I found some more information to update my previous post. There is a link to earlier releases of this data located here: http://pages.uoregon.edu/jpiger/us_recession_probs.htm
Click on the excel link all the way at the bottom for earlier releases of recession probabilities. There is much to look at, but at first glance, it is obvious that this data is subject to large revisions. For example, look at June 2010's initial reading of 7.2%. Thereafter is is revised downward to 0.2% on the most recent release. So, my last post is inaccurate. That being said, the last initial reading in the teens that was revised downward was Feb 2008. If you look at the data surrounding that report, we notice that there are revisions upward for earlier dates. So, this indicator is not a real time indicator due to the revisions. Look at Oct 2008's initial reading of 16.1 compared to the 99.2 for September. Then, look at Jan 08's initial reading of 4, which is revised up to 31.5 by November 08.

The aforementioned data plus knowing what was going on at that time leads me to believe this data series is not a real time indicator (similar to mass revisions to GDP calculations). The last recession began in December 2007. The initial reading for that month was 4.3%. That reading saw continuous monthly revisions upward until Feb 2009 with a reading of 23.4% (our recession indicator), which is too late based on where the market was at (sub 1000 SPX). Feb 2008 had an initial reading of 16.3%. Every month until Oct 2009 had readings above 10% and upward revisions. If this data series and chart is to be useful (identifying recession), we'll need to see a continual revision upward of the Aug, Sept, Oct data. The next data release will be very telling. Not to see if there is an over 20% reading, but if their is continual upward revision.

Traderbob58 said...

Very good research. Thank you for taking the time to delve into the data and to summarize what you found.

Excellent and thanks,


The information in this blog is provided for educational purposes only and is not to be construed as investment advice.