Folks might remember my obsession with this analysis. I made a blithe remark this evening that led to a rather lively discussion. In short:

I: “Anytime there is more than one y/y down month, we see a downturn in the market soon after.”

Other: “Bullshit.”

And that was the end of the discussion.

So, of course I had to prove my point. I went back to the data. Turns out I was wrong.

This happens frequently so I am really not that bent out of shape over it. When I originally looked at the y/y info, I was struck by how infrequently the negative datapoints showed up. This time around, as I tried to prove my [incorrect] thesis, I was struck by the fact there was not a single negative y/y datapoint from Jan87 – Jan90.

For obvious reasons, I am abandoning this pursuit from here on. But, I do provide below a list of the months where there is more than one negative y/y usage datapoint in a row. Maybe someone can salvage my orginal idea.

6/79 – 10/79
12/79 – 1/80
4/80 – 5/80
2/81 – 3/81
8/81 – 9/81
5/82 – 5/83
8/84 – 9/84
1/86 – 2/86
11/90 – 12/90
2/91 – 4/91
12/91 – 1/92
5/92 – 8/92
7/94 – 8/94
11/94 – 1/95
7/96 – 8/96
2/97 – 3/97
5/97 – 6/97
10/98 – 11/98
9/99 – 10/99
2/01 – 6/01
11/01 – 2/02
6/03 – 7/03
10/03 – 11/03
1/05 – 2/05
4/05 – 5/05
5/07 – 7/07
8/08 – 11/09

Tweet