Here is a regression of the Conference Board's Consumer Confidence number vs. GDP. Here I've regressed the change in both as the variables. R^2 is paltry 0.053.
Now if you want to make the regression statistics look good, you can run it with the straight levels of both, as shown below.
This comes up with a reasonable 0.618 R^2. But hopefully anyone who actually passed Econometrics can see the flaw right away. There are four data points dominating the plot. And guess what? All four points are the most recent four points! Run the regression taking these out...
And you go right back to a plot with a very weak relationship, 0.049 R^2.
Its long been my view that I'd rather watch what consumers do than what they say. I can't remember how many times I've seen surveys of consumers asking things like do you plan on spending less this Christmas? Saving more for your children's college? These questions are like asking they people in your office if they are going to stick to their new diet. They'll all say yes, but will it actually happen?
Not that there isn't plenty to worry about when it comes to consumers. I'm getting tired of warning that consumers aren't even going to spend enough to create inflation. But forget about consumer confidence! It doesn't mean anything to anybody.
Agreed that Consumer Confidence is a pretty weak indicator, but did you try lagging either one? (I could see a better R^2 with GDP being a predictor of Consumer Confidence...)
ReplyDeleteSorry to be off topic, but I just found out about "Star Wars: The Concert" and the show is coming my way in November. Looks amazing:
ReplyDeletehttp://www.starwarsinconcert.com/
Bill Gross is selling RMBS and stocking up on 5-10 year treasuries.
ReplyDeletehttp://www.bloomberg.com/apps/news?pid=20601087&sid=a5qkMIPH67tQ
I guess this means he agrees w/your prior posting on RMBS being overpriced here.