« How to cope with an autistic economist | Main | Why is CPP funded, but OAS is Pay As You Go? »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Thanks for a very interesting post. I recently asked a former student, who works on Bay St, why we hear so little discussion about the LFS hours worked number. I was told that it's not taken very seriously because of measurement error. The US releases its version of the LFS and SEPH surveys at the same time, and the establishment survey is viewed as a more reliable measure of both hours worked and employment. But your correlation matrix with GDP makes it look like the LFS numbers in Canada, especially for hours worked, are pretty informative (even if they are measured with error). At least for monthly GDP.

Which reminds me, we never got to do the projections for the year thing.

Angelo said what I was going to say. Must remember to pay more attention to hours worked.

Interesting - everyone was revising up expectations after Friday's strong trade report. I had 1.5% revised to 1.8% after Friday.

Yeah, I really should expand this project to incorporate other numbers.

Have you considered principal components? First two months of GDP plus a first principal component of select variables?

I've never been able to make sense of principal components models. I'd probably go with a (dynamic) factor model.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad