The New York Fed has just published a study on the predictive power of the yield curve in 3 months vs. 10 year Treasury bonds. (warning: pdf. For an html friendly summary with graphs, see Prof. Mark J. Perrys' claims that the NY Fed research means the recovery has already begun!).
The study updates previous research dating from the 1980's onward to the effect that a negative spread between the 3 month and 10 year Treasury yields (negative means 3 month Treasuries pay more interest than 10 year Treasuries) is means economic contraction - a recession - 1 year later. Conversely, a positive spread means economic expansion 1 year later.
Based on that, the New York Fed says that the recession is over! I disagree.
It is true that in both inflationary and deflationary times a negative interest rate spread means contraction one year later. During inflationary times, a positive yield curve means economic expansion one year later. What isn't true is that in deflationary times a positive yield curve has any predictive power at all.
The problem with the New York Fed study is that it covers data beginning only with 1960. These include only inflationary post WW2 recessions. As this graph shows, there was a positive yield curve during the Great Depression of 1929-32, and prior to the deflationary Recessions of 1938 and 1949 (3 month rates are in green, 10 year rates in red).
I examined this in detail in my series on Economic Indicators during the Roaring Twenties and Great Depression.
First quarter 2009 GDP is going to be reported in about 3 weeks. I will state with confidence that it will be negative and the New York Fed will be proven wrong.