A lot attention has been paid to the yield curve recently, specially the 10 years minus 2 years US treasure spread (spread). In fact many papers have been written on the subject and the important prediction function that the spread has on US economic cycles and recession. However since the introduction of the ZIRP the nature of the spread seems to have changed. Looking through the data since 1978 until ZIRP its possible to notice that the most important driver of the spread has been the short term interest rate, more specifically the Fed funds, and not the long tail. But after ZIRP things changed and because the short term is anchored by the Fed funds the spread has been driven by the long tail. So, looking for the spread nowadays could be misleading in terms of recession prediction. Sure a recession is always possible but this time it wont be caused by the usual driver of the spread, what reduces its importance in my view.
Fed Funds, curve flattening and recessions:
the flattening of the curve and FED funds hiking cycle is very similar. But to measure better the effect of Fed Funds and the 10 Years UST on the spread I ran a couple o simple regressions below.
First the impact of FED Funds on the spread since 1978:
and from 1978 until 2007 just before the crisis...
things are pretty much different for the UST 10Y:
However, since 2010 things changed and UST is the main driver of the spread now, not surprisingly once the 2Y has been anchored by the Fed funds.
What can be seen here also...
If the the spread is important because most of it is driven by the short term rates it must be important to question its meaning now that the driver has changed. NBER has a paper that through a much deeper analyzes also finds more prediction power on the short term rates instead (http://www.nber.org/papers/w10672.pdf).