"This is interesting beuscae my view is that if I am to win the bet, it will likely be before QE3 is announced unless the initial amount of purchases is small and/or the recession turns out to be severe. My view is based on the Sumnerian reasoning that (in addition to the fact that the Fed is always late) aggressive easing will tend to push up yields in expectation of stronger economic conditions ahead rather than causing yields to fall. Looking back at QE1, I see that the collapse in the 10-year yield from 3.82 to 2.07 occurred between 13 Nov and 18 Dec 2008. QE1 was announced on 25 Nov, by which time yields had already hit 3.14. And QE1 did not include any purchases of Treasuries before 18 March 2009. You could argue that an expectation of MBS purchases caused a portfolio effect that benefited Treasury prices in late 2008, but I think it more likely that people were just buying bonds in Nov 2008 beuscae they thought we were going into a depression. I see that Treasury yields did fall half a percent on 18 March 2009 when the Fed announced a doubling of QE1 including $300bn of Treasuries. But yields by then were already well up from their late 2008 lows. QE2 was similar yields had nearly hit their lows by the time Bernanke hinted at QE2 on 27 Aug 2010."