I like how they report that "a majority of the panel expects the yield to go up." And, if there is one way to get a majority to vote a certain way in your survery, that would be to only give them one possible answer! It doesn't look like the survey respondents minded, though, as they appear to be firmly in favor of rising rates.
Here is a graph of actual, observed 30 year treasury rates since 2008 (with QE episodes shaded):
How many QE's do we have to experience before the survey allows us to vote for the only outcome that we have actually observed?
Depending on where you set the start and end dates, 30 year treasuries have lost about 2 basis points a week when QE has been off and gained about 3 basis points a week when QE has been on.
Is this a case where the market gives an efficient price for something, in the face of a stated consensus opinion that is wrong?....so wrong that you aren't even allowed to give the right answer?
For the record, I expect long term rates to be fairly stable, with a slight upward bias. But, it seems pretty clear to me that more QE would push that rate up. I have argued that QE3 could have been reducing long term rates through a sort of uncertainty discount, so that its ability to increase rates has been muted until recently. But, my impression of the consensus is that the survey expectations are simply coming from a liquidity effect.
Addendum: It could be that an answer of falling rates was available, but was only left out of the report of results because nobody chose it, so my commentary should be drawn back slightly.