When I eyeball the graph, 2002 to 2011 really jumps out. For about a third of the graph the two move in lock step.

It's a weird sort of inverted relationship when the variables spend so much time moving in the same direction.

Here's another take:

“If you look at the relationship between [mortgage] rates and [home] sales and home prices, the relationship is almost zero,” said Sam Khater, deputy chief economist at CoreLogic.

https://www.curbed.com/2018/2/20/17029746/interest-rates-ris...

Maybe all these people are biased, or lying about simple numbers, or it's some big conspiracy.

The alternative explanation is just that the economy really has more than two variables.

That second hypothesis doesn't seem so preposterous to me.