Though most of residential construction is in the dumps and unresponsive to changes in the interest rates this is not at all the case for home remodeling. Its doing exactly what the models predict  hitting record highs amid record low real interest rates. From Calculated Risk.

Right now the explanation that I am playing with is that what we think of Keynesian Effects work principally through the interaction of bonds and land. The interaction between rent and interest as it were.

Structures in general and residential structures in particular are extremely sensitive to these effects because they are created by the intersection of land and bonds markets.

Remodeling can be funded in some cases out of cash and other cases out of lines of credit that operate more like credit cards than mortgages. The coupon is not fixed and that may be important.

Indeed, the predominance of fixed coupons is something that should interest us deeply. We are explicitly creating a sticky price. Why is this optimal?