Justiniano, et al (2019): Credit Supply and the Housing Boom
JPT emphasize a loosening of lending constraints, rather than borrowing (i.e. collateral/LTV) constraints, as driving the housing boom. In particular, they argue that you need a loosening of lending constraints to match the observed decline in interest rates over the boom years. Looser borrowing constraints alone actually tend to push up interest rates because they increase the demand for credit. For example, they note that in Favilukis, et al. (2017), an infusion of foreign capital is required to reverse the rise in interests rate that accompanies the looser borrowing constraints.
Intuition
They consider the supply of mortgage credit in the US to have been constrained prior to the boom. That is, lenders were willing to lend more, but could not do so, for some reason. As this lending constraint loosens during the boom, the supply of mortgage credit increases and the interest rate falls. A lower interest rate makes borrowing more desirable, which in turn increases the value of collateral (housing) that makes borrowing possible. Thus, house prices rise purely because consumers want to bring consumption forward. To do so, they need to borrow more. But the amount they can borrow is constrained by the collateral value of their house. Accordingly, the value of housing as collateral for loans increases. Obviously, I’m very uncomfortable with this mechanism.
In contrast, consider a model with only borrowing constraints and no supply constraints. As borrowing constraints loosen, the demand for credit rises and interest rates tend to rise. The rise in interest rates dampens the desire to borrow and the collateral value of housing does not increase as much as the case above. This logic suggests that lower interest rates combined with looser collateral constraints would boost house prices a lot. I guess their innovation is that their model endogenously generates lower interest rates via exogenously loosening credit supply constraints.
Model
Their model is probably too stylized to be applied quantitatively. It features impatient (borrower) and patient (lender) households. Borrower households face a collateral constraint tied to the value of their house. And lenders face an exogenous upper bound on how much they can lend; this is the lending constraint (which is their main innovation). Mortgages are one-period bonds. The interest rate is determined endogenously by bond market clearing. They assume linear utility in consumption, so there is no risk premium for housing as in Favilukis et al.
