Figure 8b shows that the mortgage interest rate (IMORT) rose (consistent with responses to a demand shock) and the indicator of underwriting tightness (UWPC) declined (consistent with responses to a shock to mortgage supply) in response to a positive shock to mortgage balances.
On the other hand, it lowered UWPC. Again, that adds support for UWPC's serving as an indicator of underwriting tightness: as incomes rose, the likelihood of borrowers having payment problems receded; and as incomes also carried house prices upward, lenders may well have found it optimal to reduce their underwriting standards.
To the extent that underwriting standards systematically rise with mortgage interest rates, the IR in Figure 10b further supports UWPC as a useful indicator of underwriting.
More intriguingly, both UWPC and IMORT tended to fall (at least for the first two years) following an upward shock to house prices.
Finally, Figure 12 displays the estimated responses to estimated shocks to (the estimated indicator of) underwriting, UWPC. In Figure 12a, an increase in UWPC, interpreted as a tightening of standards, led both to lower GDP (GAP) and to lower GNHP.
Unlike many conventional indicators, the underwriting indicator that we constructed, UWPC, tracks the increasingly lax underwriting in the mid-2000s, followed by the extreme tightening of effective underwriting standards during the financial crisis that began in 2007.
The brevity of the sample resulted in part from the absence of data before 1996 for at least two of the five input variables that we used to construct the indicator of underwriting, UWPC. A consequence of the limited data is that sampling errors for the estimated responses loom unusually large.