Changes in Bank Credit Supply and CI Loan Growth
Table 1 presents the results of regressing CI loan growth on the tightening index and the reported change in CI loan demand for the period from 1990 to 2019. The estimates in column 1, which omits the quarter fixed effects, imply that a one standard deviation increase in the tightening index is associated with a 3.13 percentage point drop in CI loan growth in the following year.
Columns 2 and 3 add macroeconomic controls and quarter fixed effects to the specification, respectively. 7 The predicted effects of tightening ents are accounted for, indicating that banks tend to tighten during times of stress, when balances would fall even without a change 
in supply conditions. However, the coefficient on the tightening index is still economically and statistically significant. 2 percent relative to other banks over the same time.
In column 4, we add the reported change in standards to our specification. This approach allows us to assess whether the tightening index-which accounts for reported changes in loan terms and standards-adds information beyond what is reflected in the standards measure. The coefficient on standards is negative, indicating that tighter standards are associated with weaker growth even when controlling for the broader index, but it is not statistically significant. 8 The coefficient on the tightening index edges down when standards are included but remains statistically significant. These results indicate that the tightening index, by accounting for changes to both terms and standards, improves the measurement of loan supply compared to using standards alone.
Lastly, column 5 shows that the effects on loan growth are asymmetric. The specification includes a term that interacts the tightening index with an indicator for whether the index is greater than 0, denoted Tightening Index + . A one standard deviation tightening is predicted to reduce CI loan growth by 2.2 percent relative to a bank at the mean, whereas a one standard deviation easing is only predicted to increase growth by 0.3 percent. Thus, tightening tends to have a much stronger effect on loan growth than easing.
In all these regressions, reported increases in demand are associated with stronger CI loan growth. In column one, we find that when banks report that demand strengthened, CI loan growth is on average 3.3 percentage points higher in the following year. This relationship drops to around two percentage points when we include macroeconomic variables and quarter fixed effects (columns 2 and 3) but remains statistically significant.
Table 1: Effect of Tightening on CI Loan Growth
How do these effects vary over time? The previous results assessed the effects of credit conditions on growth over the subsequent year. We now estimate a set of regressions to trace out the effects of changes in supply and demand over different time horizons:
Figure 1 plots the dynamics for CI loan balances around a reported tightening in supply (left) and increase in demand (right). When the tightening index increases by one standard deviation, CI loan balances edge down in the following quarter. Balances then progressively decline over the next year and are down about 2 percent 6 quarters after tightening. Meanwhile, when banks report an increase in demand, there is an immediate one percentage point jump in CI loan balances in the following quarter. This is likely a result of most CI commitments being lines of credit rather than term loans. An increase in demand will be associated with more draws immediately, causing balances to rise. However, tightening standards will mean fewer commitments being made at the time of tightening, followed by a slow decline in balances as other loans mature or payoff.