I have seen some angst recently over declining growth in commercial bank lending. See, for example, the Wall Street Journal:
Bank loans across all categories are increasing 4.6% annually, the slowest pace since 2014, according to weekly Fed lending data from March 1. The trend is particularly marked in business loans, which are increasing 3.9% annually, a rate that is a nearly six-year low.
A number of factors are at play, including rising interest rates; bankers also said some business clients put borrowing on hold before the U.S. election and aren't confident enough to jump back in.
The slowdown is noteworthy because it is occurring when many metrics show the U.S. economy strengthening.
Looking at the weekly data, there does on the surface look to be some reason for concern:
These low rates of growth are rarely seen outside of recessions. Still, optical econometrics suggests this is more of a lagging than leading indicator. Moreover, we have another indicator that also exhibited behavior only seen in recessions. Spot the odd man out: Recall a year ago when weak industrial production numbers raised recession concerns that proved unfounded. We could be seeing something similar in bank lending. Consider that industrial production might be a leading indicator for bank loans:
Here I focus on the post-1984 period (the Great Moderation). Optical econometrics again suggests to me that lending lags industrial production. To quantify that a bit more, I converted the data to log differences (multiplied by 100), and ran it through a 13 lag vector autoregression. Granger causality tests (the f-tests here) indicate that loans (DLOANS) do not cause (or are predictive of) industrial production (DIND):
Impulse response functions (in this case, the responses are converted to impacts on the levels of the variables) illustrate the dynamics of the system:
The impact of a shock to industrial production on commercial lending (lower left chart) is delayed six months and then builds gradually over the next 18 months. The impact of a shock to lending on industrial production (upper right chart) is negligible. Ordering of the variables does not affect these results. If I use the full sample (data begin 1947:1), both variables Granger cause each other, but the impact of loans on industrial production in the short-run is minimal and dies out in the long-run:
Bottom Line: The fall in commercial lending growth looks more consistent with a lagged impact from the industrial slowdown that weighed on the US economy last year than with a warning about future activity. Something to keep an eye on, to be sure, but if past history is a guide, it is more likely than not that lending will pick up over the next year.
