|
We analyze the impact of universal banking on the real economy, by comparing the performance of a benchmark linear VAR model with a nonlinear projection process (Jordà, 2005) which tracks shocks asymmetries. We divide bank shocks into two categories—i.e., credit (loans) and fee-based shocks—and show that, based on U.S. data, fee-based shocks seem to have a significant feedback effect on real GDP and on the stock market. Bank fee-based activities are more sensitive to external shocks than traditional business lines (loans) and, more importantly, they are responsible for most of the feedback effects from the banking sector to the real economy and stock markets—especially during crises. Our results also indicate that in normal times, even if nonlinearities are less at play, the feedback effects remain significant, particularly since the 1997 regulatory changes. |
