Author
Bankers and the communities they serve worry that changes to supervision and regulation (S&R) post-financial crisis will lead to fewer community banks. They fear that higher intensity S&R equals a faster rate of consolidation. I worry that too few community banks could mean less robust local communities in the Ninth District if those mergers lead to higher costs or fewer available banking services.
Have we seen the pickup in consolidation resulting from more intense S&R? This question cannot be answered directly. We do not witness the world, including the level of consolidation, absent more intense S&R.
But the threat of increased consolidation is real. We should therefore try to identify it even in the face of significant challenges. We try to make that identification each quarter at the Minneapolis Fed. That is, we try to figure out if the rate of consolidation exceeds what we would expect if historical patterns of consolidation continue. This process allows us to “raise the red flag” if consolidation is accelerating beyond what history would suggest (recognizing that we would ideally identify and stop all unnecessary consolidation before it happens).
So far, consolidation in the states in the Ninth District generally falls within historical predictions. Note that this outcome does not mean that the government has calibrated S&R correctly. We should still try to fix cases where S&R is too intensive.
Instead, these results mean that an already rapid rate of consolidation in the past continues. I will provide a few more details about our consolidation monitoring effort. I encourage you to visit our web page devoted to “Tracking Community Bank Consolidation.”
Our Consolidation Monitoring Effort
Each quarter, we forecast the number of community banks that would remain in a state, the Ninth District, and the United States a year from the forecast. We use multiple models to develop the forecasts to make our process more robust. Three of the models are similar in that they project from an overall pattern of consolidation over time into the future. Think of these models as “top down” or “aggregate” forecasts. We use another model that estimates the future of each individual bank, which we then “add up” to get a total number of banks. Think of this model as a “bottom up” forecast.
We monitor consolidation by comparing our forecasts with what actually happens. A more rapid pace of consolidation, potentially due to more intense S&R, would show up as fewer banks than we forecasted. Our last forecast used data through the end of 2014 and made a forecast for the end of 2015. We also compared what happened in 2014 to our forecast made at the end of 2013.
The number of banks that existed at the end of 2014 matched the forecast pretty well for most of the models. The models underestimated consolidation in some cases, which is worth highlighting given our focus. However, the amount of undershooting seems relatively small given the inherent uncertainty in any forecast.
A few examples of this performance make the point. For the district as a whole, there were 603 banks at the end of 2014, down from 632 at the beginning of the year. Our models produced forecasts of the number of banks between 604 and 626. In some states, the forecasts lined up with what occurred. South Dakota had 67 banks at the end of 2014, down from 70. Our forecasts ranged from 69 to 67.
The goal of the monitoring is not to produce the most accurate forecast. Instead, we want to create a system for potentially identifying consolidation that seems excessive or too rapid relative to what history would suggest. Experience outisde of historical norms might suggest that government action, in the form of S&R, encourages too much consolidation. We should raise concern if that outcome occurs.
