Whilst trawling around the Internet looking for data for some research that I am presently doing, I came across this research from Beverly Hirtle at the New York Federal Reserve. The paper attempts to correlate the relationship between network size and branch performance, and uses much more recent data than other papers I've seen in this area.
The results are interesting. Let me quote directly:
Our results suggest that institutions with mid-sized branch networks have lower deposits per branch than institutions with both larger and smaller branch networks. They also hold lower levels of small business loans per branch than smaller organisations.
In other words, the more branches a bank has, the less performing each branch is. At least, it seems that's the case until you get large enough that your multi-market network externalities start to kick in. In the Hirtle typology, large enough is a bank with 501 or more branches. Mid sized is anywhere from 101 to 500. It appears there is some kind of "hump" that occurs as institutions begin to ramp up their service by adding more points of presence.
A while back I posted on what I call the Service Arms Race, where I hypothesised that there might be a point at which increasing investment in service propositions in retail banking might yield decreasing returns. I suggested that in a game of this kind, only those with the deepest pockets could win. Deepest pockets? Try a bank with 501 or more branches.
In the mean time, the mid-sized branch networks are not getting deposits from customers at a lower cost than the larger banks:
... after controlling for asset size, mid-sized branch networks have no deposit expense advantage relative to institutions with larger branch networks (though lower costs than institutions with smaller networks)
Considering that a very large bank will have wholesale sources of funding that are unavailable to other institutions, they probably won't participate as much in the high rate drive to get deposits. However, this quote suggests to me that the mid-sized banks are forced to compete for deposits with the smaller banks, who will typically offer higher rates.
So as you grow, your cost to get money either increases or stays the same, but your branch performance goes down. Now that's a hump.
Finishing up the research, however, is the startling conclusion that no significant relationship was found between branch network size and profitability. This, it seems to me, is more evidence supporting my Arms Race hypothesis. Surely an incremental investment in service and convenience ought to result in better profitability for an institution? Since we know that scale economies reduce the cost of a branch as the overall network grows, why doesn't a bigger branch network result in better performance?
There are two explanations that spring to my mind: firstly, that customers get served less efficiently as the network grows. This follows, since it is likely that a customer won't have the same kind of personal interaction they'd get if they just used their local branch regularly. The second explanation, though, is more challenging: what if customers actually don't want a relationship or more service? A while back, I posted on research that would tend to suggest this.
In any event, the whole thing is largely accademic in the short term. Banks are opening more branches, and banking strategists do say that the competetive battleground is service.
The question in my mind is for how much longer this will be the case.
Recent Comments