The idea for this blog post started with my curiosity about economy of scale in the banking industry. I performed a small research to study this subject. But I ended up talking about the broader topic of operating efficiency.
My research is based on data I collected from FDIC Quarterly Banking Profile, financial data of Wells Fargo (WFC), perhaps the best bank in the U.S., and data of all Texas banks that have more than 15 years of data. Banks in Texas tend to be run more conservatively than the industry so this is a good place to look at.
These Texas banks are:
Cullen/Frost Bankers (CFR): $28 billion asset
Prosperity Bancshares (PB): $22 billion asset
Texas Capital Bancshares (TCBI): $17 billion asset
International Bancshares (IBOC): $12 billion asset
First Financial (FFIN): $6 billion asset
Southside Bancshares (SBSI): $5 billion asset
The Right Efficiency Metric Is Based on Earning Assets, Not Revenue
Analysts usually use the efficiency ratio. It’s the ratio of noninterest expense to total revenue. However, I think this metric is irrelevant to a bank’s efficiency. A low-cost bank that makes very safe loans and thus less interest income can have a high efficiency ratio (analysts associate high efficiency ratio with high cost).
To compare cost, I use Noninterest Expense/Earning Assets. Earning Assets include short-term investments like money at the Federal Reserve, securities, and loans. We can also use Noninterest Expense/Deposit but Earning Assets tend to be proportional to Deposit. I use Noninterest Expense/Earning Assets because the data is more widely available. Let’s call it “Operating Cost.”
There’s one problem with Operating Cost. A bank like Wells Fargo sells a lot more financial products than other banks so it has more noninterest expenses. Using Operating Cost is unfair for Wells Fargo. So, I also look at Net Operating Cost.
Net Operating Cost = (Noninterest Expense – Noninterest Income)/Earning Assets.
The Data Tells a Different Story
I found a similar trend when I typed data of Wells Fargo and Frost. Both have lowered Operating Cost over time.
Frost’s Operating Cost from 1991 to 2014 was: 5.77% 5.82% 5.40% 5.00% 4.84% 4.78% 4.83% 4.85% 4.83% 4.94% 4.87% 4.48% 4.07% 4.13% 4.09% 4.02% 4.08% 4.10% 3.86% 3.49% 3.33% 3.02% 2.92% 2.74%.
Wells Fargo’s Operating Cost from 1991 to 2014 was: 5.31% 6.29% 6.59% 6.20% 5.70% 6.12% 6.04% 6.75% 5.70% 5.75% 5.49% 5.16% 5.40% 4.96% 4.96% 4.99% 5.12% 4.33% 4.47% 4.71% 4.48% 4.31% 3.81% 3.43%.
Over the period, Wells Fargo grew its earning assets from $37 billion to $1.4 trillion, and Frost grew its earning assets from $2.7 billion to $24 billion. The data seems to indicate that banks have lower cost when they become bigger.
But it’s not that simple. Data of other banks started making me confused. For example, Prosperity is much smaller than Wells and Frost but has much lower cost. Its Operating Cost from 1996 to 2014 was: 2.79% 2.81% 2.76% 2.70% 2.81% 2.71% 2.53% 2.30% 2.25% 2.34% 2.11% 2.53% 2.47% 2.24% 2.09% 1.97% 1.81% 1.74% 1.85%
In 1996, Prosperity had only $238 million in earning assets, or 1/15 of Frost’s $3.8 billion, but Prosperity had 1.99% lower operating cost. In 2014, Prosperity had $18 billion earning assets, or 2/3 to Frost’s $24 billion, but its operating cost advantage had declined to 0.89%.
Of the 7 banks that I collected data on, 4 banks including Wells Fargo, Frost, Prosperity, and First Financial have consistently reduced their operating cost. Three banks including Texas Capital, International Bancshares, and Southside haven’t really reduced their operating cost.
In summary, operating costs of the 7 banks are now:
Prosperity Bancshares: 1.85%
Texas Capital: 2.26%
International Bancshares: 2.64%
First Financial: 2.71%
Wells Fargo: 3.43%
And NET operating costs are:
Wells Fargo: 0.57%
International Bancshares: 0.97%
Prosperity Bancshares: 1.16%
First Financial: 1.40%
Texas Capital: 1.92%
So, the questions are:
1. Why did some banks reduce cost while others didn’t?
2. How can smaller banks have lower operating cost if economy of scale is strong?
Regarding #2, it’s true that this isn’t an oranges-to-oranges comparison. Different banks have different business models and different cost cultures. For example, Frost offers greater services than other Texas banks. It makes more Commercial and Industrial loans than other Texas banks. C&I loans have a very long sales cycle. So, Frost’s business model suggests a higher cost base. But I can’t help doubting economy of scale when looking at the data.
Big Banks Are Efficient in Selling Many Financial Products
The industry data gave me some clues. I collected operating cost of all FDIC-insured institutions categorized by asset size. There are 4 categories: banks over $10 billion in assets, banks with between $1 billion and $10 billion in assets, banks with between $100 million and $1 billion in assets, and banks below $100 million in assets. We have comparable data back to 2002.
Operating cost was pretty flat in all categories as shown in the following graph:
The graph shows that banks with more assets tend to have lower operating cost. But banks with more than $1 billion in assets have pretty similar operating cost. Specifically, the median operating cost of each category is:
Below $100 million in assets: 3.98%
$100 million to $1 billion in assets: 3.48%
$1 billion to $10 billion in assets: 3.37%
Over $10 billion in assets: 3.42%
The data suggests some economy of scale but it’s not significant above $1 billion. But net operating cost shows the advantage of big banks more clearly. Median net operating cost of each category is:
Below $100 million: 2.63%
$100 million to $1 billion: 2.23%
$1 billion to $10 billion: 1.83%
Over $10 billion: 0.98%
Big banks make much more noninterest income than smaller banks. Big banks seem more efficient in selling other financial products. But putting this advantage aside, I still doubt that economy of scale is significant on the cost side. Prosperity is a good counterexample. In 2006, it had only $3.7 billion earning assets but its net operating cost was just 1.19%. Its operating cost of 2.11% was incredibly low despite its small size.
3 Factors in Operating Efficiency
I think there are 3 key factors in having low cost
2. Unit-level economy of scale
Factor #1 was discussed above. Economy of scale on the cost side is significant when size is below $1 billion. Above $1 billion, its benefit is unclear on the cost side but obvious on the noninterest income side.
Regarding factor #2, unit-level economy of scale means efficiency gained by having high deposit per branch, high deposit per account, or high local market share, etc.
Frost is a great example of reducing costs by growing deposit per branch faster than inflation. Frost’s deposit per branch declined from $110 million in 1994 to $79 million in 1995 because it acquired some branches with low deposits. However, from 1995 to 2014, deposit per branch grew consistently from $79 million to $196 million. That’s a 4.9% annual growth over a 20-year period. And $196 million per branch is very high. Wells Fargo averages just $134 million per branch. Most banks average much less than $100 million in deposits per branch.
The clearest leverage from higher deposit per branch is occupancy cost. For Frost, occupancy cost as a percentage of earning assets has declined from 0.54% in 1995 to 0.23% in 2014. But there’s also leverage of other operating expenses. For example, JP Morgan Chase (JPM) makes about $1 million pre-tax profit per retail office. Frost made $3.7 million pre-tax profit per branch in 2014. Using a total hypothetical to illustrate the point, assume a branch manager costs $100,000 in salary, JPM would have to pay 10% of its branch’s pre-tax profit to its branch manager while Frost has to pay only 2.7%. If a branch manager costs $50,000 a year, Chase would be paying 5% of branch profits to its manager while Frost would be paying 1.4% of its profits.
I actually think high deposit per branch is more important than total size in driving down costs for banks with over $1 billion in assets. Frost offers much greater services than other banks and it usually has lower fees. But high deposit per branch allows it to have a competitive net operating cost.
First Financial matches Frost’s net operating cost of 1.40%. But it isn’t big, and it doesn’t have high deposit per branch. It has only $6 billion in assets and averages $77 million deposit per branch. However, First Financial tends to have 30-40% market share in very small markets. So, I suspect that high local market share is part of the reason why it has low cost. I think high local market share can result in efficiencies in support system.
To be fair, Frost’s account-related fees as a percentage of deposit are about 0.4% lower than First Financial’s. And Frost offers greater services than First Financial. But both have 1.40% net operating cost. So, Frost is actually more efficient than First Financial. Therefore, I think deposit per branch is a stronger force in reducing cost than local market share.
International Bancshares doesn’t really have low costs. Its net operating cost is very low at 0.97%. But its account-related fees are about 1.58% of total deposit. That’s about 1.12% higher than Frost and Prosperity Bancshares. Adjusting for this gap, International Bancshares actually has pretty high cost. That leaves Prosperity Bancshares as the most efficient bank in Texas. Actually, Prosperity Bancshares is well known in the industry for its low cost.
Prosperity averages $72 million in deposits per branch. That’s not high. Prosperity actually grows through acquisitions so growing deposit per branch is perhaps not its top strategy to reduce costs. I think Prosperity has high market share in some markets but I doubt its local market share is anywhere as high as First Financial’s. So, I think the reason for its low cost is culture.
Prosperity is really focused. It isn’t interested in selling other financial products. It just focuses on growing deposit and the loan portfolio, and maintaining a lean operation. American Banker wrote an article about Prosperity’s CFO David Hollaway, who “tracks every penny.” Prosperity’s CEO once joked that “David doesn’t pay for anything.” David Hollaway even cut things like hot chocolate, popcorn, or facial tissue for employees.
To examine a bank’s operating efficiency, I suggest a 3-question checklist:
1. Does it have high noninterest income?
2. Does it have high deposit per branch?
3. Does it have a low-cost culture?
As shown by big banks, noninterest income is a great way to reduce net operating cost. Quality of noninterest income is important. I’m not interested in account-related fees. I would love to see low account-related fees (good for customers) and a lot of income from other financial products such as trust, investment, and insurance brokerage.
Deposit per branch is an important factor in branch economics. High deposit per branch helps reduce cost at the branch level. It’s more important than total size-based economy of scale or local market share.
I use “culture” for the lack of a better word. It may mean a CEO’s action or CFO’s action. In the case of Prosperity, it actually acquired many banks and its CFO cut excess expenses during the integration process.
Culture is hard to judge. In theory, it doesn’t result in a durable low-cost advantage because competitors can try to cut costs too. In practice, it can be difficult for other banks to cut costs to extremes like David Hollaway. Many managers may never question the kind of “unnecessary” expenses that David Hollaway cut. So, culture can be a durable advantage. It’s just harder to logically prove than the advantage from things like high deposit per branch.
Finally, any qualitative study must be backed up with quantitative evidence. Fortunately, industry data is widely available for U.S. banks. So, it’s pretty easy to study a bank’s operating efficiency.