Fixing Bank Branch Counts That Include Drive-Throughs and Offices Without Tellers

On this page

Problem statement

The Summary of Deposits calls every place a bank does business an office, and a branch count that takes all of them includes drive-up windows, trust offices, administrative buildings, mobile branches and "home banking" offices that are websites. The column that separates them, BRSERTYP, holds a numeric code with no label in the data. This page decodes it from the FDIC's own definitions, counts each type with its deposits, checks the limited offices by name, and shows how much the choice changes a count, between states and between years.

Quick answer

Label the codes and count offices and median deposits by type:

TYPES = {11: "full service, brick and mortar", 12: "full service, retail", 13: "full service, home banking", 21: "limited, administrative",
         22: "limited, military facility", 23: "limited, drive-through or detached", 29: "limited, mobile or seasonal", 30: "limited, trust"}
t = sod.groupby("BRSERTYP").agg(offices=("UNINUMBR", "size"), median_deposits_m=("deposits", lambda s: s.median() / 1e6))
t.index = t.index.map(lambda k: f"{k} {TYPES.get(k, '?')}")
print(t.round(1).to_string())
                                       offices  median_deposits_m
BRSERTYP                                                         
11 full service, brick and mortar        71242               84.5
12 full service, retail                   2388               14.6
13 full service, home banking              182                7.0
21 limited, administrative                 287                0.0
22 limited, military facility                2               15.8
23 limited, drive-through or detached     1449                0.0
29 limited, mobile or seasonal             357                0.0
30 limited, trust                          190                0.0

The codes come in two families: 11 to 13 are full-service offices, 21 to 30 limited-service ones. Full-service brick-and-mortar branches are 71,242 of the 76,097 offices, with a median of $84.5 million in deposits, and retail offices in shops 2,388 with $14.6 million. Every limited type except military facilities has a median of zero: drive-throughs, administrative, mobile and trust offices take no deposits of their own, because the money is booked at a parent branch.

Horizontal bars on a log scale of offices by service type with their median deposits: 71,242 brick-and-mortar branches with a median of 84.5 million dollars and limited offices with medians of zero. Horizontal bars on a log scale of offices by service type with their median deposits: 71,242 brick-and-mortar branches with a median of 84.5 million dollars and limited offices with medians of zero.
Office types in the Summary of Deposits, 2025.

What the limited offices are, and what the choice changes

1. What the limited offices are

for k in [23, 21, 13]:
    print(k, sod[sod.BRSERTYP == k].NAMEBR.str.lower().str.extract(r"(drive|motor|atm|operations|online|internet|digital|trust)")[0].value_counts().head(3).to_dict())
23 {'drive': 486, 'motor': 102, 'operations': 6}
21 {'operations': 23, 'trust': 10, 'drive': 2}
13 {'internet': 14, 'online': 11, 'digital': 4}

The names confirm the codes. Of the drive-through and detached offices, 486 have "drive" in their name and 102 "motor", as in motor bank; administrative offices include operations centres and trust offices; home-banking offices are called internet, online or digital branches. These are not places a customer can open an account at a counter.

2. How much the count changes

full = sod.BRSERTYP.isin([11, 12])
print(f"all offices: {len(sod):,}; full-service walk-in offices: {full.sum():,} ({full.mean():.1%})")
by = sod.groupby("STALPBR").apply(lambda d: 1 - d.BRSERTYP.isin([11, 12]).mean())
print(by.sort_values(ascending=False).head(5).map("{:.1%}".format).to_string())
all offices: 76,097; full-service walk-in offices: 73,630 (96.8%)
STALPBR
MP    16.7%
DE    13.0%
NE     8.4%
OK     7.2%
GU     6.2%

Counting full-service walk-in offices, codes 11 and 12, keeps 73,630 of the 76,097 offices, 96.8%. The share of limited offices differs between places: a sixth of offices in the Northern Mariana Islands, 13.0% in Delaware, mostly mobile, administrative and trust offices, and 8.4% in Nebraska and 7.2% in Oklahoma, mostly drive-throughs. A state comparison of all offices overstates those places.

Horizontal bars of the share of offices that are limited-service or home banking: Northern Mariana Islands 16.7%, Delaware 13.0%, Nebraska 8.4%, Oklahoma 7.2%. Horizontal bars of the share of offices that are limited-service or home banking: Northern Mariana Islands 16.7%, Delaware 13.0%, Nebraska 8.4%, Oklahoma 7.2%.
Where limited-service offices are most common.

3. Online 'branches' that hold deposits

o = sod[sod.BRSERTYP == 13]
print(f"home-banking offices: {len(o)}, deposits ${o.deposits.sum() / 1e9:,.0f} billion")
print(o.nlargest(3, "deposits").assign(bn=lambda d: (d.deposits / 1e9).round(0))[["NAMEFULL", "CITYBR", "STALPBR", "bn"]].to_string(index=False))
home-banking offices: 182, deposits $906 billion
                         NAMEFULL     CITYBR STALPBR    bn
Capital One, National Association Wilmington      DE 274.0
         Charles Schwab Bank, SSB   Westlake      TX 210.0
                        Ally Bank      Sandy      UT 151.0

The 182 home-banking offices hold $906 billion of deposits, as much as all of Florida's branches. Capital One's internet branch in Wilmington books $274 billion, Charles Schwab's bank $210 billion and Ally Bank, in Sandy, Utah, $151 billion. They count as full-service offices in the codes, but they have no counter and no neighbourhood; leave code 13 out of branch maps and keep it in deposit totals.

A count that changed with the definition

The data used on this page

The FDIC's Summary of Deposits for 2015 and 2025, every office of every FDIC-insured institution with its deposits as of 30 June, from the BankFind Suite API, downloaded on 28 October 2026 and saved on this site, because banks amend past reports and the API serves the revised figures. Public domain. Deposits are reported in thousands of dollars.

from pathlib import Path
import zipfile

import geopandas as gpd
import numpy as np
import pandas as pd
import requests

Path("data").mkdir(exist_ok=True)
# FDIC Summary of Deposits for 2015 and 2025 (every branch, deposits as of 30 June) from the BankFind Suite API, with
# the Census Bureau's 2024 county population estimates, 2020 tract centres of population, counties and states.
# Saved 28 October 2026: the API serves revised years as banks amend their reports.
# Live: https://api.fdic.gov/banks/sod?filters=YEAR:2025&limit=10000&offset=0
ARCHIVE = "fdic-sod-2026-10-28.zip"
if not Path(f"data/{ARCHIVE}").exists():
    Path(f"data/{ARCHIVE}").write_bytes(requests.get(f"https://data.spatialworkflow.io/{ARCHIVE}", timeout=600).content)
Z = zipfile.ZipFile(f"data/{ARCHIVE}")
for f in ["cb_2024_us_county_20m.zip", "cb_2024_us_state_20m.zip"]:
    Path(f"data/{f}").write_bytes(Z.read(f))
def read_sod(year):
    d = pd.read_csv(Z.open(f"fdic-sod-{year}.csv"), dtype={"ZIPBR": str})
    d["STCNTYBR"] = d.STCNTYBR.astype(str).str.zfill(5)          # county FIPS, stored as a number
    d["deposits"] = d.DEPSUMBR * 1000                              # reported in thousands of dollars
    return d
sod = read_sod(2025)
print(f"FDIC Summary of Deposits 2025: {len(sod):,} offices, ${sod.deposits.sum() / 1e12:,.2f} trillion in deposits")
FDIC Summary of Deposits 2025: 76,097 offices, $18.10 trillion in deposits

Example 1: a count that changed with the definition

old = read_sod(2015)
for y, d in [(2015, old), (2025, sod)]:
    print(y, f"all {len(d):,}; full service {d.BRSERTYP.isin([11, 12]).sum():,}; drive-through {(d.BRSERTYP == 23).sum():,}")
2015 all 93,272; full service 89,566; drive-through 2,448
2025 all 76,097; full service 73,630; drive-through 1,449

Between 2015 and 2025 all offices fell from 93,272 to 76,097 and full-service offices from 89,566 to 73,630, a similar fall, while drive-throughs fell faster, from 2,448 to 1,449. A trend in "branches" is robust to the definition here, but a level is not: always say which codes you counted.

Paired bars of 2015 and 2025 office counts: all offices 93.27 and 76.10 thousand, full service 89.57 and 73.63, drive-through 2.45 and 1.45. Paired bars of 2015 and 2025 office counts: all offices 93.27 and 76.10 thousand, full service 89.57 and 73.63, drive-through 2.45 and 1.45.
Office counts by definition, 2015 and 2025.