Those markets have more children and more suburban family households, and a branch count will not tell you which are underserved. Brief 1 showed that growth since 2021 has gone to places with more room, where new households keep arriving. This brief asks who those households are, and how to size a market by its people instead of its branches. Everything here builds on the same ten years of FDIC branch data and the latest demographic updates. Near the end, a lookup tells you whether your own county is overbranched or underbanked.
We rank the 200 large metros by the share of national deposits they gained or lost from 2021 to 2026. Large means 100,000 or more households. Deposits here are counted inside the $500M office cap. No office counts for more than that, so a bank that books a huge sum at one office does not distort a market.1 Throughout, we compare the 40 metros at the top of that ranking with the 40 at the bottom.
The top 40 grew households 8.1% from 2021 to 2026, against 3.4% for the bottom 40.2 What else separates them is smaller than you might expect:
Median metro in each group. The children gap holds after controlling for density and household growth. The homeowner gap does not, so it travels with the same markets rather than adding something new.3 Average household income is lower in the top 40 ($117K against $137K). Brief 1 covers income, so we stop there. If the story were wealthy professionals, income would run the other way. This is association across metros, not proof that children or ownership attract deposits.
Tapestry is Esri's commercial household segmentation. It sorts every household into one of 61 segments, such as Flourishing Families or Urban Chic, and groups those segments into 12 LifeMode groups.4 Here is the gap in household mix between the top 40 and the bottom 40, in percentage points of households:
LifeMode group share of households, top 40 minus bottom 40, percentage points
Among the 61 segments, the largest gains in the top 40 are Flourishing Families, Boomburbs, Up and Coming Families, Southern Satellites, Classic Comfort. The largest shortfalls are Neighborhood Spirit, Top Tier, Uptown Lights, Urban Chic, Trendsetters. Segment names are Esri's labels, not ours. The pattern is family and suburban segments up and dense urban segments down, which fits what Brief 1 found on density. Because group mix travels with density and income, we read this as a description of who lives in the markets gaining share, not as a cause of deposit growth.
From 2021 to 2026 the country added 5.4% more households and 3.2% more people, so the average household shrank from 2.61 to 2.56 people.5 In 627 of 3,144 counties, 19.9% of them, households grew while population fell. They hold 27.9 million households, 20.7% of the country.
The shrinking household is not what separates the two groups. Average household size fell by 0.049 people in the top 40 and 0.043 in the bottom 40, a gap that is not statistically significant. It is everywhere, which is why households, not residents, are the unit that tells you how many accounts a market can hold.
Projections put total household income growth at $2.80 trillion from 2026 to 2031.6 Only $0.59 trillion of that, 21.1%, comes from new households arriving. The other $2.21 trillion, 78.9%, comes from higher incomes in households that exist today, whose average income is projected to rise from $122,692 to $138,579. Dollars are not adjusted for inflation, and these are projections.
Where projected household income growth comes from, 2026 to 2031
That does not weaken Brief 1. It adds a second household to serve: the one already in the market, whose income is growing. A plan that serves only arrivals leaves most of the projected growth to someone else.
We call a market underbanked when it has 1,400 or more households per branch, and overbranched when it has fewer. Branches here are bank offices plus credit union branches.7 Credit unions are 22.8% of all branches.
Most counties have fewer households per branch than the 1,400 line. Overbranched counties number 2,298, 73.1% of the total, and together they hold 42.5% of households. The 846 underbanked counties hold the other 57.5%. Among the 200 large metros, 103 are underbanked and 97 are overbranched.
Underbanked metros, by how much deposit share they gained from 2021 to 2026
Gaining deposit share does not mean thin coverage. Most of the metros that lost the most share are underbanked. Most of the top 40 are not underbanked. The 17 underbanked markets in that group are where to look first, because they pair household growth with thin coverage:
| Metro | Households | Branches | Households per branch |
|---|---|---|---|
| Riverside-San Bernardino-Ontario | 1,519,543 | 587 | 2,589 |
| San Antonio-New Braunfels | 1,057,761 | 538 | 1,966 |
| Atlanta-Sandy Springs-Roswell | 2,471,340 | 1,300 | 1,901 |
| Orlando-Kissimmee-Sanford | 1,125,867 | 604 | 1,864 |
| Tampa-St. Petersburg-Clearwater | 1,441,344 | 803 | 1,795 |
| Charlotte-Concord-Gastonia | 1,184,814 | 661 | 1,792 |
| Jacksonville | 721,664 | 406 | 1,777 |
| Austin-Round Rock-San Marcos | 1,071,655 | 609 | 1,760 |
The eight thinnest of the 17, both charters. Underbanked line: 1,400.
Counting credit unions changes the picture in 73 metros, which sit at 1,400 or more on bank offices alone and fall below the line once both charters count. That is why a branch count from one charter misleads, and why households are the better yardstick.
The CDFI Fund is the Treasury program that certifies community development financial institutions (CDFIs) and runs the New Markets Tax Credit, which steers private investment to low-income communities. Since 2001 it has placed $71.0 billion across 5,074 census tracts.8 Eligibility runs on income and poverty, not on whether anyone can reach a bank. A banking desert is a tract with no bank or credit union branch within a distance the Federal Reserve sets by place type.
Share of tracts that received New Markets Tax Credit investment, by 2025 branch access
A banking desert is less than half as likely to have been reached, and it received 46 cents on the dollar per household ($276 against $600). The gap is widest in the suburbs, which hold 66.3% of deserts:
Share of banking deserts that received investment, by place type
Suburban deserts are usually not the poorest tracts, so they may not meet the income tests that steer the program. We did not test eligibility tract by tract. The largest group of deserts is the least reached, and that is where some of the households you are sizing may sit.
Type a county name to see its households, its bank and credit union branches, and how it compares with the 1,400 households per branch line.9 Fewer than 1,400 is overbranched. 1,400 or more is underbanked.
Look up any county at cucollaborate.com/where-the-money-moved-who-the-households-are.
A new branch is one answer, and it is the most expensive one to reverse. Shared branching, a smaller format, mobile service, and digital enrollment are others. We hold no cost data for the alternatives, so Brief 3 does not guess. It shows what a new branch has needed to pay and gives you a break-even tool that runs on your own costs. Brief 3 follows next week.
Name two or three. In 30 minutes we will size each one in households, show its household mix, and tell you whether it is underbanked once both charters count.
Book a meeting