Gov. Josh Shapiro of Pennsylvania, Gov. Gretchen Whitmer of Michigan, Gov. Gavin Newsom of California and Gov. JB Pritzker of Illinois. A state bank would have earned their states $317 million, $318 million, $2.26 billion and $738 million in 2025, in that order. Credit…Kriston Jae Bethel for The New York Times; Sarah Rice for The New York Times; Jim Vondruska for The New York Times; Jamie Kelter Davis for The New York Times

For the last few weeks I’ve been prepping for the Public Banking Institute’s national conference in Philadelphia, where I’m delivering the opening keynote on Oct. 22. Somewhere in that prep an experiment got out of hand and turned into a working paper, now prepped for academic journal submission. This is that paper, told the way I’d explain it to any friend.

Also, a quick thank you to those of you who have contributed via Buy Me A Coffee, subscribed, and/or purchased anything from our store. The ER only had the space to do this research because of people like you.

The question I started with was simple. North Dakota has owned a bank since 1919. The Bank of North Dakota earned the state $231.8 million in 2025, a 17.2% return on investment, and S&P Global Ratings rates the bank A+. Every few years somebody in another state proposes copying it, and there have been varying degrees of progress, regression, and further progress. California passed a 2019 law letting its cities and counties charter public banks. In Philadelphia, City Council voted 15-1 in 2022 to create a public financial authority. In Washington State, Sen. Bob Hasegawa has introduced state bank bills since 2011.

So I asked myself what might be helpful to those advocating for a public bank? Here in Washington State, we are dealing with a substantial budget deficit. Many states are having the same problem and if I could show how that budget deficit can be addressed without cutting social safety nets or raising taxes, that would be a huge selling point for state reps who are genuinely interested in solving problems and serving the public. So I explored the feasibility of a formula where you plug in a handful of numbers that every state publishes anyway and get an answer, the same formula for every state, and the output would tell you how much additional revenue a state would earn if they had a bank built on the Bank of North Dakota’s model.

Here’s what I came up with. The bank would hold public funds which the state treasurer already manages, meaning the state’s own cash plus whatever local governments keep in the investment pool the treasurer runs for them. That bank would pay every depositor the same amount of interest which their sitting funds earn today, so no agency and no city takes a loss with this public bank being formed.

The bank then lends 47 cents of each dollar by purchasing shares of loans the state’s community banks make to businesses and farms, at the prime rate, and keeps the other 53 cents in Treasury bills. The state puts in $1 of capital for every $19 of cash, out of money it already holds. That’s how the bank works.

Here’s how one dollar moves through it, at 2025 rates. A depositor’s dollar earns the Treasury bill rate today, 4.17%. The bank lends 47 cents of that dollar at prime, 7.37%, and keeps the other 53 cents in Treasury bills, which pay the same 4.17% the bank owes the depositor, so those 53 cents neither make nor lose anything. On the 47 cents lent, the bank collects 7.37%, pays the depositor 4.17%, loses 0.71 cents per dollar to operating costs and bad loans at the Bank of North Dakota’s own 2025 rate, and keeps the 2.49% that is left. That comes to 1.17 cents earned on each public dollar, or $11.7 million for every $1 billion. The state’s capital, one dollar in 19, earns the bill rate too, which adds another 0.22 cents, and the total is $13.9 million for every $1 billion of public cash. Multiply by a state’s cash and you have its profit figure.

Then I had to find out whether the formula was accurate. To get there, I did four things.

First, I checked the formula with Washington State’s own numbers for every year from 2016 through 2025, with the $11.96 billion of state cash and $26.93 billion of local government cash the Washington State Treasurer held in fiscal 2026, $38.9 billion in all. Bill rates went from 0.04% in 2021 to 5.21% in 2023 across those years, and the bank made money in all 10 years, earning between $417 million and $570 million a year, because prime stayed 2.8 to 3.2 points above bills the whole time.

Second, the Bank of North Dakota has particularly low defaults and failed loans, so I used the national charge-off rate on all commercial bank loans instead, for every year back to 1985, because the national figures are more accurate for the entire nation and a more conservative and risk-averse number to use. The bank made money in 39 of those 41 years.

Third, I fed the formula the Bank of North Dakota’s own audited numbers for 2024 and 2025. Once I added back the pieces the formula leaves out, the bank’s bonds and its idle cash, the result matched what the bank actually earned.

Fourth, I ran a sensitivity test, moving each of the formula’s four inputs one at a time to see which one changes the answer the most, and the loan rate does. If the bank’s loans paid what the Bank of North Dakota’s loans actually paid in 2025, 5.28%, Washington State’s figure falls from $541 million to $160 million. The $541 million depends on one rule written into the bill: the bank buys only floating-rate loans that pay prime or better. Those loans exist in every state’s market today, because the Federal Reserve describes prime as “one of several base rates used by banks to price short-term business loans.” Every one of those tests is in the working paper with the arithmetic shown.

Here are the results. If the people of Washington State had owned this bank in 2025, that bank would have earned $541 million, and $4.8 billion over the last 10 years. Then I did the same thing for every other state, pulling each state’s public cash reserves from its own treasurer’s reports, and the source for every state is named in the paper’s appendix. Together the 49 banks would have earned $18.6 billion in 2025 and $165 billion from 2016 through 2025. Each line gives what the bank would have earned the state in 2025 and what it would have earned over the decade.

• California: $2.26 billion in 2025; $20.10 billion over the decade
• Texas: $1.78 billion; $15.89 billion
• Florida: $1.27 billion; $11.34 billion
• New York: $1.04 billion; $9.29 billion
• Georgia: $836 million; $7.44 billion
• Illinois: $738 million; $6.58 billion
• North Carolina: $631 million; $5.62 billion
• Ohio: $616 million; $5.48 billion
• Virginia: $577 million; $5.14 billion
• South Carolina: $558 million; $4.97 billion
• Washington State: $541 million; $4.82 billion
• Tennessee: $511 million; $4.55 billion
• Oregon: $475 million; $4.23 billion
• New Jersey: $449 million; $4.00 billion
• Maryland: $421 million; $3.75 billion
• Minnesota: $404 million; $3.60 billion
• Michigan: $318 million; $2.83 billion
• Pennsylvania: $317 million; $2.82 billion
• Arizona: $301 million; $2.68 billion
• Louisiana: $292 million; $2.60 billion
• Wisconsin: $288 million; $2.57 billion
• Connecticut: $253 million; $2.25 billion
• Oklahoma: $249 million; $2.22 billion
• Kentucky: $218 million; $1.94 billion
• New Mexico: $217 million; $1.94 billion
• Missouri: $208 million; $1.85 billion
• Indiana: $208 million; $1.85 billion
• Hawaii: $202 million; $1.80 billion
• Colorado: $197 million; $1.76 billion
• Alabama: $193 million; $1.72 billion
• Utah: $181 million; $1.61 billion
• Arkansas: $170 million; $1.52 billion
• Nevada: $155 million; $1.38 billion
• Massachusetts: $153 million; $1.36 billion
• Mississippi: $150 million; $1.33 billion
• Iowa: $149 million; $1.33 billion
• Idaho: $148 million; $1.32 billion
• West Virginia: $128 million; $1.14 billion
• Kansas: $121 million; $1.08 billion
• Montana: $112 million; $1.00 billion
• Nebraska: $109 million; $970 million
• Delaware: $86 million; $770 million
• Wyoming: $84 million; $750 million
• Alaska: $57 million; $500 million
• New Hampshire: $50 million; $450 million
• South Dakota: $50 million; $450 million
• Maine: $48 million; $430 million
• Rhode Island: $21 million; $190 million
• Vermont: $21 million; $180 million

A few of those figures are certainly low due to conservative and risk averse assessments from incomplete datasets. To explain further, nineteen states run no investment pool for their local governments, so for them I counted only the state’s own cash. For Utah, Arkansas, Massachusetts and Wyoming I haven’t yet separated the local share of the pool the state runs, so those four get state cash only returns calculated. The paper lists every one of those limits.

Here is what all of that taught me. Nobody doubts you when you say that a new private bank handed $38.9 billion in deposits on day one would be very profitable. A bank the public owns makes money the exact same way. Let’s do what works, people. It really is as simple as looking at the gap between what borrowers pay and what depositors earn, kept for the owner.

Almost nobody I’ve seen addresses how long these plans take to open a bank, and that timeline could well be a poison pill for public banks. As states put together public banking plans, they frequently put together five to ten year plans to get a public bank open and fully operational. For example, San Francisco’s 2019 feasibility study put a city bank’s break-even at 10 to 56 years, depending on the model; six years later and there is still no bank. North Dakota’s Legislature created its bank in the 1919 session, that bank opened July 28 of the very same year, still 1919, with $2 million of starting capital.

I imagine the slow version of this roll out is seen as the careful one, but consider something. A bank earns its money on the spread between what borrowers pay and what depositors earn, starting with the first loan it buys, so a bank that takes deposits and spends a decade getting ready pays its staff and its rent the whole time with no spread coming in.

No private bank opens that way, and no private board would approve a plan that pays costs for ten years before the revenue starts, because that plan is the one that fails. The Bank of North Dakota’s initiating legislation required the state open the bank and have it operational within a 90 day window. A bank would not approve a business loan for a bank that presented a business plan requiring a 10 year roll out to earn revenue on hundreds of millions of dollars in day one startup capital so why the hell are we acting like that is the way we should start a public bank?

I also want you to know what you’re going to hear when you ask some of your legislators for a state bank. The banks that hold the public money today pay a great deal of money for very smart people to come up with completely logical and rational reasons why it can’t be done. They say “you’ll lose in court.” “The state can’t run a bank.”

They have excuses but what do we have? We have a century long working use case, and North Dakota heard every one of those same excuses in 1919.

In 1915 North Dakota’s farmers asked the Legislature for a state-owned grain elevator, and a legislator told them to “go home and slop the hogs.” The farmers formed the Nonpartisan League that same year, and the League’s organizers canvassed the entire state nonstop making their case. In 1918 the candidates the League backed won the governor’s office and control of the Legislature.

Opponents then forced the bank onto a statewide referendum, and on June 26, 1919, voters upheld the bank by a 23,256-vote majority. Taxpayers sued the governor, and on June 1, 1920, the United States Supreme Court upheld the bank in Green v. Frazier, writing that “with the wisdom of such legislation and the soundness of the economic policy involved we are not concerned.” The bank has earned the state money for that century since.

The farmers were told no, so they replaced the politicians who said it, and the bank was open within the year.

CALL TO ACTION You can send this article to your own state senator and state representative and ask them directly: can our state do this, and will you file it? Your state’s figure is in the list above, and the one-page summary gives their office the seven findings behind it on a single page, with the working paper behind that for anyone on their staff who wants to check the arithmetic. Find your state legislators at OpenStates. Email works, and so does a call. Your own words work better than anything copied from here, because a legislative office reads a hundred identical messages as one message and a letter that sounds like a constituent as a hundred.