A Bank Policy Institute analysis published May 8, 2026, projects that yield-bearing stablecoins could destroy up to $4 trillion in US bank deposits by 2030 and cut bank lending by roughly $2.7 trillion in the same period.

The note, written by BPI senior fellow Bill Nelson, takes aim at a February 12, 2026 Reuters commentary by Stephen Gandel. "The first problem with the banks' argument is that it's not possible to pull deposits from the system on aggregate," Gandel had written.

Nelson disputes that framing. He cites a 2025 paper that models deposit displacement under explicit assumptions and concludes the destruction is dollar for dollar.

How does Cong's model work?

The note draws on "Stablecoins and Banking: Deposit Dynamics, Financial Stability, and Regulatory Design" (2025) by Lin William Cong, whose research received financial support from Coinbase, Paradigm, PayPal and Stripe.

"Stablecoin adoption is believed to expand substantially, reaching roughly four trillion dollars by 2030," Cong writes. Nelson's analysis maps that projection onto a 19 percent net deposit decline and an equivalent drop in lending.

Cong built his framework on a 2022 paper by Jonathan Chiu, Seyed Mohammadreza Davoodalhosseini, Janet Jiang and Yu Zhu, originally designed for interest-bearing central bank digital currencies.

The model treats deposits and interest-bearing stablecoins as perfect substitutes. Households calibrate payment balances to yield, so every dollar into stablecoins removes a dollar of deposits at the new equilibrium.

The threat alone, with zero tokens circulating, would push deposits up by $300 billion — about 1.5 percent of the current $19 trillion US deposit base — as banks compete harder for funding. Once tokens circulate, the same model reverses that gain.

Nelson arrives at $3.7 trillion in net deposit destruction after subtracting that initial competition-driven lift.

How does Fed research compare?

Nelson concedes Cong's outcomes are "implausibly large." He contrasts them with "Banks in the Age of Stablecoins," published December 17, 2025 by Federal Reserve Board economist Jessie Jiaxu Wang.

Wang estimates deposit and lending destruction in a range from $65 billion to $1.26 trillion. The high end materializes if stablecoin issuers gain access to Federal Reserve master accounts.

The spread reflects different assumptions about substitutability and central-bank access. Nelson flags Cong's premise that over 9,000 US depositories act like a fragmented market with local monopoly power as unrealistic.

Why does transaction-tracing mislead?

Nelson argues the "deposits just move around" conclusion comes from tracing one chain of payments and stopping there. That step ignores portfolio rebalancing and the destruction of deposits when loans are repaid.

Deposits are created when banks extend loans and destroyed when borrowers repay, the analysis explains. With yield-bearing tokens meeting transaction needs, households settle at a smaller banking system, lower T-bill yields, and higher deposit and loan rates.

What does the Genius Act add?

A separate ECB paper, "Toss a stablecoin to your banker," cited by Nelson, adds a second channel. If issuers park reserves at banks under the Genius Act, liquidity rules will treat those deposits — likely far above the $250,000 FDIC cap — as unstable.

Banks must back them with an equal stock of liquid assets. Those reserves drop out of the loanable pool, an outcome that converges with Cong's on the lending side.

What comes next?

Nelson's note lands as Congress weighs stablecoin legislation and federal agencies draft companion rules. The framing matters for two pending decisions: whether issuers can earn interest on reserves and whether they gain Fed master accounts.

The eventual rulebook will determine whether the realized impact tracks Cong's $4 trillion ceiling or Wang's $65 billion floor.