What if the same technology that promises faster payments and smarter money movement also made it more expensive for ordinary people and businesses to borrow? That question has been sitting with me since I first read the latest set of back-of-the-envelope calculations from a pair of economists. Their numbers are not predictions of bank failures or mass deposit flight. They are estimates of how much long-term interest-rate exposure the banking system might be willing to carry once deposits become far more mobile.
I’ve spent enough time watching balance-sheet trends to know that small shifts in deposit behavior can ripple outward in surprising ways. A 10 percent change in rate sensitivity or average deposit life does not sound dramatic until you translate it into hundreds of billions of dollars of reduced capacity. Suddenly the conversation moves from “cool new feature” to “who ends up paying more for a mortgage or a business line of credit.”
Why Tokenized Deposits Matter More Than Most People Realize
Tokenized deposits are simply regular commercial bank deposits represented on a blockchain or similar ledger. They stay liabilities of the issuing bank. Nothing about the underlying claim changes. What does change is the speed and the automation. Instant settlement, programmable instructions, and the possibility that software agents could move balances the moment a better yield appears elsewhere.
In my view, that last point is the real pivot. Most of us still think of deposits as fairly sticky. People leave money sitting in checking accounts for years even when rates elsewhere look better. The friction of paperwork, login portals, and mental bandwidth keeps balances in place. Tokenization and smart contracts can strip a lot of that friction away. Whether large numbers of depositors actually use the new tools remains an open question. The research simply asks what happens if they do.
The Core Numbers Behind The Warning
Using recent system-wide balance-sheet data, the economists estimated that U.S. banks held roughly seven trillion dollars of long-term interest-rate exposure. About 80 percent of that exposure, or 5.8 trillion, was supported by the behavioral stability of deposits other than large time deposits. Those deposits effectively act as a long-duration funding source even though they are contractually short-term.
One modeled scenario assumed a 10 percent increase in deposit rate sensitivity. Under that assumption the system’s capacity to hold duration risk fell by about 700 billion dollars measured in ten-year Treasury equivalents. A second scenario shortened average deposit life by 10 percent and produced a reduction of roughly 580 billion dollars in maturity-transformation capacity. These are not forecasts of actual loan losses. They are measures of how much less long-term exposure banks might choose to keep on their books if funding becomes less predictable.
Instant settlement would allow deposit holders who prioritize yield to switch banks almost instantaneously.
That single sentence captures the practical concern. Once the technical barriers drop, the behavioral barriers may follow. I’ve found that when technology removes friction, habits often change faster than models expect.
How Banks Might Respond To Less Stable Funding
Banks have several levers. They can raise the rates they pay on deposits to keep money from leaving. That raises funding costs and squeezes the net interest margin that supports lending. They can hold larger buffers of reserves and government securities, which reduces the share of the balance sheet available for longer-term loans. Or they can issue more wholesale term debt to replace the lost duration of deposits. Each path tends to push the cost of credit higher for borrowers.
Perhaps the most interesting aspect is how these responses interact. Higher deposit rates attract or retain funds but compress margins. Larger liquidity buffers protect against outflows but lower the overall return on assets. Extra wholesale debt preserves lending volume yet introduces a more expensive and less flexible funding mix. None of the options is free.
- Higher deposit rates to reduce switching incentives
- Increased holdings of reserves and short-term government securities
- Greater reliance on term wholesale funding markets
- Potential shrinkage in the volume of longer-term loans relative to the size of the system
Any combination of the above would likely raise the average cost of credit across the economy. Mortgages, commercial loans, and consumer credit could all feel the pressure over time.
Early Evidence From Instant Payment Systems
A useful, if imperfect, comparison comes from research on Brazil’s real-time payment network. Banks that saw heavier use of instant payments tended to shift their asset mix toward more liquid government bonds and away from loans. The parallel is not exact. Instant payments are not the same as tokenized deposits, and the Brazilian banking structure differs from the American one. Still, the direction of the adjustment is consistent with what the new estimates suggest: when funding becomes more flighty, balance sheets become more liquid and less oriented toward long-duration lending.
I keep coming back to that pattern. Technology that improves payment convenience can simultaneously change the economics of credit creation. The two effects do not cancel each other out; they coexist, and the net impact on borrowers depends on how banks and regulators adapt.
What Banks Are Already Building
Despite the funding risks, large institutions continue to invest in shared tokenized deposit infrastructure. Projects aimed at 24/7 settlement, automated workflows, and interoperability are moving forward. Regional and community banks are also exploring collective platforms so they are not left behind. The design choices matter enormously. Greater interoperability makes payments smoother and competition for deposits more intense at the same time.
In my experience watching financial innovation cycles, the institutions that move first often shape the standards that everyone else must live with. That creates both opportunity and pressure. Banks that can offer attractive programmable features may retain sticky customer relationships. Those that cannot may face faster outflows once the technology is widely available.
The Difference Between Capacity And Actual Outflows
It is worth stressing again what the 700 billion and 580 billion figures do not mean. They do not represent deposits expected to leave the banking system. They measure a possible reduction in the amount of long-term interest-rate risk banks are willing to hold. The distinction is important. Capacity can shrink even if total deposit volumes stay roughly stable, simply because the remaining deposits behave more like short-term wholesale funding.
Think of it as the difference between the size of a reservoir and how reliably the water stays there. Tokenization does not necessarily empty the reservoir. It can make the water more restless. Banks then decide how much of their long-term lending they are comfortable supporting with that restless funding.
Possible Paths Forward For The System
Several outcomes seem plausible. Banks might successfully raise deposit rates just enough to keep most balances in place while accepting thinner margins. They might accelerate the shift toward more liquid asset portfolios and accept a smaller role in long-term credit markets. Or they might lean more heavily on capital markets for term funding, transferring some of the maturity transformation outside the traditional banking system.
Regulators will watch deposit beta, liquidity coverage, and the growth of wholesale funding closely. Differences across bank sizes could become more pronounced. Larger institutions with better technology and deeper capital markets access may adapt more easily than smaller ones. That raises questions about competitive balance that go beyond the pure interest-rate calculations.
I’ve found that the most durable innovations are the ones that solve a real customer problem without creating larger systemic frictions. Tokenized deposits clearly solve problems around speed, programmability, and settlement finality. The open issue is whether the funding-side consequences can be managed without a lasting increase in the cost of credit.
Why The Timing Feels Significant
Interest rates have already spent several years at elevated levels. Banks have lived through a period of rising deposit betas and competition for funds. Adding a technology layer that further increases mobility arrives on top of an already more sensitive funding base. The marginal impact of tokenization may therefore feel larger than it would have in a low-rate, high-friction environment.
At the same time, the infrastructure build-out continues. Shared networks, industry alliances, and pilot programs keep advancing. The technology is not waiting for perfect clarity on the funding implications. That mismatch between technical progress and economic understanding is familiar; it has appeared with almost every major financial innovation of the last two decades.
Practical Implications For Borrowers And Savers
For households and businesses the near-term effect may be subtle. Deposit rates could become more competitive as banks fight to retain balances. Loan rates might edge higher over time if funding costs rise or if banks choose more conservative asset mixes. The net result for any individual depends on whether they are a net saver or a net borrower and on how quickly the system adapts.
Savers who actively monitor yields could benefit from the new mobility. Those who leave money idle may still see better rates simply because banks feel greater competitive pressure. Borrowers, especially those seeking longer-term fixed-rate credit, could face a less favorable environment if the maturity-transformation capacity of the system contracts.
Looking At The Assumptions More Closely
The estimates rest on a handful of simplifying assumptions. Average deposit life of roughly four years, a 10 percent shift in sensitivity or duration, and aggregate matching of assets and liabilities. Real-world outcomes will vary by institution, by customer segment, and by the specific design of tokenized systems. Some deposits may remain sticky because of payroll relationships, payment history, or simple inertia. Others may become highly mobile.
Agentic software that monitors rates and moves funds automatically remains more theoretical than widespread at the moment. Yet the technology path is clear enough that ignoring the possibility feels unwise. Once the tools exist and the user experience is smooth, adoption can accelerate quickly.
In my own reading of similar transitions, the early adopters often create demonstration effects that pull the rest of the market along. If a meaningful share of sophisticated depositors begins to treat balances as highly liquid yield-seeking capital, the rest of the deposit base can start to look more rate-sensitive even if many customers never open a blockchain wallet.
Balancing Innovation And Stability
The tension is not new. Faster payments, more programmable money, and greater customer control are genuine improvements. The question is whether the funding model that has supported long-term lending for decades can absorb the change without raising the price of credit. The research does not claim to have a definitive answer. It simply quantifies one possible channel through which costs could rise.
Banks that design their tokenized offerings with thoughtful retention features, clear rate strategies, and robust liquidity management will be better positioned. Those that treat tokenization purely as a product feature without considering the balance-sheet consequences may discover the costs later and under less favorable conditions.
Perhaps the most useful takeaway is that technology rarely leaves the economics of banking unchanged. Every increase in speed and flexibility on the liability side invites a corresponding adjustment on the asset side or in the pricing of risk. The 700 billion and 580 billion figures are best understood as early markers of the scale that adjustment might reach under plausible assumptions.
What Remains Uncertain
Several variables will determine the actual outcome. How widely depositors adopt automated yield-seeking tools. How banks choose to compete on rates versus services. How regulators calibrate liquidity and capital rules for a more mobile deposit base. How capital markets absorb any additional demand for term wholesale funding. And how customer behavior evolves once the novelty of tokenization wears off.
None of those factors is fixed today. That leaves room for a range of results, from modest margin compression to a more material re-pricing of longer-term credit. The research simply shows that the potential magnitude is large enough to deserve serious attention while the infrastructure is still being built.
I’ve watched enough financial technology cycles to know that the loudest early claims often overstate benefits and understate second-order effects. Tokenized deposits are no exception. The convenience is real. The funding implications are also real. Managing both at once will require careful design, clear data, and a willingness to adjust as evidence accumulates.
Putting The Scale In Perspective
Seven hundred billion dollars of reduced duration capacity is a significant number even in a multi-trillion-dollar banking system. It is large enough to influence aggregate credit conditions if it materializes. At the same time it is small enough that thoughtful adaptation by banks and policymakers can probably contain the impact. The difference between those two statements is the difference between a manageable evolution and a more disruptive shift.
The coming years will show which path is more likely. In the meantime the estimates provide a concrete starting point for discussion rather than abstract concern. They turn a vague worry about “less sticky deposits” into measurable quantities that balance-sheet managers and risk officers can stress-test against their own portfolios.
That practical quality is what makes the work useful. Speculation about technology is easy. Translating the speculation into approximate dollar impacts on the system’s ability to take duration risk is harder and more valuable. Whether the actual numbers turn out higher or lower than the initial estimates, the framework itself helps focus attention on the right questions.
A Longer View Of Deposit Behavior
Deposit stickiness has never been purely contractual. It has always contained a large behavioral component. People leave money in place because of habit, convenience, relationships, and the mental cost of moving it. Technology that lowers those costs changes the behavioral equation even if the legal terms of the deposit stay the same.
We have already seen versions of this with online savings accounts and rate comparison tools. Tokenization and programmable settlement take the process further by removing remaining delays and enabling automated responses. The direction of travel is clear. The speed and the ultimate magnitude remain open.
In that sense the new research is less a warning against the technology itself and more a reminder that balance-sheet consequences travel with product innovation. Ignoring those consequences does not make them disappear. Measuring them early improves the odds of managing them well.
As the shared networks move from announcement to live operation, the real data will start to arrive. Deposit betas, average lives, and liquidity metrics will either confirm the direction of the estimates or show that behavioral inertia is stronger than expected. Until then the prudent stance is to treat the calculated reductions in duration capacity as a credible risk scenario rather than a remote possibility.
The conversation is no longer about whether tokenized deposits will exist. It is about how the banking system will fund long-term credit once they become common. That is a more complicated and more important question, and the recent estimates give it the concrete framing it needs.
Banks, borrowers, and policymakers all have a stake in the answer. The technology is advancing on its own timetable. The economic adjustments will follow on theirs. Keeping the two in reasonable alignment is the practical challenge that now sits in front of the industry.
