'Code as Law': The tokenization of financial assets and the paradox of programmability

In the rapidly evolving landscape of financial technology, the tokenization of financial assets — in which physical-world financial assets are made exchangeable on the blockchain — has emerged as a transformative innovation, drawing significant attention from major financial institutions and regulatory bodies worldwide.

As market participants race to represent traditional assets digitally on programmable platforms, they face a fundamental paradox — the paradox of programmability — that may reshape financial intermediation itself. Regulators, meanwhile, will continue to struggle to draw the line between finance and technology.

Programmability in the context of tokenized assets refers to the ability to encode specific rules and automated actions directly into the digital tokens themselves. This means that transactions and processes involving these assets can be executed automatically based on pre-defined conditions, much like the "if this, then that" logic used in computer programming. This feature sets tokenization apart from simply representing assets digitally, as it introduces a layer of automation and removes the need for manual intervention or traditional intermediaries.

The paradox of programmability, first observed in decentralized finance (DeFi) systems that catalyzed the use of smart contracts interacting with cryptographic tokens, has two key aspects: while programmability offers enhanced efficiency and transparency through automation, it simultaneously constrains intermediaries' ability to make discretionary decisions. Indeed, with most financial markets, regulators readily identifiable entities to hold to account with entity or activity-based regulation. With tokenized assets, as with broader digital asset markets, code and market conduct may be as important.

Asset tokenization as a concept remains ill-defined since Ethereum, the often-referenced blockchain platform that inspired the tokenization movement, is actually an account-based shared ledger. Regulatory bodies have emphasized the need for permissioning in tokenization to distinguish the effort from DeFi. However, a permissioned version of such a ledger run by a centralized intermediary is perhaps not so different from existing ledgers already used for recording book-entry settlements, such as the ones maintained by the Depository Trust & Clearing Corporation for securities settlement or the Federal Reserve for Fedwire funds transfers. In short, the real benefits of tokenization may not be reaped, unless the blockchain ledgers on which it rides are open, permissionless and support programmability.

In this context, programmability is likely the key defining feature of tokenization that goes beyond the mere digital representation of assets on a shared ledger. As highlighted in a recent report by the Committee on Payments and Market Infrastructures (CPMI), "Digital tokens cannot exist independently of the programmable platform, since the issuance, recording and transfer of tokens relies on the execution of functions on the platform."

The "if this then that" logic that defines the programmability of shared ledger platforms can both reduce and amplify risks. On one hand, programmability enables automated risk controls, such as payment versus delivery protections that ensure assets are only transferred when corresponding payments are received. On the other hand, programmability can also magnify risk differences and accelerate run dynamics in ways not yet seen in traditional financial systems.

This amplification occurs because programmable platforms enable the automation of rational individual behaviours in a frictionless, transparent and high-velocity financial environment. When market participants can instantly and automatically react to even small differences in risk exposures between similar assets, these minute variations get magnified through cascading effects. The stablecoin sector has already demonstrated this phenomenon, where small deviations in perceived reserve backing quality led to rapid shifts in secondary market prices and occasional run dynamics as automated trading strategies responded to these signals.

Traditional financial institutions, particularly commercial banks pursuing tokenization strategies, will need to carefully consider these dynamics. The automation and instant settlement capabilities that make tokenization attractive could also make their deposit bases more volatile and sensitive to subtle variations in asset quality or risk profiles. This susceptibility to risk signals is particularly acute with money tokens or stablecoins, which represent a form of digital demand deposit on an underlying balance sheet.

The Financial Stability Board has highlighted key risks that emerge from tokenized deposits. As noted in their recent report, "Designing tokenised deposits [as a transferable claim] could introduce unique financial stability implications if not addressed by regulation and supervision, including the potential for a secondary market and a deviation from par that could affect the ‘singleness of money’ and which could undermine confidence in financial markets. In addition, the programmability feature of DLT-based tokens could lead to automatic transactions based on specific pre-determined triggers. In the case of tokenised deposits, such automatic transactions could increase herding behaviour or the risk of (programmed) bank runs."

These risks point to a fundamental transformation in banking models. The tokenization of deposit-like liabilities will likely push institutions toward narrower forms of banking, where they maintain only highly restricted sets of short-term, high-quality assets as backing. This shift is driven by programmability's tendency to amplify even minor risk differences, necessitating that transferable liabilities among regulated institutions are limited to those with highly standardized and fungible asset backing - similar to proposed rules for payment stablecoins in the U.S. We can already observe this trend in early tokenization projects on public blockchains, such as tokenized money market funds. While these projects have shown rapid growth, their relatively small scale demonstrates how tokenization naturally gravitates toward narrow banking models with reduced banker discretion.

The amplification of risk differences through programmability also creates a clear imperative to separate payment and savings functions. To prevent programmed reaching-for-yield behaviour and automated bank runs, non-yield-bearing payment tokens, such as stablecoins, will likely need to be strictly segregated from interest-bearing yield coins such as tokenized funds used for savings purposes.

At its core, programmability transforms code into enforceable legal contracts or specific market functions when interacting with tokenized asset values. This "code as law" characteristic inherently constrains intermediaries' discretionary choices. While this constraint helps reduce moral hazard in financial institutions systemically, it also means individual intermediaries will face increasing commoditization. With limited discretion over asset composition and other operational choices, institutions will find it harder to differentiate themselves, likely leading to an erosion of traditional market power as their functions become increasingly automated.

This transformation points to a broader shift in financial markets. Tokenization is poised to increase liquidity in previously illiquid assets while simplifying the balance sheet structure of financial institutions in terms of asset composition.

The result will be a movement away from traditional balance-sheet-heavy intermediation in credit and payment activities toward more market-based price discovery and risk sharing, fundamentally changing how risks and capital are allocated across the economy. As tokenization gains momentum, regulators would do well in regulating the activity and not the technology.