Crypto Recreates Intermediaries Despite Decentralization Promise, CEPR Report Finds

Bitcoin (BTC) was built on a radical promise: money that works without trusted intermediaries. Yet 17 years after the first block challenged the post-crisis banking order, a growing body of research and market evidence suggests the crypto economy is rebuilding the very gatekeepers it set out to remove—only now under more technical labels such as stablecoin issuers, custodians, auditors, or DeFi ‘block builders’.

That is the central argument highlighted in a recent report from the Centre for Economic Policy Research (CEPR), which assesses blockchain’s original ambition against how today’s on-chain finance actually functions. The verdict, in the report’s framing, is a paradox: decentralization succeeded at maintaining shared ledgers without a central authority, but modern crypto markets increasingly reintroduce centralized points of control where trust—and risk—concentrates.

The early achievement remains hard to dispute. Bitcoin proved that a global network of participants who do not know or trust one another can still agree on a single history of transactions, adding blocks roughly every 10 minutes and preventing double spending without relying on banks or governments. Ethereum (ETH) expanded the model by popularizing smart contracts—code that automatically executes agreements—turning the decentralization of money into a broader decentralization of ‘contract execution’.

But as blockchain systems scaled beyond experimentation into payments, trading, credit, and tokenized assets, the market began filling what traditional finance would call “institutional gaps.” The result is not necessarily a failure of the technology; it is a reminder that finance runs on layered trust. The question is where that trust is anchored—and whether users understand the new dependencies that come with it.

Stablecoins: the return of issuer risk

No segment illustrates the shift more clearly than stablecoins. They are often marketed as a bridge between blockchain settlement speed and the price stability of fiat currencies such as the U.S. dollar. In practice, however, most leading stablecoins are not “purely” decentralized instruments. They are issued by private entities, backed by off-chain reserve assets such as Treasury bills, bank deposits, and money market instruments, and supported by relationships with custodians, banking partners, and auditors.

That architecture introduces concentrated risk. If reserves are opaque, users are forced back into ‘issuer trust’. If reserves are highly concentrated at a small number of institutions—or tied to less liquid assets—redemption promises can come under stress during market shocks. And if an issuer can delay or restrict redemptions, the product begins to resemble a private bank liability more than a frictionless digital dollar.

The CEPR report’s broader point is not that stablecoins are inherently harmful; rather, they can create a mismatch between branding and reality. A token may move on a public blockchain, but the core of its stability rests on centralized balance sheets and off-chain governance.

DeFi: code replaces humans, but power still centralizes

Decentralized finance (DeFi) was supposed to eliminate brokers by replacing them with algorithms. Yet DeFi has developed its own intermediary layers—some of them less visible to everyday users. Transparency is a strength, but it also creates new avenues for exploitation. Because transactions are publicly broadcast before being finalized, sophisticated actors can observe pending orders and insert trades before and after them to extract profit, a practice often described as a ‘sandwich attack’. In effect, traditional front-running has reappeared in on-chain form.

At the infrastructure level, Ethereum’s transaction supply chain has also evolved. Specialized actors known as ‘builders’ assemble transaction bundles into blocks, and their role can become a locus of influence. While the network is open in theory, block construction power can gravitate toward a small set of well-capitalized or technically advanced players—recreating the dynamics of concentrated intermediation, only behind more complex machinery.

The outcome is a familiar pattern: humans disappear from the interface, but economic incentives and coordination still create intermediaries. In many cases, they re-emerge with more opaque accountability than traditional financial institutions.

Tokenization meets the real world’s legal rails

Tokenization of real-world assets (RWA) faces a related constraint. Converting real estate, bonds, equities, vouchers, loyalty points, or other claims into tokens can improve transferability and potentially liquidity. But a token does not automatically enforce legal rights in the off-chain world. Owning a real estate token does not, by itself, update a property registry; holding a tokenized bond does not guarantee that a court will recognize the claim without a supporting legal framework.

Once assets touch the real world, intermediaries return: notaries, depositories, courts, supervisors, and ‘oracles’ that feed external information to smart contracts. The trust problem shifts rather than disappears.

South Korea’s policy dilemma: beyond “allow or ban”

The debate has immediate relevance in South Korea, where policymakers and industry are simultaneously discussing a digital asset framework law, a won-denominated stablecoin, tokenized securities, and broader RWA tokenization initiatives. Much of the public discussion still defaults to a binary question—whether to allow or restrict new products—when the more consequential issues are design and accountability.

Key questions, according to the CEPR analysis and echoed by market structure concerns, are operational: Who issues the asset? Who manages the ledger operations users depend on? Who holds reserves, and in what instruments? Who verifies external data? Who is legally responsible when something fails?

For a won stablecoin, reserve ratios alone are not the core safeguard. More important is whether reserves are held in sufficiently safe and liquid assets, whether they are excessively concentrated at specific institutions, whether they can be verified frequently and credibly, and whether user redemption rights are legally enforceable. If an issuer can unilaterally limit redemptions in a crisis, the system risks becoming a source of financial instability rather than a payments upgrade. In that sense, a won stablecoin would function less as a crypto product and more as part of national payment infrastructure.

The same logic applies to tokenized securities and fractional investment products. The challenge is not merely splitting ownership into smaller units; it is ensuring those fragmented rights—dividends, interest, voting, redemption—can be exercised in practice. Where on-chain records and off-chain rights conflict, dispute resolution mechanisms must be clear. Otherwise, tokenization may increase disputes faster than it increases liquidity.

DeFi, too, requires a more sober lens than the slogan “no intermediaries.” Effective oversight depends on identifying who controls critical choke points: web front ends, oracle systems, transaction ordering, and liquidity provisioning. Finance ultimately revolves around power and responsibility. If power concentrates while responsibility disperses under the banner of decentralization, the result begins to look less like innovation and more like regulatory arbitrage.

A market that keeps inventing new middlemen

The broader implication is not that blockchain has failed. Bitcoin and Ethereum delivered foundational breakthroughs, and stablecoins and tokenization could still reshape payments, settlement, and asset transfer. Countries that shut the door entirely risk falling behind in financial infrastructure competition.

But enthusiasm is not the same as illusion. Decentralization is not an object of faith; it is a system design problem that demands clear rules, transparency, and accountability. Markets do not tolerate vacuums. When one intermediary is removed, another tends to emerge—whether a bank, issuer, custodian, oracle operator, or block builder. Different names do not eliminate familiar risks.

For South Korea, the next phase of digital asset competition may hinge less on token prices than on how effectively ‘trust architecture’ is built into law, supervision, and market structure. The central task is no longer whether intermediaries can be eliminated, but which intermediaries are acceptable, which must be constrained, and how critical trust functions are kept within enforceable public rules.