Russia's 1,250% Risk Weight: Inside the Basel Clone Behind the 1% Crypto Cap

1,250% risk weight. Under an 8% minimum capital adequacy ratio, that arithmetic resolves to a 100% capital charge. A Russian bank holding one ruble of Bitcoin on its own book must fund that ruble with a ruble of regulatory capital. Not leverage. Not a buffer. Full deduction.

That is the number buried in the Bank of Russia's draft prudential framework for cryptoasset exposures, circulated for public comment ahead of a targeted Q4 2026 issuance. The figure everyone quotes is 1% — the ceiling on a credit institution's crypto exposure relative to its own funds. The 1% gets the headline. The 1,250% does the work.

I have spent the better part of a decade reading capital rules the way most people read price charts. This draft is not the work of a regulator improvising. It is a high-fidelity localization of the Basel Committee's cryptoasset standard, and it is considerably more sophisticated than the coverage implies.

Context: where this rule actually comes from

In December 2022 the Basel Committee on Banking Supervision published its prudential treatment of cryptoasset exposures. The framework splits the asset class in two. Group 1 covers tokenized traditional assets and stablecoins that pass strict redemption and stabilization tests. Group 2 covers everything else — Bitcoin, Ether, and the long tail that does not qualify.

Group 2 splits again. Group 2a assets can, under conditions, receive a modified treatment with a recognized hedging regime. Group 2b — the catch-all — carries a 1,250% risk weight and a hard exposure limit of 1% of Tier 1 capital. That is where Bitcoin lives by construction.

Russia's draft maps onto that skeleton almost joint for joint. Two ratios instead of one. A penetration-tested exposure numerator. A restricted hedging and netting regime. A dual risk-weight track. The 1% ceiling. Anyone who has read the Basel text and then reads the Russian text will recognize the lineage immediately.

What the draft adds is local architecture, and the local architecture is driven by forces that have nothing to do with prudential elegance.

Sanctions have severed conventional cross-border rails. Ruble-denominated capital controls are tightening. Demand for crypto settlement and store-of-value among Russian households and corporates keeps climbing anyway, because the underlying need — moving value across a hostile border, holding something the banking system cannot freeze — does not disappear when a regulator disapproves of it. The Bank of Russia has spent years resisting that demand. This proposal is the compromise. Crypto does not vanish. It enters through a gate the state can watch.

Timeline matters more than most coverage admits. The draft is not law. Issuance is targeted for Q4 2026, with a ten-day implementation window after publication. Reporting obligations — turnover plus the N31 and N32 ratios — begin January 2027. That is a multi-year runway between announcement and anything resembling enforcement. Anyone pricing this as an immediate event is misreading the clock.

The dual-ratio architecture, and why the denominator binds

First architectural decision: two ratios, not one.

N31 applies to a single credit institution, denominated by that institution's own funds. N32 applies at the bank group level on a consolidated basis, denominated by group capital.

The elegance is in how the denominator binds the numerator. A group cannot inflate its tolerance for crypto exposure by shuffling capital between subsidiaries, because the consolidated ratio captures the aggregate position regardless of where it sits. A subsidiary cannot hide exposure behind a parent guarantee, because the solo ratio captures it against the entity's own funds. Two perspectives, one exposure.

I ran this exact check in 2017, when I bypassed the marketing decks on twelve high-profile ICOs and audited the contracts directly — including a preliminary pass at Golem's allocation mechanics. The pattern I learned then holds for regulators: the moment a rule has only one measurement point, someone builds a wrapper to move the measurement. N31 and N32 exist so that there is no single point to move.

Second decision: the numerator penetrates.

Covered exposure is not spot holdings. The draft language reaches direct and indirect investments, derivatives priced off crypto, and any loan, bond, guarantee, repo, or credit facility whose settlement or value depends on crypto. It also reaches foreign digital instruments — offshore-issued digital assets and digital rights.

That is anti-circumvention by construction. A bank that wants crypto beta cannot reach it through a structured note. It cannot reach it through a total-return swap. It cannot reach it through a repo book, a guarantee, or a credit line collateralized by tokens. The exposure follows economic substance, not the legal wrapper.

I chased this failure mode in 2020, when I led a team to scrape early OnyxDAO governance votes and cross-reference them against Uniswap liquidity pools. The protocols that collapsed were not the ones with obvious bad numbers on the dashboard. They were the ones where real risk sat inside a container no metric measured. Russia's draft appears to close that container — including the offshore container, which is a sentence worth reading twice.

Foreign digital instruments inside the capital numerator is not a prudential provision. It is a capital-control provision wearing a prudential provision's clothes. A Russian bank cannot escape the ratio by parking exposure in offshore stablecoins or foreign-issued digital rights. It has to count them, and the 1,250% weight follows them home.

Third decision: netting and hedging are deliberately throttled.

Long and short positions can only be netted within eligible lower-risk categories, and even then only with conditions attached around the underlying asset, settlement mechanics, maturity, and freeze or liquidity risk. Direct holdings and higher-risk exposures cannot be fully offset by hedges. The mapping onto Basel's differentiated Group 2a and Group 2b handling is close enough that I would assign it moderate-to-high confidence.

Translation for anyone who has run a desk: a bank cannot short a perpetual against a spot book and declare itself flat. It cannot claim hedge effectiveness on an instrument that can be frozen at the exchange layer. Regulators learned in 2008 that netting agreements are only as good as the counterparty's willingness to honor them in a crisis. This draft does not relearn that lesson.

Fourth decision — and this is the one that carries the policy.

Non-liable client custody positions carry a 50% risk weight. Proprietary exposure, plus any client position where the bank bears the loss, carries 1,250%.

Do the arithmetic.

50% × 8% = a 4% capital charge on custodied client assets. Survivable. Pricable. A custody business can build a fee schedule around that.

1,250% × 8% = a 100% capital charge. That is not a capital requirement. That is a prohibition wearing a ratio's clothing.

This is the central engineering insight, and it is not being reported clearly. The dual track is not a gradation of severity. It is a binary. The Bank of Russia is telling its banking system that it may hold crypto for other people and may not hold crypto for itself. The 1% exposure ceiling becomes almost redundant under a 100% capital charge — no rational balance sheet builds a position priced at full deduction — but the ceiling exists as a hard backstop for the entity that decides to do it anyway.

The definitional hinge is liability. Who absorbs a seizure? A freeze? A loss? The draft introduces digital depository institutions as a category, and whether a client's position lands inside N31 or N32 turns on that single legal question: does the depository bear the loss, or does the client?

This is where implementation risk concentrates. Legal characterization disputes are the tax on sophisticated rules. I watched this play out in 2022, when I went straight to the Solana ledger after FTX collapsed and traced roughly $1.2 billion in transfers to Alameda-linked accounts inside 48 hours. The on-chain record was unambiguous. What was ambiguous — and what took months of litigation to resolve — was the legal characterization of each flow. Balance-sheet metrics are only as good as the classification feeding them. Russia's draft leaves a wide surface for exactly that dispute.

The comparison that matters

Canada's 2027 capital rules, per the same coverage cycle, provide relief for cross-exchange hedging — a bank can get recognition for offsetting positions across venues. Russia's draft does the opposite. It confines netting to qualified lower-risk categories and refuses full offset for direct and high-risk holdings. Same Basel parent. Two very different children. Canada treats hedging as a risk mitigant. Russia treats it as a loophole.

The EU's MiCA sits in a different layer entirely, regulating the market rather than the bank capital stack.

And here is what the 50% custody weight is not. It is not 0%. Cash and sovereign debt sit at or near zero. A 50% risk weight states plainly that crypto custody is a medium-risk activity, not a neutral one. The Bank of Russia is permitting a business. It is not endorsing an asset. Anyone claiming this draft treats custody as routine is reading a rate card that does not exist.

The contrarian angle: protection with a condition attached

The claim circulating is that Russia's harsh 1% cap actually protects bank customer assets. It does — conditionally, and for reasons that have almost nothing to do with the customer.

The logic chain runs: protect bank capital, protect the banking system, indirectly protect depositors. Under that chain, the 1,250% weight is a firewall between depositor money and crypto volatility. Fine. The chain holds. Now break the condition.

The 50% custody weight protects the client only if two things are simultaneously true: the depository does not bear the loss, and the depository is operationally sound. If a state-linked Russian custodian becomes the concentrated holder of key material — and the accompanying arrangements point in that direction — then the protection is only as strong as one custodian's key management, one custodian's governance, one custodian's solvency under a sanctions regime that could tighten at any moment.

That is a single point of failure provisioned into national infrastructure. In 2021 I traced coordinated wash-trading wallet clusters across Ethereum and Polygon to a single entity, published the transaction hashes, and forced exchanges to pause affected pairs. The lesson was not that manipulation exists. The lesson was that concentration is fragile in ways that stay invisible until the instant they become catastrophic. Decentralization is not a slogan in custody. It is an operational risk parameter.

Two things the protective framing obscures entirely.

First, the 1% cap will not reduce Russian crypto activity. It will relocate it. Demand is driven by sanctions evasion and store-of-value need, and demand does not respond to a bank capital ratio. When compliant channels are expensive, slow, or monitored, activity migrates to peer-to-peer, to offshore venues, to gray corridors. The practical result of a strict bank cap in a high-demand jurisdiction is a colder official channel and a hotter unofficial one.

Second, the word protection is doing double duty. A rule that caps bank exposure at 1% of own funds and prices proprietary holdings at a 100% capital charge keeps crypto out of the banking system's risk surface. That is bank-system protection. Customer asset protection is a downstream effect — contingent on custody design that the same regulator is simultaneously centralizing.

Follow the capital, not the narrative. The capital says the state wants a pipe, not a portfolio.

Takeaway: four signals to track, none of them the headline

Watch the final text. Q4 2026 issuance, ten-day implementation. Ratios can move in a consultation window, and a draft that reads harsh today can soften.

Watch the N31 and N32 disclosures beginning January 2027. If no Russian bank reports meaningful custody exposure, the digital depository category is decorative architecture.

Watch the key-custody arrangements. If state-linked banks concentrate key material, the 50% custody exemption stops being a commercial permission and becomes a monitoring pipe.

Watch offshore and peer-to-peer volumes. That is where you will learn whether the cap protected customers or simply relocated them.

Code doesn't lie. Neither do capital ratios. The 1,250% weight is the real policy. Everything around it is commentary.