The Hidden Russian Bitcoin Flow

The Hidden Russian Bitcoin Flow

A Kaido Lab field report on sanctions-era settlement networks, mining, OTC liquidity, and the search for structurally identifiable BTC-specific order flow

Research date: 4 August 2026 Research status: hypothesis-development and identification design; no claim of proven structural causality Scope: Russia-linked state, state-adjacent, and large private settlement systems involving mining, A7/A7A5/Grinex-like networks, OTC liquidity hubs, stablecoin bridges, cross-border netting, and Bitcoin market microstructure

Executive summary

Russia now has the legal and institutional components required for a ring-fenced cryptocurrency settlement system: cryptocurrency is permitted for cross-border corporate settlement, domestically mined Bitcoin has publicly been used in foreign trade, mining is subject to registration and energy-location controls, and large alternative-payment networks have emerged around state-owned or state-linked institutions. The Bank of Russia’s July 2026 framework preserves the prohibition on domestic cryptocurrency payments while allowing exporters and importers to transact across borders without volume restrictions and through any wallet or cryptocurrency. citeturn1search0turn3search0turn7search3

A7 is the clearest publicly documented institutional case. The U.S. Treasury describes A7 as a Russian cross-border settlement provider owned by sanctioned Promsvyazbank and Ilan Shor, and identifies Grinex as a successor infrastructure created after the March 2025 disruption of Garantex. Treasury states that customer balances were transferred through the ruble-backed A7A5 token and that Grinex facilitated billions of dollars in cryptocurrency transactions. citeturn2view0turn4search1

Blockchain investigations indicate that the crypto component of this architecture is real but is not primarily a Bitcoin system. TRM identifies two principal roles: converting USDT into ruble liquidity through Moscow cash-market infrastructure, and using A7A5 as an internal accounting and book-balancing instrument. Elliptic found that A7A5’s main exchange pairs connected rubles to USDT, while Chainalysis observed predominantly weekday activity consistent with business rather than retail use. citeturn2view2turn2view4turn2view5

This is the report’s central distinction:

[ \boxed{ \text{A large Russia-linked crypto settlement system is confirmed} } ]

but:

[ \boxed{ \text{A large, persistent BTC-specific residual flow is not confirmed} } ]

The strongest BTC-specific mechanism is the use of newly mined Bitcoin to pay foreign suppliers. When BTC is purchased in the open market and subsequently sold by the recipient, the market receives a buy impulse followed by a sell impulse. When freshly mined BTC is transferred and the recipient sells or short-hedges it, there is no preceding open-market purchase. That can create one-sided external supply:

[ \text{electricity and ruble costs} \rightarrow \text{new BTC} \rightarrow \text{foreign supplier} \rightarrow \text{spot sale or derivative hedge} ]

Russia’s finance minister confirmed in December 2024 that Russian-mined bitcoins were already being used in international transactions. Reuters later reported that BTC, ETH, and USDT were being used in a small but growing part of oil-trade settlement with China and India. citeturn3search0turn3search1

The crucial variable is not gross crypto settlement volume. It is:

[ \boxed{ Q_t^{residual}

Q_t^{gross}

• Q_t^{internal\ netting}
• Q_t^{inventory}
• Q_t^{offsetting\ hedge} } ]

Russian officials have explicitly described the use of internal netting, crypto, and gold to resolve cross-border obligations. TRM’s reconstruction of A7 also finds extensive circular transactions associated with internal book settlement. Therefore, a network may process tens of billions of dollars while creating little immediate global BTC order flow. citeturn3search2turn2view2

The report’s prior is:

Question	Assessment
Does a large Russia-linked alternative settlement architecture exist?	Confirmed
Does it use public blockchains and crypto liquidity?	Confirmed
Are stablecoins and internal ledgers currently more important than BTC?	Strongly supported
Is mined BTC used in foreign trade?	Confirmed, scale unknown
Can mined-BTC settlement create one-sided BTC supply?	Mechanically valid
Is that supply presently large enough to drive global BTC trends?	Unproven and probably intermittent
Can the flow create short-lived basis, lead–lag, volatility, or BTC/ETH residuals?	Plausible and testable
Is the official CBR USD/RUB fixing the causal driver?	Unlikely; more plausibly a downstream or common-factor variable
Is deployable alpha already established?	No

The most promising research target is not a broad “Russia factor.” It is a state-dependent composite:

[ \boxed{ \text{BTC/RUB residual basis}

• \text{probabilistic young-coin flow}
• \text{venue lead–lag}
• \text{sanctions substitution} } ]

A robust finding would require four layers of validation:

1. a measurable anomaly;
2. out-of-sample predictive power;
3. a credible Russia-linked flow chain;
4. a natural experiment or documentary confirmation that separates settlement flow from ordinary crypto-market activity.

The investigation can be conducted without interacting with sanctioned entities. Execution should remain on lawful, screened, regulated venues; sanctioned platforms, wallets, counterparties, and prohibited payment routes should be used only as observational labels where legally permissible. OFAC states that sanctions obligations apply to virtual currency as they do to fiat and that U.S. persons generally may not transact with or facilitate transactions involving blocked persons. The UK, EU, and United States have specifically sanctioned components of the A7A5, Grinex, and related ecosystem. citeturn4search0turn4search10turn4search2turn4search3

Evidence base and institutional timeline

What is confirmed

Russia’s experimental legal regime for cross-border crypto settlement was enacted in 2024 after payment friction increased under sanctions. Reuters reported that the law was intended to allow international cryptocurrency payments and that the Bank of Russia expected initial transactions before the end of 2024. In December, Finance Minister Anton Siluanov stated that settlements involving Russian-mined Bitcoin had begun. citeturn3search8turn3search0

Reuters subsequently reported a settlement chain used in some Russia–China and Russia–India oil transactions: a buyer paid yuan or rupees to an intermediary, the funds were converted into cryptocurrency, the crypto was transferred through additional accounts, and it was ultimately converted into rubles. The report described BTC, ETH, and USDT, but also stressed that crypto remained a small fraction of Russia’s much larger oil trade. citeturn3search1

In July 2025, the head of Russia’s financial-monitoring service publicly described netting, gold, and cryptocurrency as active cross-border payment mechanisms. Under netting, domestic agents can satisfy exporters and importers locally while reducing the number and value of externally transmitted payments. citeturn3search2

The A7 ecosystem then became a visible institutional layer. Treasury’s August 2025 action stated that A7 provided cross-border settlement platforms, was owned by Promsvyazbank and Ilan Shor, and had worked with Garantex-linked infrastructure to create and distribute A7A5. The same action stated that Grinex was built by Garantex personnel after the March 2025 law-enforcement disruption and that customer positions were restored with A7A5. citeturn2view0

The Department of Justice states that the March 2025 Garantex operation seized domains and servers, obtained customer and accounting databases, and froze more than $26 million in funds. This event offers a potentially useful natural experiment because it abruptly disrupted a known Russia-linked crypto-liquidity node. citeturn4search1

A7A5 expanded rapidly after its January 2025 launch. Elliptic calculated more than $100 billion in aggregate on-chain transfers by January 2026 and $17.3 billion in exchange volume, mostly in A7A5/RUB and A7A5/USDT pairs. Those numbers cannot be interpreted as equivalent to unique economic payments: TRM found that approximately one-third of a measured A7A5 volume subset reflected circular transfers in a narrow cluster, consistent with book balancing. citeturn2view5turn2view2

By mid-2026, sanctions and infrastructure failures had substantially impaired A7A5. Elliptic reported a roughly 96% decline in activity, cessation of issuance, loss of the primary bank-card on-ramp, and failure of Grinex as the token’s principal liquidity venue. This creates another natural experiment: did displaced activity vanish, migrate to other stablecoins, move to more opaque OTC settlement, or substitute toward BTC? citeturn2view3

The broader A7 organization nevertheless remained active. Reuters reported in June 2026 that A7 had become a leading Russian cross-border payment company and intended to expand internationally, although it did not publicly disclose its operating methods. citeturn3search3

Regulatory structure as of August 2026

The Bank of Russia’s July 2026 framework establishes a sharp functional separation. Domestic payments in cryptocurrency remain prohibited. Non-qualified investors may purchase selected liquid cryptocurrencies only after testing and up to ₽300,000 annually through each intermediary. Exporters and importers, by contrast, may use any cryptocurrency and wallet for cross-border payments without volume limitations, directly or through intermediaries. A transition period for market licensing runs through 1 July 2027. citeturn2view6

Mining is also becoming more observable. The Federal Tax Service operates registers for industrial miners and mining-infrastructure operators. Registration materials include electricity-supply documentation and delivery points; registered entities report mining activity, and the registry itself is not public. Individuals below a stated household-consumption threshold may mine without industrial registration. citeturn7search0turn7search3

This does not prove that authorities direct mining output into foreign-trade settlement. It does mean that the state can possess information that outside researchers lack: operator identity, power-delivery points, production declarations, infrastructure-provider relationships, and tax records. That informational asymmetry is a central unstated condition of the stronger state-coordination hypothesis. citeturn7search3turn7search4

The Bank of Russia also already conducts broad crypto-market monitoring. Its 2024 Financial Stability Review estimated ₽4.8 trillion in cryptoasset flows attributable to Russian investors during the reviewed period and estimated that Bitcoin represented 69% of Russian users’ exchange-wallet balances. Its 2025 review estimated ₽933 billion of average monthly exchange balances attributable to Russians, of which 62% was Bitcoin; the central bank explicitly cautions that these figures are model-based estimates using exchange flows, web traffic, and bounce rates rather than direct account-level observation. citeturn10view0turn10view1

Institutional timeline

timeline
    title Russia-linked crypto settlement and enforcement timeline
    2024-07 : Russian lawmakers approve cross-border cryptocurrency settlement framework
    2024-09 : Experimental legal regime begins
    2024-12 : Finance minister confirms use of Russian-mined BTC in foreign trade
    2025-01 : A7A5 launches
    2025-03 : Garantex infrastructure disrupted and assets frozen
    2025-03 : Reuters reports BTC, ETH and USDT in some oil-trade settlement chains
    2025-07 : Russian financial-monitoring chief confirms netting, gold and crypto mechanisms
    2025-08 : OFAC sanctions Grinex, A7, Old Vector and related entities
    2025-08 : UK sanctions Grinex, Old Vector and associated infrastructure
    2025-10 : EU introduces transaction ban targeting A7A5 and linked providers
    2026-01 : Elliptic reports A7A5 exceeding $100B aggregate on-chain transfers
    2026-04 : A7A5 bank-card on-ramp fails; Grinex reports major breach
    2026-06 : A7 announces continued operation and intended global expansion
    2026-07 : Bank of Russia announces comprehensive two-contour regulatory framework
    2026-09 : Scheduled start of new retail and cross-border rules
    2027-07 : End of licensing transition period

The dates and institutional events above are documented by the Bank of Russia, Reuters, the U.S. Treasury, the Department of Justice, Elliptic, the UK government, and the Council of the European Union. citeturn3search8turn3search0turn4search1turn2view0turn4search2turn4search3turn2view3turn3search3turn2view6

Causal architecture, price channels, and unstated assumptions

The complete settlement architecture

A sanctions-constrained settlement system need not transfer cryptocurrency directly from the Russian exporter to the foreign buyer. It can function as a multilayer network composed of domestic ruble collection, internal ledgers, foreign treasury pools, stablecoin conversion, BTC inventory, OTC brokers, exchanges, and local payout agents. TRM’s investigation explicitly identifies a domestic-liquidity function and an internal-accounting function in A7’s crypto operations. citeturn2view2

flowchart LR
    RI[Russian importer] -->|RUB liability| DA[Domestic settlement agent]
    RE[Russian exporter] -->|Foreign receivable| DA
    RM[Registered or private miner] -->|Newly mined BTC| DA

    DA --> IL[Internal ledger or netting layer]
    IL -->|A7A5-like book balancing| IL
    IL --> TP[Foreign treasury pools]

    TP -->|USDT / BTC / other asset| OTC[OTC broker or liquidity hub]
    OTC --> FS[Foreign supplier]
    OTC --> FX[Foreign fiat payout]
    FS -->|Hold| H[BTC inventory]
    FS -->|Sell spot| EX[Global exchange]
    FS -->|Short futures/perpetuals| DER[Derivative venue]

    EX --> P[BTC spot price]
    DER --> P
    H --> P

    DA -->|Residual RUB demand| RUB[RUB shadow and official rates]
    TP -->|CNY / AED / USD demand| RUB

The architecture has four economically distinct balances:

[ \text{trade balance} \quad \text{crypto inventory balance} \quad \text{fiat liquidity balance} \quad \text{derivative hedge balance} ]

A gross settlement figure becomes relevant to BTC only after internal cancellation:

[ Q_t^{BTC,res}

Q_t^{BTC,gross}

• Q_t^{BTC,netted}
• \Delta H_t^{dealer}
• Q_t^{derivative\ equivalent} ]

Import settlement using market-purchased BTC

If an agent purchases BTC and sends it to a supplier who sells it, the open market experiences two metaorders:

[ \text{buy}{t_0} \rightarrow \text{transfer} \rightarrow \text{sell}{t_1} ]

The net long-run quantity effect may approach zero, but the two legs need not cancel in price terms because they can occur at different times, venues, participation rates, volatilities, and liquidity states. Crypto price differences may persist inside no-arbitrage bands created by fees, collateral, funding, transfer delays, capital controls, and user-access restrictions. citeturn6view3

Likely signal:

[ \text{positive impulse} \rightarrow \text{temporary basis} \rightarrow \text{reversal} ]

Import settlement using newly mined BTC

When the BTC originates in mining, no open-market purchase is needed:

[ \text{electricity}

• \text{hardware}
• \text{RUB operating costs} \rightarrow BTC_{\text{new}} \rightarrow \text{foreign payment} ]

If the recipient sells:

[ Q_t^{market}<0 ]

If the recipient pre-hedges:

[ Q_{t-k}^{perp}<0 \quad\text{before}\quad Q_t^{spot}<0 ]

This is the strongest candidate for structurally asymmetric BTC supply. The mechanism is supported by the confirmed use of Russian-mined BTC in foreign trade, but there is no public estimate of its scale, recipient liquidation rate, or hedge policy. citeturn3search0

Export settlement and internal recycling

A foreign buyer may acquire BTC or USDT to pay a Russian exporter:

[ \text{foreign fiat} \rightarrow BTC/USDT \rightarrow \text{Russian settlement network} ]

The Russian side can then:

[ \begin{cases} \text{sell for RUB}\ \text{hold inventory}\ \text{finance an importer}\ \text{post collateral}\ \text{swap into another asset} \end{cases} ]

When exporter receipts finance importer obligations, gross crypto volume can rise without corresponding exchange flow. This is consistent with Russian officials’ description of netting and with TRM’s finding that circular A7A5 transfers function as internal book settlement. citeturn3search2turn2view2

Price-impact function

A classic Bitcoin metaorder study reconstructed more than one million orders and found that market impact approximately followed a square-root relationship over multiple scales, with significant post-execution decay for mechanically driven flow. The dataset is historically old and should not be used to calibrate current impact without re-estimation, but the functional form remains a defensible starting prior. citeturn5search4

[ \widehat I*{t,h}

Y*{t,h}\sigma*{t,h} \operatorname{sgn}(Q_t^{BTC,res}) \left( \frac{|Q_t^{BTC,res}|}{V*{t,h}} \right)^\delta ]

with:

[ \delta \approx \frac{1}{2} ]

The interaction with liquidity is essential:

[ \frac{|Qt^{BTC,res}|}{V{t,h}} ]

A flow too small to matter during a liquid U.S. session can become visible during weekends, holiday hours, exchange outages, collateral stress, or thin order books.

Unstated assumptions

The hypothesis usually smuggles in assumptions that must be made explicit.

Assumption	Why required	Consequence if false
A material share of Russian trade uses BTC rather than only stablecoins or fiat netting	Establishes sufficient BTC exposure	Signal becomes too sparse
BTC is obtained from identifiable Russian-linked mining or settlement inventory	Enables attribution	Flow cannot be separated from global miner activity
Foreign recipients sell or hedge a meaningful fraction	Generates market pressure	Settlement becomes neutral or bullish
Internal netting is incomplete	Leaves external residual	Gross volume has little price effect
Transactions reach price-forming venues	Converts settlement into public order flow	Effects remain OTC and invisible
Labels are available before the trade signal	Prevents look-ahead bias	Backtest is invalid
Russia-linked flow is not perfectly anticipated	Allows predictive edge	Price moves before observable data
Venue latency exceeds execution latency and costs	Makes lead–lag tradable	Spread closes before entry
USDT/RUB and BTC/RUB quotes are executable	Makes basis economically meaningful	Basis is merely advertised noise
ETH or another hedge removes common crypto beta	Isolates BTC-specific response	Residual retains broad-market exposure
Regulatory changes alter concentration or routing	Creates natural experiment	September 2026 is only a legal date
Sanctions events are exogenous to immediate BTC returns	Supports event identification	Event study is confounded
Exchange-attribution data are accurate enough	Identifies liquidation endpoints	Young-coin signal becomes false attribution
Derivative exposure can be approximated from public data	Captures pre-hedging	Spot-only model misreads direction
Market impact is sufficiently persistent	Allows execution after detection	Signal is academically real but untradeable

Hypothesis registry and verification routes

The following claims range from strongly plausible to currently unprovable. None should be presented as fact without the specified evidence.

Comparative hypothesis table

Hypothesis	Present status	Expected BTC signature	Strongest verification route	Primary falsifier
Freshly mined Russian BTC finances imports and is sold by suppliers	Plausible; use of mined BTC is confirmed, liquidation is not	Young UTXOs followed by exchange inflow and negative BTC residual	Mining-origin graph, supplier or OTC labels, exchange deposit timing	No exchange-bound path; recipients retain or recycle BTC
Settlement agents buy BTC before payments and recipients sell later	Plausible	Buy impulse followed by reversal	Coupled order-flow and on-chain event study	No paired buy–sell structure
Suppliers pre-hedge with perpetual futures	Plausible	Perp weakness and OI change before on-chain transfer	Pre-event derivatives local projections	Price impact begins only after spot deposit
A7A5-like tokens are internal ledgers while BTC is an external bearer bridge	Plausible but unproven	Stablecoin imbalance precedes BTC bridge transfer	Temporal graph sequencing across token and BTC clusters	No cross-chain sequencing
Stablecoin enforcement causes substitution into BTC	Strongly plausible in theory; unproven in this network	Rising BTC share after freeze or seizure events	Event study around issuer freezes and exchange disruptions	Migration occurs only into other stablecoins or cash
Mining registration consolidates output into fewer settlement-compatible operators	Plausible secondary effect	Higher payout-address concentration and larger transfers	HHI analysis around registry and regional restrictions	Mining exits Russia or remains fragmented
Retail limits reserve liquidity for wholesale settlement	Speculative policy interpretation	Fewer small trades, larger weekday OTC flow, tighter institutional spreads	Difference-in-differences after September 2026	Retail migrates offshore; wholesale metrics unchanged
State-linked banks contract directly for mined BTC	Currently unproven	Stable miner-to-agent paths and periodic volume ratios	Contracts, leaks, litigation, financing records, repeated on-chain paths	No persistent linkage
Official USD/RUB and BTC react to the same hidden trade imbalance	Plausible	Crypto-implied RUB rates lead official adjustments	Local projections controlling for CNY, oil and DXY	No incremental predictive information
Settlement activity follows invoice, shipping, tax, or mining calendars	Plausible	Weekday, month-end or quarter-end clustering	Calendar and Hawkes-process tests	No stable seasonal structure
“Clean” young BTC receives an OTC compliance premium	Possible but weakly evidenced	Different destinations or private quote premium by coin age	Private OTC quotes and provenance-controlled regression	No premium after risk and size controls
A7A5’s 2026 impairment displaced activity into BTC	Open empirical question	BTC-linked flow rises after April 2026	Substitution-ratio event study	Total activity collapses or moves only to USDT alternatives

Route to test mined-BTC import settlement

The verification chain should progress through increasingly demanding stages:

[ \text{coinbase output} \rightarrow \text{pool payout} \rightarrow \text{probable Russian miner} \rightarrow \text{settlement intermediary} \rightarrow \text{foreign counterparty} \rightarrow \text{exchange deposit} ]

A credible result requires more than “young coins moved.” Mining pools aggregate miners from many countries; coinbase ancestry alone is not geographic evidence. The Russian attribution score must combine time-valid labels, public ownership information, pool payout structures, energy-facility evidence, sanctioned-address links, transaction patterns, and any lawful documentary data.

The hypothesis gains support if the following sequence repeats:

1. newly mined BTC enters a high-confidence Russian-linked cluster;
2. the cluster transfers consistent commercial-size amounts during business-oriented windows;
3. receiving addresses connect to foreign treasury or OTC infrastructure;
4. coins enter a price-forming exchange;
5. signed spot flow and BTC-specific returns become negative;
6. the effect scales with volume divided by contemporaneous market liquidity.

It is weakened if the coins remain in custody, move only among self-controlled wallets, are transferred to long-term holders, or have no relationship to exchange-bound flows.

Route to test stablecoin-to-BTC substitution

A7A5 offers a rare quasi-experiment because its access to global liquidity was progressively constrained by U.S., UK, and EU sanctions, and then by the failure of its principal bank-card and exchange gateways. Elliptic’s reported 2026 decline provides a measurable treatment window. citeturn2view0turn4search2turn4search3turn2view3

Define:

[ Share_t^{BTC}

\frac{F_t^{BTC}} {F_t^{BTC}+F_t^{USDT}+F_t^{A7A5}+F_t^{other}} ]

Estimate:

[ Share_t^{BTC}

\alpha

• \beta_1 PostSanction_t
• \beta_2 PostOnrampFailure_t
• \Gamma X_t
• \varepsilon_t ]

Evidence for substitution requires:

[ \beta_1>0 \quad\text{or}\quad \beta_2>0 ]

and preferably a corresponding rise in paths from known or suspected settlement clusters into BTC. A mere increase in global BTC activity is insufficient.

Route to test state coordination

The strong form of the hypothesis is:

State-linked banks or settlement operators systematically obtain mining output under private contracts or administrative coordination.

This is currently unprovable from public price data. On-chain repetition can support coordination but cannot establish state instruction.

Potential proof sources include:

• leaked contracts or internal settlement instructions;
• court filings, insolvency records, or criminal indictments;
• financing agreements between banks and mining firms;
• tax or customs disclosures;
• sanctions-designation narratives naming specific transactions;
• invoices linking mined BTC to import obligations;
• repeated address paths whose volumes match documented trade settlements.

Falsification is difficult because absence of public evidence does not prove absence of coordination. A practical standard is to downgrade the hypothesis if several years of high-quality graph data reveal no persistent miner-to-settlement concentration and if identified miners behave like ordinary global treasury operators.

Route to test regulatory liquidity reservation

The claim that retail restrictions intentionally reserve crypto liquidity for corporate settlement is a political-economy interpretation, not an official policy statement. The Bank of Russia frames the rules as investor protection, market regulation, monetary sovereignty, and provision for cross-border business payments. citeturn2view6

A testable weaker version is:

Regardless of intent, retail restrictions change the composition and temporal structure of local crypto liquidity.

Expected post-regime observations include:

[ \downarrow \text{small-ticket frequency} ]

[ \uparrow \text{median transaction size} ]

[ \uparrow \text{weekday concentration} ]

[ \uparrow \text{counterparty concentration} ]

[ \downarrow \text{retail weekend variance} ]

A difference-in-differences design should compare small versus large quotes, retail payment methods versus institutional OTC, and Russia-exposed pairs versus control pairs in countries without a simultaneous regulatory break.

Data architecture, labeling, and pipeline design

Required datasets

Layer	Minimum viable source	Institutional source	Core fields	Main limitation
Bitcoin blockchain	Self-hosted Bitcoin Core, mempool.space API	Coin Metrics ATLAS, TRM, Elliptic, Chainalysis	Inputs, outputs, UTXO age, blocks, fees, address paths	Geographic identity is not native to the chain
Mining attribution	Public pool tags and payout templates	Commercial entity labels, private pool data	Coinbase tags, payout addresses, pool shares	Pools serve miners across jurisdictions
Spot trades	Exchange public archives	Kaiko or comparable institutional feed	Timestamp, price, size, aggressor side	Venue coverage and historical corrections
Order books	Exchange WebSockets	Kaiko L2 or direct licensed feeds	Depth, spread, imbalance, replenishment	Storage-intensive; outages and spoofing
Derivatives	Public futures APIs	Institutional derivatives data	Funding, OI, basis, liquidations, trades	OI changes do not reveal trader identity
RUB crypto prices	Licensed exchange or broker quotes	Compliant OTC quote archive	Bid, ask, size, payment rail, fill status	Advertised P2P quotes are often non-executable
FX benchmarks	CBR, compliant market data	Institutional FX and NDF feeds	USD/RUB, CNY/RUB, CNH, fixing methodology	Official fixing can lag shadow markets
Sanctions events	OFAC, EU, UK official lists	Compliance vendor feeds	Entity, wallet, effective time, jurisdiction	Labels become public after historical transactions
Trade calendars	Customs aggregates, shipping and commodity data	Vessel, invoice, customs, bill-of-lading providers	Loading dates, payment windows, value	Exact settlement dates are usually private
Russian regulatory events	CBR, FTS, ministries	Legal databases	Effective date, scope, transition period	Announcement and implementation can diverge

Coin Metrics ATLAS provides standardized transaction, address, and block queries; mempool.space exposes public transaction, block, fee, mining, and mempool APIs. Binance publishes historical trade files, while Kaiko provides institutional trade and order-book coverage across many venues. These capabilities are suitable for lawful market-observation pipelines, subject to venue terms and local regulation. citeturn8search0turn8search2turn8search5turn8search7

Event-time data model

Every datum should carry both an event time and a knowledge time:

event_time       = when the transaction, trade, or legal event occurred
observed_time    = when the pipeline received it
label_known_time = when an address attribution became available
publication_time = when a report or designation was published

A historical wallet must not be treated as “known Russian-linked” before the date on which the label became available to the researcher. Failure to preserve label_known_time creates severe look-ahead bias.

Sample SQL for triangular basis

WITH btc_rub AS (
    SELECT
        date_trunc('minute', ts) AS minute,
        weighted_mid(price_bid, price_ask, executable_size) AS btc_rub_mid,
        SUM(executable_size) AS btc_rub_depth
    FROM local_quotes
    WHERE pair = 'BTC/RUB'
      AND quote_status = 'executable'
      AND sanctions_screen_passed = TRUE
    GROUP BY 1
),
usdt_rub AS (
    SELECT
        date_trunc('minute', ts) AS minute,
        weighted_mid(price_bid, price_ask, executable_size) AS usdt_rub_mid,
        SUM(executable_size) AS usdt_rub_depth
    FROM local_quotes
    WHERE pair = 'USDT/RUB'
      AND quote_status = 'executable'
      AND sanctions_screen_passed = TRUE
    GROUP BY 1
),
btc_usdt AS (
    SELECT
        date_trunc('minute', ts) AS minute,
        SUM(price * size) / NULLIF(SUM(size), 0) AS btc_usdt_vwap
    FROM global_trades
    WHERE pair = 'BTC/USDT'
      AND venue IN ('COMPLIANT_VENUE_A', 'COMPLIANT_VENUE_B')
    GROUP BY 1
)
SELECT
    b.minute,
    LN(b.btc_rub_mid)
      - LN(g.btc_usdt_vwap)
      - LN(u.usdt_rub_mid) AS raw_log_basis,
    LEAST(b.btc_rub_depth, u.usdt_rub_depth) AS local_capacity,
    cost.total_bps,
    (
      LN(b.btc_rub_mid)
      - LN(g.btc_usdt_vwap)
      - LN(u.usdt_rub_mid)
      - cost.total_bps / 10000.0
    ) AS net_log_basis
FROM btc_rub b
JOIN usdt_rub u USING (minute)
JOIN btc_usdt g USING (minute)
JOIN execution_cost_model cost USING (minute);

Sample SQL for exchange-bound young coins

WITH spent_utxos AS (
    SELECT
        tx.txid,
        tx.block_time,
        inp.prev_txid,
        inp.prev_vout,
        u.created_time,
        u.value_btc,
        EXTRACT(EPOCH FROM (tx.block_time - u.created_time)) / 86400.0
            AS coin_age_days
    FROM bitcoin_inputs inp
    JOIN bitcoin_transactions tx
      ON tx.txid = inp.txid
    JOIN bitcoin_utxos u
      ON u.txid = inp.prev_txid
     AND u.vout = inp.prev_vout
),
scored_paths AS (
    SELECT
        s.txid,
        s.block_time,
        s.value_btc,
        s.coin_age_days,
        labels.p_ru,
        labels.p_mining_origin,
        labels.p_settlement,
        destinations.p_exchange_within_24h,
        EXP(-s.coin_age_days / :age_half_life_days) AS age_weight
    FROM spent_utxos s
    JOIN probabilistic_entity_labels labels
      ON labels.txid = s.txid
     AND labels.known_at <= s.block_time
    JOIN path_destinations destinations
      ON destinations.txid = s.txid
)
SELECT
    date_trunc('hour', block_time) AS hour,
    SUM(
        value_btc
        * p_ru
        * p_mining_origin
        * p_settlement
        * p_exchange_within_24h
        * age_weight
    ) AS weighted_exchange_pressure_btc
FROM scored_paths
GROUP BY 1;

Probabilistic on-chain labeling

from dataclasses import dataclass
from math import exp
from typing import Iterable

@dataclass(frozen=True)
class Evidence:
    source: str
    log_likelihood_ratio: float
    known_at_unix: int
    expires_at_unix: int | None = None

def posterior_probability(
    prior: float,
    evidence: Iterable[Evidence],
    event_time_unix: int,
) -> float:
    """
    Bayesian-style evidence combiner.
    Only evidence known by event time is used, preventing look-ahead bias.
    Correlated evidence should be grouped or down-weighted before calling.
    """
    if not 0.0 < prior < 1.0:
        raise ValueError("prior must be between zero and one")

    prior_odds = prior / (1.0 - prior)
    log_odds = __import__("math").log(prior_odds)

    for item in evidence:
        if item.known_at_unix > event_time_unix:
            continue
        if item.expires_at_unix is not None and event_time_unix > item.expires_at_unix:
            continue
        log_odds += item.log_likelihood_ratio

    odds = exp(log_odds)
    return odds / (1.0 + odds)

def settlement_flow_score(
    btc_amount: float,
    coin_age_days: float,
    p_ru: float,
    p_mining: float,
    p_settlement: float,
    p_exchange_liquidation: float,
    direction: int,
    age_decay_days: float = 30.0,
) -> float:
    if direction not in (-1, 1):
        raise ValueError("direction must be -1 for supply or +1 for demand")

    age_weight = exp(-coin_age_days / age_decay_days)
    return (
        direction
        * btc_amount
        * p_ru
        * p_mining
        * p_settlement
        * p_exchange_liquidation
        * age_weight
    )

Label hierarchy

A defensible label system should separate identity from behavior:

Label dimension	Example evidence	Output
Geographic linkage	corporate ownership, facility records, legal filings	(p^{RU})
Mining origin	coinbase ancestry, pool payout pattern	(p^{mine})
Settlement role	repeated commercial-size routes, leaked ledger linkage	(p^{settle})
Exchange destination	known deposit address or probabilistic peel chain	(p^{exchange})
Sanctions exposure	official listing, direct or indirect graph exposure	compliance flag
Knowledge validity	publication and discovery timestamps	backtest eligibility

No single heuristic should determine identity. Time-zone activity, round amounts, and pool membership are weak evidence and should receive low weights unless corroborated.

Econometric identification, event studies, and expected signals

Identification ladder

The project should distinguish five levels of evidence.

Level	Result	What it proves	What it does not prove
Anomaly	Basis, flow, or lead–lag is statistically unusual	Market structure contains a pattern	Russian settlement caused it
Prediction	Pattern forecasts returns out of sample	Usable information exists	Structural causality
Flow linkage	Labeled path precedes market response	Specific flow is associated with response	Economic purpose of transfer
Natural experiment	Exogenous disruption changes exposed markets	Stronger causal evidence	Full state coordination
Documentary proof	Contract, invoice, ledger, or official record	Transaction purpose and counterparties	Generalizability to all flow

Residual-return construction

The dependent variable should not be naked BTC return. A BTC-specific residual can be estimated against crypto and macro factors:

[ r*{t,h}^{BTC,res}

r*{BTC,t,h}

• \beta*{ETH,t}r*{ETH,t,h}
• \beta*{NDX,t}r*{Nasdaq,t,h}
• \beta*{DXY,t}r*{DXY,t,h}
• \beta*{CNH,t}r*{CNH,t,h} ]

The hedge coefficients must be estimated using only prior data, with regime adaptation and regularization. ETH removes some common crypto risk, while Nasdaq, DXY, and offshore yuan help control broader risk, dollar-liquidity, and China-linked macro conditions.

Core local-projection model

[ r_{t,t+h}^{BTC,res}

\alpha_h

• \theta_{1,h}A_t^{RUB}
• \theta_{2,h}\widetilde Q_t^{young}
• \theta_{3,h}L_t^{venue}
• \theta_{4,h}Z_t^{sanctions}
• \theta_{5,h}U_t^{inventory}
• \Gamma_hX_t
• \varepsilon_{t,h} ]

Estimate separately for:

[ h \in { 5m,15m,1h,4h,24h,72h,7d } ]

Controls (X_t) should include volatility, spread, aggregate order-flow imbalance, funding, open interest, ETF flow where relevant, options skew, weekend status, hour of day, exchange outages, month-end, and broad sanctions news.

Newey–West errors are not enough when observations overlap heavily. Use block bootstrap or non-overlapping samples as robustness checks, and evaluate false-discovery-adjusted significance across horizons and signal variants.

Event-study templates

Unhedged supplier liquidation

[ \text{young BTC transfer} \rightarrow \text{exchange deposit} \rightarrow \text{negative signed spot flow} \rightarrow \text{negative BTC residual} ]

Event return:

[ CAR*i(\tau_1,\tau_2)

\sum*{\tau=\tau1}^{\tau_2} r{i,\tau}^{BTC,res} ]

Compare event groups by:

• coin age;
• transfer size;
• label confidence;
• exchange destination;
• market liquidity;
• Moscow, Asian, European, and U.S. trading windows;
• pre- and post-regulatory regime;
• pre- and post-A7A5 impairment.

Pre-hedged settlement

[ \text{short perp} \rightarrow \text{on-chain receipt} \rightarrow \text{spot liquidation} \rightarrow \text{short close} ]

Expected pattern:

[ \Delta OI*{t-k}>0, \quad Funding*{t-k}\downarrow, \quad r_{BTC,t-k}^{res}<0 ]

followed by possible basis normalization after spot delivery.

Dealer inventory accumulation

[ \text{mined BTC} \rightarrow \text{non-exchange settlement wallet} \rightarrow \text{reduced expected supply} ]

The likely price response is not necessarily immediate. Inventory withholding can create a positive supply surprise relative to a miner-liquidation model.

Expected triangular-basis signal

The chart below is a synthetic illustration, not an empirical claim. It shows the expected shape if mined BTC is locally liquidated below the synthetic BTC/USDT × USDT/RUB price and the discount subsequently converges.

A valid empirical chart must use executable, size-matched bid and ask prices rather than last trades or advertised P2P offers.

Expected young-coin signal

The next synthetic chart illustrates a probabilistic young-coin exchange-pressure spike and a corresponding BTC-specific residual response.

A real event study should report confidence intervals, number of independent events, label-confidence strata, liquidity normalization, and the result after excluding events whose labels became known only later.

Expected venue lead–lag

The following synthetic chart illustrates a local BTC/USDT venue moving first and a Western BTC/USD venue converging later.

BTC/USDT and BTC/USD cannot be compared without correcting for USDT/USD, carry, funding, collateral differences, and each venue’s normal equilibrium spread. Research on cryptocurrency price differences emphasizes that fees, stablecoin deviations, market access, and regulatory segmentation create non-zero no-arbitrage bands. citeturn6view3

Structural-break design

The September 2026 regime can be tested as:

[ Y_t

\alpha

• \beta_1Signal_t
• \beta_2Post_t
• \beta_3Signal_t\times Post_t
• \Gamma X_t
• \varepsilon_t ]

where:

[ Post_t = \mathbb{1} {t\geq1\ September\ 2026} ]

Interpretations:

| Result | Meaning | | ---------------------------------------------- | ------------------------------------------------------------------------ | ------------ | ---------------------------------------------------------- | | ( | \beta_3 | ) increases | Flow becomes more concentrated or more observable | | ( | \beta_3 | ) decreases | Licensed intermediaries improve internalization or netting | | Local signals vanish but offshore signals rise | Activity migrates rather than contracts | | No break | Law has little immediate operational effect or implementation is delayed | | Break appears before legal date | Participants front-run the regime; date is endogenous |

The formal implementation date should not be assumed to equal the actual behavioral treatment date. Use change-point detection and legal implementation milestones alongside the statutory dummy.

Placebo and falsification suite

A credible study should be designed to fail.

Test	Implementation	Failure interpretation
Random-time placebo	Shuffle event dates within matching hour and weekday strata	Signal may be ordinary seasonality
Asset placebo	Apply to ETH, LTC, and unrelated assets	Effect is broad crypto beta, not BTC-specific
Age-matched UTXO placebo	Replace candidate coins with random UTXOs of similar size and age	“Young coin” variable is non-specific
Geography placebo	Compare with non-Russian P2P markets under similar inflation or controls	Signal reflects generic local friction
Label-time placebo	Remove labels unavailable at the historical event time	Original result used future information
Quote-quality placebo	Compare advertised quotes with filled or executable quotes	Basis is quotation artefact
Sanctions-event placebo	Use false event dates and unrelated designations	Event response is generic news volatility
Cost stress	Multiply all estimated costs by 1.5, 2, and 3	Edge is operationally fragile
Venue-outage exclusion	Remove exchange outages and depegs	Result is infrastructure noise
Alternative hedge	Replace ETH hedge with multivariate factor hedge	Result depends on arbitrary benchmark
Reverse causality	Model settlement timing as a function of prior BTC return	Agents may time transactions after price moves
Multiple-testing control	Apply family-wise or false-discovery adjustments	Apparent significance may be data mining

The strongest falsifier is conditional independence:

[ r_{BTC,t+h}^{res} \perp A_t^{RUB},\widetilde Q_t,L_t \mid \text{global factors, USDT stress, liquidity and prior returns} ]

If the Russia-linked variables add no incremental out-of-sample information after these controls, the tradable version of the hypothesis should be rejected even if the settlement network itself is real.

Trading formulas, execution constraints, and risk

The formulas below are research constructs. They are not invitations to trade with sanctioned entities or transfer assets through prohibited routes. Their intended use is to extract information from public, licensed, or vendor-provided data and execute only on legally accessible, compliant venues.

Triangular BTC/RUB residual basis

Synthetic ruble BTC price:

[ P*{BTC/RUB,t}^{syn}

P*{BTC/USDT,t}^{global} P_{USDT/RUB,t}^{local} ]

Raw basis:

[ A*t^{RUB}

\ln P*{BTC/RUB,t}^{local}

• \ln P_{BTC/USDT,t}^{global}
• \ln P_{USDT/RUB,t}^{local} ]

Executable net basis:

[ A_t^{net}

A_t^{RUB}

• C_t^{local}
• C_t^{global}
• H_t^{legal}
• H_t^{counterparty} ]

where (H) denotes non-linear haircuts for operational, legal, banking, and counterparty risks.

Research interpretation

[ A_t^{net}>0 ]

can indicate local BTC scarcity or BTC-specific acquisition demand.

[ A_t^{net}<0 ]

can indicate miner liquidation, dealer inventory reduction, or urgent conversion from BTC into RUB or USDT.

Predictive formulation

[ \widehat r_{t,h}^{BTC,res}

\alpha_h

• \beta_h z(A_t^{net})
• \Gamma_hX_t ]

Implementation notes

Use side-specific prices:

[ A*{buy}^{net}

\ln Ask*{BTC/RUB}

• \ln Bid_{BTC/USDT}
• \ln Bid_{USDT/RUB}
• C ]

[ A*{sell}^{net}

\ln Bid*{BTC/RUB}

• \ln Ask_{BTC/USDT}
• \ln Ask_{USDT/RUB}
• C ]

Match the quote size across all legs, reject stale observations, and model fill probability. An apparent 2% basis based on a tiny local offer cannot support a large strategy.

Costs and risks

• bid–ask spread on three legs;
• local payment-rail premium;
• transfer and withdrawal restrictions;
• quote cancellation;
• bank-account interruption;
• stablecoin depeg;
• inability to short or transfer;
• sanctions screening;
• exchange custody risk;
• legal prohibition on particular counterparties or routes.

The safest implementation is informational: use the local residual basis as a predictor for a hedged position on compliant global venues rather than attempting physical cross-border convergence.

Young-coin settlement-impact residual

Probabilistic flow:

[ \widetilde Q_t

\sum_j q_j p_j^{RU} p_j^{mine} p_j^{settle} p_j^{liq} s_j e^{-age_j/\tau} ]

Expected impact:

[ \widehat I*{t,h}

Y*{t,h} \sigma*{t,h} \operatorname{sgn}(\widetilde Q_t) \sqrt{ \frac{|\widetilde Q_t|} {V*{t,h}} } ]

Observed BTC-specific move:

[ R*{t}^{res}

r*{BTC,t}

• \widehat\betat r{ETH,t} ]

Edge:

[ E*t^{young}

\widehat I*{t,h}

• R_t^{res}
• C_t
• \Lambda_t^{label} ]

where (\Lambda_t^{label}) is an explicit model-risk penalty increasing as attribution confidence falls.

Position

[ w*{BTC,t}

clip \left( \frac{E_t^{young}} {\widehat\sigma*{t,h}^{2}}, -w*{max},w*{max} \right) ]

[ w*{ETH,t}

-\widehat\beta_t w*{BTC,t} ]

Implementation notes

Do not trade at first on-chain appearance if historical tests show that recipients commonly pre-hedge. Build separate models for:

• transfer announcement or mempool detection;
• confirmation;
• exchange deposit;
• first abnormal derivative activity;
• first abnormal signed spot flow.

The correct entry event may precede or follow blockchain confirmation.

Costs and risks

• address-label error;
• exchange-deposit false positives;
• coin-age contamination from wallet consolidation;
• hidden derivatives;
• slow confirmation;
• mempool replacement;
• global news overlap;
• hedge-basis risk;
• model decay after public discovery;
• inability to reconstruct the true metaorder.

Cross-venue lead–lag convergence

Define a corrected venue spread:

[ L*t

\ln P*{BTC/USDT,t}^{A}

• \ln P_{BTC/USD,t}^{B}
• \ln P_{USDT/USD,t}
• \overline L_{t,W}
• Carry_t ]

where (\overline L_{t,W}) is the rolling equilibrium spread rather than zero.

Russia-linked activation filter:

[ G_t

w_1z(A_t^{RUB})

• w_2z(\widetilde Q_t)
• w3z(\Delta Premium{USDT/RUB,t})
• w_4z(Z_t^{sanctions}) ]

Trade only if:

[ |Gt|>g \quad\land\quad |L_t|>\ell \quad\land\quad Depth_t>D{min} ]

Position

If (L_t>0):

[ \text{short BTC exposure on venue A}

• \text{long BTC exposure on venue B} ]

If (L_t<0), reverse the pair.

Expected profit:

[ \Pi*{t,t+h}

-\operatorname{sgn}(L_t) (L*{t+h}-L_t)

• Fees
• Funding
• Borrow
• Slippage ]

Implementation notes

Both legs should be prefunded. Transferring collateral after the signal will usually destroy the edge. Execution should use synchronized clocks, depth-aware order sizing, kill switches, and venue-specific risk limits.

Correct explicitly for:

• USDT/USD deviations;
• futures basis;
• funding accrual;
• collateral currency;
• mark-price construction;
• exchange-specific index composition;
• normal time-of-day spread.

Costs and risks

• convergence failure;
• exchange default;
• leg risk;
• mark-price divergence;
• forced deleveraging;
• funding spikes;
• withdrawal suspension;
• false activation by unrelated USDT stress;
• latency asymmetry;
• regulatory access differences.

Stablecoin-enforcement substitution factor

Settlement-asset share:

[ S_t^{BTC}

\frac{F_t^{BTC}} {F_t^{BTC}+F_t^{USDT}+F_t^{A7A5}+F_t^{other}} ]

Enforcement intensity:

[ Z*t^{enf}

\sum_k \omega_k Shock*{k,t} e^{-(t-t_k)/\lambda} ]

Expected substitution:

[ \widehat{\Delta S_t^{BTC}}

\alpha

• \beta Z_t^{enf}
• \Gamma X_t ]

Directional BTC demand must then be separated by source:

[ Q_t^{substitution}

Q_t^{market\ purchased}

• Q_t^{inventory\ deployed}
• Q_t^{recipient\ liquidated} ]

An increase in BTC settlement share is not automatically bullish. If the system substitutes into existing mined inventory and recipients liquidate, BTC share can rise while price pressure is negative.

Trading implementation

Use the substitution estimate only as a regime prior:

[ P(Regime_t=BTC\ fallback) ]

Then condition the basis, young-coin, and venue models on that probability. Do not trade a sanctions headline directly without flow confirmation.

Miner and settlement-inventory surprise

Accounting model:

[ M_t

E_t

• F_t
• \Delta H_t ]

where:

• (M_t): estimated Russia-linked mining output;
• (E_t): direct exchange deposits;
• (F_t): settlement transfers;
• (\Delta H_t): inferred inventory change.

Open-market supply:

[ S_t^{open}

E_t

• \gamma_tF_t
• \delta_tD_t ]

where (D_t) is spot-equivalent derivative hedging.

Supply surprise:

[ U_t

S_t^{open}

• E[ S_t^{open} \mid hashrate, fees, price, energy\ constraints, seasonality ] ]

Signal:

[ Signal_t^{inventory}

-\frac{U_t} {\sqrt{V_t}\sigma_t} ]

Implementation notes

The model should be estimated hierarchically because (\gamma_t), the recipient-sale ratio, is unobserved. Use posterior distributions rather than point estimates:

[ \gamma_t \sim Beta(a_t,b_t) ]

Update the posterior as labeled transfers reach or avoid exchanges.

Risks

The accounting identity can fail empirically because the modeled cluster set is incomplete. Missing wallets appear as false inventory retention; mislabeled exchanges appear as false liquidation.

Composite signal and abstention

A production strategy should combine signals only after each survives independent out-of-sample testing:

[ F_t= \begin{bmatrix} A_t^{net}\ \widetilde Q_t^{young}\ L_t^{venue}\ S_t^{BTC}\ U_t^{inventory}\ Liquidity_t\ LabelConfidence_t \end{bmatrix} ]

[ \widehat r_{t,h}^{BTC,res}

\Theta_h^\top F_t ]

Risk-adjusted position:

[ w*t^{raw}

\frac{ \widehat r*{t,h}^{BTC,res} }{ \widehat\sigma_{t,h}^{2} } ]

Epistemic haircut:

[ w*t^{final}

clip(w_t^{raw},-w*{max},w_{max}) \left( 1-U_t^{epistemic} \right) ]

Trade condition:

[ | \widehat r_{t,h}^{BTC,res} |

C_t^{all-in}

• Buffer_t^{model} ]

The abstention state is not optional. Most observations should produce no trade.

Compliance boundaries and prioritized research roadmap

Legal and ethical constraints

The project should be designed as market research and sanctions-risk observation, not as participation in evasion infrastructure.

OFAC’s virtual-currency guidance states that sanctions obligations apply equally to virtual and fiat currencies and that industry participants must avoid direct and indirect dealings prohibited by sanctions. Its Russia-related guidance warns that U.S. persons and service providers generally may not transact with or facilitate transactions in which blocked persons have an interest, and non-U.S. persons may also face exposure when causing U.S. sanctions violations or engaging in evasive conduct. citeturn4search0turn4search10

The UK sanctioned Grinex, Old Vector, and related entities in August 2025. The EU’s nineteenth sanctions package introduced measures against A7A5 and linked crypto providers, and the U.S. Treasury designated A7, Grinex, Old Vector, and associated entities. Exact obligations vary by citizenship, residence, corporate nexus, venue jurisdiction, and the location of infrastructure and personnel. citeturn4search2turn4search3turn2view0

Operational rules should include:

• no account, wallet, or payment relationship with designated entities;
• no transfer designed to test or exploit sanctioned liquidity routes;
• no provision of execution, routing, or optimization services to sanctioned actors;
• no evasion of exchange geofencing, KYC, AML, or reporting rules;
• legal review before storing or processing sensitive vendor labels;
• screening of counterparties and wallet exposure before any transaction;
• retention of evidence showing that sanctioned-network data were used only for observation;
• separate compliance approval for each execution venue and legal entity.

Public blockchain analysis can also create false allegations. Probabilistic labels should not be presented as definitive attribution, especially when they concern individuals or companies not officially designated. Research publication should report confidence levels, correction procedures, and vendor-label limitations.

Prioritized roadmap

Milestone	Deliverable	Required datasets	Estimated effort	Go/no-go criterion
Basis observatory	Clean BTC/RUB, USDT/RUB, BTC/USDT residual basis	Executable local quotes, global spot, fees	2–4 engineer-weeks	Stable, reproducible basis after quote-quality filters
Market microstructure store	Normalized trades, L2 books, funding, OI	Public archives plus institutional feed	4–8 engineer-weeks	Clock synchronization and realistic fill simulator
Event and sanctions registry	Machine-readable official timeline	OFAC, EU, UK, CBR, DOJ, Reuters metadata	1–3 researcher-weeks	Event timestamps and jurisdiction tags audited
Bitcoin graph foundation	UTXO age, pool attribution, exchange paths	Bitcoin node, ATLAS or equivalent	6–12 engineer-weeks	Reproducible coinbase-to-destination graph
Probabilistic label layer	Versioned labels with knowledge time	Commercial analytics, OSINT, sanctions data	8–16 mixed weeks	No future-label leakage; confidence calibration
Baseline econometrics	Local projections and placebo suite	Basis, market, event, graph data	4–8 quant-weeks	Out-of-sample incremental predictive value
A7A5 substitution study	Pre/post disruption asset-share analysis	Ethereum/TRON data, A7A5 labels, BTC graph	4–10 researcher-weeks	Measurable shift beyond general market controls
September regime study	Regulatory structural-break report	Local quote and activity data before/after Sep. 2026	3–6 months elapsed data	Detectable change in size, timing, or concentration
Paper-trading engine	Cost-aware hedged strategy simulation	All prior layers	4–8 engineer/quant-weeks	Positive net expectation under doubled costs
Independent replication	Separate implementation and label audit	Frozen research dataset	3–6 reviewer-weeks	Results reproduce without original analyst discretion
Limited live pilot	Small, compliant venue-neutral book	Screened execution venues	2–3 months	Stable slippage, no compliance exceptions, bounded drawdown

Research phases

Foundation

Build the basis monitor and global market store first. These provide useful information even if the Russia-specific causal hypothesis fails.

Primary deliverables:

[ A_t^{RUB}, \quad USDT/RUB\ premium, \quad BTC/ETH\ residual, \quad venue\ spreads ]

Do not begin with wallet attribution. The risk of expensive forensic work is unjustified until the simpler market variables show stable anomalies.

Identification

Add event-time sanctions data and run pre-registered event studies around:

• the Garantex disruption;
• OFAC designation of A7 and Grinex;
• UK designations;
• EU A7A5 measures;
• A7A5 bank-card on-ramp failure;
• Grinex failure;
• September 2026 regulation;
• major mining restrictions.

The pre-registration should specify windows, controls, exclusions, and primary outcomes before looking at results.

On-chain enrichment

Only after the market-level signal survives should the project acquire expensive entity labels and build the young-coin graph.

The primary output is not a binary list of “Russian addresses.” It is:

[ p^{RU} \times p^{mine} \times p^{settle} \times p^{liquidation} ]

with full version history.

Structural verification

The strongest verification effort should seek lawful documentary evidence: commercial disputes, public tenders, financing records, leaked datasets whose provenance and legality have been reviewed, court cases, corporate disclosures, and future sanctions narratives.

Without documentary evidence, the report should continue to say “Russia-linked probabilistic settlement flow,” not “state-directed Bitcoin sale.”

Deployment

A live strategy should be launched only if:

[ IR_{OOS}>0 ]

after:

• doubled estimated costs;
• label-time restrictions;
• sanctions screening;
• venue-outage exclusions;
• multiple-testing correction;
• hedging error;
• execution latency;
• realistic capacity limits.

Final deliverables

Deliverable	Publication value	Trading value	Causal value
BTC/RUB triangular-basis dashboard	High	Medium	Low
Sanctions and regulatory event database	High	Medium	Medium
A7/A7A5 network chronology	High	Low	Medium
Young-coin probabilistic-flow index	High	Potentially high	High if labels are credible
Venue lead–lag model	Medium	High	Low to medium
Stablecoin-to-BTC substitution index	High	Medium	Medium
September 2026 structural-break study	High	Medium	Medium to high
Documentary miner-to-settlement case study	Very high	Medium	Very high
Cost-aware paper-trading report	Medium	Very high	Low
Independent falsification report	Very high	High	High

Bottom line

The broad hypothesis survives deep scrutiny in a qualified form:

Russia has built and continues to develop a large alternative cross-border settlement architecture that uses netting, state-linked institutions, stablecoins, public blockchains, OTC liquidity, and, in at least some cases, domestically mined Bitcoin.

The evidence does not currently justify the stronger claim that this architecture creates a persistent, dominant, or centrally scheduled BTC price factor.

The most plausible BTC-specific mechanisms are:

[ \text{newly mined BTC paid to suppliers and liquidated} ]

[ \text{market-purchased BTC producing buy–sell settlement cycles} ]

[ \text{pre-hedging in perpetual futures} ]

[ \text{temporary substitution from sanctionable stablecoins into BTC} ]

[ \text{dealer inventory imbalances after incomplete internal netting} ]

The most important negative evidence is equally clear: public forensic work describes A7A5 as an internal accounting and ruble-to-USDT bridging system, while USDT appears to be the principal liquid cryptoasset in Russian sanctions-related settlement. A7A5’s enormous gross on-chain volume includes circular book-balancing flows and therefore cannot be mapped directly to market demand. citeturn2view2turn2view5

The central empirical object is therefore:

[ \boxed{ Q_t^{BTC,residual}

\text{unnetted}

• \text{unwarehoused}
• \text{unhedged}
• \text{exchange-reaching BTC flow} } ]

The first-order research stack should be:

[ \boxed{ BTC/RUB\ residual\ basis } ]

[ \boxed{ young\ coin\ exchange\ pressure } ]

[ \boxed{ corrected\ BTC/USDT\ venue\ lead\text{–}lag } ]

[ \boxed{ stablecoin\ enforcement\ substitution } ]

A positive result would show that these variables forecast BTC-specific returns out of sample, after realistic costs and global-factor controls, and that the result strengthens around independently identified Russia-linked settlement events.

A negative result would be equally informative:

Russia’s alternative settlement system may be large and strategically important while remaining largely irrelevant to tradable BTC price discovery because stablecoins, OTC inventories, internal netting, and derivatives absorb the flow before it reaches public order books.

That is the correct standard. The political narrative is not the signal. The signal, if it exists, is the measurable transformation between mining output, internal liabilities, dealer inventory, recipient hedging, and the global Bitcoin order book.