Skip to main content

> FINANCIAL_INTEGRITY_MANUAL_03

Manual 03: Sanctions Screening, PEP Identification & Asset Freezing

Real-time pre-authorization screening, multilingual phonetic transliteration matching (Cyrillic, Arabic), and statutory asset freezing protocols.

Canonical Engineering Manual #03|TinyCTO Financial Integrity

Sanctions Screening, PEP Identification & Asset Freezing

Real-time pre-authorization screening, multilingual phonetic transliteration matching (Cyrillic, Arabic), and statutory asset freezing protocols.

#3.1 Real-Time Sanctions Screening Architecture

Sanctions compliance represents a strict-liability statutory mandate. In contrast to transaction monitoring (which permits retrospective risk-based alerting), sanctions screening requires pre-transaction interception to prevent designated entities from receiving or transferring assets.

Consolidated Sanctions Registers

Institutions must continuously ingest and reconcile designations from multiple sovereign and multilateral authorities:

  • United Nations Security Council (UNSC): Consolidated List (ISIL/Al-Qaida, Taliban, 1718 DPRK Sanctions Committees).
  • US Office of Foreign Assets Control (OFAC): Specially Designated Nationals and Blocked Persons List (SDN), Sectoral Sanctions Identifications (SSI), Non-SDN Menu-Based Sanctions (NS-MBS).
  • European Union (EU): Consolidated Financial Sanctions Database (CFSP).
  • National Sanctions Authorities: UK HM Treasury / OFSI, Türkiye Resmi Gazete (Presidential Decisions under Law No. 6415/7262).

#3.2 Multilingual Phonetic & Fuzzy Matching Algorithms

Sanctions evaders deliberately modify spellings, omit middle names, or exploit alternative transliterations of non-Latin scripts (Cyrillic, Arabic, Chinese, Farsi). Standard exact string lookups fail completely against real-world payment messages.

Fuzzy Transliteration Pipeline

  1. Token Normalization: Stripping corporate suffixes ('Inc', 'GmbH', 'LLC', 'A.Ş.'), removing honorifics ('Sheikh', 'General', 'Dr.'), and converting all tokens to uppercase UTF-8.
  2. Phonetic Encoding: Multi-pass phonetic transformations using Double Metaphone and Beider-Morse Phonetic Matching (BMPM) specifically tuned for Slavic, Arabic, and Germanic surnames.
  3. String Similarity Distance: Calculating blended similarity metrics combining Jaro-Winkler (which favors prefix matches) and Levenshtein edit distance with token-order insensitivity:
Score(S1,S2)=α⋅JW(S1,S2)+(1−α)⋅LevenshteinRatio(Tokens(S1),Tokens(S2))\text{Score}(S_1, S_2) = \alpha \cdot \text{JW}(S_1, S_2) + (1 - \alpha) \cdot \text{LevenshteinRatio}(\text{Tokens}(S_1), \text{Tokens}(S_2))
  1. Match thresholds are set conservatively (≥85%\ge 85\%) for pre-authorization queues to ensure zero false suppression of sanctioned targets.

#3.3 Statutory Asset Freezing & Anti-Tipping-Off Execution

When a confirmed sanctions match occurs:

  • The transaction must be halted instantly at the payment switch level.
  • Account balances are placed under immediate judicial restraint.
  • Under strict anti-tipping-off provisions, front-line personnel must never alert the customer of the sanctions block. Customer inquiries are handled using standardized, neutral administrative scripts.
  • The MLRO submits a formal notification to the national sanctions authority within 24 hours of confirmation.