Transaction Processing

1. Self-custody of assets
2. Access to financial infrastructure
3. Conceptually non-custodial services
4. Value management layer = common resource
5. Platform approach to commerce

Revisit: What makes blockchain-based finance different from TradFi?

{

{

Transaction Visualization

Key Insight: there is no built-in privacy!

Blockchain has no privacy

Example:

• you buy something online using a stablecoin and ship it to your home address
• the merchant can now
  • link your name and address to your account
  • knows everything you ever did with this account
  • knows everything you ever will do with this account
  • can use data analytics to map you and other contact

• intrinsic feature of public blockchains is that information gets stored by public addresses.
• logic of smart contracts is visible to all
• => firms have need of secrecy, individuals have rights to privacy

Root problem

Solutions

• large numbers of wallets/addresses
• coinjoin
• ring signatures
• zero knowledge proofs
  
   

Blockchain has no privacy

So no privacy. So why is everyone concerned about money laundering and crime?

But KYC for AML/CFT is all abound

• criminals are real and they are using blockchains 
• tax evasion, ransomware payments, ransom payments (real kidnappings!)
• for tokenized assets firms need to know
  • their shareholders for communications
  • avoid politically exposed persons

How can we have compliance that keeps criminals out?

Basic FINTRAC/FINCEN Rules for Money Services Businesses

• Basic rule: for any virtual currency transaction money services businesses need to verify and record the identity of the person involved
   
• Transactions over $10,000 need to be reported
   
• lots of smaller rules surrounding suspicious activties

\(\Rightarrow\) the rules are "tight"

Tornado cash: a privacy Solution

Banned addresses (usually by OFAC order)

August 8, 2022: OFAC sanctions Tornado Cash

The Common View

Zero Knowledge Proofs

Basic Premise

A mechanism that proves to one party (the VERIFIER) that another party (the PROVER) possesses some knowledge, without revealing the knowledge itself or any other information that can be used to re-construct it   

This is a probabilistic statement not a math proof  

Examples

WHY?

Computational

Integrity

I performed the computation

you asked me to

Toy example #1 for a zero knowledge proof

Verifier Victor

Prover Peggy

How? Toy Example 2

Toy example #2 for a zero knowledge proof

Privacy solutions

Current Compliance Solutions

Decentralized Compliance Networks

Underlying Idea

• user onboards with public address
• performs compliance checks etc
• obtains private address(es)
• does business privately

Compliance violation occurs 

• Users:

  • Choose their revokers.

  • Users have full control over their privacy.

• Revokers:

  • Initiate requests for de-anonymization based on suspicious activity.

  • Must post verifiable public requests to trigger the process, ensuring transparency.

  • Can be entities like Chainalysis, TRM Labs, or other trusted individuals decided through the public governance portal.

• Guardians:

  • Vote to approve or deny the de-anonymization request.

  • Operate with a threshold mechanism (e.g., 6 of 10 must agree to proceed).

  • Guardians cannot see the transaction details themselves. Only Revoker can see it after receiving threshold permission from Guardians.

Economic Research on Privacy

Economic Research on Privacy

"Privacy-Enhanced Payment Systems", Capponi, Lee, Zhu, 2025, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5136803

• Multidimensional approach to privacy:
  • Identity privacy – are user identities connected to account(s)?
  • Transaction privacy – is tx history auditable?
• Modeling a payment system environment:
  • Operator selects rules governing privacy-enhancements
  • → Irrevocable commitments ex-ante (e.g. smart contract)
  • Users (good and bad) create and use accounts for various reasons
  • Enforcer attempts to trace TXs back to users

Economic Research on Privacy

"Privacy-Enhanced Payment Systems", Capponi, Lee, Zhu, 2025, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5136803

• Multidimensional approach to privacy:
  • Identity privacy – are user identities connected to account(s)?
  • Transaction privacy – is tx history auditable?
• Modeling a payment system environment:
  • Operator selects rules governing privacy-enhancements
  • → Irrevocable commitments ex-ante (e.g. smart contract)
  • Users (good and bad) create and use accounts for various reasons
  • Enforcer attempts to trace TXs back to users

Model Sketch

• Two types of privacy
  • identity (on/off)
  • transaction (trace/no-trace)
• Good users
  • suffer utility loss when traced
• Bad users
  • benefit if not traced, lose when traced
• Enforcer
  • Chooses tracing probability (not possible with private identity)
• System designer:
  • chooses system to maximize welfare (utility of bad guys is negative for welfare)

Model Sketch

Result Sketch

• With identity privacy, tracing cannot reasonably occur
• tracing decision rule as a function of transaction size is complex (assuming bad guys don't "innovate")

trace

don't trace

Result Sketch when bad guys innovate

• tracing decision rule as a function of transaction size is simple

trace

don't trace

Pecking order

identity privacy

on

off