Private Settlement in Blockchain Systems 

Alfred Lehar & Motahhareh Moravvej-Hamedani

Discussion by Katya Malinova

DeGroote School of Business

McMaster University

Central Bank Workshop

December 2022, Sydney

Zhilin Wang, Qin Hu, Yawei Wang, Yinhao Xiao (2022), Transaction pricing mechanism design and assessment for blockchain, High-Confidence Computing

Private Transaction

Comments: miner incentives and behaviour

Private settlement on the Bitcoin vs Ethereum

Bitcoin blockchain (this paper):

  • 6% of transactions settle privately
  • Smaller miners
  • (MUCH) larger (active) users
  • Private users (5-6% of total):
    • at least one "private" transaction
    • but only use 1.002 miner -- once private, always private?
  • Lower (by 20%) and less volatile fees

Private settlement on the Bitcoin vs Ethereum

Ethereum

  • E.g. Capponi, Jia, Wang (2022) "The Evolution of Blockchain: From Public to Private Mempools" (theory and empirics). 
    • Private pools:
      • Higher overall welfare (more users submit transactions)
      • But higher fees
      • Higher validator welfare 
  • Weintraub, Ferreira Torres, Nita-Rotaru, State (2022) "A Flash(bot) in the Pan: Measuring Maximal Extractable Value in Private Pools"
    • Flashbots [private pool] miners (in 2022): over 99.9 % of the hashing power in the Ethereum network
    • Miners: make more than x2 what they did before joining Flashbots

What problems does the private settlement solve?

 

Bitcoin blockchain (this paper):

  • For the miners: fee volatility 
    • get a steady stream of "private" transactions
    • willing to mine for lower fees ("certainty equivalent")
  • For the users: high fees
    • attractive to fee-sensitive but time-insensitive users

Comment 1: can we get a better idea of who these users are & how much are they saving?

What problems does the private settlement solve? 

Ethereum

  • Ordering of transactions is important
    • E.g., price of a decentralized exchange (DEX) trade is determined by the funds on DEX at the time of settlement
  • No time priority
    • Miners have discretion over ordering within a block (typically by fees)
  • \(\to\) Visibility of transactions in a public mempool is a problem
    • E.g., can be "front-run" (by miners or other users)
  • \(\to\) MEV = miner (maximal) extractable value

Detour: MEV extraction example = "sandwich attack" on a DEX trade

Buy low (goes first for $$$ fee)

Sell high (this is last: $ fee)

Front-run \(\to\) pays a higher price 

\(\to\) Evil bot earns MEV

\(\to\) users are willing to pay higher fees for private settlement to avoid being "front-run" & get the "low" transaction price

Buy

Buy

Sell

Private settlement protocol on Ethereum: Flashbots protocol

  • Private mempool (not a one-to-one contract arrangement)
  • Any miner can join
    • vetting procedure, kicked out for misbehaviour 
    • No exclusivity: miners that join monitor both public and private 
  • Users: 
    • Con: higher execution/timing uncertainty (fewer miners monitor the private pool)
    • Pro: avoid being front-run
    • \(\to\) May be willing to pay higher fees
  • Miners:
    • join private \(\to\) see both \(\to\) weakly higher payoff from joining (Capponi et al)
  • Empirically: higher fees, all large miners join (Weintraub et al)
    • \(\approx\) 50% of blocks include Flashbot transactions

Doesn't apply to BTC blockchain

This can be arranged on the BTC blockchain (?)

Any general insights that apply to private settlement across different blockchains?

Comments: miner incentives and behaviour

 

  • (Toy) model in the paper:  
    • Miners that provide private settlement do not monitor public pools
    • No impact on the miner size from going public to private
    • No user decisions 
  • Can similar insights be obtained from first principles? 
    • Smaller miners face higher income (fee) volatility?
    • Accept smaller fee to lower it (certainty equivalent type argument?) 
  • Alternatively - build up the model to allow more choices?

Zhilin Wang, Qin Hu, Yawei Wang, Yinhao Xiao (2022), Transaction pricing mechanism design and assessment for blockchain, High-Confidence Computing

Private Transaction

Private Transaction

Comments: miner incentives and behaviour

To what extent do miners "specialize" on private vs public?

Time trends in who offers private?

Comments: miner incentives and behaviour 

Extent of specialization?

  • If do not have to "switch" to private, then it's "extra" income & not just risk/uncertainty reduction 
    • \(\to\) always willing to provide private settlement, no matter the fee
  • The average private transaction fee \(\approx\) std dev (11-12 satoshi)
    • Explore further how these are set?
      • ​E.g., do fees for private transactions differ across miners?
      • Across users?

Comments: who are the users?

  • Who are  the users of private transactions?
    • Conjecture: fee-sensitive, time-insensitive.
    • Conjecture: layer 1 applications
  • Can you tease out more information?
    • E.g., exchange wallets?
    • Can Cardiff University Bitcoin Database (CUBiD) help?
  • What are the incentives for the less active users?
    • The average for all private users is under 1.7 transaction
      • < 20% have 10+ transactions
    • How easy is it to find private settlement for just 1-2 transactions? Worth it?
    • Why would a miner offer a discount for a one-off transaction?
    • Back-of-the-envelope fee savings computation?

Comments: who are the users?

  • Do these settle exclusively privately or only occasionally?
    • Implications for security/single point of failure
    • "Bulk discounts"
      • agreements when the user must send everything? 
      • akin to payment for order flow
  • Do active vs occasional users pay different fees for privately settled transactions?
  • Do private users have transactions with miners?
    • Curious if they may provider services to miners (loans?) and receive lower fees in exchange for these services?

@katyamalinova

malinovk@mcmaster.ca

slides.com/kmalinova

https://sites.google.com/site/katyamalinova/