Detect vote buying in bribe markets
Bribe markets (Votium, Hidden Hand, Paladin) pay delegates to vote a certain way. That is not always illegitimate, but an undeclared, concentrated dependency means the vote reflects the payer, not the delegate's judgement. Detect and quantify it.
Procedure
Pull the delegate's votes for a gauge or proposal:
graphql { votes(first:100, where:{voter:"$DELEGATE"}) { proposal { id } choice } } against the DAO's subgraph, or read VoteCast logs.
Pull the same delegate address' bribe receipts from the market. On Votium the claim events are Claimed on the vlAURA wrapper; on Hidden Hand read the BribeDeposited events for the round.
Correlate: for each round, compare the delegate's vote direction with the bribe offering the largest per-gauge payout. A delegate who votes with the top briber in over 80% of rounds is effectively a paid voter.
Quantify concentration. Sum bribe value received in the last N rounds and divide by the delegate's estimated monthly operating cost; a ratio above about 1 means the income depends on bribes.
Check disclosure: search the delegate's forum thread for a bribe policy. Undisclosed paid voting is the finding, not the payment itself.
Detect self-dealing: a party that both deposits bribes and votes through a delegatee it controls is buying its own proposal.
Report per-round: bribe payer, amount, delegate vote, and whether the correlation held across rounds.
Fetch the delegate statement and check whether bribe income is disclosed; silence is the finding.
Compare the delegate's actual vote against their pre-vote stated intention; a flip that matches the top bribe is the signal.
Aggregate across delegates: if several top voters side with the same briber, treat it as coordinated paid influence.
Pitfalls
Treating every bribe as corruption; disclosed, capped bribe markets are a legitimate incentive design — flag the undisclosed and the concentrated, not the mechanism.
Correlating a single round and calling it a pattern; delegates often agree with the largest briber by shared interest, not payment.
Ignoring proxy claimers: the on-chain recipient is a claim contract, and you must walk the transfer to the delegate to attribute it.
Assuming a bribe changed the vote; unless the delegate's vote differed from their historical lean, the bribe may have bought nothing.
Missing off-chain payments (tokens sent to a treasury, OTC deals) that never touch the market contract.
Attribution error: a claim contract receives the bribe and forwards it later, so the immediate recipient is not the voter.
A delegate who abstains on bribe-heavy gauges may be avoiding the conflict; note abstention, do not treat it as clean.
Using USD bribe values without the rate and date makes the ratio to operating cost non-reproducible.
Verification
graphql { votes(where:{voter:"$DELEGATE"}) { proposal { id snapshot } choice } }
# join against the market's Claimed/BribeDeposited events per round; >80% agreement is the flag
Report the round-by-round payer, amount, and vote, the agreement rate, and whether a bribe policy is disclosed, with both datasets cited.
1---2name: detect-vote-buying-in-bribe-markets3description: Use when assessing whether DAO votes are being bought, or before relying on a delegation as independent. Correlates bribe-market deposits with delegate voting and flags concentrated paid influence.4metadata:5 source: https://skillmd.com/skills/jiayaoqijia/detect-vote-buying-in-bribe-markets6---78# Detect vote buying in bribe markets910Bribe markets (Votium, Hidden Hand, Paladin) pay delegates to vote a certain way. That is not always illegitimate, but an undeclared, concentrated dependency means the vote reflects the payer, not the delegate's judgement. Detect and quantify it.1112## Procedure13141. Pull the delegate's votes for a gauge or proposal:15 `graphql { votes(first:100, where:{voter:"$DELEGATE"}) { proposal { id } choice } }` against the DAO's subgraph, or read `VoteCast` logs.162. Pull the same delegate address' bribe receipts from the market. On Votium the claim events are `Claimed` on the vlAURA wrapper; on Hidden Hand read the `BribeDeposited` events for the round.173. Correlate: for each round, compare the delegate's vote direction with the bribe offering the largest per-gauge payout. A delegate who votes with the top briber in over 80% of rounds is effectively a paid voter.184. Quantify concentration. Sum bribe value received in the last N rounds and divide by the delegate's estimated monthly operating cost; a ratio above about 1 means the income depends on bribes.195. Check disclosure: search the delegate's forum thread for a bribe policy. Undisclosed paid voting is the finding, not the payment itself.206. Detect self-dealing: a party that both deposits bribes and votes through a delegatee it controls is buying its own proposal.217. Report per-round: bribe payer, amount, delegate vote, and whether the correlation held across rounds.22238. Fetch the delegate statement and check whether bribe income is disclosed; silence is the finding.249. Compare the delegate's actual vote against their pre-vote stated intention; a flip that matches the top bribe is the signal.2510. Aggregate across delegates: if several top voters side with the same briber, treat it as coordinated paid influence.2627## Pitfalls2829- Treating every bribe as corruption; disclosed, capped bribe markets are a legitimate incentive design — flag the undisclosed and the concentrated, not the mechanism.30- Correlating a single round and calling it a pattern; delegates often agree with the largest briber by shared interest, not payment.31- Ignoring proxy claimers: the on-chain recipient is a claim contract, and you must walk the transfer to the delegate to attribute it.32- Assuming a bribe changed the vote; unless the delegate's vote differed from their historical lean, the bribe may have bought nothing.33- Missing off-chain payments (tokens sent to a treasury, OTC deals) that never touch the market contract.3435- Attribution error: a claim contract receives the bribe and forwards it later, so the immediate recipient is not the voter.36- A delegate who abstains on bribe-heavy gauges may be avoiding the conflict; note abstention, do not treat it as clean.37- Using USD bribe values without the rate and date makes the ratio to operating cost non-reproducible.3839## Verification4041 graphql { votes(where:{voter:"$DELEGATE"}) { proposal { id snapshot } choice } }42 # join against the market's Claimed/BribeDeposited events per round; >80% agreement is the flag4344Report the round-by-round payer, amount, and vote, the agreement rate, and whether a bribe policy is disclosed, with both datasets cited.