bounty-hunters-pay-for-performance-economics-and

Posted by:

admin 15 years, 11 months ago

There have been some interesting discussions about incenting judges to set bails appropriately and on the impact commercial bail bondsmen have on the 'failure to appear' rates. I first read about it in the Financial Times Undercover Economist column. The original post I found on Marginal Revolution

Here are some of the findings from the MR post:

• A whopping one-quarter of all felony defendants fail to appear at trial. Of these some thirty percent can't be found after a year.
• The police are overrun with unserved arrest warrants for failure to appear and typically devote little time to the task.
• As a result, FTA appear rates are some 28% lower for those released on commercial bail compared to those released on their own recognizance.
• When a defendant does FTA he is about 50% more likely to be caught and is caught much sooner if a bounty hunter is on his trail compared to if only the police are involved. (Both of these effects are after controlling for other relevant factors, of course).

The UE post talks about a proposed method to improve bail setting by judges:

Steven Landsburg, an economist at the University of Rochester, has suggested one way of doing that. A judge would get a fat bonus for every suspect he released, and a stiff financial penalty if the suspect machinegunned anyone. The fatter the bonus, the more the judge will want to risk releasing subjects, and the more the system will favour liberty over public safety. Whether the bonus is small or large, however, the judge has a strong financial incentive to think carefully about which suspects are worth the risk to society - and his bank account.

The Hartford (the UE), notes that in the US bail bondsmen perform that role:

Steven Landsburg, an economist at the University of Rochester, has suggested one way of doing that. A judge would get a fat bonus for every suspect he released, and a stiff financial penalty if the suspect machinegunned anyone. The fatter the bonus, the more the judge will want to risk releasing subjects, and the more the system will favour liberty over public safety. Whether the bonus is small or large, however, the judge has a strong financial incentive to think carefully about which suspects are worth the risk to society - and his bank account.

There is an added benefit to using third-party bail dealers: There are no incentive issues around bonus timing. You might find judges getting very cautious about their bonus pool around payout time. You can smooth this with pooling, but there would still be times like retirement that are hard to manage.

I was wondering if the missing piece between software publishers and bug bounty hunters, which I had on my mind, thanks to the good folks at Emergent Chaos (or EC, to stick to convention) is the bail bondsman (or apparently more likely bonsdswoman). This would require a new licensing model, I think. It's not exactly clear to me how it would work, but perhaps the software purchaser could post of reward for bugs found before they buy the license. The bail dealer would then contract with the bounty hunters to find as many bugs as possible before the contract date. Any bugs found and patched before the software is purchased should reduce the TCO of the software for the buyer. The bail dealer would need to be very good at predicting bug rates and severities and remediation cost after purchase and deployment. If the cost of the bail is less than cost of the bounties, they win. One big problem is that bug submission process would have to be handled by a third party.

Perhaps if the bail dealer took money from both the software house and the purchaser. There could be four payment scenarios: A high fee from the purchaser if there are a low number of bugs and a high fee from the software publisher if there was a high number of bugs (or a high potential cost of remediation). They get also would get a bigger fee based on the number of bugs found. See the little spreadsheet:



• Tags:
• Security and Economics,
• Information Security
Current rating: 1
• • 1
  • 2
  • 3
  • 4
  • 5
Share on Twitter Share on Facebook
• ← Can the WiKID Server and/or Token be embedded?
• on-the-short-tenure-of-cisos-and-low-frequency →