[under construction] 02b. The Forward - deterrence straight from the underwriter's office
Previous: 02. The Gist - Using economics and game theory to bust criminals of any sophistication
“No. Absolutely not. No,” Dr. Lee snaps, looking over the underwriter’s draft term sheet. “‘Underwriter may remotely shutter the fermenter?’ Really? Our CFO will never allow it.”
Senior Underwriter Marcus, who studied alongside Lee many years ago, doesn’t bat an eye. “Dana, there is a clay lens breakout in Grid-7 North that could turn into a solvent superhighway if your fermenter breaks. You could pollute 200 acres of prime farmland,” he says, leaning back, “and guess who would be on the hook for it? Us. For you. For $4 million - conservatively.”
Dr. Lee has just recieved an offer letter from Ecovault Organics, an agribusiness unicorn promising to “reinvent nitrogen fixation”. The job comes with the standard disclaimer in 99% of reputable firms in this industry that she carry her own personal liability insurance before she can so much as swipe in to the coffee room. It’s about as normal to her as driver’s insurance. Marcus, personally aware of Dr. Lee’s brilliance in hexazene polymer degradation pathways, waived the firt-time consultation fee just to get her in the door at Tri-County Liability Underwriters. But Dr. Lee is starting to get the feeling this isn’t really about winning her business.
“You’re already asking for a quarter of my paycheck,” Dr. Lee mutters. “Least you could do is be nice about it.”
A moment of silence. Marcus looks far into the distance. Then back at Dana. “Could you close the door, and pull the curtains, please?”
Under the banker’s lamp Marcus takes out a small manila folder, simply labeled “ECOVAULT - CODE SIX”. A grainy drone image of earth-moving equipment lies beside Grid-7 North, outside the permitted berm. A dated invoice shows a Foamcrete backfill with what both of them can recognize as woefully inadequate chemical resistance.
“Oh my God,” Dana mutters. If Ecovault’s site prep is that sloppy before production, what corners will be cut once the fermenter is humming at 24 bar?
“Now you understand why we’re asking for a quarter of your paycheck,” Marcus says, eyes downcast. “We got these from our tipsters’ desk in Carcosa. The CEO himself authorized prepayment to the tipster to get his hands on the originals, they’re in our vault now along with about a dozen other documents.”
Dr. Lee switched from sitting at the hardwood desk chair to lying on the old linen couch in Marcus’s office, and raked her memory over the coals. How could she have been so stupid? How could she ignore the warning signs? They wouldn’t even show her lidar maps of the area she would be serving. The azoxy radical detectors were developed in-house and were always “awaiting certification”. What competitive advantage could that possibly give? And let’s not even talk about the tri-containment sump debacle…
“If you have all this information,” Dr. Lee asked Marcus quietly, “why haven’t you come forward publicly yet?”
“We don’t have enough proof for an open and shut case,” Marcus said. “And - we’re not warmakers. Bounty hunting is hard work. We’re just trying to be smart about the risks we underwrite.”
Dr. Lee felt a bittersweetness three months later, hearing Ecovault filed for bankruptcy. The technology was genuinely novel and exciting, but due to their execution (so she reasoned) insurers wouldn’t touch any of their new hires with a ten-foot pole without aggressive monitoring policies in place, which meant new hires wouldn’t touch Ecovault with a ten-foot pole either. After the dissolution of the company, a small group from Tri-County Liability Underwriters including Marcus and the CEO were rumored to have approached the founder of the now-failed startup, and negotiated a clandestine exchange of the bona fide evidence in their vaults for much more than they paid for it.
In our previous essay we focused on who pays the fine, who grabs the bounty, and some of the more interesting interactions that fall out of that. We paid special emphasis on how whistleblower-paid bounty mechanisms can effectively detect and prosecute even organized crime, both in theory and in practice, with some examples from the SEC’s whistleblower program as well as the False Claims Act. This is already a considerable improvement on the baseline.
Less flashy but arguably even more important for our purposes is the role of the private insurer in a fine-insured bounty regime. Few individuals or start-ups can leave tens of millions idle on a balance sheet just in case they run the wrong side of a bounty hunt, so those who run a non-negligible risk often elect to buy insurance instead. Holding insurance may be required by law, as in the case of drivers being required to hold driver’s insurance in the United States; or they may be technically voluntary but required in practice to be considered a legitimate business, as in aviation insurance, medical malpractice insurance, marine insurance, cyber insurance, nuclear liability insurance, or practice builder’s risk insurance.
The instant an agent, like Dr. Lee above, signs up for an insurance policy, a professional player with real skin joins the battlefield, one that:
- Prices risks up-front: Premiums rise or fall with how dangerous your activities appear.
- Polices clients continuously: Every report, audit, or sensor feed that can lower expected losses boots profit, so underwriters actively look for red flags and warning signs well before cashing in on an accident. Indeed, “veterans” from whatever field of endeavor the bounty is aimed at would be highly sought-after assets for such firms, as in the case of Marcus above, precisely because their expertise makes the signals that much more valuable.
- Encourages innovation: A large reinsurer might cover dozens or even hundreds of firms in a given niche. A single clever detection scheme which noticeably moves the needle can be rolled out firm-wide overnight and benefit all clients.
The result is a distributed compliance network, that scales faster than any centralized regulator could hope to.
Exhibit A: The unreasonable effectiveness of aviation insurance on making plane flights safer
Many people know that travelling by plane is, bar none, the safest form of travel out there, by almost any metric you care to measure them by. On average one injury occurs every ten billion passenger miles flown, or every 400,000 round-the-Earth trips. But it wasn’t always this way. Aviation insurance underwriters routinely refuse to insure airlines unless they meet strict maintenance, training, and operations standards. Already back in 1968, G. I. Whitehead Jr.’s seminal paper The Role of the Insurance Company in Air Safety found that risk‑based premium models led insurers to clamor for technological improvements from airlines across the whole complex engineering stack. Everything from straightforward engine improvments, to “black boxes” recording flight and cockpit data, to the installation of weather radar, to the ultra-reliable autopilots 99% of your average flight runs on today is in part an invention created thanks to the constant financial pressure these underwriters put on aeronautics engineers to get their act together.
Aviation is of particular interest to us here because the worst-case scenario is truly catastrophic. The combined civil and regulatory financial exposure for a 200 to 300-passenger airplane suffering catastrophic hull loss and killing everyone on board easily exceeds $100 million. Even that sky-high number is orders of magnitude lower than where it should be, given modern estimates of the value of human life. Despite that, it still remains an extremely credible risk which is so large it can only be averted via sustained and profound efforts at deterrence. Even a 1% risk reduction is worth millions to the airline industry, and that’s for this situation alone - nevermind that such improvements tend to reduce risk across the board for all kinds of lesser accidents.
The importance of skin in the game
A bounty hunter fundamentally makes money on finding people who have done or ideally are currently doing bad things, proving they did so, and then taking their money. Conversely, a liability insurer makes money on finding people who are not currently doing and ideally never have done bad things, proving they are and have not done so, and then taking their money. They are almost opposites, save for one detail: Your insurer greatly wants to keep the relationship going.
Insurance is a heavily competitive industry no matter where you find it. There are a lot of nerds out there who love the idea of making money quasi-ex nihilo by simple virtue of being smart, objective, and consistent, and they flock to fields like this. Typical underwriting profit margins are in the single-digit percentages. Even a 0.1% increase in loss probability can dwarf this profit margin, if the expected cost of the loss is great enough. Nassim Nicholas Taleb would call this “skin in the game”, and it is notably absent from traditional criminal justice approaches - osome unscrupulous peacekeepers may even wish for more crime in their cities so they can justify larger budgets from the state.
Before any policy is granted at all, an insurance underwriter operating in a FIB regime would perform agressive vetting of their applicants. Fly-by-night operations will find it very difficult to get any insurance policy at a rate they can afford unless they have very good reasons for being that way.
But this is just the beginning. As the insuerer-client relationship continues, the insurer may well decide to press the client to help them sleep better at night by e.g. implementing live telemetry, mandatory logging, and secure-by-design infrastructure. Depending on the insurer’s percieved risk of the client, these may be required to get a policy at all, or they might simply allow the client to purchase a policy on more favorable terms - similar to how life insurance agencies will often offer you a better deal if you submit to a pre-offer checkup by a nurse.
Such proactive monitoring policies also give the insurer a way to combat adverse selection. Even if you already know that the access logs your insurer is asking for will never reveal anything troublesome, the mere act of agreeing to share them is a huge green flag in your favor. The insurer can now credibly move you out of the “low-risk, low-certainty” group to the “low-risk, higher-certainty” group, which gives you access to better rates. Done at scale, that “low-risk, low-certainty” group eventually contains mostly people who are not agreeing to these methods because they actually do have something to hide, and so their premiums will go up.
This feedback loop is one of those special little ways in which insurers quietly nudge best-practices upon the vast majority of entrants in a market - not with a hammer, but with a scalpel, bit by delicate bit.