November 05, 2016

Monetising the dispute

Discussing how we move forward on legal tech has raised the usual (dry, ho-hum) issue - programmers cost money, especially ones doing (dry, ho-hum) code in the field of (dry, ho-hum) law. So I was encouraged in a recent convo to stop asking for charity and instead show some revenue potential for those building the tech.

No matter that the law and courts is the original loss leader of democratic tradition, and also a good way to keep farmers from acting like wolves, the courts are still something that has to be paid for. Somewhere, somehow.

In terms of business models, it's a challenge to find an incentive system to put legal engineers to work, earn a return, and deliver a better result for society. Which is a shame because the new ideas of financial cryptography are showing some screaming improvements to be made.

The classical model for lawtech is to sell software to law firms and earn a support contract. But I don't think this works in the new distributed ledger world because of an assumption I have -- we are moving to a world of non-lawyer arbitrations, in which parties are bound by network contract into a loose association with expert peers to hear complaints. In the future, we'll take our disputes to our own membership, rather than the courts.

OK, so this is a clear assumption of mine, but it is the one I'm making. Now, how to make this work?

In a future world of online dispute resolution of financial contracts, the arbitrator has to be paid. Typically, the arbitrator might assess the fee for each case. I suppose this can be done on a percentage basis or a flat fee basis, but I'd suggest that doesn't concern us right now.

That fee needs then to be split - the infra has to be paid for. Let's suggest that it be broken up into these components:

a) direct fee for arbitrator
b) fee for forum, that runs the archive, the case manager, the arbitrator selection, etc.
c) software

We could also assume that the software client used by the arbitrator would automatically manage the remit back to foundation at b) and software supplier at c), because we now live in a world of integrated digital currencies. This isn't a hard thing to program up, once we've agreed on the details.

Would that be enough to make the business cycle start? I'm not sure, but we could go further.

How about we establish a rules-light approach? Now, the interesting thing about this is that it's .... light on rules! And we know what happens when there aren't enough rules - conflict! Which in this case is a good thing.

Arbitrations - or courts - are very good at two things - resolving conflicts (of course) and establishing rules for the future. These latter are called precedents, and they can be very helpful in setting the scene.

For non-lawyers, let me add some more context: An Arbitrator is empaneled on what is probably the most powerful civil setting that exists - to resolve disputes that could run into billions of dollars. The gravity of the ruling requires a scientific approach to the most unscientific of problems - the disputes of man.

Out of this scientific investigation emerges a set of facts, the application of prior code (rules & law) over those facts, and a reasoned decision based on the foregoing. In short, if the situation follows A, B, C, then Rule X, Y, Z applies would be found in legislation or a prior ruling.

For the most part, the law or precedent already exists, but on occasion, the problem is novel, so the present Arbitrator has to create the test herself. Once she has ruled, this ruling can become a precedent that binds or informs future cases. The mechanics vary, but the efficiency is clear - if the original Arbitrator got it right why not follow?

And so the rules get built out. From nothing to an edifice of self-correcting governance, a bit like the tiny programming language of Forth which is basically a tiny VM and a lot of Forth that builds up the environment bit by bit. Notice here how a single precedent becomes a rigourous base for the future, a document that many rely upon. Precedents are bricks on which trade is built; the system is capable of building its own edifice.

That's gotta be worth something. We should encourage more of this, and indeed it was the vision in CAcert that we encourage the Arbitrator to write precedents and fill out the rules. My experience with CAcert tells me that concept was worth a lot -- Since 2007 and 400 closed cases, 20 of them or 5% created precedents, which combined are probably as voluminous as the essential body of policy.

How would we turn this into not only a concept that is valuable but one that also returns value?

What if the precedent were like a Ricardian contract? I mean this in the sense of a property right, with a clause in the precedent that says what to do if this contract were to be invoked - cited or relied upon - in another case. Let's imagine that Alice the Arbitrator has declared the following in her ruling, soon to become an important precedent:

Licence to Rely. A future decision may cite and rely on this ruling by customary fee remitted to the Foundation, under the rules established by the Policy on Precedents as Property.

In this design, I could earn money by establishing really good precedents. When Alice comes along to a new case, and relies on the precedent, she cites my precedent and some of her fee gets sliced across to me.

We'll need some safeguards of course. Could this be abused? Of course. But - actually the positives outweigh the negatives. I want my precedent to be used, so I'm encouraged to write it carefully and solidly. I want others to rely on it, so I write if for the general case as well as the particular case. I won't get any revenue unless future cases find my work compelling. If I spam the precedents book with marginally different cases, I'll get boycotted. There is substantial negative or controlling feedback built into the system.

If I refer to another precedent, I've probably got to slice a fair fee across to it. So we need to adjust the above fee split:

d) fixed portion for precedents, falls across to general insurance fund (not the Arbitrator) if not spent.
e) general insurance fund to handle bigger losses.

A well written and spot-on precedent could make an Arbitrator rich. But isn't that a fair result? If the Arbitrator has slaved through a case and established the rule, which then goes on to guide future cases, business and society, isn't it reasonable that reward flows?

Recall, that the Arbitrator cannot make others rely. Every future case is incentivised to rely on the merits, an Arbitrator is judged on his today-ruling. A ruling has no need to rely - only the efficiency of doing so. So each new event is a due diligence vote over prior work - in effect the entire system would be due diligence.

I think getting the details of a monetary feedback loop over precedents will take some whiteboarding. The details are pretty messy. But it also is enticing. If we think about the entire academic machinery, how cites are obsessed over by tenure committees, and fraudulently rigged in false conferences and chummy cross-citing "peer-review" groups, this mechanism is similar - we are accounting for good work. Why not make precedents linked to money? Why not make it be money?

ps - there is one good criticism I've been handed by Eva, which is this - a good precedent creates a rule which informs beyond the case. That is, the best precedent solves problems outside the forum, and this aspect of goodness is actually superior in benefit than the precedent cited in forum. And this has no part in the feedback loop.

It's a good critique - and why I asked for critique from Eva :) and to be frank, I've no answer to it yet. Thoughts?

Posted by iang at November 5, 2016 08:19 AM
Post a comment

Remember personal info?

Hit preview to see your comment as it would be displayed.