Blockchains as title registries

Using blockchains as authoritative registries for land and other titles is a popular proposal, but there are multiple problems which need to be resolved first. A twitter thread by @oliverbeige.

Kate Sills makes a very interesting argument in favor of using blockchains to “decentralize” (actually, to privatize) land registries. I have a whole bunch of comments regarding title registries in general, and land registries in particular.

Blockchains have this popular appeal as shared, transparent databases that use a participatory consensus-finding process to resolve conflicts. As such, land registries and similar registries (e.g. for music, trademarks, patents, vehicle…) sound like attractive use cases.

To make an actual case for using a blockchain is much trickier, both in theory and in implementation. It’s a bit easier if the property to be registered is intellectual rather than physical, but most of the boundary problems apply to mindspace just as they apply to physical space.

Blockchains make sense in situations where mutuality (rather than hierarchy) is required as basis for conflict resolution, and folding the resolution process into the transaction (“ex intra”) either eliminates or significantly reduces the occurrence of conflicts ex post.

For this to work on a blockchain, the problem domain must be computationally resolvable, or must be rephrased to become computationally resolvable. Otherwise it makes little sense to expend the extra cost of running a blockchain. Resolution should happen on a different layer.

With accounting blockchains the problem domain is comparatively simple. Any transaction has to fulfills the accounting equation (debit = credit), and in cases of near-simultaneous transfers out of a single account a commonly accepted precedence ordering has to be imposed.

“Land” (or music, etc.) poses a number of problems that make current blockchains much less applicable. I expect and hope that further research will tackle these problems, but first it’s necessary to make them explicit.

1. Ambiguous boundaries. Blockchains might make sense if parcels of land are fixed and boundaries are unambiguously defined. Actual claims to property fulfill neither — see e.g. the requirements of the UK HM Land Registry for establishing boundaries.

2. Legacy rights and precedence. Blockchains are best applied in “greenfield” scenarios, where an immutable record from genesis can be established. But no computational algorithm embedded in a blockchain can resolve disputes arising from competing claims older than the registry.

3. Out-of-domain disputes. Conflicting claims on land often arise not from ambiguities in commercial transfers, but from inheritance disputes or (private or public) expropriation, or simply shifting landscapes. Blockchain consensus doesn’t do well on that.

4. Recording conflicting claims. To deal with unresolved competing claims, current title registries must have a mechanism to record and maintain records even if they are either ambiguous or directly conflicting. Blockchains are expressly designed to reject ambiguity. This is esp. relevant for patent, copyright or trademark databases, where scope cannot be delineated in advance in a way that an exhaustive, mutually exclusive mapping of scenarios to claims becomes possible. With land, such a mapping could at least be considered in theory.

5. Blockchain consensus works on three principles: i. Proposing an invalid solution is a costly and almost surely unsuccessful endeavor. ii. (Hash-weighed) majority is congruent with validity. iii. Majority converges toward (near-)unanimity to avoid forks. Title disputes are resolved by law especially because (ii) cannot be guaranteed IRL. Valid claims are not necessarily supported by a majority, especially if the majority can be acquired with hash power. To create a consensus algorithm that can resolve this is the holy grail.

6. Enforcement. Property cannot be disconnected from the force required to defend it, so public blockchains only make sense as a substitute for public, centralized land registries as long as they can compel enforcement (private or public) to act in accordance with their rulings.

About the proposed solution to let ratings agencies compete for the most credible estimate of the credibility of property claims/transfers: This is clearly a privatized solution, not a decentralized one, especially knowing that ratings markets have strong economies of scale.

Ratings agencies are of course self-interested actors, and also prone to herding. An oligopoly of private players who don’t inject enough diversity of opinion into the market doesn’t sound like an appealing solution.

Also, there’s a strong empirical correlation between level of industrialization and public enforcement of property. The direction of causality is somewhat complex, but an argument in favor of private ordering should make a strong claim why such a solution is more efficient.

And of course, the existence of such a market for competing ratings agencies would make blockchains an unnecessary and costly complexity…

I write about how technology shapes the world we live in.