Nostrato: How relational contracts on NOSTR might underpin the foundations of most economic trade in the real world
Prelude & tl;dr: Though bitcoin is itself trustless, any and all economic activity that is tied to the real world has an element of trust and uncertainty for there is no such thing as an atomic exchange with physical objects, including software services rendered from someone else’s computer. The premise of this essay is that Relational Contract theory in economics indicates a game-theoretically stable way to establish and quantify trust with Nostr. As I argued in a previous post (Bitcoin defunds the State; Nostr actually replaces it), economic trading partners are able to form stable trading relationships without necessarily needing recourse to an entity that has a monopoly on violence. Essentially, the idea is that trust is valuable in business, and it is established through a sequence of events where each party recognizes there is mutual value to the relationship. Herein, I haphazardly attempt to move toward a semi-formalized explanation of how that might play out. This was written partly in response to Ray Yousef et al CivKit proposal, which I did not see as providing an actionable framework. Moreover, I am highly skeptical of escrow based solutions over Relational Contracts. I do not know that what I lay out is valid, but in the event that it is, my hope is that it might be useful to someone. I tried to get feedback on this, but it seems people find it confusing so I realized I should just get it out there. Thank you to the person who suggested I should add a slight introduction. Hopefully this suffices.
This forms the basis of a simple relational contract model that exhibits stability without 3rd party escrow or threat of violent countermeasures in response to a breach of contract (whether such service is delivered by a large entity, or a competitive, distributed market with blinded information between the bilateral parties and their 3rd party). We might also call this arrangement a sort of stateless statefulness for the state of the relationship exists, but it can be held in its entirety by the market participants without any state-specific entity.
1 Problems with Relational Contracts
Relational contracts have two principal challenges/downsides:
Switching costs lower competitiveness: if there is insufficient ability for market participants to choose their local, optimal preference, efficiency of local events is lower. How substantial is this loss?
Bootstrapping the relationship: we described a relational contract with a future value needing to exceed a certain idiosyncratic threshold, so how do you get off zero?
Costliness of destroying relationship: if it is easy enough to start a new relationship with other parties, and the expected future value of continued trade is not at stake, deterrence is limited. Can costliness and bootstrappability co-exist?
The Theory of the Firm (and its actual existence in the real world) proves beyond a shadow of a doubt that the notion that unconditional optimization of participant switchability is wrong. This is obvious when one considers that switching costs are far from zero, and even the smallest of tasks that might be handed to a contractor as an atomic action comprise a series of discrete steps (imagine, for instance, that the delivery of a good were to be bid on a per millimeter of transport basis for the sake of “more perfection competition”– such a notion is obviously absurd).
To bootstrap a relationship, a market participant could engage in costly signaling via the provable burning of funds matching T (Ei ), for a smaller total transaction size, after which they can engage in a sequence of those, and gradually augment the size after their accumulated expected future value grows – in this method, there is a practical downside to destroying the relationship as well, as it requires a sequence of trades that may confer lower profit to the participant. Moreover, bootstrapping algorithms may also design parameters where relatively newer participants offer lower fees – similarly, when a participant values greater certainty of delivery (say, if you are buying emergency insulin that is needed immediately, rather than a sandwich you could survive without), only counterparties with the greatest “proof-of-honest-work” (if such a thing can exist) are able to command the highest premiums for their service/goods.
Now, these observations are not meant to claim irrefutably that such approaches are tenable, but that there are approaches that appear to have dynamics that directionally improve upon the central challenges. Moreover, it is important to recognize that Relational Contract Theory is far more than just theory, as it actually is the nature of the real world today in a great number of instances. Not only does it exist, but it exists in more places than one might think, and this may be related to the fact that its emergent nature means it has no marketing arm. State-violence-based systems are often assumed to be the enforcement mechanism simply because they could be invoked by counterparties at any moment – the reality is that recourse to violence is usually not the true mechanism and it is purely propaganda the breadth of relationships that would cease to function without the state. The stability of contracts and supply chains due to relational factors is particularly extraordinary when considering they are usually not explicitly designed with their ability to function reliably in the absence of the State, yet this emerges regardless. When designed intentionally, it will be even more robust.
2.1 The Collapse of Relational Contracting
In small hunter gatherer tribes, the dynamics of relational contracting were rather different, with a localized web-of-trust within the tribe that was a dense enough network that reputational data could travel with high fidelity, allowing a sort of fungibility such that there could be localized switching with relational levels of low overhead enforcement costs. However, with the advent of agriculture, the State emerged, supplanting much of relational contracts, and the world has been rife with scammers ever since there has been an effectively in finite pool of new suckers. Defecting becomes a favorable strategy when scams can be distributed at scale and retribution is too costly. However, if the honest participants are able to better leverage the upside of cooperation, the relative value accrual to honest participants will grow, and defecting will become a less favorable strategy. This is the promise of a system that is able to both offer the competitive efficiency of low switching costs, with the low overhead enforcement costs of relational contracts. This is the promise of relational contracts with participant fungibility over Nostr.
3 Sybil-Resistant Proof-of-Trade-History
Consider a plurality of agents on Nostr attempting to engage in a particular class of transaction with a real world component, for which there may be a number of clients optimized for that class of transaction. The client may carry out the task of ranking and matching offers, such that the closing of an offer is broadcast to all relays aware of the offer so it is not double-closed. Following the closing of an offer, it is impossible to determine the “correct” amount of bitcoin to go from one party to the other even if one has a good estimate of what the value will be since the contract may go unilaterally unfulfilled, but more often, because it is useful to be able to enter into a contract into which the actual costs cannot be determined until completion. However, there will be some amount of bitcoin that moves in one direction, which can be demonstrated via a Zap or other mechanism that leaves signed historical data. Alas, such transactions are susceptible to Sybil attack since a user can generate a large number of Nostr pubkeys and enter into fictitious activity.
This is resolved via a voluntary rake on transactions that counterparties engage in where some percentage of each transaction (e.g. 1-5%) is sent as a note to a relay with a new type of reputational data format that proves this rake is associated with one’s principal transaction, along with a note sent to other relays that is able to provide an attestation of the rake paid. A user is incentivized to reduce single points of failure so they will likely distribute their rakes over time to different relays.
Of course, this then shifts the Sybil problem onto the relays since how do you know the attacker hasn’t also spun up a large number of relays to generate fraudulent reputational histories? In the steady state, there is huge relational upside to being an honest reputation holder. The more you are a go-to source to check reputation, the more people will send their reputational rakes to you, which is far more lucrative than a Sybil attack. Therefore, this problem actually reduces to one of how to bootstrap an honest reputational relay, and besides the initial bootstrap from zero such relays, this is necessary to ensure a competitive environment. A solution would be for a relay to signal its intent to long term value capture via a provable burn of bitcoin.
With this method, it seems reputational attestation data can achieve a level of useful trustfulness, purely from relational means, with zero reliance on any sort of web-of-trust or know-your-partner in-real-life models. Another feature of this reputation that can now be assigned to a Nostr pubkey is that the value of it is rather quantifiable and bounded, so it can bear no monetary premium – moreover, the “double-spendability” of a Nostr private key means that Nostr identities that have accrued reputational value cannot be traded or sold, monetary premium aside.
That reputational information can be Sybil resistant does not make it perfectly useful since reputational information is analog and can be falsified. One method of increasing honesty in reviews would be to introduce costliness and have a parallel review system where one component is based on ranking the quality and the other with the honesty – there could be a binary fraudulent/theft/criminal blacklist option to select, which bears a cost to make such a claim to reduce the odds it is made spuriously, and this can parallel a 1-10 stars quality score. Data could be structured in such a way that positive re views being weighted are related to rakes you paid, while negative reviews are related to those your counterparty pays so it is not possible to simply pay a low rake when intent on underperforming (it may also be necessary/optimal to have rakes paid upon instantiation of a contract rather than upon completion).
The binary option may be the layer of information upon which webs-of-trust can be useful, where a user might select to blacklist people that a set of known acquaintances have blacklisted via this costly, binary declaration. Such a filtration system would suffice to ensure that destroying relationships have a cost, while the established trade history is easily viewable on relays with zero dependencies on trust-webs, so it would appear we have solved the overall problem of fungibility in relational contracts with Nostr.
4 Next Steps and Areas of Applicability
Of course, we may readily admit that I am hand-waiving throughout and this ought to be more formalized. I expect someone familiar with experience in agent-based modeling could run some experiments to see whether this appears to work, and if so, it may make the case for developing actual services more compelling. The lucrativeness of the reputational raking relays alone should offer a fairly compelling incentive to prove or disprove this conjecture.
The areas in which this sort of reputational scheme is strictly necessary is not without bounds. For instance, both e-commerce and physical store fronts are able to operate without issues since their physicality or brand inertia is large enough that switching costs make dishonest activity unfavorable for sellers, and bitcoin’s settlement finality makes dishonesty much harder for buyers. There is nothing preventing online or physical stores from independently electing to transact in bitcoin, and storefronts are not monopolized so these markets are already fairly competitive. The same is true for international business payments and supply chains where nothing is stopping counterparties from electing to use bitcoin. So even if the reptutational schemes here could offer some utility in those areas, it is not an impetus for adoption. Finally, the sale of less fungible items is also an area where there is minimal need, for Craigslist users can settle in bitcoin, there is the ability to inspect goods, and the Non-Simultaneous Exchange Problem does not arise.
Instead, it is marketplaces controlled by a small handful of companies that control discoverability and inhibit freedom of speech and interaction between market participants leading to highly suboptimal outcomes, and extracting substantial rents. Moreover, these oligopolies/monopolies leverage regulatory capture (i.e. organized violence) to avoid expending resources to improve their offerings. These are the well-known two-sided-marketplace apps, which often coordinate transactions that must have low latency and have low total transaction value thereby making it difficult for people to effectively coordinate these sorts of agreements over a completely open-ended messaging protocol or bulletin board (e.g. if you tweet out a request for a car ride, good luck finding a driver and having a good experience). Uber, Airbnb, Lyft, TaskRabbit, DoorDash, Instacart, Fiverr, UpWork, or Turo. These often have counterparties facing a septuple rent extraction of nested State and indirectly-State-sponsored entities of 3rd party corporation, under 3rd party payment infrastructure, under 3rd party app store, under a taxing municipalty, under a taxing country, under a taxing state, under a taxing nation. This is where there is enough upside for p2p for people to cross the chasm – moreover, some of these reflect a number of geographically localized services where critical mass network effects are not needed globally, but only within a certain region for a service to run smoothly.
These are also the domains where full time gig work is a viable option, such that these tools being brought to Nostr gives individuals a comprehensive alternative option to their fiat life, a true exit option, all while being the most powerful tool for driving bitcoin adoption since bitcoin inherently must be the native currency of this economic system. There will be no reasoning about why bitcoin, it will simply be the only choice so people will use it without thinking about it. Another interesting thing that will happen is that these networks of free agents will have high exposure to bitcoin and low exposure to dollars, in stark opposition to large corporations that will be competing for labor – in recent years, corporations are increasingly failing to be able to plan long term due to the speed with which the dollar devalues and increasing churn of employees so the forces driving the theory of the firm, while still present, are competitively weaker – coupled with the dollar exposure, will eventually tip things over the edge, and the dollar exposure is present due to that being the standard of value employed by their counterparties to whom their are implicit (relational) agreements for future payments. The free agents can swiftly change pricing, but the only way corporate entities will be able to handle the continued devaluation is agreements with their counterparties negotiated around composite currency value standards that require holding a basket of goods to remain market neutral with respect to those obligations and thence will begin the end game of fiat, if it is not otherwise already triggered. Hence, my steadfast bullishness on Nostr.
P.S. I have recently started working on a new aerospace & defense startup... If that is something you might be interested in, please send a reply to this email.