Liability rules and property rules as a framework to view virtual world RMT

Calabresi and Melamed proposed a three-way taxonomy of entitlements – inalienable entitlements, which cannot be traded between willing sellers and willing purchasers; property rules, which are enforceable by injunction to prevent non-consensual takings; and liability rules, where objectively measured damages are the only remedy, leaving open the possibility and expectation that the benefit will be taken non-consensually by a person who values it more than the objective measure.

So, does any of this help us frame disputes about entitlements in virtual worlds?

slot machine - cobalt123

Image by cobalt123, CC NC-BY-SA

Imagine we have to determine what we should do about virtual property. Calabresi and Melamed tell us that there are five ways we can allocate the entitlement. Leaving aside an outright prohibition on the transfer of virtual property, we could grant a property right to the publisher, which would mean that the participant only acquires any interest in virtual property through a negotiated agreement with the publisher – i.e., the ToS. This is certainly how publishers currently see themselves (or would like to be seen). While it may be appropriate for Bartle-World, this model ceases to be appropriate where the interests of the player base are not so homogenous and the actions of the publisher give rise to expectations of legitimate participant interests in virtual property (see Bragg v Linden).

Alternatively, we could grant a property entitlement to the participant, which would prevent publishers from confiscating or devaluing virtual property. This is unlikely to be a workable system, because the publisher will not be able to make any changes to the platform without obtaining consent from each and every participant – although consent could be given in the EULA, which, assuming low transaction costs, would give us much the same situation as that above – in either scenario, participants or platform owners would pay for the rights to deal with the property as they want to. The problem here really becomes one of achieving this bargain with the entire population in advance, and then not being able to modify the bargain without the consent of all involved – a process which is vulnerable to large negotiation costs and strategic behaviour.

The third method would be to grant a liability entitlement to virtual property to the participant. The participant would be entitled to deal with their property as they see fit, but the publisher would be able to make changes to the platform as required, provided they compensate the participants for their loss. For example, this seems like a workable solution to the problem of 'property' and 'currency' in Second Life – where participants certainly feel that they are entitled to the value of the money they convert to Linden Dollars, and the items they then buy, but Linden Labs needs the ability to modify its rules (eg banning casinos) or enforce its rules (eg Bragg). In this situation, such a change to the value of a participant's virtual holdings may be compensable but not prohibited.


Alternatively, we could grant an entitlement to virtual property to the publisher, but support this only with a liability rule. This would mean, for example, that the publisher would not be able to prevent RMT – the participant could deal with the property he or she possesses at will, but would be under an obligation to compensate the publisher at an objectively determined rate. Would this be an attractive solution? In an appropriate situation it may be desirable to allow a participant to commodify their virtual property, to sell it on the open market, and then compensate the publisher for its loss. This approach prevents the platform owner from unilaterally prohibiting RMT, but also prevents the participant from appropriating the whole value of the virtual property by transferring it and retaining the proceeds. This may be useful in recognising that the value of virtual property is not created or determined by either party alone. In this scenario, the obligation to pay a share of the value to the platform owner without being prevented from dealing with the property may be attractive enough to all parties.

Are these models helpful? I think they provide an interesting framework for examining the allocation of entitlements and the protection those entitlements are given. In reality, the entitlements above need to be much more atomic, and the models much more complicated and mixed than the brief blurb I have provided. For example, in any one platform, the right to sell virtual property to a third party for real world cash may be a liability rule which belongs to the platform owner, the right not to have one's account terminated and virtual property confiscated may be a property entitlement of the participant, and the right not to have one's property interfered with by rule or design changes may be a liability entitlement of the participant. Our choice as to whom entitlements are initially allocated and how those entitlements will be protected still depends on our policy goals.