The early implementations of metaverses have seen NFT markets fraught with rampant speculation. As the industry evolves, many of these NFTs are likely to develop substantial utility due to the unique privileges they bestow within their respective metaverses. This shift also comes with an increased risk that speculative activity will undermine many projects’ long-term viability. Speculators generally will choose to purchase assets without actively using them, which limits access for more active participants by reducing the supply and increasing market price. While speculator-owned assets could achieve increased utilization through rental agreements, speculators are still able to extract wealth without providing a real benefit to the economy. Rent-seekers must be cut out to improve economic efficiency. This issue warrants research around potential economic mechanisms to limit the negative impact of speculators on the market for non-fungible assets and facilitate economic growth.
This paper proposes the Bid-Based Land Value Tax, a potential taxation method for non-fungible assets that uses principles from Georgist Land Value Taxes and Harberger taxation systems to combat speculation and value assets without the need for external bureaucracy. It also confers additional benefits around improving buy-side liquidity and incentive alignment by allowing DAOs to collect tax revenue from asset ownership, rather than just sales. This mechanism uses asset bids for tax assessment while leveraging tax revenue to incentivize buy-side liquidity and price discovery in a way that can be realistically applied in metaverse economies. After a survey of popular anti-speculation mechanisms suggested in the past, we will propose our preferred mechanism.
A commonly cited method for counteracting speculative activity is a Land Value Tax (LVT), championed by the 19th-century author and advocate Henry George. The concept originated for physical land but can be applied in the metaverse as well on a variety of non-fungible assets. The direct comparable is metaverse land with a specified location, which provides scarcity and utility based on proximity to desirable locations. Even if additional land is created, it would not necessarily be near something players are interested in and therefore not satisfy excess demand for land. However, other types of NFTs that have desirable utility and a permanently fixed supply (for example due to a commitment to early investors to not create more assets with that exact utility) can also benefit from an LVT.
An LVT is implemented by separating the value of the land or base asset from any improvements the owner is able to make, then implementing a tax only on the base asset. This serves to disincentivize land speculation by making it costly to sit on an asset without putting it to use. This should result in the land being owned by whoever can use it most productively. Land Value Taxes also push asset owners to make improvements to their land as they need to offset the land tax, but are not taxed on any increase in valuation from these upgrades. LVTs still require an asset valuation method, which is combined with a specified tax rate to calculate the tax expense. The following valuation methods could also be implemented in a non-LVT scenario with both land and improvements taxed, and are often discussed in that context, but this hurts investment efficiency (getting the owner to improve the asset), as additional taxes on improvements disincentivize their production.
Valuations for the asset and improvements can be based on comparable sales of these separate components, as discussed by Lars Doucet for virtual land. Hedonic regression can derive each attribute of an asset’s contribution to the overall valuation and then set taxes based on that assessment. This method may be complex to build and require additional bureaucratic components during implementation, such as an appeals process, in order to placate participants. The system could also be exploited by users who wash-trade assets with certain traits to reduce their own taxes or increase them for competitors.
Another option for valuing the land/base asset is to set fixed-term leases for the land and auction the leases off each period. Doucet suggests using a Vickrey Auction, which is a type of sealed-bid auction where the winner pays the amount of the second highest bid as an incentive to set their bid at their true valuation. While this leads to high allocative efficiency (the person who values the asset most gets it) at the time of the auction, that is not necessarily the case in the periods between auctions and comes with the tradeoff that ownership can be easily lost at any auction. This tradeoff conflicts with the potential design goal of true asset ownership, which is prevalent in many metaverse projects.
Proposed by Arnold Harberger and popularized by Posner and Weyl in Radical Markets, Harberger taxes are based around self-valuation by the asset owner. Taxes are collected based on the valuation the owner chooses, but with the catch that anyone can buy them out at any time at that price. Harberger taxes theoretically lead to excellent allocative efficiency and benefit those who wish to be able to buy up multiple assets simultaneously (for example getting metaverse land in close proximity without other owners noticing and raising prices).
Some drawbacks here include many users having limited ability to accurately self-assess asset value and frequently re-assess as the market shifts. The potential for immediate loss by the owner can create a poor user experience, making Harberger taxation unlikely to be implemented in practice. Participants’ valuations also could be overstated due to their level of risk aversion.
In order to mitigate the downside of forced sales with Harberger Taxes, Fabricio Nakata proposed a modified version where the asset owner had the option of accepting or rejecting the bid. If they reject the offer, then they are taxed with the valuation of that offer and the bidder gets rewarded with a portion of the additional tax revenue. The owner can choose to lower their assessment but must put up money in case a higher bid occurs that they reject. This method adds some additional complexity to implement and still involves some reliance on participants being able to accurately self-assess valuations.
In 2017, Anthony Zhang shared a paper on depreciating licenses. Similar to Harberger taxes, the taxes are set based on the market resale value. However, unlike Harberger taxes, the asset depreciates by a certain percentage, for example, 10%, annually. The government auctions off the “depreciated” part (10%) and the asset owner either purchase the 10% back or someone else purchases 10% from the government and 90% from the current owner. This method encourages the right selection of owner of the asset – where an owner using the asset for productive purposes is more likely to repurchase the depreciated 10% while an owner holding on to the asset for rent-seeking purposes is likely to sell the 90% instead of rebidding for the 10%, that might return less value than sitting on the unproductive asset. A downside to this method can be that the valuation and corresponding taxation of an asset is only updated as frequently as auctions occur, which can lead to market prices deviating significantly from valuations in the meantime in volatile markets (such as the markets for most NFTs). This can be mitigated with more frequent auctions, but this places an additional burden on the owner to participate or risk losing their asset.
In 2019, Vitalik Buterin discussed a harberger variation focused on bids when setting the valuation for taxes. In this system, “users pay the Harberger tax by prepaying funds into a contract, and there exists a separate mechanism where users can submit bids for any item of property; each bid must be fully collateralized. The tax rate that the current owner pays is computed based on the value of the highest bid. The owner has the ability to sell to the highest bidder at the bid price at any time (or the sale happens automatically if the prepaid tax contract runs out of money)”. This bidding method gives control over sales to the current asset owner but does rely on the existence of buy-side liquidity for efficiency in asset valuation and liquidation.
More recently, Buterin made a similar suggestion for a demand-based recurring pricing strategy for taxing ENS domains where if a bid were open for sufficiently long, then the domain valuation would rise up to that level and if there are no bids the valuation would gradually decay. A potential tradeoff here though is that an asset holder could potentially end up paying much higher taxes than warranted if the overall market dropped and there was no longer a bid open for them to accept (even with the gradual depreciation in valuation).
The valuation method proposed by Buterin can be expanded upon in three ways to create the Bid-Based Land Value Tax model. Firstly, apply the Land Value Tax method of splitting bids into two components: the bid on the underlying land/base asset and the bid for the house/improvements on top. The tax rate is only levied on the land component and owners can choose to accept bids separately for the land or house with these components split into separate NFTs. Metaverse assets are particularly adaptable for this type of component separation, relative to physical real estate. The outcome is that the asset owner is taxed using the current highest land bid only as the valuation.
Secondly, allocate a portion of the tax revenue towards the bidder with the current highest unaccepted bid for that asset, creating an incentive for participants to make competitive bids and leave the bids open over time. This improves buy-side liquidity for non-fungible assets and drives price discovery. If there are multiple highest bids within a tax period, tax revenues are distributed in proportion to the revenue each bidder generated, based on the size and duration of their bid. Take an example where bidder A caused the owner to pay $10 in taxes during the first part of a tax period and bidder B held the highest bid for the remainder of the period, causing $30 of tax obligation. If 50% of tax revenue is allocated to the top bidder, then bidder A would receive $5 and bidder B would receive $15. If the land owner accepts a bid on the land component, then that bidder becomes the new owner and tax-payer, while the next highest bidder becomes the new valuation reference and revenue recipient.
Lastly, it is important to clarify the taxation and liquidation protocol for implementing a bid-based taxation method. Expanding on the prepayment concept, asset owners do not always need to have funds in a prepayment contract (as early pre-payment reduces capital efficiency), they just need to add sufficient funds into the contract to cover their tax obligations for that period any time before a fixed tax due date. If the asset is revenue-generating, earnings can be directed towards this contract to automatically cover tax obligations. In the event of a failure to pay on the due date, the land component is automatically liquidated through a sale to the current highest bidder, with a cut redirected to the owed bidders out of the sale revenue. If a forced sale is not desirable in a given situation, land with back taxes can be temporarily stripped of utility until the debt is paid. An interest rate can be set on this debt such that it grows at a rate higher than a speculator’s estimate of appreciation and if the owner sells an asset with back taxes, sale revenue will be redirected to cover them.
To summarize the mechanism, the current owner of the asset incurs a tax obligation based on the current highest open bid from another user for the base asset multiplied by a specified x% tax rate. At periodic intervals, the owner must pay those taxes, with a portion of that revenue going towards the bidders who held the highest bid at various times during the period. If the tax is not paid at the deadline, then the base asset is sold to the current highest bidder with a portion of the revenue directed towards paying off back taxes. Bids are separated between bids for the base asset and for improvements. These can be accepted independently and only the bid for the base asset is used for determining the tax obligation for the current owner. If the current owner accepts a bid for the base asset, then that bidder becomes the new owner and starts accruing a tax obligation based on the new highest bids going forward.
A simple first example of how a Bid-Based Land Value Tax (BBLVT) could be implemented is for land in a game with a limited supply, which aligns well with the original focus of Land Value Taxes on physical land. In order for this metaverse to be successful, people need to build structures and experiences on it that improve the experience for other players. However, the limited nature of land brings in a wave of speculators who have no intention of building, only to hoard land until they can sell it for more later or extract rent without providing value to the ecosystem. This behaviour has been shown to kill a game’s growth in past MMOs. In order to stop this crisis, a BBLVT is implemented and incentives suddenly shift. Bidders enter the market seeking tax revenue and a chance at finally getting a house for a reasonable price. Faced with taxes on their unproductive land plots, speculators accept bids and give up their speculation attempts. These bidders will be those who are willing to put the land to the most efficient use, likely through development that creates some revenue opportunity and overall benefit to the ecosystem.
But a BBLVT can be applied outside of land to most scarce assets that have some important utility. Consider a game in which a particular scarce NFT (let’s call it a “Hero”) is required in order to play the premium game mode, which has revenue potential through tokens and/or other NFTs. One type of market participant, “players,” purchase Heroes and actively use them in-game. However, given the limited nature of Heroes (perhaps the limit is based on a desire to avoid diluting initial holders through a supply increase), speculators step in. They buy up Heroes, but have no intention of playing; they just want to sit on them until the price goes up or extract rent without adding value. The game is dying due to a lack of players in the premium mode, so the team implements a BBLVT. Offers are then split between bids for the base Hero and bids for any upgrades that it has, with taxes only on the base bids, so that players are not penalized for playing and upgrading their hero with increased taxation. If needed, upgrades could be split into NFTs separate from the base hero. That means the owner does not have to accept both the base and improvement bids from a particular bidder. If a bidder has a high offer on the upgrades and a low offer on the Hero, the owner can choose to accept just the upgrades bid and sell the upgrades as a separate NFT. Those who can put the Heroes to the best use put pressure on speculators to sell through high bids (with their corresponding taxes) and Heroes end up in the hands of players who will use them.
As a practical example of a current issue, a BBLVT could be used as a fairer method for setting renewal fees for ENS domain names. As outlined in Buterin’s ENS article linked above, ENS renewal fees are quite low and a large number of speculators have bought them up. These squatters do not intend to build websites on the domains, which is a poor outcome for the growth and long-term success of the ENS system. The current fee structure does not adjust upwards based on the demand for each domain, providing sub-optimal revenue to the ENS DAO (another poor outcome for ecosystem growth). In this scenario, the ENS domain acts as the land, while the website built on top is the separate improvement. In a BBLVT implementation, bids are separated between the domain and website, with taxes levied based on the valuation from the current highest bid for the domain.
Incentivized bidders provide liquidity and price discovery, with the valuations from their bids helping set taxes that better match the variation between names. Costs ramp up on higher quality names and owners always have a choice between accepting a bid or being taxed at that valuation. Valuations vary over time as bids reflect market fluctuations. Minimum renewal fees (e.g. the current fees) can still be implemented as the baseline when there are no bids in place. The long-run result is ecosystem growth through more domains allocated to those who build websites, exposing the world to ENS, and increased DAO revenue that allows for additional initiatives.
Note that there are nuances to all of these scenarios and additional changes or caveats may be needed, but we hope they provide clarity on how a BBLVT could be used in practice to alleviate the negative pressure of speculators on a project’s long-term viability.
The BBLVT model provides a method to combat speculation and value non-fungible assets while owners retain the right to choose if the asset is sold. As opposed to a Harberger tax system, owners are not penalized for poor valuation skills or forgetting to update their valuation. This more palatable method should have a better chance of actual implementation, especially given the property-rights-focused nature of NFT ownership.
Bid-based taxes create predictability and benefits for the owner’s future planning, which is lacking with Harberger taxation, as pointed out by Vitalik Buterin in his response to Weyl and Posner’s Radical Markets.
This model also benefits from not needing bureaucracy for assessing non-fungible asset values. There are no additional organizations needed for performing valuations or reviewing appeals for valuation that may arise in a manual or automated assessment-based valuation model. This simplifies implementation and acceptance by participants.
Tax revenue can be collected from ownership, rather than just sales, and the portion of revenue that does not go towards incentivizing bidders can be used to improve the project ecosystem. That revenue can be used on grants for public goods or targeted incentives for the beneficial user activity. DAOs generally act as a governing body in these virtual worlds and this opens up a method of growing their treasuries (and corresponding impact) outside of just a sales tax on transactions. Instituting this tax on ownership better aligns the DAO’s incentives for facilitating long-term participation and allows for sufficient tax revenues in markets with low turnover.
Creating incentives for buy-side NFT liquidity improves the viability of further financialization in this asset class, such as NFT-collateralized lending with reduced slippage on liquidation. In the process of providing this liquidity, bidders are also in frequent competition with each other to provide the highest bid (without bidding over their own valuation for the asset), which improves price discovery as bids are public. This liquidity is also useful for facilitating liquidations without the need for a separate auction system, as bidders are essentially in a perpetual auction that is only forcefully settled if the owner does not pay their taxes at the end of the tax period. Greater buy-side liquidity also decreases risks for asset holders and expands the audience of investors interested in that asset. These benefits of increased liquidity help justify the expenditure of tax revenue on maintaining bids over time, rather than running a fixed-period auction.
Another key benefit of a BBLVT is high allocative efficiency. While bidders may be partially or primarily attracted by the tax revenue earned, they also must factor in the probability that their bid is accepted, which rises as they approach the current owner’s own private valuation. As the chance of the bid being accepted increases, a larger portion of the bidder’s expected return is based on their expected return as an asset owner rather than as a bidder collecting tax revenue. Therefore, all else equal (such as risk preferences), those who have a higher expected benefit from owning the land (which is crucially driven by active usage), will have a higher expected return and willingness to pay.
BBLVT implementations should also have high investment efficiency. Taxing only the base asset, not any improvements, avoids any disincentive for activities beneficial to the ecosystem, like using and building on the asset. With owners retaining the choice to sell, the risk of losing the asset suddenly is not a deterrence to investment in asset improvement.
Given the inability to instantly purchase assets, this taxation method is not useful for the Harberger tax use case of eliminating holdouts, who refuse to sell and block projects looking to acquire multiple assets. This is likely a necessary concession for the cause of usability when eliminating forced sales.
With bidding extended in perpetuity, rather than a fixed auction period, bidders should be concerned with the opportunity cost of their bid. The taxation revenue directed towards bidders must be high enough, relative to other opportunities, to incentivize users to participate. This does divert tax revenue away from other uses.
Setting the specific tax rate to be applied to the underlying asset valuation is outside the scope of this paper and open to further research. As general guidelines, the tax rate (or at least the portion of the tax directed towards the highest bidder) should lead to a tax burden slightly lower than the financial benefit of owning the land for any given time period. It should be high enough to be a reasonable alternative to other methods of deploying capital. The tax should push the value of the base asset to near zero to discourage inactive investment, but not far enough to risk a negative value.
We encourage further research and discussion around the concept of Bid-Based Land Value Taxes, as well as other mechanisms for combating speculation (especially in metaverse contexts). Virtual worlds provide an excellent testing ground for proposed economic mechanisms and are flexible enough to even enable approaches infeasible in the physical world. As the size of metaverse economies grows, the importance of properly aligning economic incentives will only increase. Especially as more assets gain cash flow potential, effective methods for taxing ownership will be crucial to supplement sales taxes for DAO revenue. A Bid-Based Land Value Tax can provide a realistic method to deter speculators and maximize economic growth with efficient asset allocation, incentives for buy-side liquidity, and aligned motivation for asset improvement.
For assistance with implementing a BBLVT or other economic mechanisms in a Web3 context, feel free to reach out to Economics Design.