All Articles

Exploring Harberger Tax Rates in Virtual Collectibles & Patronage Markets

In the creation of patronage as an asset class, one creates virtual collectibles whose beneficiary earns revenue at a specific rate from an always specified self-assessed price by the owner.

For example:

One creates a virtual, collectible water bottle. When owning this virtual bottle, you always have to specify a sales price. At a specific rate (say 5% per annum), the owner has to pay a fee towards the beneficiary (say, a water charity) to keep it. Anyone can buy it from the current owner at the self-specified price.

This ownership model is called Common-Ownership Self-Assessed Tax (COST) or Harberger Tax.


One of the outstanding questions in these patronage markets is the choice of rates.

In literature (, recommend Harberger taxes are quite low (closer to the expected turnover rate), because it’s all for physical items where turnover is expensive, and investment into the assets is desirable (eg, renovating a home). Literature also makes a distinction in rates between assets that are prone to monopolistic exploitation, such as spectrum versus family heirlooms.

Matthew Prewitt did great research towards thinking about intellectual property, delineating it into natural and artificial capital. The gist is, is that when people invent, the rates should be low to allow the inventor to exploit their ideas, before eventually accruing a tax rate closer to enabling the public’s ability to exploit it too. A similar proposal of ratcheting rates was put forth by Arjun Narayan.

However, there’s no research for virtual items. Virtual items would have quite broad uses. For example, virtual items that have use-value (eg, a digital blockchain sword in a game or a virtual piece of land in Decentraland) would have a lower turnover rate because someone actually wants to use it. Some of these virtual items might even have investment capabilities (upgrading with effects, aesthetics or new uses). In these realms, you still need to consider the trade-offs of Harberger Tax on users and the world. If people can corner a virtual land market, for example, then Harberger Tax helps. If the game developers seek novel funding, then for certain, scarcer items in the game, Harberger Tax can raise funds for the team and the virtual economy itself.

Pure & Patronage Collectibles

Patronage collectibles could have additional use-value: eg if funding a gorilla sanctuary, one can get the privilege to vote on new names for baby gorillas. However, in some circumstances, the patronage collectible has no other use-value, except the ownership of it, itself. It’s the act of collecting, alongside the ability to exclude access to others in being a patron. In other words, the lowest common-denominator in use-value is ability to exclude others from donating towards projects and accruing status through it. This, I would call a pure collectible (it only exists to be collected).

It gets into a strange territory, because right-of-use is low, and the issue of monopoly extraction is also low. A question for the latter: Does society suffer if markets in charity are cornered? Is it even capturable at all? However, the act of creating this collectible at all, above and beyond its use-value of donation, collecting & status, is that it facilitates trading/speculation. The is the key ingredient of patronage as an asset class, because it adds additional self-interest to donate to common or public goods.

So, in this case, perhaps the most important facet in these markets to choose a tax rate that is not so much about the asset itself (if it has little use-value), but rather the ability for it to facilitate this self-interest and trade whilst maintaining positive externalities of donating to common good.

The trade-offs thus acquire new dimensions:

Too high prices (low tax rate) and most of the value transferred is being paid as profit to traders. To pay $100 pa to a charity on a low tax rate of 0.01%, the price would be $1,000,000. In order to donate $100 pa, you’d have to put down $1,000,000, given to the previous owner. This capital outlay is too much and thus donating directly achieves the same ends if donating is the core goal.

Too low prices (high tax rate), and traders lose their self-interested opportunity for potential profit above and beyond the value of their donations. To pay $100 pa to a charity on a high tax rate of 10,000%, the price would be $1. The potential profit on trading is low even though more traders could participate.

For $100/year:

0.01% tax -> Price: $1,000,000
10,000% tax -> Price: $1

Thus: because for these collectibles, patronage being the primary purpose, higher rates are better. This is because it moves the needle towards the intended purpose: funding donations through a new market for self-interest. The money flows towards the charities at the end of the day. Too much angling towards profit for speculators and the positive externalities from trade becomes less.

But what that percentage is the best is still unknown. Is it 50%, 100%, 200% or 10,000%? If the goal is to facilitate trade, then reducing mental transactions becomes a meaningful factor. Since you are combining trading into a donation system, then anchoring towards existing patronage systems could help. People are familiar in donating per year or per month.

Thus: either 100% per annum or 100% per month (1,200% per annum) makes the most sense. When a user sees the current price to buy access to donate, they know it would be an equivalent amount they will donate per annum or per month. My gut feeling lies in 100% per annum.

Outstanding Questions & Incentives

There’s quite a lot of levers to pull in the design of market like this. For example, if a patronage market like this exists, and there’s little use-value besides the exclusion of access to donations, then is a collectible even necessary? Could one design a market just where users trade/buy/sell slots to access the ability to be a patron? My thinking is that, yes, if you remove collectibility, there’s still self-interest to trade such a market. If you trade and lost money, you still donated to a project. If you trade and made money yourself, everyone won. If it just a pure slot, then it does run more of a risk of playing into a game of hot potato. Even if there’s positive externality, you want to ultimately reduce speculators in just buying to sell to someone else for a higher price. However, this game might still be meaningful in and of itself, because of the excessive positive externalities it produces.

Additional incentives only adds additional demand:

  • Adding a collectible to the mix adds aesthetic meaning.
  • Surfacing status adds willingness to donate above and beyond trading.
  • Adding additional use-values to the slot/collectible adds more demand.

Ultimately, when a user approaches a Harberger taxed market for patronage (at 100% pa), they will considering the following (excluding additional incentives around status & collecting):

  • I can donate $20 pa directly.
  • OR, I can buy an option for say $10, up the price to $20 and donate the total across a year. If someone buys the option from me before 6 months, I will have made a profit.

Without incentives for collecting or status, it creates a market for speculation that allows interested parties to leverage themselves. If the cause becomes more worthwhile (in part because it becomes necessary OR the owner promotes it through self interest), someone will buy it from them at this higher price.

I am still not sure if it is kept this simple (only trading slots), whether incentives are sufficient. It is such a difficult bag of incentives, because incentives for donating, status, altruism and markets are all different. Adding a market to altruism for example, already radically alters it. One tweak in either direction and these ideas can be dead out of the water or be radically successful. This does not even include how all these incentives play out in terms of UX and complexity.

It definitely remains very interesting. Perhaps, I should start playing around with cadCAD (Complex Adaptive Dynamics Computer Aided Design) to figure out what happens.