ANT Demand Modeling Framework

This document describes a basic framework for thinking about the demand for the Aragon Network Token (ANT) based on its current and potential future utility value.

Currently, the primary utility of ANT is governance. Holders of ANT are able to participate in the AGP process, allowing them to make proposals which include allocating funds from the Aragon Association’s treasury and shaping the strategy and direction of the project as a whole.

So far the AGP process has been used to primarilly fund teams building Aragon related infrastructure and products including the Aragon Client, developer tools, and Aragon Network service protocols like the Aragon Court. In broad strokes, these initiatives are intended to drive adoption of Aragon, grow the market for Aragon Network services, and ultimately drive demand to hold ANT.

Some of these initiatives can be considered public goods for the Aragon Community. They are made available to all Aragon users without any direct monetization or fees. For example, the Aragon Client and Aragon Developer Tools are open source and made freely available – but they are intended to grow the market for Aragon Network services.

Some initiatives like Aragon Fundraising are public goods with defaults that directly benefit the Aragon Network and ANT demand. For example, the deployment template and on-boarding experience for organizations using Aragon Fundraising uses both DAI and ANT as bonding curve collateral. This creates a liquidity bridge between organizations with ANT as a hub, and it locks ANT as collateral as demand for the organization token grows.

Finally, some initiatives like the Aragon Court are private goods which must be paid for in order to receive benefits. In the case of the Aragon Court, subscription fees and dispute fees are captured by the protocol and used to incentivize jurors to stake and adjudicate disputes. Jurors are required to stake ANJ, a protocol specific “work” token, which is linked to ANT via a bonding curve.

In the future there may be other Aragon Network services which may be monetized in different ways, but the basic framework of linking a derivative token to ANT via a bonding curve seems likely to be a recurring theme, as it enables the services to be designed with independent protocol incentives, such as participation rate targetting. Maintaining this separation allows for more effective protocol design and reduces the complexity and risk for the average ANT holder.

Bonding Curves relate a collateral token (e.g. ANT) to a bonded token supply (eg ANJ supply * price) using a ratio parameter called connector weight. As the market price of the bonded token increases the bonding curve will pull in additional collateral by creating an arbitrage opportunity for traders, similarly if the market price of the bonded token decreases (relative to collateral), the bonding curve will decrease the supply of bonded tokens in order to release collateral. This creates a fairly simple relationship between the price of a bonded token and demand for bonding curve collateral because the contract ensures that the following invariant is always true connector weight * market cap of the bonded token = value of collateral . For a detailed breakdown of the math behind bonding curves I recommend reviewing Bancor’s white paper.

Using this relationship we can create a basic model for ANT demand as a function of aggregate demand for all bonded tokens connected to ANT, each bonded token or category of tokens (in the case of fundraising) can be evaluated separately and independently using an appropriate valuation model, and the aggregate result can help inform our understanding of long-term demand for ANT. In order to focus on the individual initiatives, the current price of ANT is assumed to be known and fixed, allowing the model to extrapolate based on that how much ANT would be locked up by the bonding curve as a result of demand for a particular service described in the model.

As an example I’ve created individual demand models for both the Aragon Court and Aragon Fundraising.

Court Draft:

Fundraising Draft:

My intention with this is to create a general framework to help analyze various initiatives that Aragon and ANT holders could pursue and fund. I wouldn’t consider myself an expert in this type of modelling, and encourage people to help improve both the assumptions and methodology used here.

6 Likes

will the staking bonding curve look like this for ANJ, ARA and / or Aragon Fundraising?:
Bonding%20Curve

Could you give an example for this statement?

can you trade ANJ or ARA on exchanges without touching the bonding curve? yes but there will be always arbitrage opportunities because of the bonding curve right? so even if people by lots of ARA on exchanges and increase the price of ARA there then there is an arbitrage opportunity to sell ARA on exchanges and buy ANT to stake it in the bonding curve because there the price is lower. But who determines the % of the connector weight? the ANT holders or the market or how?

why do you think that the connector wight will be likely between 20% and 30%?

Im not sure because there are two curves in this chart.

This chart from the bancor whitepaper may help:

With low connector weight, you end up with a curve where as the price increases the supply begins to reach an asymptote. If we think of the curve as a market maker/exchange as the the curve asymptotes its you have an increased spread, greater price slippage, and similar properties to a market that has a low depth of liquidity.

With a high connector weight, you get closer to a peg, there is relatively low price sensitive/very little slippage between the connected tokens. In the extreme case of 100% connector weight you have a one to one peg.

We want to balance the connection between ANT (because this is how investing in the development and maintainence of these systems pays off for the Aragon Network and ANT holders) with the need to create semi-isolated economic systems. If we don’t have that isolation the economic systems either wont work, or they will become incredibly complex and determining potential failure modes related to the dynamics of multiple systems overlapping and interacting becomes extremely difficult.

The 20-30% number is something that feels like a reasonable balance, and its likely a pretty large window in terms of what will work. But would welcome further analysis on what the right value for the connector weight should be.

With Aragon Fundraising its possible to have the bonded token be transferrable or non-transferrable. Minting/Burning the token using the bonding curve is not a “transfer” as it interacts directly with the mint/burn functionality of the token contract for the bonded token. So its possible to have a token which is not easily trade-able on external markets and this could have interesting properties for system design…

In the case of ANJ or ARA, there doesn’t seem to be a strong reason why we would want to prevent trading on other exchanges. The goal of the curve is create a relationship between the demand for ANT and the demand for the bonded token (ANJ/ARA), as the price offered moves along the curve, slippage/price sensitivity will increase (on the curve), but trade volume can simply move to external markets. We need the curve to have that shape because we need that partial isolation between the economic systems, but it’s not necessarily desirable for to force a low liquidity depth on that market.

so even if people by lots of ARA on exchanges and increase the price of ARA there then there is an arbitrage opportunity to sell ARA on exchanges and buy ANT to stake it in the bonding curve because there the price is lower.

Yes, even if there is an external market, the market price will be reflected in the bonding curve because traders can arbitrage the “automated market maker” functionality. When the do the market maker will mint/burn bonded tokens and hold/release ANT to keep a ratio between the value of the bonded token and the relative value of ANT held as collateral.

This is a parameter set up when the curve is initialized. It likely won’t change–it should be treated as an invariant.

For some of these systems there will be protocol parameters (not necessarily the connector weight) that could be changed through governance. The governance process may be different depending on the parameters, and likely would involve some combination of ANT holders and participants in the protocol involved in making those changes. For example, for Aragon Chain there may be some parameters governed entirely by validators, some parameters that are governed by a double approval of both ANT holders and validators, and some which may be entirely governed by ANT holders.

I tend to think of the governance over parameters as an important political question, but separate question from the lower level protocol/system design.

Hope that’s helpful!

1 Like

thank you i know understand this system a lot better and i feel confident in this bonding curve token network and think this is a very elegant way to balance autonomy and interdependence between the different token systems within Aragon.

2 Likes

I struggle to understand why the value flowing through the network should be partitioned among multiple classes of tokenholders?

Looking at this through the capital structure irrelevance principle (https://en.wikipedia.org/wiki/Modigliani–Miller_theorem), if changing the mix of assets in the capital structure of the network doesn’t change valuation shouldn’t the simplest possible structure be used?

In the case of ANJ, what are the issues in having jurors stake ANT instead of ANJ?

In my mind, the relationship between ANT and ANJ, is similar to a stock and convertible bond, with the connector weight being the largest influence on the leverage factor. Where a corporation would issue bonds to buyback it’s stock (or instead of issuing more stock) and effectively allocate a portion of cashflows to servicing the debt. In a tax-free world that would not change the value of the firm, however in the corporate world taxes do exist and so this is commonly done as the interest paid is tax-deductible which lowers the cost of capital and thus raises the value of the firm. In the case of a network, it’s not a taxable entity and I don’t think this structure brings those same tax benefits to network participants.

It seems the best reason to introduce such a structure is if it introduces positive cryptoeconomic incentives for network participants that could not be achieved without it. I struggle to see what those incentives are.

2 Likes

Hey @bgits,

Welcome to the forum :slight_smile:
I don’t think that the reasoning behind token classes in Aragon was done with financial value optimization in mind. A convertible bond doesn’t provide access to a right to work or service, does it? (ANJ does). My understanding is that we’re trying to optimize utility and UX for stakeholders of Aragon.

2 Likes

A work token does differentiate from a convertible bond. I don’t understand how creating multiple classes of tokens would be better UX then using ANT for the juror pool?

1 Like

This has been extensively debated on this thread although I don’t think a consensus was reached on the question.

Some argue that creating different tokens will make the system more complex and create confusion for the users. Others support that this will make each sub-system easier to build, manage and optimize for a given use-case.

2 Likes