How should a protocol like Uniswap or Compound structure its treasury? This post covers the principles of protocol treasury management, fiscal policy, asset allocation, ways to diversify the treasury, and responsibilities of a treasury committee.
Principles of treasury management
Infinite time horizon
Inflows should exceed outflows
Diversified asset allocation
A protocol treasury could have a different approach: be dissolved after a certain number of years like the Gates Foundation. Arguably, dissolving a treasury is a sign of even greater decentralization than having a community treasury. It means that the protocol is self-sustaining and does not depend on a single treasury.
But having a diversified treasury that exists in perpetuity can be effective, especially in bear markets. If markets are down 40% but you still need to solve an important technical problem, a diversified treasury can fund that without relying on broader public goods funding. I lean towards structuring a treasury with an infinite time horizon.
The annual budget for a treasury should be decided based on the protocol’s needs and its treasury size. Ideally, a protocol’s annual budget should be less than its expected revenue (e.g. Uniswap fee revenue) and treasury investment income.
One of the most valuable things a protocol can fund is ecosystem development, which can eventually serve as a competitive edge. Pilot programs by Uniswap and Compound have proposed $750,000 per quarter in grants that might be reduced or expanded based on their effectiveness.
As the recent Compound proposal states, the main needs of a protocol are:
It is risky for protocols to only have their own governance tokens in their treasury. If Uniswap has only UNI in its treasury and the token price falls by 40%, its balance sheet would also shrink by 40%. This is dangerous because it occurs at a time when the protocol needs to ramp up spending, whether it is LP subsidies, token buybacks, or core development. This could lead to a further downward spiral for the token.
A protocol's treasury should be diversified over time. Diversification should be planned based on the treasury’s expected annual outflows, liquidity needs, and risk-return of assets.
Here is a sample allocation (without knowing the numbers above):
Protocols can also allocate a portion of the treasury for token buybacks when the token is highly undervalued and there aren't other effective ways to deploy capital.
There can be ways to diversify a treasury that are trust-minimized and does not require a group of people to have multisig access. Protocols can leverage the existing on-chain asset management stack:
How can protocols diversify their treasuries?
The simplest method to diversify is to sell some of the protocol's tokens to buy other assets (e.g. ETH, DAI, etc.). Treasuries can’t be diversified overnight. But plans can be made to reach a steady-state allocation over several years.
There is currently a negative perception associated with selling the protocol's token because it could signal that the treasury has lost confidence in the protocol. This has a short-term impact on price. Selling large amounts of the protocol’s tokens would involve slippage, especially if it’s not done thoughtfully. However, this can be avoided if the treasury has a stated target policy allocation and dollar cost averages into other assets over several months or years.
The treasury could also receive other crypto assets as inflows. Few methods include:
Index Coop is a good case study. They recently published a thoughtful treasury report and made a proposal to purchase ETH with INDEX and DPI in its treasury. Index aims to implement a smart treasury at Balancer with INDEX / ETH at 80/20.
Protocols should set up a 5-7 member treasury management committee, which will work on the following:
The treasury committee can be voted on and removed by token holders.
We will increasingly see specialization emerge for protocols. Token holders will approve big decisions and elect or nominate managerial stakeholders with expertise (individuals, DAOs). Managerial stakeholders will make smaller, more frequent decisions and be held accountable by token holders.
We will have several experiments to figure out the right balance between managerial effectiveness and community control. In both cases, clear processes and useful transparency is critical.
Protocol treasuries need to be structured with an infinite time horizon. This requires thinking effectively about: