This post is kept up to date here.


  • We are in the growth stage of DeFi where new protocols are showering participants, and themselves, with governance tokens.
  • Rewards from one protocol are re-invested in another protocol to earn more rewards, and so on. The system is multi-layered.
  • The value of these tokens is often purely a promise of a great DeFi future, with no underlying income attributable to governance tokens.
  • DeFi will have peaks and troughs. When we hit a trough, sentiments on the great DeFi future and the value of these tokens will plummet, at least temporarily.
  • The rational/cynical approach to this system is to offer a platform to invest in these tokens and take a profit share rather than participating oneself. This is selling shovels rather than searching for gold. This is platforms like and .
  • Now, however, the shovel makers are themselves being lured into complicating their own governance tokens. Harvest.Finance just passed a governance proposal to this effect.
  • Those that ditch their shovels in search of gold may pound sand.

Anatomy of a DeFi Protocol

Step one in a DeFi protocol is the governance token, which is typically used for voting on governance matters. The supply of the token is some times fixed, sometimes not, and it is often distributed to some or all of the following:

  • Founders of the protocol - upfront or on an ongoing basis.
  • Developers/employees.
  • Participants - for example, you earn COMP tokens if you lend or borrow crypto on the Compound platform.
  • Current holders of tokens - yes, this is a funny one - often existing holders of tokens receive more tokens. This doesn't make a lot of sense in many cases - more discussion on why below.

As a note, many platforms have no explicit income attributable to the governance token. For example:

  • COMP has no income whatsoever, just the right for governance to change the rules in the future.
  • Uniswap (and I like and own Uniswap) has no income to the token at present, but can turn on a 0.05% transaction fee if governance decides.

Just two notes on terminology:

  • Tokens that are issued at the start of the protocol are often called a "pre-mint".
  • Tokens that are issued over time are often called "emissions".

Anatomy of a Gold Rush & Multilayering

This is best explained by an example. Let's look at the USDC strategy in

Ok, what's happening here? So:

  1. You invest USDC on Harvest, and they give you some FARM tokens for that.
  2. Harvest takes the USDC and invests it in Idle. Harvest gets some IDLE for doing that. Harvest sells the IDLE for USDC and reinvests that for you (this is called auto-compounding).
  3. Now IDLE takes that USDC and invests it in Compound, earning you some COMP. Farm also sells that COMP for USDC and autocompounds it for you.
  4. Compound now lends out your USDC and you get some interest. The interest is really an afterthought. Let's see why by looking at the breakdown of the 61.65% yield you get on the USDC pool in Harvest:
  • FARM  tokens provide 9.86% of the yield.
  • If you look on the Compound app, you can see interest on USDC is 10.8% at the time of writing.
  • This means that of the 51.79% of the yield provided by IDLE and COMP tokens, 10.8% is provided by interest, and 40.99% is provided by earning the IDLE and COMP tokens.

So of the 61.65% yield, the vast majority of that is earned from selling off governance tokens earned as rewards for participating in platforms.

How to Sell Shovels

Taking FARM as the example, the way to sell shovels is to offer investing/auto-compounding services - as described above - and to take a cut of the profits. Thus, you can get the upside, but not have to put collateral at risk.

Harvest.Finance does this as follows:

  • Harvest allows you to invest your own funds (e.g. USDC) at your own risk via the Harvest platform.
  • By pooling your funds with others, you save on the transaction fees that are very high on the Ethereum network right now.
  • In return, you give Harvest.Finance a 30% share of your profits.

Harvest gets paid the profits in the form of all kinds of tokens (COMP, IDLE, CRV - whatever the multitude of strategies is generating for investors in reward tokens). As a shovel seller, Harvest doesn't want to hold on to these tokens - that would be risky - it wants to sell them off and pass the returns on to the owners of the platform, i.e. the holders of the FARM governance token. Now, there are a few ways to do this:

  1. Directly send the earned tokens (COMP, IDLE etc.) to each FARM token holder. This is messy, FARM holders would then have to sell the tokens themselves, which would cost a lot in transaction fees.
  2. Harvest swaps all rewards tokens for a single crypto like USDC and distributes that crypto to FARM token holders. This is a better idea than idea #1. It would be cheaper (transaction wise) for Harvest to convert the reward tokens in bulk than to put that burden on individual FARM holders. The USDC would be ordinary income to FARM holders, so there would be a tax consequence, but at least it would be denominated in dollars and reasonably easy to track.
  3. Harvest uses the profit share tokens earned (COMP, IDLE etc.) to buy back FARM governance tokens - driving up the market value of FARM, which has a limited maximum supply. This is currently what Harvest does.

To recap, Harvest autocompounds interest and rewards for investors on its platform. In return, it takes a 30% share of profits. Harvest takes those rewards tokens and uses them to buy back FARM tokens, driving up the price of its governance token. It is analogous to a company like Apple buying back its stock with profits.

A note on FARM Emissions as Incentives

While FARM buys back its tokens with profits, it also prints tokens to reward developers and incentivise investors to use its strategies. (If you go back to the example above, you can see that the USDC strategy earns FARM tokens as rewards.) I'm not a huge fan of this because it means that people who invest in FARM with USDC are typically selling off any FARM tokens they earn and this comes at a cost to FARM holders. However, I see the argument that FARM emissions effectively serve as a discount on the 30% profitshare fee and plays a role in incentivising early adopters.

Losing sight of shovels

Where I think Harvest loses sight of selling shovels is by issuing new FARM tokens to existing FARM holders who have "staked" their FARM tokens in the platform. Staking just means you lock your FARM into a contract on that allows you to vote (although you don't have to vote) and allows you to earn these newly issued FARM tokens. Of course, if you're not staked you lose out on these emissions, so nearly everyone who owns FARM also stakes their FARM. (The staking percentage is always very high - 75%+.)

Providing FARM emissions to existing holders has no net effect other than costing transaction fees and imposing income tax costs and tracking burdens on FARM holders:

  1. Let's say I own 1% of all outstanding FARM tokens. This entitles me to 1% of the profitshare earned by Harvest (distributed via the buyback mechanism).
  2. By issuing new FARM shares to all FARM holders on a pro rata basis nothing changes in terms of what I own. I still own 1% of outstanding FARM tokens and I'm still entitled to 1% of the profitshare!

However, emissions cause the following problems:

  1. Emissions cost gas (transaction fees) for the platform.
  2. Emissions are income for FARM holders. FARM holders have to pay ordinary income tax on this income and keep track of the emissions (which means tracking every receipt of FARM and the price of FARM at that time).
  3. By requiring FARM holders to stake in order to earn emissions, they have to pay gas in order to stake. If all FARM tokens were allowed to vote and emissions were abolished then FARMers would save the costs of having to stake. In any case, the vast majority FARM tokens are staked, which makes staking rather pointless but costly.

There is the criticism that you can't show buybacks as an APY in the same way as emissions, but you can. If the platform is buying back 10 FARM per month, and the outstanding level is 1000 tokens, you can calculate a rate of return to show online.

Really losing sight of shovels

Now that gas costs are insanely high, FARM holders understandably don't want to have to pay the high fees for staking and unstaking FARM tokens. They also want to use their FARM tokens to lend out on other platforms (which I think is risky) but currently would have to unstake their FARM to do this and lose out on emissions.

This has led to a request for staked FARM to be tokenized as iFARM. In other words, the recently passed governance proposal calls for the creation of iFARM which is a staked version of FARM that will be tradeable.

In my opinion, this is trying to use a small plaster (Bandaid) to stem a large bleed. There is no benefit to requiring FARM to be staked, in fact it is costing FARMers gas because we are forcing them to stake and it means FARM holders can't use the tokens on other platforms without unstaking. Rather than simply dropping the requirement for staking and dropping emissions to FARM holders, the platform is deciding tokenize the problem.

Now, in addition to staking costs and tax complexities being firmly embedded into the Harvest platform there is the additional problem of having both iFARM and FARM tokens which will:

  1. Make the Harvest software more complicated, i.e. add to technical debt.
  2. Split Uniswap liquidity between two tokens rather than one.
  3. Make the case for having FARM approved on a lending platform like Aave more complicated.

In Conclusion.

FARM is the shovel not the gold-digger. Let's keep the platform simple while other platforms are busy layering.

Disclosure: I own staked FARM, staked Aave, Uniswap, locked Curve, locked Celo and BTC. You can find my DeFi mini-course and hear about my investment approach at .