R.I.P. Liquidity Mining: Abridged

After the two infamous installments of R.I.P. Liquidity Mining, the purpose of this final edition is to distill the data & information from both works down into the definitive scripture of the life and unforeseen demise of our dearly departed, Liquidity Mining, and to articulate the next chapter of economic models in the book of DeFi, brought to you by Tapioca.


Liquidity Mining’s humble beginnings can be traced back to October 1st, 2017, with IDEX. IDEX rewarded IDEX tokens to users who filled limit orders (provided liquidity). Shortly thereafter in 2018, the centralized exchange, FCoin, debuted “Transaction Fee Mining”, which awarded FT tokens to traders as an 80% reimbursement for its BTC/ETH trading fees, spiking it’s volume by 7000%- surpassing Binance, Huobi, and OKex combined in trading volume, until…

The First FT Token That Went to Zero
The First FT Token That Went to Zero

After failed beginnings, liquidity mining was formally brought to prominence with the 2019 stakingrewards.sol contract, created by Synthetix through a grant accepted by none other than Anton Bukov, now known as the founder of 1inch. The contract was used to incentivize Synthetix’s Uniswap V1 pool of sETH/ETH through inflating the supply of SNX to fund the rewards. From 2020-2021, Synthetix paid out half a billion dollars in free SNX rewards (source: tokenterminal). Many however place the true beginnings of Liquidity Mining to May of 2020, with Compound’s infamous COMP Liquidity Mining Program. The Liquidity Mining program skyrocketed Compound’s TVL by 600%, and is thought of as the spark which ignited 2020’s DeFi Summer. Both Compound and Synthetix have both since tried unsuccessfully to end their Liquidity Mining program(s).

Following Compound’s meteoric “success”, we saw the rise of many new derivatives of Liquidity Mining, such as pool1/pool2 staking, being rocketed into the mainstream of DeFi with the viral Yam Finance. “pool1 staking” is otherwise known as the practice of staking (depositing) governance tokens (or LP tokens in pool2) for more of said governance tokens, proliferating the term “useless governance token”. Yam captured $750m in TVL in August of 2020 with its yield farms mixed with a splash of Ampleforth rebase mechanics, used to acquire yCRV. SushiSwap then created the famous “MasterChef.sol” contract as a vampire attack against the top AMM, Uniswap, reaching $1.2 billion in TVL. The MasterChef contract became one of the most forked contracts in history, being used to distribute token rewards to liquidity providers. Finally, Olympus combined Yam’s borrowing of Ampleforth’s rebasing mechanics with its own bonding model, a constant sale of discounted OHM tokens for stablecoins & LP tokens, as the crescendo to the (totally not) negative sum yield farming extravaganza of DeFi Summer.


The basis of the DeFi protocol economic model since it’s emergence has been predicated on liquidity mining- giving away free tokens as rewards to incentivize liquidity providers to deposit their capital in order for the protocol to capture TVL.

Let’s look at two critical statistics.

“A whopping 42% of users that deposit capital into a liquidity pool with liquidity mining incentives on the day it launches, exit within 24 hours. 16% leave within 48 hours, and by the third day, 70% of users withdraw their funds from the contract.” - Nansen

The profitability of many blue chip DeFi protocols as of July 2022 - Bankless.
The profitability of many blue chip DeFi protocols as of July 2022 - Bankless.

As you can see from the graphic above, DeFi protocols which employ liquidity mining are deeply unprofitable when we look at a protocol as a business, its fees as its revenue, and incentives as expenses. Liquidity Providers (LP’s) know protocols employing liquidity mining are operating at monstrous losses, even when compared to the plight of a normal tech start-up, and therefore also know that the value available to be “farmed”- predicated on the value of the token emissions- is fleeting. This is the reason why LP’s are observed by Nansen in such high percentages following the same ultra short-term script of: deposit > farm & dump > withdraw to find the next liquidity mining geyser to continue the never ending negative sum harvest.

Liquidity Mining’s goal is to capture TVL. What is TVL, anyway? TVL, or “Total Value Locked”- a metric created by DeFiPulse in 2019, is the capital currently deposited into a protocol’s smart contracts. Two questions: Is this liquidity actually locked? No. If we all put $1 million dollars into a Cardboard Box temporarily, does that mean the Box has $1m in value? Not really, the $1m is really a liability, as it’s owed to us. Let’s also take for example- if I offered you a million dollars for any amount of time I please- could be 12 seconds (one Ethereum block) or 10 years, versus $100,000 for one year guaranteed, which would you select? Time is Money, and without considering capital custody time with incentive models, we’re also thereby not able to capture real and estimable value, which is how yield farmers are not only enabled but incentivized to obtain something for nothing.

To hammer home the point- let’s say we open a coffee shop next to Starbucks together, charge the same amount per cup of coffee ($10), but then attach gift cards (token incentives) to every cup, enabling customers to come back to claim $100 in free coffee (liquidity mining), what would happen? Shitcoin Coffee would be the most successful coffee shop in existence temporarily, and catastrophically fail in a longer time frame, sound familiar? What would the customers do? Buy the first cup of coffee purposefully (deposit liquidity) just to obtain the gift card (token incentives), so that they could redeem them and drink for free, or sell the gift cards to others (yield farm), and never come back (withdraw) once the Gift Card’s funds run out (analogous to the token reward becoming worthless).

Game Theory wise, yield farming is actually the strictly dominant strategy in liquidity mining, creating, and most importantly incentivizing, a negative feedback loop which paradoxically incentivizes the extremely deleterious and parasitic strategy (farm & dump) as the most beneficial & logical strategy to be employed. Once the token has lost enough value to users employing the dominant strategy, the protocol can no longer inflate forever to rent liquidity, as the protocol becomes economically disabled with a worthless reward token. Giving away something permanently for free (tokens), for nearly nothing temporarily (rented liquidity) doesn’t work, no matter what package the Sisyphean task is bundled in- vested emissions, celsius style membership rewards, shifting the onus of incentives around (bribing), etc.

Call Option Incentive =/ Solution

“To get value, you must give value”.

Even though many looked at Olympus as the peak of insanity in DeFi Summer, it had an extremely powerful innovation made possible by Olympus’s bonding mechanism: Protocol Owned Liquidity (POL). The problem with Liquidity Mining & TVL boils down to the following piece of data: While Compound’s TVL increased by 600% with its liquidity mining program, Compound only captured $45m in fees with its $400m in emissions (expenses) (source: tokenterminal) from 2020-2021- nearly $10 in expenditures (emissions) for every $1.00 in revenues (fees). TVL has nearly zero value as it’s liquidity rented (liabilities) from the mafia (yield farmers), while POL’s value is indisputable- Liquidity is King. The advent of POL was the dawn of what many called “DeFi 2.0” (cringe), with headliners such as TribeDAO (FEI), Tokemak, and Frax. Generally in “DeFi 2.0", instead of employing liquidity mining, liquidity providers would receive an incentive that allowed them to purchase market discounted governance tokens. If Compound’s nearly $400 million dollars in COMP emissions from 2020-2021 were instead COMP call options or bonds, bearing an average of a 50% discount to market price, Compound would have brought in $200 million dollars in POL- seeding the largest Treasury in DeFi by a factor of two. Olympus still stands as one of the largest Treasuries in DeFi today (source: defillama).

While that is obviously a marked improvement, it does not address the parasitic nature of yield farming, nor many other issues. Call Options & Bonds do have many intrinsic benefits, and a few of their own problems (namely UX), but most importantly alone do not solve the three largest problems needed to create a self perpetuating and sustainable economic model. Tapioca’s economic model is much more than “call option incentives”.

The three requisites we posited were required to create our desired economic model are:

  1. Firstly, employing a sustainable incentive model which maximizes economic growth while minimizing the negative effects of the growth stimulus itself.

  2. The protocol’s (business’s) economy needs to promote real value creation (Fees & POL) to enable self-perpetuation (true flywheel) which reduces dependence on external LPs, and maintain profitable unit economics, which I define as a protocol having less expenditures (token incentive expenses) than balance sheet inflows (POL capture) and revenue (fees)- which is the basis of the real yield movement pioneered by GMX.

  3. Lastly, the model must employ game theory in a manner that incentivizes a well-defined positive feedback loop. DeFi is nothing more than an MMO Economic Wargame, and the protocol must target & reward what it deems to be the most beneficial strategies (generally longer term value provision), and deter the employment of the most deleterious strategies (generally mercenary liquidity provision, or parasitic yield farming).

    Let's not forget, DeFi = Yield, and DeFi has been in a functionless state for almost two years due to an incentive problem, or lack of attractive supply rates for LP’s to be incentivized to provide capital, especially where short term U.S. Treasury Bills yield multiplicatively over Aave or many other major protocols. Secondarily, decentralized CDP stablecoins have had an inability to scale and/or to enforce their price alignment to the U.S. Dollar, especially without becoming a “USDC Wrapper” (backing itself with USDC/USDT, losing censorship resistance), such as with Aave’s depegged GHO stablecoin, due to a dire supply side incentive misalignment. To address this, decentralized stablecoin issuers like Maker (DAI) have been chewing through protocol profits to incentivize DAI demand & supply side incentives via the Enhanced DAI Savings Rate (eDSR).

Source: makerburn.com
Source: makerburn.com

These equally dire problems that Tapioca’s success depends with it’s USDO stablecoin & lending markets both boil down to an incentive problem. Tapioca was designed from the ground up to meet the three aforementioned requirements of a self-perpetuating value capture focused economic model with DAO Share Options (DSO) ft. oTAP (call option incentive), Time Weighted Average Magnitude Lock (twAML), while being targeted on POL & “real TVL”, or truly locked value.

The Actual Solution

Game Theory is incredibly important to economics, as we can determine potential outcomes of economic models via backward induction through the perspective it affords. For example, Liquidity Mining creates a strictly dominant strategy that incentivizes all players (users) to adopt a parasitic yield farming strategy, while Olympus’s memetic “3,3” creates an infinitely repeating Prisoner’s Dilemma. Olympus through backward induction loses subgame perfect equilibrium, a theory created by John Nash that calls for no player having an incentive to change their strategy, which is why users became incentivized to -6 (Sell) instead of continuing to +6 (Stake). Tapioca’s economic model is based around Rubenstein’s Bargaining Model, in which players (users) make alternating offers of lock time of TAP or lent USDO to the protocol to receive rewards (twTAP output ratios & call option discount levels, respectively). If players stop negotiating (locking), the AML (Average Magnitude Lock), a user controlled measuring stick that quantifies the value of a proposed lock against the current consensus of economic activity (current locks) in the Tapioca protocol, begins decaying until users begin locking again. Thus, the higher the AML is, the more open locks there are, and the higher the lock time will be required to get larger rewards, and the lower the AML, less open lock positions, the lower the lock time will be required to get larger rewards. Only users with sufficient offers (at least 0.10%* of total platform liquidity) affects the AML measuring stick. The oTAP discount level to TAP market price goes from 0% to 50%, and the twTAP output for an input of TAP + Time Weight (lock duration) goes from 0% to 100%. twTAP abolishes Curve’s vote escrow (ve) fixed “loyalty tiers'' (100 CRV locked for 1 year = 25 veCRV, 100 CRV locked for 4 years = 100 veCRV) in favor of dynamic output ratios to constantly ensure the minimum reward needed to incentivize new locks is met. twTAP ensures the circulating TAP supply is constantly diminishing, being the only way for users to capture a pro-rata share of Tapioca’s ecosystem fees, paid in USDO. Curve’s gauge model was also employed for twTAP lockers to obtain the ability to direct the flow of oTAP incentives to their desired lending market(s).

Sustainability is aided via call option incentives, first employed by Andre Cronje with Keep3r Network, which I opine are the perfect incentive vehicle from a stimulus standpoint as they are a right to purchase an asset at a specific price (strike price), within a specific time frame (expiry). The lowest strike price = the price floor, and these options expire each weekly epoch. The oTAP options can go “OTM” (out of the money, or unprofitable to exercise) if the strike price of the users oTAP goes below the market price of TAP, and/or simply expire, capping the deleterious effects of the stimulus. Liquidity Mining on the other hand can be thought of as an infinite expiry, $0.00 strike price call option. The unredeemed TAP from each epoch rolls over to the next- extending the incentive program well past the six years it would take to emit the majority of the TAP available to DSO per it’s predefined design- therefore, the circulating TAP supply doesn’t inflate unless the option(s) are exercised, and the incentive program can constantly extend itself. The overall global number of oTAP shares available in each weekly epoch are implemented along an exponential decay curve, which means the potential inflation of TAP is in constant decline. Lastly, TAP has a fixed supply of 100,000,000, bringing an end to DeFi protocols infinite inflation, as Tapioca no longer needs to retain rented liquidity (TVL). twAML, however, is the key to sustainability, as it enables Tapioca to constantly ensure the monetary reward (twTAP output ratio & oTAP discount level) are always in line with the minimum requirement of the consensus of economic actors to continue economic growth (snowballing TAP locks and locked lending positions) at the lowest cost basis to the protocol. Lastly, codename “FL” is planned to establish on-demand liquidity at TAP’s price floor, further defending the floor, heightening POL capture, and enabling the Tapioca DAO to reclaim TAP from the market at a near positive cost basis. If a user doesn’t want to commit to a locked lending position or a locked TAP position, they will only receive real yields (or have no share of protocol revenue or governance in the case of twTAP lockers), as temporary liquidity (TVL) isn’t valued by the protocol.

Profitable Unit Economics are met by Tapioca’s economic model from the fact that the free market controls the value of time via the AML, reaching allocative efficiency of incentives- or incentives that are distributed in equilibrium with user requirements and retain the maximum benefit to the producer (protocol) and users. Tapioca’s focus on capturing “Real TVL” - actually locked value, targets the stimulus at creating a positive feedback loop of capturing real revenues (fees) & balance sheet inflows (POL) with expenditures (incentives). This is because the free market controls the AML, and thus the reward size is always in harmony with the minimum economic stimulus required to continue economic growth. Capturing locked capital which can be extracted for fees will create a calculable cost basis of revenue (fees) for the Tapioca ecosystem:

Revenue [Average Fees Per Lock] - (Expenditures [Average oTAP Per Lock] - Balance Sheet Inflow [Average POL Captured Per Lock]) = Cost Basis.

The most important component to Tapioca’s unit economics is the POL captured by every redeemed oTAP incentive. The need for external liquidity will linearly diminish, especially important in the case of USDO LP on AMM’s, giving the DAO enormous control over USDO’s programmatic alignment to the U.S. Dollar. Thereby this is why TAP inflation was also designed to linearly diminish, decreasing expenditures while revenues stay constant via POL replacing external liquidity. Tapioca’s POL will also begin to grow in value, and create more and more fees itself, which will increase revenue.

Death begets Salvation- Genesis 022424



The Original Installments:

Subscribe to TapiocaDAO
Receive the latest updates directly to your inbox.
Mint this entry as an NFT to add it to your collection.
This entry has been permanently stored onchain and signed by its creator.