What kind of tokens are we talking about?

We will mainly discuss tokens and cryptocurrencies playing a determining role in the state of the network it’s linked to, whether by being essential to the functioning of the consensus mechanism, necessary for on-chain governance or enabling actors to work towards incentive alignment. Payment or discount tokens and security tokens have generally less effect on the underlying network.

What is considered a fair and efficient initial distribution?

Since fairness is relative to the scope and scale of what it’s about, let’s not go into too much detail and only consider a few essential points. In the current context, it is important to associate fairness with the notion of efficiency.

A fair distribution should ensure that:

• a reasonable amount of people hold the token (that number proportionally increasing with the market cap of the project)
• the biggest holders do not have possession of too large a majority of tokens
• there are no large price discrepancies to acquire the tokens

An efficient distribution should imply a distribution of the tokens to participants:

• who are involved and are ready to potentially produce work for the project
• with different profiles (speculators, investors, small token holders)
• who understand and are aligned with the incentive mechanisms that the token powers

To what scope does it apply?

Having an optimized distribution from the launch of the network is a tremendous advantage to start building resilient decentralization for the long run. However, the importance and effects that the decentralization of a token can have to depend on the role that it plays in the network. Indeed, token distribution does not necessarily affect the safety or the performance of a blockchain.

In proof-of-work (POW) systems, hash power, which is an off-chain parameter, is often considered as the relevant metric when assessing the health of the chain. As a result, a large concentration of BTCs in the hands of a few specific actors doesn’t necessarily threaten the resilience or performance of the network. One could argue that having large amounts of a POW network tokens would provide the holder with the power to manipulate prices. Since price is often correlated with the hash rate, large holders could have a remote influence on protocol performance. For the purpose of this analysis, this scenario won’t be considered. Even though legacy wallets today own a few percents of all the BTCs (totaling over 1 billion in USD value at current rates) and would destroy the price if they were all dumped on the markets at once, it’s a very unlikely scenario.

The disconnect between the participants who technically support the network and the distribution of the tokens can represent a powerful advantage for the miners, though it can quickly become a liability for the network. Mining facilities business model supports the cost to assign large computing resources to run the POW algorithms and the electricity to run them for a certain time, in order to acquire the tokens of the network, and resell those hoping for a higher price than the costs incurred. By not having a stake in the network, mining cartels attacks are a real threat to any POW network. The larger mining pools never risk more than the cost of electricity and the price of their mining setups diminishing.

In a proof-of-stake (POS) system, the initial holders who receive coins at the network launch obviously don’t have any incentive to redistribute their coins and make the network more decentralized since it is directly correlated to their wealth and ability to generate more wealth through staking. We shouldn’t rely on a natural redistribution from big holders of their staking power to different actors in the network. This would involve us to consider their bias and convictions, which may not be the network’s health, and wouldn’t support adversarial views.

A great piece by Vitalik from early 2017 goes into depth on the meaning of decentralization and the reasons why it matters.

Why is decentralization important?

Decentralization can actually be quantified and represented by different coefficients. This very detailed piece from Balaji S. Srinivasan and Leland Leeintroduces the minimum Nakamoto coefficient, based on the Gini coefficient and Lorenz curve, and quantifies the minimum number of entities required to compromise a decentralized system.

The importance of a systems decentralization can be described by the following elements:

• Network security and resilience against Sybil attacks: A blockchain is only as secure as the cost of performing a 51% attack on the network. In POS systems, being able to perform such an attack would require attackers to effectively purchase enough tokens, resulting in them holding a very high stake in the network. When a token is properly distributed amongst many stakeholders, it becomes far more difficult for an attacker to gather that critical mass. Moreover, the process of acquiring such a stake will inevitably drive the price up as demand grows, making it more difficult to accumulate. The more decentralized the holding of tokens is, the harder it becomes from an external attacker to get to this critical mass. Certain POS blockchains implementing on-chain governance have mechanisms to expel such an actor from the validator set, effectively punishing it by destroying its entire token holdings, but this goes beyond the scope of this article.
• Long term sustainability: The more decentralized a network is from the early days, the harder it will be for attackers to acquire the necessary weight to influence network decisions, for example through the governance of price manipulation. As the network gains traction, users improve its technology, and the price of the token, when crypto economics is properly designed, should follow that growth.
• Incentive alignment: A well-designed network must compel every actor involved to act in their own interests and correlate this competition (in hash power, total stake, amount of delegation, size of masternode, the security of validator…) with an enhancement of network performance. When a few actors share too large a portion of the total token supply, the weight they potentially have on many aspects of the network is disproportionate. Even though one might argue that the financial incentives built at the core of a network are not sustainable enough to be the only thing holding the network together in the long run, they are necessary and efficient mechanisms, ensuring every actor has skin in the “same” game and fostering rational economic behaviors, where everyone is accountable for their mistakes.
• Governance: On-chain governance can tie the voting weight of a player to the relative amount of tokens they are holding, or transfer that voting power to the entity they are delegating their tokens to. A wider distribution of tokens leverages the wisdom of a larger number of people regarding the ledger amendment.
• Token categorization: According to the SEC, a token is considered security until it is “decentralized” enough. Ethereum can be used as an example of this. Ether tokens once considered security-ish is now mostly sorted in another category of tokens by different regulators? A good piece on this subject was in The Block last December:

Decentralization is an important parameter that reflects many aspects of the health of a network. However, it is not an end in itself. A perfectly even distribution between all network participants with efficient incentive alignment alone does not ensure the performance of the underlying technology.

Now, let’s dive into a few examples of how tokens have been distributed in the past

• Bitcoin: Since the genesis block on January 3, 2009, BTCs have been distributed as a reward for mining blocks thanks to its POW algorithm. The reward miners receive for mining a block (excluding transaction fees) started at 50 BTC and halves every 210,000 blocks. The block reward is how new bitcoin is “minted” or brought into the economy until all 21 million BTCs are eventually minted in 2140: after that, the rewards shall solely be constituted of transaction fees. The bitcoin distribution model can be considered fair as the earliest contributors received better rewards as the mining competition was low and the network needed to be bootstrapped.
• Ethereum: Ethereum was one of the earliest ICOs. A significant portion of total ether was effectively “premined” to be sold to investors during a token sale. The team dedicated a large amount of these premined tokens to be vested and redistributed to the team and advisors, as well as to different entities responsible for handling the future development of the project. After the initial distribution of tokens, a POW mining reward distribution model similar to bitcoin took over and is still active today. The Ethereum Foundation and other entities involved in the development of the project are planning to eventually switch to a POS model after 6–7 years of existence.
  The process of sorting out the participants involved in the network through POW, before switching to POS is giving a sufficiently long time to participants to figure out the role they want to play in the network. As Ethereum’s network effect remains one of the strongest in the blockchain space, it is very likely that the remaining network participants, after the switch from POW to POS has occurred, will be highly involved in the general health of the blockchain.
• ZCash / Beam MW:
  Zcash’s monetary base is similar to Bitcoin, with a maximum of 21 million coins mined over time through POW, with the exception that 10% of all rewards, called the “Founder’s Reward”, are rerouted to the stakeholders in the ZCash company — founders, investors, employees, and advisors. This allows for long term alignment of the actors involved in the health of the chain.
  BEAM MW recently opted for a similar model where 20% of mined coins in the first year and 10% for the following 4 years are deposited directly into Beam’s treasury.
  While these solutions can be criticized for fair reasons, it is still an interesting way to keep the development team involved in the project for the long term.
• Grin adopted a linear emission rate of 60 GRIN per block, with 1 block being mined every minute via POW. Grin does not have a maximum supply, and the high inflation rate of the first 10 years will change to low inflation rates in the future. This choice was made following the ideology that the MimbleWimble protocol outlined, which was to become sound money with a strong store of value property while being a medium of exchange.
• Many 2017 ICOs largely underestimated the importance of token design and distribution, resulting in absurd distributions of tokens today, and sometimes effectively unusable networks, where control is divided among only a few actors, not always committed to the long term sustainability of the project.

Innovative ways to distribute tokens to participants that matter

In the past couple of years, some projects strongly highlighted the importance of initial token distribution and made a significant effort to find innovative ways to initially distribute tokens to participants who are genuinely involved in the network.

The challenge at stake is to create efficient distribution mechanisms that would put the tokens in the hands of people willing to produce work for the network (through staking or other active network participation mechanisms), to use the product as intended, to actively promote the project, and more.

Let’s highlight a few of these innovative solutions

Edgeware’s Lockdrop

Edgeware is a smart-contract blockchain built on Polkadot. They introduced the notion of Lockdrop: “Ether holders are able to lock their tokens for as short as 3 months or as long as one year, with longer timelocks corresponding to receiving proportionally more Edgeware tokens.”
More than a smart-contract platform, Edgeware wishes to experiment with on-chain governance in a network with high usage and effective incentive mechanisms.

The only way to initially get EDG tokens is to lock Ether in the Ethereum smart-contract developed by the team. It enforces people that are genuinely interested in the project and would want to have an impact on the decisions affecting the chain early-on to renounce the opportunity costs of ETH during the duration of the locking period. Opportunity cost can be calculated with the different lending services, compounding the tokens… At the end of the lockup period, participants end up owing both EDG tokens and the ETH they had locked.

This mechanism efficiently incentivizes actors to get involved in the health of the network without asking for direct investment. The team counts on the fact that the token distribution at the end of the locking period will represent a strong base of participants that can steer the project in the best direction in the long run.

On the flip-side, this could leave space for a competitors chain with large ETH holdings to take a large position in lockup contracts and eventually have a significant influence over the network. If hypothetical pos-chain-xyz has tens of thousand of ETH in reserve and Edgeware is a direct competitor to pos-chain-xyz, pos-chain-xyz could decide to lock thousands of ETH for 3 months and end up with a very large amount EDG tokens proportionally to the rest of the network.
Another caveat, however unlikely, is the possibility of Ethereum undergoing a clivant fork during the middle of the lockup period.

Numerai’s targeted distribution

Numerai is a good example of a project that realized the current distribution of their tokens implies high centralization and control over the network. They initially didn’t do any ICO or token sale, but rather minted and distributed their tokens to actors they believed would be the bedrock of their value proposition: the distributed network of 19,000 data scientists that contribute work and help improve the calculations it uses to make real bets on the stock market.

After a couple of years, the Numerai team decided to cut token supply by 10 Million to become ‘Decentralized as F*ck’

Livepeer’s Merkle-Mine

Broken down in this article by Kyle Samani from Multicoin, Livepeer came up with a very innovative solution to distribute their tokens after noticing the inefficiency of airdrops. Livepeer’s token LPT is a work token which you are required to stake in order to produce work that is then rewarded. This work can be of different natures — in Livepeer’s case, transcoding. In order to distribute LPT tokens to participants with the technical capabilities to produce this work, Livepeer introduced the Merkle mine, a permissionless and decentralized process complicated enough to sort out participants with relevant technical skills — “the Merkle mine requires participants to generate a cryptographic proof and to send that proof to a smart contract in exchange for LPT”.

NuCypher’s Worklock

Similar to Livepeer, the NuCypher’s project introduced a distribution method called Worklock where a participant looking to receive NU tokens needs to lock ETH in an escrow smart-contract which then distributes NU tokens proportionally to the amount of ETH locked. The ETH is released after the participant performs work using the NU tokens. This effectively incentivizes participants to use the NU token for its intended purpose, here running a staking node on the NuCypher network.

FOAM’s Proof of use

FOAM’s goal is to offer a secure decentralized location serves as an alternative to existing centralized GPS solutions.

In their token sale, every participant is “expected to contribute as Cartographers through the static proof of location functionality or signaling mechanisms to prove use” to effectively unlock the tokens bought. Additionally, purchasers had to properly register to take part in the sale, ensuring that the tokens were sold to participants that intend to use them to curate the network map and contribute to the long-term development of the FOAM protocol.