I went for a deep dive into how DeFi protocols/organisations interact with legal compliance and I came across an interesting Medium article written by @lex_node [2020] “Defining Decentralization for Law” where he proposed both a more flexible standard and a bright-line safe harbour for assessing whether a blockchain system is decentralised as a matter of law. First, he established that assuming the law applies differently to a “sufficiently decentralised” system than to a centralised system, it is imperative that lawyers, judges, regulators and others involved in the legal industry develop a common understanding of what “decentralisation” means. Otherwise, they will not know how to apply the law this the proposal: The Flexible Test (the “real” decentralisation test) The powers that centre a blockchain must be evaluated: - Validation power: Validation power refers to the ability of persons participating in or relying on the network to read and validate its data. Decentralisation of validation power requires that all relevant software code be publicly available for review and operation. However, Validation power on systems with heavier state(computationally challenging for enterprise state computers to download and validate data) should be considered to be less decentralised than on systems with lighter state, but may still be sufficiently decentralised for purposes of the law. - Consensus power:(this one is a bit tricky) Consensus power refers to the ability of persons to write data to the blockchain (who can produce blocks). Therefore, you judge decentralisation not by how blocks are produced, but by who can realistically control the means of block production. a. Block production is always permissioned and scarce. PoW → only people with specialised hardware can mine. DPoS → only elected validators can produce blocks. PoS → only people with stake can be selected. So by design, this power is concentrated. That’s normal. b. The danger is not how validation works; it’s whether a single group can dominate it. If one mining pool, a cartel of stakers, or a coordinated validator set can control block production, then consensus is not decentralised. Control doesn’t require 51%, it depends on the chain’s internal rules. EOS: control = 34% of validators (enough to block consensus). PoW: majority hashpower controls reality. c. However: concentration of consensus power is not always fatal. Even if miners/validators are concentrated, they may still act honestly because they are financially incentivised not to kill the chain. But legally? Regulators don’t care about intentions....