Hey everyone, this thread contains many good points and questions, and I hope to cover them. Discussions will roll on for a while. Big thanks to everyone who has taken the time to give their views so far.
I want to start by addressing the tax considerations, and whether it is appropriate for the design of neuron staking to be sympathetic to them.
The case of a Seed investor who thought they faced major losses
There was an engineer in California, who persuaded his mum to participate in the 2017 public “Seed” fundraiser. She had not sold anything at all. However, her CPA, who was a leading CPA in the Bay Area, told her that increases in maturity would be seen as equivalent to income by the IRS, and that whenever the maturity of her neurons had increased in the past, an income tax event had occurred. He claimed that each time maturity had increased, ordinary income was received equal to the “equivalent” number of ICP multiplied by the market price of ICP the time. Given the enormous early highs in the price of ICP, which were likely created by price manipulation on FTX in the hours before Genesis, she was told she had received a huge amount of income, even though she had received no cash receipts from sales, and owed an unaffordable amount of income tax. Consequently, she immediately began using her accumulated maturity to produce ICP, which she then liquidated on an exchange, and began liquidating her retirement accounts to cover the shortfall in funds needed to cover her assumed income tax liability. In addition, she placed all her neurons into dissolve mode to escape further financial danger as soon as possible.
I spoke to the engineer after events were well in motion. I tried explaining that in my view, it was very unlikely that increases maturity were taxable ordinary income. There were two firewalls: (1) that receipt of staking rewards would not be taxable income, since newly minted tokens created through staking were the product of collaborative creative effort, and normally, such product does not create income until sold, and (2) since maturity is only an attribute of neurons, its application is subject to the current design of the NNS, which self-updates, which means maturity cannot possibly be seen as equivalent to ICP tokens on the immutable ICP ledger.
Naturally, these arguments made informally had little effect, and they felt it necessary to heed the advice they were getting. Most holders of ICP do not have the necessary experience with tax law to draw their own conclusions. Meanwhile tax advisors hear the word “maturity” and instantly equate it to coupons on a bond, interest paid on a bank balance, or something similar. They have insufficient understanding of decentralized networks and blockchain and in the absence of external guidance draw faulty equivalences. Sophisticated tax lawyers will often draw different conclusions, but they are out of reach for most ICP holders.
Why I provided a tax analysis and proposal to harden the system
I became aware of others in similar situations located around the world. Today, on an ongoing basis, there are people who are continually dumping ICP, even though this is not their preferred course of action, because they fear being unable to pay down income tax liabilities generated by maturity. We must feel for all individuals in this position, and more broadly, this is not good for the network’s decentralized economy.
For such reasons, I realized that an analysis of the current situation should be provided, to provide those in this situation with input material that might be supplied to tax advisors and/or tax lawyers. In addition, I realized that the design should be further hardened, to make it even more difficult for any tax authority around the world to claim that increases in maturity represent ordinary income that should be immediately taxed. The post I made to propose changes, which is often referenced in this thread, can be found here.
In this thread, some are concerned that designing a tax efficient system is illegal “tax avoidance”. Thankfully they are mistaken.
Whenever financial professionals design business frameworks, their designs always take into consideration the tax consequences, and this should also be the case when decentralized protocols are designed, since a key purpose is to allow widespread participation. Considering tax consequences in design is very clearly a form of tax planning, rather than tax avoidance. Tax avoidance involves people making efforts not to pay taxes that are legally due. Tax planning, by contrast, involves designing systems that work with existing tax frameworks in order to legally protect and further the interests of participants, which tax authorities encourage. Proposing design tweaks with the objective of ensuring those staking neurons are not subject to unfair income tax liabilities, when they have received no cash compensation to pay down those liabilities, is very clearly tax planning.
The IRS has since accepted arguments made in the proposal
Please see the tax analysis provided and nature of the proposal (or just search for “Taxation Of Proof-Of-Stake Staking Rewards” in the post if you have limited time). What’s important is that the arguments involved are derived axiomatically from hard case law. As mentioned in the foregoing, there are two firewalls, which provide powerful foundations for legal argument where tax authorities unfairly claim increases in maturity represent income (and which can inform CPAs/accountants and tax lawyers advising neuron holders).
The first firewall lays out the arguments why the receipt of staking rewards, generally, should not be treated as ordinary income, and why income is only generated when tokens received are sold, exchanged or otherwise disposed of. The argument depends on accepted case law in the United States, but can be extended to nearly every tax system around the world. Thankfully, since the analysis and proposal was posted, it appears the IRS (the “Inland Revenue Service” is the primary tax authority in the USA) have accepted the arguments involved, and will not treat the receipt of staking rewards as income tax events. I have updated the post with links to this breaking news.
When my analysis and proposal was first posted, nearly every CPA/accountant in the USA would have advised that when tokens are received as staking rewards they immediately generate income tax liabilities. However, the correct tax treatment of such things derives from the law, and tax authorities cannot ignore the law when defining tax treatment. In this case, the IRS wanted to treat staking rewards as ordinary income, but presumably they were swayed by the same legal considerations that I made in my arguments. In this case, the law will hopefully have prevented those involved in staking in the USA being exposed to harmful financial situations.
Why we must care about these firewalls
Although the first firewall appears to have succeeded in the USA, we should not be lulled into believing that the second firewall is not important. Firewalls should be made as numerous and strong as possible to protect participants in the Internet Computer’s decentralized ecosystem. Furthermore, just because the IRS in the USA appears to have stopped at the first firewall, that does not mean tax authorities in other jurisdictions around the world will do the same, and we need to extend protection as far as possible.
It’s tempting to make the ideological claim that tax considerations should play no part in decentralized network design. However, in practice that limits participation, ensuring that only community members in special locations like Puerto Rico, Portugal and Monaco can stake with financial safety. This surely makes the argument incompatible with the basic aims of decentralization. If it can be helped, we should do everything we can to provide for broad participation and protect our community and the decentralized token economy created by the network.
Helping evolve an adaptive, self-directed blockchain
We should aim to constantly refine and improve the decentralized protocols of the IC so as to lay ever better foundations for a thriving and growing network economy. That the IC self-directs and evolves using the NNS will potentially eventually prove a decisive advantage - and from day one, it has always been a clearly stated aim of this project to create an adaptive blockchain (see this early post proposing a “Blockchain Nervous System”). We cannot shy from proposing improvements to the NNS framework and tokenomics without abandoning a founding objective of the project.
Because the IC is adaptive, rather than enshrine ossified blockchain dogmas, it seeks to “win” through rapid evolution - including some degree of active experimentation. We can already see the advantages of this paradigm in the rapid release of new technical advances and features, such as the provision of capabilities derived from “chain key” cryptography to smart contracts that will allow them to create transactions on any blockchain without holding private keys, and turnkey Service Nervous System DAO functionality for dapps, among many other things. No other blockchain on earth advances nearly as fast as the Internet Computer, and eventually, once the price manipulation, FUD and press boycotts fade, and the incredible services being built by our organically growing community begin to make their mark, this will begin to tell.
Thinking about the most controversial feature
In the proposal, when maturity is used to produce new ICP, the quantity of ICP created some days in the future is modulated by ICP price trends. This makes it near impossible for tax authorities to draw a simple equivalence between increased maturity and ICP tokens received and thus claim that maturity represents income received that has created a tax liability. I hope that the foregoing highlights why this is important for large parts of the community, in order to help prevent them being exposed to unavoidable tax liabilities they have no control over, and also for the wider community, who suffer when others feel forced to dump ICP on the markets to mitigate the risks. However, there is also another dimension to this feature.
The modulation feature increases the amount of ICP produced from maturity by up to 5% if the markets are rising, and decreases the amount of ICP produced from maturity by up to -5% if the markets are falling (to understand the full details, please refer to other posts in this thread). This is possible because the NNS is made aware of the price of ICP and various fiat exchange rates by the submission of proposals, which enables it to transform ICP into cycles at a constant rate, where ICP worth 1 SDR produces 1 trillion cycles, such that the cost of computation remains constant for those building dapps and services (note: although currently the NNS learns of external exchange rates through proposals submitted to it, soon it will inform itself by using HTTP outcalls that pass through consensus). As well as hardening the second firewall, this feature also creates incentives for neuron holders.
The system creates a small incentive to use maturity to produce and sell ICP when the market is rising, since more ICP will be received, than when it is falling, when less ICP will be received. In a more mature market, this (or at least some more sophisticated later variant created through the ongoing evolution of the network) might exert a significant stabilizing effect on the market, which is beneficial for adoption of ICP. A concern about this feature, which has already been raised in this thread, is that it is unfair for those wishing to sell during a market decline to receive less ICP than those selling when the market is rising. However, in practical terms, there is a strong counter argument.
Those wishing to sell, will exchange ICP for some other thing of value, most often fiat currency that can be used in the off-chain world. If the overall volume of selling in a downmarket is reduced by this feature, then the price those selling receive for their ICP may be higher. Simply put, often times while those using maturity to produce ICP for sale may receive, say, 5% less ICP in a steeply falling market, the price they receive for their ICP might be 5% or greater higher, and thus overall they will actually do better.
Such economic reasoning is difficult to make work, and its incorporation into tokenomnics schemes pushes into uncharted territory, but that does not mean designs should not take such reasoning into consideration. Decentralized protocols create public networks that survive and thrive by merit of their decentralized economies (even if today, token price often largely derives from conceptions and misconceptions carefully cultivated in the popular imagination, since this will eventually change). The Internet Computer is not a simple blockchain like Bitcoin - it aims to provide a universal public compute layer for the world, so that everything can be built using smart contracts. Today, the remarkable efficiencies of its architecture allow smart contracts to serve HTTP, and data and data processing to be moved on-chain, among other advances. Innovation and evolution should not be limited to things like chain key cryptography, but also involve its tokenomics, even though this too is a complex and difficult area.
We must constantly optimize, not least because the Internet Computer holds continual evolution as a core feature through the Network Nervous System. This allows for iteration where some measured risks can be taken, because the network is adaptive and designs can be refined, on an almost continuous basis, R&D contributions allowing, or if necessary, even reverted if they are not working. We should avoid be distracted by sacred cows and popular blockchain dogmas. The ability of this network to continually evolve along all design axes will prove a decisive strength. The Network Nervous System is what makes the Internet Computer blockchain self-directed and adaptive, and we should also embrace its own rapid evolution, and the evolution of next-generation tokenomics.