Original Title: "V God: Nathan Schneider Discusses the Limitations of Crypto Economics, Blockchainization Should Not Be Equated with Financialization"

Author: Vitalik Buterin, Co-founder of Ethereum
Translator: Yangz
Source: 8btc

Nathan Schneider (at the University of Colorado Boulder) recently published an article describing his views on crypto economics, particularly regarding the limitations of governance methods in crypto economics and what can be used to enhance its practicality. Of course, this is a topic that resonates with me ([1][2][3][4][5]), so I am pleased to see someone taking the blockchain space seriously as a knowledge tradition and engaging in discussions on these issues from different unique perspectives.

The main question that Nathan's work attempts to explore is simple. A large body of intellectual work has criticized what they call the "economization," "neoliberalism," and similar bubble concepts, arguing that they erode the values of democratic politics and thus leave many people's needs unmet. (However) the world of cryptocurrencies is very economic (with many tokens flying around and many functions assigned to these tokens), very new (this space is only 12 years old!), and very free (freedom and voluntary participation are at the core of the whole thing). Do these criticisms also apply to blockchain systems? If so, what conclusions should we draw, and how should we design blockchain systems to address these criticisms? Nathan's answer is: more mixed approaches that combine ideas from economics and political science. But what does it actually take to achieve this, and can it bring us the results we want? My answer is: yes, but there are many nuances involved.

What are the criticisms of neoliberalism and economic logic?

At the beginning of Nathan's article, he briefly describes the criticisms of the overuse of economic logic. That is to say, he does not delve further into potential critiques himself but tends to point to other materials that have already explored this issue in depth:

The economics of crypto economics has raised a series of particular anxieties. Critics have long warned that the expansion of economic logic crowds out the vibrant political space of public life. From the Zapatistas in southern Mexico (Hayden, 2002) to political theorists like William Davies (2014) and Wendy Brown (2015), the desire of "neoliberalism" for economics to guide all aspects of society represents a threat to democratic governance and humanity itself. Here is Brown's perspective:

Neoliberalism transforms every domain and endeavor of humanity, and humanity itself, according to a particular economic image. All actions are economic actions; all domains of existence are framed and measured in economic terms and metrics, even if those domains are not directly monetized. In the rationality of neoliberalism and the domains it governs, we are all economic agents, and economic agents are everywhere (p. 10).

For Brown and other critics of neoliberalism, the rise of economics signifies the decline of politics, the space for collective decision-making for the common good, and the means to achieve it.

At this point, it is worth noting that the "neoliberalism" being criticized here is different from the "neoliberalism" that people cheerfully promote as part of the "neoliberal project"; what is being criticized here is a mindset of "sufficient bipartisan trade can solve all problems," while the "neoliberal project" advocates for a mix of markets and democracy. But what is the essence of the criticism that Nathan points out? Everyone behaves more like economic agents, but what is the problem with that? For this, we can take a detour and peek at the source, namely Wendy Brown's Undoing the Demos. This book helps provide a list of the most important "four harmful effects" (the following is a reformatted and abridged direct quote).

1. Exacerbating Inequality: In this case, the top tier accumulates and retains more and more wealth, while the bottom tier is effectively pushed onto the streets or into the slums of the increasingly urbanized and suburbanized world, and the middle tier works more hours but receives less pay, fewer benefits, and less security…

2. Brutal or Unethical Commercialization of Things and Activities Considered Unsuitable for Marketization: Claiming that marketization contributes to human exploitation or degradation, […] restricting or stratifying access to things that should be widely accessible and shared, […] or because it enables some inherently terrible or severely damaging things to the planet.

3. The Increasingly Close Relationship Between Corporate and Financial Capital and the State: The dominance of corporations over political decision-making and economic policy is growing.

4. The Rise of Financial Capital and the Destruction Caused by Freedom to the Economy: Particularly the inherent instability of bubbles and other severe fluctuations in financial markets.

Much of Nathan's article unfolds along the lines of analyzing how these issues affect governance mechanisms within DAOs and specific cryptocurrency spaces. Nathan focuses on three key issues:

• Plutocracy: "Those who own more tokens than others have more [I would say, disproportionate] decision-making power…"

• Limited Engagement with Various Motivations: "Crypto economics only sees a portion of the people involved. Concepts such as self-sacrifice, responsibility, and honor are fundamental characteristics of most political and business organizations but are difficult to simulate or approximate with the incentive designs of crypto economics."

• Positive and Negative Externalities: "Environmental costs are typical externalities—unseen by the feedback loops understood by the system, it is conveyed to users as incentives… Funding 'public goods' is another example of externalities—also a threat to the sustainability of crypto economic systems."

For me, the naturally arising questions are: (1) To what extent do I agree with this criticism, and how does it align with my own thoughts, and (2) What implications does this have for blockchain, and what do blockchain protocols need to do in practice to avoid these issues?

How do I view the general criticism of neoliberalism?

I disagree with some of the points but agree with others. I have always been skeptical of the criticism of "vulgar and unethical commercialization," as it often seems that the authors are trying to wash their own disgust and aesthetic preferences into grand ethical and political ideologies—this is a common sin of all such ideologies, often more so on the right (here's a random example) than on the left. Back when I had much less money and sometimes walked an entire hour to the airport to avoid taxi fare, I remember thinking that I would love to be compensated for donating blood or participating in clinical trials. So the idea that such transactions are inhumane exploitation has never been appealing to me.

But at the same time, I am far from being a Walter Block-style defender of all local voluntary two-party transactions. I have written in various articles expressing concerns similar to some of the items on Wendy Brown's list:

• Multiple articles have condemned the evils of ticket buying and even the general financialization of governance.
• The importance of funding public goods.
• The failure modes of financial markets caused by subtle issues like capital efficiency.

So where do my own objections to mixed finance and governance come from? It’s a complex topic, and my conclusion largely stems from my own years of trying to find a financially stable governance mechanism that has failed.

Finance Does Not Prevent Collusion

In the standard assumptions derisively termed "spherical cow economics," people tend to focus on the unrealistic nature of complete information and complete rationality. However, the unrealistic assumptions hidden in the list seem more misleading to me, and that is individual choice: the belief that each agent is making their own decision independently, with no agent having a positive or negative stake in another agent's outcome, and no "marginal games"; the only thing that can see each agent's decision is the black box we call "mechanism."

This assumption is often used to guide complex devices like the VCG mechanism, whose theoretical optimality is based on the perfect argument that since the price each player pays only depends on the bids of other players, each player has no incentive to make bids that do not reflect their true value to manipulate the price. In theory, this is a perfect argument, but once the possibility of two players being allies or adversaries outside the mechanism is introduced, it completely collapses.

Economics and economics-inspired philosophy are very good at describing the complexities that arise when the number of players "playing the game" increases from one to two (see Murray Rothbard's story of Crusoe and Friday in Ethics of Liberty as an example). But what this philosophical tradition completely overlooks is that increasing to three players adds even more complexity. In interactions between two people, they can ignore each other or fight or trade. In interactions among three people, a new strategy emerges: any two of the three can communicate and unite against the third. Three is the smallest denominator where it becomes possible to talk about 51% attacks that can make outsiders victims.

When there are only two people, more coordination can only be a good thing. But once there are three people, wrong coordination can be harmful, and the techniques to prevent harmful coordination (including decentralization itself) can become very valuable. And it is this management of coordination that is the essence of "politics."

From two people to three, the introduction of imbalanced coordination brings the possibility of harm: this is not just about "the relationship between individuals and groups," but rather "the relationship between individuals, groups, and the world."

Now, we can understand the pitfalls of trying to use this framework to understand "finance." Finance can be seen as a set of patterns that naturally emerge in many systems that do not attempt to prevent collusion. Any system that claims to be non-financial but does not actually make efforts to prevent collusion will ultimately acquire financial characteristics, if not worse. To understand why this is the case, we can compare two scoring systems we are all familiar with: money and Twitter likes. Both of these scores have value for external reasons, both inevitably become symbols of status, and both are numerical games that people spend a lot of time optimizing to try to achieve higher scores. However, their behaviors are vastly different. So what is the fundamental difference between the two?

The answer is simple: it is the lack of an effective market to facilitate agreements like "if you like my tweet, I will like yours" or "if you pay me with other currencies, I will like your tweet." If such a market existed and was easy to use, then Twitter would completely collapse (something akin to hyperinflation would occur, with the possible outcome being that everyone runs bots to like every tweet in exchange for rewards), and even the currently illegal market for buying likes poses a significant problem for Twitter. However, for money, "if you send me Y, I will send you X" is not an attack medium; it is just a boring old currency exchange transaction. If a Twitter clone that does not prevent similar markets "overinflates," turning into a platform where everyone likes everything, if this Twitter clone tries to prevent overinflation by limiting the number of likes each user can give, likes will behave like a currency, and the end result will resemble how Twitter's recently added tipping feature behaves.

So what is the problem with finance? Well, if finance is the optimization and structuring of collusion, then we can use our existing economic tools to find where finance causes problems to understand which mechanisms will break if you introduce collusion! This is important. Unfortunately, voting governance is a core example of this type; I have introduced the reasons in "Beyond Token Voting Governance" and many other occasions. Worse still, cooperative game theory suggests that there may be no possible way to create a governance mechanism that is completely resistant to collusion.

Thus, we arrive at a fundamental dilemma: the cyberpunk spirit is fundamentally about creating maximally unchangeable systems that work with as little information as possible about who is participating ("On the internet, nobody knows you're a dog"), but creating new forms of governance requires systems to have richer information about their participants and the ability to dynamically respond to attacks to maintain stability in the face of actors with unpredictable motives. If this cannot be achieved, it means everything looks like finance, which also means… years of over-representation of concentrated interests and all the problems that arise from it.

On the internet, nobody knows you're a dog (Image Source). But what does this mean for governance?

The Core Role of Collusion in Understanding the Difference Between Kleros and Ordinary Courts

Now, let’s return to Nathan's article. The distinction between financial and non-financial mechanisms is key to the article. Let’s start by describing the Kleros court:

Jurors earn rewards by correctly choosing the answers they expect other jurors to independently select. This process realizes the concept of the "Schelling point" in game theory (Aouidef et al., 2021; Dylag & Smith, 2021). Such a jury does not deliberate or collectively seek the common good; its members unite through self-interest. Before coming to the jury, the factual basis of the case should not come from official agencies or respected news organizations but from anonymous users who are similarly bound by the rewards they seek. Prediction markets themselves are predicated on this assumption: when people gain or lose money-equivalent benefits in this process, they make better predictions. Here, the political discourse around presidential elections has been thoroughly transformed into an economic cluster.

The implicit critique is clear: the ultimate motivation of the Kleros court is not to make decisions based on its "true" correctness or incorrectness but rather based on its economic interests. If Kleros decides whether Biden or Trump wins the 2020 election, and a Kleros juror is very fond of Trump, pre-committed to voting for him, and bribes other jurors to vote the same way, then other jurors are likely to yield to Kleros's consistency incentives: if a juror's vote aligns with the majority, they will be rewarded; otherwise, they will be punished. The theoretical answer in this regard is the exit right: if the majority of Kleros jurors vote to declare Trump the winner, then a minority can fork off from Kleros, claiming Biden won, and their fork is likely to achieve a higher market price than the original. Sometimes, this is indeed feasible, but as Nathan points out, it is not always that simple.

However, whether from social media networks or protocols, exiting may not be as easy as it seems. The continued dominance of early blockchains like Bitcoin and Ethereum indicates that crypto economics also favors incumbents.

But alongside the implicit critique, there is also an implicit promise: ordinary courts can, to some extent, transcend self-interest and "collectively seek the common good," thereby avoiding some of the failure modes. What exactly allows the financialized Kleros court to lack what non-financialized ordinary courts retain, making them stronger? One possible answer is that courts lack the explicit compliance incentives of Kleros. However, if you simply replicate Kleros as is, removing the consistency incentives (for example, the rewards for voting do not depend on how you vote), without doing anything else, you may create more problems. Kleros judges may become lazy, but more importantly, without any incentives to choose how you vote, even the slightest bribe can influence a judge's decision.

Thus, we now begin to discuss the real answer: the key difference between the financialized Kleros court and the non-financialized ordinary court is that the financialized Kleros court is, well… financialized. They do not make explicit efforts to prevent collusive behavior. On the other hand, non-financialized courts do indeed prevent collusion in two key ways:

• Bribing judges to vote in a specific way is explicitly illegal.
• The judges' positions themselves are irreplaceable. They are specific, carefully selected individuals who cannot simply sell or reallocate their entire judicial authority and salary to others.

The reason political and legal systems can function is solely due to the extensive hard thinking and work behind the scenes that isolates decision-makers from external incentives, and if they are found to accept external incentives, they are explicitly punished. The lack of external incentives allows internal incentives to shine through. Furthermore, due to the lack of transferability, governance rights can be entrusted to specific actors whose intrinsic motives we trust, avoiding governance rights always flowing to the "highest bidder." But in the case of Kleros, the lack of hostile external motives is not guaranteed, and transferability is inevitable, so overly powerful internal and external motives (compliance incentives) are the best solution they can find to address the issues.

Of course, the "last resort" that Kleros relies on, namely the right of users to fork away, itself depends on the occurrence of social coordination—this is a chaotic and difficult institution often mocked by crypto economic purists as "proof of social media," its role being precisely because public discussions have many informal collusion detection and prevention mechanisms everywhere.

Understanding Collusion in DAO Governance Issues

But what happens when there is no single correct answer, and they can expect voters to converge? This is where we move from adjudication to governance (yes, I know adjudication also has inevitable gray edge cases. Governance just occurs more frequently). Nathan writes:

Economic governance is not new. Joint-stock companies traditionally operate based on plutocratic governance—more shares equal more voting power. This arrangement is economically effective in coordinating shareholders' interests (Davidson and Potts), even if it may exclude external factors like fair wages and environmental impacts…

In my view, this actually acknowledges too much. Once you abandon the assumption of no collusion in spherical cows, economic governance is not "effective" because it is itself susceptible to the influence of 51% of stakeholders colluding to liquidate the company and distribute resources among themselves. The only reason this situation does not frequently occur in "real life" is that decades of shareholder regulation have been explicitly established to prohibit the most common abuses. Of course, this regulation is non-"economic" (or, in my jargon, it makes corporate governance less financialized) because it is an explicit attempt to prevent collusion.

It is noteworthy that the solutions Nathan advocates do not attempt to regulate token voting. Instead, they seek to mitigate the harms of its weaknesses by combining it with other mechanisms.

The Graph does not rely on direct token voting like other protocols but uses a board-like mediation layer, the Graph Council, which includes representatives from the main stakeholder groups of the protocol. In this case, the proposal may benefit one stakeholder group over others, and decisions by the council require the agreement of multiple stakeholder groups. Meanwhile, snapshot voting puts pressure on the council to execute the will of token holders.

In the case of 1Hive, the anti-financialization safeguards are described as purely cultural:

According to the slogan that repeatedly appears in 1Hive discussions, "Come for the honey, stay for the bees." That is to say, although the economy occupies a prominent position when people first encounter and explore 1Hive, participants understand that the community's primary value is interpersonal, social, and non-economic.

I personally hold skepticism towards the latter approach: it may work well in fun-oriented communities with low economic value, but if this approach is attempted in a more serious system with broad open participation and sufficient stakes to invite resolute attacks, it will not survive long. As I wrote above, "Any system that claims to be non-financial but does not actually make efforts to prevent collusion will ultimately acquire financial characteristics."

[Edit/Correction 2021.09.27: Someone pointed out to me that besides culture, financialization is also subject to (i) conviction voting and (ii) limitations on jurors executing contracts. In the long run, I am skeptical of conviction voting; many DAOs use it today, but in the long run, it can be defeated by encapsulated tokens. On the other hand, the covenant is quite interesting.]

This money is called honey. But is calling money honey enough to differentiate its role from that of money? If not, how much more do you need to do?

The Graph's solution is largely an example of preventing collusion: participants are carefully selected, coming from different constituencies, trustworthy, and upright individuals who are unlikely to sell their voting rights. Therefore, if this approach successfully avoids centralization, I am optimistic about it.

So how can we more generally address these issues?

Nathan's article argues:

A napkin sketch of classical, never-realized liberal democracy (Brown, 2015) would describe a market (managed through economic incentives) wrapped in politics (managed through deliberation of the common good). Economics has its place, but this system is not entirely economics; the rules guiding the market, as well as the rules that allow the market to exist in the first place, are democratically determined based on citizens' civic rights rather than economic power. By embedding democracy into the system's foundation, it is possible to overcome the various limitations that crypto economics is prone to, such as countering plutocracy through mass participation and making external factors that markets may not see more visible.

There is a key distinction between blockchain political theory and traditional nation-state political theory—over the long term, nation-states may have to learn from blockchains. The political theory of nation-states talks about "markets embedded in democracy," as if democracy is a foundational layer that encompasses all societies. In reality, this is not true: there are multiple nations, each of which allows trade with external nations that they cannot regulate to some extent. Individuals and companies can choose in which countries to live and do business. Therefore, markets are not only embedded in democracy but also revolve around democracy, and the real world is a complex interaction between the two.

Blockchain systems do not attempt to counter this interconnectivity but embrace it. Blockchain systems do not have the capacity to regulate "markets," that is, the general ability of people to trade freely. But what they can do is regulate and construct (even create) specific markets, establishing patterns of specific behaviors whose incentive mechanisms are ultimately set and guided by built-in anti-collusion institutions, capable of resisting pressure from economic actors. In fact, this is also the direction Nathan ultimately aims for. He actively discusses the design of Civil, believing it is an example of this spirit.

The failed Ethereum-based project Civil attempted to use crypto economics to protect journalism from censorship and the degradation of professional standards (Schneider, 2020). Part of the system is the Civil Council, a committee of prominent journalists serving as a supreme court to adjudicate the practices of network newsrooms. Token holders can earn rewards by successfully challenging the practices of newsrooms; the success or failure of challenges ultimately depends on the judgment of the citizen council, which is designed to be unaffected by economic motives. In this way, the crypto economic enforcement market serves a non-economic social mission. This design can allow crypto economic networks to serve purposes that cannot be reduced to economic feedback loops.

This fundamentally resembles an idea I proposed in 2018: Prediction Markets to Scale Content Moderation. Rather than running low-quality AI algorithms on all content for moderation, generating a lot of false positives, it would be better to establish a public small prediction market on each post, where if enough people participate, a high-quality committee can intervene to adjudicate, and participants in the prediction market will be punished or rewarded based on whether they correctly predicted the outcome. Meanwhile, posts predicted to be deleted will not be shown to users who have not explicitly chosen to participate in the prediction game. This kind of open yet responsible moderation has precedents. Slashdot's meta-control can be said to be a limited version of it. This more financialized meta-regulation through prediction markets may yield better results, as the incentive mechanisms invite capable and professional participants to engage.

Nathan subsequently expands on this:

I believe pairing crypto economics with political systems helps overcome the limitations that individually plague crypto economic governance. Introducing purpose-driven mechanisms and temporal controls can compensate for the blind spots of token economies. But I am not advocating for a complete rejection of crypto economics. I am not saying that this politics must occur in every application and protocol. Liberal democratic theory allows for multiple forms of association and business within democratic structures; similarly, politics may only be necessary at key leverage points in the ecosystem to overcome the limitations of crypto economics individually.

This seems roughly correct. As Nathan points out in his conclusion, financialization has benefits because it attracts a large amount of motivation and energy to build and participate in systems that would not otherwise exist. Moreover, preventing financialization is very difficult and costly, and doing less where it is most needed tends to be most effective. However, if the incentive mechanisms of financialized systems are structured around an ultimately non-financialized system, then it is indeed more stable.

Prediction markets avoid the core plutocratic problems inherent in token voting because they introduce personal accountability: users who vote in favor of a decision that ultimately proves wrong suffer more losses than those who vote against it. However, prediction markets require some statistics they are measuring, and measuring oracles cannot achieve security solely through crypto economics: at least community forks are needed as a backup to prevent attacks. And if we want to avoid the chaos of frequent forks, some other explicit non-financial mechanisms at the center are a valuable option.

Conclusion

In his conclusion, Nathan writes:

However, the autonomy of crypto economic systems from external regulation may make them more susceptible to uncontrolled feedback loops, where narrow incentives overwhelm the public good. The designers of these systems have shown admirable capability in designing many kinds of crypto economic mechanisms. However, to achieve the institutional range that advocates of crypto economics hope for, it needs to leave space for less economic forms of governance.

If crypto economics needs a political layer and is no longer self-sufficient, what use is crypto economics? One answer might be that crypto economics can serve as a foundation for ensuring more democratic and value-centered governance, where incentives can reduce reliance on military or police force. Through mature design, combined with fewer economic purposes, crypto economics may transcend its initial limitations. Politics also needs crypto economics… By combining crypto economics with democracy, both legacies seem to benefit.

I broadly agree with both conclusions. The language of preventing collusion can help understand why the purism of crypto economics severely limits the design space—"finance" is a class of patterns that emerge when systems do not attempt to prevent collusion. When a system does not prevent collusion, it cannot differentiate between different individuals, nor can it differentiate between different numbers of individuals: as long as there is a "position" that exerts influence, the owner of that position can sell it to the highest bidder.

The gavel on Amazon. If these were truly NFTs with relevant judicial authority, the world could be very interesting, but I certainly wouldn't want to be a defendant!

On the other hand, a defense-centered design language is an undervalued way to think about some of the advantages of blockchain-based design. Nation-state systems typically handle threats with one of two totalizing mindsets: closed borders or conquering the world. The closed-border approach attempts to make a hard distinction between the "internal" that the system can regulate and the "external" that the system cannot regulate, severely limiting the flow between the internal and external. The conquering-the-world approach attempts to extraterritorialize the preferences of one nation-state, seeking a state where no unwelcome activities can occur anywhere in the world. Blockchains are structurally unable to adopt either approach, so they must seek alternatives.

Fortunately, blockchains do have a very powerful tool that makes security under these porous conditions actually feasible: cryptography. Cryptography allows everyone to verify that some governance procedures are being executed according to the rules. It leaves verifiable evidence trails for all actions, although zero-knowledge proofs allow mechanism designers to freely choose which evidence is visible and which is not. Cryptography can even prevent collusion! Blockchains allow applications to exist on a substrate that their governors cannot control, enabling them to effectively implement certain techniques, such as ensuring that every rule change only takes effect after a 60-day delay. Finally, free forking is more practical than most centralized systems, and the economic and human costs of forking are much lower.

Blockchain-based devices can offer many things to the world that other kinds of systems cannot. On the other hand, Nathan is completely correct that blockchainization should not be equated with financialization. There is a lot of space for blockchain-based systems that do not look like money, and we do indeed need more such systems.

Source link: www.8btc.com