Blockchain Governance. Hi, this is Peter Zeitz, research… | by Peter Zeitz | 0x Blog


Hi, this is Peter Zeitz, research fellow on the 0x core team. Formerly, I was an assistant professor of economics at the National University of Singapore Business School and prior to that a post-doctoral fellow at Stanford, where I arrived after earning an Economics PhD from UCLA. I have been interested in crypto for a long time. I first learned about Bitcoin during my time at Stanford in April 2011. This very quickly led to the purchase of some GPUs and the creation of a small Bitcoin mining node from my San Francisco apartment, which, as I later discovered, was less than a block away from Ross Ulbricht’s alleged Silk Road command center. Anyways, having lingered in this space for a long time as a hobbyist, I was very excited by an opportunity to pursue my interest full time as a governance researcher at 0x. Upon arriving, I quickly learned that the talent and set of personalities assembled here are pretty awesome.

I am now delighted to begin sharing some of my research on governance with the broader 0x community. In this essay, I focus on how the extent of horizontal competition and the assignment of project ownership rights affects the incentives of blockchain development teams to collaborate. The essay may be a bit long and dry, but if you are interested in a deep dive into blockchain governance I think you’ll find the content interesting. In parallel with this work, I have been working on plans to operationalize decentralized governance systems at 0x. In the next couple of weeks, I will provide some details on our plans to utilize a community grant program to implement experiments with decentralized governance. Some of our emerging governance ideas are still too preliminary to release to the public, but in the coming months I expect that we will have a lot of exciting work to disclose.

Here are some bullet points that summarize the essay’s argument:

  1. Sustaining collaborative innovation among independent projects is a governance problem.
  2. A core part of governance is determining an optimal division of rights to control project decisions across various stakeholders.
  3. Blockchains frequently require participation of multiple independent platform layers to deliver functionality to end users.
  4. Some of these platform layers operate in winner-take-all markets and have the potential to acquire significant market power.
  5. Platform layers that acquire market power and that are controlled by third-party investors will run into conflicts with upstream and downstream projects over rights to collect revenue from end users. These conflicts will make it difficult to sustain cooperation in areas such as interoperability and open source licensing.
  6. Consolidation of platform ownership within a single entity is one way of resolving these conflicts.
  7. In an alternative resolution which preserves decentralization, platform layers that expect to acquire market power could transfer ownership rights to end users. User-owned platform layers have a mandate to operate as non-profits and therefore do not compete with upstream and downstream layers over revenue. This allows non-profits to sustain collaboration with upstream and downstream layers under a much broader set of circumstances.
  8. User-owned platforms encourage third party participation to a greater degree than investor-owned platforms, but are likely to face difficulties attracting initial investment.
  9. To balance these tradeoffs, projects could finance initial operation via equity investment and then gradually sell ownership to users as usage expands. This transfer can be implemented via token mechanics.

Governing Blockchains as Public Infrastructure

Blockchain platforms have witnessed an explosion of collaborative innovation, where small independent teams of developers create narrow layers of functionality that interact with one another to create an end user experience. Currently, project teams in our ecosystem align towards a shared goal of building out open infrastructure and persuading a common pool of users to adopt it. As long as this bundle is valuable principally due to its prospects of future usefulness, it will remain easy for independent teams to sustain cooperation in areas such as open source licensing, interoperability, and competitive resource pricing. Eventually, however, projects in the ecosystem will need to assert rights to collect revenue from platform users in what will be an inherently competitive process. When this transition occurs, there is a risk that cooperative efforts to develop the blockchain as public infrastructure will break down.

Sustaining cooperative innovation in the blockchain ecosystem will be a governance challenge and not a technological one. Blockchains have the potential to deliver public network infrastructure that anyone can build on without seeking permission from a central owner. However, some of the projects that contribute to this infrastructure operate in markets with strong network externalities and have the potential to become natural monopolies. A monopolist owner of critical infrastructure would be in a position to extract rents for blockchain access in much the same manner as the owner of a private network. In this essay, I argue that projects that have the potential to become monopolies should seek to gradually transfer ownership and control to end users. User ownership would commit projects to operation as nonprofit entities similar to public utilities.

What is accountability and why does it matter for the governance of organizations?

Governance describes a set of rules, processes, and procedures through which an organization is controlled. The purpose of governance is to ensure accountability of an organization’s manager to one or more classes of patrons, defined as a group who supplies the organization with resources. In most private organizations, patrons elect a small group of board members who monitor the organization on their behalf and the board in turn appoints a manager to handle the organization’s affairs. Direct control over almost all organizational decisions resides with the manager. Nevertheless, the manager’s authority is constrained by the need to justify any decisions she makes to the board and to patrons that vote in board elections. The indirect control that patrons exercise over managerial decision-making through a threat of potential dismissal is known as accountability (Bainbridge 2008).

The specific class of patrons that an organization is held accountable to differs across governance systems. A diverse range of accountability structures are present in real world organizations and the selection of an appropriate structure depends on an organization’s circumstances. For example, in democratic states where membership is coincidental and involuntary, governance structures attempt to maintain accountability to diverse groups. In private organizations, individuals affiliate on the basis of perceived similarity and often prefer to exit organizations where they form a minority. As a result, private organizations tend to represent a narrowly-defined class of individuals and adopt governance structures that enforce accountability to a single class of patrons (Hansmann 2006). Within the universe of private organizations, many different types of single stakeholder accountability are represented. For example, joint-stock companies enforce accountability to equity holders, producer cooperatives to groups of member firms, worker cooperatives to employees, consumer cooperatives to consumers, and open source communities to volunteer developers.

How do organizations choose who to be held accountable to?

A grant of exclusive voting rights serves to protect patrons from risks of opportunistic behavior by other members of an organization. Frequently, once a patron makes a resource commitment to an organization, it becomes difficult for him to exit the organization without incurring a substantial loss. As a result, the organization may promise patrons favorable terms of resource provision to convince them to make a commitment. Once a commitment has been made, however, an organization may wish to renegotiate these terms to take advantage of the patron’s loss of bargaining power. The patron thus faces a risk of ex-post opportunism which can prevent him from agreeing to contract with the organization in the first place. Assigning ultimate control of the organization to patrons alleviates risks of ex-post opportunism and increases their willingness to supply resources (Williamson 1985). When risks of ex-post opportunism primarily affect a single class of patrons, granting exclusive voting rights to this group enhances a firm’s access to resources and confers a competitive advantage.

The joint-stock company provides the most well-known example of how this theory can be applied to determine an optimal allocation of voting rights. In the textbook situation where an organization as a joint-stock company is optimal, firms are assumed to utilize equity investment and to operate in competitive input and product markets. Equity investors in the firm face a risk of managerial opportunism since the market value of their holdings depends on a firm’s expected earnings and can be driven to zero if a manager is lazy or dishonest. Granting equity investors the right to vote out management is beneficial for organizations because it increases their willingness to invest and reduces an organization’s effective cost of capital. By contrast, due to the assumption that the firm operates in competitive input and product markets, the firm’s customers, suppliers, and employees are protected from opportunism by the ability to seek out competing offers from other firms. Since risks of ex-post opportunism are borne solely by equity investors, it is socially optimal for the firm to assign exclusive voting rights to this group.

For monopolies initial assignments of voting rights can have long-term consequences.

The assignment of voting rights raises interesting normative questions when one considers an organization that threatens to become a monopoly. In this case, the bargaining power a monopoly holds in its relationships with customers and suppliers results in several groups facing risks of ex-post opportunism. It is generally not possible to assign voting rights in a way that simultaneously protects the interests of suppliers, customers, and equity investors. Instead, stable assignments of voting rights will tend to favor one group. In order for an assignment of voting rights to be stable, the interests of voting members need to be sufficiently well-aligned that members agree on critical issues. Furthermore, the benefit that patrons obtain from controlling voting outcomes must be significant enough that they prefer to hold voting rights rather than sell them to third parties (Hart and Moore 1998).

Under monopoly, there can exist multiple stable assignments of voting rights. When this occurs, an organization can commit to ongoing operation on behalf of a particular class of patrons by assigning initial voting rights to this group. More formally, an organization will have some ability to determine its own accountability structure when (1) a class of patrons that is granted exclusive voting rights will prefer to retain these rights rather than auction off voting rights to a third party AND (2) there are multiple assignments of voting rights that satisfy (1) (Volker et al 2004).

Problems caused by accountability of monopolies to third-party investors.

Assigning voting rights to equity holders is advantageous from the perspective of a firm seeking funding because it commits the firm to acting on investors behalf. However, when an organization operates as a monopoly, accountability to third party investors introduces a range of inefficiencies. In general, equity investors will prefer to price platform services on the basis of long-term profit maximization rather than maximization of social benefit or end user utility. Operation of the platform so as to maximize profit leads to inefficiently high access fees that cause the platform to be underutilized relative to a socially efficient outcome. A second type of inefficiency arises when the existence of a lucrative winner-take-all market provokes excessive entry. When excessive entry occurs, early entrants offering similar end user functionality may compete to win the market on the basis of distributing a war chest of subsidies. In this case, the total cost of subsidies and other incentives can greatly exceed what would be necessary for a single firm to gain traction.

The severity of distortions arising from monopolies increases greatly when independently-owned platforms build on top of one another in what is known as a sequential monopoly. Under a sequential monopoly, end user applications may require payments to multiple platform layers that wield strong bargaining power in order to deliver full functionality. When these layers are under separate ownership, their owners fail to internalize effects of pricing decisions on usage volumes. In particular, if an application requires access to multiple layers and the owner of this application reduces fees to promote broader adoption, the best response for owners of upstream and downstream layers is typically to raise prices in order to extract additional rents. As a result, attempts to promote platform adoption by keeping fees low require ongoing cooperation from each participant in a chain of value creation. At some point, if the owner of one layer in the chain chooses to begin monetizing usage, cooperation breaks down and the owners of other layers are forced to shift to a monetization strategy as well. If this problem is not addressed, the presence of sequential monopoly is likely to eventually constrain developer and end user adoption of blockchain platforms.

In order to resolve this problem, platform layers that threaten to become monopolies should adopt governance structures that commit them to acting on behalf of one or more classes of end users. These governance structures are unique in that they can allow organizations to commit to operation as nonprofits. This commitment would ensure that creators of downstream applications benefit from their own creative work and thus encourage third-party innovation. If efforts to transition platforms to user ownership fail, a second-best resolution of the sequential monopoly problem is consolidation of the value chain under a single owner. Following consolidation, the resulting ownership structure of a ‘second-best’ blockchain platform might closely resemble that of Google, Facebook, or Amazon.

Who owns a project after a utility token ICO?

Many blockchain platforms have eschewed traditional equity finance and instead relied on crowd sales of utility tokens to raise capital. The set of incentives established in these crowd sales depends on how token sales are structured, whether utility tokens confer a right to vote on organizational decisions, and whether projects incorporate inflation in token mechanics. To a large extent, these issues are still up in the air. At one extreme, it is possible to structure a utility token sale so that it becomes functionally equivalent to equity finance. In this case, the problems of sequential monopoly are not addressed by shifting from a traditional IPO to a utility token ICO. At the other extreme, it is possible for a project that does an ICO to adopt governance structures that assign accountability to a class of end users in a manner akin to a consumer or producer cooperative. These structures can allow organizations to commit to keep fees low, but tend to make it more difficult to raise capital. Finally, projects could adopt structures that lie on a continuum anywhere between these two extremes.

In a traditional crowdfunding scenario, consumers agree to pay a fixed price in exchange for a nontransferable right to utilize a good or service. If later participants can enter the crowd sale on the same terms as initial backers and if all crowd sale participants acquire exclusive voting rights, the governance structure that emerges is a classical consumer cooperative. Under this structure, initial backers of a platform are unable to extract rents from future generation of users. Since all users acquire equal voting rights regardless of when they join, governance is highly democratic. Unfortunately, consumer cooperatives face severe bootstrapping problems. Since they operate under rules that preclude speculative investment, it is only worthwhile to participate as an initial backer if a critical mass of other backers also agrees to participate at the same time. If this bootstrapping hurdle can be overcome, ongoing growth of a consumer cooperative platform is likely to be self-sustaining since the benefits new users derive from membership increase as more users join.

Utility token sales differ from traditional crowdfunding in that they permit initial backers to resell usage rights to third parties. The bulk of participants in these crowd sales likely acquire usage rights primarily in the hopes of reselling them to future generations of users at a higher price. On the surface, these token sales appear to differ from investor ownership in that the resale of usage rights is not centrally coordinated. However, if token holders acquire exclusive voting rights and thus gain control over platform rules, one would expect them to vote to use the platform to organize the resale of usage rights so as to maximize their collective profit. In other words, regardless of whether tokens represent an equity claim or a transferable right to acquire goods and services, token holder voters face the same set of incentives and would be expected to vote for identical pricing decisions.

How can projects balance their need to raise capital from third-party investors with the competing goal of remaining accountable to end users?

Perhaps the most fundamental difference between the two models is that in the consumer cooperative case the price of usage rights remains constant for both initial and future participants, whereas in the alternative case where resales are allowed, initial backers expect usage rights to appreciate. This prospect of utility token appreciation allows third-party financiers to extract rents from platform adoption. Under this structure, utility token holders face an incentive to restrict the free circulation of tokens so as to raise the market price of services to the monopoly level. Just as in a traditional monopoly, this leads to a socially inefficient under provision of services. At the same time, however, permitting the resale of usage rights allows projects to raise much larger sums in an ICO than would be possible through traditional crowd sales. Thus, allowing resales overcomes the bootstrapping problem faced by traditional cooperatives.

Many blockchain projects may want to locate on a continuum between these two extremes. In this case, a project would strike a balance between maintaining an adequate incentive for initial investment and committing to act on behalf of one or more classes of users. Since these projects will likely find it more difficult to raise capital, one might be skeptical of the strategic viability of this approach. However, the cooperative form also has some strategic advantages. Adopting a (partial) cooperative structure allows a project to commit to maintaining at cost pricing in the long-term. This commitment strengthens incentives of upstream and downstream businesses to join the platform (Rey and Tirole 2007).

When adoption by upstream and downstream business is vital for a platform’s success, it can be rational for a for-profit company to sell ownership to users in order to create a cooperative. The most famous example of this strategy is VISA’s origin story. In the mid 1960s, Bank of America was licensing rights to utilize its payment processing services and issue credit cards with the BankAmericard imprint to third party banks throughout California. Major regional banks feared that BofA would gain too much power as its network expanded and that this would ultimately lead to progressive increases in the licensing fees charged for use of its technology. To address these concerns, BofA changed the name of its card product to VISA and sold ownership of VISA to a cooperative association of member banks throughout the US. A recent analysis suggests that the restructuring of VISA as a cooperative resulted in a larger proportion of the benefits of credit card technology accruing to consumers (Weyl 2009).

How can blockchain projects structure incentives to encourage user ownership?

This discussion leads to the practical question of how blockchain projects could progress towards user ownership. One way of accomplishing this is through the introduction of token mechanics that provide users who stake tokens with a stream of usage rights. Under this model, token ownership would provide benefits analogous to those of a gym membership or a magazine subscription. For example, a platform adopting this structure could rebate a portion of any external fees token holders incur when accessing platform services. Examples of these fees include charges for blockchain storage, computation, decentralized exchange, oracle use, etc. Rebates could be capped at a fixed percentage of staked tokens, so that user staking requirements would remain proportional to frequency of use. Furthermore, if these rebates were funded through token inflation, the resulting structure would transfer a stream of value from third-party holders of the token to user owners. The strength of the incentives for user ownership created through this system depend on the proportion of fees rebated and the platform’s current usage volume. In general, ownership would tend to flow from third-party investors to user owners as usage of the platform expanded. However, the strength of incentives for this transition could be left up to individual projects. In the limit, a project offering very generous rebates financed by very rapid inflation would closely resemble a traditional user cooperative.

What type of user ownership should be encouraged when a project has multiple classes of users?

A very important final point to make is that the definition of ‘user’ and ‘user cooperative’ is ambiguous for many blockchain projects. As a rule, the blockchain applications that are most likely to progress towards a monopoly market structure are multi-sided platforms. These platforms match two or more classes of users together in order to facilitate an exchange between them. Any of these classes of patrons could potentially be designated as intended user owners. For example, for 0x, there are at least three potential classes of user owners: relayers, market makers, and takers. Often, there will be trade-offs that strongly favor assignment of ownership to particular group(s) of users. I expect these trade-offs to be nuanced and project specific, so that each individual project stands to benefit by tailoring the assignment of ownership rights to its own needs.