Many in the Cardano community were surprised by recent developments in the Catalyst Fund 14 voting process. This round introduced an alternative voting method aimed at creating a fairer balance of influence among voters, but an unexpected interaction in the voting mechanics led to outcomes that differed from what was intended. The decision was ultimately made to revert to the previous system to ensure fairness and continuity for all participants. This moment highlights both the challenges and the value of experimentation in decentralized governance. Let’s take a closer look at what happened.
In Fund 14 the Catalyst program specified an allocation of voting creditsaccording to the cubic (⅓) root of the wallet ADA holdings. A miscalculation of an approval threshold resulted in only a dozen or so projects qualifying for funding. While the approval threshold, in which a minimum required number of votes determines eligibility, is separate from the ranking of proposals, the leadership decided to revert to the linear method of prior funds for this round’s allocation.
First, it helps to separate two related but distinct concepts: the allocation of voting credits (how much influence each voter has) and quadratic voting (QV) (how that influence can be expressed across choices). QV can and has been successfully implemented with both democratic (i.e., one-person-one-vote (1p1v)) and corporate (i.e., one-coin (or other asset) one-vote (1c1v) allocations. Let’s review the QV method, then return to the issue of allocation of voting credits. Traditional voting methods, such as democratic 1p1v and corporate share voting, are known to be inefficient, that is, poor allocators of value. Arrow’s Impossibility Theorem famously proved the difficulties of social choices based on ranked decisions, including binary decisions. Providing voters with the ability to continuously rank their choices can overcome these limitations, but a new problem arises. Votes do not correspond linearly to the power to modify the outcome of a decision. The breakthrough by Lalley and Weyl was to show that a quadratic scale in which the square root of voting credits assigned to a particular choice provides an efficient pricing of value across a broad range of assumptions about the distribution of voting credits among participants. From this theoretical foundation, an array of experimental evidence has further strengthened the role of QV in improving social choice.
In addition to efficient outcomes, blockchain communities face a second issue regarding the decentralization of decision-making. Blockchain communities share both similarities and differences with public forums and private enterprises. Permissionless, layer one chains such as Cardano are intended to be public platforms for global commerce. Nevertheless, ownership of coins, such as ADA, is a prerequisite for such commerce. Thus, a balance between open, global access to participation and the rights of ownership is sought with regard to the design of governance structures, including the allocation of development funds. 1c1v is the simplest and most widely used method of governance voting, since it provides both open participation and the security of significant ownership costs. Nevertheless, alternatives have been sought due to the danger that centralized decisions will lead to corrupt outcomes that restrict resources and diminish long-term participation.
The allocation of voting credits for blockchain communities could range from one-wallet-one-vote (mathematically, this is equivalent to raising the asset held to the power zero), 1c1v (equivalent to raising the asset to the power one), or a fraction between 0 and 1. Any fractional power will modify who has the power to influence the decision outcome, but none of these allocations will achieve the efficiency gains provided by the preferential choices inherent in the QV method. Thus, although the use of a square-root allocation has been colloquially referred to as quadratic voting, the association is misleading. Given the weaker theoretical foundation of credit allocations, modesty in the use of such methods should be the rule.
Thus, regarding the implementation problems with the Cardano Catalyst Fund 14, this incident serves less as a failure of theory and more as a valuable reminder of how easily theoretical findings can be misapplied in practice. While the quadratic pricing, which facilitates preferential choice, could have been used as a first-order guidance for allocation of credits, any deviation from this should have sided with a modest change, such as 2/3rds or 3/4ths first, given the lack of theoretical and experimental evidence. A secondary question worth examining is, given that the failure mechanism was the approval threshold, why wasn’t the threshold either removed or reverted to prior methods? The threshold does not impact the ranking, which could still have been based on the publicly announced ⅓ power. Were issues directly related to the ranking also evident?
The Photrek team remains committed to advancing high-quality governance in Cardano and other communities. More details of our research in this area are detailed in the paper, Sociocratic Pluralism: Governance for Blockchain Ecosystems. We welcome further discussion with the community on lessons from the Catalyst Fund 14 voting and broader innovations in governance design.