DAPPCON 2019: Governance Design for Positive Stakeholder Engagement - Stephanie Hurder (Prysm Group)

Transcript

good morning everybody thank you for waking up lunch it's good to see you all it's a man name is Stephanie hurt her I'm a founding economist at prism group as was just that how's this sound level on this too loud good okay so today I'm gonna be talking about blockchain governance so prism group where a blockchain economics and governance advisory firm so we're led by PhD economists have a PhD from Harvard and we help blockchain and distributed ledger projects figure out their incentive design their governance design their token economics and all that other good stuff so coming from an economics background one of the things I

think is most amazing about blockchain is rethinking how we make collective decisions right this is basically what governance is and there are a lot of claims made about governance and the ability to decentralize and automate governance that to me are really fantastic and a great example of this is the Dow I'm sure most of you are familiar with the Dow and one of the claims that was made about it you know when it was first launched was the following that had autonomous governance and I've taken this from the documentation of the platform the eath held by the Dow will never be centrally managed Dow token holders are able to vote on important decisions relating to

the management of the Dow including the power to redistribute its eath among themselves right so there was this very bold claim that we now had a platform which was going to be used for investing where there was no longer need for a central decision-maker and as I said you know based on a lot of the economics that I'll talk about to me this was really impressive and so we spend a lot of our time thinking about you know what is governance what do we mean by governance and is it possible to have truly decent Roenick decentralized autonomous governance and so this is what I'm gonna be talking about today so I'm gonna be talking about the way we think about governance and its design

having worked with dozens of projects from startups to Corp it's um and my talk is gonna have two parts so the first is what exactly are we talking about when we talk about governance I think you know projects have different definitions and we're sort of converging but it helps to start out and say you know what exactly do we mean by governance and secondly I'm going to talk about our conception of the components of governance and listening to the previous talk I actually kind of wish that that could have gone in parallel because there are some really great examples that were brought up that illustrates some of the points that we're gonna make here so first you know what is governance and

stepping back you know governance to us is one part of the overall economic design of any platform so as you know a lot of blockchain projects are rethinking how economic systems work it's one of the most amazing parts of working and blockchain right and so when we think about what are the different economic components of a blockchain platform we use this house framework to organize our thinking so starting you know the house starts from the bottom and builds its way up you know starting at the bottom you have you know the value proposition what is it that your project is trying to do this is typically something that hopefully the founding team has figured

out before we begin our discussions next there's raising capital you know how are you going to fund the development of whatever it is you're building and then there are three pillars that we think of as forming the incentive layer of the platform so there's contract design there's market design and there's information systems and all of these are different levers that can be used together to shape the incentives of the different stakeholders who come to use whatever your project is and a way to illustrate this is the following suppose that I'm building a blockchain based based version of eBay how many in here have used a bay ever so most of you are

familiar with how eBay works suppose we're making you know a new open Bazaar and we're going to a blockchain base to eBay and one of the core tenants of success for this new platform is that people need to come and behave in good faith right they're gonna buy and sell goods and you need them to actually go and do what they're supposed to do and not cheat the people on the other side of their transactions so what are the some of the levers that we can use to incentivize good behavior so first we can use contract and transaction design we can make payment contingent on measures of whether you've actually sent what you're supposed to do or paid what you're supposed to and we can have escrow we can have dispute

resolution these are all tools that help to align the incentives of both sides with the goal of doing what is in the best interests of the platform as a whole what other tools could we use to incentivize good behavior I mean we could use market design you know so for example in any decentralized marketplace they're going to be tools that help buyers and sellers find each other we could say you know if you cheat someone that you previously transacted with in any subsequent search you're gonna be penalized in the search list right you're gonna come up way lower in the list than you normally would have due to your bad behavior this is an additional incentive if you care about making any

you know profitable trades in the future for you to you know behave in the way you're supposed to and finally there are information systems right so most of the peer-to-peer platforms that we use on a regular basis have some kind of reputation where the data enclosed in the reputation comes from the other users but the design of that reputation was implemented by the platform right so if you think about eBay's rating or Airbnb the structure of the data that's collected and how its aggregated and displayed was determined by the platform founders right and then as the platform works in data is collected and used in the following way and so we can also incentivize good behavior by saying you

know if you are terrible to all your uber drivers you're gonna have a rating of four and no one's ever gonna pick you up again right so these are all the different tools that can be used in combination to shape incentives on a platform in our view a token is a tool that interacts with these different incentive levers to facilitate function on the platform right and we think about today's not talking about token economics but we're thinking about you know do you need your own token what is its monetary policy need to be that so that when it interacts with the incentive layer you get the outcomes that you want and so why did I go through all of this to us governance is the roof on the top of the

house right so once you've thought about what are the operational rules of your platform what are the different incentive levers what is the token going to do governance is the is the set of rules that allows this to evolve over time and so the more precisely you've thought about what are the incentive levers and the token economics of your platform the better you can design your governance another question we get all the time is you know these platforms are new how are you gonna do economic design for something that hasn't existed before and there's actually many many Nobel laureates who have a tackled problems

like this in fields called contract theory market design and it's basically a the economics version of a scientific process right so you start with economic theory and game theory and you say you know what are our hypotheses about how we think this system should work can we collect observational evidence that suggests that we're on the right track and doesn't immediately contradict what we think the system design should be then can we run experiments and these can be experiments in the laboratory these can be experiments in the field but can we run some kind of you know test of the system to make sure that the outcomes are going to be in line with what we want and then finally can we launch you know and this this approach has been used in everything from

designing school system matching to spectrum auctions to tax systems right and so although we're thinking about how do you design new economic mechanisms in a in an entirely new setting you don't just need to you know close your eyes and squeeze your nose and jump off the cliff there are ways to test the mechanisms that you're thinking of using to ensure that or to maximize the probability that you're going to get the outcomes that you want so all that being said you know why am I talking about this for the first five minutes because it helps frame you know what we think about the role of governance is on a platform so on any blockchain platform

you have two different types of of rules you have operational rules and you have governance so operational rules are the incentive layers I was talking about the token economics these are the written down rules that specify how the platform is going to work on a regular basis right this is what's in your code is the operational rules the governance is the part that fills in the missing pieces in the operational rules it's the piece that says you know something hasn't yet been specified but we need to make a decision the rules aren't specified in the code we've already written how are we going to come to a decision about what to do and so

one very classic example of this is the dow hack right so this is back in 2016 the dow has launched a very clever person has stolen a 50 million bucks or about a third of the money from the from the dow and there's a question of you know what do we do how do we handle this right are we going to undo this particular transaction or are we going to you know leave it the way it is and this was a question that wasn't covered in any of the operational rules right and so for those of you who know the story there was an entire series of ad hoc decisions that had to be made to figure out what was going to happen with the community right and the decision

wasn't unanimous there was a hard fork but this type of activity is exactly why you need governance right so that something happens it wasn't thought of in the original you know operational rules of the platform how are we going to think about what to do and some people think about you know they hear this definition of governance and they say okay that makes sense but my platform isn't gonna have disasters like the Dow right I only you only need governance when you have an absolute disaster I'm not planning to have any of those so I don't really need governance governance is also helpful when your platform is unexpectedly successful so

one of my favorite examples of this was the launch of crypto kitties which actually has part of its story based in Berlin so as many of you know crypto kitties launches it completely overwhelms see fear in blockchain there's this huge transaction lag this huge queue and you know the to many of the users it just seems like the watching has stopped working what do we do so the interventions that solve this crisis had two parts so first the nodes decide to raise the network gas limit this is something that's within their purview there's contention but they take a vote in their race to the network gas limit but you still have this huge backlog of transactions whose gas prices were just too low now right they had

been submitted with gas prices that were no longer competitive and absent any kind of action these transactions were never going to get processed so a group of core developers and other you know folks get together they're actually based in Berlin at the time for some unrelated conference and they develop a button that allows you to resubmit transactions with a more competitive gas price so this allows users to resubmit their transactions and starts to clear out the backlog and so the first what I think is so interesting about this example is that the first intervention amazing the network gas limit was something that had been planned for the second piece was not right and so it just happened to be that a critical mass

of decision-makers were in the right place at the right time to come up with an appropriate intervention but hopefully you wouldn't want to rely on that amount of luck in subsequent cases how many does a baton hair still own crypto kitties how many crypto Kitty owners do we have we have a couple I've got one so another question we get all the time is you know do we really need governance if we write sufficiently great operational rules is governance necessary right because maybe four years ago there were a lot of unforeseen

disasters that we had never seen in any platform but as we you know see more problems can't we just anticipate them and not really need governance as much anymore I mean the answer to this is no it's not possible to eliminate the need for governance and the reason is because of an economic concept called contractual incompleteness and what contractual incompleteness says is that it's impossible to write a contract that anticipates everything in the future that could possibly happen right no matter how much time you spend drafting a contract no matter how much time and how much detail and thought you put into all of the different cases that are

anticipated there's always something that could happen that you didn't think of when you wrote it you know even if you wanted to spend that much time it's probably not a good use of your time to sit and spend you know weeks trying to think of everything that could possibly happen and in some cases when a truly unexpected event happens the different parties might want to renegotiate so right so we might get into a situation where you know we originally thought that you know if something really strange happened this is what we would want to do that's what actually happens and that the parties look at each other and say no way actually now that this has happened we we all agree we want to do something else and so there's always

going to be a need for governance because there's always going to be something that can happen that your contract didn't take into account or didn't account for when it was written and so in some you know we think of governance as being necessary and essential for two purposes the first is updating operational rules right so you have operational rules on your platform maybe it's you know how much block producers are paid or you know the inflation rate of tokens and you decide as a group that you want to upgrade something make a change the second is when your system has failed

you when a situation arises that wasn't accounted for in the original code and you as a community you need to decide what you're going to do next and before I delve more into the different parts of governance one thing I want to point out is that it's not possible to avoid economic design right so something we hear frequently is I'll you know I've got my it's hard enough to get code written I don't really want to think about the economics of this so I'm just gonna wait and the thing is in a crisis situation I mean something's happened and a decision needs to be made

about what to do quickly someone is gonna step up to make that decision right and so then the question is a platform founder is do you want that person who steps up to be a person that you've delegated in advance right so you've decided in advance that this is someone or a set of people that we trust who are going to decide the next steps in the best interest for the platform or do you want you know whoever can accumulate as much the necessary political power to be able to make that decision and so it's in your best interest as a platform founder to specify someone in advance who will make these decisions for you rather than just hoping that someone with good intentions

shows up at the right time you know one thing that gets pointed out to me a lot is that the vigilante hero looks way more interesting than the representative but all right so that being said you know I want to spend the rest of the time talking about the different elements of governance and this is motivated by an observation that we had over the past you know years so that there's a lot of emphasis when we talk about governance on voting systems and we'll hear many dialogues that say you know it's a long debate about are we gonna use quadratic voting or majority rule and you know what's the quota gonna be and are we

gonna have low turnout waiting or you know status quo bias or stuff like this and if you think about a lot of the decision-making processes that we engage with on a regular basis there's actually way more to them than voting and if you think about what are the pieces that make these systems function well the voting is just one tiny piece so in developing this framework which we call the wheel the goal is to enumerate what do we see as the different pieces of governance that need to be considered for a governance system to work well so these are the seven elements of governance that we discuss we shape it as a wheel because typically you have to

go through this a couple of times to get a system that works that you like so you start start at the top and you go through and you say okay is this sort of a an aligned system of decision making probably not you have to refine it a little bit so starting at the top the first step is figuring out you know what are the scope of decisions that this process is going to be used to make you might have different governance processes for different types of decisions who's going to participate in that decision-making process who were the stakeholders policy research and development so one of the things I really like from the previous talk was he was saying that you know developing proposals takes work and resources and so he was doing most of the development

himself you have to think about where these resources going to come from by what means are upgrades or proposals submitted how is information about these proposals going to be distributed both in terms of what they are and the potential impacts what are the decision-making procedures that's the voting and then how is implementation enforced right how do you make sure that the decision that you come to is actually implemented by the platform so I want to talk just briefly about you know what are examples of a couple of these and why are they important so the first is the scope of decisions why might you need different decision-making processes for different

types of decisions you know at a fundamental level the reason we need governance is because we have a group of people with different opinions about what the group should do right if everybody agreed about every action in every circumstance we don't need a governance process we just need to pick a person and we just do whatever that person says because we know that that person has the same opinions as the entire group right so we need a governance process because one of the two things one of these two things has happened and you actually have different you you'll want different types of outcomes depending on which situation you're in so what does an example of this could we go to the next slide there

we go what if we have different preferences so you know my when I was growing up my dad and I always used to argue about what kind of ice cream to get he likes chocolate I like strawberry and in situations like this you know there's a fairly easy compromise that can be taken when you have people with different preferences you try to come up with something that's a mix right so the solution is we end up going and getting that weird striped ice cream that's you know chocolate strawberry and vanilla in the middle and so compromise is able to satisfy both of us to a reasonable extent when we want different things from the system so when there are different preferences frequently you can say okay what is something in the middle

that will be sufficiently satisfying to both parties in contrast when you have different information this is a type of situation where for example we all want the same outcome we're all on the board of directors of a company and we want to maximize shareholder value because we're in the u.s. and this is what we care about at least for this example but we differ in how we think this is best done right so some of the people on the board think that the best path to maximizing shareholder value is investing in research and development the other side prefers to acquire another company and the thing about this type of situation is that a compromise isn't really going to work right if you say ok we're going to take the money and split it in half

and have half as big an R&D team and then acquire a company of half the size that's not really going to have a good outcome in fact that's probably be worse than picking one or the other right so in a situation like this your decision-making process maybe won't want to come something to a compromise decision you want to pick one or the other in terms of stakeholders in what I think one of the most interesting developments over the past six months to a year is the number of platforms that are developing some kind of representative decision-making process right so councils groups of block producers you know elected bodies of various types that are going to make

decisions on behalf of the entire community and from our point of view this kind of decision-making process makes a lot of sense you know the vast majority of decisions in our daily lives whether it's in you know federal governance or corporate governance are not made by the entire community they're delegated to a set of decision makers and from an economic point of view this makes a lot of sense you know why is this why might we want to delegate decision-making the first is that you know as an economy we thrive on division of labor right there's too many different things in the world to consider to be involved in all of them

right so we divide up tasks some people do something some people do others and there's a whole set of things that honestly I don't really care about enough to spend time and energy you know time energy thinking is fine our finite resources and so we might just want to say you know it's not worth it for everyone to be involved in everything the second is specialization so for example you know suppose that I find out I have a devastating disease and I want to crowdsource potential treatments for my disease I'm probably not going to just go into downtown Berlin and ask the first hundred people I see what I should do if I'm gonna crowd source information

I probably want to talk to specialists okay doctors who have expertise in the area of my illness and finally information distribution systems and this is actually my favorite area of the wheel so if you think about a lot of decision-making processes from you know voting for candidates to selecting policies we rely on people giving us information about what these pop what the implication of the different policies are as well as if we have a representative government we rely on aggregated information about

the decisions that our representatives are making so that we can figure out whether they're there doing what we want them to do so an example of this was this spreadsheet that exposed the collusion between EOS block producers so for those of you who aren't familiar block producers on use we're colluding with each other and this was exposed by a spreadsheet that was collected by Hobie who I believe is one of the block producers and so it wasn't until you know this information came out that people realized that the decision-makers in the group of block producers weren't quite as independent as people thought and so rather than relying on one of the users in a leak to expose this kind of

information proactively setting up for example voting results if you're developing a platform and you're going to have a representative council proactively developing this kind of information system allows users to see what representatives are doing and to trust you know what's going on and so I think I have about one more minute so let me just finish two questions that I get a lot so the first is is on chain governance optimal and the second is can governance by itself present prevent hard Forks and the first piece is is on chain governance optimal

I think that having some portions of governance on chain is very helpful but to us the most important piece is that you've accounted for the different required pieces of the system so whether that's on chain or off chain you know the main goal is that you have a transparent well laid out system with all the different component parts and however that's implemented is is fine and can governance prevent hard Forks in term I also really liked from the previous talk was this concept of ragequitting right when do you get hard Forks you get hard Forks when a set of users is so upset by the outcome or disagrees cares about the outcome enough

and has a sufficient mass that they're willing to fork and start their own community and so we've done some research on this they won't have time to cover but basically what we find is that voting processes by themselves are not good at preventing hard Forks if you want to minimize the probability of hard Forks you need to leverage the other pieces of the wheel so you need to think about the policy proposal process information distribution that's going to be far more effective at preventing hard Forks then switching from majority rule to quadratic voting all right so as I said I'm Stephanie herder I'm with prism group if anyone's interested in discussing the industry or

governance or any other fun pieces this is the Khan my contact information and also that of Reed Cataldo who is another founding partner who's here alright thirty seven thirty seconds over you