Decentralised Model in Detail - Decision Making Models
Next: Money, tokens, financial flows
Decision Making Models
Finding the Right Decision Model
How to find the right decision-making model for the decentralized world? Decisions are being done on two different levels: in decentralized local communities and in global, fee-to-join decentralized autonomous organizations or Internet Native Organizations.
Local decision making is easier. In small communities everyone can just gather together to discuss and vote or make unanimous decisions. In somewhat larger local communities like city level, representative democracy might work fairly well as the decision makers live with the rest of the community and are known. But if the cities are big, this bond starts to break up.
It's time to think a little bit more.
Voting is not a good method for decision making but for dispute resolution. In representational democracy the dispute is over who gets to make all the decisions, in company setting the choices are directly presented.
In traditional model all the essential decisions are made before the vote. Voters never see what other options are possible or if there would be a larger problem in society worth solving. This is called opportunity cost – what else we could have done with the funds.
Neither can voting be used to make decisions about the myriad of small details in the proposal. Nor do voters usually do rigorous background test to verify that the presented data is valid and without error.
To get good ideas and make good decisions, a number of things can be done
Training for decision makers and anyone wanting to propose ideas. See more on post on public decision making.
Training how to use and analyze publicly available data from society and access various interface (APIs) and data sets
Making society’s data openly available so people can form their own opinions from validated information
Then it all starts from generation ideas. We’ve covered alternatives
Open Ministry/Municipality is the concept of crowdsourcing ideas for change. If an idea has enough support, it goes for official handling.
Citizen’s Automation Budget where taxpayers directly control part of the tax funds, band together and fund automating some part of the public operations bringing costs down.
In Pyramid Model small group of people discuss about something. Each base level group nominates representative to the next level until a small set of decision makers each representing their subtree makes final decisions. Decisions are explained back the tree.
Machine Learning can help by grouping and summarizing ideas from a wide range of people in the “pyramid” and communicate back what others have though of. Or ML models can summarize laws and principles from other countries.
Before going to formal voting, the proposals need to be validated. This can happen
On discussion forums people comment and improve proposal ideas before they are formalised. Society can pay for some legal experts to help formulate ideas and check how existing legislation affects them,
Prediction markets are a way of gamifying the crowd’s opinion. People bet how likely some change is to succeed and winner take in the other side’s bets. Gives an idea how much of public speaking is advertising (as viewed by participants)
Simulations should be available showing some kind of idea how fragile the proposal is to changes in input conditions and what it benefits might be on a good day. Token engineering is a method for building simulators and seeing where and how often even the best laid out plans of mice and men go awry.
Key Performance Indicators (the target improvement) need to be defined for simulations (and for final implementation) and what happens when they are not met.
Field trials
Field trials show how real people behave in real world. Its known that people change their behavior when they know that they are in the test group so this is not the final truth but indicative.
Voting models
There is a myriad of models for who can vote and for counting votes.
We’ve covered some of them such as token (share) voting, person voting, squared votes, conviction voting, using gini-index in balancing between token and person voting (sogur model). See for example Decentralized Organisations - Decision Making and Decision Making in a Decentralised World and
It’s been shown that there is no perfect voting model. So called Arrow’s theorem says that there is no perfect solution to governance and the party that forms the questions and outlines how the problem should be solved, direct the decision to a particular direction.
In more detail it says that there is no electoral system based on members ranking options that satisfies three fairness criteria:
If every voter prefers alternative X over Y, the group prefers X over Y
If preference between X and Y remain unchanged (but other preferences change), the preference order between X and Y remains unchanged
There is no dictator, no single one having power to choose the outcome
Fundamentally participants have different interest and worldviews and it is sometimes impossible to find alternatives that satisfy all. There may not be a correct solution to a problem that everyone agrees.
Better information helps to build shared understanding and people to agree on a wider set of things that are true, but issues will remain where disagreement exists and interest differ.
The post on solving complex problems proposed one way to find common ground when people have dissimilar interests, based on Alexander de Haan’s book Solving Complex Problems: Professional Group Decision-Making Support in Highly Complex Situations
Implementation
Governmental Kanban is an implementation method where one of the original implementors participates to the daily work of implementing the change
If the implementation requires third parties like enterprises to do something, a reference implementation in code is released as part of implementation to showing in clear terms what is needed to fulfil the requirement (at the same time it saves costs)
Monitoring
Once the change is implemented, it needs to be monitored to see what the impact is. Some people find and use loopholes only after a longer period of time
When things go pear shaped, corrective actions are needed. This results in alarms and some form of process handling the error. This was covered in xxx
Defence of decision making
The decision process can be corrupted even when monitoring is implemented. People have limited time to follow up all developments leading to specialization and formation of elites that have deep know-how on a particular topic. Elites tend to subvert the system by slowing introducing changes that favor them.
To prevent this, the foundational rules need to have defenses against the very people who are best equipped to run and take it forwards.
We covered this by referring this a16z post where recommendations included constant upheaval of elite class for example with term limits and by setting up the system so that there are at least two competing teams of elites permanently.
The various ways to subvert decision making system and mitigations were discussed in the post on decentralized decision making
More Defence for Decision Making
In the article on decentralised decision making we covered additional topics on how governance can grow pear shaped and what’s possible to mitigate.
First “Sogur Way” was based on notion that both person and share (stake) based voting models have fundamental flaws. Person based model does not reward adequately people who work for the community and have skin in the game and stake based can result in the wealthy and powerful having most voting power.
They adapted voting power dynamically based on how equally the resources are distributed based on Gini-coefficient as a proposed solution
The a16z article took the machiavellian approach based on notion that there are always elites (key persons doing central decisions either directly or indirectly) and human condition is such that sooner or later they start bending rules in favour of themselves rather than the community. They recommended having methods like term limits (no one can rule too long) and constructing two combatting teams of elites prevent or slow down elite take over (divide and conquer written in constitution serving all).