Why Direct Democracy and Liquid Democracy continue to breakΒ DAOs.

Do you want to know why Direct Democracy (DD) and Liquid Democracy (LD) systems keep breaking DAOs and will continue to break them? πŸ’₯

( hint: they both have fatal flaws. πŸ› )

Years of thinking broken down below πŸ‘‡

Listen to the audio version of this post hosted forever on Arweave here: arweave.net/1v9k4cVGhjW2GLHGH5o0EvsDXZrqsOveqoYjNPdOt4Q

Here's the TLDR:

Direct Democracy (DD) bugs:

  1. Apathy (voters don't care)

  2. Ignorance (voters don't know)

DD's naturally drift to become LD's.

Liquid Democracy (LD) bugs:

  1. Ignorance problem (proxies don't know)

  2. Bandwidth problem (too many voices)

  3. Apathy problem (voters set and forget)

  4. Corruption/purge phases (power corrupts)

Who am I?

I've been studying/thinking about governance for 10 years, and watched dysfunctional DAO after DAO execute poorly, grid lock, and collapse for the last 5 years.


Direct Democracy (DD)

At first glance DD's "everyone voting on everything directly" sounds great.

But it's not.πŸ’₯

Let's start with Apathy:

Apathy is a state of "lack of interest or concern".Β  πŸ˜‘

Apathetic voters are eligible to vote, yet they lack the interest or concern to vote.

Apathy is a lack of concern or interest and can have deadly consequences for the DAO.
Apathy is a lack of concern or interest and can have deadly consequences for the DAO.

DD DAOs usually start good with a group of highly engaged voters. πŸ₯³πŸŽ‰

However, over time the honeymoon period wanes πŸ₯±, and people get voter fatigue.

Why do DD's get voter apathy again and again?Β  Because we're failing to see that we're asking too much of voters. πŸ˜“ πŸ“š

Why would I repeatedly spend hours of my time reading proposals where my vote only counts for 0.04% of total voting power? Most wouldn't. 😡

Example: Today we have multi-billion dollar DD DAOs with under 5% voter turn out. πŸ—³That's less voter turnout than traditional nation states or corporations. πŸ“‰ Let that sink in for a second. 😬

Low voter turnout in a DD is not healthy πŸš‘ because:

  1. Bad actors can more easily form cartels that have an outsized influence on proposals, which can lead to poor quality proposals or cash grabs getting passed.

  2. Higher risk of DAO gridlock/death.

Paying for voter engagement πŸ’Έ

A common thought is: couldn’t you just pay voters for engagement to solve apathy?Short answer is no, but why?

Because (value paid to voters - cost of engagement) needs to be a large enough positive number that the majority of members decide it's worth their time to engage regularly.

The amount of payment needed to sustain πŸ”voter engagement long term is likely too high for the DAO to sustain, (like paying every voter a part time wageπŸ’°) in order to carefully πŸ”Ž read, research and review each and every proposal. πŸ“š

Additionally, when you pay voters for engagement, this opens up a negative incentive towards voting quickly without research to farm voter engagement rewards. Farming further dilutes the voters who carefully research their votes, which is suboptimal.

Paying for engagement is not a viable solution for apathy. 🚫

To sum up the Apathy problem in DD:

At first, people are excited and vote a lot 🀩 -> People get tired πŸ₯± -> People stop votingπŸ˜‘ -> Less voters opens door for bad things to happen πŸšͺ -> Then bad things do happen. πŸ”₯🏠πŸ”₯

Or even more simply:

Apathy in Direct Democracy = Death. πŸ’€


Next, let's look at the Ignorance problem in DD:

When "everyone is voting on everything directly" you run into a problem of people voting on things they may know little or nothing about. πŸ€”

Even if you are an expert πŸ€“ in the field that concerns the proposal, you will be out voted by the majority of voters who are not experts in that field.

Ignorance Problem: Experts get out voted by non-experts, which cause suboptimal choices to be made and those lead to suboptimal outcomes.
Ignorance Problem: Experts get out voted by non-experts, which cause suboptimal choices to be made and those lead to suboptimal outcomes.

πŸ€‘πŸ˜ˆπŸ€­πŸ€€πŸ§πŸ€©πŸ€ πŸ₯± Β  > Β  πŸ€“πŸ€“πŸ€“

To illustrate the point let's tell a story:

Let's say someone submitted a snazzy dev proposal for a DD DAO, which sounds very convincing and has the majority of voters excited.

However, a specialized dev notices some subtle red flags in the proposal that most people don't catch.

The dev happens to know the feature promised in the fraudulent proposal can't actually work that way. The expert dev knows this is a cash grab.

The dev tries to warn but since its highly specialized, few understand and the less specialized majority voters pass the proposal and the DAO gets bamboozled. 🚨 πŸ’Έ

Simple vs Complex in Direct Democracy (DD)

Ignorance in DD isn't a problem when the voting issues are simple:

"We're shipwrecked, should we build a hut?" βœ…

However, once you add complexity to decisions:

"What's the best way to build a hut while using the least materials possible?" πŸ˜΅β€πŸ’«

then ignorance kicks in.

When you ask a DD to vote on complex topics like:

"Here are three proposals for building a nuclear power plant. Please vote for the best one."

DD's are nearly guaranteed to have the majority of voters voting on issues they lack the knowledge to vote effectively on.

In DD, complexity of questions leads to poor decisions being made due to ignorance.

But lack of complex decisions being made skillfully also leads to poor performance.

Either way you slice it, if "everyone votes on everything directly" you will always get sub optimal performance.

To sum up the Ignorance problem in DD:

When everyone votes on everything, and everyone can't know everything. Then the majority of voters will vote on things they don't know and make poor decisions.

Repeated poor decisions lead to bad outcomes for the DAO.

The suboptimal choices stack up and pull the DAO off course.
The suboptimal choices stack up and pull the DAO off course.

Bugs for Direct Democracy are:

  1. Apathy πŸ₯± = cartels, hacks, gridlock, death.

  2. Ignorance 🀷 = fraud/bamboozlement, and guaranteed sub-optimal performance.

Is this as bad as it sounds? How many high functioning DD DAOs have you seen actually work? There's your answer.

This leads us to the next argument:


Direct Democracy + Entropy = Liquid Democracy

A Direct Democracy will naturally drift to become a Liquid Democracy over time.

As voters in DD's get tired they naturally ask:

β€œWhat if we let voters proxy their vote to someone who is both smart and will spend the time and effort required to engage with all proposals?” (AKA convert from Direct Democracy to ➑️ Liquid Democracy)

At first this seems to be a positive step forward.

All those tired voters can proxy their vote to someone they trust to do the work. And if the proxy they delegate to does a bad job, voters will simply change their proxy to someone else.

Since a proxy can have a lot more voting power -> that proxy will feel like the time they spend reviewing a proposal is more worth their time -> thus proxy/voter participation increases -> which can solve the apathy problem.

Because of these positives on the surface, this will cause dysfunctional DDs to adopt LD in an effort to make the DAO less dysfunctional. Problem is, LDs have just as many if not more problems as DDs.


Problems of Liquid Democracy

Firstly, the ignorance problem is still not solved in LDs. Just because a proxy has been delegated votes, does not mean they will be an expert on all proposals they review.

Even if you could fix the ignorance problem with a team of experts supporting the proxy, LDs have a new problem added into the mix that DDs do not have:

The bandwidth problem = too many voices for one person to accurately represent.

The bandwidth problem is simple:
The more voters delegating to a proxy, the more competing voices for that proxy there will be.

The bandwidth problem = too many voices for one person to accurately represent.
The bandwidth problem = too many voices for one person to accurately represent.

β¬‡πŸ˜πŸ˜£πŸ€ πŸ™„πŸ§πŸ˜’πŸ€“πŸ˜žπŸ˜€πŸ₯±β¬‡

πŸ˜΅β€πŸ’«Proxy

β¬†πŸ€‘πŸ˜ˆπŸ₯³πŸ€”πŸ€ͺπŸ˜‡πŸ˜’πŸ€­πŸ₯ΊπŸ˜Žβ¬†

Result: some voter signal is ignored.

There is a trade off between ultra powerful proxies which may have enough resources to hire experts, but are so overwhelmed by the bandwidth problem that they cannot represent all voices so usually end up doing what's best for themselves.

OR

You have a large number of proxies which can more accurately (but not perfectly) represent the voice of their voters, but are less resource backed or expertly informed, causing them to make sub optimal choices similar to a DD.

⬅️ Many weak proxies  ↔️ Fewer powerful proxies ➑️

The bandwidth problem is not a boolean on or off value, it's a gradient.
The bandwidth problem is not a boolean on or off value, it's a gradient.

Bugs πŸ› of DD are nested into LD

Additionally, you still have LD voters who need to decide which proxy to delegate to which means:

  1. Apathy problem still exists (Voters will get tired of spending time choosing and monitoring the right proxy)

  2. Ignorance problem still exists (Voters don't always know who they should delegate to)

Meaning, the problems the DAO was trying to escape from in DD still exist in a 2nd β€œnested” layer state inside the LD system.

Enjoying the content? Subscribe for more by clicking the button.

Corruption

Now we get to the ultimate fatal error for LD's and all representative democracies:

Representation + Entropy = Corruption

Corruption problem = proxies/representatives have a misaligned incentive to choose actions that keep themselves in power, at the cost of the people they represent.

Corruption comes in the forms of:

  1. Bribing for proposals to be passed πŸ’°

  2. Pressure from other proxies to support their proposals🀐

  3. Doing anything necessary for a proxy to stay in power 😈

Proxies/Representatives have a misaligned incentive to choose actions that keep themselves in power, at the cost of the people they represent.
Proxies/Representatives have a misaligned incentive to choose actions that keep themselves in power, at the cost of the people they represent.

Because we live in a constantly changing world, a proxy is either gaining πŸ“ˆ or losing πŸ“‰ power, and thus not actively seeking to expand power will likely lead to a loss of power.

Proxies can use their Power to gain Money, Money can be spent on Influence, Influence can be used to expand Power.

More power ➑️ more money ➑️ more influence ➑️ more power.... round and round. πŸ”
More power ➑️ more money ➑️ more influence ➑️ more power.... round and round. πŸ”

The need proxies have to keep power feeds the corruption 🦠 as it grows and festers like a cancer in its' LD DAO host and as power becomes more centralized in the hands of fewer proxies, corruption and its adverse side effects systematically rise.

Over time corruption becomes worse, proxy choices are made more for the reason to stay in power rather than to benefit the members or the mission of the LD DAO.

Corruption can be extremely damaging to the LD DAO in financial, legal, and reputational terms.

The dysfunction brought on by corruption 😈 lowers the morale of the LD DAO 😞. And as stated before, there's no guarantee voters keep up to date with if proxies are acting honestly.

All this causes other members who were paying attention, to leave out of frustration, anger, or disillusionment.

Eventually one of two things happens:

  1. The LD DAO becomes so corrupt that it no longer functions and kills itself.

  2. It goes through a purge cycle where the voters who care about the LD DAO change their proxy on mass to a populist proxy type promising a return to honesty. πŸ”

Another way to look at the outcomes for the DAO are like cancer in a person:

  1. Person dies from the cancer. πŸ’€

  2. Person removes the cancer from the body. πŸ”ͺ

Except in this case, the cancer will always return because the LD system naturally creates the cancer. πŸ”

In other words, if the LD DAO is lucky enough to survive the extreme damage caused by corruption, then it will be doomed to repeat the cycle again when the new proxies eventually become corrupted by the same system which corrupted the previous power group.

To sum up the bugs in Liquid Democracy:

  1. Doesn't solve the Ignorance problem (lack of knowledge to vote)

  2. Bandwidth problem (too many voices for one person to represent)

  3. Doesn’t solve voter delegation apathy (voters may not pay attention or respond to proxy misbehaviour)

  4. Corruption problem (power corrupts, which causes highly destructive corruption/purge cycles)


Conclusion

My main argument in this post is that our DAOs are failing because both DD and LD are fundamentally broken to their core and will always lead to sub optimal outcomes.

Given enough time, DD and LD systems naturally become more and more dysfunctional. The common action for dysfunctional DD's is to move to LD, while a dysfunctional LD swings between corruption/purge cycles.
Given enough time, DD and LD systems naturally become more and more dysfunctional. The common action for dysfunctional DD's is to move to LD, while a dysfunctional LD swings between corruption/purge cycles.

Since 2018, I’ve watched DAO after DAO fail. We need to recognize the insanity of spinning up the same old dysfunctional DD or LD DAO over and over only to get the same result.

What's the solution?

We need to innovate. πŸ§ͺ

What are the stakes if we don't? The dysfunction or corruption never end and humanity pays the price.

Dysfunctional group decisions have huge consequences.
Dysfunctional group decisions have huge consequences.

The price of not acting to make better governance models could be beyond our comprehension:

  • In the worst case, it might not be hyperbole to say that dysfunctional governance is an existential threat to our species.

  • While in the best cast, trillions of trillions of future value is either created πŸ“ˆ or destroyed πŸ“‰ if we do or do not build better governance systems.

Better governance could lead to:

Better decisions being made more often, better usage of human talent, better capital allocation, better technology R&D, better living conditions, a better humanity. 🌻

When you improve the quality of decisions made by the group, you get better outcomes for the group.
When you improve the quality of decisions made by the group, you get better outcomes for the group.

What can we do? πŸ—πŸ‘·πŸ› 

  1. Believe that governance can be improved.

  2. Talk about it and tell others.

  3. Help others who are working on governance.

  4. Run governance experiments.

  5. Think all we can on better governance.

  6. Share this post to help spread a governance wildfire πŸ”₯πŸ”₯πŸ”₯

Lastly, while this post has been focused on problems of existing systems, I'd like to leave you with something that excites and fills me with hope for the future.

Introducing Outcome Democracy (β€œOD”)

I'm building a new governance primitive which I coined Outcome Democracy (β€œOD”) back in 2018 and we've got a small team working on it at an org called MotherDAO.

OD operates like a DD, in the sense that everyone gets to vote, however it's nothing like you are probably imagining.

Counterintuitively, you don't vote on what or how something should be done. Instead, you vote on the destination or "outcome" you'd like to reach. 🏝 β¬… πŸ›³

Outcome Democracy targets 🎯 improvements for Apathy, Bandwidth, Corruption, and Ignorance problems.

I'm excited to explain OD to you in the next post. πŸš€


Join the MotherDAO community: https://discord.gg/KMfUndbgpr

Follow MotherDAO on Twitter: @mother_DAO

Original Twitter thread from the author: https://twitter.com/AFurmanczyk/status/1567576463251210240

Audio version of this post is hosted forever on Arweave here:

arweave.net/1v9k4cVGhjW2GLHGH5o0EvsDXZrqsOveqoYjNPdOt4Q

πŸ™ Thanks for reading -❀️- Andrew Furmanczyk πŸ™‹β€β™‚οΈ

Subscribe to afurmanczyk
Receive the latest updates directly to yourΒ inbox.
Mint this entry as an NFT to add it to your collection.
Verification
This entry has been permanently stored onchain and signed by its creator.