Hey Arbitrum community,
We’ve been tracking Arbitrum governance health daily since January as part of ChainSights, an independent governance analytics platform covering 53+ DAOs.
Here’s where Arbitrum stands today (March 18, 2026):
- Overall Score: 6.2/10 (C) — #17 of 53 DAOs, #3 in the Infrastructure category
- Human Participation Rate: 9.9/10 (A+) — excellent signal of genuine human engagement
- Delegate Engagement Index: 1.9/10 (F) — the critical weak point
- Power Dynamics Index: 4.8/10 (D) — voting power concentration needs improvement
- Last Proposal: March 12 — “Automate the Consolidation of Idle Funds into the Treasury Management Portfolio” (6 days ago)
Arbitrum ranks above the Infrastructure average (5.6) and above the all-DAO average (5.9), but the delegate engagement score tells a different story. The DEI has dropped significantly over the past month — from ~7.0 in mid-February to 1.9 today.
What this means: Arbitrum has the participation infrastructure, but delegates are disengaging. The governance “muscle” is there — it’s just not being used.
For delegates: We built a free Governance Score Card that benchmarks your DAO’s governance health — no signup required. Run a quick check, and if you want the detailed PDF with your rank, benchmark comparison, and key insights, you can grab that too.
→ Check Arbitrum governance health: Check your DAO
→ Full Arbitrum profile: Arbitrum Profile
Happy to answer questions about the methodology or the data.
— Mario (ChainSights)
2 Likes
Mario 
this data is genuinely valuable, and the timing is striking.
The gap between Human Participation (9.9/10) and Delegate Engagement (1.9/10) tells a very specific story: ordinary community members are showing up, but the people with actual voting power are disengaging.
This is precisely the problem I’ve been researching independently. I published an analysis two days ago arguing that token-weighted governance structurally misaligns voting power from genuine participation:
https://paragraph.com/@0x2603ea2e2953c61a948da819549aeb18ea36e890/
The DEI dropping from 7.0 to 1.9 in one month during an active Security Council election is a data point worth the community taking seriously.
Question for you: does ChainSights track whether delegate disengagement correlates with token price movements…..? My hypothesis is that delegates participate when financial incentives align and disengage when they don’t. If that correlation exists in your data, it would significantly strengthen the case for contribution-based voting models.
MconnectDAO Freelance Researcher | Based in India | Still learning but paying attention.
Disclosure: I have no financial relationship with any protocol, candidate, or entity mentioned. I hold no governance tokens. I am an independent researcher. @ChainSightsOne @Arbitrum
1 Like
@MconnectDAO — appreciate the depth here, and your Paragraph piece is on my reading list.
To your question: ChainSights doesn’t currently cross-reference token price data — we’re purely governance-layer. But your hypothesis maps directly to what we’re seeing in the DEI signal. The 1.9/10 during an active Security Council election is exactly the kind of anomaly that shouldn’t exist if financial incentives were the primary driver.
Which actually challenges your hypothesis in an interesting way: if delegates participate when incentives align, they should be most active during elections with real stakes. The opposite happened here.
Either the incentive structure is broken at the delegate level, or the disengagement is strategic. Both scenarios are worth measuring. I’d be curious what your analysis found on the contribution-based voting side — if you’re seeing structural misalignment in the token-weighted model, our DEI data might be useful evidence.
2 Likes
That counter-point is worth sitting with.
If delegates should be most active
during high-stakes elections and the
opposite happened, the financial
incentive explanation breaks down.
Which leaves two possibilities as you
noted: structural failure or strategic
disengagement.
My research leans toward structural.
Token-weighted models reward holding,
not participating. A delegate with
large token weight has little
accountability pressure to show up
consistently because their voting power
does not diminish with absence.
Contribution-based models attempt to
fix this by tying influence to
demonstrated engagement over time. The
challenge is measurement
participation can be gamed just as
easily as token accumulation.
If your DEI data includes delegate-level
breakdown, it would be interesting to
check whether disengagement is
concentrated among high-weight delegates
or distributed across the board. That
distinction would help separate
structural misalignment from broader
apathy.
Happy to share my analysis in more
detail if useful.
1 Like
The structural argument is compelling, and the accountability asymmetry is exactly what DEI captures indirectly — but you’re right that we don’t yet break it down to delegate level.
Current DEI measures aggregate engagement across the active delegate pool. We can see that high-weight delegates are disengaging as a group, but not isolate whether concentration at the top is driving the number. That’s a real gap in the data.
Your distinction matters: structural misalignment would show disengagement concentrated among top-weight delegates. Broad apathy would show it distributed. Those are very different governance failure modes.
I’d be interested in your analysis — particularly if you have delegate-level participation data from Arbitrum. If we can combine your structural framework with our longitudinal DEI signal, there might be something worth publishing together.
2 Likes
Thank you for engaging with this.
I don’t currently have delegate-level
participation data from Arbitrum
my framework is based on structural
analysis of the token-weighted model
rather than raw on-chain data.
If your DEI signal can be broken down
by delegate weight tier, that would
be the missing piece. The combination
could make a strong case for
contribution-based accountability
models.
Happy to explore this further…..
2 Likes
Good to know the boundaries of each dataset. Delegate-weight-tier breakdown is technically feasible on our end — not live yet, but buildable.
Let’s take this off the forum. Drop me a DM here or on X (@X_Sem_pre) and we can figure out if there’s a structured collaboration worth pursuing.
1 Like