Final edit: The 3.75 billion period 1 committed Algo value is incorrect. I published a new post summarizing how I arrived at the erroneous amount, how I identified the issue, and how it was solved, here. I'm going to leave this post up for transparency.

Edit: Some reddit feedback rightly pointed out that the algorandstats.com governance period 1 page doesn't show an initial commitment Algo count of 3+ billion. I traced the discrepency to an Algorand wallet account that made several very large period 1 commitments -- the "mysterious whale" I discuss later in this post. I'm not sure if this account is a test account of some sort and should be ignored, so I'm going to reach out to the Algorand Foundation to get clarification. If you have an answer, please comment! I'll update this post when I get more information. 

There are already fewer governors eligible for Algorand's second governance period than the first at an equivalent time -- and its only been a little more than week since registration closed. Those who remain by April should expect a handsome reward for their participation if the governor attrition rates are similar. 

Here's why I think there's a good chance period 2's rewards will be higher, as well as other observations I stumbled across while examining Algorand's governance data. 

Primer on Algorand Governance

Algo holders commit a number of Algos in order to have voting rights on measures brought forward by the Algorand Foundation. Check out an earlier post -- "Algonauts: don't forget to commit your Algos for Governance period 2!" -- for more details.

Governance period 2 started Dec. 24, 2021 and won't close until March 31, 2022. Period 2's registration closed on Jan. 7, and governors cannot move their committed Algos for the period's duration to remain eligible. 

Data sources

All governance data was queried from the Algorand Foundation's governance API on Jan. 15, 2022 with Python. 

Committed Algos: period-1 vs period-2

Algorand's first governance period began with over 72k governors and 3.75 billion committed Algos. Over 22k governors lost eligibility before the first period finished, leaving 1.71 billion Algos committed by the governors who remained eligible to split the 60 million Algo rewards pot. 

Algorand's second governance period closed registration with 68.5k governors -- 3k less than the first period's equivalent count. Period 2's governors to date committed 3.15 billion Algos, and those remaining by April will split the 70.5 million Algo rewards pot. 

Governance period 2 has fewer governors with fewer committed Algos to split a larger rewards pool -- and the period's barely started. Less committed Algo and a larger rewards pot mean more rewards for the governors who remain eligible at the period's end. More period 2 governors will certainly lose eligibility and move their committed Algos between now and the second period's closing date March 31.

I'm not sure if we'll see period 1's changes repeated during period 2 however. Period 2 governors would have more time to understand the consequences of moving their committed Algos, especially if they participated in period 1. I would not be surprised if period 2's lower participation actually means more informed engagement. 

Algo whales: period-1 vs period-2

We can divide each governor's committed Algo amount by the period's total committed Algo to get an idea of which accounts hold the most governance "power." The resultant scatterplot reveals a single account held over 35 percent of period 1's committed Algo with a near 1400 million Algo commitment before it lost eligibility later in the period. 

The 1400 million Algo commitment was by far the largest single commitment of period 1 or period 2, so much so that it pretty much ruins the plot by grossly extending the X and Y axes. I wonder who or what controlled the 1400 million Algo account. I'm sure a lot of the larger commitments represent exchange wallets, but it seems abnormal that one exchange would control more Algo than its peers by a factor of 10. 

Solving period 1's fat whale mystery will have to wait however, and we can get a more informative graph by restricting the X and Y axis bounds to more reasonable values. 

Two things stand out to me:

• Period 1's largest governors all lost eligibility at some point before the period's end. The largest governance address that remained eligible held just under six percent of period 1's total commitment. 
• Period 2's distribution contains no governance address with more than five percent control. This does not mean however that period 2's governance is more decentralized, because a single entity could always spread their Algo commitment across several addresses. 

Final thoughts

The Algorand blockchain has really caught my interest, especially the Algorand governance API. There's a few more governance analysis ideas I want to pursue but I'm always looking for more -- drop a comment if you're curious about more details on this post's contents, or other Algorand aspects. 

Thanks for reading! 

EDIT: The original post contained plots that incorrectly converted microAlgos to Algos. The plots are now fixed. You can view the git pull request that implemented this fix here. 

Thumbnail photo by Kanchanara on Unsplash. 

This analysis was completed in Python. View the source code here: https://github.com/simplyrangel/crypto-publish0x, in the 2022-01-post-2 directory.