Stake Distribution of Avalanche Validators

The graph above shows the cumulative stake distribution among the Avalanche validators: Since (at the moment), there is no progressive taxation on the larger validator rewards, the shown GINI inequality coefficient is applicable to wealth (i.e. stake) and income (i.e. rewards):

1. The red line plots the cumulative stake of validators including delegations. The shown GINI coefficient is applicable to this plot of total weights only.

2. The blue line plots the cumulative stake of validators excluding delegations, where its GINI coefficient is not shown, but it is obviously in the same ball park as for the red one.

3. The diagonal dotted line is the "frontier of decentralization", i.e. if all stakes were distributed equally among all the validators, then the red and blue lines would overlap with this dotted diagonal.

4. And finally, the vertical dashed-dotted one is the "line of control" at the 30% -vs- 70% split (w.r.t. total stakes): Hence, the many smaller validators on the left hand side of the line control only 30% of all stakes, while the few larger validators on the right hand side control a whopping 70%!

GINI and Wealth

Despite having such an extreme wealth GINI, Avalanche suprisingly seems to feature less wealth inequality than the global average of 88.5% (in 2019) and about the same as for the United States of 85.2% (also in 2019). Although extreme, these two comparisons are perhaps an indication, that there should not be a wealth tax (i.e. partial confiscation and redistribution) on the existing larger validator stakes.

GINI and Income

However, if we compare the same GINI in terms of rewards (on the stakes), then it is obvious that Avalanche displays significantly more income inequality than the global average of 38% (in 2002) and also more than the United States of 47% (in 2014). This gap is a clear indication, that the Avalanche validator community should introduce progressive taxation on the larger validator rewards (but not stakes), and then redistribute the collected tax among the smaller ones.

Prerequisite(s)

Name        : python
Version     : 3
Description : Next generation of the python high-level scripting language
URL         : https://www.python.org/

Name            : jq
Version         : 1.6-4
Description     : Command-line JSON processor
URL             : https://stedolan.github.io/jq/

Name            : geoip
Version         : 1.6.12-2
Description     : Non-DNS IP-to-country resolver C library & utils
URL             : https://www.maxmind.com/app/c

Name            : geoip-database
Version         : 20210921-1
Description     : GeoIP legacy country database (based on GeoLite2 data created by MaxMind)
URL             : https://mailfud.org/geoip-legacy/

Name            : geoip-database-extra
Version         : 20210921-1
Description     : GeoIP legacy city/ASN databases (based on GeoLite2 data created by MaxMind)
URL             : https://mailfud.org/geoip-legacy/

Installation

$ ./setup.sh ## setup virtual python environment

$ source bin/activate ## activate virtual python environment

[avalanche] $ ./setup.py install ## setup python dependencies

Fetch Stake Distribution

$ ./stakes.sh -h

Usage: stakes.sh
    [-a|--api-host=${API_HOST-https://api.avax.network}]*
    [-h|--help]

$ ./stakes.sh ## fetches stake distribution

Plot Stake Distribution

[avalanche] $ ./stakes.py -h

Usage: stakes.py
    [-h] [--seed SEED] [--gini-00] [--gini-33] [--gini-66] [-s]
    [-x EXPONENT] [data_path]

Plots Avalanche stake distribution

positional arguments:
  data_path             path to validator data (default: None)

optional arguments:
  -h, --help            show this help message and exit
  --seed SEED           random generator seed (default: 1)
  --gini-00             plot equal distribution (default: False)
  --gini-33             plot uniform distribution (default: False)
  --gini-66             plot log-logistic distribution (default: False)
  -g, --group           group by reward address (default: False)
  -e, --extended        use extended validators (default: False)
  -x EXPONENT, --exponent EXPONENT
                        distribution exponent mapper (default: 1.0)
  -s, --show            show plot (default: False)

[avalanche] $ ./stakes.py  ## plots stake distribution

List Stake Distribution

$ ./json/validators-list.sh -h

Usage: validators-list.sh
    [-m|--min-n=${MIN_N-0}] [-M|--max-n=${MAX_N}]
    [-g|--group-by-address] [-h|--help]

$ ./json/validators-list.sh -g < ./json/$(date +'%Y-%m-%d')/validators-ext.json

Copyright

© 2021, Hasan Karahan, MSc ETH Zurich.