open-government

A framework for an open source algorithmic government with voluntary participation

Background

What some define as a "government" in it's current state is an organization with a monopoly on force. That is, regardless whether a government's actions are beneficial for society or to what degree of force is required for an orderly society, all power current governments wield derives from this monopoly. Since most people can agree that monopolies are generally bad for consumers (tax-payers or citizens in this case), then it’s at least rational to explore the possibility of competing governments with voluntary participation by citizens.

Such a system has historically been difficult. Throughout history, governments are replaced by new ones when it becomes incapable managing its affairs, resulting in either domestic uprisings or foreign takeovers. These transitions are extremely expensive and rarely peaceful. The new regimes rarely give up the newfound, highly lucrative monopoly after fronting the cost of the revolution/war with their human or monetary capital.

The American Revolution attempted to create a non-monopolistic government by limiting the government’s power with the Constitution (separating power between three branches of the government and granting high level of autonomy to states) and the Bill of Rights (granting individuals with inalienable* rights). The ever westward expansion further limited the government’s control of those seeking new opportunities or evading the government’s grasp. Coincidentally, the federal government’s power has grown increasingly after the last major westward migrations in the mid 1800’s.

Given the size and opacity of governments today, it’s fair to say that no citizen knows everything his government does, much less agree with it. If there are any major dissatisfaction, an average citizen has the following choices (in order of increasing cost):

1. Tolerate it - Most of us do this.
2. Influence legislators to change or become one - legislators have limited time, not enough to give to every average citizen; difficult to achieve and once achieved, one’s views may have changed
3. Exit - Relocate to a geographic location, possibly a different nation, which is expensive both in monetary and social capital
4. Forcefully implement change - AKA Revolution/rebellion, as the cost is treason if failed, this only becomes a reasonable option when the other options have failed for a large portion of the population

We can perhaps model the history of governments with an overly simple Markov chain in which each state is a level of citizen satisfaction rate. State 0 is completely satisfied which we assume to be the case at the start of any new regime. States 1-4 correspond to above list. A state can only move up or down 1 step at a time, with exception to state 4 (on the edge of a rebellion), which can go to state 0 when the rebellion succeeds. Each arrow in the Markov chain indicate the probability of moving between states in each iteration. These transition probabilities were arbitrarily assigned while being reasonable* and simple. We assign high values to all self-transitions except state 0 because governments generally stay the same. A fully satisfied population is rarely the case so we assigned state 0’s self-transition to be a low value (0.2). The transition from 1->0 and 4->3 are assigned 0 because there is no incentive and it’s beyond saving respectively. It’s clear that for the majority of the time (both in history and our model), we’d reside in either state 1 or 2. If this model is a reasonable approximation, then we can say that the objective of an effective government is to maximize time spent in the lower states. We assume a government at state 2 is equally likely to change for better or for worse, while in state 3, the system has become more biased towards unsavory politicians.

This Markov chain can be analyzed to make further conclusions about the society we’d live in. For example, an equilibrium distribution which represents the amount of time people spend in certain levels of satisfaction with their government can be calculated with the following python code.

import numpy as np
a = np.matrix('0.2 0.8 0 0 0; 0 0.9 0.1 0 0; 0 0.1 0.8 0.1 0; 0 0 0.1 0.7 0.2; 0.4 0 0 0 0.6')
b = np.matrix('0.2 0.8 0 0 0; 0 0.9 0.1 0 0; 0 0.1 0.8 0.1 0; 0 0 0.2 0.7 0.1; 0.4 0 0 0 0.6')
prob_a = np.linalg.matrix_power(a,100)[0]
prob_b = np.linalg.matrix_power(b,100)[0]

>>> print a
[[ 0.2  0.8  0.   0.   0. ]
 [ 0.   0.9  0.1  0.   0. ]
 [ 0.   0.1  0.8  0.1  0. ]
 [ 0.   0.   0.1  0.7  0.2]
 [ 0.4  0.   0.   0.   0.6]]
>>> print b
[[ 0.2  0.8  0.   0.   0. ]
 [ 0.   0.9  0.1  0.   0. ]
 [ 0.   0.1  0.8  0.1  0. ]
 [ 0.   0.   0.2  0.7  0.1]
 [ 0.4  0.   0.   0.   0.6]]
>>> print prob_a
[[ 0.02564103  0.51282051  0.30769231  0.1025641   0.05128205]]
>>> prob_a[0,3] + prob_a[0,4]
0.15384615384654135
>>> print prob_b
[[ 0.01492537  0.47761194  0.35820896  0.11940299  0.02985075]]
>>> prob_b[0,3] + prob_b[0,4]
0.14925373134320222

prob_a is a vector in which each component is the expected percentage of time someone is expected to reside in the corresponding state. As previously claimed, most time will be spent in states 1 and 2. This model also predicts a total of over 15% of the time we can expect to live in states 3 and 4, representing very unfavorable conditions. When giving the state 3 government the benefit of the doubt by swaping probabilies of changing for better or for worse, the bad times still approach 15%.

As most of us have lived in periods mostly considered state 1 and 2 that value may be surprising. Generating 20 random walks using randwalk.py starting from state 1 and ending when reaching 0 or 20 iterations using b Markov matrix result in the following scenarios. In over half of these cases, someone with experience of 20 time steps won’t experience state 3 or 4. (20 steps was picked as average lifespan of a person is ~80, governmental change occurs on a scale of years, for example 4 - a single term of US presidency.)

>>> from randwalk import randwalk
>>> for i in range(1,21): print ('%02i : ' + randwalk(20,1,'good')) % i
... 
01 : 111111111223333333223
02 : 111111111111111122222
03 : 111122111111222222111
04 : 111111111112333321111
05 : 1111112223333340
06 : 111111111111111111112
07 : 111111111111122111111
08 : 111111111111111222332
09 : 112233223222222333333
10 : 111111122222221111111
11 : 111111111111111111112
12 : 111111111111111111111
13 : 111111111111111111222
14 : 111111111122211122211
15 : 111111111123322111111
16 : 111111111111222222332
17 : 111111111222211111111
18 : 111111122233333222233
19 : 111222222222222222222
20 : 111111111111122222222

Globalization and the internet have lower the cost of relocation and enabled easy global communication. The emergence cryptocurrencies now allows greater numbers of unsatisfied citizens to exit his nation state economically, although not necessarily physically. Together, these advancements may produce a significantly different Markov chain in the not-immediate-but-perhaps-foreseeable future, one without state 4 and the resulting period of despair.

To take it one step further, suppose governments are completely transparent, taxes (or "membership/subscription fees") are voluntary, and anyone can propose changes to his government for review (state 2). If the changes are quickly accepted, then citizens would only be briefly in state 2 before going back to state 1. Now suppose governments can be created at a trivial cost. A rejection of proposed change (or even a delayed decision) then either results in immediate exits, either to an existing competing government or new clone of the original with added updates. We effectively remove state 3 and dramatically alter governments willingness to change (in order to preserve citizenship and their voluntary tax contributions).

Calculating the equilibrium distribution on this new reduced cycle, we easily see the hypothetical benefits.

>>> c = np.matrix('0.2 0.8 0; 0 0.9 0.1; 0.1 0.8 0.1')
>>> print c
[[ 0.2  0.8  0. ]
 [ 0.   0.9  0.1]
 [ 0.1  0.8  0.1]]
>>> prob_c = np.linalg.matrix_power(c,100)[0]
>>> print prob_c
[[ 0.01234568  0.88888889  0.09876543]]

If this sounds like an open source software project to you, it's because it basically is. Open Government is a project to create a framework for such a government. By leveraging cryptocurrencies, trustless contract arbitration, and competition, governments can become lean, mean service providers for a freer society.

To do

• Describe the mechanics of such a government in detail
• Provide example of hypthetical cases
• Start a test government that provide services by accepting voluntary Bitcoin tax payments

Questions of how such a government can play out or things that require further explaination? Submit an issue for concerns that can be addressed as an example.

Want to contribute? Open up an issue for any potential bugs, improvments to the Markov model, write about new ideas, or submit possible services to Bitcoin users.