Flux Political Party Forcably Deregistered (govt doing bad stats / epistemology and other bad things)

[Max]

Here is an example of what happens when you try to fight the system, especially using “from the inside out” type methods.

Posting it here b/c there are so many ways that this decision contradicts CF that it’s hard to list them all.

Summary of background/events (simplified to keep it relatively concise, note that the quote at the bottom is probably enough to get the gist if you want to skip):

• Flux is was a political party in Australia that I founded with a friend in 2015/16 (full name: VOTEFLUX.ORG | Upgrade Democracy!) – this was before I was told about FI, and actually learning about FI was due to a blog post I wrote about Popper’s Criterion. BoI inspired my going through with establishing the party; I mixed in ideas that I had before reading BoI, too.
• Political parties register by proving to the Aus Electoral Commission (AEC) that they have >1500 members (it was 500 members back then, it was increased mostly with no grace period in Sept 2021 – that’s not the AEC’s fault, that was when a new bill was passed).
  • The bill was supported by our 4 major parties: the Liberals (generally: conservative, right wing, not the good kind), Nationals (similar to the liberals but they get the rural vote; in a coalition with the libs), Labor (our major left-wing party, often holds government on its own), the Greens (a small party comparatively, ~1/150 seats in the lower house, but they have like ~10/70 seats in the senate). There are only ~6-12 members of the senate that aren’t members of major parties, and a few independents in the lower house.
• Membership testing:
  • Parties are allowed to submit 1650 members maximum as evidence of validity. (Excess members are discarded)
  • A statistical test is done to ascertain membership eligibility – members are contacted and asked to confirm membership.
  • If a threshold is passed for membership denials then this means the party fails the test; if there are that many denials or fewer then the party passes. At best, a 9.09% denial rate is tolerated.
  • If a party fails then it’s deregistered (taken off the register of political parties; the party can’t run in elections)
• In Jan 2022 Flux was notified that we’d failed our membership test (undertaken December 2021).
  • this test is only for non-parliamentary parties (those that don’t have a member of parliament in either the lower or upper house – only 1 member is required)
• When this happens the party gets a chance to respond, which often results in a second membership test.
  • After either the party response is rejected or the conclusion of the 2nd membership test, the AEC publishes a statement of reasons that explains their decision, along with the decision.
• Flux responded in Feb, and another membership testing procedure was initiated.
• Flux failed its second membership test too.
• Flux was deregistered on March 24th (I was not notified – it’s likely that someone was, but probably only 1 person)

I omitted the points we raised in the response b/c the statement of reasons explains enough.
Source (authored by Ms Reid): https://web.archive.org/web/20220326093243/https://www.aec.gov.au/Parties_and_Representatives/Party_Registration/Deregistered_parties/files/statement-of-reasons-voteflux-org-upgrade-democracy-s137-deregistration.pdf (this contains membership test results and other miscellany)

Flux’s Feb response: https://xertrov.github.io/aec-membership-test-simulator/docs/Response-to-AEC-rego-20220213.pdf

I’ve talked before about Flux’s mistreatment by govt.

Select quotes from the AEC’s statement of reasons (note: under point a): Reid omitted a > before the second paragraph, so it is a quote in the orig doc even though it doesn’t look like it below.

25. I have considered the statement lodged by the Party on 13 February 2022, setting out
    reasons why the Party should not be deregistered.

• a) “In your January 13 correspondence, you said:
  I am notifying you under s 137(1)(b) of the Electoral Act that the Electoral Commission is
  considering deregistering the Party, as the Electoral Commission is satisfied on
  reasonable grounds that the Party does not have at least 1,500 members.”
• b) “We do not believe that such a decision would be based on reasonable grounds. This is
  because the AEC’s methodology is flawed. Below, we will detail why we believe that this
  is the case and provide reasoning and evidence to back this up.”
• c) “We have 3 arguments supporting our case. Each argument is individually sufficient to
  show that a decision (by the AEC) to deregister the Party would not be based on
  reasonable grounds; each argument is a decisive criticism of the current methodology.
  • The statistical method used fails ~10% of the time for borderline cases.
  • The statistical method uses an artificially limited sample size and thus does not
    estimate party membership, though does (roughly) measure membership attrition.
  • We have sufficient membership and provide evidence. Attached is a list of 4680
    members. Each entry was, at some point, verified against the electoral roll.”
• d) “Unless each of these criticisms can be addressed, we do not believe that a decision by
  the AEC to deregister the Party would be based in reality.”
• e) “Please find attached a list of 4,680 members of the Party. Note that these are the
  subset of members for whom we have been able to pre-validate their electoral roll details
  at some point.”
• f) “Based on the observation that 4,680 is a lot more than 1,500, we reject any decision to
  deregister the Party on the grounds that it is in conflict with reality.”

26. I reject the reasons outlined by the Party in its statement provided on 13 February 2022 for
    the following reasons.

27. The Party failed membership testing for exceeding the maximum number of permitted
    denials according to the ABS methodology used by the AEC. It did not fail membership
    testing due to having an insufficient number of members being identified on the electoral roll.

28. The Electoral Act defines an elector as someone that is on the Commonwealth Electoral
    Roll. Section 123 of the Electoral Act prescribes that an eligible political party, not being a
    Parliamentary party, has ‘at least 1,500 members’. The requirement is not to be solely ‘an
    elector’ but to be a member of the party.

29. The Party challenges the validity of the AEC’s membership testing process. This process
    has been developed by the AEC to support the delegate’s consideration of whether a party
    has sufficient members. It is based on sampling methodology designed in consultation with
    the ABS and provides a valid methodology to satisfy a delegate of a party’s membership.
    The Electoral Commission has previously concluded that the methodology ‘was appropriate
    for membership testing, including because it was rational, fair and practical in all the
    circumstances.’^[1]

30. I consider that the membership testing results outlined above provide a more robust method
    for ascertaining whether a party has satisfied the requirements of the Electoral Act than a
    statement provided by the party.

31. In summary, I remain satisfied that the Party does not have at least 1,500 members based
    on the outcomes from membership testing both membership lists of 7 December 2021 and
    13 February 2022.

32. Accordingly, in my capacity as a delegate of the Electoral Commission, I have deregistered
    VOTEFLUX.ORG | Upgrade Democracy! under s 137(6) of the Electoral Act and the
    particulars of the Party have been cancelled from the Register under s 138 of the Electoral
    Act.

The document that Reid cited is actually for the old testing method (prior to Sept 2021) that used 500 members, not 1500. So I wonder if it was her expert knowledge of statistics (which she implicitly rejects having) that enabled her to judge whether the new method was equally valid.

The objections Flux raised are one’s that other parties have tried to but didn’t have the statistical knowledge to know what was wrong. Like the respondents (from political parties) knew something was wrong, but didn’t know how to word it very well.

Also note WRT the first point under c) (“The statistical method used fails ~10% of the time for borderline cases.”), it failed 10.5% of the time (the maths are in Flux’s response), and the AEC has, in the past, claimed the method has a >90% confidence. (It’s unclear how they count or calculate this, they’ve never said.)

edit: I missed this quote and screenshot which provides more context:

22. The membership list submitted by the Party on 13 February 2022 contained 4,680 names of
    individuals that the Party considers to be current members (referred to as ‘members’ below).
    As a delegate of the Electoral Commission, I instructed that the top 1,650 names be tested
    to conform with the AEC’s membership testing parameters. The following results were found
    after initial membership testing against the electoral roll.

---

I don’t mind criticism for this post, but I’m posting it in Other deliberately. Thought it was worth sharing.
If it’s appropriate to move to Unbounded that’s okay, but I don’t think it matters much.

I have more to say but this is enough for now.

---

1. https://www.aec.gov.au/Parties_and_Representatives/Party_Registration/Registration_Decisions/2021/notice-of-decision-with-reasons-SUPA.pdf ↩︎

[Elliot]

Their statistics are ridiculous.

Flux’s response is bad PR. It’d be better to take a simple, short point with almost no math and focus attention on it.

Suppose a party has a list of million members but only around half of them are currently in the party. It’d be very hard for them to pass the test (submit 1650 names at least 1500 of which are active now) even though they obviously have over 1500 current members.

That’s only two sentences and I didn’t even edit it.

BTW you could try to pre-verify your own members but if you have a million people you probably have enough political opponents to sabotage you (tell you they are active members, but then if phoned say they’re not). And that’s just for free. People could also be paid to sabotage you.

So in one place it says you had an 83% success rate on their random sample of 53 people. 83% of 4500 is 3735 which seems like a reasonable estimate of your membership and is above 1500.

But this doesn’t matter. Your registration doesn’t matter.

What matters is you tried to fight Power and get your own Power with basically no plan.

Or a very naive plan: read the current rules and assume they will be followed and let you win something (and I think you were failing even at that).

But Power has the Power to change the rules when the rules don’t suit them. The system is rigged. (Not that the current change is about you. You were not even a threat. If you were actually a threat, then they’d rig things against you.)

You want things and Power does not want you to have those things.

Come up with a better plan or get out of politics.

They are actually doing you a favor by not stringing you along longer so that you waste more resources on bad plans.

If you aren’t going to actually befriend Power or persuade them or persuade 10% of the country, but you want to affect stuff, then you need to get some power of your own (not a recommendation – see e.g. Gail Wynand). Or, better, make the planning trees Goldratt explained (current reality cause and effect, transition, future reality, etc), figure out constraints and how to affect them, etc.

[Max]

@Elliot

Flux’s response is bad PR. It’d be better to take a simple, short point with almost no math and focus attention on it.

I didn’t understand the first bit. Does PR mean public relations? And, do you mean PR with the govt or more broadly?

One way I read it is like: it’s less easy for regular ppl to understand what went wrong b/c Flux’s response was overly complicated.

WRT what’d be better – I didn’t think Flux would have passed the 2nd test either, or that the AEC would contradict their past statements.
I agree that a short single point is good for explaining things to e.g., members in a notification email.

@Elliot

Come up with a better plan or get out of politics.

I’m not sure if my original post came off as like particularly pro-Flux or something – it’s mostly meant as documentation.
I’ve been in favor of wrapping things up for a while now. Reading AS helped. I didn’t want to just resign and leave friends in the lurch to clean up tho.

[Max]

There is an update:

I have lodged an application for review of the decision to deregister Flux.

I feel that this is, in some ways, my time at the cabin.

I philosophically oppose my own actions, but I want to do them, still. The advantage of doing them is that I can say “it is done.”
There is also an aspect of, like, ‘a captain goes down with his ship’ – I can’t claim to be the captain, but I feel some ownership, still. I hope my meaning is clear enough.
I want to see this through.

You can read my request for review here: https://xertrov.github.io/aec-membership-test-simulator/request-review-flux-deregistration. note, there is some missing context due to the nature of the request, I suggest you read up-to and including section 2 of the following for background if you would like a briefing: AEC Party Membership Test Methodology is Rigged! A Statistical Analysis of AEC Methodology and Graphs (of PMFs) | aec-membership-test-simulator – additionally, the linked request is addressed to the three people who make up the Australian Electoral Commission (excluding supporting staff): the Australian Statistician who heads up the ABS (aus bureau of stats), a federal judge, and the electoral commissioner themselves.

I don’t expect the application to succeed. If it does succeed, Flux is going to voluntarily deregsiter as a political party anyway. (There’s a draft email and post to members about this that will go out soon). It’s mostly about doing it on our own terms.

[Lebowski]

@Max

my time at the cabin

I assume this is a local saying of some sort but I’m not familiar with it’s intended meaning. I don’t think you’re literally spending time at a cabin related to the Flux deregistration.

[alanforr]

In Atlas Shrugged Dagny Taggart resigns from Taggart Transcontinental and goes to a cabin in the woods to think about her future. But then a disaster happens at TT and she comes back.

[Lebowski]

I don’t understand how the Rand reference would explain

@Max

I feel that this is, in some ways, my time at the cabin.

The Rand reference would suggest something more like that working on Flux has put an end to his time at the cabin.

If he’d said what I suggest (working on Flux has put an end to his time at the cabin) I still wouldn’t have initially gotten the reference. I would just have understood it with the explanation you gave. Maybe there’s some nuance about it I’m still not aware of.

[Max]

@Lebowski

The Rand reference would suggest something more like that working on Flux has put an end to his time at the cabin.

yeah, it’s not perfect, but it’s what I was referring to. I think it’s fine for you to read it like that.