There seems to be some confusion as to what effect ALBA will have on the distribution of seats in the Scottish Parliament. As teased yesterday, I’ve built a toy model to illustrate that the answer is unambiguously “it depends”. 🧵🏴󠁧󠁢󠁳󠁣󠁴󠁿 /1
Depends on what? On a number of things, but specifically (1) on the performance of ALBA itself, which may be either a drag or a bonus for the pro-indy camp, (2) on *where* ALBA is getting list votes from, and (3) on the overall performance of the indy camp. /2
As a premise: I’m not a Scotland specialist. I’m interested in electoral systems and I think this example shows key points about the mechanics of strategic coordination in mixed-member systems. I.e. that coordination is (1) possible, (2) consequential and (3) really hard. /3
First off, I took the latest BMG poll, simulated an election on uniform natl swing. Then I re-run the election with ALBA, assuming it gets all its votes from the SNP and is distributed like the SNP. This is how ALBA’s addition changes the number of seats for pro-indy parties. /4
The effect is non-linear: it’s negative if Alba is betwee 1.5 and 5%, as it’s not gaining seats but is undercutting the SNP in the lists. After that, the effect becomes positive, as the SNP plateaus (i.e. wins only const. seats), and ALBA starts winning in the regional lists. /5
What happens if ALBA also gets votes from the Greens? I don't know - or care - about the plausibility of this scenario (let's call it the “Green Leakage"), but let's just think of these voters as strategically minded independence single-issue voters. /6
Here 20% of ALBA voters would have voted Green in a scenario where Salmond’s party didn't exist. We see that the ‘slump’ for low ALBA shares here is both deeper and longer. ALBA needs to win over 10% to be a net positive! This is where ALBA could do real damage. /7
Let’s now consider the third variable: i.e. how the effect of ALBA varies conditional on the overall performance of the independence camp. First, I simulated 250 possible election results, using the distribution of polling estimates available since January 2021. /8
Then, I compared these results with the counterfactuals of ALBA winning 4% or 10% of the vote. In the scenario where ALBA only drains SNP list votes, the effect is quasi-linear, but ALBA is (relatively) a little more helpful when pro-indy parties perform relatively poorly. /9
Under the Green Leakage scenario, this point is clearer. The difference btw splitting list votes in two (SNP and Grn) or three (SNP, Grn, ALBA) is smaller as the indy-camp gets bigger: e.g if the SNP gets over ~55% they would win list seats without strategic coordination. /10
Finally, we can bring together all these variables in one graph, which draws on 250 simulated elections and 30 different levels of ALBA performance, for a grand total of 7500 estimates. Blue means that ALBA is helpful to pro-indy parties, orange that it’s a net negative. /11
The tile graph shows how absolute changes in pro-indy seats due to ALBA vary as a function of pro-indy parties’ performance *and* ALBA performance. ALBA is always a negative if it's below 5%; it’s somewhat more helpful for a relatively worse showing of the indy camp. /12
In relative terms, if pro-independence parties do poorly, but ALBA does very well, it may mean over 15% additional seats over the counterfactual. Note also that for very high performance of the indy camp, ALBA is superfluous: the SNP would win list seats anyway. /13
Finally: what about Salmond's claims about the "Independence supermajority"? Well, they're not *that* far-fetched. An efficient distribution of the pro-independence list vote could easily yield over 80 seats for the camp. /14
That's all folks, I believe. Find the data and code here: some parts (e.g. the simulation from polling) are very rudimentary and please don't bully me because my coding is crap. But it was fun to have a first stab at this. Electoral systems are fun! /15 dropbox.com/sh/qwrep3f699f…

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Leonardo Carella

Leonardo Carella Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @leonardocarella

26 Mar
I regret to announce that instead of doing my actual job or enjoying my Friday evening, I've built an R model to simulate how ALBA performance in the lists would affect the distribution of seats conditional on SNP performance in the constituencies and Green/SNP transfers to ALBA.
A snippet (clean code and more stuff tomorrow): on a good night for the SNP (in which they'd win 62 constituencies: uniform swing on latest BMG poll), ALBA would waste pro-indy votes as long as it stays between 1.5 and 5%. After that 'slump', ALBA starts benefiting the indy camp.
I think there's more to this: ALBA could do some real damage if they split pro-indy list votes with the Greens or if the SNP has a bad night in the constituencies and needs list votes. Will think about it and let you know :)
Read 6 tweets
19 Mar
Europe has a science problem.

56% of French and 17% of Italians have used homeopathy, which is considered a medication under the law in those countries. There is no evidence, anywhere, in the scientific literature that homeopathy is any better than a placebo.
Glyphosate was declared non-carcinogenic by the EU's EFSA and ECHA, as well as the US's EPA. It's the most used weedkiller in the world and it's crucial to ensure reliable food production. It's banned in France, the Netherlands and Austria.
GMO crops are the linchpin of agricultural modernisation in the developing world, and they're credited with saving millions from starvation. The overwhelming scientific consensus is that existing GMO food poses no greater risk to human health than conventional food.
Read 7 tweets
11 Mar
Ogni giorno in Italia muoiono ~2000 persone. La probabilità di morire per un individuo in un giorno è quindi 2k/60m in media. Se 150k persone ricevono il vaccino in un giorno, ci si aspetta che 150k*(2k/60m) = 5 persone muoiano nello stesso giorno in cui vengono vaccinate.
Le morti di infarto sono circa 120k all'anno, quindi circa 330 al giorno. Quindi ci si dovrebbe aspettare che 150k*(2k/60m) = 0.825 ≈ 1 persona muoia d'infarto nello stesso giorno in cui viene vaccinata.
Questo è senza contare che c'è una relazione positiva fra rischio d'infarto e priorità di accesso ai vaccini (dovuta, per esempio, all'età). Quindi non solo non c'è prova che gli incidenti isolati riportati dalla stampa non sono in rapporto di causa-effetto con il vaccino...
Read 4 tweets
27 Feb
On a first read of Worlds of Welfare, students often have troubles figuring out what it means that the conservative (continental European) welfare systems are hierarchical and status-differentiating. The Italian vaccination programme is set to become the new canonical example.
"Sicily: trainee lawyers also included among priority groups" catania.liveuniversity.it/2021/02/26/vac…
In Tuscany, university lecturers and admin are being vaccinated, but doctoral students are up in arms that they were excluded. pisatoday.it/cronaca/precar…
Read 4 tweets
25 Feb
I am terrified at the potential mess they could make of House of Dragons. The Dance of the Dragons is great raw material, but it doesn't come with extensive source dialogue, and lends itself to the temptation of going full-on CGI at the expense of the (highly complex) plot.
The fact that the two main characters - Rhaenyra and Alicent - are Strong Female Leads™ also comes with the risk of repeating the mistakes of the Dornish plot, where insistence on the Girl Power stuff is just one-note and results in flat, unlikeable characters.
Another way of screwing it up could be just overimposing the Starks/Lannisters boring goodies-vs-sexy baddies dynamics on the Blacks-vs-Green civil war. No. The Blacks are not the goodies. Give us two sides of deeply flawed characters to cheer for.
Read 4 tweets
8 Jan
A word of caution: many aspects of the social, political and informational mess that led to extremists storming the US Capitol are present in Western Europe too. (1)
I. Political Violence

Wednesday didn’t happen in a vacuum: it happened in a context of rising right-wing terrorism, aimed at their adversaries in the elite (pipe bombs mailed to Obama, Biden, Clinton, DNC) and among the public (Charlottesville, El Paso, Pittsburgh, Charleston).
Europe has seen its share of the same violence.

Against parliamentarians: Jo Cox, Labour Party MP, and Walter Lübke (CDU politician in Hesse) have been assassinated. Plots against other politicians, in Germany and the UK, led to further convictions. (3)
Read 24 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!