Marios Richards

Is the kind of question I could answer if euref75Vote had appeared as a question in the BES.

Sadly, it didn't so I don't know but it certainly *sounds plausible*.

There's a blunt empirical point - Germany has a highly regionalised/localised state - even a social insurance system! - the UK is ultra-centralised.

Which of these states is ultimately governed better?

I think it's pretty obviously Germany - and I'm not complimenting Germany there!

(iii) hedging/robustness - sometimes shit people get into power, having multiple points of power/negotiation makes your system trade off a bit of efficiency for a bit of robustness

(ii) cognitive resources - if you think local/regional government is inefficient ... try watching a central government trying to substitute for tens of regional govts/thousands of local govts

It's necessary because

(i) you have to negotiate with local power brokers/bourgeoisie/homeowners - the only choice (in a democracy) is to do it well or badly - and a local council is much better placed to do that than an MP

It's a good argument against the *wrong* defence of local/regional government.

Local/regional government *does* have a great deal of extra costs/local power broker beak-wetting and it's not inherently superior because it's local or anything.

I think 75 was less about immigration cutting the population in Social Liberal/Authoritarian and a bit more about "Are you in a currently protected sector you think would be less protected in the EU?" (e.g. fishing).

40 years is quite a bit of time, but I think the main is that they were different electoral events dividing the population on different lines.

I haven't actually seen any good data on this - BES data datasets back to 1964, but not I think any questions on how people voted in the 76 eutref.

It's such an obvious "why does the larger level of govt not simply eat the smaller" that trap that normally both institutions act to prevent that - but in the UK it lined up with an adversarial pol divide.

Nowhere does *completely* in modern democracies - but there's a broad continuum of how much.

Part of the problem is that regional inequality drives inter-regional redistribution ... which in turn drives reduces the power if regional government to act to invest to reduce that inequality.

... whereas in science you can just say "I don't know, my intuition isn't worth shit, let's look".

> I certainly think there are more than two value-axis

This is where, in my opinion, science beats out philosophy.

Philosophers will fill pages with speculation about the intuition, their thoughts about ideology and legitimacy ...

But, in practice if you sit down and treat it like you would <any raw data>, do you year 1 exploratory analysis within a few seconds it's clear it's "two-dimensions-and-small-change".

> certainly think there are more than two value-axis along
> which people politically disagree

Well, this is an empirical question.

Human political attitudes *could* be 4 dimensional - there's no absence of political scientists/psychologists looking for more structure/dimensions.

Presidential systems work best with that but the Wilders style "the member c'est moi" is the closest accommodation in parliamentary systems.

I think not so much - Radical Right parties have brands (and much personnel) that only seem to trend down so unlike 'normal' parties it's actually good for them to be constantly dying.

Tldr; The Overton Window matters *a lot* in relation to *policy*, but much less with *coalitions/voting behaviour*.

E.g. social attitudes on Gay Marriage have absolutely shifted *dramatically*.

But if you normalise them - look at who is relatively more/less suportive - and plot them on the political compass ... they're pretty static.

Re: "Overton Window" - that's specifically - notionally! - about policies.

Policy space certainly zips around.

But attitude space does not and most of the time all you need to do is normalise.

There are all sorts of issues with the details - one issue with this specific set is that they are unbalanced (all Do You Agree towards Ec Left/Soc Auth) which is probably why the scale distribution is weird.

But that **mostly** disappears after normalising onto a unit plane.

... but actually the items are hard to improve on because they were picked because they line up with the results of PCA on human political attitudinal data.

And that stuff - by design and empirical fact - doesn't change much.

So when I say "political compass" I'm referring not to whatever it is that the trademark website does but to semi-standard items used for a long time in political science.

Everyone's first reflex - mine included - is a conviction that they can surely do better ...

The data is also repeated over a set of (fieldwork) waves, so technically we have 5 dimensions:

ec_left_right * soc_lib_auth * like * party * wave

But you want to look one wave at a time and break likes down into singular voting intention/behaviour.

Schematically, it's just 3D data *per party* (respondent 1234, has pos x_1234 on economic left-right, y_1234 on social lib-auth, likeCon (0-10, dks), likeLab, likeLD, likeGrn ....).

The fitted planes are just fitted planes in that 3D space.

Data access requires promising to agree not to try to deanonymise/share the data without that requirement - i.e. I believe sharing subsets with people is fine *done privately* and *after they've also clicked the 'promise not to share/deanonymise'*.

Yes and No - full dataset is up here (not mine and completely public):

www.britishelectionstudy.com/data-objects...

I'm specifically looking at political compass values (mostly just al_scale/lr_scale variables from al1-5/lr1-5 items) against like<Party> (likerts with DKs).

(Binary representation - in practice, it's less "no representation possible" but "always at risk due to trust deficit")

But Auth-Left representation isn't really compatible with liberal democracy - either it's someone like Nick Griffin, scamming for bits of money here and there or it's someone they don't feel can represent them because they're on an MP salary.

Auth-Right representation isn't a problem - they look at MPs and see people economically like themselves.

Lib-Left/Lib-Right representation isn't a problem - they don't care about the MP's person, they care about their ideological values.

There's that - great clip btw - but I think there's also a fundamental, unavoidable problem with the Auth-Left representation.

Authoritarians are selfish and struggle to believe they can be represented by people not like them - and their personal circs is their economic ideology.