Thread by @ProfPaulPoast

According to The Polity Project, the US is no longer a democracy.

Why does that matter? Why is the Polity Project wrong?

[THREAD]

Why does the Polity score matter?

There are a host of other democracy measures out there, such as @freedomhouse or @vdeminstitute.

But the Polity Project offers what is probably the most widely used social science measure of a country's regime type.

The Polity Project assigns to every country in the world from 1800 to the present an annual "Polity score" from -10 to 10.

10 = Purely Democratic
-10 = Purely Autocratic

Critically, the Polity project has traditionally considered a score of 6 or higher to indicate that a country is a democracy (source: systemicpeace.org/polityproject.…)

How is the Polity score computed? By coding various "authority characteristics" of states. For a democracy, this means three broad aspects: (1) citizens can express preferences, (2) constraints on the executive, (3) guarantee of civil liberties. Polity focuses on #1 and #2.

A key component in the Polity coding is "Executive Constraints". You can see this by looking at the weights in the codebook

Given this background on the Polity score, why is the Polity Project wrong to lower the 2020 US Polity Score to below 6?

I'm definitely not the only one to go "wait, what?" over this decision. For example, @benwansell has a very good thread on his problems with the decision (i.e. Polity project might have been seeking out clicks)

https://twitter.com/benwansell/status/1347173302993883136

For me, the decision is head scratching based on their own stated reasoning (found on the homepage of the project's website: systemicpeace.org/index.html). Let's break it down.

First, they reduced the Polity score to 8 in 2016 due to "polarized" or "factionalized" competition. According to the code book, this means 👇

I could be wrong, but doesn't this description pretty well capture a long term feature of US politics. Why only notice it in 2016? 🤔

Second, the coders started reducing the "Executive Constraints" variable in 2019 and, again, in 2020.

They claim that Trump was able to avoid Congressional oversight. Sure, but so have MANY Presidents. I mean, the phrase "Imperial Presidency" is not new

amazon.com/dp/B0085TKO04/…

If anything, 2020 showed that Trump IS a constrained executive: that is EXACTLY why he's become frustrated and desperate...

politico.com/news/2020/12/1…

...which is a point I made a little less than a month ago

https://twitter.com/ProfPaulPoast/status/1337558159842488320

In other words, Trump's behavior is a REFLECTION of US executive constraints being present and functioning, not due to their weakening or absence.

As another example of Trump being constrained, think back to his "Travel Ban". Remember how he had to rewrite it three times? Yeah, that's executive constraints.

nbcnews.com/news/us-news/w…

Now, is the US a 10 on the Polity scale in 2020? No. But a 5? Probably too low.

But there is a more more fundamental issue regarding the Polity project and its coding of the US.

The Polity Project has coded the United States as having a Polity score of 6 or higher

That graph should IMMEDIATELY raise questions:

- US had a Polity score of at least 8 prior to the Civil War?
- US had a Polity score of 10 before Women's suffrage?
- US had a Polity score of 10 before the VRA in 1965?
- Trump's election caused a sudden drop in the Polity score?

Indeed, @annemeng_ argues that the US probably shouldn't have been coded a democracy prior to 1965 at the earliest (I wouldn't go that far, but the point is VERY well taken)

https://twitter.com/annemeng_/status/1347274747105705987

You might wonder, "was the US scored high in the mid-19th century because Polity is coded RELATIVE to other countries at that point in time?"

Seems reasonable, but the answer is "No".

Making Polity suitable for longitudinal analysis (i.e. over time) was done in the 1990s (from the Polity codebook)

To illustrate, in a 1995 @JPR_journal piece, Jaggers and Gurr, founders of the project, showed trends in the Polity score (and its components) over time

Here is the paper
journals.sagepub.com/doi/abs/10.117…

So there are reasons to already be skeptical of the coding of the United States even BEFORE the Polity Project's recent decision.

Now, you might say that this is unfair to the Polity project: one could probably quibble with the coding of any polity score for any country.

That is true, as @JeffDColgan pointed out about Iran's coding in a @Journal_of_GSS piece (and how that coding was likely infused with "Western Bias"): academic.oup.com/jogss/article-…)

But this is about more than "quibbles": the US is the home country of the Polity project. It's the country where the founders of the Polity project have the most direct knowledge. If they can't code IT in a way that passes face validity, well...🤷‍♂️🤦‍♂️

In short, I *could* be okay with Polity=5 for the US in 2020 if the Polity Project ALSO adjusts their scores for, say, the rest of American history. But I don't see that happening. Until it happens, one is justified being very skeptical of results based solely on Polity.

[END]