Thread by @ProfPaulPoast

Does @NATO matter?

Okay, that's too big of a question. How about: did NATO help Eastern European states achieve and maintain democracy?

That's a question I'm considering in a working paper. Summary thread👇

This question is obviously asking one to consider the "counterfactual".

Counterfactual analysis has become popular lately when thinking about NATO. A good example is this recent paper by @KimberlyMarten

cambridge.org/core/journals/…

With respect to NATO's role in supporting democracy, there are a few views on the subject:

First, @dmgibler & Jamil Sewell in @JPR_journal say "Yes, NATO helped democracy" (by creating a peaceful environment).

journals.sagepub.com/doi/abs/10.117…

Second, Dan Reiter in @Journal_IS said, "Nope. If anything, NATO makes the situation harder for democracy" (by, for instance, antagonizing Russia)

mitpressjournals.org/doi/abs/10.116…

Third, there's me & @jurpelai in our @UChicagoPress book. We argue that most of the hard work for democracy is done well before states are even close to being offered NATO membership:

press.uchicago.edu/ucp/books/book…

So how are we to adjudicate b/w these claims?

As it turns out, now is an ideal time to reevaluate these claims. We now have as many years of data since NATO entry for the Baltic states as had passed between independence and NATO entry (13 years).

https://twitter.com/ProfPaulPoast/status/1114506236408496129

In addition, the Varieties of Democracy project is now producing measures of regime type that not only well capture the core components of democratic systems, but offer a measure that is more refined than the commonly applied Polity data.

v-dem.net/en/

That V-Dem measure is the "Liberal Democracy Index", which runs from 0 (highly illiberal) to 1 (highly liberal).

To show how it compares to Polity, consider this figure. It plots the "Liberal Democracy Index" and the Polity 2 score for the United States from 1789 to 2017.

Polity is largely impervious to the fact that the US had legalized slavery and women did not have the right to vote (it's score is 8 or above in both the 1850s & in 2017).

But the "Liberal Democracy Index" is dramatically lower in the 19th century compared to today.

Stated differently, the "Liberal Democracy Index" appears to have more face validity: it seems to "measure what is is intended to measure"

cambridge.org/core/journals/…

So how do I make use of the "Liberal Democracy Index" to evaluate the effect of NATO on Baltic democracy?

I follow several steps:

First, I assume that the Index values for the three Baltic states in each year provide a range of "possible" Index values for those states. For example, here are the scores for 1992:

- Estonia: 0.38
- Latvia: 0.54
- Lithuania: 0.73

This is a pretty big range of values.

Second, I randomly draw a number that lies within these values (for 1992, that would be b/w 0.38 and 0.73).

Hence, I view this randomly drawn number as a plausible "counterfactual" Index score for a Baltic state in 1992.

Third, I repeat the random draw a bunch of times (say, 1,000). This gives me 1,000 plausible values for the Index score for 1992. I then compute the median, 0.95 lower bound, and 0.95 upper bound.

I then repeat these steps for each year (from 1991 to 2017). This generates DATA

I can then plot the 0.95 confidence intervals and median value for the entire 1991 to 2017 time period. These are shown in the below figure.

The red line indicates 2004, the year the Baltics entered NATO.

What do I see? I see that not much differed before or after NATO entry

Of course, this could be because ANTICIPATION of NATO entry forced the Baltics to become stable liberal democracies (and then continued afterwards).

Fortunately, we have a way of testing if this is the case: the Baltics weren't the only states to enter NATO in 2004

In other words, I can perform the same steps using Bulgaria, Romania, Slovenia, and Slovakia -- the other four 2004 NATO entrants!

Here is the resulting graph:

It shows that the range is MUCH wider for these four countries. To me, this suggests that being a stable democracy was not essential for entering NATO in 2004.

The graph also shows wide variation AFTER NATO entry. This suggest that remaining a liberal democracy is also not necessary to stay in NATO.

Of course, we know that from the experience of Poland and Hungary (and Turkey before that).

brookings.edu/blog/order-fro…

My conclusion: NATO was neither necessary to make the Baltics democratic, nor necessary for them to remain democratic.

NOW, I'm not saying that, in reality, NATO didn't help. But NATO might be better thought of as sufficient, not necessary, for democratic support.

(end)