Sunday, May 23, 2021

A feature, not a bug, part 2

 My last post was about a measure of party "illiberalism" from the Varieties of Democracy project.  It showed a substantial increase for the Republican Party over the period covered by the data (1970 onward), especially after 2000.  I said "I'll look more closely at the data in the future to see if there are any other cases of a major party in an established democracy dramatically increasing its illiberalism score," so here it is.  There are a lot of parties in the dataset, many of which existed for only a few elections, so I began by limiting it to ones with values for seven or more elections.  Then I fit separate linear trends for each one.  The largest positive (towards greater illiberalism) estimated trends were for the following:

Party                                                                Nation

Fidesz                                                              Hungary
Nationalist Republican Liberal                       Panama
Internal Macedonian Revolutionary Org        Macedonia
Democratic Party                                            Mongolia
Indian People's Party                                       India

Laka-CMD                                                    Philippines
Slovenian Democratic Party                         Slovenia
Senegalese Democratic Party                        Senegal
Brazilian Social Democracy Party                Brazil
Republican Party                                          United States

Swiss People's Party                                      Switzerland
New Slovenia--Christian People's Party        Slovenia
Northern League                                            Italy

The "leader" is Fidesz, which is a well-known case of a party that went from liberal to authoritarian.  The #5 party (also known as the BJP) is currently in power in India.  I don't know anything about the others in the top 9, but none of them are in what I would regard as established democracies.  The Republicans are #10 and Swiss and Italian parties are at 11 and 13.   

According the scores, the Swiss People's Party started as a fairly normal democratic party, but took an illiberal turn in 1991, and has stayed the same since then.  

The Northern League (often just called the League) always has had high illiberalism scores, but has increased substantially starting in 2008.  

What exactly do these scores mean?  People who are supposed to be experts in the politics of the country are asked "have leaders of this party used severe personal attacks or tactics of demonization against their opponents?," "to what extent was the leadership . . . clearly committed to free and fair elections with multiple parties, freedom of speech, media, assembly and association?" "according to the leadership . . . how often should the will of the majority be implemented even if doing so would violate the rights of minorities" and "to what extent does the leadership . . . explicitly discourage the use of violence against domestic political opponents."  Each rating was on a scale with five categories.  They they are combined using a "Bayesian item response theory measurement model" and put on a 0-1 scale.   The experts were asked to give ratings for each election, but this was all done at once in 2019-20--that is, people were being asked to look back at earlier elections. The Democratic and Republican scores on the individual items (with higher values meaning less support for liberal democracy):



 The first one is low and constant in both parties until 2008.  This may seem surprising, but I think it shows that the respondents paid attention to a further "clarification" included with the question--"includes dehumanizing opponents or describing them as an existential threat or as subversive, criminal, or foreign agents."  That made it clear that it didn't include attacks on an opponent's policies, judgment, or ability, even severe or unfair ones.


On the second question, Republicans start with a somewhat higher score, then jump in 1986 and again in the 21st century.  The Trump years don't stand out as much, and there's a move towards democratic principles in 2010.  This pattern doesn't make much sense to me, even with the clarification:  "parties show a full commitment to key democratic principles if they unambiguously support freedom of speech, media, assembly, and association and pledge to accept defeat in free and fair elections."  Taken literally, I think that Republicans supported all of these until 2016, when Donald Trump wouldn't commit to accepting defeat--even then, they didn't try to place significant restrictions on freedom of speech, media, assembly, and association.  Perhaps respondents were using this question to get at something that isn't directly covered by any of the questions--the growth of "constitutional hardball," the practice of going as far as you can while staying within the letter of the law regardless of traditional practices or principles of fairness.   

The third one, about violating the rights of minorities, doesn't seem very good to me--it's hard to distinguish from the ideological content of policies.  Since the Democrats put more emphasis on the value of equality, the Republicans could be expected to score higher, and they do.  However, there's not much change for either party, so this doesn't make much contribution to the difference in trends.  

Finally, rejection of political violence is pretty straightforward.  

Last time I mentioned that I thought that the current scores for the Republicans seemed too high when comparing them with parties in some other nations. Given the procedure, I can see why this happened--the respondents were asked to focus on their own nation, not to compare it to others.  As a result, there would be a tendency to make use of a wider range than if they had do do international comparisons.  

You could argue that the ratings just reflect the biases of the "experts"  most of them are academics, and most academic political scientists are liberal:  "since most academics are liberal, so are V-Dem’s ratings—even on dimensions of democracy that are ostensibly objective."  But this isn't a convincing argument.  Some of the figures I've given in this post count against it, but a bigger thing is that these ratings seem to have predictive power.  They were finished in the beginning of 2020, and suggested that the Republicans were not very committed to democracy.  After the election of 2020, the Republicans acted like a party that wasn't committed to democracy--many of them tried to overturn the election results, and few explicitly accepted them.*  In contrast, the Democrats have continued to act like a normal democratic party--they haven't engaged in payback or changed the rules to make it harder for the Republicans to compete in the future.  

*Of course, a lot of that was just grandstanding, but that's what they acted like, whether or not they were sincere. 

No comments:

Post a Comment