Like legislators in the US Congress, the member states of the United Nations (UN) are politically divided on many issues such as trade, nuclear disarmament, and human rights. During the Cold War, countries in the UN General Assembly tended to split into two factions: one led by the capitalist United States and the other by the communist Soviet Union. In this exercise, we will analyze how states’ ideological positions, as captured by their votes on UN resolutions, have changed since the fall of communism.

This exercise is based on Michael A. Bailey, Anton Strezhnev, and Erik Voeten. “Estimating Dynamic State Preferences from United Nations Voting Data.” Journal of Conflict Resolution, August 2015.

The data is called unvoting.csv and the variables are:

Name Description
CountryName The name of the country
idealpoint Its estimated ideal point
Year The year for which the ideal point is estimated
PctAgreeUS The percentage of votes that agree with the US on the same issue
PctAgreeRUSSIA The percentage of votes that agree with Russia/the Soviet Union on the same issue

In the analysis that follows, we measure state preferences in two ways. Note that the data for 1964 are missing due to the absence of roll call data.

First, we can use the percentage of votes by each country that coincide with votes on the same issue cast by the two major Cold War powers: the United States and the Soviet Union. For example, if a country voted for ten resolutions in 1992, and if its vote matched the United States’s vote on exactly six of these resolutions, the variable PctAgreeUS in 1992 would equal 60 for this country.

Second, we can also measure state preferences in terms of numerical ideal points.
These ideal points capture what international relations scholars have called countries’ liberalism on issues such as political freedom, democratization, and financial liberalization. The two measures are highly correlated, with larger (more liberal) ideal points corresponding to a higher percentage of votes that agree with the US.

Question 1

We begin by examining how the distribution of state ideal points has changed since the end of communism. Plot the distribution of ideal points separately for 1980 and 2000 - about ten years before and after the fall of the Berlin Wall, respectively. Add median to each plot as a vertical line. How do the two distributions differ? Pay attention to the degree of polarization and give a brief substantive interpretation of the results. Use the quantile function to quantify the patterns you identified.

un= read.csv('~/Desktop/POL-2504/unvoting.csv', header=TRUE)

un1980 = subset(un, subset = (Year == 1980))
un2000 = subset(un, subset = (Year == 2000))

hist(un1980$idealpoint, xlim = c(-3, 3),
     xlab = "Ideal Point",
     main = 1980)
abline(v = median(un1980$idealpoint))

Question 2

Next, examine how the number of countries voting with the US has changed over time. Plot the average percent agreement with US across all counties over time. Also, add the average percent agreement with Russia as another line for comparison. Does the US appear to be getting more or less isolated over time, as compared to Russia? What are some countries that are consistently pro-US? What are the most pro-Russian countries? Give a brief substantive interpretation of the results.

years = unique(un$Year)

plot(x = years, y = tapply(un$PctAgreeUS, un$Year, mean, na.rm = TRUE), 
     type = "l",
     main = "US/Russia Agreement Over Time",
     col = "blue",
     ylim = c(0, 1),
     xlab = "Year",
     ylab = "Average percent agreement")

lines(x = years, y = tapply(un$PctAgreeRUSSIA, un$Year, mean, na.rm = TRUE),
      col = "red")
text(2000, 0.9, "Russia")
text(2000, 0.1, "USA")

Question 3

One problem of using the percentage of votes that agree with the US or Russia as a measure of state preferences is that the ideological positions, and consequently the voting patterns, of the two countries might have themselves changed over time. This makes it difficult to know which countries’ ideological positions have changed. Investigate this issue by plotting the evolution of the two countries’ ideal points over time. Add the yearly median ideal point of all countries. How might the results of this analysis modify (or not) your interpretation of the previous analysis?

Question 4

Let’s examine how countries that were formerly part of the Soviet Union differ in terms of their ideology and UN voting compared to countries that were not part of the Soviet Union. The former Soviet Union countries are: Estonia, Latvia, Lithuania, Belarus, Moldova, Ukraine, Armenia, Azerbaijian, Georgia, Kazakhstan, Kyrgyzstan, Tajikistan, Uzbekistan, and Russia. The %in% operator, which is used as x %in% y, may be useful. This operator returns a logical vector whose element is TRUE (FALSE) if the corresponding element of vector x is equal to a value contained in vector y. Focus on the most recently available UN data from 2012 and plot each Post-Soviet Union state’s ideal point against the percentage of its votes that agree with the United States. Compare the post Soviet Union states, within the same plot, against the other countries. Briefly comment on what you observe.

Question 5

We have just seen that while some post-Soviet countries have retained non-liberal ideologies, other post-Soviet countries were much more liberal in 2012. Let’s examine how the median ideal points of Soviet/post-Soviet countries and all other countries has varied over all the years in the data. Plot these median ideal points by year. Be sure to indicate 1989, the year of the fall of the Berlin Wall, on the graph. Briefly comment on what you observe.

Question 6

Following the end of communism, countries that were formerly part of the Soviet Union have become much more ideologically diverse. Is this also true of the world as a whole? In other words, do countries still divide into two ideological factions? Let’s assess this question by applying the k-means clustering algorithm to ideal points and the percentage of votes agreeing with the US. Initiate the algorithm with just two centroids and visualize the results separately for 1989 and 2012. Briefly comment on the results.

Question 7

We saw earlier that the median post-Soviet country joined the liberal cluster after the fall of communism. Which countries, possibly from outside the Soviet Union, followed suit? Conversely, are there countries that exited the liberal cluster after the end of communism? Identify the countries which were in the non-liberal cluster in 1989 but belonged to the liberal cluster in 2012. Again, the %in% operator may be useful. Then, do the reverse, so that you can see which countries exited the liberal cluster. Briefly comment on what you observe.