When are similar regimes more likely to form alliances? Institutions and ideologies in the post-communist world

Abstract

Following the collapse of the old communist regimes, 28 post-communist countries chose from among three main foreign security arrangements: commonwealth of independent states (CIS)/collective security treaty organization (CSTO) membership, north atlantic treaty organization (NATO) membership, or neutrality. What explains these choices? We are most interested theoretically in the role played by regime type. The alliances literature typically uses a narrow institutional theory of the effects of regime type, which implies that more democratic regimes are more attractive alliance partners than more authoritarian regimes. Post-communist area specialists will be aware that this institutional theory fails to explain the apparent tendency of more authoritarian post-communist regimes to join the CIS/CSTO. We develop a broader ideological theory of how regime type affects alliances, in which political institutions are complemented by substantive ideological and policy goals. Applying the ideological approach to the post-communist world, we define and measure two main ideological regime types – liberal nationalist regimes and neo-communist authoritarian regimes. Multinomial logit regressions indicate that more democratic, liberal nationalist regimes are more likely to affiliate with NATO, whereas more authoritarian, neo-communist regimes are more likely to join the CIS/CSTO. Moreover, the desire of neo-communist authoritarianism regimes to affiliate with the CIS/CSTO is as strong or stronger than that of neo-liberal democracies to affiliate with NATO – largely because NATO is more reluctant than Russia to accept aspirants. We conclude that the ideological approach to regime type may offer significant explanatory value as a refinement of the institutional approach.

Keywords::

• post-communism
• alliances
• democracy
• authoritarianism
• regime type
• ideology

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

^1. There are mixed findings in the empirical literature. Siverson and Emmons (1991) and Leeds (1999) find that democracies are more likely to be allies with other democracies. Simon and Gartzke (1996), on the other hand, find that democracies are more likely to be allies with non-democracies. Lai and Reiter (2000) find that both democracies and authoritarian regimes are more likely to be allies with one another. See also Werner and Lemke (1997).

^2. Owen (2005) theorizes that regimes of mutually threatening ideological types would be expected to form homogenous alliance blocs against one another.

^3. A third regime type, Islamist authoritarianism, is also a possibility in some former Soviet Republics in Central Asia. But so far this option only exists in the form of internal political oppositions.

^4. Openings at a different point in time might have led local elites to view their institutional and policy options differently. For example, in a world-historical context such as the 1930s, right-authoritarian, fascist-style models might have been much stronger contenders as vehicles to pursue either reform nationalist strategies or to camouflage old regime elite efforts to preserve power and privilege.

^5. Russia is excluded because some of the hypotheses associated with control variables (discussed below) are relational, and Russia's relations with itself are not comparable to its relations with other countries. For example, Russia's ‘distance from itself’ is not comparable to another country's distance from Russia.

^6. War or frozen conflict with Russia perfectly predicts failure to seek and join an alliance with Russia within the CIS/CSTO. War or frozen conflict with either Russia or other countries perfectly predicts failure to join NATO. See below.

^7. Estimates are calculated in the R software package using the program of Colby et al. (2010).

^8. These and other variant results referred to in the main text are available from the authors upon request.

^9. The GUAM states are Georgia, Ukraine, Azerbaijan, and Moldova, and were later joined, at different times, by Uzbekistan.

^10. The average wait for admitted countries has been between 11 and 12 years. For countries still waiting to be admitted, as of 2012 (the last year in our dataset), it had been about 15 years.

^11. NATO's 1995 official statement, Study on NATO Enlargement, emphasizes preparing military forces to contribute effectively to the alliance's common defense, strengthening democratic institutions, and resolving any international and internal disputes peacefully, consistent with the norms of other Euro-Atlantic institutions and the United Nations Charter. Article 7, Chapter 1, of the Study states that,

Enlargement will occur through a gradual, deliberate, and transparent process, encompassing dialogue with all interested parties. There is no fixed or rigid list of criteria for inviting new member states to join the Alliance. Enlargement will be decided on a case-by-case basis and some nations may attain membership before others (see http://www.nato.int/cps/en/SID-EBB51EA9-BBEA610E/natolive/official_texts_24733.htm?).

^12. The first countries admitted were the Czech Republic, Hungary, and Poland, in March 1999. The second cluster, entering in March 2004, included Bulgaria, Estonia, Latvia, Lithuania, Romania, Slovakia, and Slovenia. The third and most recent group – Albania and Croatia – joined in April 2009.

^13. Freedom House scores are available at http://www.freedomhouse.org. Polity IV data are from http://www.systemicpeace.org/polity/polity4.htm.

^14. Gibler and Wolford (2006) discuss two-way causation between democracy and alliances.

^15. The index of pre-communist political and economic attainments and share of time at war are taken from Horowitz (2003). Share of time at war is updated through 2012 by the authors.

^16. Updated dataset available at http://www.pcr.uu.se/research/ucdp/datasets/ucdp_prio_armed_conflict_dataset/.

^17. COW power data are not available for Montenegro. Prior to independence, Montenegro was integrated with Serbia and had a similar level of economic development. Hence, Montenegro's power is coded as a population-weighted share of Serbia's power.

^18. See https://www.cia.gov/library/publications/the-world-factbook/.

^19. We also created an alternative measure of ethnic Russian population shares, which in countries other than Ukraine and Belarus includes Ukrainians and Belarusians along with Russians. Outside of Ukraine and Belarus, such minorities were apt to identify politically with ethnic Russians. The results are almost identical for this variant measure of ethnic Russian population share.

^20. Specifically, the data are for net imports or exports relative to consumption. For example, a country importing all of its consumption has a score of 100; one exporting the equivalent of all of its consumption has a score of − 100; and one exporting four times its consumption has a score of − 400.

^21. Available as net energy imports at www.indexmundi.com/facts. Data are not available for 2012 and sometimes also not for 2011. In such cases, the figure for the last available year (2010 or 2011) is used.

^22. The first-stage results use the raw data. The second-stage results are based on independent variables recentered to zero means and rescaled into standard deviation units, so that the coefficients are standardized.

^23. In predicting NATO aspiration, note the negative sign for choosing NATO in the previous period. The lagged dependent variable is coded ‘1’ for choosing Russia in the previous period, ‘0’ for remaining neutral in the previous period, and ‘–1’ for choosing NATO in the previous period. So the estimated negative coefficient for the lagged dependent variable for choice between NATO and neutrality, multiplied by the ‘–1’ coding for choosing NATO in the previous period, indicates a positive impact of choosing NATO in the previous period on the current period choice.