Georgia’s Foreign Policy Trilemma: Balance, Bandwagon, or Hedge? Part 1

Georgia is a small, partly free democracy in a tough neighbourhood, and NATO membership remains an unfulfilled promise. While Russia is widely perceived as the main threat to Georgia’s security, the appropriate strategic or political response to the threat is not obvious. What options does Georgia have when faced with a powerful rival on its border, and what public support is there for these options?

Research on the foreign policy options of small states have generally focused on elite attitudes and structural determinants, but – with some exceptions – less so on the relationship between public opinion and foreign policy. This gap matters as public preferences for a country’s foreign policy (should) be important to decision-makers in a democracy. While foreign policy generally ranks lower than domestic concerns in the publics’ priorities, high stake questions of national security do cut through, and a foreign policy that lacks public support will be harder to sustain. This leads to the question, do threat perceptions predict foreign policy preferences?

Foreign policy analysts often talk about geopolitical choices in terms of balancing and bandwagoning: a state can either form alliances with other countries to balance against the threat of a stronger state, or bandwagon with the country that threatens them in the hope that the threat is mitigated by aligning themselves with the more powerful country.

Georgia’s options can seem like a stark choice between Russia and the West: should Georgia balance against Russia and deter the threat by developing closer alliances with Europe and the US, or bandwagon with Russia and forge closer relations with their northern neighbour? Each option has its advocates and detractors. Balancing can be seen as either a rational response to guarantee survival or as unnecessarily antagonistic. Bandwagoning can be seen as either the pragmatic management of geopolitical realities or appeasement and capitulation.

A third option is for Georgia to hedge its bets and strike a path between balancing and bandwagoning between Russia and the West. This risk management strategy may be well-suited for small states caught between Great Powers, including Russia’s neighbours who want to escape the influence of the region’s dominant power and protect themselves from East-West tensions, while working within the constraints of geography. Armenia’s attempts at a ‘multi-vectored’ foreign policy reflect this logic as well as the difficulties of implementation.

In the March 2016 CRRC-Georgia/NDI survey, respondents were asked to choose between four exclusive options: “In your opinion, Georgia's foreign policy should be Pro-Western; Pro-Western, however we should maintain good relations with Russia; Pro-Russian, however we should maintain good relations with the EU and NATO; or Pro-Russian?”. Georgians clearly favour a pro-Western foreign policy but there is a large constituency that does not want to accept a binary choice between the West and Russia. More than two thirds want what amounts to a hedging option, that is, a foreign policy that leans towards one side but without sacrificing good relations with the other side.


In theory, perceived threats to security should be an important determinant of foreign policy preferences in general, and of preferences over balancing, bandwagoning and hedging in particular. As the chart below suggests, Russia is by far the single most commonly identified threat to Georgia.

To test whether threat perceptions are related to these foreign policy preferences, a multinomial regression analysis is used with the above question about foreign preference as the dependent variable, and the threat perception question, in addition to demographic characteristics and domestic political preferences, as the independent variables.

If Georgians have a preference for balancing, there should be an association between identifying Russia as the main threat and a preference for a pro-Western foreign policy orientation that deters that threat, i.e., aligning with the European Union and NATO. The results provide evidence in support of this. As the table shows, those who identify Russia as the main threat are far more likely to support a pro-Western foreign policy (top left quadrant).



If, conversely, Georgians had a preference for bandwagoning, we would see an association between identifying Russia as the main threat and a preference for closer relations with Russia, i.e. a pro-Russian orientation – this is not the case. Contrary to the bandwagoning logic, we do not see support for a pro-Russian foreign policy amongst those who see Russia as the main threat. Rather, it is those who do not identify Russia as the main threat who are more likely to support a pro-Russian foreign policy. Notably, running the model without the threat perception variable made very little difference to the effect of the other variables on foreign policy preference.

However, as noted above, most respondents, when given the option, expressed a preference for a foreign policy that hedged between a purely pro-Western or pro-Russian orientation. The second part of this blog post, which will be published next Monday discusses this in greater depth.

Note: The analysis uses a multinomial logistic regression. The dependent variable is the foreign policy preference. The base category is ‘Pro-Western’. The table shows the predicted probabilities for the following independent variables (with base category in parentheses): threat perception (all responses other than Russia as main threat), education (higher than secondary level), party support (Georgian Dream–Democratic Georgia), and country direction (Georgia is not changing at all). The other independent variables are sex, age group, settlement type, and ethnic minority domain. Party support was recoded into four categories: GD-DG, United National Movement, No party/Don’t know, and Other. Country direction was recoded as wrong direction, right direction, or no change.

Replication code of the full analysis is available here, including alternative model specifications. The data used are available here.