Do sanctions lead to a decline in civil liberties?

Just before the outbreak of the Peloponnesian War, the Athenian empire levied sanctions on the city of Megara, excluding Megarians from the empire’s market and harbors. For many scholars, the economic consequences of those sanctions were a decisive factor in the outbreak of the Peloponnesian War (Brunt 1951).

The recent debate over the appropriateness of imposing sanctions in order to destabilize the Maduro regime in Venezuela is suggestive: proponents of the use of sanctions claim that immiserizing the general population will destabilize the regime and eventually improve the conditions in the country. On the other hand, opponents consider the measures to be ineffective since they give the regime the opportunity to use propaganda and repression to strengthen its position further.

With that in mind, our baseline analysis includes a general category of sanction episodes and not only those with the goal of reducing repression and destabilizing the targeted regime.

To examine the robustness of our results, we also examine UN-imposed sanctions as well as sanctions imposed by all states.

Choi and James (2016) provide evidence that US military interventions also are primarily prompted by the same reason.

Specifically, Peksen and Drury (2009, 2010) find that sanctions have a negative effect on the level of democracy, as measured by the Freedom House index of civil and political liberties, which in the long run is equivalent to approximately a one-point decline in the index on its 1–7 cardinal scale. Peksen (2009) and Wood (2008) find similar effects using an index of respect for human rights and an index of political terror, as dependent variables, respectively.

Nooruddin (2002) deals with the problem of selection by estimating a selection model in the spirit of Heckman (1979). Similarly, Jing et al. (2003) estimate a two-stage model wherein sanctions are imposed according to an endogenous political process and the model therefore controls effectively for both the selection and the confounding factors problem. Here we follow another approach, mimicking randomization.

Kaempfer et al. (2004) consider a policy of using the sanctions to target directly the means of repression as more effective, since it increases the price of repression and thus destabilizes the regime.

The countries in our sample are listed in the “Appendix”.

The political liberties ideals involve, as Gastil (1982, p. 7) points out, the “rights to participate meaningfully in the political process. In a democracy this means the right of all adults to vote and compete for public office, and for elected representatives to have a decisive vote on public policies”. Civil liberties ideals involve a series of various economic, political and civil liberties enjoyed by the citizens of the country, such as freedom of expression and belief, association and organization rights, rule of law and personal autonomy and economic rights. Similarly, in the words of Gastil (ibid.), “Civil liberties are rights to free expression, to organize or demonstrate, as well as rights to a degree of autonomy such as is provided by freedom of religion, education, travel, and other personal rights”.

As robustness analysis, we also examine the effect of sanctions imposed by the United Nations and sanctions imposed by all states.

The United States participated in 46.5% of all sanctions imposed and issued 56.4% of all sanction threats from 1946 up to 2004.

Variable definitions, sources and descriptive statistics are given in the “Appendix” (Table 8).

See also Angrist and Kuersteiner (2011) for the technical details.

The weights used are calculated according to the inverse probability weighting scheme.

The three models discussed above rely on three basic assumptions to estimate the ATET: (i) conditional Independence, i.e., after conditioning on the covariates, the outcomes (civil liberties) are conditionally independent of the control-level potential outcome, (ii) overlap, i.e., each treated observation has a positive probability of being allocated to each treatment level and (iii) i.i.d., which in our setting rules out interactions among countries in each period. If we were to estimate the ATEs (Average Treatment Effects), those assumptions would have taken a more restrictive form, requiring both conditional independence and overlap to hold for both treatment statuses (Wooldrige 2010). For more details on the assumptions, see Imbens and Wooldridge (2009) and Angrist and Pischke (2009). To inspect visually whether the overlap assumption holds in Fig. 3 in the “Appendix”, we present the smoothed densities, using a standard Epanechnikov kernel, of the estimated propensities of receiving sanctions between the two groups. As the reader can verify, considerable overlap is found among treated and control propensities. What is more important, the control observations cover the support for all treated observations, when either the Freedom House or the empowerment rights index is used. That evidence provides support for the required overlap assumption and gives suggestive evidence in favor of our empirical strategy.

We conceive the doubly robust method as being more general, since it combines the two other methods and we rely solely on it for the robustness analysis. The associated impulse response functions for the regression adjustment and the inverse probability weighting methods are shown in the “Appendix”. The remaining results are available from the authors and are qualitatively similar to the ones presented here.

The finding also verifies the analysis of Kaempfer and Lowenberg (1999), who show that unilateral sanctions are more effective in reducing repression than multilateral sanctions.

We should note that in 58 out of 67 sanction episodes in our sample, the United States also participated as a primary or secondary sender.