“Leave It as It Is”: International Network Effects on Protected Lands

Abstract

The protection of a portion of a country’s land is vital for sustainable economic growth and biodiversity, though land protection also imposes important costs. States have set aside a growing proportion of their land for protection over time, with wide variation among states. What explains this variation? Theoretically, we argue that states exist internationally in dependence networks with each other and that those networks provide pathways for influence on a state’s environmental choices. A state’s dependence network is the other states with which it regularly exchanges valued goods. We find strong evidence that increases in protected lands among trade partners and international organization partners (both weighted by importance) increase a given state’s protected lands, with substantive effects larger than domestic-level variables like democracy. This paper expands our understandings of the ways that states may influence each other on environmental policy.

La protección de una parte del terreno de un país es vital para el crecimiento económico sostenible y para la biodiversidad, aunque la protección de la tierra también conlleva unos costes importantes. Los Estados han estado reservando para su protección una proporción, cada vez mayor, de sus tierras a lo largo del tiempo. Sin embargo, existe una amplia variación a este respecto entre los Estados. ¿Qué explica esta variación? De forma teórica, argumentamos que los Estados existen internacionalmente en redes de dependencia entre sí y que esas redes proporcionan diversas vías que influyen en las elecciones ambientales de un Estado. La red de dependencia de un Estado está formada por otros Estados con los cuales intercambia regularmente bienes valiosos. Encontramos pruebas sólidas de que los aumentos en el número de áreas protegidas entre los socios comerciales y entre los socios de las organizaciones internacionales (ambos ponderados por importancia) aumentan también la cantidad de áreas protegidas de un Estado determinado. Estos aumentos conllevan unos efectos sustantivos mayores que las variables a nivel nacional como la democracia. Este artículo amplía nuestra comprensión acerca de las formas en que los Estados pueden influir mutuamente en la política ambiental.

La protection d’une partie des terres d’un pays est cruciale pour une croissance économique durable et la biodiversité, mais elle impose aussi des coûts importants. Au fil des années, les États consacrent une part croissante de leur territoire à la sauvegarde, mais elle varie grandement de l’un à l’autre. Comment expliquer ces différences ? Sur le plan théorique, nous affirmons que les États existent à l’international au sein de réseaux d’interdépendances qui ont tendance à influencer les choix environnementaux étatiques. Le réseau d’interdépendances d’un État regroupe les États avec lesquels il échange régulièrement des marchandises de valeur. Nous trouvons nombre d’éléments probants pour justifier qu’une augmentation de terres protégées chez des partenaires commerciaux et des organisations internationales partenaires (en pondérant les deux en fonction de leur importance) débouche sur une augmentation des terres protégées dans un État donné. Cet effet se révèle plus important que les variables nationales comme la démocratie. Cet article enrichit notre compréhension des moyens par lesquels les États peuvent s’influencer mutuellement en politique environnementale.

Keywords:

• Dependence networks
• environmental protection
• international networks
• protected lands

Acknowledgments

Previous versions of this article were presented at the 2019 Annual Meeting of American Political Science Association and an Environmental Politics and Governance workshop in October 2020. We thank the participants for their feedback, especially Elizabeth Albright, Danilo Freire, and Cesar Martinez-Alvarez. We recognize the assistance of Joanna Burstedt, Chad Faulkner, Abbey Higham, Jonathan Moyer, and Adam Roberts; and the editors of International Interactions and anonymous reviewers for their comments and suggestions. All mistakes remain our own.

Notes

1 We use the IMF Direction of Trade Statistics.

2 We show the formula for Trade Network in the main text, and the formulas for other networks in the Supplemental Information.

3 This means that we start our analysis in 1991, using data from 1990. Econometrically, this is a spatial autoregressive model (Franzese and Hays 2008). As Cook, Hays, and Franzese (2023, 61) note, “first-order time lags seem to suffice in most applications” with annual data.

4 Guler, Guillén, and Macpherson use quadratic weights on trade, positing that “strong ties may have a substantially higher impact than … weak ties” (2002, 221). Prakash and Potoski (2007) also use quadratic weights. We use weights to the 10^th power.

5 As an example, for Finland in 2010, its top 3 IO partners are Sweden, Denmark, and Norway, which make up 1.26%, 1.19%, and 1.18% of its IO Network, respectively. After the power-law transformation, they make up 19%, 10%, and 9% of its IO Network, respectively. This approach is similar to, but more general than, including only the top three countries. It allows there to be a different number of “top” countries for each state.

6 If we use a weaker transformation (such as a quadratic) or no transformation, our results are stronger for IO Network. However, this is likely absorbing a global effect. If we weight IOs by importance (Ingram, Robinson, and Busch 2005) we get a stronger effect. And if weight IO partners by GDP, we also obtain a stronger effect. However, weighting by GDP (or importance) is likely picking up economic factors that we want to estimate separately with trade and aid.

7 We cannot merely set Aid Network to 0 for those countries that do not receive aid. This would be equivalent to stating that all countries from which another country received aid had no protected lands. (And we do not set those countries to missing, which would drop many country-years from the model.)

8 Because we are using a model with country fixed effects, we use the “within” standard deviation for each variable, that is, how much the variable changes within a country. The short-term effect is the coefficient (reported in Table 1). The long-term effect is the variable coefficient divided by the complement of the coefficient of the lagged dependent variable. Note that the long-term effect of a variable can be statistically significant even if the short-term effect is not (and vice versa). We use the delta method to calculate the standard errors.

9 If one country increased its protection of lands, then its partners would protect more lands, which would induce those countries’ partners to protect more lands. Our model picks up the first effect, but not the second.