Abstract:
This thesis empirically investigated the impacts of institutions on the environment. Using the instrumental variable (IV) strategy, the effects of both formal and informal institutions and their interactions were estimated on mitigating the emissions of carbon dioxide across 140 countries over the span of 25 years (1990-2014).
In order to reduce environmental pollution, the conventional collective-action theory calls for an external formal power to monitor and punish people who do not take proper actions. However, there is a lack of sound institutional analysis in empirical studies. There is no consensus among scholars on the efficacy of formal institutions, as both positive and negative effects are documented according to different environmental indicators. In explaining such contradictory results, scholars mainly rely on the risks of free-riding activities that exist in universal collective action problems like mitigating emissions. To minimise free-riding risks, the updated collective action theory introduces reciprocal cooperation, maintaining which depends on virtues, social norms, and high interpersonal trust. With more trust, less protection against and monitoring of free-riding is required. However, the role of informal institutions is broadly ignored in environmental analyses, which indicates the need for empirical testing of the updated theory.
To build strong institutional foundations, I adopted the inclusive SES framework as the conceptual foundation and reshaped it to fit into current cross-country empirical research, taking into account factors such as research questions, focal level of analysis, institutional theories and empirical specifications. In searching the literature for possible estimation techniques suitable for examining the institution-environment nexus, keeping in mind that institutions are inherently endogenous, I noted that the IV strategy has not been employed. Scholars have mainly drawn on simple OLS methods. Data availability is another common issue that I identified in the empirical analyses, especially in the case of poor countries. To rectify the problems and obtain reliable estimations, I constructed a panel dataset and drew on the FE-IV estimator. The proposed empirical specification proved to be compatible with the conceptual framework, cross-country level of analysis, and nature of the institutional and environmental problems.
Nevertheless, existing empirical studies lack proper time-variant instruments for formal and informal institutions. Relying on the current institutional economics literature, I used the colonial origins of countries as instruments for formal institutions. The instrumental variables were then interacted with time to denote the difference in developmental paths of countries over time, which is caused by having different colonial origins. I further extended the literature by constructing a variable named distance to conflict zones for instrumenting the employed measure of trust. This study benefitted from the use of seven different institutional measures, six of which represent formal institutions, including political, legal and economic systems. A new variable, religious tensions, was employed for quantifying the level of within-nation trust; compared to the conventional measure, this allowed for far more observations and could be better explained by the constructed instrument.
The thesis carried out the following four empirical studies. First, to test the conventional theory, the impacts of formal institutions were studied. The second study extended current knowledge by incorporating formal and informal institutions into the analysis for the purpose of testing the updated theory. This study represents the main contribution to the existing research. The interaction of institutions was then added to the model specification in the third study. Overall, the results confirmed that carbon emissions are effectively mitigated in countries with stronger formal and informal institutions. The robustness of estimations were further checked in the fourth study, using different estimation techniques (i.e., two-step Sys-GMM), different samples of countries (i.e., resource-rich and resource-poor countries), and different dependent variables (i.e., emissions stemming from different fossil fuels and sectors). The results support the EKC relationship.