Non-governmental organizations, boomerangs, and forest loss: a cross-national analysis

ABSTRACT

On the one hand, some research finds that higher numbers of domestic non-governmental organizations (NOGs) are related to less forest loss, while others find that higher numbers of international NGOs are related to less forest loss. Moreover, some find that neither domestic nor international NGOs have an impact on forest loss. We argue that Keck and Sikkink’s insight regarding a boomerang pattern of influence between domestic and international NGOs is the missing piece of the puzzle regarding past contradictory findings. While the activities of domestic NGOs should be successful at reducing forest loss given their local accountability and closeness to affected populations, their efforts could be more effective when combined with the resources and power of international NGOs. To test this hypothesis, we use ordinary least squares regression analysis for a sample of 75 low- and middle-income nations to see how the interaction between domestic and international NGOs impacts forest loss. In support of the ‘boomerang effect,’ we find domestic NGOs are related to less forest loss in low- and middle-income nations with higher rather than lower numbers of international NGOs. The findings suggest that international NGOs enhance the ability of domestic NGOs to decrease forest loss by providing financial and organizational resources.

KEYWORDS:

• Non-governmental organizations
• boomerang
• forest loss
• a cross-national analysis

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. We find high variance inflation factor scores in the model that includes the interaction term. This is not surprising and rather is somewhat expected. However, we feel that it does not present a problem here. This is because multicollinearity increases the standard errors of the variables that are highly correlated (Allison 2012). In such an instance, it means the standard error for the interaction term may be larger, thereby leading to a more conservative test of statistical significance than otherwise the case (Allison 2012).

2. A sample of 75 low- and middle-income nations after listwise deletion of missing data is in line with the samples of other cross-national studies of forest loss including Tester (2020) and Restivo, Shandra, and Sommer (2018). It is also not surprising because we exclude nations in the Middle East and North Africa because they have little to no forest cover like these and other studies.

3. We replicated the model using domestic and international non-governmental organization variables standardized by population size in millions. Both measures – total organizations and organizations standardized by population – capture the extent to which the domestic or international environmental community is active within a country. The first measure accounts for the fact that organizations are not limited by a country’s population size as well. The second measure accounts for the tendency of more populated countries to have more organizations working within their borders. The results are substantively similar to what is reported in this article.

4. We replicated the analyses using the Polity IV (Marshall et al. 200) measure of democracy rather than the Freedom House (2005) scales. The results are substantively similar.

5. Following Jorgenson and Burns (2007), we included rural and urban population growth in the models. We find that higher levels of rural population growth correspond with increased forest loss. However, urban population growth does not predict any significant variation in forest loss. The other results remain stable and consistent across these model specifications.

6. There is a very real concern that endogeneity may be influencing our results in regard to the non-governmental organization variables. Thus, we calculate Davidson and MacKinnon (1993) augmented endogeneity test for the international and domestic non-governmental organization variables. This procedure is carried out in the following way. First, we run an ordinary least squares regression with each endogenous variable as the dependent variable as a function of the exogenous variables in the original regression model and an instrument. We use the threatened number of bird species in a country as our instrument because Lewis (2000) argues that it represents a recipient nation’s need (i.e., high levels of biodiversity loss) for intervention by international and domestic non-governmental organizations. Second, we calculate residuals from this equation. Third, we rerun the original regression equation with forest loss as the dependent variable as a function of the endogenous variable, its residuals, and the exogenous independent variables. If the coefficient for the residuals reaches a level of statistical significance, then endogeneity may be a problem in the analysis and the tests of significance of the ordinary least square estimates are inefficient and the coefficients are biased (Davidson and MacKinnon 1993). The coefficients for the residuals for international and domestic non-governmental organizations do not reach a level of statistical significance so we conclude that endogeneity should not be problematic in our analysis.

Additional information

Notes on contributors

Samia Tasmim

Samia Tasmim is a graduate student of Sociology at the State University of New York at Stony Brook. Her research focuses on how different policies and institutions impact environment and development.

Jamie M. Sommer

Jamie M. Sommer is an Assistant Professor of Sociology at the University of South Florida. Her research on how the interaction between internal and external factors at the cross-national level impact environmental issues is published inThe Sociological Quarterly, Sociological Perspectives, and Society & Natural Resources.

Kristen Shorette

Kristen Shorette is an Assistant Professor of Sociology at the State University of New York at Stony Brook. Her work on cross-national inequalities in environmental disruption has appeared inSocial Forces, The Journal of World-Systems Research, and Sociology Compass. 

John M. Shandra

John M. Shandra is a Professor of Sociology at the State University of New York at Stony Brook. His research on how organizations impact environment and development outcomes has appeared in several journals includingSociological Inquiry, Rural Sociology, and the International Journal of Comparative Sociology.