Functional interrelations of governance elements and their effects on tropical deforestation - combining qualitative and quantitative approaches
von Richard Fischer
Datum der mündl. Prüfung:2020-11-20
Betreuer:Prof. Dr. Lukas Giessen
Gutachter:Prof. Dr. Lukas Giessen
Gutachter:Prof. Dr. Carola Paul
EnglischThe global rate of annual net forest loss has slowed from 7.8 mio hectares in the 1990s to 4.7 mio hectares between 2015 and 2020. Nevertheless, the area of the world’s forests continues to decrease specifically in the tropics. Improved governance has come into the focus as a means to help reversing trends of tropical deforestation. Yet, “good governance” remains a normative, broad and often unspecified concept consisting of a wide range of elements and implicit value judgements. Specific knowledge is missing on the relative importance of single governance elements, on their interdependencies and their specific effects on deforestation. Forest governance research to date has a strong focus on qualitative approaches. This study aims to (i) elaborate on and implement mixed methods for forest governance measurements, (ii) determine functional relationships between forest governance elements based on quantitative data, and to (iii) substantiate and quantify governance effects on tropical deforestation. The presented research develops a new method called quantitative content analysis with standardized scores. The method is applied in a literature review comprising 28 reviewed publications. This review classifies governance elements based on the framework of the World Resource Institute. It quantifies effects on deforestation for single elements by Likert scores. In addition, this study presents a harmonized landscape level governance assessment methodology which is developed and implemented in research areas covering approximately 500,000 hectares in Ecuador, Zambia and the Philippines. Both methods combine qualitative and quantitative approaches and are shown to be applicable and operational. In order to analyze functional relationships between governance elements, principal component analysis (PCA) is applied to all data sets. The results show two general main mechanisms. Firstly, there is a joint positive loading of governance elements on the first principal component for all data sets. Governance elements thus function synchronously. They are expressions of a similar underlying mechanism. Secondly, results show that for the review data structural and agency related governance elements are grouped on specific principal components. These components together describe 38% of the variation of governance elements. For the first time, governance functioning is described by these two aspects based on an empirical data set. However, for neither of the landscape level data sets such a functional structure - agency description is possible. Instead, country specific governance elements have major importance in the landscape level data sets. Effects of governance on deforestation are analyzed by multiple regression analysis for the data sets from Ecuador and Zambia. Deforestation rates are calculated based on satellite data and are used as target variable. In addition to governance elements, context data on deforestation drivers are assessed in the landscapes and used as explanatory variables. The different models explain approx. 50% of the variation in deforestation. Direct drivers such as agriculture and infrastructure explain largest shares of deforestation. However, an additional positive effect of single, country specific governance elements can be substantiated. The study concludes that for forest governance research mixed method approaches need stronger consideration. Data transformation into quantitative scores enables generalization of knowledge based on a multitude of case studies. Existing studies thus gain added value and should be considered before new field research is implemented. New policy requirements and research questions, however, will necessitate new field studies. These studies need to rely on harmonized approaches to which this study makes an important contribution. The synchronous function of different governance elements is encouraging for development work and policy. The joint positive loading of governance elements on the first principal components can motivate to concentrate on governance core features that are relevant in the specific context. Possible synergies between governance elements need to be further researched. The structure - agency approach can help to select relevant elements. The study shows that both, structure and agency aspects need to be considered. The REDD+ approach is a prominent example for this. Within REDD+, structural measures - in so-called readiness phases, as well as agency related incentives - through so-called results based payments need to complement each other. The structure - agency dualism only became visible within the pan tropical review data set. This can imply that the landscape level is not sufficient to tackle governance issues. Multilevel governance is obviously required spanning from international, to national and local levels. Direct deforestation drivers like agriculture and infrastructure had stronger effects on deforestation as compared to governance measures, which are regarded as underlying factor. A governance focus alone can thus not compensate effects of direct drivers. However, without governance measures work on direct drivers may not be successful. Compared to a normative “good” governance approach, the presented analytical approach explores causalities. It is outcome oriented. Based on such an approach, measures can rely on jointly agreed aims instead of input and value oriented principles. This can facilitate development work, because specifically the informal values often differ between actors.
Keywords: Governance; Deforestation; Principal Component Analysis; Multiple linear regression; Ecuador; Zambia; The Philippines; Forest; Tropics