The Comparable Sustainable Development Index (CSDI) for an evidence-based sustainability policy

On a macroeconomic level, sustainability goals have been operationalised. In order to monitor and assess the progress of achieving these objectives, data collection and generation is crucial. Sustainability indicators and... [ view full abstract ]

On a macroeconomic level, sustainability goals have been operationalised. In order to monitor and assess the progress of achieving these objectives, data collection and generation is crucial. Sustainability indicators and indices are the most successful assessment method in measuring sustainability. However, existing reporting systems and indices do not always include all three sustainability dimensions or do not apply to microeconomic, mesoeconomic and macroeconomic level at the same time. To overcome these shortcomings, we developed a holistic, comparative sustainability reporting system and the Comparable Sustainable Development Index – the CSDI. This work in progress investigates sensitivities of the CSDI’s methodological approach, aiming to determine the most efficient method such that the index will serve as a reliable tool in management control systems and evidence-based sustainability policies. An efficient index is defined as an index whose underlying preference order is independent of admissible transformations of the variables which describe the states of each sustainability dimension. The first step in calculating the CSDI is raw data collection of sustainability key figures like the gross value added, CO₂ emissions and number of apprentices, followed by missing value imputation. Single time series imputation as well as multiple imputation by an expectation-maximising bootstrapping algorithm is performed. Hereafter, the CSDI’s special feature, i.e. applicability to any aggregation level and cross-unit comparability, is obtained by firstly transforming microeconomic and mesoeconomic data into macroeconomic categories and secondly by standardising to the size of the object of investigation in terms of gross value added or employment. The absolute deviation around the median is utilised to detect outliers and re-scaling is applied, accounting for the indicators’ diverse units. The indicators’ weights are worked out by three multivariate statistical techniques: principal component analysis (PCA), partial triadic analysis (PTA) and the maximum relevance minimum redundancy backward criterion (MRMRB). The last calculation step is aggregation, which includes aggregation within and across each sustainability dimension. It is conducted by the geometric mean, rewarding good and punishing bad outcomes. The CSDI has been computed for branches of the German economy as well as the overall economy. The data analysis has revealed that missing value imputation by time series regression yields less volatile and thus more plausible results than the more sophisticated multiple imputation technique. Regarding the determination of weights, the MRMRB criterion outperforms the PCA and PTA as non-linear correlations are factored in and therefore, contentual richer indicators are weighted more heavily. The choice of the geometric mean is especially remarkable for the chemical industry: its bad performance in the environmental dimension cannot be offset by its good performance in the economic and social dimension. The goal of the CSDI is the comparison of sustainability outcomes in all three dimensions of economic units regardless of their aggregation level. Several methodological approaches have been explored and further work will involve sensitivities of outlier detection and treatment as well as omitting/adding indicators. Despite general limitations of such an index, e.g. non-inclusion of the supply chain, non-avoidance of slight value judgement, the CSDI will become a reliable tool in management control systems and evidence-based sustainability policies.