Search Results (3)

Shapley values are the most widely used point-valued solution concept for cooperative games and have recently garnered attention for their applicability in explainable machine learning. Due to the complexity of Shapley value computation, users mostly resort to Monte Carlo approximations for large problems. We take a detailed look at an approximation method grounded in multilinear extensions proposed in 2021 under the name “Owen sampling”. We point out why Owen sampling is biased and propose unbiased alternatives based on combining multilinear extensions with stratified sampling and importance sampling. Finally, we discuss empirical results of the presented algorithms for various cooperative games, including real-world explainability scenarios. Full article

In developing countries, where economic expansion depends on extractive activities such as agriculture and mining, water quantity and quality considerations need to be examined in tandem with GDP growth and poverty reduction efforts. Poorest households in the Tupiza watershed in Bolivia are located in rural areas where water access for irrigation and safe drinking water is becoming increasingly scarce. Small-scale unregulated mining offers an alternative for revenue making in rural households, although wastewater from industry threatens water quality and new technologies to reduce water pollution are not implemented in this region yet. This study analyses water access and poverty linkages using the Multidimensional Poverty Analysis (MDPA) framework from the Swedish International Development Cooperation Agency (Sida) to guide a survey and to develop indicators using the Shapley-Owen decomposition method and multivariate regressions. A set of household-level policies were included to predict the influence of these policies on poverty reduction estimates. Results have shown that remote communities in rural areas have the lowest value of multidimensional poverty and for some of these communities mining activities represent an alternative that could be considered if their water access conditions worsen over time. While mining can bring better monetary benefits, it can cause the degradation of ecological flows from the produced wastewater. Under the current technologies and processes, it can pose negative impacts on water quality and threatens the public health of these communities. Full article

Understanding how growth factors contribute to explaining the large differences in growth rates across countries remains an important research agenda. The common approach to exploring this issue is based on the use of multiple linear regression analyses. This work contributes to growth literature by applying a new perspective based on the use of variance decomposition procedures: Shapley–Owen–Shorrocks and Oaxaca–Blinder. These methodologies have four main advantages with respect to traditional methodologies: they make possible the quantification of the relative contribution of each factor to economic growth, they allow us to estimate the efficiency in the use of the endowments of each factor, they can be used with any functional form and they can be used with estimation methods that are robust regarding endogeneity issues. We illustrate these advantages by analyzing the causes of the economic growth gap between Latin America and East Asia over the period 1980–2014. We find that the economic growth divergence between the two regions can be primarily explained by the differences in institutions and physical capital. In addition, the results indicate that the higher East Asian performance is not only due to its higher levels of endowments in these factors, but also to the higher efficiency in its use. We connect our results with the 2030 Agenda for Sustainable Development. Full article