AStA Wirtschafts- und Sozialstatistisches Archiv最新文献

Productivity growth in the advanced economies has been slowing down for some time. This is surprising insofar as large-scale technological impulses are expected as a result of the digital revolution. An analysis for Germany points to significantly weakening productivity impulses from technological progress and in particular from capital formation. In order to understand the weaker capital stock effects, gross fixed capital formation is briefly examined according to different components. Starting from the established types of investment and capital, we discuss how the empirical coverage of capital as a production factor can be advanced. The focus is on a more broadly defined digital capital stock. Comprehensive indicators to describe the extent and development of the general digitization also offer conceptual starting points for defining and measuring digital capital. In this way, it could be possible to estimate what digital capital is actually available for domestic production, how it develops over time and, ultimately, what contributions to growth and productivity it makes. The paper lists a series of requirements necessary for an implementation. The dynamics of the capital stock depend on whether and to what extent current investments exceed retirements of fixed assets. In any case, rising investment per unit of labor does not necessarily indicate an increase in capital deepening. However, modern capital goods in particular are characterized by a high rate of innovation and correspondingly high retirements. This leads us to expect that even comparatively high investments based on a concept that has been comprehensively expanded in terms of digitization will not necessarily trigger correspondingly high capital stock and productivity impulses.

National Statistical Organizations (NSOs) increasingly draw on Machine Learning (ML) to improve the timeliness and cost-effectiveness of their products. When introducing ML solutions, NSOs must ensure that high standards with respect to robustness, reproducibility, and accuracy are upheld as codified, e.g., in the Quality Framework for Statistical Algorithms (QF4SA; Yung et al. 2022, Statistical Journal of the IAOS). At the same time, a growing body of research focuses on fairness as a pre-condition of a safe deployment of ML to prevent disparate social impacts in practice. However, fairness has not yet been explicitly discussed as a quality aspect in the context of the application of ML at NSOs. We employ the QF4SA quality framework and present a mapping of its quality dimensions to algorithmic fairness. We thereby extend the QF4SA framework in several ways: First, we investigate the interaction of fairness with each of these quality dimensions. Second, we argue for fairness as its own, additional quality dimension, beyond what is contained in the QF4SA so far. Third, we emphasize and explicitly address data, both on its own and its interaction with applied methodology. In parallel with empirical illustrations, we show how our mapping can contribute to methodology in the domains of official statistics, algorithmic fairness, and trustworthy machine learning.

Little to no prior knowledge of ML, fairness, and quality dimensions in official statistics is required as we provide introductions to these subjects. These introductions are also targeted to the discussion of quality dimensions and fairness.

Data collection and the availability of large data sets has increased over the last decades. In both statistical and machine learning frameworks, two methodological issues typically arise when performing regression analysis on large data sets. First, variable selection is crucial in regression modeling, as it helps to identify an appropriate model with respect to the considered set of conditioning variables. Second, especially in the context of survey data, handling of missing values is important for estimation, which occur even with state-of-the-art data collection and processing methods. Within this paper, we provide an Bayesian approach based on a spike-and-slab prior for the regression coefficients, which allows for simultaneous handling of variable selection and estimation in combination with handling of missing values in covariate data. The paper also discusses the implementation of the approach using Markov chain Monte Carlo techniques and provides results for simulated data sets and an empirical illustration based on data from the German National Educational Panel Study. The suggested Bayesian approach is compared to other statistical and machine learning frameworks such as Lasso, ridge regression, and Elastic net, and is shown to perform well in terms of estimation performance and variable selection accuracy. The simulation results demonstrate that ignoring the handling of missing values in data sets can lead to the generation of biased selection results. Overall, the proposed Bayesian method offers a holistic, flexible, and powerful framework for variable selection in the presence of missing covariate data.