Evaluating Economic Impacts of Automation Using Big Data Approaches

Abstract

As automation is increasingly driven by advanced technological integration, quantitatively evaluating its economic impacts becomes crucial. This paper studies the effects of automation on three economic outcomes: transactions, sales, and costs. First, we use big data approaches to distinguish transaction distribution patterns across various temporal segments. These methods employ survival and mean residual functions to cluster transaction distributions and customer traffic data over time. Empirical evidence provides distinct clusters, distinguishing high and low customer traffic. Second, we illustrate how automation can lead to higher forecast accuracy in sales. This approach utilizes stochastic error distance for comparing forecast error distribution functions. Lastly, we study the impact of automation on costs through a probabilistic model. The results suggest that while labor costs increase due to retraining and longer hours, a potential reduction in turnover and waste costs can offset these rises. The impacts of automation and the applicability of methods are demonstrated through Monte Carlo simulations and empirical studies.