Abstract

The rapid digitalization of the global economy has accelerated the accumulation of intangible assets, particularly code repositories and Artificial Intelligence (AI) models. Despite the growing recognition of data as a critical economic asset, national statistical offices (NSOs) face significant methodological challenges in measuring these non-monetary digital flows. In addition, the evolving cost structures in modern enterprises, expenditures on code, software development and AI models frequently exceed those on physical assets, underscoring the urgency of capturing these values.
To address these gaps, this paper proposes a methodological framework that leverages open-source repositories as nontraditional data sources to estimate the economic valuation of data and knowledge as an asset as well as to measure digital services and asset trade. Our approach leverages open-source platforms: we analyze code repositories (e.g., GitHub) to quantify software development services and AI model/dataset repositories (e.g., Hugging Face) to measure AI-related services and assets. We estimate the economic value of contributions using proxy metrics (such as lines of code committed or model downloads), weighted by a skilled wage proxy to reflect production cost. We utilize geographic residence metadata from repository owners, contributors, and users to enable identification of trade flows and benchmarking against other regions or countries.
Initial findings from the Indonesian case study show that our estimated digital services and asset trade closely align with Indonesia’s ICT development trends. The results show the trade flows by countries and reveal that Indonesia is a net importer of digital assets, with a large portion of its open-source digital tools and content originating from more technologically advanced economies.
This research contributes to validating open-source repositories as a scalable new data source. Furthermore, the proposed framework helps bridge current data gaps in official statistics. It offers policymakers, statisticians, and data scientists a novel methodology to capture the economic value of intangible assets and AI-related economic activity by providing a micro-level measurement approach for macro-level indicators.