Breaking News — A groundbreaking study published today in Research Policy reveals that the programming languages used by developers on GitHub can predict a nation's GDP, income inequality, and carbon emissions with accuracy that traditional economic data cannot match. Using data from the GitHub Innovation Graph, four researchers have created a new measure called “software Economic Complexity Index (ECI)” that finally quantifies the “digital dark matter” of the global economy.
“For the last fifteen years, economists have measured national complexity by looking at physical exports, patents, and research papers. But software was completely invisible — code doesn’t go through customs,” said Sándor Juhász, a research fellow at Corvinus University of Budapest and co-author of the paper. “Now, with GitHub’s data, we can see the productive knowledge embedded in every git push across borders.”
The study, by Juhász, Johannes Wachs (Corvinus University and Complexity Science Hub), Jermain Kaminski (Maastricht University), and César A. Hidalgo (Toulouse School of Economics and Corvinus University), applies the established Economic Complexity Index to the diversity and ubiquity of programming languages used in each country. The result: software complexity explains variations in economic outcomes that conventional metrics miss.
Background
Traditional economic complexity measures rely on tangible goods, research articles, and patents — all of which miss the vast amount of knowledge embedded in software. Code flows via cloud services, package managers, and collaborative platforms like GitHub, leaving no customs record. This “digital dark matter” has long been suspected to be a powerful economic signal, but no dataset large enough existed to test it.
The GitHub Innovation Graph, which tracks the number of developers per economy pushing code in each programming language (based on IP addresses), filled that gap. The researchers used Q4 2025 data from the Innovation Graph to build their software ECI and correlated it with GDP per capita, the Gini coefficient, and CO₂ emissions per capita across dozens of nations.
What This Means
“Countries that develop a diverse set of programming languages — similar to how they export a diverse set of products — tend to have higher incomes and lower inequality,” explained Jermain Kaminski, assistant professor at Maastricht University. “But software complexity predicts these outcomes even after controlling for traditional measures, meaning it captures unique information about a country's digital capabilities.”
The findings have immediate policy implications: governments can now use real-time developer activity to track which high-skill industries are emerging, spot vulnerabilities in their tech ecosystems, and guide investments in digital education. For businesses, the data offers a leading indicator of national innovation potential. “This isn't just an academic exercise,” said Johannes Wachs, director of the Center for Collective Learning. “If you want to know where the next wave of economic growth will come from, look at where developers are pushing code — not just where factories are built.”
The study is part of a broader effort to make open-source collaboration visible as an economic force. GitHub's Innovation Graph continues to release quarterly data, enabling ongoing research. As digital economies expand, ignoring software complexity is like navigating with a map that omits oceans.