The Economics of AI and Robotics: A Scenario Analysis of Automation, Employment, Poverty, Taxation and Income Distribution in a Hypothetical Three-Sector Economy
Year: 2026
Type: preprint
Abstract: This study examines how the deepening of artificial intelligence (AI) and robotics, defined as the progressive replacement of human workers by machines, affects output, employment, poverty and income distribution. Drawing on the economics of prediction and general-purpose technologies, the task-based theory of automation, growth theory, the literature on factor shares and Keynesian demand constraints, we construct a hypothetical economy, Solvania, with 30 million inhabitants, GDP per capita of US$25,000, 6% unemployment, a 5.3% poverty rate and three sectors: agriculture (30% of output and employment), automobile manufacturing (20%) and services (50%). A macro-sectoral model linked to a 150,000-household microsimulation, implemented in Python, traces a gradual scenario, in which automation reaches 25% of jobs within five years and 50% by year 20, and a rapid scenario, in which it reaches 80% in agriculture and manufacturing and 60% in services within ten years. Both paths raise GDP strongly, to US$1.57 trillion by year 20 (gradual) and US$1.39 trillion by year 10 (rapid), but the gains are distributed very unequally. In the rapid scenario unemployment reaches 47.2%, the labour share falls from 60% to 19.9%, the Gini coefficient rises from 0.369 to 0.621 and poverty more than triples to 17.3%, while median income rises by only 8%. A policy package combining an automation levy that finances a universal dividend, universal income support, accelerated reinstatement and work sharing reduces poverty to 0.7%, lowers the Gini to 0.408 and raises GDP by a further 16%. The results are robust to alternative assumptions and support redistributive and labour market institutions established before the fastest phase of automation. JEL classification: D31, E24, H23, I32, J23, J31, O33
Source: Research Square
Author: Muhmmad Muqeem Islam Muqeem
Institution: National School of Public Policy
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7235981719
