Loading the current manuscript…
Working paper / September 2026
The Upside of Being Replaceable: AI, Market Structure, and Growth
A model of market structure, growth, and distributionDigital edition
Abstract
Loading the current manuscript…
From the paper to the trajectories
Two long-run regimes.
Different transitions.
Explore RSI activation under an existing monopoly: four substitution elasticities and two research productivities. Follow growth, prices, and the distribution of income over the simulated horizon.
The reader also includes the experiments on introducing monopoly and losing AI-dominated growth.
The working paper and supplementary results
Explore the simulations →RSI activation under an existing monopoly
Explore the transition
What changes when AI research is more productive, or when AI and labor are easier to substitute? Compare the eight stored trajectories from the paper’s two RSI-activation exercises.
For the separate changes in market structure, read From competitive provision to monopoly with RSI and When monopoly eliminates AI-dominated growth, including their figures.
Loading the stored simulation data…
Negative years show the pre-RSI balanced-growth reference. RSI is activated unexpectedly at year 0; the stocks of capital and AI efficiency remain continuous, while choices can jump. The years are model time, not calendar forecasts.
The four economies have different initial capital stocks, chosen to place each on its pre-RSI balanced-growth path. They share the other parameters within each exercise. A point at year 500 is a simulated observation, not necessarily a long-run limit.
Parameters, sources, and scope
Open research
Read. Check. Replicate.
The manuscript, numerical methods, saved simulation data, and web-edition source are available in the same repository.
01↗02↗03↗
Source repository
The paper and its computational materials.
Replication guide
Environment setup, commands, numerical checks, and figure generation.
About this digital edition
Sources, conversion checks, and the automatic update workflow.
The explorer displays saved results; it does not solve a new equilibrium in your browser. Research productivity and substitution elasticity can be selected only from the scenarios studied in the paper.