Draft:Uneven Growth by Automation's Impact Theory

Uneven Growth by Automation's Impact Theory refers to a set of interconnected economic ideas and empirical findings developed to explain how recent waves of automation—particularly in industrial robots and intelligent software even at private level—have led to divergent effects on employment, wage distribution, productivity, and inequality. The theory is not related to accelerationism, though it expects non-linear increase in innovation. It is inheritantly related to other Technological unemployment analyses, though it follows strict econometrics estimates. Per-se, it does not criticize the growing use in society of intelligent machines, as they are part of a liberal view of progress, though it simulates and estimates the inequalities and suggests aspects to be considered in future policies. The theory was introduced by the economists Pascual Restrepo and Daron Acemoglu, and it has been introduced in several key publications in economic journals between 2018 and 2020.

Overview

The theory suggests that automation technologies, while potentially boosting productivity and economic growth, often displace human labor in specific tasks, contributing to wage inequality and structural unemployment. This process produces uneven growth: while some sectors, regions, and workers benefit, others experience technological unemployement and economic displacement.

In the theory, automation is seen within a "task-based" model rather than being primarily labor-augmenting or capital-enhancing. It eliminates the need for human labor in certain jobs, particularly those that are routine or codifiable, while occasionally inventing new tasks where people still have a comparative advantage. In the short to medium term, however, the replacement frequently surpasses the creation.

Key Theoretical Foundations

The framework was formally introduced in The Race Between Man and Machine: Implications of Technology for Growth, Factor Shares and Unemployment. This paper quantifies through simulations defined by mathematical models, how new automation technologies can displace labor and how new task creation can offset this effect.

The simulations at the core of the theory rely on the estimate of the production function, which is expressed as:​

Y = Π(I,N)·[Γ(I,N)·(A_L L)^(σ−1)/σ + (1−Γ(I,N))·(A_K K)^(σ−1)/σ ]^(σ/(σ−1))

Where:

• Y: Total output
• Π(I, N): Total factor productivity, depending on automation level I and range of tasks N
• Γ(I, N): Share of tasks performed by labor (increasing in N, decreasing in I)
• A_L, A_K: Productivity of labor and capital, respectively
• L, K: Labor and capital inputs
• σ: Elasticity of substitution between labor and capital

This equation models the reallocation of tasks between capital and labor as automation progresses and new tasks have been introduced.

Building upon this work, Automation and New Tasks: How Technology Displaces and Reinstates Labor elaborates the idea of a dynamic cycle in which technological progress both destroys and creates tasks for labor. They discover, however, that job creation is slower and less reliable, whereas task removal is typically more quick and predictable..

Empirical Evidence

The paper Robots and Jobs: Evidence from US Local Labor Markets offers direct empirical validation with retrospective data. In this work, it has been estimated that each additional robot per 1,000 workers decreased the employment-to-population ratio by roughly 0.2 percentage points and earnings by roughly 0.42 percent in the impacted local labor markets, according to a study that examined the effects of growing industrial robot adoption in the United States between 1990 and 2007.

These findings are further supported in subsequent analysis, where the observation was that high-skilled workers in non-routine cognitive functions have seen increases in their wages, while middle-skilled individuals performing repetitive tasks have witnessed disproportionate income reductions.

Lastly, using a macroeconomic model to quantify the distributional consequences of automation across different income and wealth groups, it was possible to quantify that unless automation is supported by robust redistributive laws or institutional frameworks, it may worsen already-existing disparities.

Policy Implications

The theory has also been applied in the context of public policies, arguing that the U.S. tax system inadvertently encourages automation over human labor by allowing capital deductions and providing fewer incentives to retain workers. Indeed, taking all those perspective evaluations, it is possible to suggest that correcting these biases could reduce automation-induced inequality and improve employment outcomes, and indeed this theory should be taken into account by policy makers as suggested by the authors.

References

This page will be placed in the following categories if it is moved to the article namespace.

Categories:

• Distribution of wealth

This page will be placed in the following categories if it is moved to the article namespace.

Categories:

• Economic inequality
• Libertarian theory

This page will be placed in the following categories if it is moved to the article namespace.

Categories:

• Social inequality