The Productivity J-Curve: How Intangibles Complement General Purpose Technologies

Erik Brynjolfsson, Daniel Rock, Chad Syverson · American Economic Journal Macroeconomics · 30 Dec 2020

Where this sits in the debate

Will AI actually boost productivity and growth? · its finding: it depends (why studies differ)
How fast will the impact arrive? · its finding goes against the claim (why studies differ)

Direction of the finding as read from the abstract by an AI model (AI-written).

Key takeaway AI-written

Argues that general-purpose technologies like AI require large, often-intangible complementary investments that official statistics undercount, producing a 'Productivity J-curve': measured productivity growth is understated early in a technology's diffusion and overstated later as those investments pay off. Applying the model to U.S. data on computer hardware and software, the authors estimate 2017 total factor productivity was 15.9% higher than official figures once intangible capital is included — a key reference point for economists arguing today's modest measured AI productivity gains understate its eventual impact.

The paper in three steps AI-written

Written by an AI agent from the paper's abstract, which is the primary source and is quoted below.

The question

Do official statistics mismeasure productivity growth while a general-purpose technology like AI is being adopted?

The approach

The authors build a model in which general-purpose technologies require large, poorly measured intangible investments, and apply it to US data on intangibles related to computer hardware and software.

The findings

The model implies productivity growth is underestimated in a new technology's early years and overestimated later, when the intangible investments pay off: a Productivity J-curve. Adjusting for hardware and software intangibles gives a US TFP level 15.9% above official measures by the end of 2017.

How to read it: theory. A theoretical model — its value is the mechanism it isolates; ask whether the assumptions fit the real economy.

Limitations and counterpoints AI-written

The empirical application covers computer hardware and software intangibles up to 2017, not AI specifically, and depends on how the unmeasured intangible investment is estimated.

Abstract

General purpose technologies (GPTs) like AI enable and require significant complementary investments. These investments are often intangible and poorly measured in national accounts. We develop a model that shows how this can lead to underestimation of productivity growth in a new GPTs early years and, later, when the benefits of intangible investments are harvested, productivity growth overestimation. We call this phenomenon the Productivity J-curve. We apply our method to US data and find that adjusting for intangibles related to computer hardware and software yields a TFP level that is 15.9 percent higher than official measures by the end of 2017. (JEL E22, E23, G31, L63, L86)

About this publication

Academic journal. Published in an academic journal — typically peer-reviewed before publication, though indexing alone doesn't guarantee review depth.

Published: 2020 (71 months ago)
Citations: 584 (OpenAlex)

584 citations recorded by OpenAlex — too few 2020 papers in this collection for a fair comparison.

□ added to this collection in 2026.

Chad Syverson, Daniel Rock, Erik Brynjolfsson signed the wemustactnow.ai letter · author affiliation matches seed institutions stanford, stanford university, university of chicago, university of pennsylvania — part of why this paper is in the collection.

Cited elsewhere in this guide

State of the debate: How fast will the impact arrive?