The economics of technology adoption

Electric motors were available to factory owners by the 1880s. Yet for roughly four decades, the productivity of American manufacturing barely moved. The economic historian Paul David showed why: factories were built around a single steam engine turning a central shaft, and simply swapping in an electric motor changed little. The gains arrived only when owners redesigned the whole factory, giving each machine its own motor and rearranging the floor around the flow of work.

The story captures the central lesson of the economics of adoption. A technology is not a product but a bundle of costs, risks and changes that someone must decide to take on. Understanding that decision explains why good technologies spread slowly, why some people rationally refuse them, and what governments and firms can do about it.

Adoption as an investment

Economists treat adoption as an investment. A farmer, a firm or a household compares the expected benefit with the full cost, which includes the price, the effort of learning, the disruption of changing routines and the risk that it will not work. They also weigh the value of waiting: if a technology is likely to get cheaper or better, or if others’ experience will reveal whether it works, holding off can be the smarter move.

This framing overturns a common assumption, that people who do not adopt are ignorant or resistant. In a study of hybrid maize in Kenya, the economist Tavneet Suri found that returns and costs vary widely between farmers. Some gain a great deal, while others, because of poor access to markets or unsuitable soils, gain little. Farmers who stayed away were often making a sensible choice. The same technology can be a bargain for one user and a waste of money for the next.

Five forces that shape adoption

Force	How it works	Example
Information and learning	People adopt as they see others succeed and learn how to use it	Farmers in India copied neighbours who had made high-yield seeds work
Network effects	A technology is worth more as more people use it	Messaging, payment and charging networks
Complementary investment	The gain depends on changing other things too	Redesigning factories for electric motors
Finance and risk	High upfront costs and uncertain returns deter adopters	Solar panels, machinery, irrigation
Behavioural frictions	Present bias and procrastination delay worthwhile choices	Fertiliser use in Kenya

Information. Foster and Rosenzweig studied Indian farmers adopting high-yield seeds and found that farmers became more profitable both from their own experience and from watching their neighbours. Adoption is a social process of learning, which is why early adopters matter so much and why a visible failure can set a technology back.

Network effects. Some technologies are valuable mainly because others use them. Katz and Shapiro showed how this creates tipping points and a risk of lock-in: users hesitate to commit to a network that may not win, and a technology with a head start can dominate even if a rival is better. This is why standards and compatibility matter so much.

Complementary investment. As David’s dynamo showed, the full gains often require reorganising work, retraining people and building new supporting infrastructure. The same pattern appears with computers and, more recently, with artificial intelligence: the technology arrives first, and the productivity gains follow when organisations learn to use it.

Finance and risk. Many beneficial technologies, from irrigation pumps to solar panels, cost a lot upfront and pay back slowly. Credit-constrained and risk-averse households may rationally pass them by. This links directly to the cost of capital discussed in the climate finance debate.

Behavioural frictions. Even when a technology pays, people may delay. Duflo, Kremer and Robinson found that Kenyan farmers often failed to buy fertiliser when they had the cash, but that a small, time-limited discount offered right after harvest raised use substantially. People plan to adopt later and then do not.

A technology is not adopted when it is invented. It is adopted when its benefits outweigh its costs for the person deciding.

Why the same technology spreads at different speeds

The same forces explain why a technology can take off in one place and stall in another. Take a rooftop solar system. Where electricity is expensive and unreliable, the benefit is large. Where borrowing is cheap, the upfront cost is easy to bear. Where installers are plentiful and neighbours already have panels, information is abundant and the risk is low. A country or community with all three can move quickly, while one lacking them can stay far behind, even though the panels are identical.

This is also why average figures can mislead. A technology may be available everywhere yet used intensively by only a small part of the population. The economists Diego Comin and Martí Mestieri have argued that differences in how intensively technologies are used, rather than whether they have arrived, explain much of the gap between rich and poor countries. Closing that gap means working on the decision, on its costs, its risks and its complements, rather than just on supply.

What speeds adoption up

Because adoption is a decision, it can be influenced at each point.

• Cut the cost and the risk. Subsidies, loans paid back in instalments, guarantees and trial periods reduce the stakes. Pay-as-you-go models, such as for solar systems, break a large purchase into small payments.
• Show it working. Demonstration plots, pilot projects and trusted local users shorten the learning period. Visible success can bring forward the steep part of the adoption curve.
• Build standards and compatibility. Shared standards ease the fear of choosing the losing side and make networks larger sooner.
• Support the complements. Training, repair services, supply chains and infrastructure determine whether the technology delivers. Without them, adoption can stall or disappoint.
• Design for the way people behave. Timing offers to coincide with when people have money, making the technology the default, and reducing the effort of signing up all help.

What it implies

Understanding adoption changes how we judge slow uptake. Where a technology spreads slowly, the cause is often not resistance but a rational response to high costs, uncertain returns or missing complements. The remedy is to change the conditions, not to blame the users.

It also shows why invention is only the first step. As argued in the diffusion and resilience article, the gains from technology depend on how widely and deeply it is used. Policymakers and firms that want faster adoption should ask who is deciding, what they stand to gain and lose, and what is standing in their way.