AI, Productivity and Firm Resilience: Evidence from Indonesia
This study analyzes Indonesian manufacturing firms to reveal that AI adoption significantly boosts productivity, particularly for medium-sized firms in technology-intensive sectors and those recovering from economic shocks, while demonstrating diminishing marginal returns as firms increase in size and input intensity.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the modern economy, factories are the engines that drive nations forward, turning raw materials into the goods that sustain daily life. For decades, economists have watched closely as these engines evolved, seeking to understand what makes them run faster and more efficiently. A central question in this field is how new technologies change the way companies work. When a factory adopts a new machine or a smarter software system, does it simply produce more, or does it fundamentally change the nature of the work itself? This question becomes even more urgent when looking at developing nations, where the path to prosperity often depends on how quickly local businesses can upgrade their tools. In recent years, a specific type of technology known as artificial intelligence has emerged as a powerful force. Unlike simple machines that repeat a single task, this technology can learn, adapt, and make decisions, promising to reshape productivity across the globe. Yet, while much is known about how these tools work in wealthy countries, there is a gap in our understanding of how they function in places like Indonesia, where labor costs are lower and digital systems are still maturing.
A new study by Al Bahits Annef from Nagoya University steps into this gap to examine exactly how artificial intelligence affects the performance of manufacturing firms in Indonesia. The researcher analyzed data from thousands of factories over a two-year period that coincided with the global pandemic, a time when economic activity was severely disrupted. By comparing companies that had adopted these intelligent systems with those that had not, the study reveals a clear and significant benefit: factories using artificial intelligence were substantially more productive. Specifically, the research found that adopters produced between 17 and 32 percent more output per unit of input than their non-adopting counterparts. This gap was not a fluke; it held true even when the researcher accounted for differences in company size, capital, and the number of workers, suggesting that the technology itself was the driver of the improvement.
However, the story is not as simple as "more technology equals more success for everyone." The study uncovers a surprising twist: the benefits of artificial intelligence are not evenly distributed. In fact, the technology appears to work best for medium-sized companies, particularly those in industries that are already technologically advanced. For these firms, adopting artificial intelligence acted as a powerful equalizer, allowing them to close the gap with larger competitors. In contrast, very large firms in traditional sectors saw smaller gains, and the study suggests that the more capital and labor a company already uses, the less additional benefit it gets from adding artificial intelligence. This pattern indicates that the technology does not simply reinforce the dominance of the biggest players; instead, it offers a unique advantage to agile, mid-sized firms that can integrate these tools without the heavy coordination costs that often slow down massive organizations.
The timing of this research also provides a rare glimpse into how technology helps businesses survive crises. The data covered the years 2020 and 2021, a period when the pandemic caused widespread shutdowns and supply chain chaos. The findings show that artificial intelligence did not prevent factories from feeling the initial shock of the crisis in 2020; productivity still dipped as operations were disrupted. However, by 2021, as the economy began to recover, the factories that had adopted the technology bounced back much faster and stronger than those that had not. This suggests that while the technology cannot stop a storm from hitting, it acts as a crucial tool for rebuilding afterward, allowing firms to reorganize their operations and overcome bottlenecks like labor shortages or supply delays.
These results challenge the common assumption that only the largest, most established companies are best positioned to benefit from high-tech upgrades. Instead, the evidence from Indonesia suggests that artificial intelligence is a flexible tool that can boost efficiency for a wide range of businesses, provided they have the right size and context to use it effectively. For policymakers and business leaders in developing economies, the lesson is clear: supporting the adoption of these technologies in medium-sized firms could yield greater overall economic growth than focusing solely on the largest industrial giants. The study confirms that while the path to digital transformation is complex, the payoff in terms of resilience and productivity is real, offering a tangible way for emerging markets to strengthen their economic foundations in an uncertain world.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.