Technological Complexity Based on Japanese Patent Data
Rintaro Karashima, Hiroyasu Inoue · 2025
Read Technological Complexity Based on Japanese Patent Data by Rintaro Karashima, Hiroyasu Inoue (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-11932.txt.
Introduction
Rintaro Karashima and Hiroyasu Inoue measure how complex national technologies are, in a 2025 paper based on Japanese patents. Nations want to find key fields to grow, yet a field's depth is hard to judge from simple counts. Older work rated regions and noted only whether they specialized in a field. This study rates firms instead, giving a finer view.
The authors map a network linking many firms to either 35 or 124 technology classes. Data span fiscal years from 1981 to 2010. Their index captures how rare and how advanced each field is. Results stay stable under the finer split and show clear regional strengths. Firms and planners can use the scores to shape strategy.
Catalog record
| CREATOR | Rintaro Karashima, Hiroyasu Inoue |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-11932
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Technological Complexity Based on Japanese Patent Data Rintaro Karashima, Hiroyasu Inoue Abstract As international competition intensifies in technologies, nations need to identify key technologies to foster innovation. However, the identification is difficult because a technology is independent, and therefore it has complex nature. Here, this study aims to assess the patent technological fields by applying Technological Complexity Index from a corporate perspective, addressing its underutilization in Japan despite its potential. By utilizing carefully processed patent data from fiscal years 1981 to 2010, we analyze the bipartite network which consists of 1,938 corporations and 35 or 124 technological fields. Our findings provide quantitative characteristics of the ubiquity and the sophistication for each patent fields, the detailed technological trends that reflect the social context, and methodological stability for policymakers and researchers, contributing to targeted innovation strategies in Japan. keywords Keyword1, Keyword2, Keyword3 Introduction In the twenty first century, the global landscape of innovation has grown markedly competitive. Nations around the world have designed diverse industrial policies to foster innovation, recognizing it as a vital factor for sustaining economic competitiveness. Despite the abundance of strategies, one persistent difficulty to identify the core technologies and actors that drive innovation remains. The difficulty arises from the fact that innovation is an intangible phenomenon emerging from an interconnected ecosystem composed of countries, cities, corporations, and technologies. Understanding this complex environment requires methodologies capable of pinpointing the central elements that contribute to innovation amid intricate interdependencies. Globally, different regions and institutions have approached this challenge from various angles. In the European Union, the concept of Smart Specialisation [ 1 , 2 ] has helped identify key domains for policy support, while in Japan, science and technology strategies continue to evolve [ 3 ] . For supporting them, a range of indicators have been proposed to assess innovation across manufacturing, services, and advanced materials [ 4 , 5 ] . Some studies focus on either national-level [ 2 , 6 ] , regional-level [ 7 , 8 , 9 , 10 ] or corporate-level analyses [ 11 , 12 , 13 ] . Moreover, recent research has addressed diverse sectors such as Industry 4.0 [ 2 ] , pharmaceutical biotechnology [ 14 , 15 , 16 , 17 ] , and service innovation [ 4 ] , reflecting the breadth of modern innovation ecosystems. Among these approaches, a notable stream of research has emerged from the study of product complexity derived from Hidalgo-Hausumann (HH) algorithm [ 18 ] . The algorithm quantifies sophistication by applying network analysis to trade data and it has been adapted to different contexts such as economic, patents, and technological capabilities [ 19 , 20 , 21 , 22 , 23 ] . The algorithm also has proven instrumental in highlighting technologies and capabilities that hold particular importance. Building on this foundation, Balland and Rigby [ 24 ] developed the Technological Complexity Index (TCI) to measure the complexity of technological fields using patent data. The TCI captures the nuanced interplay between technological ubiquity (i.e., how few players are capable of producing it) and sophistication (i.e., ubiquitous and produced by key players) [ 7 ] . Various extensions and complementary metrics have been proposed, including fitness-based approaches [ 25 , 26 , 27 ] and additional frameworks for classification and measurement [ 28 , 29 , 30 , 31 , 32 ] . More recently, the method has been extended to Asian contexts and corporate-level analyses, supported by the increasing availability of detailed patent databases [ 33 , 34 ] . Despite these advancements, comprehensive applications of these methods in the Japanese context remain limited. In addition, although Japan has experienced prolonged economic stagnation, often described as the “Lost Decades”, it can hardly be denied that Japanese government cannot effectively identify the technologies that will drive renewed innovation and international competitiveness [ 35 ] . Although some studies, such as the work on automotive suppliers [ 13 ] , have employed related approaches, there is a clear lack of a broad analysis that would illuminate the full landscape of Japanese technologies through the lens of complexity. This study addresses the existing gap by applying the HH algorithm to patent data from the Patent Office of a country, which encompasses a diverse range of technological fields. A bipartite network is constructed to link domestic corporations with these technological domains, and a subsequent projection into a monopartite network facilitates the derivation of TCI. Represented by the second largest eigenvector of the network adjacency matrix, TCI enables a simultaneous evaluation of both technological spillovers and scarcity. An analysis of temporal trends over the period 1981 to 2010 demonstrates that the proposed method yields a robust and detailed depiction of the technological landscape, thereby surpassing conventional quantitative measures such as simple patent counts. The assessment reveals detailed aspects of technological capabilities and their evolution over time. This study contributes to a comprehensive understanding of the innovation ecosystem by identifying the central technological fields through the lens of complexity. The insights obtained herein may inform the development of more effective innovation policies and strategies, and they establish a methodological framework that can be extended to comparative studies in other contexts. The rest of this paper is organized as follows. In Results Results section, we present the ubiquity and complexity of technological fields presented by the patent data, followed by a detailed analysis of technological trends. In the Discussion Discussion section, we provide a critical analysis of our findings and discuss the implications of our results. Finally, in the Methods Methods section, we detail the data processing and methods used in our analysis. Results Characteristics of Technologies In this study, we employ the HH algorithm to derive quantitative indicators that characterize each technology (see Methods Methods ). First, we define ubiquity (or degree centrality) K_{T,0} as the number of corporations connected to a technology; a lower value indicates that the technology is more exclusive. Similarly, the average diversity K_{T,1} , which measures the variety of corporations connected to a technology, implies that a higher value reflects diversified publication, whereas a lower value suggests specialization. In many cases, the average of each indicator is used as a threshold to distinguish high from low values [ 18 , 36 ] . Another key indicator is the TCI, which represents the sophistication required to produce a technology. A threshold of zero divides TCI into high and low categories: a high TCI implies that a technology demands a complex combination of unique inputs and advanced technical expertise which explains why only a few advanced corporations can publish it competitively, whereas a low TCI indicates reliance on widely available capabilities, allowing many corporations to participate. To analyze the structure of corporate patent publication as depicted by the bipartite graph, we focus on the relationships among these three indicators. The mapping of technologies is based on the Schmoch system [ 28 ] (Panel A in Fig. 1 ), where K_{T,1} and K_{T,0} are divided by their mean values, and TCI is segmented by a zero threshold. Panel B in Fig. 1 shows that there is no correlation between K_{T,0} and K_{T,1} (Pearson correlation coefficient r=0.064 ), suggesting that technologies widely held by corporations do not necessarily correspond