When policy changes the direction of innovation
Danyang He’s PhD research shows that government efforts to shape digital platforms – through regulation, guidance, standards and economic support – do not necessarily reduce innovation. Instead, they can influence whether companies pursue risky breakthroughs or gradual improvements.
Having recently defended her thesis and earned PhD degrees from the University of Chinese Academy of Sciences and Copenhagen Business School, Danyang He has completed a project centred on a key question: what happens to innovation when governments become more involved in shaping digital platforms?
‘I wanted to understand how the relationship changed as platforms became more important – and what those changes meant for innovation,’ says Danyang He.
Her research follows Chinese policy from 2000 to 2025, as digital platforms became central to payments, shopping, communication and data use. Government priorities also broadened from helping digital markets grow to addressing competition, privacy, security and platform responsibility.
Danyang He grew up in Ningxia in north-western China and studied communication at Beijing Foreign Studies University before completing a master’s degree in strategic marketing at Imperial College London. She then began her PhD within the SDC Social Sciences research theme.
Jiale Sheng, Stine Haakonsson, Danyang He, Jing Jillian Wang
Following 25 years of policy
Research on regulation and innovation often begins with a single law and compares what happened before and after it. Danyang He took a wider view, examining how different forms of government action accumulated and changed over time.
The study combines 239 national policy documents, 674 provincial documents and digital patent data. She read key policies and Five-Year Plans closely and used text analysis and econometric methods to identify patterns across the larger collection.
To compare one year with another, she measured what she calls policy intensity. This reflects how far-reaching, detailed and consequential government action was – not simply how many documents were published.
‘Important policies do not appear in isolation. They are introduced for a reason, and other measures often follow. Looking across the whole period allowed me to see how government involvement developed from year to year,’ says Danyang He.
Stronger policy activity was generally followed by increases in patent applications one year later and patent grants after two years. This does not show that every policy produced new inventions, but it suggests that government action may shape expectations before the results become visible in patent data.
The surprising impact of soft instruments
Danyang He divided government action into binding rules, financial measures such as subsidies and tax incentives, and soft instruments – measures that guide companies and signal the government’s priorities.
These could include development plans pointing to priority technologies, recommendations on platform compliance, data governance and how algorithms are used to rank or recommend content, normative appeals, voluntary or contractual agreements or consultations and training that help companies interpret new requirements. Unlike regulations, they do not compel companies to take a particular action.
The result surprised Danyang He. She had expected direct financial support to show a clear effect on innovation. Instead, soft instruments had the clearest and most consistent positive relationship with innovation.
‘I expected economic instruments to matter because they provide companies with funding and resources. Instead, they showed no detectable average effect in this part of the analysis, while soft instruments had the clearest and most consistent positive relationship with innovation.
Their signals can reduce uncertainty by showing companies which technologies and practices are likely to receive continued political attention. I had not expected that,’ says Danyang He.
Soft instruments and regulation could also reinforce each other. Rules defined what companies had to do, while guidance helped them understand how to respond. But when both were used very intensively, companies could face overlapping or even contradictory signals and more compliance work without a similar increase in innovation.
SDC Social Science symposium in North Zealand August 15-16 2024
Innovation may continue – but change course
The most important finding concerns something total patent numbers cannot show. A more intensive policy environment was not necessarily linked to less innovation overall. Instead, it was associated with a different balance between types of innovation.
Companies appeared to move away from uncertain projects with the potential to produce major technological leaps and towards improving existing technologies. The shift was clearest in industries where technological development was particularly uncertain.
A policy can therefore appear successful if companies continue to produce patents, while beneath the total they are choosing safer and more predictable projects.
‘We should not evaluate policy only by asking whether innovation goes up or down. We also need to ask what kind of innovation companies are encouraged to pursue,’ says Danyang He.
SDC Social Science symposium in North Zealand August 15-16 2024
Two research cultures
Working across China and Denmark also changed the way Danyang He approached the project. She encountered a strong emphasis on quantitative methods and practical implications in China. At Copenhagen Business School, she was pushed to explain the theoretical choices behind the analysis and make greater use of qualitative approaches.
‘At the beginning, combining the approaches was difficult. By the end, I could see that they helped me understand different parts of the same problem,’ says Danyang He.
Danyang He is already extending the work in a new research project. She is examining why financial instruments produced no detectable average effect in her PhD analysis and whether their influence changes at different stages of policy development. Longer term, she hopes to compare how digital platforms are governed across different national settings. Completing the double PhD has also given her a stronger sense of confidence after a demanding process.
‘After the defence, many people congratulated me and said that completing one PhD is difficult, but I had completed two. I feel very proud of myself, and it has made me more confident. I faced many challenges during the process, but I made it,’ says Danyang He.
Adam Frost, Tara Qian Sun, Wang Haiyan, Zhao Hong, Danyang He