Do AI development zones improve firm-level innovation efficiency? Evidence from China (with Xiaoying Yao)
Do government efforts to foster artificial intelligence translate into more efficient innovation by firms? This paper investigates the impact of China's AI Development Zone program on the innovation efficiency of strategic emerging industry (SEI) firms, paying particular attention to both proportional and distributional effects. Drawing on a panel of 1,009 A-share listed firms over the period 2013-2024, we employ the staggered difference-in-differences estimator of Callaway and Sant'Anna (2021) in conjunction with stochastic frontier analysis. Our results show that AI Development Zone designation increases local firms' innovation efficiency by approximately 6%. These gains arise immediately following designation, persist over time, and remain robust to alternative specifications and quality-adjusted measures of innovation efficiency. We further document a reduction in within-city efficiency dispersion, driven primarily by convergence among initially less efficient firms rather than improvements at the technological frontier. The effects are largely homogeneous across firm, industry, and city characteristics, indicating that the policy's benefits are broadly shared and do not disproportionately accrue to particular groups of firms.
Intellectual Property Rights and the Direction of Creative Activity: Evidence from Renaissance Venice (with Stefano Comino and Clara Graziano)
Intellectual property rights (IPRs) are commonly viewed as a tool to stimulate innovation and creativity, yet comparatively less is known about how their design affects the direction of creative activity. We study this question using historical data from the sixteenth-century Venetian printing industry and exploiting a major institutional reform introduced in 1517. Before the reform, printing privileges could be granted without a systematic assessment of the novelty of the protected work, allowing publishers to obtain exclusive rights over previously published titles. The 1517 reform introduced novelty as a requirement for obtaining protection. We show that, following the reform, book production in Venice increased relative to other Italian cities, with the increase being particularly pronounced for works by contemporary authors. We rationalize these findings through a simple theoretical model in which weak requirements for obtaining intellectual property protection distort the allocation of creative effort toward the exploitation of existing knowledge and content. By strengthening novelty requirements, IPR policy can instead redirect creative activity toward exploration strategies, more valuable in changing/unstable environments.
The effects of taxation on differentiated products markets (with Anna Rita Bennato) Available at CCP 23-06.
In this study, we analyse the impact of differentiated and homogeneous ad valorem tax rates on two product qualities within imperfectly competitive markets. Our focus is on the tax's application to product quality and its implications for prices and welfare. We find that with a symmetric tax rate increase, net prices diverge when the tax is quality-differentiated and converge when it is uniform across qualities. Competition mitigates these effects, promoting convergence under differentiation and divergence under homogeneity. A differentiated tax system generates higher tax revenues and producer rents, but reduces consumer surplus, highlighting the trade-offs inherent in tax policy.
Unintended distributional impacts: Concentration, and competition policy (with Steve W. Davies and Junjun Zhang) Original short paper CCP 22-01.
This paper contributes to the empirical landscape regarding the distributional impact of competition and competition policy. Employing a unique blend of data from the national survey of household expenditure and the UK business structure database, it uncovers three noteworthy findings. Firstly, it reveals that individuals with lower income levels tend to rely more heavily on product purchases and services from markets characterised by higher concentration, compared to their wealthier counterparts. Secondly, it identifies a significant negative correlation across products and services, linking the income elasticity of demand to the concentration of the industries that supply them. Thirdly, the paper sheds light on the impact of regulation and competition enforcement by competition agencies. Specifically, it demonstrates that regulated sectors or markets subject to competition enforcement measures have experienced reduced concentration, particularly benefiting consumers with lower incomes.
Cleaning up? The effect of quality contribution in Amazon’s attempted iRobot acquisition (with Junjun Zhang)
Amazon’s proposed acquisition of iRobot is analysed using an intuitive model of platform-driven quality allocation, demonstrating that platform ownership can amplify quality asymmetries - especially after acquiring a high-quality seller. Structural estimates based on US robotic vacuum data reveal that even modest, algorithmically driven boosts to iRobot’s ratings post-merger can substantially increase both firm and platform profits, as well as shift market shares. These findings underscore the competitive importance of algorithmic quality contributions and indicate that merger policy should also account these non-price effects in digital markets.
Non-price Effects of M&As in the App Market (with Michael Kummer and Junjun Zhang)
Using data from 8 million apps and 1.8 million developers on the USA Google Play Store (2015–2019), this study investigates the non-price effects of 1,387 mergers and acquisitions (M&As) involving US app developers from 2016 to 2017, addressing the growing interest in digital markets. The analysis shows that M&As reduce new app launches and purchases while also limiting app removals and sales, leading to a net decline in app variety. However, M&As increase app updates, indicating a strategic focus on enhancing the quality of existing products. Acquisitions drive a sharper decline in new launches compared to mergers, which show minimal impact. Horizontal M&As reduce app removals and sales, while non-horizontal M&As restrict market expansion by curbing new and purchased apps. These findings highlight distinct strategic differences between horizontal and non-horizontal M&As, underscoring the role of app substitutability and complementarity in shaping post-M&A outcomes.