Industrial policy will play a key role in navigating the energy transformation, and the global automotive industry is already adapting. This column examines the impact of industrial policies on electric vehicle innovation. Using both country- and firm-level records, the authors compile a database of global industrial policies and patents for the automobile industry, highlighting the targeted policies that increased electric vehicle patent applications over a five-year period. At the same time, the data confirm the importance of path dependence in the direction of technology change in the automobile industry.
Industrial policies have garnered renewed interest among major economies (Millot and Rawdanowicz 2024) and have been widely adopted to promote government targets in energy transformation and navigate challenges such as climate change. One prominent example is electrifying the transportation sector, widely recognised as a critical strategy for reducing carbon emissions. Countries including the US and China have set ambitious goals for electric vehicle (EV) adoption, targeting 50% and 40% respectively of total vehicle sales to be EVs by 2030. To achieve these objectives, governments have implemented various industrial policies such as purchase incentives to stimulate the growth of EV and EV battery industries.
Policymakers and economists are showing increased interest in measuring the impact of industrial policies and understanding their role in fostering industrial transformation (Juhász et al. 2023). In our recent paper (Barwick et al. 2024), we add to the ongoing discussion from the perspective of technology innovation by examining industrial policies in the electrification of the global automobile industry.
We compile a comprehensive database of industrial policies and patents for the global automobile industry spanning the period 2008 to 2023. To quantitatively measure innovation, we use the European Patent Office (EPO)’s PATSTAT global patent filing data from 1980 to 2023, and classify patent applications’ fuel types into electric vehicle (EV), gasoline vehicle (GV), and general. We overcome measurement difficulties with regard to industrial policies by using two distinct approaches. First, we apply the text-based approach to the Global Trade Alert (GTA) data following Juhász et al. (2022) to quantify the number of industrial policies at the country level. Second, we compile country-by-product level EV subsidy data and construct EV subsidy exposure at the firm level for both automakers and battery cell suppliers.
Trend of industrial polices in the global automobile industry
We first use our compiled dataset to document key data patterns of industrial policies and innovation in the global automobile industry. Figure 1 illustrates a prominent trend in new automobile-related industrial policies from 2008 to 2022. Following the 2008 global financial crisis, the number of IPs increased steadily from fewer than 100 in 2009 to over 300 by 2022. Meanwhile, the share of EV-targeted industrial policies jumped from nearly zero in 2008 to nearly 50% by 2022. This upward trend reflects the global shift toward electrifying the transportation sector and underscores the increasing emphasis on supporting EVs in recent policy initiatives.
Meanwhile, our data show that developed countries dominated industrial policy implementation post-2008, with the US leading in the number of policies. Emerging economies such as Russia, Brazil, India, and China have also implemented a considerable number of industrial policies in the automobile industry, but the industrial policies in these countries appear less focused on EVs, possibly reflecting a continued emphasis on labour-intensive advantages in traditional internal combustion engine manufacturing.
Figure 1 The number of new industrial policies in the automobile sector over time
Note: Bars (left axis) show the number of new industrial policies in the automobile sector over time; dashed line (right axis) depicts the share of EV-targeted industrial policies. An industrial policy is defined as EV-targeted if 50% of its affected country-products are EV-related. Note that EV-targeted IPs have a much smaller number of affected countries and products compared with conventional industrial policies, as EV is an emerging industry.
Source: Barwick et al. (2024) using data from Global Trade Alert Database.
Transition to green technology
We also reveal a clear technological shift from GV technologies to EV technologies, especially after 2000. Newly granted inventions were predominantly focused on GV technologies before 2010, with rapid growth occurring from the 1990s through 2010. The number of GV-exclusive inventions peaked in 2011 and has since gradually declined. In contrast, EV-exclusive inventions worldwide have exhibited exponential growth since the mid-2000s, surpassing GV inventions in 2013. By 2020, the number of EV inventions had grown to three times that of GV inventions. General inventions applicable to both EVs and GVs have grown steadily over the past four decades, though they appear to have peaked in 2016 and have since trended downward. These patterns align with the growing number of EV-related policies in the last decade.
Figure 2 Global trend of newly granted automobile industry inventions
Note: Figure shows the number of automobile-related inventions from 1980 to 2018. One invention is identified as a unique technological innovation, which can be filed at multiple patent offices for multiple product classification codes. Solid green line (left axis) represents inventions that are exclusively used for EVs; blue dashed line (left axis) represents inventions that are exclusively used for GVs; red dash-dot line (right axis) represents general inventions that can be used for both EVs and GVs.
Source: Barwick et al. (2024) using data from EPO’s PATSTAT database.
Evidence on the effects of industrial polices on innovation
Our empirical analysis evaluates the impact of industrial policies on patent applications and grants within the global automobile industry, with a focus on EV-related industrial policies and EV-related patents that capture innovations in green technologies
We explore the relationship between country-level patent numbers and the cumulative number of past IPs implemented in the countries. We treat the lagged number of industrial policies as exogenous to EV innovation after controlling for the impact of EV patent stocks and other environmental regulations.
Figure 3 Empirical relationship between patents applied against cumulative EV industrial policies
Note: Figure shows a (residualised) relationship between lagged cumulative EV IPs (x-axis) and the number of new EV/GV patents applied at the country-IPC-year level (y-axis). Country is the country of residence for inventors, showing the effects of EV industrial policies on patents applied using the ‘binscatter’ methods proposed by Cattaneo et al. (2024). Green circles and solid fitted line stand for EV patent; blue diamond dots and dash fitted line stand for GV patents.
Source: Barwick et al. (2024)’s estimation using complied IPs and patents database.
Figure 3 graphically depicts two patterns related to the relationship. First, there is a clustering of data points near the origin, reflecting that a large proportion of the sample (country) has no new patent applications. Second, there is a clear positive relationship between EV patents and cumulative EV industrial policies, while no such relationship is observed for GV patents and EV industrial policies. This suggests that EV industrial policies incentivise research and innovation in clean vehicle technologies without spillover effects on non-targeted GV technologies. Our estimation shows that a one-standard-deviation increase in five-year cumulative industrial policies targeting the EV industry is associated with a 4% increase in the number of new EV patents filed. Trade and subsidy-related industrial policies exhibit stronger correlations with innovation than other types of industrial policies.
The firm-level analysis that more carefully addresses the potential reverse causality concerns by exploiting variation in subsidy exposure from demand shock shows a similarly positive impact of industrial policies on innovation.
Another important finding is that we confirm strong path dependence and knowledge spillovers in automotive innovation (Aghion et al. 2016). Specifically, knowledge stocks related to the same fuel type facilitate the generation of new patents while knowledge stocks of different fuel types hinder this process. In other words, firms with more extensive experience in EV-related technologies tend to innovate more rapidly and produce new EV patents faster. One implication of this finding is that the positive effects of industrial policies are self-reinforcing due to path dependence, suggesting that the marginal return of industrial policies in promoting EV innovation will be larger in the long run. Therefore, when assessing the costs and benefits of industrial policies, it is crucial to adopt a long-term perspective.
Our findings highlight the crucial influence of industrial policy in energy transitions. Current discussions and policy discourse often focus on the direct impact of industrial policy on the demand side, such as how purchase subsidies and charging infrastructure affect electric vehicle adoption. We demonstrate that there is a significant new pathway playing a vital role, specifically by directing supply-side innovation efforts and advancing technological frontiers. The insights gained from the global automotive industry are also applicable to other emerging and key sectors such as solar panels, renewable energy, and the semiconductor industry.
Source : VOXeu
