China\'s rising IQ (Innovation Quotient) and growth: Firm-level

China's rising IQ (Innovation Quotient) and growth: Firm-level evidence

Jing Fang a, Hui He b c, Nan Li c

https://www.sciencedirect.com/science/article/abs/pii/S030438782030136X?

Highlights

• Build a dataset uniting Chinese firm balance sheet information with firm patent data for 1998–2007.

• Both within-firm increases in patent stock and initial patenting event are associated with increases in productivity.•

• Patent subsidy programs sequentially implemented across Chinese provinces weaken this positive association over time.•

• State-owned enterprises have higher productivity-patenting elasticities than private firms, especially after the SOE reforms.

• There is an inverted U-Shaped relationship between the quality of local IPR protection and innovation.

Abstract

This paper examines whether the rapidly growing firm patenting activity in China is associated with productivity growth, and whether this association changes after the introduction of patent subsidy programs and differs across ownership types. We first build a unique dataset uniting detailed firm balance sheet information with firm patent data for the period 1998–2007. We find strong evidence that both within-firm increases in patent stock and initial patenting event are associated with increases in total factor productivity. Patent subsidy programs sequentially implemented across Chinese provinces weaken this positive association over time. State-owned enterprises (SOEs) have higher productivity-patenting elasticities than private firms, especially after the SOE reforms. Exploring regional variations in the perceived quality of intellectual property rights (IPR) protection shows an inverted U-Shaped relationship between IPR protection and innovation, consistent with the literature.

Introduction

The last two decades have witnessed astonishing growth in China's innovation input and output. R&D spending increased by 22 percent per year during 1998–2013, reaching 190 billion USD in 2013. By 2014, China's R&D expenditure, as a ratio of GDP, had exceeded the OECD average, although its GDP per capita was just one-fifth of the average OECD economy (Wei et al., 2017). Innovation output, measured by the number of domestic applications for invention patents, grew by almost 23-fold from 1998 to 2013, surpassing Japan and the United States in 2011 (Fig. I). Accompanying the rising innovation performance is China's spectacular growth, which is often attributed to its productivity improvement. Zhu (2012), for example, finds that total factor productivity growth has contributed to about 80 percent of China's per capita GDP growth for the 1978–2007 period.

The soaring number of patents held by Chinese firms (such as Huawei and Lenovo), the rapid accumulation of R&D stocks, and the success of large internet and telecom companies (such as Alibaba and Tencent) have led some to conclude that China has leaped into the world innovation frontier. Skeptics, on the other hand, contend that China's prolific patent filings are simply a response to the government-set target and various subsidies. With weak intellectual property rights (IPR) protection (Zhao, 2006), and under the dominance of inefficient and uninnovative SOEs, many question if the rapid upsurge in patenting is “fiction” (Economist, Dec 11th, 2014) or represents real changes in the technological capability and competitive edge of Chinese firms.

While a number of recent studies have strived to unravel what is behind China's explosive growth in R&D and patents (Hu and Jefferson, 2009; Li, 2012; Hu et al., 2017; Chen et al., 2018), this paper poses different questions: Is China's patent explosion associated with real productivity growth? And does the association differ across ownership types and change after the introduction of patent subsidy programs? Answers to these questions are of both policy and economic interests, as they provide valuable insight into China's transition from an investment-led to a more innovation-based growth model (Zilibotti, 2017) and the impact of past patent subsidy programs on the “quality” of patents—where quality is evaluated by their relationship with productivity. In addition, as patent statistics are often used as indicators of innovation and R&D successes in developed economies (e.g., Griliches, 1981; Jaffe, 1986; Hall et al., 2001; Bloom and Van Reenen, 2002; Hall et al., 2005; Aghion et al., 2005; Balasubramanian and Sivadasan, 2011), our work also informs future studies using Chinese patent statistics whether they are able to meaningfully capture firm's innovative and economic activity and how they are compared with evidence in advanced economies.

We address these questions by presenting evidence at the firm level using a unified dataset merging firm patenting data obtained from China's National Bureau of Statistics (NBS) with firm production and balance sheet data from Annual Surveys of Industrial Enterprises in China (ASIEC) which include all “above-scale” firms. Our main results are that within-firm changes in the patent stock of innovating firms (intensive margin) and firms' initial patent applications (extensive margin) are both significantly and positively associated with Total Factor Productivity (TFP) improvement in China. The TFP-patent elasticity, on average, is even higher than that observed in the U.S. This positive association, however, weakens over time throughout our sample period of 1998–2007, which could be related to the patent subsidy programs that were launched sequentially across different provinces and municipalities in China. The improvement in firms' productivity associated with a given change in patent stock drops when these government-led incentives for patenting are in place. Somewhat surprisingly, compared to their privately-owned peers, changes in patent stocks of state-owned enterprises (SOEs) are associated with more improvement in measured productivity, even though they produce less patent per yuan spent in R&D. This elasticity gap between ownership types is especially significant after 2001 and increases over time towards the end of the sample. Our followup analyses seem to suggest that it is not because SOEs enjoy better IPR protection, or receive more subsides from the government, or are less financially constrained, or benefit more from China's accession to WTO. Since the divergence between the TFP-patent elasticity for SOEs and that for POEs emerges after SOE reforms, it suggests that SOE reforms might have contributed to the above observation.

Our analysis proceeds as follows. First, using firm names, we develop annual links between patent applicants and firms included in the ASIEC. The matched data cover more than 1/4 million firms and almost 1.5 million firm-year observations from 1998 to 2007, representing the majority of nonindividual, nonresearch-institution, and nongovernment patentees during this period. Using this unified database, we first document some stylized facts about Chinese firms’ patenting behavior. In line with observations in developed economies, we find that the distribution of patent activities across Chinese firms is highly skewed. Only 9 percent of all firms in the merged sample applied for patents, accounting for 38 percent of value added, 42 percent of capital stock, and 27 percent of employment. Among these patent-filing firms, 6 percent engage in innovation in multiple four-digit industries, accounting for 91 percent of overall patents. Patenting firms are, in general, significantly larger in size than nonpatenting firms. They also tend to be older, have higher capital-to-labor ratios, and higher shares of new products in sales. Patenting behavior is also highly heterogeneous across industries for the merged above-scale firms. For example, an average firm files 20 times more patents in the computer industry than that in the least innovative food processing industry. The medical industry has the highest fraction (39 percent) of firms filing patents in China; while in industry of apparel, footwear, and caps, only 3 percent of firms ever filed patents.

To understand which firm characteristics are associated with patenting in China, we then estimate a count data model of patents based on the Negative Binomial specification. We find that younger and larger firms, and firms with more R&D investment, patent more. SOEs tend to file fewer patents than POEs, while exporting firms are more innovative than nonexporting firms.

We then compute the within-firm elasticity of firm productivity (and other production performance) to changes in patent stock (i.e. the accumulated number of patent applications). The elasticities of productivity to changes in patent stock are 0.017, 0.014, 0.039, and 0.026 log points for labor productivity, the Solow Residual, the OLS estimate of Total Factor Productivity (TFP), and the Ackerberg et al. (2015) measure of TFP, respectively. These elasticities are surprisingly higher and more significant than those observed in the U.S., as documented in Balasubramanian and Sivadasan (2011). The elasticity of the new product revenue share is 1.5 percent and significant, implying that innovation is also associated with the introduction of new products in China. Significantly positive changes in other production outcomes, such as size (output, value added, capital stock, and employment) and exports are also observed, but not for factor intensity or markup. Although firm entry and exit are definitely important phenomena in a fast-growing economy like China's, and often have far-reaching implications, considering only the surviving firms does not alter our findings, and the estimated elasticities are even larger than the baseline estimates. These findings also hold across different patent types (invention patent, utility model patent, and design patent) with comparable elasticities.

In line with these findings, we also observe that significant real economic changes are associated with firms’ initial patent applications. Based on a difference-in-differences (DID) analysis, using first-time patentees as the treatment group, and nonpatenting firm—selected based on the Propensity-Score Matching (PSM) method—as the control group, we find that significant improvements in productivity, size, new product shares and exports, are associated with first-time patent application events. The significant effect tends to take place after a year from the initial patenting.

We then examine the dynamics of the contribution of patent to productivity growth by allowing the patent elasticity of TFP to vary year by year. Our results show that this contribution declines steadily over time. We investigate whether the weakening elasticity over time is because more and more patent applications are induced by government-led patent fee subsidy programs and hence embody less innovation value and are less correlated with TFP growth. During our sample period (1998–2007), various patent subsidy programs were rolled out sequentially across different provinces in China. Employing data on different introduction years of the programs across various regions (Li, 2012), we find that it is indeed the case. When the patent subsidy program is in place, the elasticity drops on average by 0.017 log point, implying a significant reduction in the TFP-patent elasticity.

Next we investigate whether a firm's ownership status plays any significant role, as ownership is a uniquely important element in understanding firm performance in China (e.g., Hsieh and Klenow, 2009; Song et al., 2011; Zhu, 2012; Chang et al., 2016). We find that the aforementioned positive association between patent application and productivity growth is significantly stronger for SOEs than for their private-owned peers, especially after 2001. This result is not driven by firm entry and exit, neither by changes in firm ownership. This finding is to some extend unexpected, especially considering that SOEs are generally regarded as uninnovative and less effective with their R&D investment.

To understand this somewhat surprising result, we carry out the following further analyses. First, since the conventional role of patents is to deter copying and pre-empt unauthorized entry, patent may generate IPR protection value for the firms in addition to its technological value. To test whether it is the higher protection value enjoyed by SOEs that contributes to the above observation, we turn into another alternative measure of technology—R&D expenditure. Unfortunately, the ASIEC data we have access to only provide R&D expenditure for three years—2001, 2005, and 2006. Nevertheless, when both R&D and patents are included in the estimation of TFP returns to innovation, with the first approximating the unobserved technology and the second representing the IPR protection premium, we observe that the TFP return to R&D is again significantly higher for SOEs while the TFP return to IPR protection is non-differentiable between SOEs and POEs. This evidence suggests that the higher TFP return to SOEs’ patents does not come from the patent protection premium but may reflect true technological value of SOEs innovation.

Second, since this divergence of TFP-patent elasticities between SOEs and POEs emerges after 2001, coinciding with China's accession to WTO, we next investigate whether this event disproportionately benefits exporting SOEs and drives the above result. However, when splitting SOEs into exporting and nonexporting SOEs, we fail to see significant differences. The elasticity of nonexporting SOEs is still significantly higher than the elasticity of POEs.

Third, given the well-documented misallocation of resources between SOEs and private enterprises, this discrepancy could be the result of the considerable credit support SOEs receive from the government, as well as their favorable access to bank loans (which might reflect in their higher leverage ratios). We test the role of government subsidy and leverage ratio in explaining the TFP-patent elasticity, but find that both are insignificant in explaining the elasticity gap associated with the ownership status.

Since the gap emerges following a sequence of SOE reforms, it suggests that reforms might have contributed to the observation. Under the slogan “Grasp the Large and Let Go of the Small”, Hsieh and Song (2015) have shown that reforms intended to strengthen large SOEs, but privatized or closed loss-making SOEs, and have contributed positively to TFP growth. In addition, they document that TFP growth of SOEs was faster than that of private firms for our sample period. Similarly, our analysis also shows that these large innovating SOEs experienced higher TFP growth than innovating POEs, suggesting that the SOE reform may have improved both SOEs’ innovating behavior and TFP growth, as well as the relationship between these two.

Lastly, we use survey data on the perceived quality of local IPR protection across 66 prefectures in China, and show that there is an inverted U-shaped relationship between the strength of IPR protection and firms' innovation and TFP growth. This is consistent with the IPR literature which establishes that stronger IPR protection initially strengthens firms' incentive to innovate and increases innovation, but once reaching a certain level, it could discourage innovation by hindering knowledge diffusion or deterring innovation and entry of smaller firms (O'Donoghue and Zweimüller, 2004; Bessen and Maskin, 2009; Lerner, 2009; Furukawa, 2010; Gangopadhyay and Mondal, 2012).

Related Literature This paper contributes to several strands of literature. First, it is related to the large literature using patent data for economic research on productivity and innovation, which dates back to Schmookler and Brownlee (1962), Griliches and Schmookler (1963), and Scherer (1965). The empirical evidence so far has been concentrated in advanced economies, where high-quality patent data are available and attempts have been made to combine them with firm production data. For example, Balasubramanian and Sivadasan (2011) developed a detailed concordance between NBER patent data and U.S. Census data to examine the consequences of firm patenting. Similar research on developing economies, however, is scant. To our knowledge, this is the first paper evaluating the quality of Chinese patents by examining the relationship between patents and TFP using a large firm-level dataset matched with patent statistics. We observe many similarities between Chinese patent data and the U.S. observations, but also point out important differences in the following sections. Our results thus support the previous literature that uses patents as meaningful proxies of innovation. In addition, the positive association between patents and productivity growth validates the prevailing approach in the literature, which uses changes in TFP or the introduction of new products as measures of Chinese firm innovation (e.g., Aghion et al., 2015).

Second, this paper contributes to the growing literature on various aspects of China's innovation activity. Most empirical studies use aggregate secondary data at the provincial level (e.g., Cheung and Ping, 2004; Li, 2012). A few studies using disaggregated firm-level data focus on listed firms (e.g. Choi et al., 2011; Lin et al., 2010; Boeing et al., 2016) or large and medium size enterprises (e.g., Hu and Jefferson, 2009; Hu et al., 2017) or are limited to a particular region (Lei et al., 2012). One exception is Xie and Zhang (2015), which document several basic patterns of Chinese firm patenting behavior by making a similar effort as ours of matching ASIEC and patent data at the firm level.

Many existing works study the factors behind the explosion of China's R&D expenditure and patent applications. For example, based on large and medium size enterprises data, Hu and Jefferson (2009) find that a combination of rising foreign direct investment, changing ownership structures in Chinese industry, and pro-patent legislation contributed to China's patent boom during 1995–2001. In their follow-up research, Hu et al. (2017) relate the more recent 2007-11 patent surge to noninnovation-related motives for acquiring patents. Consistent with ours, they find the positive correlation between patents in force and labor productivity becomes weaker over time. However, their patent data are only available for five years. Li (2012) finds that patent subsidy programs, together with R&D intensification and a pro-patent legal change fostered the jump in Chinese patents from 1995 to 2007. Recently using a difference-in-differences approach, Fang et al. (2017) find that patent filing increases after SOE privatizations and this increase is larger in areas with strong IPR protection, suggesting that institutions matter for innovation. Chen et al. (2018) analyze the effects of tax cuts for R&D and find large responses of reported R&D and increase in firm productivity.

Third, this paper also adds to a recent literature differentiating innovation behavior of firms with different ownership. Using publicly listed Chinese firms data, Boeing et al. (2016) find that throughout their entire sample period of 2001–2011 POEs are more effective with their R&D spending in increasing TFP. However, for the earlier period of 2001–2006 they find a higher TFP-returns of patents for SOEs than POEs, in line with our result, although this observation is reversed for their later sample period when policy-induced patenting strategies were introduced. Wei et al. (2017) find that innovation productivity is higher for private firms than for SOEs, while the latter receive more government subsidies, and identify the potential misallocation of R&D resources. Corroborating our evidence, Wei et al. (2017) also argue that the Chinese patent explosion is associated with a real, robust improvement in patent quality, based on patent approval rates and comparisons between Chinese patent citations and those of other countries that are patenting in the United States.

The rest of the paper is organized as follows: Section 2 provides an institutional background on China's patent system and describes the data construction and measurement of key variables. Section 3 presents evidence on the determinants of patent activity and examines the relationship between changes in a firm's patent activity and production performance. Section 4 discusses the over time change of the relationship mentioned above and provides the evidence of its link to the introduction of patent subsidy programs. Section 5 investigates the role of state ownership, and Section 6 discusses the role of IPR protection in China and studies how they matter for firms' patenting behavior and productivity growth. Section 7 concludes.

 

Section snippets

China's patent system

China's patent law was first introduced in 1984, put into effect the following year, and has since been amended several times (September 1992, August 2000, and December 2008) to comply with international standards and to facilitate its development into an innovative economy. China's State Intellectual Property Office (SIPO) grants three types of patents: invention patents, utility model patents, and design patents. Broadly speaking, an invention patent protects technical solutions or

Determinants of Chinese firms’ patenting behavior

Before evaluating the relationship between firm patenting with firm performance, we first ask what factors account for Chinese firms' patenting behaviors. As argued in Hu et al. (2017), if patenting decision is driven by technological advancement, as opposed to non-innovation related incentives that affect the propensity to apply for patent, one should expect a tight link between R&D input and patent application. Therefore, studying the determinants of Chinese firms’ patenting provides another

Overtime Changes and the role of patent subsidy programs

As discussed in Section 2.2, during our sample period of 1998–2007, different provinces across China sequentially introduced patent subsidy programs which have been found to cause the increase in the number of patent filings for Chinese firms (Li, 2012; Lei et al., 2012), as they largely reduce the cost of seeking patent protection and increase the overall return of patenting. If patent applications were induced by government policies rather than by commerical proposes, it is reasonable to

The role of state ownership

A unique feature of the Chinese economy is the prevailing existence of SOEs, which are often viewed as less productive than POEs, and not completely driven by profit maximization purposes (Song et al., 2011). A simple comparison between an average SOE and an average POE presented in Fig. V reveals that despite lower R&D intensity, POEs generally have higher patent intensity (number of patent applications-to-sales ratio). Thus, consistent with Wei et al. (2017), the average SOE produces much

The role of intellectual property right protection

Our result so far suggests that Chinese firms' rapidly growing patents are “real”, as patents are found to be associated with real within-firm improvement in production performance, especially productivity. This finding may seem to be at odd with the widely shared perception of that China's IPR protection is rather weak and does not incentivize firms to engage in meaningful innovation activities (e.g. Massey, 2006; United States Trade Representative, 2017). Even though China's IP laws and the

Conclusion

Is firm patenting in China accompanied by real changes in firm production performance, especially firm productivity? This paper answers the question by constructing a unique dataset uniting detailed firm balance sheet data with patent application data for the period 1998–2007. We find compelling evidence that increases in patent stock are associated with increases in firm size (output, sales, and employment), export performance, and more interestingly, firm productivity and the revenue share of 

Declaration of competing interest

None.

Acknowledgement

We would like to thank the following people for their helpful comments and discussions: Daniel Garcia-Macia, Ann Harrison, Zheng Liu, Shang-Jin Wei, Daniel Xu, Jianhuan Xu, Xiaodong Zhu, Fabrizio Zilibotti, and participants of the China Economics Annual 2015, NBER Chinese Economy Meeting 2016, IMF-Atlanta Fed Research Workshop on China's Economy 2017, Bank-of-Canada-University-of-Toronto-IMI Conference on the Chinese Economy 2017, and seminars at Wuhan University and the Huazhong University of