This paper investigates the role of R&D investment in shaping the relationship between financial
constraints and aggregate total factor productivity (TFP). I study a dynamic model in which R&D
investment, which affects productivity evolution endogenously, is subject to financial constraints. I parameterize the model with production, innovation, and balance sheet data. The estimated model implies sizeable static TFP losses caused by capital misallocation and dynamic TFP losses from distorting R&D investment. The accumulation of internal funds reduces the static TFP loss gradually.
In contrast, because R&D has a persistent effect on productivity, the dynamic TFP loss rises initially and declines later. Compared to a model with exogenous productivity, innovation investment
makes firms less able to use self-financing to reduce TFP losses, and prolongs the transition. Endogenous productivity growth amplifies the gains in TFP and output from financial reform, and
leads to a longer-lasting consequence from a credit crunch. Improving the pledgeability of intangible assets in China to be the US level reduces the static TFP loss only 0.4%, but the dynamic TFP loss
This paper proposes that processing trade, which played an important role in China’s export
miracle, not only leads goods to be “Made in China,” but also “Created in China.” Using unique
transaction-level trade data on firms’ branding information, we document four main findings.
First, there exists a significant share of exporters that engages in both ordinary and processing
export activities, and they exhibit superior performance in various margins. Second, even within
firms, there is a tight link between firms’ export mode choice and brand ownership—own branded
products are typically exported under ordinary trade regime while products under other firms’
brands are exported under processing trade regime. Third, there is a price premium associated
with own-branded products. Fourth, Chinese firms intensify their branding activities when faced
with favorable processing trade policies upstream. To rationalize these findings, we present a
simple theoretical framework where firms with multi-attributes endogenously determine their
specialization within the production network.
We study a stylized model of multi-product firms with firm-product level heterogeneity with Hicks-neutral production technology. The productivity process allows flexible correlation between the production efficiency of different products. We characterize the moment conditions implied by the model and show that the production functions are non-parametrically non-identified without observing either the allocation of inputs or exogenous input price variations. If the production functions are in a parametric family and are point identified, we can use standard GMM method to estimate the parameter of interest. If the model is partially identified, set estimation method based on the moment equality conditions can also be used. In the case of the Cobb-Douglas production functions, we show the optimal input allocation rules can be solved in closed form. We also argue that the derived moment conditions identify the production functions. Monte Carlo evidence shows that our estimation strategy performs well. We then apply our methodology to a sample of agricultural goods manufacturing firms. The result shows the production functions of the multi-product firms differ from that of the single product firms even for the same product. We also find that multi-product firms produce their core (resp. peripheral) products at higher (resp. lower) technical efficiency than single-product firms. Lastly, we find a strong positive co-movement between the productivity shocks of different product lines.
Bank Competition, Policy Distortions, and Margins of Innovation [Update Coming Soon!]
with Mengbo Zhang
We study a model in which firms differ in their defaulting risks to investigate the relationship between bank competition and risky investment. Banks face asymmetric information when lending to firms and compete in a Betrand-Edgeworth game. In equilibrium firms with higher defaulting risks are credit rationed. We show that increasing banking competition enhances investment by firms with low defaulting risks at the extensive margin, and stimulates investment by firms with large defaulting risks. Employing a deregulation policy for city commercial banks in China, we find empirical evidence consistent with predictions of the model.
We study the impact of anticorruption efforts on firm performance, exploiting an unanticipated corruption crackdown in China’s Heilongjiang province in 2004. We compare firms in the affected regions with those in other inland regions before and after the crackdown. Our main finding is an overall negative impact of the
crackdown on firm productivity and entry rates. Furthermore, these negative impacts are mainly experienced by private and foreign firms, while state-owned firms are mostly unaffected. We present evidence concerning two potential explanations for our findings. First, the corruption crackdown may have limited bribery opportunities that helped private firms operate. Second, the corruption
crackdown may have interfered with personal connections between private firms and government officials to a greater extent than institutional connections between state-owned firms and the government. Overall, our findings suggest that corruption crackdowns may not restore efficiency in the economy, but instead
lead to worse economic outcomes, at least in the short run.
We extend the empirical framework by Peters et al. (2017) to include both R&D and patents in the productivity evolution. We provide a decomposition of the benefits of R&D into the patent and non-patent components, and a novel measure of the patent value conditioning on the firm's R&D investment. Using a sample of Chinese high-tech manufacturing firms, we find that (1) 47.8% to 67% of the benefits of R&D investment comes from non-patent R&D activities; (2) On average an invention (a utility model) patent causes around 0.76 (0.66) percent increase in the firm value; (3) The start-up costs of R&D are around ten times as large as the maintenance R&D costs. The counterfactual analysis shows that the lump-sum subsidy is more effective than the proportional subsidy in increasing the expected firm value and innovation probability. R&D continuers respond more actively than the R&D starters to the R&D subsidy.
As a developing economy, China's unprecedented patenting surge is puzzling. We study China's patent surge and its driving forces using a novel and comprehensive merged dataset on patent applications filed by Chinese firms. We find that R&D investment, FDI, and patent subsidy have different effects on different types of patents. First, R&D investment has a positive and significant impact on patenting activities for all types of patents under different model specifications. Second, the stimulating effect of foreign direct investment on patent applications is only robust for utility model patents and design patents. Third, the patent subsidy only has a positive impact on design patents. The results imply that FDI and patent subsidy may disproportionately spur low-quality patents.
This paper investigates the relationship between imports and innovation by importing
firms. We first construct a theoretical model in which imports stimulate innovation
through cost-reducing knowledge spillovers. We then employ a combined micro dataset
of Chinese manufacturing firms to estimate the effects of imported intermediates on the
firm’s R&D investment. The dataset allows us to construct firm-year level instruments for
importing and exporting that are uncorrelated with the innovation decision of the firm.
Our estimations find that: (1) importing intermediates tends to increase importing firms’
R&D intensity; and that (2) exporting also increases importing firms’ R&D intensity. Examining
the channels through which importing affects innovation, we find that importing
from high-income sources has a greater impact on innovation. High-tech firms tend to
experience greater increases in innovation intensity, as do private firms. Our results are
supported by a series of robustness checks.
This paper asks whether banks help or grab the enterprise in the real economy. Using the firm-
level data on Chinese enterprises during 2001–2007, we find that interest payment of private
enterprises is negatively related to the return on sales (ROS) and asset growth, which implies a
detrimental effect of bank loans on private firms' performance. But this linkage is significantly
positive for state-owned enterprises. Focusing on private enterprises, the grabbing impact from
banks is strongest for firms without government subsidies, with low production values, with
small size, or with low capital intensity. Our results are robust to alternative estimation
approach and variable specifications. To conclude, the bank-centred financial system in China
has assisted in the development of state-owned enterprises, while the development of private
enterprises has been impeded by Chinese banks.