Essays on globalization, market power and innovation
Publication Date
July 25, 2023
Creator
Kazmi, Syed Fahad Hasan
Abstract
Competition between firms is an essential ingredient for an economy to function prop- erly. A higher level of competition through the process of firm entry into markets allows for prices to better reflect the costs of production being incurred, thereby benefitting consumers. In an economic environment with the dwindling competition, firms gain market power and thus are able to command higher prices. This has negative welfare implications for consumers along with a potential for creating misallocation of resources. It also discourages business dynamics, reduces a firm’s incentive to invest in capital and has negative implications for innovation in the economy. While it has been extensively documented that global market power is on the rise, there is little research on the effects globalization can play in mediating and taming market power. The effect of market power on a firm’s incentive to innovate is also an area of research that has received little attention.
Economic theory suggests that the markup is a key measure of market power and that its relationship with trade is rich and complex. Trade liberalisation can reduce markups via a decline in the residual domestic demand but also increase it via several channels. Trade-induced increases in competition leads to more concentrated markets via entry and exit, putting upward pressure on markups. Market shares reallocation toward larger, more powerful firms, increase the aggregate markup. In the first and second chapter of my thesis, we use a large episode of trade liberalisation in Spain to test this rich set of transmission mechanisms linking trade and markups. We use firm- level data from survey of Spanish manufacturing firms, the Encuesta sobre Estrategias Empresariales (ESEE), to estimate markups via the production function approach used in De Loecker and Warzynski (2012). The overall effect of reductions in Spanish import tariffs on firm-level and aggregate markups is pro-competitive but we find evidence of offsetting effects via the other channels. In particular, we show that firms with high intangible investment experience a weaker reduction in markups. Supporting the theoretical insight that the feedback effect via concentration is stronger with higher barriers to entry. Increases in markups are also produced by reallocations effects but the results are weaker, suggesting that the link between trade and markups is mostly driven by changes at the intensive margin.
The third chapter looks at the relationship between trade, competition (domestic and foreign) and innovation, which is complex. Schumpeter (1942) emphasizes a business stealing effect in which increased entry by foreign firms into domestic markets (via trade) reduces the duration of monopolies and hence the innovation incentives for those incumbents. A counteracting economic force introduced by Aghion et al. (2005a) and Akcigit et al. (2018) is the escape-competition effect in which through strategic interactions between multiple firms in a given product line, incumbent firms have an incentive to move ahead of their competitors through innovation. In our paper, we find evidence of both of these counteracting forces at play. These theoretical predictions inform our empirical analysis of the effects of firm markups on innovation in the Spanish man facturing industry in the period 1990-2010. In these years, firm investment in Research & Development (R&D) expenditure witnessed a substantial increase. We use firm-level data from survey of Spanish manufacturing firms, the Encuesta sobre Estrategias Empre- sariales (ESEE), to estimate markups via the production function approach (De Loecker and Warzynski, 2012). We find a statistically significant negative relationship between firm markups and their corresponding R&D intensity. This relationship does not change for firms far or near the technology frontier. We find that firm market power is also negatively correlated with product and process innovation. Firms with lower market power are equally likely to engage in product and process innovation. This is an interesting result as it is contrast to Aghion and Howitt (1996), who find that competition induces firms to reallocate resources away from process innovation towards product innovation. When instrumenting for firm market power with output tariffs, we no longer find a statistically significant negative link with firm R&D intensity. This seems to suggest that between the two competing economic forces; escape competition effects and business stealing effects, the former wins out, albeit this relationship is not statistically significant. Our results imply that the escape-competition effect could be one of the many channels attributed to the increase in R&D expenditure for firms across the Spanish manufacturing sector.
Closely related to our work in the first and second chapter is the empirical literature on the impact of trade on markups. Early studies of the effects of trade on firm-level markups focused on reductions in output tariffs, finding evidence of pro-competitive effects (e.g. Harrison, 1994; Levinsohn, 1993). Recently, De Loecker et al. (2016) extend the analysis to reductions in tariffs on intermediate inputs, generating a decline in costs that can increase markups if firms do not pass it entirely to consumers. Studying an episode of large scale liberalisation in India, they find that while the reduction of tariffs on final goods, output tariffs, has pro-competitive effects on markups, declining input tariffs have the opposite effect. In the first chapter, we complement their findings showing that trade can have anti-competitive effects above and beyond the incomplete pass-through. Our finding suggests that lower output tariffs can be anti-competitive as well, as trade can increase firm-level markups, via a concentration effect, and the aggregate markup, as a consequence of reallocations of market shares toward more productive (high markup) firms. To the best of our knowledge we provide the first empirical test of the concentration and reallocation effects of trade on markups in the first chapter.
In the third chapter, on the relationship between competition and innovation there is a large body of work that illustrates the complex relationship between the two. Using sector-level UK data, Aghion et al. (2005a) find an inverted-U relationship between competition and innovation. A similar analysis conducted by Hashmi (2013) for the U.S. reports a negative relationship between competition and innovation. There are other firm-level analyses confirming the inverted U-shape relationship for large firms in different developed economics, namely France Askenazy et al. (2013), while a positive relationship is found in Spanish data in a study conducted by Beneito et al. (2015). Aghion et al. (2009) suggest that foreign entry encourages innovation for firms in technological advanced industries and discourages it for firms in laggard sectors.
On the relationship between innovation response and trade liberalization there is a wide array of empirical literature that looks at this problem. There is special focus on the impact of import competition on the innovation response of firms to escape competition (see Bloom et al. (2016), Iacovone et al. (2011); Autor et al. (2016); Hombert and Matray (2018). Aghion et al. (2019) show that R&D-intensive firms exhibit higher sales growth, profitability, and capital expenditures than low-R&D firms when faced with Chinese competition. Fresard and Valta (2016) disentangle the horizontal and vertical components of the Chinese Import Shock at the firm-level and analyse its effects on employment, sales and innovation. De Loecker et al. (2016) find that input tariff liberalization in India has benefited domestic firms as they have enjoyed lower production costs while they have simultaneously raised markups. They argue that additional short-run profits accrued to firms may have spurred innovation in Indian manufacturing, particularly through the introduction of many new products. This tends to suggest that higher (and not lower) markups can be associated with higher innovation. Our analysis is a departure from this as we look at the effect of a reduction in output tariffs (not input tar- iffs) on markups and how that shapes innovation outcomes. On the export competition side, Aghion et al. (2021) analyse how the timing, quantity and quality of French patents generated by French exporters respond to demand shocks in the export markets served by those firms – and how this varies with different characteristics. We contribute to this literature by analysing the relationship between changes to firm-level markups (caused by trade liberalisation) and how that impacts their innovation responses.
The empirical results in De Loecker and Eckhout (2020), Bajgar et al. (2019), and Diez et al. (2021) suggest that the increase in markups is a global phenomenon. Several re- cent papers provided explanations for these facts based on technological or institutional channels. Gutierrez and Philippon (2018) use a political economy model to show that countries in a single market tend to promote supranational regulation enforcing stricter competition. They test and validate the prediction of the model showing that European institutions enforce more competition. Another research line instead explores channels operating via innovation and technological change. Aghion et al. (2017) show that the IT revolution has enabled superstar firms, with high markups, to expand thereby in- creasing the average markup in the economy. Similar results are obtained in De Ridder (2020) where the source of rising market power is the increase in the use of intangible inputs. Akcigit and Ates (2019) focus instead on the slowdown of the diffusion of technology from frontier firms to laggards. Our research in chapter 2 finds that markups are declining, in stark contrast to existing literature. We find that markups for the Spanish manufacturing sector are declining throughout the sample period between 1990 and 2010. While the rise in markups globally have been attributed to factors such as techno- logical change and institutional reform, our research in chapter 1 and 2 argues that trade can play a mediating role in shaping market power outcomes.
From a methodological perspective (used in the empirical analysis), the production function estimation approach outlined by De Loecker and Warzynski (2012) and eventu- ally employed in papers such as De Loecker et al. (2016) and De Loecker and Eckhout (2020) has been criticized as of late. De Ridder et al. (2022) discuss the use of revenue- based measures in the data available to the researcher to estimate firm-level markups and the restrictive approach of estimating Cobb-Douglas variants of the production function (both following De Loecker and Warzynski (2012)), both of which create biases in the markup estimates. To address these problems in the thesis, first, we deflate our revenue- based measures in the ESEE firm-level dataset using firm-specific price changes, to back out a rough measure of firm-level quantities. Second, we estimate translog variants of the production function, in addition to Cobb-Douglas, to corroborate our findings - our results remain the same. Raval (2023) critiques the production function approach by showing that the choice of inputs used to back out markup estimates from the data gives different and at times opposite results. For every dataset, markups estimated using labor are negatively correlated with markups estimated using materials, exhibit greater dispersion, and have opposite time trends. Raval (2023) continues to find stark differences in markups estimated using energy and non-energy raw materials. This thesis acknowledges the critique posited by Raval (2023). In our case, both labour and material share firm-level markups estimated from the production function are positively correlated and have similar time trends - in stark contrast to the findings from Raval (2023). In our robustness section in chapter 2, we also show that using labour as a flexible input in the production function yields similar results, when testing for the firm- level pro-competitive effect, to their material share counterparts (which constitute as the benchmark measure of firm-level markups in our analysis).
The theoretical channels we highlight using our model are not new. The procompetitive and incomplete pass-through channels feature in a large class of theoretical models with variable markups. Arkolakis et al. (2019) show that in monopolistically competitive models with endogenous markups obtained departing from CES preferences, the horse race between those two channels crucially depends on the choice of preferences. Introducing heterogeneous firms opens up the possibility of a positive effect of trade on aggregate markups via the reallocation channel (e.g. Melitz and Ottaviano, 2008). Oligopolistic trade models feature these channels as well (e.g. Brander and Krugman, 1983; Atkeson and Burstein, 2008; Edmond et al., 2015). Moreover, Venables (1985) shows that in this class of models free entry generates the concentration effect of trade in the case of unilateral liberalisation. As in the model presented in chapter 1, Impullitti et al. (2021) find that the concentration effect attains also with multilateral liberalisation. The scope of the model we use in chapter 1 is not to uncover a new transmission channel of trade to markups but rather to provide a tractable unified framework that includes many existing channels and use it as a theoretical guide for our empirical investigation.
What follows are chapters 1, 2 and 3 of my thesis. Chapter 1 details the theoretical framework that goes onto inform our empirical analysis on the relationship between trade liberalization and firm markups in chapter 2. Chapter 3 looks the relationship between trade, firm market power (competition) and their corresponding innovation efforts. A conclusion to the thesis is presented after the end of chapter 3.
Item Type
ethesis
Thesis Type
PhD
Associated Schools / Departments
School of Economics (UK)
eprints ID
73582
UoN Repository URI
Except where otherwise noted, this item's license is described as
File(s)![Thumbnail Image]()
Name
Final_merged_signed.pdf
Type
Full-text
Description
Examined
Size
2.08 MB
Format
Adobe PDF
Checksum (MD5)
be35ea3a404e615b9d7ee552dd1c692c