**1. Introduction**

A frequently asked question is, do environmental regulations reduce a firm's productivity and employment? While no strong conclusion has emerged, countries have strengthened their domestic and international environmental regulations. Although studies on the e ffects of environmental regulations on a firm's employment and productivity have increased [1], they mainly focus on the effects of polluting industries and do not consider the heterogeneity e ffects among di fferent sectors. The debate on environmental regulations should be analyzed by distinguishing between green and non-green sectors. The two sectors are defined as industries producing goods and services related to the environment or resource recycling.

One segmen<sup>t</sup> that is the focus of policy is the green sector. Environmental regulations can create employment or so-called green jobs in this sector; this is important as these industries are regarded as the drivers of sustainable development.

This study examines how manufacturers respond to regulations measured by the Low-Carbon Green Growth (LCGG) strategy. Korea's environmental regulations were strengthened in the 2000s according to international standards. After 2010, the Government of Korea intended to simultaneously achieve both environmental and economic goals through the LCGG project [2]. Environmental regulations were strengthened after President Lee proclaimed 'Low-Carbon Green Growth,' as the government's goal for reducing greenhouse gas (GHG) emissions by 27–20 percent by 2020 relative to the 'business as usual' scenario of 2005. The Act on Low-Carbon Green Growth was enacted on 13 January 2010 to promote the development of the national economy by laying down the necessary foundation for low carbon, green growth and by utilizing green technologies and green industries as new engines of growth. This research is interested in the e ffects that these regulations had on economic outcomes defined as employment and labor productivity.

Greenstone [3] reviewed the e ffects of CAAA (Clean Air Act Amendment) on economic growth and industrial activities using US manufacturing plant-level data. US counties received non-attainment or attainment designations according to the air quality standards after CAAA was implemented. His study suggests that environmental regulations led to emitters in non-attainment counties who were subject to stricter regulatory oversight (treatment groups) losing output, jobs, and capital investments as compared to emitters in attainment counties. His results also sugges<sup>t</sup> that there was a trade-o ff between environmental and economic outcomes. This is contrary to Porter's [4] argumen<sup>t</sup> that environmental regulations can promote investments and technology development. Morgenstern et al. [5] and Lanoie et al. [6] confirm the positive e ffects of regulations on employment and productivity respectively. The e ffects of environmental regulations on a firm's performance are still controversial because in addition to the enterprises' extensive margins of entry and exit the regulations also lead to labor reallocations between regulated and non-regulated firms asymmetrically depending on the type of regulation [7]. So, the total economic outcomes in the general equilibrium framework depend on the characteristics of the industries and the applied environmental policy.

In this study, we test three hypotheses. First, the Porter Hypothesis (PH) is tested in terms of labor productivity and employment in the polluting industry. The polluting industry's performances will decrease more than that of the less-polluting industry under environmental regulations in keeping with Greenstone [3] and Walker [8]. Second, dividing the firms by another criterion, namely whether they produce environmental goods, the green sector might benefit from the regulations. Lastly, this paper tests whether the negative e ffects of the regulations on a polluting establishment can be o ff-set if it is included in the green sector. We use Korea Statistics' manufacturing surveys from 2004 to 2015, which include the periods before and after LCGG's implementation. Manufacturing is identified as the major emitter of harmful substances in the production and processing of paper, rubber, chemicals, and petroleum refining. Hence, it responds sensitively and di fferently, depending on the emissions in the processes and the types of products produced. In addition, green manufacturers account for about 20 percent of this industry so it is useful to examine cross-sectional variations between these establishments as well.

Figure 1 shows the increased energy and environmental expenditure in the manufacturing industry from 2004 to 2016. The energy and environmental expenditure includes pollution abatement and control expenditure (PACE) and is a proxy for domestic environmental regulations since the expenditure is not directly related to the firms' profit maximization behavior but to environmental pressures by the authorities [9]. Expenditure increased sharply during the LCGG project period after 2010. This trend is useful for examining the e ffects of environmental regulations on economic growth over time using the di fference-in-di fferences method.

**Figure 1.** Energy and Environmental Expenditure of Korean Manufacturing (Source: Statistics Korea).

This study contributes to the environmental economic field in which studies that inspect the link between environmental policies and their effect on the economy are scarce outside the US. Environmental policies covered in literature using US data include Clean Water and Clean Air Acts [10,11], Clean Air Act Amendment [7], and Cross-State Air Pollution Rule [12]. Our empirical results sugges<sup>t</sup> that employment and labor productivity were impacted by stricter environmental standards under LCGG in Korea and their effects were asymmetric depending on the industry's features. These findings have policy implications for 'the environment versus jobs' debate and the possibility of sustainable development in terms of productivity.

The rest of this study is organized as follows. Section 2 reviews previous studies on environmental regulations and firms' competitiveness. Section 3 documents the characteristics of industries according to CO2 emissions and environmental classifications. Section 4 explains the theoretical framework of the relationship between regulations and a firm's performance and sets the empirical equation. Section 5 reports the results of the estimation analysis for employment and labor productivity. The conclusion is given in Section 6.

### **2. Literature Review**

In the literature, the effect of regulations on a firm's economic outcomes has been examined in terms of productivity, investments, employment, and international trade [1]. Several studies also focus on the manufacturing industry since it is regarded as the main culprit responsible for emitting toxic substances and as such is a target of environmental regulations. After the famous Porter Hypothesis [4] researchers have tested whether this hypothesis can hold.

The following studies imply that the Porter Hypothesis holds at least weakly: Lanoie et al. [6] found the hypothesis to be consistent with Quebec manufacturing data. Their findings sugges<sup>t</sup> that the direct effect of regulations on total factor productivity (TFP) growth was negative, but the lagged regulatory variable had a positive productivity effect. They also confirm that this effect was stronger when the industries were more exposed to international competition. Jaffe and Palmer [13] and Johnstone et al.'s [14] studies focus on investment activities in industries. Jaffe and Palmer [13] sugges<sup>t</sup> that lagged stringency of environmental regulations measured by pollution control expenditure

spurred R&D activities by using a three-digit industry level and a fixed effects model. Their results were the opposite when the model was estimated using the pooled ordinary least squares (POLS) method neglecting industry heterogeneity. Johnstone et al. [14] and Rubashkina et al. [15] found that innovations based on patents and R&D were not obstructed by environmental policies.

Yang et al. [16] support the Porter Hypothesis using data for Taiwanese manufacturing plants by using pollution abatement fees and R&D expenditure. In their study, capital expenditure was not significantly related to R&D in the case of pollution abatement. Molina-Azorín et al. [17] analyzed the relationship between a firm's performance and environmental practices in the Spanish hotel industry and found that a strong commitment to environmental practices was linked to the hotels' higher performance levels.

Greenstone [3] used US manufacturing data with fixed effects models like Jaffe and Palmer's [13] study, but his results are the opposite. Ozone regulations had the strongest contemporaneous negative effect and the overall effect of the regulations on the industry's TFP was also negative. The dynamic effects of regulations captured by the variations between attainment and non-attainment counties under CAAA were negative and the opposite of PH. According to Wagner et al. [18] the hypothesis did not hold in the European paper industry. Yuan and Xiang [19] maintain that the hypotheses for both weak and strong versions were not supported by data on Chinese manufacturing industries, but in the short term environmental regulations improved labor productivity and environmental and energy efficiency. Using German data, Rexhäuser and Rammer [20] show that PH held only in those innovations which increased a firm's energy efficiency. Thus, empirical research on environmental regulations' effects on a firm's competitiveness has reached no conclusive results [21].

The crowding-out effect in investments between clean and dirty industries is one of the reasons why it is hard to predict the net effect of regulations in an economy. Wang and Shen [22] examined the effects of environmental regulations in China, separating the industries on the basis of cleanand pollution-intensive production. They concluded that clean production industries had higher environmental productivity under regulations than dirty industries. Using the system generalized methods of moment (GMM) and threshold regression estimation methods, they also found that environmental regulations and environmental productivity had an inverted U-shaped relation. Gray and Shadbegian [10] and Kneller and Manderson [23] concentrated on the crowding-out effect between green and traditional investments (technology). According to them, environmental regulations promoted green investments while increased investments in green technology crowded-out existing investment activities. Gray and Shadbegian [10] explored investment decisions of US paper plants using a multinomial logit model. They found that pollution abatement costs and expenditure (PACE) used as a proxy of environmental investments crowded-out productive investments. Kneller and Manderson [23] also found this substitution in UK's manufacturing industry data using the two-step system, GMM. More pollution abatement pressures increased environmental research and development (R&D) and investments in environmental capital, but an increase in environmental R&D tended to crowd-out non-environmental R&D.

These asymmetric effects of environmental regulations occur not only at the industry or plant level but also at the common border between countries. Alpay et al. [24] used data from the food manufacturing industry in the US and Mexico during the North American Free Trade Agreement (NAFTA) to capture the effects of environmental regulations on each country's productivity under free trade. They argue that stricter environmental regulations in Mexico enhanced the industry's productivity growth thus corroborating PH, while US pollution regulations had an insignificant effect on the US food industry.

When it comes to employment, the impact of green innovations in literature is ambiguous. Kunapatarawong and Martínez-Ros [25] investigated the relationship between innovation activities and employment using firm-level Spanish panel data. They sugges<sup>t</sup> that green innovations and employment were positively linked, especially in a dirty industry. Morgenstern et al. [5] found a weak positive relationship between stringent environmental regulations measured by PACE and jobs using a

structural model in four polluting manufacturing industries. They used aggregated data from pulp and paper mills, plastic manufacturers, petroleum refiners, and iron and steel mills to estimate the structural model based on a translog cost functional form. Yamazaki [26] and Walker [8] focused on the re-allocative e ffects of regulations. Yamazaki [26] showed that in British Colombia employment fell in carbon-intensive and trade-sensitive industries, but jobs increased in clean service industries with the revenue-neutral carbon tax. Walker [8] found that workers reallocated from the regulated sector to other industries within the same labor market under CAAA. Aldy and Pizer [12] investigated power sector regulations and found that these a ffected jobs negatively, but they raised electricity rates and led to higher production costs in all US firms.

Previous research using Korean data includes that by Kang and Lee [27], Kim and Ha [9], Kang and Jo [28], Jung [29], and Lee and Choi [30]. According to Kang and Jo [28] the enforcement of the Indoor Air Quality Control Act enabled about 652 people to find jobs during 2015-19. The following papers show that the PH holds using proxies for environmental variables; Kang and Lee [27] concluded that PACE, a proxy for environmental regulations, increased R&D and this e ffect was more significant in the stringent regulation period (1992–2001) than during the less stringent regulation period in 1982-91. They also used industry-level manufacturing data and fixed-e ffects estimation. Kim and Ha's [9] research is similar to ours in terms of data. They used the ratio of energy and environmental expenditure to measure environmental pressures. The e ffects of environmental pressures on TFP were positive, especially in polluting industries through a lagged regulation proxy.

Studies assessing the e ffects of environmental regulations not the proxy variables are scarce in case of Korea, because domestic environmental regulations are not strict. The Low-Carbon Green Growth (LCGG) Act sets the fundamentals of environmental policies including the Emission Trading Scheme (ETS). Article 46 of the Act states that the introduction of a cap and trade system for emissions of GHG [2] and the pilot project for ETS was undertaken in January 2010. The carbon market will stimulate incentives for firms to reduce their emissions; as previously announced the Act was implemented in 2015. The GHG Inventory and Research Center was also established to construct GHG data and manage emissions according to Article 45 [2,31]. Lee and Choi [30] show that ETS implemented in 2015 promoted technical changes in the manufacturing industry, but this top-down approach by the governmen<sup>t</sup> may have limits in enhancing environmental e fficiency. Jung [29] suggests that LCGG was based on economic growth rather than ecological modernization and green jobs were only created in traditional environmental protection and pollution reduction areas. While this government-initiated project for achieving both economic growth and environmental conservation has been controversial [32], environmental and energy expenditure in the country reflects increased investments in the environmental parts (see Figure 1). We can directly analyze the e ffects of environmental regulations by dividing periods into before and after LCGG.

This study contributes to literature in two ways. First, it extends the field by shedding light on the relationship between regulations and firms' economic performance by distinguishing the sectors within the polluting industry. Second, it uses LCGG for measuring stringent environmental regulations which should be helpful in assessing policy. Therefore, the results sugges<sup>t</sup> new academic research and important policy implications according to the types of sectors.
