Energy Efficiency and Productivity: A Worldwide Firm-level Analysis.

• Author: Montalbano, Pierluigi

1. INTRODUCTION

   Improving energy efficiency is a growing policy priority for many countries in the world and, as pointed out by the International Energy Agency (IEA, 2014), is widely recognized as one of the most cost-effective and readily available means of addressing numerous energy-related issues such as energy security, the socio-economic impacts of high energy prices, and climate change. The launching of the so-called "Agenda 2030" (1) has introduced the energy issue to the Sustainable Development Goals (SDGs), namely "ensure access to affordable, reliable, sustainable, and modern energy for all" (SDG 7). Among other things, SDG 7 establishes the need to prioritize energy-efficient practices and, specifically, to double the global rate of improvement in energy efficiency by 2030 (Target 7.3). (2) The main motivation for this collective action is that energy-related emissions--especially produced from fossil fuels, such as coal, petroleum, and natural gas--increase greenhouse gas (GHG) emissions and other air pollutants, thus worsening global warming. (3) Since manufacturing and industrial activities are among the primary energy users and are thus also responsible for C02 emissions (Abdelaziz et al., 2011), industrial firms have been put under political and social pressure to re-examine their energy awareness practices and move towards greater energy efficiency (Okereke, 2007). A common argument is that energy efficiency and competitiveness are intrinsically connected and efficiency enhancement occurs without negative economic consequences, promoting economic performance (EU, 2012; United States Congress, 2013; World Bank, 2009; Eichammer and Walz, 2011). In this perspective, increasing energy efficiency can also be seen as a tool for enhancing productivity growth (Schurr et al., 1960; Schurr, 1982; Jorgenson, 1984). By reducing production costs, energy efficiency provides a competitive advantage to firms as well as the possibility to develop new products or processes. Furthermore, it can boost compliance with environmental and quality standards. This is also consistent with the so-called "Porter Hypothesis" (Porter and van der Linde, 1995) where more stringent environmental regulations are supposed to trigger innovation, which, in turn, may offset regulatory costs, thus ensuring better economic performance. In this perspective, energy efficiency can support not only SDG 7 but also world progress towards SDG 8 which calls for greater economic growth and productivity (World Bank, 2020). (4)

   A major drawback to the empirical literature in this field is that, to our knowledge, there have been very few academic studies that measure the contribution of energy efficiency to firm performance. Most studies are carried out at the aggregate level or focused on developed economies, whereas relatively few firm-level analyses have been carried out for developing countries. The global dashboard set out to monitor the achievement of the SDG 7 also applies indicators of "National Energy Intensity" and provides aggregate empirical correlations with GDP growth (World Bank, 2020).

   Relying on the World Bank Enterprise Survey (WBES) dataset, here we explore the relationship between energy efficiency and productivity using firm-level data that are available for a large sample of countries, mainly developing countries, during the period 2006-2018. Thanks to the WBES, we can rely on panel data for a subsample of firms. Panel data are essential in these kinds of exercise since they allow to control for differences in unobservables, both common shocks across time as well as time-invariant differences across firms or sectors. To the best of our knowledge, this is one of the earliest and most comprehensive studies exploring the effect of energy efficiency on productivity at the firm level in terms of geographical coverage. To this end, we apply a Cobb-Douglas production function that expresses firm productivity as a function of labor, physical capital, human capital, productive inputs, and a technology shifter proxied by a standard process of economic innovation, exporter status, and our variable of interest, that is energy efficiency. (5) We also control for industry and firm heterogeneity (Doro et al., 2010; Mulder and de Groot, 2012; Grossi and Mussini, 2017) and apply alternative measures of energy efficiency.

   Our findings show a positive relationship between energy efficiency and firm-level productivity worldwide, although with some degree of heterogeneity in firm size, industry, and geographical regions. These results are robust to a set of sensitivity tests and when using different techniques. They are consistent with those obtained by previous firm-level analyses, although the latter are focused on a narrower set of countries, do not control for panel dimension, and analyze different time-spans (Cantore et al., 2016; Montalbano and Nenci, 2019). This work provides empirical support for the messages conveyed by international institutions regarding the positive relationship between environmental actions and firm performance, thus supporting the efforts to improve the private sector's energy efficiency. It shows that the implementation of Agenda 2030 is not only a fundamental step for sustainable development, but also a tool for fostering firm-level productivity. However, the detected heterogeneity by sector, size, and geography suggests some caution in adopting one size fits all solutions. Incentive schemes aiming at fostering the productivity of firms should be fine tuned on firm characteristics and allow for appropriate forms of flexibility.

   The paper is organized as follows: Section 2 briefly summarizes the literature, Section 3 introduces the methodology and provides some stylized facts on energy intensity by firm characteristics at region and industry level, Section 4 presents the empirical analysis and reports the outcomes, and Section 5 concludes and provides policy implications.

2. LITERATURE REVIEW

   Over the last decades, the literature on the relationship between environment and productivity has increased considerably. A major drawback is that most studies in the field are either carried out in developed economies or at the aggregate level, whereas a very limited number of works look at developing countries and firms. The classic studies investigating the relationship between energy and productivity are Schurr et al. (1960); Schurr (1982), and Jorgenson (1984). All these focused on the United States' industrial sectors and showed a general inverse relationship between productivity and energy intensity. This inverse relationship was explained not solely on the basis of substitution of expensive energy for more expensive labor, but also with technical change. Later works called those claims into question (see Murillo-Zamorano, 2005, for a critical view on this issue). More recently, however, international organizations advocated for an "energy productivity bonus", that is, the "bonus wealth" created by annual energy intensity improvements (IEA, 2019; World Bank, 2020).

   The debate on the impact of energy efficiency on productivity at the firm level has mainly centered on the so-called "Porter hypothesis". Porter (1991) and Porter and van der Linde (1995) stated that more stringent but properly designed environmental regulations might create incentives for firms to innovate, increase efficiency, and subsequently enhance their performance. The literature differentiates between "weak" and "strong" versions of the Porter Hypothesis: the "weak" version states that environmental regulation may lead to innovation; the "strong" version adds that the regulation can improve firms' competitiveness and productivity (Jaffe and Palmer, 1997). Up to now, scholars have not reached relatively consistent conclusions on the empirical evidence of the "strong" version of the Porter Hypothesis. Some works concluded that environmental regulation policy has led to a reduction in productivity due to the higher costs that firms may face (Gray and Shadbegian, 1995, 2003; Becker, 2011; Dechezlepretre and Sato, 2014) or showed a negative and/or a not significant effect of environmental regulation on productivity growth (Cohen and Tubb, 2015; Kozluk and Zipperer, 2014; Hille and Mobius, 2019). Others highlighted the positive effects of this kind of regulation on productivity (Hamamoto, 2006; Yang et al., 2012; Jorge et al., 2015; Qiu et al., 2018). The discrepancy between these results is probably caused by the fact that these studies suffer from a lack of generality since they usually analyze very specific regulations or industries in a single country setting (Albrizio et al., 2017) as well as not having a uniform standard for measuring performance (Zeng et al., 2010).

   With regard to developing economies, most of the literature has focused on the relationship between the so-called "energy constraints" (6) and productivity. Of these studies, Hulten et al. (2006) showed that growth in electric generation capacity positively affected productivity growth in Indian manufacturing in the previous twenty-year period. Fernandes (2008)--using data from a survey of large manufacturing firms in Bangladesh--documented that power supply problems affect firms' productivity growth. Eifert et al. (2008) showed that energy constraints increase firms' indirect costs, partly explaining low firm productivity in Africa. Arnold et al. (2008), using a sample of manufacturing firms from ten Sub-Saharan African countries, investigated the relationship between firm productivity and access to electricity services. They found that energy reliability problems exert a significant negative effect on firms' productivity. Escribano et al. (2009) assessed the impact of energy infrastructure quality on the productivity of African manufacturing firms showing a strong negative effect. Drawing on Indian firm-level data, Abeberese (2013) analyzed the effect of infrastructure...