The relationship between environmental
performance and firm value for firms in the utility
sector, between 2010 – 2015.
Amsterdam Business School
Name Alien van der Vliet Student number 10781153
Program Economics & Business Specialization Finance & Organization Number of ECTS 12
Supervisor dr. I.J. (Ilko) Naaborg Completion 27-06-2017
Abstract
It has become increasingly relevant for companies to enhance their
environmental performance. A substantial part of the emissions responsible for climate change is caused by the generation of energy by the utility industry. In this paper the relationship between environmental performance and firm value in the utility sector in Europe and the United States is investigated by using multiple regression analysis. Results show that environmental performance has a small negative influence on firm value. This supports Neoclassical Economic theory, which states that increasing environmental performance leads to higher costs. The social implication of these results is that firms can increase their environmental performance, without losing much of their firm value. If governments were to provide small subsidies, this might be sufficient to compensate utility firms for this loss in firm value.
Table of Contents
1. Introduction ... 2
2. Literature Review ... 5
2.1. Theoretical Framework ... 5
2.2. Empirical Evidence ... 6
2.3. Summary and Hypothesis ... 10
3. Methodology and Data ... 11
3.1. Methodology ... 11
3.2. Data and Descriptive Statistics ... 13
4. Analysis ... 14
4.1. Empirical Results ... 14
4.2. Robustness Check... 16
5. Conclusion and Discussion ... 21
References ... 23
1. Introduction
Environmental concerns are becoming more and more relevant in our current society. The climate is changing and countries worldwide have set up an agreement to tackle this problem, the Paris Agreement. This agreement became effective on 4 November 2016. It’s central aim is to keep the global temperature rise of this century below 2 degrees Celsius (UNFCC, 2016). To reach this target, several measures need to be taken, among which a reduction in the emission of Green House Gasses.
One of the largest sources of greenhouse gas emissions is the utility sector, which, according to the United States Environmental Protection Agency (2015), accounted for about 29 percent of U.S. greenhouse gas emissions in 2015. A study by Climate Analytics (2017) found that in order to meet the commitments made in the Paris Agreement, the majority of coal-fired power plants in the European Union need to be closed within 15 years. An alternative source of energy could be renewable energy. Utility firms would have to increase their renewable energy production while decreasing coal-fired energy
production. This transition could be costly, but according to the Natural Resource Based Theory it could also present firms with opportunities for a competitive advantage (Hart, 1995). If this is the case, it could encourage firms to make environmental efforts. It is therefore interesting to investigate how
increasing environmental performance influences a utility firm’s financial performance.
A lot of research has already been done regarding the relationship between environmental performance and firm value in general, but it has been relatively inconclusive. Therefore, the focus has been shifted from ‘Does it pay to be green?’ to ‘When does it pay to be green?’ (Albertini, 2013). The aim of this new direction is to identify in which circumstances it does or does not pay to be green. The utility industry has not been extensively studied thus far, so
determining whether it pays to be green in this industry would contribute to this new direction of research.
This paper will try to find an answer to the question of how
utility industry using 93 electric, gas, water and multi-utility companies in the United Stated and Europe. A list of the companies used in this sample is given in Appendix 5. Europe and the United States are viewed as one population in this study. Both western countries have highly individualistic cultures (Heine, 2011), are democratic and maintain a free market economy. From the 1970’s,
environmental protection became a priority for both the U.S. and Europe, and between 1970 and 1995 major environmental laws came into effect (Albertini, 2013). It is expected that the relationship between environmental performance and financial performance in these two regions is similar. A meta-analysis by Albertini (2013) confirms that both regions show on average a positive relationship between environmental performance and financial performance, although the relationship appears to be somewhat stronger for U.S.-Canada than for Europe. Considering the fact that there are also differences between the two regions, for example the exact specifications of environmental laws, some divergence in the results can be expected. Therefore, a sub-sample analysis will be performed, carrying out two separate regressions for the U.S. sample and the European sample.
A time span of 5 years will be applied, using data from 2010-2015. It can be argued that before this period, the public view of environmental
responsibility was inherently different. Until ten years ago, the primary focus of corporate managers was the maximization of shareholder profits (Kim, Park, & Ryu, 2017). Now, because of increasing public concern for the sustainability of their business methods, environmental responsibility has become an important aspect of their job. The financial crisis of 2008-2009 and the period leading up to the crisis are excluded from the analysis, because the relationship between environmental performance and financial performance was found to be significantly different in times of crisis (Gallego-Álvarez, García-Sánchez, & da Silva Vieira, 2014). During the crisis period, a synergy was found between environmental and financial performance, which implies that, in times of
economic crisis, companies that care about Corporate Social Responsibility have higher financial performance.
To perform this research, data about the environmental and financial performance of the selected firms is needed. Environmental performance is
operationalized by the Environmental Score variable from Thomson Reuter’s Asset4 database, available in Datastream, which is based on multiple facets of environmental performance. Firm performance is measured by Tobin’s q, which is calculated by dividing market value by asset value. Multiple regression will be used to model the relationship. Control variables will be added for size, yearly time trend and continent.
In previous empirical research a wide variety of measures for
environmental performance and financial performance have been used. The variable Environmental Score that is used to measure environmental
performance in this paper, which is based 70 key performance indicators, has not been used in empirical research in the utility sector thus far. Similar
variables have been used in research regarding other industries and it has been found that this type of variable more strongly and more positively predicts the relationship between environmental performance and financial performance (Albertini, 2013).
The theoretical implications of this study might be further confirmation of one of both theories underlying the relationship between environmental
performance and financial performance. A positive relationship provides support for the Natural Resource Based View (Hart, 1995), which argues that increased environmental performance can lead to opportunities for a
competitive advantage. A negative relationship supports the Neoclassical Economic View, which states that the higher costs associated with increasing environmental performance leads to a decrease in firm value. Social implications of this study might be an increase in awareness of the consequences of
environmental performance for firm value, regarding firms in the utility sector. Should there be a positive relationship, this might be an incentive for utility firms to increase their environmental performance. If a negative relationship is found, governments must rely on regulations such as the Paris Agreement or provide subsidies to encourage utility firms to increase their environmental
performance.
This paper is organized as follows. Chapter 2 will provide a theoretical framework for the research question and elaborate on earlier empirical research. Chapter 3 will give information about the data and methodology. In Chapter 4
the main analyses and a robustness check are performed. Conclusions are given in Chapter 5, together with a discussion of the implications of this study.
2. Literature review
2.1 Theoretical Framework
The central theory in the field of environmental economics is the Natural Resource Based View (NRBV). This theory was developed by Hart (1995) to incorporate the natural environment into the already existing Resource Based View. The Resource Based Theory (RBT) identifies firm resources and
capabilities that are costly to copy as key sources of competitive advantage. For a resource to contribute to a sustained competitive advantage, it must be valuable, rare, imperfectly imitable and non-substitutable (Barney, 1991). An increasingly important drawback of this theory is the exclusion of the ecological environment. The Natural Resource Based View, as formulated by Hart (1995), is “a theory of competitive advantage based upon the firm’s relationship with the natural environment.” NRBV is composed of three interconnected strategies: pollution prevention, product stewardship and sustainable development, and predicts two types of competitive advantage: differentiation advantage and cost advantage (Hart & Dowell, 2011). Pollution prevention seeks to prevent
emissions and waste instead of cleaning it up at the end, which is the case with pollution control , and is associated with lower costs because of increased efficiency. Product stewardship elaborates on this by including the products “life cycle”, this involves reducing the ecological impact and use of resources
throughout the firms entire value chain. This could lead to a competitive
advantage by setting industry standards or by strategic pre-empting, for example securing exclusive access to resources. Sustainable development is much
broader and includes social and economic concerns. Regarding product
stewardship and sustainable development there is not much research available, although there seems to be a positive link between the implication of product stewardship and the inclusion of external stakeholders (Hart & Dowell, 2011). Which, in its turn, was found to benefit waste reduction management and energy conservation programs (Sharma and Vredenburg, 1998). As to pollution
prevention, more research is available. According to the NRBV, pollution prevention can lead to a competitive advantage, because it depends on the tacit skill development of employees who are involved in continuous-improvement efforts, which makes it difficult to observe and therefore not easy to duplicate (Hart, 1995). Consequently, this theory predicts environmental performance to create opportunities for competitive advantage, and hence to have a positive influence on firm value.
Another theory related to this subject is Neoclassical Economic Theory. From a Neoclassical point of view, higher environmental performance would lead to higher costs, unproductive investments and a possible loss of competitive advantage, and thus lower financial performance (Walley & Whitehead, 1994). Palmer, Oats and Portney (1995) explain that, although investments in
environmental performance might be beneficial from a societal standpoint, purely from a financial point of view it can be really expensive. Walley and Whitehead (1994) argue that opportunities for win-win situations of
environmental and financial performance, as suggested by Porter and van der Linde (1995), rarely exist. In most cases, responding to environmental
challenges is costly and it hardly ever pays itself back, let alone yields profits. Neoclassical economic theory predicts that high environmental performance is likely to decrease firm value.
The contradiction between these two theories is an important debate in this field of research. There has been a lot of research regarding the relationship between environmental and financial performance over the past decades, but no conclusion has been reached yet about the direction of this relationship.
2.2 Empirical Evidence
A lot of research has already been done regarding the relationship between environmental performance and firm value, but it has been relatively inconclusive. In a meta-analysis of 52 studies over a 35 year period, starting in 1975, Albertini (2013) finds that slightly more than half the surveyed studies show a positive relationship between environmental performance and firm value. The results of different studies are influenced by multiple factors, among which differences in research methodologies, data collection and analysis,
sampling size, and industrial context. Moreover, the relationship between environmental performance and firm value seems to vary per region. The average positive correlation between environmental performance and financial performance appears to be stronger for the U.S.-Canada than for Europe, and even stronger for the rest of the world (Albertini, 2013). A possible explanation for this is that environmental regulations in Europe are more strict than in the U.S., whereas flexible environmental regulations provide greater incentives to innovate.
The variables used to measure environmental performance also influence the results of the study. Albertini distinguishes three different categories of measurement variables: Environmental Management Variables (EMV’s), Environmental Performance Variables (EPV’s), and Environmental Disclosure variables (EDV’s). EMV’s are measures of a firm’s environmental management processes, and attitudes and objectives towards environmental responsibility. This includes strategies such as pollution prevention and product stewardship, which are part of the NRBV (Hart, 1995). EPV’s are monetary and physical measures of environmental performance, for example the volume of renewable energy produced or the decrease of toxic waste and emissions. EDV’s can be environmental awards or announcements, such as newspaper articles that signal good (or bad) environmental performance. Albertini (2013) finds that EMV’s are stronger moderators in the relationship between environmental and financial performance and that using EMV’s leads to a more positive relationship than when EPV’s are used. This finding is in line with the predictions made by the NRBV (Hart, 1995). While environmental performance represents the
elimination of waste and inefficient use of resources, which can result in lower costs, environmental management goes beyond this and includes the views of management regarding the environment and strategies that can cover the entire value chain of the company. This not only results in lower costs, but can also lead to another type of competitive advantage: differentiation advantage.
Evidence in favour of the NRBV is provided by King and Lenox (2002), whose study confirms that pollution prevention can lead to financial gain. Pollution control, on the other hand, which means that pollution is cleaned up
end-of-pipe, was not linked to increased profits. This is because pollution prevention requires a firm to make environmental investments and therefore creates opportunities for a competitive advantage, whereas pollution control only fights the symptoms. Factors that influence the effect of a pollution prevention strategy on firm value have also been identified: organizational capabilities and managerial cognition (Hart & Dowell, 2011). Firstly, firms are more likely to profit from pollution prevention if they possess strong innovation capabilities, in particular continuous improvement (King & Lenox, 2002). The findings suggest that managers tend to underinvest in pollution prevention. Secondly, the profitability of environmental strategies is influenced by
managerial attention and the framing of environmental issues (Hart & Dowell, 2011). If managers believe profitable opportunities exist, they are more likely to profit from them (King & Lenox, 2002). Sharma and Vredenburg (1998) confirm that firms who regard environmental responsiveness as an instrument to
increase shareholder value, are more likely to take a proactive environmental viewpoint.
Several studies support this positive relationship between environmental performance and firm value (Aragón-Correa, Hurtado-Torres, Sharma, & García-Morales, 2008; Karagozoglu and Lindell, 2000; Wahba, 2008). For example, Klassen and McLaughlin (1996) investigate the effect of environmental management on stock prices. Environmental performance is measured using Environmental Disclosure Variables. They find that weak environmental performance, indicated by environmental crises, led to negative abnormal returns, and that strong environmental management, indicated by
environmental awards, led to positive returns. The positive effect was smaller for firms in environmentally dirty industries. This study shows that the market values environmental management.
Konar and Cohen (2001) also find a positive relationship between environmental performance and firm value, using publicly traded firms in the S&P 500. Firm value was measured by Tobin’s q. Environmental
(mal)performance was measured in two ways: by the aggregate pounds of toxic chemicals emitted per dollar revenue of the firm, and by the number of
showed that both measures of environmental (mal)performance were
statistically significant and have a negative impact on Tobin’s q. The results of this study give evidence for a negative effect of poor environmental performance on intangible-asset value for publicly traded firms in the S&P 500.
Support for the Neoclassical view of environmental performance is provided by Sueyoshi and Goto (2009), who investigate the implications of the U.S. Clean Air Act which was implemented in the electric utility industry in 1995, using panel data of 167 firms from 1989 to 2001. They conclude that the U.S. Clean Air Act negatively influenced firm financial performance, measured by Return on Assets (ROA). They explain that this could be caused by higher capital costs in the short term, which makes it difficult for utility firms to attain both financial performance and environmental performance at the same time. However, the relationship between environmental performance and financial performance of a firm is found to be influenced by the supportiveness of environmental regulations (Karagozoglu and Lindell, 2000; Albertini, 2013). When regulations are flexible and provide the right incentives, it could help firms attain an environmental competitive advantage. Thus, perhaps these findings are not applicable to all environmental regulations. On the other hand, the study found that, on a long-term horizon, the total amount of accumulated facility investments for environmental protection might positively relate to financial performance, but this was not statistically proved.
Another study that finds a negative relationship between environmental and financial performance for the utility industry was performed by Ruggiero and Lehkonen (2017), who investigate the effect of renewable energy production on firm value for 66 large electric utility companies over the period 2005-2014. They measure firm value in three ways: Return on Equity (ROE), ROA and Tobin’s q. The main explanatory variable is the volume of renewable energy. Control variables for size, risk, capital intensity, growth, carbon intensity and yearly time trend were added to the regression equation, of which size, risk and yearly time trend were found to be significant. The analysis shows a negative relationship between the volume of renewable energy and firm value, for all three dependent variables. An explanation for this might be retrieved from
Neoclassical Economic Theory, meaning that increasing the volume of renewable energy leads to an increase in capital costs, which has a negative impact on profits. However, in this study, the variable used to measure environmental performance is the volume of renewable energy. This variable could be
categorized as an Environmental Performance Variable (EPV) (Albertini, 2013) and might not be the only relevant factor in a firms environmental footprint. An EPV only measures a physical aspect of environmental performance, whereas an EMV would measure a firm’s attitudes toward and use of environmental
opportunities. When using this type of variable, the relationship could look differently, because firms are scored on their overall environmental
performance, instead of a single aspect.
There are also studies that find neither a positive, nor a negative
relationship between environmental performance and financial performance. Yu, Ting, and Wu (2009) find no statistically significant relationship between firm environmental performance and financial performance for European companies across different industries, using three measures for environmental performance and several measures for financial performance, including ROA, ROE and
Earnings per Share (EPS). Similar results are found by Gallego-Álvarez et al. (2014). They investigate the relationship between environmental performance and financial performance for 855 international companies from 2006-2009 in sectors of intensive greenhouse gas emissions. Financial performance is
measured using ROA and environmental performance is measured by the ratio of greenhouse gas emissions to volume of sales. The results show no influence of environmental performance on financial performance.
2.3 Summary and Hypothesis
All in all, the relationship between environmental performance and firm value seems ambiguous. It can vary depending on which industry and country are investigated, and which variables are used (Albertini, 2013). The NRBV predicts a positive relationship between environmental performance and financial performance, whereas Neoclassical economic theory predicts a negative relationship. Because of the empirically unsettled relationship between
environmental performance and financial performance, the focus has been shifted from ‘Does it pay to be green?’ to ‘When does it pay to be green?’
(Albertini, 2013). The aim is to find out in which specific situations it pays to be green. This paper hooks on to that by investigating whether it pays to be green for firms in the utility industry.
In this paper, the relationship between environmental performance and firm performance will be estimated for the utility industry, with a sample period of 2010-2015 and using firms from the U.S. and Europe. The previously
mentioned studies performed with regard to the utility sector show a negative relationship between environmental performance and financial performance. However, these studies use Environmental Performance Variables as a measure of environmental performance, whereas Albertini (2013) found that using an Environmental Management Variable leads to a more positive relationship. In this paper environmental performance will be measured using an EMV, therefore it is possible that a positive relationship between environmental and financial performance will be found.
Based on theory and empirical research, both a negative and a positive relationship seem plausible. The aim of this paper is to find out whether a negative or a positive relationship is supported by the data.
3. Methodology and Data 3.1. Methodology
To investigate the relationship between environmental performance and firm value, multiple regression will be used. The regression equation looks as follows:
𝑇𝑂𝐵𝐼𝑁𝑄𝑖𝑡 = 𝛽0+ 𝛽1(𝐸𝑁𝑉𝑆𝐶𝑂𝑅𝐸)𝑖𝑡+ 𝛽2(𝑆𝐼𝑍𝐸)𝑖𝑡+ 𝛽3(𝐸𝑈𝑅𝑂𝑃𝐸)𝑖𝑡 + 𝛽3(𝑇𝐼𝑀𝐸)𝑖𝑡+ 𝜀𝑖𝑡
Firm value will be assessed by using Tobin’s q, which is a measure of intangible firm value, this is in line with earlier research (Konar & Cohen, 2001; Ruggiero & Lehkonen, 2017; King & Lenox, 2002). Tobin’s q is computed by dividing the market value of a firm by the replacement value. If a firm has no
intangible asset value, the market value will equal the replacement value and Tobin’s q will be equal to 1. Tobin’s q will increase if the intangible asset value of the firm increases (Konar & Cohen, 2001). Since Replacement Value of a firm is difficult to measure, the book value of total assets is used, in accordance with earlier research (Ruggiero & Lehkonen, 2017). Hence, Tobin’s q is calculated by dividing the Market Value by the Total Assets. Both market value and total assets are found in Compustat.
The main explanatory variable, environmental performance, is measured using the variable Environmental Score from Thomson Reuters Corporate Responsibility Ratings (“TRCRR”), provided by Asset4, a leading global provider of Environmental, Social and Governance (ESG) data. This variable is defined as a measure of ‘a company's impact on living and non-living natural systems,
including the air, land and water, as well as complete ecosystems’ (Asset4 ESG Data Glossary, 2013). The variable displays to which extent a company uses environmental opportunities to generate long term shareholder value and avoids environmental risks. This variable could be categorized as an Environmental Management Variable (Albertini, 2013). The score is calculated based on 70 Key Performance Indicators (KPI’s), which can be divided into three main categories: emission reduction, product innovation and resource reduction (Thomson
Reuters, 2013). The variable is scored from 0 to 100 and measures
environmental performance relatively to other firms in an industry. Zero means that the firm has very poor environmental performance relative to its peers and 100 means that the firm scores exemplary. The measure is designed to have a mean and median close to 50.
Control variables
Firm size significantly influences the relationship between environmental performance and firm value (Ruggiero & Lehkonen, 2017). Size and Tobin’s q were found to be positively related, which means that bigger firms have a higher value of Tobin’s q. A control variable for firm size will thus be added to the regression model. In accordance to earlier research (Gallego-Álvarez et al., 2014; Sueyoshi & Goto, 2009) the proxy used for firm size is total assets, which is extracted from Compustat. To be able to include this variable in the model, it is
needed to convert the values of total assets for European companies to USD. The official average yearly exchange rates from the Worldbank (n.d.) are used for this. An overview of the exchange rates can be found in Appendix 1.
To control for the influence of country specific characteristics on Tobin’s q, a dummy control variable for continent is added. The variable EUROPE takes the value of one is the company is European and zero is the company is U.S. based.
Another variable that could influence the results is yearly time trend. To control for effects of macro-economic cyclicity, a variable is added for time trend. Dummies are used for the years 2011 to 2015, leaving 2010 as the reference year to avoid a dummy trap.
Hypotheses
Since there are arguments for both a negative and a positive relationship between environmental performance and firm value in this sector, the aim of this paper is to find out which direction is supported by the data. The statistical hypotheses are: 𝐻0: 𝛽𝐸𝑛𝑣𝑠𝑐𝑜𝑟𝑒 = 0
𝐻1: 𝛽𝐸𝑛𝑣𝑠𝑐𝑜𝑟𝑒 ≠ 0
3.2. Data and Descriptive Statistics
The data for the explanatory variable has been acquired from Datastream. Datastream provides ESG data from Asset4, a Thomson Reuter business. This data is available for over 4 600 public companies worldwide, divided over 52 industries (Thomson Reuter, 2013). The data was selected by searching in the constituent lists of the Asset4 database and selecting the industry categories ‘electricity’ and ‘gas, water, and multiutilities’. This procedure has been followed for both the U.S. and Europe. The time period of 2010-2015 was indicated. For the companies available, the Asset4 ESG variable ‘Environmental score’ was downloaded to an excel file. For all these companies the Global Company Key was downloaded and this variable was used to look up the matching financial data from Compustat. From this database, the market value, debt, equity and total asset value were downloaded. This was extracted as an excel file and copied
and pasted into the existing excel file using the Global Company Key to match it with the environmental data.
The file contained a total of 549 data points for 93 companies. Before performing the regression, the data was checked for errors. A scatter plot, as can be seen in Appendix 2, Plot 1, showed that missing values of ENVSCORE were treated as if they have a value of 100. This could lead to a bias in the results. Therefore, the missing values were deleted. The dependent variable, Tobin’s q, is calculated by dividing market value by total assets, missing one of those
variables can lead to a biased q. Therefore, missing values were deleted from the sample. This resulted in a final database of 454 data points for 93 companies, 39 of which are European companies. A list of the companies in this sample can be found in Appendix 5.
After the deleting missing values, a scatterplot was made, as can be seen in Appendix 2, Plot 2, which showed that there were no big outliers. Descriptive statistics are given in Table 1. The mean Tobin’s q is .545, which by definition means that the companies in this sample are undervalued (Konar & Cohen, 2001). Since the measure is designed to have a mean of 50, the mean environmental score of 67.411 means that the sample scores higher than average.
Table 1, Descriptive Statistics
4. Analysis
4.1. Empirical Results
Regression results are presented in Table 2. Firstly, a regression is run with only the explanatory variable, to be able to observe how the coefficient changes when control variables are added. The coefficient ENVSCORE in the first regression (1) is negative but not significant. In regression (2) and (3) control
N Mean SD Minimum Maximum
Tobin’s q 454 .545 .179 .068 1.252
Environmental Score 454 67.411 25.700 9.6 94.57 Total assets in USD 454 39084.89 55571.79 150.449 355423.1
variables for size and continent are added respectively. Both variables are significant at a 1% level. However, the coefficient of size is estimated to be close to zero and does not affect the value of the coefficient of ENVSCORE by much, although it changes the direction of the relationship, and does not add any significance to this coefficient. Adding a control variable for continent to regression (1) makes the coefficient of ENVSCORE significant at a 1% level. Adding both EUROPE and SIZE, as in regression (4), increases the value and the significance of the coefficient of ENVSCORE with respect to regression (1) and (2), but it is less significant than in regression (3). In regression (5), dummy variables are used to control for yearly time trend. Of these dummy variables, only the coefficient of year 2014 turns out to be significant at a 10% level. Adding all control variables to the regression, as in model (6), increases the significance of the yearly time trend for 2014 to 5%.
Table 2. Regression results
Robust standard errors in parentheses. ***, ** and * refer to 1%, 5% and 10% significance levels
It can be noted that, whenever it is significant, the coefficient of ENVSCORE is negative and fairly close to zero. This implies that the
environmental performance of a utility firm does not have very much effect it’s firm value, measured by Tobin’s q. Using regression (6), an increase in
environmental score of 1 results in a decrease of Tobin’s q of 0.0007. The (non-existing) average firm in this sample has a Tobin’s q of .545 and an
environmental score of 67.411. An increase in environmental score to 68.411 results in a Tobin’s q of .5443, which equals a decrease of 0.13 percent. This may be statistically significant, but the effect is not very economically significant.
Controlling for size does not seem to have a large effect. The coefficient of SIZE is significant, but extremely close to zero in all regressions. This could be caused by presence of total assets in the variable Tobin’s q. Because the dependent variable is a ratio, it is already to some extent scaled to the size of a firm. Although this variable is statistically significant in all regressions, it can be argued that it is not economically significant in this model, because it is
extremely close to zero.
The coefficient EUROPE is significant and relatively big in all regressions. It accounts for the majority of the explained variance, as 𝑅2 becomes much larger when this coefficient is added to the regression. Adding EUROPE to the
regression also makes the coefficient of ENVSCORE significant at a 1% level, whereas without EUROPE it is insignificant. This implies that part of the variance in Tobin’s q is explained by the continent the company is based in. Apparently, country specific characteristics play a big role in the determination of Tobin’s q.
The yearly time trend does not seem to have very much effect on Tobin’s q, since only the year 2014 was significant. This does not correspond to earlier research. However, earlier studies use sample periods including the financial crisis, whereas the sample period of this paper, 2010-2015, was relatively stable. This could explain the insignificance of the time trend control variables. The results of this thesis suggest that yearly time trend might not be a relevant variable in explaining Tobin’s q in this industry for all time periods.
4.2 Robustness Check
4.2.1 Assumptions of the Model
Before a multiple regression can be performed, the assumptions of the model need to be verified. Only then can it be statistically justified to use this model. In this section the five regression assumptions will be tested. These five
regression assumptions are: linearity, normality, independence, heteroscedasticity, and multicollinearity.
The assumption of linearity is tested by analysing two graphs: the P-P plot and the residuals scatter plot, which both can be found in Appendix 3. The
standardized residuals in the P-P plot follow a linear pattern. Naturally, because the standardized normal distribution is not uniform, the first half of the graph shows a concave pattern, whereas the latter half of the graph shows a more convex pattern. The residuals scatter plot does not give any reason to doubt the validity of the linearity assumption, because no pattern in the residuals
themselves can be detected.
Normality is assessed by graphing the residuals in a histogram. It can clearly be seen, in the graph in Appendix 4, that the residuals approximately follow a normal distribution with 𝜇 = 0 and 𝜎 = 1. The assumption of normality has been met, because the residuals are approximately normally distributed with a mean of -0.03 and a standard deviation of 1.021, which implies a near perfect standard normal distribution.
The independence assumption implies that autocorrelation between the residuals, as a first order effect, is not present. In this model this is especially relevant and important, because data from several years - a time series - is used. The Durbin-Watson statistic is 1,778, which means the independence
assumption is not violated, because a result between 1,5 and 2,5 implies no autocorrelation.
Inequality of variances or heteroscedasticity is not present in the model. As can be seen in the residuals scatter plot, the variance does not significantly decrease or increase. Additionally, because heteroscedasticity robust standard errors are used, this assumption is no longer relevant.
Multicollinearity is tested by analysing the Variance Inflation Factor (VIF) of each distinct independent variable. To calculate the VIF factor each
independent variable is regressed as dependent variable, with all the other variables as independent variables. Whenever the R-squared becomes
significantly high, the VIF factor will increase. This means multicollinearity is present, because one independent variable explains another independent variable in the model. For this sample, the VIF factor is smaller than 1.2 for each
dependent variable, which means that multicollinearity is not present in the model.
4.2.2 Subsample analysis
To examine whether the relationship between environmental
performance and firm value is different between the United States and Europe, separate regressions are performed. The sample is divided into two separate samples, one for the U.S. and one for Europe. First some descriptive statistics of both continents are shown in Table 3 and Table 4.
Table 3, Descriptive Statistics United States.
Table 4, Descriptive Statistics Europe.
N Mean SD Minimum Maximum
Tobin’s q 220 0.629 0.136 0.117 0.866
Environmental Score 220 76.305 22.573 11.76 94.57 Total assets in USD 220 51190.24 74227.06 2692.068 355423.1
It can immediately be noted that all variables have higher means in the European sample compared to the United States sample. This difference in environmental performance is possibly caused by the leading position of Europe in
environmental regulations (Yu, Ting, & Wu, 2009). The more stringent regulations may have driven European companies to increase their
environmental performance further than U.S. based companies. It is not clear why European companies also have higher asset values and Tobin’s q. It could be the case that, since both environmental score and Tobin’s q are higher, the relationship between environmental performance and financial performance is the same for Europe and the U.S.., and the European market values the relatively
N Mean SD Minimum Maximum
Tobin’s q 234 0.466 0.180 0.068 1.252
Environmental Score 234 59.049 25.696 9.6 94.08 Total assets in USD 234 27703.8 23602.74 150.449 121156
high environmental performance. Of course, since Tobin’s q can be influenced by many things, there are plenty of other possible explanations for this difference.
To test if the relationship between environmental performance and firm value is different for the United States and Europe, two separate regressions are performed. The regressions for the United States are shown in Table 5 and the regressions for Europe are shown in Table 6. The first regression (1) includes only the main explanatory variable, ENVSCORE. The second regression (2) adds the control variable size, and the third regression (3) adds control variables for yearly time trend to model (1). In regression (4), all variables are added to the model.
Table 5. Subsample regression results for the United States.
Robust standard errors in parentheses. ***, ** and * refer to 1%, 5% and 10% significance levels
In Table 5, it can be observed that the coefficient ENVSCORE is negative and significant in regression (1), at 5% level. The explained variance in this regression is very low. Adding the control variable size to the regression increases the explained variance, but the coefficient of ENVSCORE becomes insignificant. This results in size predicting Tobin’s q, instead of environmental performance. When control variables for yearly time trend are added to the
model, in regression (3), the value and significance of the coefficient of
ENVSCORE remain stable with regard to regression (1). The yearly time trend variables 2013 and 2014 are significant, at 10% level and 1% level respectively. Adding these variables triples the explained variance in regression (3) with respect to regression (1). In a model with all variables, as in regression (4), the coefficients of the control variables are significant, but that of ENVSCORE is not. This coefficient is only significant when no control variables are added to the model, as in regression (1). The insignificance of the coefficient of ENVSCORE in other models implies that there is no relationship between environmental performance and firm value for U.S. based firms.
Table 6. Subsample regression results for Europe
Robust standard errors in parentheses. ***, ** and * refer to 1%, 5% and 10% significance levels
In Table 6, the first thing that can be noticed is that the coefficient of ENVSCORE is negative and significant at 1% level in all four regressions. When the control variable firm size is added in regression (2), the explained variance more than doubles. This implies that, for European companies, the size of the firm explains a large part of the variance in Tobin’s q. Nevertheless, the
lower intangible firm value. The control variables for yearly time trend are not significant for the European sample, as can be seen in regression (3), which means that the variance in Tobin’s q cannot be explained by the yearly time trend. When in regression (4) all variables are added, this resembles regression (2). From these regressions it can be concluded that environmental performance negatively predicts firm value, although the effect does not seem to be very large.
The regression results are different for the U.S. and Europe. The coefficient of ENVSCORE is significant in all regressions using the sample of European companies, whereas it is only significant in regression (1) for the U.S. sample. Additionally, the explained variance is much higher in the regressions using the European sample.
5. Conclusion and Discussion
In this paper the relationship between environmental performance and financial performance in the utility sector was investigated using multiple regression analysis. Tobin’s q was used to measure financial performance and the Environmental Score variable from TRCRR was used to measure
environmental performance. The results of the analysis showed that environmental performance has a negative relationship with firm value.
However, the coefficient was very small, so although it is statistically significant, it can be argued that it has little economic meaning. The results of this study do not conform to the expectations based on the NRBV and empirical research. Instead, evidence is provided for the Neoclassical Economic Theory in this industry.
Because earlier empirical research has been relatively inconclusive, the focus has been shifted from ‘Does it pay to be green?’ to ‘When does it pay to be green?’. It seems that in the utility industry, investigated in this paper, the NRBV does not hold and hence it does not pay to be green. The mean Tobin’s q in this industry is fairly small, which indicates that firms are on average undervalued. Perhaps this means that the utility sector is not sensitive to changes in intangible asset value.
An explanation for the results is Neoclassical Economic theory. Considering that higher environmental performance typically results in
increased costs, it is likely to decrease firm value. Unlike the predictions of the NRBV, the increased environmental performance may not have contributed to a competitive advantage, or at least not to a great extent. The increase in cost may have outweighed any competitive advantages.
A limitation of this research is that there are not many control variables. Most likely, other variables influence Tobin’s q in this regression. Earlier
research into this industry identified some control variables that are significant. Those variables include firm size and yearly time trend, which were used for this model. The other significant control variable, risk, was measured as a ratio of debt to assets. The variable debt was not available for a large part of the U.S. firms in this sample. Therefore this control variable could not be added to the regression equation. It may be the case that some of the variance in the data can be explained by adding more control variables, and that it would make the effect of environmental performance on firm value more clear.
Another limitation of this research is that companies from the United States and Europe were treated as one sample, while there turned out to be quite some differences between both regions. Separating the sample into two
regressions, as put out in the subsample analysis, changed the coefficients extensively.
A suggestion for further research would be to investigate the nature of the difference between Europe and the United States. Staying on the topic of when it pays to be green, it could be interesting to identify which specific characteristics cause the variation between both regions. This could aid politicians and policy makers in their decisions regarding environmental strategies.
Social implications of this paper are that utility firms cannot benefit financially from an increase in their environmental performance. This implies that firms may need an external incentive to increase environmental
performance, for example government regulations or subsidies. Such a subsidy need not be large, because the results suggest that investing in environmental performance does not cause a big decrease in firm value.
References
Albertini, E. (2013). Does Environmental Management Improve Financial Performance? A Meta-Analytical Review. Organization & Environment, 26(4), 431–457.
Aragón-Correa, J. A., Hurtado-Torres, N., Sharma, S., & García-Morales, V. J. (2008). Environmental strategy and performance in small firms: A resource-based perspective. Journal of Environmental Management, 86, 88–103.
Asset4 ESG Data Glossary (2013). Retreived from:
https://uvalibraryfeb.wordpress.com/, at 31-5-2015.
Barney, J. B., Ketchen, D. J., & Wright, M. (2011). The future of resource-based theory: Revitalization or decline? Journal of Management, 37, 1299-1315. Climate Analytics (2017). A stress test for coal in Europe under the Paris
Agreement. Retrieved from: www.climateanalytics.org/publications. EPA (n.d.). Sources of Greenhouse Gas Emissions. Retrieved from:
https://www.epa.gov/ghgemissions/sources-greenhouse-gas-emissions, at: 08-06-2017.
Gallego-Álvarez, I., García-Sánchez, I. M., & da Silva Vieira, C. (2014). Climate Change and Financial Performance in Times of Crisis. Business Strategy and the Environment, 23, 361–374. DOI: 10.1002/bse.1786
Hart, S. L. (1995). A Natural-Resource-Based View of the Firm. The Academy of Management Review, 20(4), 986-1014.
Hart, S. L., & Dowell G. (2011). A Natural-Resource-Based View of the Firm: Fifteen Years After. Journal of Management, 37 (5).
Heine, S. J. (2011). Cultural psychology. New York, NY: WW Norton & Company. Karagozoglu, N. & Lindell, M. (2000). Environmental Management: Testing the
Win‐Win Model. Journal of Environmental Planning and Management, 43 (6), 817-829. DOI: 10.1080/09640560020001700
Kim, H., Park, K., & Ryu, D. (2017). Corporate Environmental Responsibility: A Legal Origins Perspective. Journal of Business Ethics, 140, 381–402. DOI 10.1007/s10551-015-2641-1
King, A. & Lenox, M. (2002). Exploring the Locus of Profitable Pollution Reduction. Management Science, 48 (2).
Klassen, R. D., & McLaughlin, C. P. (1996). The Impact of Environmental Management on Firm Performance. Management Science, 42 (8), 1199-1214.
Konar , S., & Cohen, M. A. (2001). Does the market value environmental performance? Review of Economics and Statistics, 83(2), 281–289.
Owen, P. (2017, June). Tracking Trump: president tosses Paris agreement in the trash. The Guardian. Retreived from: https://www.theguardian.com/us-news/2017/jun/03/donald-trump-news-paris-agreement-covfefe-fbi, at 05-06-2017.
Palmer, K., Oates, W. E., & Portney, P. R. (1995). Tightening Environmental Standards: The Benefit-Cost or the No-Cost Paradigm? Journal of Economic Perspectives, 9(4), 119–132.
Porter, M. E., & van der Linde, C. (1995). Toward a New Conception of the Environment-Competitiveness Relationship. Journal of Economic Perspectives, 9(4), 97–118.
Ruggiero, S., & Lehkonen, H. (2017). Renewable energy growth and the financial performance of electric utilities: A panel data study. Journal of Cleaner Production, 142, 3676-3688.
Sharma, S. & Vredenburg, H. (1998). Proactive Corporate Environmental Strategy and the Development of Competitively Valuable Organizational
Capabilities. Strategic Management Journal, 19(8), 729-753.
Sueyoshi, T., & Goto, M. (2009). Can environmental investment and expenditure enhance financial performance of U.S. electric utility firms under the clean air act amendment of 1990? Energy Policy, 37, 4819–4826.
Thomson Reuters (2013, august). Thomson Reuters Corporate Responsibility Ratings (TRCRR): Rating and Ranking, Rules and Methodologies. Retreived from: https://www.thomsonreuters.com, at 01-06-2017. UNFCC (2016). The Paris Agreement. Retrieved from:
http://unfccc.int/paris_agreement/items/9485.php, at: 11-05-2017. Wally, N., & Whitehead, B. (1994). It's Not Easy Being Green. Harvard Business
Review, 72(3), 46-51.
Wahba, H. (2008). Does the market value corporate environmental
and Environment Management, 15, 89-99.
Worldbank (n.d.). Official exchange rate (LCU per US$, period average). Retreived from:
http://data.worldbank.org/indicator/PA.NUS.FCRF?end=2016&locations =DK-CZ-CH-PL-GB-XC&start=2010&view=chart. At: 12-06-2017.
Yu, V., Ting, H. I., & Wu, Y.-C. J. (2009). Assessing the greenness effort for
European firms : A resource efficiency perspective. Management Decision, 47, 1065-1079.
Appendix 1
Official exchange rate (Local Currency per USD, period average)
Country Currency 2010 2011 2012 2013 2014 2015 Euro Zone EUR 0.755 0.719 0.778 0.753 0.754 0.902 United Kingdom GBP 0.647 0.624 0.633 0.640 0.608 0.655 Switzerland CHF 1.043 0.888 0.938 0.927 0.916 0.962 Czech Rebublic CZK 19.098 17.696 19.578 19.571 20.758 24.599 Denmark DKK 5.624 5.369 5.792 5.616 5.612 6.728 Poland PLD 3.015 2.963 3.257 3.161 3.155 3.770 Source: Worldbank (n.d.)
Appendix 2 Plot 1
Appendix 5
Company legal name City Web URL
A2A SPA Brescia www.a2a.eu
Acea SPA, Roma Rome www.acea.it
AES Corporation (The) Arlington www.aes.com
Alpiq Holding AG Lausanne www.alpiq.com
Aqua America Inc Bryn Mawr www.aquaamerica.com
Areva SA Courbevoie www.areva.com
Atlantica Yield plc Brentford www.atlanticayield.com
Avangrid Inc New Haven www.avangrid.com
Avista Corp Spokane www.avistacorp.com
California Water Service Group San Jose www.calwatergroup.com
Calpine Corp Houston www.calpine.com
CenterPoint Energy Inc. Houston www.centerpointenergy.com
Centrica PLC Windsor www.centrica.com
Cez A.S., Praha Prague www.cez.cz
CMS Energy Corp Jackson www.cmsenergy.com
Consolidated Edison Inc. New York https://www.coned.com/en Consolidated Water Co. Ltd Grand
Cayman www.cwco.com
Dominion Energy Inc Richmond www.dom.com
DONG Energy AS Fredericia www.dongenergy.com
Drax Group Selby www.drax.com
DTE Energy Co Detroit www.dteenergy.com
Duke Energy Corp Charlotte www.duke-energy.com
Dynegy Inc. Houston www.dynegy.com
E.ON SE Essen www.eon.com
EDF Paris www.edf.fr
Edison International Rosemead www.edison.com
Edison SPA Milan www.edison.it
EDP Renovaveis SA Oviedo www.edpr.com
EDP-Energias de Portugal SA Lisbon www.edp.pt
El Paso Electric Co El Paso www.epelectric.com Elia System Operator SA Brussels www.elia.be
Enagas SA Madrid www.enagas.es
Endesa SA, Madrid Madrid www.endesa.com
ENEA SA Poznan www.enea.pl
Enel Ente Nazionale Per
L'Energia Elettrica SPA, Roma Rome www.enel.com
ENERGA SA Gdansk www.grupa.energa.pl
Engie SA Courbevoie www.engie.com
Entergy Corp. New Orleans www.entergy.com
Eversource Energy Springfield www.eversource.com
FirstEnergy Corp. Akron www.firstenergycorp.com
Fortum Corp Espoo www.fortum.com
Gas Natural Fenosa Barcelona www.gasnaturalfenosa.com Great Plains Energy Inc Kansas City www.greatplainsenergy.com Hawaiian Electric Industries Inc. Honolulu www.hei.com
Hera SPA, Bologna Bologna www.gruppohera.it
Iberdrola SA, Bilbao Bilbao www.iberdrola.com
IDACORP Inc. Boise www.idacorpinc.com
MGE Energy Inc Madison www.mgeenergy.com
Mvv Energie AG, Mannheim Mannheim www.mvv-energie.de National Fuel Gas Co Williamsville
New Jersey Resources Corp Wall www.njresources.com NextEra Energy Inc Juno Beach www.nexteraenergy.com
NiSource Inc. Merrillville
Northwest Natural Gas Co Portland www.nwnatural.com
NorthWestern Corp Sioux Falls www.northwesternenergy.com
NRG Energy Inc Princeton www.nrg.com
ONE Gas Inc Tulsa www.onegas.com
ONEOK Inc. Tulsa www.oneok.com
Ormat Technologies Inc Reno www.ormat.com
Pennon Group PLC Exeter www.pennon-group.co.uk
PG&E Corp San Francisco www.pgecorp.com
Pinnacle West Capital Corp Phoenix www.pinnaclewest.com PNM Resources Inc. Albuquerque www.pnmresources.com Polska Grupa Energetyczna SA Warsaw www.gkpge.pl
Portland General Electric Co Portland www.portlandgeneral.com
PPL Corp Allentown www.pplweb.com
Public Power Corp S.A. Athens www.dei.gr Public Service Enterprise Group
Inc Newark www.pseg.com
Red Electrica Corp SA Alcobendas www.ree.es
Rubis SA, Paris Paris www.rubis.fr
Rwe AG Essen www.rwe.com
SCANA Corp Cayce www.scana.com
Sempra Energy San Diego www.sempra.com
Severn Trent PLC Coventry www.severntrent.com
Snam SPA San Donato
Milanese
www.snam.it
South Jersey Industries Inc. Folsom www.sjindustries.com
Southern Co (The) Atlanta www.southernco.com
Southwest Gas Holdings Inc Las Vegas www.swgasholdings.com
Spire Inc St. Louis www.spireenergy.com
Suez SA Paris www.suez-environnement.com
Tauron Polska Energia SA Katowice www.tauron-pe.pl
TerraForm Power Inc Bethesda www.terraformpower.com
UGI Corp King Of
Prussia www.ugicorp.com United Utilities Group PLC Warrington www.unitedutilities.com
Vectren Corp Evansville www.vectren.com
Veolia Environnement, Paris Paris www.veolia.com
Verbund AG Vienna www.verbund.com
WEC Energy Group Inc Milwaukee www.wecenergygroup.com
Westar Energy Inc. Topeka www.westarenergy.com
WGL Holdings Inc. Washington www.wglholdings.com Xcel Energy Inc. Minneapolis www.xcelenergy.com
