An analysis of the effect of environmental policy on firms
with high ESG ratings; Does the nature of an environmental
policy matter?
Lorenzo Pizzi
1. Introduction
With modern environmental policies starting to come up in the 1960s, the last few decades have seen a great development in the care for the environment by citizens, states and also firms. People have started to pay more attention to what they spend their money on or in which firms they invest it in, not only considering purely financial aspects, but instead also accounting for environmental ones, such as how sustainable a firm’s activities are or how much pollution they generate. For this reason, how a company organizes its activities with respect to the environment is becoming increasingly important for its performance on the stock market, which in turn can dictate whether or not the firm will be able to survive and thrive in the long term.
This thesis studies the effect environmental policy announcements have on ESG-rated companies, which are the ones that are considered most highly in terms of their attitude towards environmental, social and governance issues. More specifically, I studywhether they experience significant increases in their stock prices around announcements of such policies. The focus is on German ESG-rated firms and their reactions to both national and international environmental policy announcements. While conventional firms may react negatively to such policy announcements if they limit their freedom, high ESG-rated firms may not be as heavily affected, or may even benefit from them.
The health of the environment has become an increasingly relevant topic in the past few decades. As Renneboog et al (2008) suggest, environmental disasters such as the explosion of the nuclear power plant in Chernobyl of 1986 and the Enron oil spill of 1989 have highlighted the harmful effects that industrial development has had, and can continue to have, on the environment. This had led especially the younger generations to consider other factors apart from financials when investing: a 2017 report by Ernst and Young found that millennials are “nearly twice as likely to invest in companies or funds that target specific social or environmental outcomes”. Quite literally, they are investing in a way that will help them and their descendants have a world to live in.
This is relevant when considering the impact of environmental policy on high ESG-rated companies. If such policies benefit socially responsible firms, it could provide a financial incentive, apart from the existing ethical one, for them to be implemented. Intuitively, firms with production processes harmful to the environment do not benefit from environmental policy. Green policies would then not only discourage environmentally harmful practices, but also give a competitive edge to ESG-rated firms: this would benefit responsible investors and likely encourage others to divest from conventional companies into responsible ones unless they changed their production practices. In essence, if green policies did in fact positively
affect ESG-rated firms, one could support their implementation not just via non-financial arguments (it is good for the environment), but also via financial ones (ESG-rated firms greatly benefit), which would make it easier to justify the losses incurred by conventional companies. Furthermore, it would provide more reasons for a conventional company to become more environmentally friendly, as green policies that would usually harm them would instead boost their returns.
In general, there is some evidence that environmentally responsible firms outperform non-environmentally responsible ones. The literature on Socially Responsible Investing (SRI), which consists of investing by considering not only financial metrics, but also ones of environmental, social and corporate responsibility, provides mixed results (Galema et al. 2008; Derwall et al. 2011; Renneboog et al. 2008).
In addition, empirical literature and research present mixed evidence on whether or not environmental regulation creates or destroys value for firms. As Ramiah et al (2013) argue, firms in different sectors may very well react differently to announcements of green policy. For these reasons, I believe the topic at hand to be quite relevant and interesting. In Germany, there is a “high public concern for environmental policies”, which explains why they “…remain some of the most progressive in the world” (Bailey, 2007, p. 536-537): understanding how national and international policies affect the country’s ESG-rated firms could provide useful insight into whether these firms should be emulated by others. It would also be beneficial to understand how environmental policies and the recent wave of SRI interact with each other. I focus on several environmental policies, both national (the Second Leak and Official Draft of the German Climate Action Plan 2050, and Germany’s 2019 Climate Package) and international (the Kyoto Protocol and the Paris Agreement). To test whether these policies were actually strict enough to be able to cause a difference between high-ESG-rated and other firms, I first analyse abnormal returns of firms in highly polluting industries in general: if the stock returns of such firms are not significantly affected, then it is likely the policies considered were not influential enough or at least their effects were anticipated early enough for them to not have an impact. I then perform an event study on each policy and look at the cumulative abnormal returns of high-ESG-rated firms around the policy announcements to see if they are significantly positive.
In addition, I study whether the cumulative abnormal returns (CARs) of high-ESG-rated companies are significantly higher around national environmental policy announcements rather than international ones.
While the results of the strictness test and analysis on environmentally friendly firms seem to be significant for the Paris Agreement and the Official Draft of the German Climate Action Plan 2050, in general the results are mixed. On the other hand, there seems to be strong evidence that high-ESG-rated firms experience significantly higher CARs around national environmental policy announcements compared to around international announcements.
2. Literature Review: considerations of ethical investing and
environmental policy impacts on firms
The literature studying the impact of socially responsible investing on stock returns is certainly not lacking in quantity, but the results presented are mixed, with researchers providing different reasons for the relevance, or lack thereof, of SRI for stock returns.
Renneboog et al (2008) describe how SRI originated from religious traditions. It started to become increasingly popular from the 1990s onwards, as investors began to see how industrial development could harm the environment, especially due to events like the Chernobyl disaster of 1986 and the Enron oil spill of 1989: just in the US, the assets managed by SRI funds increased from $162 billion in 1994 to $2.3 trillion in 2005. For this reason, the authors argue that socially responsible investors expect firms to not only focus on financial performance, but to do this while maximizing social welfare.
While classical economics studies suggest that value and social welfare maximization go hand in hand, Baumol (1991) argues that firms that give importance to corporate social responsibility will succumb to the competitive pressures of the market. However, such a study was made as SRI had just begun to grow, and today it may very well prove wrong: already in 2003, Bagnoli and Walts showed that CSR should be feasible in market equilibrium. The fact is, firms do not necessarily have to make more money than other firms to receive support from investors: in another study, Renneboog et al. found that investors were making less money by investing in SRI funds in European and Asia-Pacific countries rather than if they had invested in the domestic benchmark portfolios (Renneboog et al., 2008). Together with a previous study from 2005, Renneboog et al. provide evidence that investors are willing to accept lower returns in favour of more ethical firm behaviour. On the other hand, a 2016 study by Auer on European data reports that investing in a socially responsible way does not force investors to earn lower returns. However, this is only the case when investors use specific negative screens to exclude stocks without an ESG rating. Most importantly, the author reports that, while environmental (and social) selection does not negatively impact returns, it provides no additional returns either (Auer, 2016).
Galema et al (2008) research why studies do not find a relationship between alphas and SRI. They find evidence that SRI affects stock returns not by creating positive alphas, but by lowering the book-to-market ratio of firms. Interestingly, they report that firms with high scores related to diversity and the environment experience decreases in the ratio. This decrease means that, in the Fama-French regressions often run by researchers, the alphas do not capture SRI effects. As they report, this “explains why so few studies are able to establish a link between alpha’s and SRI”, while also being consistent with theory that generally suggests SRI significantly impacts stock returns.
Bauer et al (2005) find little evidence of SRI funds significantly over or underperforming non-SRI funds. While their study suggests that UK domestic and international ethical funds significantly outperform conventional ones, they report that alphas of German SRI and non-SRI funds have insignificant differences.
Derwall et al (2011) try to explain why research on the effect of SRI on returns is mixed. They argue that the SRI movement consists of 2 groups of people: values-driven investors and profit-driven investors. They also propose two hypotheses: the shunned stock hypothesis and the errors-in-expectations hypothesis. According to the first hypothesis, values-driven investors shun non-SRI stocks in favour of SRI-stocks, making the former under-priced and
thus increasing their returns compared to the latter’s. The second hypothesis suggests that SRI stocks have higher risk-adjusted returns compared to conventional ones due to the market underestimating the effect of CSR on their performance. The two effects supposedly often cancel out in practice, leading to studies finding no significant differences between returns of SRI and conventional funds.
The authors also report that “There is… evidence that stocks of companies with positive scores on environmental and social issues outperform companies with low scores over specific periods, which supports the errors-in-expectations hypothesis” (Derwall et al., 2011). If my research were to find that firms indeed experience positive abnormal returns following environmental policy announcements, it may be that the “specific periods” the authors refer to include periods following green policy announcements.
Environmental policy impacts on firms
Ramiah et al. (2013) find that the Australian market was significantly affected by the announcement of the Carbon Pollution Reduction Scheme (CPRS), with most sectors being negatively affected. Curiously enough, the sector with the largest negative cumulative abnormal returns following the announcement was the alternative energy sector. Furthermore, they report that the biggest polluters, the policy’s targets according to the Australian government, were not significantly affected: electricity producers, at the top of the biggest-polluters list, faced no gains or losses. For this reason, the authors argue that
“green policies in their current form may not be effective” (Ramiah et al., 2013).
In a later study, Ramiah et al. (2015) looked at the effect of green policy announcements on portfolios of industrial firms in the US. They report “negative abnormal returns and increases in systematic risk for the biggest polluters”, although more environmentally responsible firms were not significantly affected, which again raises the question of policy effectiveness. Overall, they found that environmental policy announcements in the US, on average, greatly affected US and non-US stock returns, but that this effect is not necessarily positive for “greener” firms and negative for conventional firms.
Another important study is by Dowell et al. (2000). They find that “adopting stringent global environmental standards is positively associated with a higher firm value” (p. 1069). While this finding may not necessarily apply to the national climate policies discussed in this paper, the international Kyoto Protocol and the Paris Agreement should fully qualify. One could say that announcement of the policy is not the same as its adoption. However, it should be plausible to assume that the effect of adopting the policy will be very similar to when the policy is announced (at least in terms of positive or negative, if not in terms of magnitude) and, unless the policy implications undergo significant changes between announcement and adoption date, its full effect will already be reflected on stock prices upon announcement. The authors also argue that one of the reasons for their finding may be that “poorly managed and less competitive firms may tend to adopt lower environmental standards” (Dowell et al., 2000). Combining these two observations, we would indeed expect high-ESG-rated firms to experience positive abnormal returns following announcements of green policy. This is supported by papers like Cohen et al. (1995), that find firm profits and environmental performance to be strongly correlated.
A recent paper by Batten et al. (2016) studies the impact of climate change on central banks and the economy in general. They also run an event study to examine the reaction of the market to events related to investing in carbon-related assets. The authors find that “these events generally had a negative but insignificant effect on abnormal returns for oil and gas companies, but a positive and significant effect for renewable energy companies” (p. 16). Other research and the bottom line
As discussed above, the literature on socially responsible investing seems to provide mixed results. However, an interesting study by Friede et al. (2015) aggregates the findings of more than 2000 empirical studies on the relation between ESG and financial performance. The authors report that 90% of the studies argue for a non-negative “ESG-CFP” (corporate financial performance) relation, with most of these actually finding a stable and positive relation between the two.
With respect to environmental policies, Blancard & Petit (2019) study how the market reacts to ordinary ESG news. They report that while negative events slightly decrease stock prices, positive events have no significant effects.
This paper will have quite a narrow scope. While the focus on one country is important, what I believe will be key will be the focus on the environmental factor of ESG, as I will not consider social and governance factors in my analysis. With respect to Friede et al.’s 2015 study, this paper could shed some light on how significant the “E” in ESG is on its own. The fact that I will study the impact of major, more specific environmental policies rather than that of ordinary ESG news may lead to different results from Blancard & Petit’s 2019 study.
2. Research Question and Hypotheses
In this paper, I research whether firms with high ESG ratings experience positive abnormal
returns following announcements of environmental policy. To do this, I look at the average
cumulative abnormal returns (CARs) of firms around each of the events considered. I then compare the average CARs of companies around national environmental policy announcements to those around international ones, to gain some insight on whether the nature of the announcement affects the results.
All the hypotheses I study are one-sided.
The first five hypotheses test whether the CARs of firms with high ESG ratings around a given event are significantly positive. As explained above, companies with high ESG ratings were found to outperform conventional ones in several studies. For example, Friede et al. (2015) reported a positive relationship existing between ESG and corporate financial performance in most of the studies they analysed. As mentioned previously, because one would expect highly polluting companies to suffer due to environmental policy restrictions, it would make sense for more environmentally-friendly ones to experience increases in their returns around such policies: as they would not be negatively affected by the policies unlike conventional or highly-polluting firms, investors would likely invest more into them, which would justify positive cumulative abnormal returns.
In my sixth and final hypothesis, I test whether high-ESG-rated firms experience CARs around environmental policy announcements that are higher if the policies are national compared to
if they are international. An important reason for this assumption is that, while national policies bind the country’s companies to follow them, international policies are less restrictive. The binding nature of national policies means that polluting companies will be forced to comply, thus affecting them more severely on average than if the policies were international since they cannot avoid the policy’s consequences. Again, based on the idea that investors should reasonably be expected to invest more into environmentally-friendly companies if conventional ones underperform due to environmental policy announcements, a more detrimental effect on polluting companies should translate to a more beneficial one on high-ESG companies. For this reason, national environmental policies should harm polluting firms more, and thus benefit environmentally responsible firms more, than international policies. The hypotheses I test are given below.
Separate events
Hypothesis 1H0: Firms with a high ESG rating in Germany do not experience any significant cumulative abnormal returns following the announcement of the Kyoto Protocol.
H1: Firms with a high ESG rating in Germany experience significantly positive cumulative abnormal returns following the announcement of the Kyoto Protocol.
Hypothesis 2
H0: Firms with a high ESG rating in Germany do not experience any significant cumulative abnormal returns around the announcement of the Paris Agreement.
H1: Firms with a high ESG rating in Germany experience significantly positive cumulative abnormal returns around the announcement of the Paris Agreement.
Hypothesis 3
H0: Firms with a high ESG rating in Germany do not experience any significant cumulative abnormal returns around the second leak of the German Climate Action Plan 2050.
H1: Firms with a high ESG rating in Germany experience significantly positive cumulative abnormal returns around the second leak of the German Climate Action Plan 2050.
Hypothesis 4
H0: Firms with a high ESG rating in Germany do not experience any significant cumulative abnormal returns around the official announcement of the German Climate Action Plan 2050. H1: Firms with a high ESG rating in Germany experience significantly positive cumulative abnormal returns around the official announcement of the German Climate Action Plan 2050. Hypothesis 5
H0: Firms with a high ESG rating in Germany do not experience any significant cumulative abnormal returns around the announcement of Germany’s Climate Package of 2019.
H1: Firms with a high ESG rating in Germany experience significantly positive cumulative abnormal returns around the announcement of Germany’s Climate Package of 2019.
National vs International events
Hypothesis 6H0: On average, there is no significant difference between the cumulative abnormal returns of high-ESG-rated companies in Germany around national and international environmental policy announcements.
H1: On average, the cumulative abnormal returns of high-ESG-rated companies in Germany around national environmental policy announcements are significantly higher than around international ones.
3. Data and Methodology
In this paper, the data comes from different sources. First, I identify the 30 most important ESG-rated companies in Germany via the Sustainalytics database, which contains ESG-rated companies which have high Environmental, Social and Governance scores. Specifically, I select the companies with an Environmental Score equal to or higher than 50. I then retrieve their respective daily closing prices via the Compustat – Capital IQ database, and use them to calculate the discretely compounded stock returns. Since the earliest and latest environmental policy announcements occurred, respectively, in 1997 and 2019 (The Kyoto Protocol and the German Climate Package 2019), I retrieve data going from the beginning of 1997 to the end of 2019. Both Sustainalytics and Compustat – Capital IQ were accessed through Wharton Research Data Services (WRDS).
Second, I approximate the German market returns via the daily returns of the DAX Performance-Index, which contains the 30 most prominent German companies that trade on the Frankfurt Stock Exchange, retrieved from Investing.com. As the data only includes prices, I again calculate returns by using adjusted closing prices. Lastly, I take the 10-year daily German bond returns from Yahoo! Finance and use them to approximate the risk-free rate. A problem is that there are some extreme values of the risk-free rate which could bias interpretation; these outliers are mostly observed around the Second Leak of the German Climate Action Plan 2050. One could argue these fluctuations are actually due to the event announcement, but it seems unlikely as they persist for only single days, and then quickly return to much lower and common values. For this reason, I use two different methods to test the environmental policy announcements I consider: one for the Kyoto Protocol, the Paris Agreement, the Official Draft of the German Climate Action Plan 2050 and the German Climate Package 2019, and one for the Second Leak of the German Climate Action Plan 2050. These methods are explained below.
Testing of environmental policy announcements on high-ESG-rated firms
In this thesis I carry out an event study. To do so, I follow the methodology presented by De Jong in his 2007 paper. I look at the expected abnormal returns of (a group of) companies around environmental policy announcements. These abnormal returns are then used to calculate the cumulative abnormal returns, which I use to test my hypotheses.
As mentioned previously, I use two different models to test different policy announcements. These two models differ in the way they calculate the normal returns (aka benchmark returns) which are used to find the abnormal returns of a given stock at a certain time (De Jong, 2007, p. 3-6). In both cases, I use an estimation window of 30 trading days, going from 60 days before the environmental policy announcement considered to 30 days before it; I make a separate estimation window for each event. When calculating cumulative abnormal returns, I consider the event window starting 5 trading days before the event and ending 5 trading days after it, for a total of 11 days.
For the Kyoto Protocol, the Paris Agreement, the official draft of the German Climate Action Plan 2050 and the German Climate Package 2019, I use the Capital Asset Pricing Model as the benchmark used to estimate abnormal returns. First, the excess returns are operationalized as:
𝑅",$− 𝑅&,$ =b" ∗ )𝑅*+$,$− 𝑅&,$, (1) where 𝑅",$ is the stock i’s (discretely compounded) return, 𝑅&,$ is the return of a 10-year
German bond, and 𝑅*+$,$ is the return of the market approximated by the DAX, all for a given
time t. b" thus represents the sensitivity of a stock i’s excess returns to the market. The abnormal returns are then expressed as the difference between the actual returns of a given stock, in excess of the risk-free rate, and those predicted by the CAPM, which represent the so-called “normal returns” (De Jong, 2007, pp. 3-6):
𝐴𝑏𝑛𝑅𝑒𝑡",$ = 𝑅",$− 𝑁𝑜𝑟𝑚𝑎𝑙𝑅𝑒𝑡",$,<=>? (2) Where 𝑁𝑜𝑟𝑚𝑎𝑙𝑅𝑒𝑡",$,<=>? is the predicted (normal) return of firm i at time t calculated using
the Capital Asset Pricing Model.
For the Second Leak of the German Climate Action Plan 2050, I use the market model. The abnormal returns are calculated in a similar manner to the CAPM model explained above, except in this case the risk-free rate is not considered:
𝑅",$ =a" +b"∗ 𝑅*+$,$+e",$ (3) 𝐴𝑏𝑛𝑅𝑒𝑡",$ = 𝑅",$− 𝑁𝑜𝑟𝑚𝑎𝑙𝑅𝑒𝑡",$,?CD (4) Where 𝑁𝑜𝑟𝑚𝑎𝑙𝑅𝑒𝑡",$,?CD is the predicted (normal) return of firm i at time t calculated using
the market model.
As the market model does not use the risk-free rate, it means the measure’s outliers will not be a problem: the abnormal returns of the environmental policy announcement I apply it to are thus feasibly comparable to the others.
After estimating the abnormal returns of every firm around each environmental policy announcement, I calculate their respective cumulative abnormal returns (CARs), which are simply the sum of all the abnormal returns around the given event announcement.
In the first five hypotheses, I look at the CARs of each firm around each of the five event announcements I study to see if they are significantly larger than 0. For this reason, following De Jong (2007), the hypotheses can be expressed as:
𝐻G: 𝐸)𝐶𝑢𝑚𝐴𝑏𝑛𝑅𝑒𝑡",$,L, = 0 𝐻N: 𝐸)𝐶𝑢𝑚𝐴𝑏𝑛𝑅𝑒𝑡",$,L, > 0
where i represents the firm, t the time, and e the event considered. As De Jong argues, looking at the (cumulative) abnormal returns of each firm separately is not necessarily the best method to study the effect of an event on stock returns, as they may be affected by other (idiosyncratic) information not directly connected to the event (De Jong, 2007). Analysing the average (cumulative) abnormal returns should ignore said other information and instead isolate the impact on stock returns of the event considered. For this reason, I run t-tests on the average cumulative abnormal returns across firms for each event, operationalized as follows:
𝑡L = √𝑁 ∗𝐴𝑣𝑔𝐶𝑢𝑚𝐴𝑏𝑛𝑅𝑒𝑡𝑠 L
L (5)
where e represents the event considered, N is the number of firms in the dataset, and 𝐶𝑢𝑚𝐴𝑣𝑔𝐴𝑏𝑛𝑅𝑒𝑡L and 𝑠L are, respectively, the cumulative average abnormal return and its
standard deviation across firms for event e.
Important to note is that not all of the high-ESG-rated companies I select have enough information to test all events. In fact, 4 of the 30 firms selected do not have the data required to test the Kyoto Protocol. As they have data available to test all other environmental policy announcements, I do not exclude them from the analysis. However, this means that the Kyoto Protocol is tested with observations from 26 firms, while the other announcements are tested with those from 30.
Testing on firms in highly polluting industries
As mentioned above, I first test whether the environmental policy announcements were seen as surprising and/or unexpectedly strict by looking at the CARs of firms in highly polluting industries. Thus, I first test whether the CARs of highly polluting firms around the environmental policy announcements were significantly smaller than zero: if there is enough evidence for the null hypotheses to be rejected, then it is reasonable to assume that such announcements were unexpected, either in their totality or in their strictness. If this were the case, then it would make sense to find abnormal returns in my main analysis of the high-ESG-rated firms.
This test is based on hypotheses similar to hypotheses 1-5, only in this case the alternative hypotheses are that the CARs are significantly negative. Thus, for example, the hypothesis regarding the Kyoto Protocol in this case would be:
H0: Firms in highly polluting industries in Germany do not experience any significant cumulative abnormal returns following the announcement of the Kyoto Protocol.
H1: Firms in highly polluting industries in Germany experience significantly negative cumulative abnormal returns following the announcement of the Kyoto Protocol.
The hypotheses regarding other environmental policy announcements follow the same logic. The formulas I apply to this test are the same as the ones used in the main analysis of high-ESG-rated firms. In addition, just like in the main analysis, I also use a mix of CAPM and market models to test the environmental policy announcements to avoid problems with risk-free rate outliers. Thus, the only thing that changes across the two analyses is the data used, which again is mostly retrieved via WRDS. After this test, I proceed with the study of the high-ESG rated firms.
Identifying which industries are the most polluting can be tricky: Low (1992) argues that there is no standard definition of what he refers to as “dirty industries” (p. 106). Thus, to choose which industries may classify as such, I turn to several articles.
According to Jänicke et al. (1997), the so-called “dirty industries” include “the producers of paper and paperboard, petroleum products, primary metals, stone, clay, glass and chemicals”, to which they also add the non-manufacturing “electricity production, mining and road transport” (Jänicke et al., 1997, p. 468). Based on recent literature on cosmetic ingredients as pollutants (Jardak et al., 2016; Juliano & Magrini, 2017), I also consider the industry of personal products. Others I add to the list are the clothing and footwear (Šajn, 2019), construction and building materials related (Morledge & Jackson, 2001; Cheng et al., 2006; Fan, 2017), and the semiconductor and electronic components and hardware (Holden & Kelty, 2005) industries. According to a recent briefing from the European Environment Agency, while the European industrial sector has greatly improved, its emissions still remain high (2015): for this reason, I also take firms from subsectors such as the industrial machinery one.
I next select German firms part of the industries outlined above from those listed on the Frankfurt Stock Exchange (Deutsche Börse). I then use Compustat – Capital IQ to retrieve closing prices for each of these firms that have data available for all or at least most of the years 1997 to 2019. Once again, not all companies considered have enough data to test all events, the problem mostly being focused on the Kyoto Protocol. This is a limitation of the analysis that will later be brought up again and explained.
National vs international environmental policy: is there a difference?
While national and international environmental policies differ in several ways, in my sixth hypothesis I look at whether there is a significant difference between the CARs of high-ESG-rated firms around national and international policy announcements. I thus create a dummy variable, called International, which is equal to 1 if the environmental policy announcement is international (thus either the Kyoto Protocol or the Paris Agreement) and 0 if national (representing either the Second Leak or Official Draft of the German Climate Action Plan 2050, or the 2019 German Climate Action Package. I then regress the CARs of firms around
environmental policy announcements on a constant and this dummy variable, meaning the hypothesis is as follows:
𝐻G: bUV$LWVX$"YVXZ = 0
𝐻G:bUV$LWVX$"YVXZ < 0
The hypothesis is negatively one-sided for a couple of reasons. First of all, national policies “only” have to worry about one country, while international policies must consider the possible impact on even hundreds of them around the world. For a country like Germany, which is used to a certain level of environmental policy at home, handling an international policy that must also consider less environmentally-competent countries may be easier. The second and even more important reason, already mentioned in Section 3, is that while firms must adhere to national policy, international environmental policies like the Kyoto Protocol are often just “recommendations”. For both of these reasons, firms – whether high-ESG-rated ones or those in highly polluting industries – should be affected more by national than international policies. As previously discussed, this means that high-ESG-rated companies should experience higher CARs around national environmental policy announcements compared to around international ones, making the coefficient of the International dummy negative.
4. Results and Discussion
Tests on firms in highly polluting industries
Table 1 | Firms in highly-polluting industries | Mix of CAPM and Market Model
Table 1 above shows the average CARs around each event and the respective t-tests for significance when looking at firms in highly-polluting industries and using a mix of the CAPM model and market model to estimate normal returns. As previously mentioned, these tests are meant to try and understand whether or not the environmental policy announcements I consider were (reasonably) unexpected by the market either in their totality or at least in their strictness. If such events were actually unexpected, one would expect the average CARs, and thus the t-statistics, to be significantly negative. From these results, it seems that none of the environmental policy announcements were unexpected, except for the Second Leak of the German Climate Action Plan 2050 whose one-sided t-statistic is significant at the 1% level.
Event Date of announcement Average CAR t-test
Kyoto Protocol 11-Dec-97 -3.568 -0.851 Paris Agreement 11-Dec-15 3.936 2.045
Second Leak 21-Jun-16 -6.279 -2.420*** Official Draft 06-Sep-16 210.592 127.711 Climate Package 20-Sep-19 5.816 1.726
Important to remember is that the "Second Leak” event is the only one studied via a market model. Although this was done to avoid the problems of outliers in the risk-free rate biasing results, it is also true that the market model can in general be seen as a more powerful test than one using the CAPM model. The “power” of a statistical test is the “probability that the test correctly rejects the null hypothesis when the alternative hypothesis is true…” (Stock & Watson, 2015). As restrictions are added to a test, its power decreases: an additional restriction increases the variance compared to the case without it, meaning the standard deviation of the error will increase as well, leading to a lower test statistic. In the case at hand, one could say that, all other things equal, it makes sense for the “Second Leak” announcement, tested via the market model, to have a more significant result compared to the other tests, which were based on the CAPM and thus have the additional restriction of the risk-free rate. For this reason, I compare the mixed model results to another one based only on the market model.
Table 2 | Firms in highly-polluting industries | Market Model only
Table 2 presents the same test on polluting firms, but this time all the environmental policy announcements were tested via the market model. While the Kyoto Protocol and Climate Package results changed, they remain insignificant. On the other hand, both the “Paris Agreement” and “Official Draft” test statistics are now significantly negative, respectively at the 1% and 5% levels. This analysis suggests that three announcements were unexpected, compared to only one under the mixed analysis. The fact that the results for the Kyoto Protocol and the Climate Package remain constant across analyses may signify that these two announcements were entirely expected: this suggests that a test on high-ESG-rated firms should return insignificant results regarding CARs. However, the differing results across analyses for the other announcements do not really support one interpretation over the other: while the CAPM analysis of the Paris Agreement and Official Draft suggests these events were expected, the market model analysis says otherwise. Due to the presence of extreme outliers in the risk-free rate around the Second Leak announcement, it would not make sense to compare its results according to the market model analysis with those of a CAPM analysis; the only analysis suggests that the policy announcement was unexpected, so the main one based on ESG-rated companies should yield significant results under the null hypothesis.
Tests on high-ESG-rated firms
Table 3 presents the average CARs around each environmental policy announcement and the respective tests for significance when considering high-ESG-rated firms and estimating normal returns using a mix of the CAPM and the market model.
Event Date of announcement Average CAR t-test
Kyoto Protocol 11-Dec-97 -4.512 -1.077 Paris Agreement 11-Dec-15 -6.045 -3.141***
Second Leak 21-Jun-16 -6.279 -2.420*** Official Draft 06-Sep-16 -2.972 -1.801** Climate Package 20-Sep-19 2.306 0.697
Table 3 |High-ESG-rated firms | Mix of CAPM and Market Model
As was expected from the test on firms in highly polluting industries, CARs around the Kyoto Protocol and the Climate Package announcements do not seem to be significantly different from 0. The tests on the Paris Agreement and the Official Draft seem to agree with the market model analysis of the firms in highly polluting industries. On the other hand, the test statistic of the Second Leak would be significantly negative in a two-sided or negatively one-sided test. With a mixed analysis, it seems that high-ESG-rated firms experienced significant CARs only around two of the 5 environmental policy announcements. As before, I also run a market-model-only analysis of the policy announcements to compare it to the mixed analysis. The results are shown in Table 4 below.
Table 4 |High-ESG-rated firms | Market Model only
As can be seen, none of the test statistics shown is significantly positive, as I instead expected based on my null and alternative hypothesis on high-ESG-rated firms. In fact, the results would mostly be significant if the initial hypothesis were two-sided or one-sided predicting negative abnormal returns. Compared to the mixed CAPM and market model analysis shown in table 3, the Kyoto Protocol and Climate Package tests remain insignificant, while the Paris Agreement and Official Draft tests change from being highly significant to insignificant under the null hypothesis.
A last analysis I run to analyse ESG firms’ CARs around environmental policy announcements is one based on simple regressions which regress the CARs of the event considered on a constant. These regressions are equivalent to the t-tests in the tables above, but they also allow to use robust standard errors. The results are shown in the Regression Output 1 below, and are based on the mixed CAPM and market model analysis reported in Table 3.
Event Date of announcement Average CAR t-test
Kyoto Protocol 11-Dec-97 -0.256 -0.186 Paris Agreement 11-Dec-15 5.563 5.308***
Second Leak 21-Jun-16 -4.106 -4.199 Official Draft 06-Sep-16 208.511 196.784*** Climate Package 20-Sep-19 1.238 1.189
Significance levels: *** p<0.01, ** p<0.05, * p<0.1
Event Date of announcement Average CAR t-test
Kyoto Protocol 11-Dec-97 -2.178 -1.569 Paris Agreement 11-Dec-15 -4.417 -4.215 Second Leak 21-Jun-16 -4.106 -4.199 Official Draft 06-Sep-16 -5.056 -4.780 Climate Package 20-Sep-19 -1.880 -1.864
Regression Output 1
Overall, the results across the ESG analyses are mostly insignificant, with the exception of the Paris Agreement and Official Draft announcements in the mixed analysis. Usually, regression outputs are based on two-sided hypothesis tests, so that would make the Second Leak CARs also significant at the 1% level. However, the hypothesis I test is (positively) one-sided, which leaves only the Paris Agreement and Official Draft (respectively an international and a national event) average CARs as significantly positive. These outcomes agree with the general idea of previous studies finding mixed results on the effect that environmental policy announcements have on stock returns; on the other hand, they don’t necessarily match with Dowell’s 2000 findings, mentioned in the literature review, that global environmental policy is positively associated with a higher firm value” (p. 1069), as the firms experienced significantly positive cumulative abnormal returns only around the announcement of the Paris Agreement.
National vs international environmental policy announcements
As mentioned previously, in my sixth hypothesis I test whether CARs of high-ESG-rated firms are significantly higher around national environmental policy announcements than around international ones. Below is the output of the regression of CARs of firms around policy announcements on a constant and the dummy variable International.
Regression Output 2
The output gives more than enough evidence to reject the null hypothesis that the nature of the environmental policy announcement – national or international – is irrelevant when calculating CARs of firms around such announcements: the t-statistic of the International
(1) (2) (3) (4) (5)
VARIABLES Kyoto Protocol Paris Agreement Second Leak Official Draft Package
Constant -0.286 5.561*** -4.199 208.5*** 0.989 (1.484) (1.055) (1.000) (1.065) (1.123)
Observations 26 30 30 30 30
R-squared 0.000 0.000 0.000 0.000 0.000 Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1 (1) VARIABLES International International -65.57*** (3.169) Constant 68.42*** (3.155) Observations 1,606 R-squared 0.143
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
dummy variable is -20.70, which supports the idea that high-ESG-rated firms do in fact experience higher CARs around policy announcements when these are national (the Second Leak and Official Draft of the German Climate Action Plan, and the German Climate Package 2019) compared to when they are international (the Kyoto Protocol and the Paris Agreement).
In general, the findings do not present any strong evidence that high-ESG-rated firms experience positive CARs around environmental policy announcements. In the main analysis based on the CAPM and market model mix, only 2 of the 5 announcements considered presented positive results, with the other 3 being insignificant; in the market model only analysis, the CARs of the firms were not significantly positive around any of the events. These findings go against the positive ones of Friede et al. (2015), and are instead more in line with Ramiah et al.’s 2015 findings in the US, where environmental policies did not necessarily positively affect “green” firms. These mixed results may be due to the fact that, according to the mixed analysis of the CARs of firms in highly-polluting industries, all policy announcements except the Second Leak were for the most part apparently either not surprising or not strict enough to warrant abnormal returns. On the other hand, it is also true that, when implementing only the market model, 2 additional policy announcements out of 5 caused significantly negative cumulative abnormal returns in the firms considered.
On a different note, the findings seem to strongly support the idea that the CARs of green firms are more significantly affected by national, rather than international, environmental policy announcements.
In the end, one could argue national environmental policies are more likely to be effective than international ones when considering a single country: this would explain why a lower fraction of international environmental policies significantly impact the returns of its firms compared to national policies. A study considering a larger number of both national and international events would undoubtedly be better to see whether this is actually true or not. Another factor to consider is the validity of the results of the Second Leak, both on its own and through its impact on the average CARs across events. An improved analysis could find that these returns were actually significantly positive, or perhaps that they were truly negative but not in the magnitude I found. Something important to point out is that the “Second Leak” and the “Official Draft” environmental policy announcements are both part of German Climate Action Plan 2050, of which I did not consider all leaks and other relevant information. A more in-depth event study of this specific Plan may prove to be useful in understanding the effectiveness of national environmental policies in Germany.
5. Conclusion
Germany is generally considered to have some of the world’s most progressive green policies (Bailey, 2007). In this paper, I ran an event study on five environmental policy announcements, three of them national and two of them international, to test whether high-ESG-rated firms in Germany were positively affected by them. To do so, I looked at the average cumulative abnormal returns across firms for each event. In addition, I also studied whether cumulative abnormal returns of high-ESG-rated firms around national
environmental policy announcements are higher than around international policies. Based on the results, 2 of the 5 policy announcements considered caused significantly positive CARs for the firms: the Paris Agreement and the Official Draft of the German Climate Action Plan 2050. Whether this outcome is enough to support the idea that high-ESG-rated firms actually benefit from environmental policy announcements is up to debate, especially as different analyses provide different results. On the other hand, the results from the analysis of the nature of the environmental policy seem to provide ample support for the idea that national environmental policy announcements have a significantly larger effect on the returns of high-ESG-rated firms compared to international policies.
There are several factors and limitations that must be considered when interpreting the results from the analyses. First of all, the number of firms used for both the main analysis and for test for strictness of the environmental policy announcements was not large. A total of thirty high-ESG-rated firms were studied, which is usually considered to be more than enough to allow for normal approximations of the distributions of the data (Stock & Watson, 2015). However, 4 of the firms selected did not have enough data to study the Kyoto Protocol. This lack of data becomes more serious when considering firms in highly polluting industries: while 28 firms were chosen in total, only 15 were able to be used to test all events, as 13 of them did not have enough data for the Kyoto Protocol. While Stock & Watson (2015) suggest that “for n>15, the difference in the p-values computed using the Student t and standard normal distributions never exceeds 0.01…” (p. 137), it is well known that a low sample size does not help in finding powerful tests. Future research should undoubtedly try and increase the number of firms examined, though this is not necessarily an easy task if the focus remains on a single country.
Another possible problem in the data may have been the difference between when the ESG scores were assigned to German firms and today. According to Sustainalytics, the ratings were last updated in August 2009. These scores may very well have changed in the past decade, meaning they may not be faithful representations of the current state of affairs. For example, a firm included in the high-ESG-rated list was RWE AG, but this same firm was actually considered as highly polluting by the Carbon Majors Report of 2017 (Griffin & Heede). Furthermore, other firms like Adidas and Puma were also given high ESG scores back in 2009, but are actually part of the (highly) polluting industries if one follows the 2019 article by Šajn. Of course, this may very well also be a problem related to how I chose the highly polluting industries. In addition, a firm could be part of a highly polluting industry, but still be environmentally friendly when compared to its peers in the same industry. In general, it would be best to try and use environmental scores that have been assigned on dates that are not too different from the environmental policies analysed: possible new research would benefit from updated environmental scores and/or better screening methods when selecting both polluting and greem companies in a way that does not excessively restrict available observations.
The last but probably most important point to note when considering how the two analyses of cumulative abnormal returns generally did not provide significant results is that environmental policy announcements may very well have been expected. On one hand, the results of the main analysis of CARs of polluting companies based on a mix of CAPM and market model lends credence to the idea that most policies (4 out of 5) were expected. On
the other hand, the market-model-only analysis suggests 3 of the policies were unexpected. This “test for expectations” is key when analysing the results because if the highly polluting firms did not experience significantly negative (or one might also argue positive via a two-sided test) CARs, then the policies were expected, meaning all information about them would be included in the price of stocks: there would be no abnormal returns as they would all be expected by the market. When looking at the environmental policies considered, one could justifiably be more predictable than another one. For example, international agreements like the Kyoto Protocol and the Paris Agreement require the discussion of conditions by many different countries, and thus likely take more time to be completed than national policies on average: this makes it more likely for their consequences to be predicted (for example due to information updates and leaks) before they are officially announced. In this case, this logic seems to hold for the Kyoto Protocol, but not so much for the Paris Agreement as is shown by the difference in results between the mixed and market-model-only CAR analyses: this may be because, while the event itself was predicted, its strictness was not.
The Climate Package remained insignificant in both analyses, although if the test for strictness on firms in highly polluting industries had been two-sided, then the result would have been significantly different from zero.
The most interesting policies of the ones considered are undoubtedly the Second Leak and the Official Draft of the German Climate Action Plan. The results of the analyses on polluting firms are significant for both policies when considering the market-model-only CARs. While the CAPM analysis of the Second Leak would not have been fruitful due to outliers in the risk-free rate, the CARs around the Official Draft would have been very significant had the test for strictness been two-sided. This result, however, may also be due to (smaller but still included) outliers in the risk-free rate, which contributed to the large values of CARs around the Official draft in the analyses of both the polluting firms and the ESG-rated firms. Overall, the results from these two policy announcements are curious. After all, they are part of a larger plan which also includes a “first leak” of information. This first leak would supposedly prepare firms for any possible successive leaks and drafts, unless, of course, these were unexpectedly strict. Further research focusing on the German Climate Action Plan 2050 would likely be interesting to study this. However, if different environmental policies are studied, I believe future research should look at a larger number and wider variety of environmental policy announcements to properly understand whether or not high-ESG-rated firms in Germany, or anywhere else, benefit from environmental policy.
References
Deutsche Börse Cash Market. Listed Companies.
Retrieved July 10, 2019, from https://www.deutsche-boerse-cash-market.com/dbcm-en/instruments-statistics/statistics/listes-companies
Baumol, W. J., & Blackman, S. A. B. (1991). Perfect Markets and Easy Virtue Business Ethics and the Invisible Hand.
Low, P. (1992). Trade measures and environmental quality: the implications for Mexico's exports. World Bank Discussion Papers[WORLD BANK DISCUSSION PAPER.]. 1992.
Cohen, M. A., Fenn, S., & Naimon, J. S. (1995). Environmental and financial performance: are they related?. Washington, DC: Investor Responsibility Research Center, Environmental Information Service.
Jänicke, M., Binder, M., & Mönch, H. (1997). ‘Dirty industries’: Patterns of change in industrial countries. Environmental and resource economics, 9(4), 467-491.
Dowell, G., Hart, S., & Yeung, B. (2000). Do corporate global environmental standards create or destroy market value?. Management science, 46(8), 1059-1074.
Morledge, R., & Jackson, F. (2001). Reducing environmental pollution caused by construction plant. Environmental Management and Health.
Bagnoli, M., & Watts, S. G. (2003). Selling to socially responsible consumers: Competition and the private provision of public goods. Journal of Economics & Management
Strategy, 12(3), 419-445.
Bauer, R., Koedijk, K., & Otten, R. (2005). International evidence on ethical mutual fund performance and investment style. Journal of Banking & Finance, 29(7), 1751-1767. Holden, J., & Kelty, C. (2005). The Environmental Impact of the Manufacturing of Semiconductors. Retrieved July 9, 2020, from: https://cnx.org/contents/Sc3- Pqne@4.35:7238FjUe@3/The-Environmental-Impact-of-the-Manufacturing-of-Seminconductors-ticket-5622
Renneboog, L., Ter Horst, J., & Zhang, C. (2005). Is ethical money financially smart?. ECGI-Finance Working Paper, (117).
Cheng, E. W., Chiang, Y. H., & Tang, B. S. (2006). Exploring the economic impact of construction pollution by disaggregating the construction sector of the input–output table. Building and Environment, 41(12), 1940-1951.
Bailey, I. (2007). Market environmentalism, new environmental policy instruments, and climate policy in the United Kingdom and Germany. Annals of the Association of American Geographers, 97(3), 530-550.
De Jong, F., & de Goeij, P. (2007). Event studies methodology. Lecture Notes. Renneboog, L., Ter Horst, J., & Zhang, C. (2008). The price of ethics and stakeholder governance: The performance of socially responsible mutual funds. Journal of corporate finance, 14(3), 302-322.
Renneboog, L., Ter Horst, J., & Zhang, C. (2008). Socially responsible investments: Institutional aspects, performance, and investor behavior. Journal of banking &
finance, 32(9), 1723-1742.
Galema, R., Plantinga, A., & Scholtens, B. (2008). The stocks at stake: Return and risk in socially responsible investment. Journal of Banking & Finance, 32(12), 2646-2654. Derwall, J., Koedijk, K., & Ter Horst, J. (2011). A tale of values-driven and profit-seeking social investors. Journal of Banking & Finance, 35(8), 2137-2147.
Neuhierl, A., Scherbina, A., & Schiusene, B. (2013). Market reaction to corporate press releases. Journal of Financial and Quantitative Analysis, 1207-1240.
Ramiah, V., Martin, B., & Moosa, I. (2013). How does the stock market react to the announcement of green policies?. Journal of Banking & Finance, 37(5), 1747-1758.
Ramiah, V., Pichelli, J., & Moosa, I. (2015). Environmental regulation, the Obama effect and the stock market: some empirical results. Applied Economics, 47(7), 725-738.
Friede, G., Busch, T., & Bassen, A. (2015). ESG and financial performance: aggregated evidence from more than 2000 empirical studies. Journal of Sustainable Finance & Investment, 5(4), 210-233.
Stock, J. H., & Watson, M. W. (2015). Introduction to econometrics.
Auer, B. R. (2016). Do socially responsible investment policies add or destroy European stock portfolio value?. Journal of business ethics, 135(2), 381-397.
Batten, S., Sowerbutts, R., & Tanaka, M. (2016). Let's talk about the weather: the impact of climate change on central banks.
Jardak, K., Drogui, P., & Daghrir, R. (2016). Surfactants in aquatic and terrestrial
environment: occurrence, behavior, and treatment processes. Environmental Science and Pollution Research, 23(4), 3195-3216.
Fan, H. (2017). A Critical Review and Analysis of Construction equipment emission factors. Procedia engineering, 196, 351-358.
Juliano, C., & Magrini, G. A. (2017). Cosmetic ingredients as emerging pollutants of environmental and health concern. A mini-review. Cosmetics, 4(2), 11.
Griffin, P., & Heede, C. R. (2017). The carbon majors database. CDP carbon majors report 2017, 14.
Capelle-Blancard, G., & Petit, A. (2019). Every little helps? ESG news and stock market reaction. Journal of Business Ethics, 157(2), 543-565.
Šajn, N. (2019). Environmental impact of the textile and clothing industry. European Parliamentary Research Service.
Appendix
Table A. Presents the CARs of each high-ESG-rated firm around the environmental policy announcements based on the main analysis (mixed CAPM and market model).
Significance levels: *** p<0.01, ** p<0.05, * p<0.10 based on a positively one-sided alternative hypothesis
Table B. Presents the CARs of each firm part of a highly-polluting industry around the environmental policy announcements based on the main analysis (mixed CAPM and market model).
ticker CAR Kyoto KyotoTest CAR Paris ParisTest CAR Second Leak SecondLeakTest CAR Official OfficialTest CAR Package PackageTest ADS -7.813 -0.186 2.429 5.308*** -0.208 -4.199 203.984 196.784*** 5.238 1.189 ALV 8.091 -0.186 4.790 5.308*** -4.131 -4.199 215.083 196.784*** 4.762 1.189 BAS -3.497 -0.186 2.675 5.308*** -3.238 -4.199 207.813 196.784*** 3.092 1.189 BAYN -4.326 -0.186 6.917 5.308*** 2.999 -4.199 206.591 196.784*** -5.214 1.189 BEI -4.930 -0.186 1.179 5.308*** 2.523 -4.199 207.780 196.784*** -1.275 1.189 BMW -8.100 -0.186 5.132 5.308*** -3.598 -4.199 205.800 196.784*** 2.144 1.189 CBK 14.634 -0.186 4.744 5.308*** -5.830 -4.199 219.105 196.784*** 0.028 1.189 CON -5.133 -0.186 5.290 5.308*** -3.441 -4.199 210.818 196.784*** -1.152 1.189 DAI -0.792 -0.186 5.267 5.308*** -1.017 -4.199 210.299 196.784*** 0.366 1.189 DB1 // // 8.320 5.308*** -0.625 -4.199 210.047 196.784*** 5.139 1.189 DBK 6.687 -0.186 2.701 5.308*** -4.405 -4.199 215.733 196.784*** -7.049 1.189 DPW // // 5.676 5.308*** -6.675 -4.199 208.169 196.784*** -1.935 1.189 DTE -9.334 -0.186 6.871 5.308*** -0.331 -4.199 206.346 196.784*** 5.759 1.189 EOAN 10.776 -0.186 -0.907 5.308*** -1.234 -4.199 188.110 196.784*** 9.654 1.189 FRA // // 10.083 5.308*** -1.080 -4.199 209.395 196.784*** 2.441 1.189 HEI -3.932 -0.186 9.509 5.308*** -6.600 -4.199 207.549 196.784*** 4.670 1.189 HEN3 0.013 -0.186 3.614 5.308*** 4.024 -4.199 208.301 196.784*** -3.775 1.189 HOT -10.711 -0.186 3.621 5.308*** -3.403 -4.199 208.464 196.784*** 2.273 1.189 LHA -1.638 -0.186 10.340 5.308*** -3.120 -4.199 219.323 196.784*** 0.293 1.189 MRK -4.837 -0.186 0.267 5.308*** -2.403 -4.199 208.266 196.784*** 8.551 1.189 MUV2 13.117 -0.186 6.597 5.308*** -1.119 -4.199 215.085 196.784*** 6.926 1.189 PUM -0.984 -0.186 -0.449 5.308*** -8.586 -4.199 209.973 196.784*** -0.637 1.189 RWE 5.641 -0.186 -0.697 5.308*** -1.657 -4.199 200.667 196.784*** 4.169 1.189 SAP -9.007 -0.186 2.114 5.308*** -0.641 -4.199 208.320 196.784*** 3.159 1.189 SDF 15.099 -0.186 17.179 5.308*** -15.236 -4.199 212.816 196.784*** -12.999 1.189 SIE 0.038 -0.186 2.928 5.308*** -3.228 -4.199 205.309 196.784*** 8.250 1.189 SZG 0.752 -0.186 7.846 5.308*** -20.393 -4.199 206.768 196.784*** -3.953 1.189 TKA -3.513 -0.186 2.608 5.308*** -6.370 -4.199 202.697 196.784*** 6.994 1.189 VOW3 -3.725 -0.186 28.283 5.308*** -14.437 -4.199 211.033 196.784*** 1.804 1.189 WCH // // 1.911 5.308*** -12.492 -4.199 204.448 196.784*** -18.042 1.189
Company Name CAR Kyoto KyotoTest CAR Paris ParisTest CAR Second Leak SecondLeakTest CAR Official OfficialTest CAR Package PackageTest
AHLERS AG -0.669 -0.851 5.003 2.045 7.570 -2.420*** 217.148 127.711 13.598 1.726 AIXTRON SE // // -35.727 2.045 -4.502 -2.420*** 214.243 127.711 -15.692 1.726 AS CREATION TAPETEN AG // // 9.455 2.045 -10.940 -2.420*** 209.060 127.711 3.605 1.726 AURUBIS AG // // -11.785 2.045 -11.182 -2.420*** 212.539 127.711 2.520 1.726 DATA MODUL AG -16.977 -0.851 6.027 2.045 -0.315 -2.420*** 213.960 127.711 14.634 1.726 DEUTZ AG 8.144 -0.851 7.527 2.045 -15.481 -2.420*** 213.527 127.711 8.793 1.726 DUERR AG 3.957 -0.851 1.369 2.045 2.214 -2.420*** 203.093 127.711 0.779 1.726 ELMOS SEMICONDUCTOR SE // // 19.691 2.045 2.141 -2.420*** 214.772 127.711 -8.921 1.726
ENBW ENERGIE BADEN // // 6.020 2.045 -6.439 -2.420*** 214.650 127.711 12.917 1.726
FORTEC ELEKTRONIK VERTRIEBS // // 8.937 2.045 0.298 -2.420*** 211.965 127.711 2.683 1.726
FUCHS PETROLUB SE 7.316 -0.851 5.755 2.045 0.657 -2.420*** 207.077 127.711 1.690 1.726
H&R GMBH & CO KGAA -75.653 -0.851 10.751 2.045 -3.241 -2.420*** 200.963 127.711 3.456 1.726
HUGO BOSS AG // // 7.943 2.045 -7.533 -2.420*** 207.180 127.711 -5.706 1.726
INIT INNOVATION TRAFFIC SE // // -7.498 2.045 2.366 -2.420*** 206.672 127.711 -4.676 1.726
JUNGHEINRICH AG -3.143 -0.851 10.531 2.045 -70.798 -2.420*** 239.081 127.711 -1.050 1.726
KOENIG & BAUER AG 1.488 -0.851 13.743 2.045 -3.953 -2.420*** 206.243 127.711 3.982 1.726
KRONES AG // // 0.402 2.045 -3.205 -2.420*** 217.051 127.711 9.712 1.726 KSB SE & CO KGAA -9.868 -0.851 7.151 2.045 -2.558 -2.420*** 208.342 127.711 3.600 1.726 KUKA AG 4.406 -0.851 9.196 2.045 1.384 -2.420*** 194.912 127.711 2.197 1.726 SCHALTBAU HOLDING AG 6.072 -0.851 12.294 2.045 -5.515 -2.420*** 192.764 127.711 18.768 1.726 SGL CARBON SE -10.824 -0.851 -2.845 2.045 -13.505 -2.420*** 204.204 127.711 -4.373 1.726 SIMONA KUNSTSTOFFWERKE AG // // 7.124 2.045 -7.225 -2.420*** 221.180 127.711 1.634 1.726 SIXT SE -10.617 -0.851 0.290 2.045 -2.531 -2.420*** 211.266 127.711 -0.501 1.726 SUESS MICROTEC SE // // -3.111 2.045 -1.233 -2.420*** 212.045 127.711 -4.522 1.726 UZIN UTZ AG // // 14.621 2.045 0.760 -2.420*** 201.108 127.711 4.208 1.726 WASHTEC AG 0.000 -0.851 -1.302 2.045 -7.795 -2.420*** 209.876 127.711 4.520 1.726
WEBER (GERRY) INTERNATNL AG 0.000 -0.851 -3.677 2.045 -9.065 -2.420*** 219.725 127.711 90.729 1.726
Significance levels: *** p<0.01, ** p<0.05, * p<0.10 based on a negatively one-sided alternative hypothesis
Variables used in the analysis of high-ESG-rated firms (variable descriptions
provided below in the labels). Returns are expressed in percentages (so 2.859
is 2.859%)
Descriptive Statistics
Variable Obs Mean Std.Dev. Min Max datadate 156000 17877.8 2341.244 13516 21913 prccd 156000 63.81 171.962 .038 3970 MktReturns 156000 235.278 8598.327 -99.983 581000 Rf_ret 156000 -.23 53.942 -3266.67 1300 set 156000 1 0 1 1 DateKyoto 156000 13859 0 13859 13859 DateParis 156000 20433 0 20433 20433 DateSecond~k 156000 20626 0 20626 20626 DateOfficial 156000 20703 0 20703 20703 DatePackage 156000 21812 0 21812 21812 group_id 156000 14.514 8.092 1 28 Datenum 156000 2798.035 1623.135 1 5820 StockReturn 156000 .027 2.859 -90.921 60.652 tKyoto 86469 195.594 65.516 22 234 td_Kyoto 86469 2687.832 1665.005 -233 5586 tdcal_Kyoto 156000 4018.799 2341.244 -343 8054 tParis 156000 4571.997 278.255 3613 4796 td_Paris 156000 -1773.962 1623.135 -4795 1024 tdcal_Paris 156000 -2555.201 2341.244 -6917 1480 tSecondLeak 156000 4701.997 278.255 3743 4926 td_SecondL~k 156000 -1903.962 1623.135 -4925 894 tdcal_Seco~k 156000 -2748.201 2341.244 -7110 1287 tOfficial 156000 4756.997 278.255 3798 4981 td_Official 156000 -1958.962 1623.135 -4980 839 tdcal_Offi~l 156000 -2825.201 2341.244 -7187 1210 tPackage 156000 5528.997 278.255 4570 5753 td_Package 156000 -2730.962 1623.135 -5752 67 tdcal_Pack~e 156000 -3934.201 2341.244 -8296 101 Kyoto_window 156000 .001 .032 0 1 count_Kyot~s 156000 6.088 5.469 0 11 EstKyoto_w~w 156000 .003 .051 0 1 count_EstK~s 156000 15.277 14.805 0 30 Paris_window 156000 .002 .044 0 1 count_Pari~s 156000 11 0 11 11 EstParis_w~w 156000 .005 .073 0 1 c~tParis_obs 156000 30 0 30 30 SecondLeak~w 156000 .002 .044 0 1 count_Seco~s 156000 11 0 11 11 EstSecondL~w 156000 .005 .073 0 1 count_EstS~s 156000 30 0 30 30 Official_w~w 156000 .002 .044 0 1 count_Offi~s 156000 11 0 11 11 EstOfficia~w 156000 .005 .073 0 1 count_EstO~s 156000 30 0 30 30 Package_wi~w 156000 .002 .044 0 1 count_Pack~s 156000 11 0 11 11 EstPackage~w 156000 .005 .073 0 1 count_EstP.. 156000 30 0 30 30 id 156000 14.514 8.092 1 28
ExcStockRe~n 156000 .256 54.02 -1303.205 3270.275 ExcMktReturn 156000 235.508 8598.533 -1299.248 581000 PredRet_Ky~2 143 .137 .51 -.629 1.669 AbnRet_Kyo~2 143 -.674 8.284 -91.189 20.586 CumAbnRet~o2 156000 -3.568 15.122 -75.653 8.144 PredRet_P~s2 308 .711 .314 .104 1.593 AbnRet_Par~2 308 .35 9.049 -35.152 20.276 CumAbnRet~s2 156000 3.936 10.184 -35.727 19.691 PredRet_Se~2 308 .097 .282 -.629 .762 AbnRet_Sec~2 308 -.557 4.676 -66.83 8.101 CumAbnRet~k2 156000 -6.279 13.728 -70.798 7.57 PredRet_Of~2 308 -40.824 .627 -43.977 -39.365 AbnRet_Off~2 308 19.147 87.106 -170.585 161.353 CumAbnRet~l2 156000 210.592 8.726 192.764 239.081 PredRet_P~e2 308 -1.709 4.373 -10.997 8.387 AbnRet_Pac~2 308 .519 5.598 -18.487 59.93 CumAbnRet~e2 156000 5.816 17.836 -15.692 90.729 AvgCAR_Kyo~2 156000 -3.568 0 -3.568 -3.568 CAR_Kyoto2~d 156000 15.122 0 15.122 15.122 KyotoTest 156000 -.851 0 -.851 -.851 AvgCAR_Par~2 156000 3.936 0 3.936 3.936 CAR_Paris2~d 156000 10.184 0 10.184 10.184 ParisTest 156000 2.045 0 2.045 2.045 AvgCAR_Sec~2 156000 -6.279 0 -6.279 -6.279 CAR_Second~d 156000 13.728 0 13.728 13.728 SecondLeak~t 156000 -2.42 0 -2.42 -2.42 AvgCAR_Off~2 156000 210.592 0 210.592 210.592 CAR_Offici~d 156000 8.726 0 8.726 8.726 OfficialTest 156000 127.711 0 127.711 127.711 AvgCAR_Pac~2 156000 5.816 0 5.816 5.816 CAR_Packag~d 156000 17.836 0 17.836 17.836 PackageTest 156000 1.726 0 1.726 1.726 Variable descriptions name varlab date Date ticker TICKER
environment_score Environment Score datadate Data Date - Daily Prices prccd Price - Close - Daily MktReturns Market Returns Rf_ret Risk-free Return
set Event study construction variable DateKyoto 11 December 1997
DateParis 11 December 2015 DatePackage 20 September 2019 DateSecondLeak 21 June 2016 DateOfficial 6 September 2016 group_id group(ticker set)
Datenum Counts number of trading days StockReturn Stock Return
tKyoto Day number on which Kyoto occurs td_Kyoto Time to/since Kyoto
tdcal_Kyoto Calendar time to/since Kyoto Protocol announcement
tParis Day number on which Paris occurs td_Paris Calendar time to/since Paris Agreement
announcement
announcement
tSecondLeak Day number on which Second Leak occurs td_SecondLeak Time to/since Plan Second leak
tdcal_SecondLeak Calendar time to/since Plan Second leak tOfficial Day number on which Official occurs td_Official Time to/since Plan official draft release tdcal_Official Calendar time to/since Plan official draft release tPackage Day number on which Package occurs
td_Package Time to/since 2019 Climate Package announcement
tdcal_Package Calendar time to/since 2019 Climate Package announcement
Kyoto_window 1 if day is one of the 11 days in event window count_Kyoto_obs Event window day count
EstKyoto_window Normal performance estimation days (30) count_EstKyoto_obs Normal performance estimation window day
count
Paris_window 1 if day is one of the 11 days in event window count_Paris_obs Event window day count
EstParis_window Normal performance estimation days (30) count_EstParis_obs Normal performance estimation window day
count (30)
SecondLeak_window 1 if day is one of the 11 days in event window count_SecondLeak_obs Event window day count
EstSecondLeak_window Normal performance estimation days (30) count_EstSecondLeak_obs Normal performance estimation window day
count
Official_window 1 if day is one of the 11 days in event window count_Official_obs Event window day count
EstOfficial_window Normal performance estimation days (30) count_EstOfficial_obs Normal performance estimation window day
count
Package_window 1 if day is one of the 11 days in event window count_Package_obs Event window day count
EstPackage_window Normal performance estimation days (30) count_EstPackage_obs Normal performance estimation window day
count
ExcStockReturn In excess of risk-free return ExcMktReturn In excess of risk-free return id group(group_id)
PredRet_Kyoto2 Normal Return AbnRet_Kyoto2 Actual-Normal Return CumAbnRet_Kyoto2 Sum of abnormal returns PredRet_Paris2 Normal Return
AbnRet_Paris2 Actual-Normal Return CumAbnRet_Paris2 Sum of abnormal returns PredRet_SecondLeak2 Normal Return
AbnRet_SecondLeak2 Actual-Normal Return CumAbnRet_SecondLeak2 Sum of abnormal returns PredRet_Official2 Normal Return
AbnRet_Official2 Actual-Normal Return CumAbnRet_Official2 Sum of abnormal returns PredRet_Package2 Normal Return
AbnRet_Package2 Actual-Normal Return CumAbnRet_Package2 Sum of abnormal returns AvgCAR_Kyoto2 Average CAR
CAR_Kyoto2_sd Standard Deviation
KyotoTest Test for significance of CARs AvgCAR_Paris2 Average CAR
CAR_Paris2_sd Standard Deviation
ParisTest Test for significance of CARs AvgCAR_SecondLeak2 Average CAR
