DISCUSSION PAPER December 2014 RFF DP 14-43 Voluntary Environmental Information Disclosure and Firm Size Evidence from the Hydraulic Fracturing Chemical Registry FracFocus Zhongmin Wang 1616 P St. NW Washington, DC 20036 202-328-5000 www.rff.org Voluntary Environmental Information Disclosure and Firm Size: Evidence from the Hydraulic Fracturing Chemical Registry FracFocus Zhongmin Wang Abstract Most enterprises are small- and medium-sized firms, but scholarship on corporate social responsibility (CSR) has focused on large corporations. In this paper, I study small, medium, and large firms' likelihood of engaging in a specific CSR activity—voluntary environmental information disclosure. I present evidence that in the oil and gas industry, large firms are more likely than small firms to voluntarily disclose to the general public information about the industrial process of hydraulic fracturing. Key Words: voluntary environmental information disclosure, corporate social responsibility, hydraulic fracturing, firm size JEL Classification Numbers: D21, M14, Q50 © 2014 Resources for the Future. All rights reserved. No portion of this paper may be reproduced without permission of the authors. Discussion papers are research materials circulated by their authors for purposes of information and discussion. They have not necessarily undergone formal peer review. Contents 1. Introduction ........................................................................................................................ 1 2. Industry Background, Data, and Results ........................................................................ 2 3. Conclusion .......................................................................................................................... 5 References ................................................................................................................................ 6 Resources for the Future Wang Voluntary Environmental Information Disclosure and Firm Size: Evidence from the Hydraulic Fracturing Chemical Registry FracFocus Zhongmin Wang 1. Introduction Suppose firms of all sizes are engaged in an industrial process that has the potential to pollute the local environment. Are small or large firms more likely to voluntarily disclose information about the industrial process to the general public? Voluntary information disclosure is a type of corporate social responsibility (CSR) activity (Kitzmueller and Shimshack 2012). According to Commission of the European Communities (2006, 2), CSR “is about enterprises deciding to go beyond minimum legal requirements and obligations … to address societal needs. Through CSR, enterprises of all sizes … can help to reconcile economic, social and environmental ambitions.” The EU definition specifically mentions enterprises of all sizes and most firms are small- and mediumsized enterprises (SMEs), but scholarship on CSR has focused on large corporations. As noted by Kim and Lyon (2011), most of the empirical work on voluntary environmental disclosure focuses on large corporations’ annual reports and appears in the accounting literature. A small number of economic studies (e.g., Arora and Cason 1996, Videras and Alberini 2000, and Kim and Lyon 2011) find that large corporations’ likelihood of participating in voluntary environmental abatement or disclosure programs increases with their size. However, it is not clear whether this finding extends to firms of all sizes. A small but growing literature in business ethics (e.g., Vázquez-Carrasco and López-Pérez 2012 and Smith 2013) recognizes that small firms differ from large corporations when it comes to CSR: they are subject to less public scrutiny and have fewer resources, but they are often deeply involved in their local communities and it is easier for them to commit to CSR because they are often managed by owners. These considerations suggest that theoretical reasoning alone cannot provide a convincing answer to the question I pose. Fellow, Resources for the Future, 1616 P St. NW, Washington DC, 20036; wang@rff.org, 202.328.5036. This reseach did not receive any outside funding. I thank DrillingInfo and Pennsylvania Department of Conservation and Natural Resources for providing data, Tom Lyon and Eun-Hee Kim for helpful comments, and Alex Egorenkov, Xu Liu, and Kuangyuan Zhang for research assistance. Any remaining errors are mine only. 1 Resources for the Future Wang In this paper, I present empirical evidence from the oil and gas industry that large firms are more likely than small firms to voluntarily disclose to the general public information about the industrial process of hydraulic fracturing, which involves the use of toxic chemicals. 2. Industry Background, Data, and Results Hydraulic fracturing is a key industrial process used to extract shale gas or tight oil. Shale gas is natural gas produced from shale formations, and tight oil is oil produced from shale or other formations with very low permeability. Shale gas and tight oil in the past decade or so has experienced an extraordinary boom in production that has “dramatically changed the energy future of the United States and potentially of the world” (Joskow 2013, 339). The top 13 US states in shale gas production in 2012, according to Energy Information Administration (EIA) data, are shown in column 1 of Table 1. Hydraulic fracturing involves injecting large volumes of fluid at high pressure into the rock to create fractures through which oil and gas may flow out. Chemicals constitute a small fraction of the fracturing fluid, but their total volume can be substantial. Concerned that those chemicals could pollute the environment, stakeholders have called for oil and gas operators to disclose the chemicals they use to fracture each well. Table 1. States, Disclosure Regulations, and Sample Disclosure regulation State Texas Effective date 02/01/2012 Pennsylvania 04/14/2012 Louisiana Arkansas 10/20/2012 01/15/2011 Oklahoma 01/01/2012 West Virginia North Dakota Michigan California Montana New Mexico Colorado Wyoming 12/11/2011 04/10/2012 06/22/2011 01/01/2014 08/26/2011 02/15/2012 04/01/2012 09/15/2010 Where to disclose? FracFocus State and FracFocus State or FracFocus State State and FracFocus State FracFocus State FracFocus State or FracFocus State FracFocus State Sample Period: 04//11/11 to 01/31/12 04/13/12 Number of wells in the sample 1,987 1,160 12/31/13 1,057 No data 12/31/11 464 12/31/13 510 04/09/12 1,497 No data 12/31/13 1,456 12/31/13 354 12/31/13 1,212 03/31/12 1,582 12/31/13 274 Notes: The listed states are the top 13 in shale gas production in 2012. The effective dates and where-to-disclose information are from state regulations. 2 Resources for the Future Wang The oil and gas industry responded to these calls by supporting the creation of the fracturing chemical registry website, www.FracFocus.org, which went online on April 11, 2011. Operators can post on this website the location of a well and the chemicals used. The 13 states in Table 1 all responded by passing fracturing disclosure regulations. These regulations may require operators to disclose the information to FracFocus (e.g., Texas), to FracFocus and a state agency (e.g., Pennsylvania), to a state agency (e.g., West Virginia), or to either FracFocus or a state agency (e.g., Louisiana). Before the regulations were passed, disclosure to FracFocus was voluntary in all four cases. After the regulations were passed, disclosure to FracFocus became mandatory in the first two cases and voluntary in the third case. In the fourth case, disclosure to FracFocus is semi-voluntary. It is much easier for the public to access the FracFocus website than to request information from a state agency; state agencies do not post disclosed chemical information on the Internet. The disclosure regulations’ timing and where-to-disclose requirements are shown in columns 2 and 3 of Table 1. I focus on the states and the periods for which disclosure to FracFocus is voluntary or semi-voluntary. Data availability limits my sample to 11 of the 13 states. The sample starting date is always the date on which FracFocus went online. The sample ending dates, listed in column 4 of Table 1, are either December 31, 2013 or the day immediately before the effective date of a regulation that mandates disclosure to FracFocus. The sample of wells for nine states comes from DrillingInfo, a market research firm whose data are often used by EIA. My sample includes all the oil and gas wells drilled into shale formations; such wells were all fractured. The Pennsylvania data come from the Pennsylvania Department of Conservation and Natural Resources, and the Louisiana data come from both DrillingInfo and the Louisiana Department of Natural Resources. I use the date on which the drilling of a well was completed to decide whether to include a well in the sample. The number of sampled wells for each state is listed in column 5 of Table 1. The number of sampled wells is quite large, at 11,553, and these wells were drilled by 388 firms. The number of wells drilled per firm ranges from 1 to 1,035. A total of 213 firms drilled 4 wells or fewer, 43 firms drilled between 5 and 9 wells, 97 firms drilled between 10 and 99 wells, and 35 firms drilled 100 or more wells. The firms that drilled only a few wells are almost always small private firms with a few employees (Independent Petroleum Association of America 2012-2013), and the firms that drilled a large number of wells are either medium-sized enterprises or large corporations. My measure of firms’ sizes is the number of wells they drilled, which is essentially a measure of their market share in drilling. There are no cases in which a firm is a large oil and gas producer but drilled only a small number of wells in my sample. I also 3 Resources for the Future Wang used the number of wells drilled in 2009 and 2010 as an alternative measure of firm size, and the results, not reported here, are quite similar. The basic information for every well disclosed at FracFocus, up to early May, 2014, was downloaded from the website using self-written Perl scripts. Legal scholars (e.g., Konschnik et al. 2013) argue that FracFocus is far from ideal (because, for example, firms’ claim of trade secrets may be too broad), but they do not dispute that the information disclosed at the website is useful. Because every well has a unique ID called the API number, I can easily identify whether a well in my sample is disclosed at FracFocus. 60 disclosure_percent 20 40 lowess disclosure_percent size 0 % disclosed at FracFocus 80 100 Figure 1. Disclosure Rate and Firm Size 0 200 400 600 Firm size (# of wells drilled) 800 1000 Figure 1 plots the percentage of wells disclosed against firm size and also shows the associated LOWESS (locally weighted scatterplot smoothing) curve. The disclosure rate appears to be an increasing and concave function of firm size. For example, the average disclosure rate for the 122 firms with a single well is 18.9 percent (23 of 122 wells), and the average disclosure rate for the 35 firms with 100 wells or more is 68.7 percent. I use the logit model to estimate how firm size affects disclosure probability. The logit model controls for time trend, which is defined as the number of days between a well’s completion date and FracFocus’ first date of operation. I do not attempt to control for factors that may explain variations in disclosure rate across states because my main finding holds up even if I run the logit model for each individual state separately. Table 2 reports the maximum likelihood estimates of five logit models. Model 1 considers the full sample of wells. Model 2 considers states and time periods without disclosure regulations, and Model 3 considers states and time periods with regulations. Models 4 and 5 further separate the states with disclosure regulations into voluntary and semi-voluntary states. Due to space considerations, the results for 4 Resources for the Future Wang the individual states are not reported here. The estimated coefficients for firm size are positive and highly statistically significant in all five models, and the estimated coefficients for the square of firm size are negative and highly statistically significant in four of the five models. These results indicate that firm size has a positive impact on the odds of disclosure in all types of states and that the marginal effect tends to decline with firm size. The time trend could be positive or negative. Table 2. Estimates of Logit Models Model 1 Model 2 No disc. Full sample Regulation Size Size^2 Time trend Constant Model 3 With disc. regulation Model 4 State agency 0.0048*** 0.0045*** 0.0082*** 0.0029*** (0.00026) (0.00029) (0.00071) (0.00083) -3.4e-06*** -3.6e-06*** -4.2e-06*** 1.6e-06* (2.4e-07) (2.7e-07) (6.7e-07) (9.7e-07) 0.00034*** 0.00054*** -0.0018*** -0.00055** (0.000086) (0.00013) (0.00022) (0.00028) -0.33*** -0.34*** 0.55*** -0.23 (0.043) (0.047) (0.14) (0.18) Model 5 Either state agency or FracFocus 0.018*** (0.0021) -0.000014*** (1.9e-06) -0.0021*** (0.00064) 1.14*** (0.33) Observations 11,553 9,017 2,536 1,574 962 Notes: Dependent variable is whether a well was disclosed at FracFocus. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 3. Conclusion My results indicate that larger firms are more likely to voluntarily disclose information at FracFocus, whether the firms are operating in states and time periods with or without specific voluntary fracturing disclosure regulations. This finding suggests that the existing result that size affects large corporations’ likelihood of engaging in CSR activities extends to firms of all sizes. It would be interesting for future research to investigate the mechanisms underlying my finding. My finding does not imply that small firms are inherently less socially responsible. Small firms may or may not undertake other CSR activities that I do not study in my paper. 5 Resources for the Future Wang References Arora, Seema, and Timothy N. Cason. 1996. “Why Do Firms Volunteer to Exceed Environmental Regulations? Understanding Participation in EPA’s 33/50 Program.” Land Economics 72(4): 413-32. Commission of the European Communities. 2006. “Implementing the Partnership for Growth and Jobs: Making Europe a Pole of Excellence on Corporate Social Responsibility.” Available at http://eurlex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2006:0136:FIN:en:PDF Independent Petroleum Association of America. 2012-2013. “Profile of Independent Producers.” Washington DC. Joskow, Paul. 2013. “Natural Gas: From Shortage to Abundance in the United States.” American Economic Review: Paper and Proceedings 103(3): 338–43. Kim, Eun-Hee, and Thomas P. Lyon. 2011. “Strategic Environmental Disclosure: Evidence from the DOE’s Voluntary Greenhouse Gas Registry.” Journal of Environment Economics and Management 61: 311-326 Kitzmueller, Markus, and Jay Shimshack. 2012. “Economic Perspectives on Corporate Social Responsibility.” Journal of Economic Literature 50(1): 51-84. Konschnik, Kate, Margaret Holden, and Alexa Shasteen. 2013. “Legal Fractures in Chemical Disclosure Laws: Why the Voluntary Chemical Disclosure Registry FracFocus Fails as a Regulatory Compliance Tool.” Available at http://blogs.law.harvard.edu/environmentallawprogram/files/2013/04/4-23-2013LEGAL-FRACTURES.pdf Smith, N. Craig. “When It Comes to CSR, Size Matters.” Available at http://www.forbes.com/sites/insead/2013/08/14/when-it-comes-to-csr-size-matters/ Vázquez-Carrasco, Rosario, and M. Eugenia López-Pérez. 2012. “Small & Medium-Sized Enterprises and Corporate Social Responsibility: A Systematic Review of the Literature.” Quality and Quantity 47: 3205-18. Videras, Julio, and Anna Alberini. 2000. “The Appeal of Voluntary Environmental Programs: Which Firms Participate and Why?” Contemporary Economic Policy 18(4): 449-61. 6