Carbon Disclosure and the Cost of Debt

Size: px

Start display at page:

Download "Carbon Disclosure and the Cost of Debt"

Hope Wilcox
5 years ago
Views:

1 Carbon Disclosure and the Cost of Debt Michael Viehs University of Oxford Smith School of Enterprise and the Environment European Centre for Corporate Engagement Stefanie Kleimeier Maastricht University European Centre for Corporate Engagement University of Stellenbosch Business School Maastricht, 11 March 2015 ECCE Webinar

Carbon Disclosure and the Cost of Debt Topic: Responsible/ethical lending Carbon emissions in the context of corporate loans Major research question: To what extent are voluntary carbon

2 Carbon Disclosure and the Cost of Debt Topic: Responsible/ethical lending Carbon emissions in the context of corporate loans Major research question: To what extent are voluntary carbon disclosure and actual CO 2 emissions related to the costs of debt? Major findings: Voluntary disclosers pay significantly lower loan spreads High polluters pay significantly higher loan spreads

3 Context: Socially Responsible Investing

4 Socially Responsible Investing (SRI) Different responsible investment approaches Exclusionary rules: Norms-based investing Inclusion based on positive screening Engagement on ESG (environmental, social, governance) issues Consequence of exclusion: Neglect effect Investors will not invest Or only violate their norms for a price (higher return) Evidence: Hong and Kacperczyk, 2009JFE; Heinkel, Kraus & Zechner, 2001JFQA, or Gollier and Pouget, 2012WP

literature on sustainability and firm-level effects Have superior CSR/ESG

5 Relevance to Investors? Meta-study by Clark, Feiner, and Viehs (2015) on SRI: Overview of the literature on sustainability and firm-level effects Have superior CSR/ESG practices any effects on firms? If so, which? Review of more than 200 sources (academic and industry) Updated in March 2015, available on SSRN!

6 From the Stockholder to the Stakeholder Highlights of the report 90% of the cost of capital studies show that sound ESG standards are associated with lower cost of capital 88% of the reviewed studies show that solid ESG practices are associated with better operational performance 80% of the studies show that stock price performance is positively related to good CSR practices

7 What s New? Evidence on the implications of superior CSR/ESG standards on cost of capital (also debt), but Using indirect measures of environmental performance Mostly US-based only Focus on investor preferences (norms), and less focus on risk Our contributions are Direct measure of environmental performance: Disclosure and level of CO 2 emissions Global sample: Firms from 87 countries Differentiation between investor preferences and risk

8 Research Question What determines a company s cost of debt beyond credit risk? Company s environmental performance and disclosure thereof? More specific research questions Does the voluntary disclosure of a company s CO 2 emissions affect its cost of debt? Is the company s actual amount of CO 2 emissions related to its cost of debt?

9 Research Hypotheses H1 Voluntary CO 2 emissions disclosure is negatively associated with corporations cost of debt. Driver: Information uncertainty

10 Research Hypotheses H2 The actual amount of CO 2 emissions by the borrower is positively associated with the cost of debt. Driver: Risk premium

11 Research Hypotheses H3 The rise in the cost of bank debt is driven by those loans which have CDP signatories as lead arrangers. Driver: Norms-constrained investor

Data CO 2 emissions from CDP (Carbon Disclosure Project) London-based not-for-profit organization Annual survey to almost 5,000 public companies worldwide Did company answer questionnaire?

12 Data CO 2 emissions from CDP (Carbon Disclosure Project) London-based not-for-profit organization Annual survey to almost 5,000 public companies worldwide Did company answer questionnaire? Scope 1 emissions: Direct emissions Scope 2 emissions: Emissions from electricity purchased (supply chain) 767 financial institutions as signatories Cost of debt from Dealscan Provided by Thomson Reuters Loan Pricing Corporation Bank loans to private and public borrowers worldwide More than 250,000 loans to almost 70,000 borrowers from 1987 to 2012 Loan, borrower & lender information

13 Resulting Sample 9,260 loans signed between 2008 and 2012 Loans to public companies These borrowers could have been targeted by CDP, e.g. could have received questionnaire 3,819 borrowers in 87 countries 6,997 loans to borrowers who received questionnaire; 2,263 loans to borrowers who did not receive questionnaire 6,616 loans for which at least one of the lead arrangers is CDP signatory; 2,644 loans for which none of the lead arrangers is CDP signatory

14 First Impressions

15 Methodology OLS regressions with clustered SEs at borrower level ln(loan spread) = β 1 Targeted by CDP + β 2 (Company Response to CDP Q) +β 3 (Signatory controls) + β 4 (Firm and loan controls) + β 5 (Country and time controls) ln(loan spread) = β 1 CO2 emissions +β 2 (Signatory controls) + β 3 (Firm and loan controls) + β 4 (Country and time controls) We always control for Dummies: Multiple-tranche deal, performance pricing, financial covenants, multiple base rate, base rate, year of loan signing, loan purpose, tranche type, loan seniority, loan security, borrower's senior debt rating, borrower industry, borrower country Environmental performance index by country and year

16 Results: Voluntary Disclosure ln(spread in bp) (1) (2) (3) (4) (5) (6) Dummy controlling for selection of firms by CDP Borrower received questionnaire *** *** *** *** *** Dummies indicating response to CDP questionnaire Borrower answered questionnaire *** Borrower provided some info ** Borrower did not respond 0.10 *** 0.12 *** Borrower declined to participate 0.10 ** 0.15 *** Borrower, loan, country, time controls Yes Yes Yes Yes Yes Yes Adjusted R-squared Observations 9,260 9,260 9,260 9,260 9,260 9,260 *** 1%, ** 5%, * 10% significance.

Results: Voluntary Disclosure ln(spread in bp) (1) (2) (3) (4) (5) (6) Dummy controlling for selection of firms by CDP Borrower received questionnaire -0.10 *** -0.03-0.11 *** -0.

15 ** Borrower did not respond Average loan 0.10 size *** = $ 446 mio 0.12 *** Borrower declined to participate 0.10 ** 0.

17 Results: Voluntary Disclosure ln(spread in bp) (1) (2) (3) (4) (5) (6) Dummy controlling for selection of firms by CDP Borrower received questionnaire *** *** *** *** *** Dummies indicating response to CDP questionnaire Borrower answered questionnaire *** Economic Significance Borrower provided some info ** Borrower did not respond Average loan 0.10 size *** = $ 446 mio 0.12 *** Borrower declined to participate 0.10 ** 0.15 *** Average spread = 249 bp Borrower, loan, country, time controls Yes Yes Yes Yes Yes Yes Spread paid by average borrower = $ 11.1 mio p.a. Adjusted R-squared Observations 9,260 9,260 9,260 9,260 9,260 9,260 *** 1%, ** 5%, * 10% significance. Spread saved by disclosing borrower = $ 1.4 mio p.a.

18 Results: Voluntary Disclosure & Signatories ln(spread in bp) (1) (2) Dummy controlling for selection of firms by CDP Borrower has been send questionnaire by CDP Dummy indicating response to CDP questionnaire Borrower answered questionnaire (D AQ ) *** *** Dummy for presence of CDP signatories among lead arrangers At least one CDP signatory is among lead arrangers (D signatory ) *** *** Interaction effect: Response to CDP questionnaire & CDP signatory D signatory *D AQ 0.08 Borrower, loan, country, time controls Yes Yes Adjusted R-squared Observations 9,260 9,260 *** 1%, ** 5%, * 10% significance.

19 Results: Voluntary Disclosure & Credit Rating ln(spread in bp) Loans without rating Loans with rating Dummy controlling for selection of firms by CDP Borrower has been send questionnaire by CDP Dummy indicating response to CDP questionnaire Borrower answered questionnaire (AQ) *** *** Borrower, loan, country, time controls Yes Yes Adjusted R-squared Observations 5,521 3,739 *** 1%, ** 5%, * 10% significance.

20 Results: Voluntary Disclosure H1 Voluntary CO 2 emissions disclosure is negatively associated with corporations cost of debt. Relevance of information uncertainty Effect for all lenders Effect stronger for unrated loans

21 Transportation, public utilities Mining Agriculture, forestry, fishing Public administration Manufacturing Wholesale trade Construction Retail trade Services Finance, insurance, real estate Actual Emissions Median CO 2 emissions (tons per $ mio of assets) Scope 1 Scope 2

22 Results: Actual Emissions & Signatories ln(co 2 scope 1 ln(co 2 scope 2 emissions) emissions) (1) (2) (3) (4) CO 2 emissions of the borrower Industry and firm-size adjusted CO 2 emissions (CO2) 0.02 * Dummy for presence of CDP signatories among lead arrangers At least one CDP signatory is among lead arrangers (D signatory ) Interaction effect: CO 2 emissions & CDP signatory ln(spread in bp) Panel A: Panel B: D signatory *CO * 0.04 Borrower, loan, country, time controls Yes Yes Yes Yes Adjusted R-squared Observations 1,022 1, *** 1%, ** 5%, * 10% significance.

Economic Significance Results: Actual Emissions & Signatories Average loan size = $ 1.61 bn Average spread = 161.5 bp Spread paid p.a. = $ 26 mio Reducing CO 2 emissions by 1% leads to spread savings of $ 1.

02 Dummy for presence of CDP signatories among lead arrangers At least one CDP signatory is among lead arrangers (D signatory ) 0.08 0.

23 Economic Significance Results: Actual Emissions & Signatories Average loan size = $ 1.61 bn Average spread = bp Spread paid p.a. = $ 26 mio Reducing CO 2 emissions by 1% leads to spread savings of $ 1.5 mio p.a. ln(co 2 scope 1 ln(co 2 scope 2 emissions) emissions) (1) (2) (3) (4) CO 2 emissions of the borrower Industry and firm-size adjusted CO 2 emissions (CO2) 0.02 * Dummy for presence of CDP signatories among lead arrangers At least one CDP signatory is among lead arrangers (D signatory ) Interaction effect: CO 2 emissions & CDP signatory ln(spread in bp) Panel A: Panel B: D signatory *CO * 0.04 Borrower, loan, country, time controls Yes Yes Yes Yes Adjusted R-squared Observations 1,022 1, *** 1%, ** 5%, * 10% significance.

24 Results: Actual Emissions, Signatories & Credit Ratings CO 2 emissions of the borrower no rating ln(spread in bp) inv. grade rating rating below inv. grade rating Industry and firm-size adjusted CO 2 emissions (CO2) Dummy for presence of CDP signatories among lead arrangers At least one CDP signatory is among lead arrangers (D signatory ) Interaction effect D signatory *CO ** 0.13 ** 0.00 Borrower, loan, country, time controls Yes Yes Yes Yes Adjusted R-squared Observations Fraction of loans without CDP signatories 15.0% 9.4% 8.8% 11.8% Median CO 2 emissions (in tons per $ mio assets, industry adjusted) *** 1%, ** 5%, * 10% significance.

25 Results: Actual Emissions H2 The actual amount of CO 2 emissions by the borrower is positively associated with the cost of debt. H3 The rise in the cost of bank debt is driven by those loans which have CDP signatories as lead arrangers. Evidence for Investor Preference Hypothesis Puzzle: Strongest effect for investment grade borrowers

26 Implications Implications for borrowers Managing environmental performance can lower cost of debt by voluntarily disclosing environmental information by reducing CO 2 emission levels However, selection of lenders matters Implications for lenders Norm-constrained investors can charge an environmental premium

27 Carbon Disclosure and the Cost of Debt Michael Viehs University of Oxford Smith School of Enterprise and the Environment European Centre for Corporate Engagement Stefanie Kleimeier Maastricht University European Centre for Corporate Engagement University of Stellenbosch Business School Maastricht, 11 March 2015 ECCE Webinar

Carbon Disclosure, Emission Levels, and the Cost of Debt *

Carbon Disclosure, Emission Levels, and the Cost of Debt * Stefanie Kleimeier Maastricht University - School of Business and Economics University of Stellenbosch Business School Open Universiteit The European