What we can learn by linking firms’ reported emissions with their financial data

Introduction

Governments have intensified their commitments to reduce greenhouse gas (GhG) emissions over the past few decades.1 These commitments are likely to affect firms in sectors with high emissions, yet little is known about the financial characteristics of these firms.

We address this knowledge gap by creating an innovative dataset that links the subset of Canadian-incorporated publicly traded firms that report emissions with their financial statements. We then present facts related to these firms and assess whether reported emissions affect their stock prices.

Key findings

• GhG emissions are concentrated within a few firms and industries. In 2019, the 10 largest emitters accounted for about one-half of all reported emissions, while firms in two industries (electricity, gas and sanitary services; oil and gas extraction) accounted for about two-thirds of GhG emissions in our sample.
• Equity financing is typically the largest funding source for firms that report emissions. In our sample of firms, equity accounts for 57% of total funding. Bonds are the second largest funding source, at 37%, followed by bank loans, at 6%.
• Increasingly, equity investors appear to be considering a firms’ GhG emissions in their investment decisions. Investors apply a higher discount to firms with larger emissions, but the discounts do not seem economically meaningful. This suggests that the risk of a sudden repricing of assets exposed to climate change remains a concern for financial stability.

However, our results should be thought of as suggestive due to significant data gaps.2 For instance, many firms do not report emissions, nor are firms required to report their exposure to the physical risks of climate change. As well, the data on emissions in our estimation ends in 2020. Since then, awareness on climate-related risks has grown, and whether our results would still hold today is unclear.

The dataset

In a first step, we create a composite emissions dataset by combining information reported by Morgan Stanley Capital International (MSCI) and Refinitiv.3 Both data providers capture the emissions that firms report in their financial statements.4 We start with MSCI, which has a broader coverage, and then use Refinitiv to fill gaps. The reported values are usually identical when a firm is covered by both sources.

Both data sources account for scope 1, 2 and 3 emissions.5 To avoid double counting, we include only scope 1 emissions in the charts that portray total emissions. The regressions we estimate include a firms’ scope 1 and 2 emissions because we assume investors would consider both when making investment decisions. We exclude scope 3 from our analysis because few firms report those emissions due to challenges in measuring them.

We then match our composite emissions dataset with firm-level annual financial statements for publicly listed firms in Canada from FactSet.6 We use International Securities Identification Numbers to link the two datasets.7 We can match about 280 firms (or 28% of firms) in FactSet with the composite emissions data we construct. This relatively low match rate is due to the limited number of firms that report emissions.8

This final dataset provides complete financial statements and the level of emissions for each matched firm over 2008–20. The data also contain some financial market information, such as the firm’s price-to-book ratio, which is a key variable in our regressions.

Facts about firms that report emissions

Our data reveal two interesting facts.

First, a small number of firms in a few industries generated most of the scope 1 emissions in our dataset. Chart 1 shows that, in 2019, the 10 highest-emitting firms (out of 280) accounted for about 50% of the total emissions in our dataset, and the 50 largest accounted for nearly 95%.

Chart 1: Relatively few firms generate the bulk of emissions

The concentration in emissions is also apparent at the industry level.9 Chart 2 shows that two industries (out of 46) accounted for just under 65% of reported emissions in 2019. Firms in the electric, gas and sanitary services industry generated the largest share of reported emissions, followed by firms in the oil and gas extraction industry.10

Chart 2: Reported emissions are concentrated in two industries

Second, firms that report emissions typically use equity as their main funding source. Chart 3 shows the composition of funding sources across industries in our dataset. Equity is the largest funding source, followed by bonds and then loans.11 Firms that report emissions in the electric, gas and sanitary services and the transportation by air industries are the exception to this rule. Bonds are their primary funding source.12

Chart 3: Equity is often the main funding source for firms that report emissions

Effects on investment decisions

We investigate whether equity investors consider a firm’s GhG emissions when making investment decisions. We focus on equity investors for two reasons:

• Equity is typically the largest source of funding for firms that report emissions.
• Equity prices may be more sensitive than the prices for bonds and loans if the transition toward net-zero emissions reduces the income of a firm that generates GhGs. This is because equity holders receive dividend payments after all required interest and principal payments are made on bonds and loans.

We estimate the impact on a firm’s price-to-book ratio from a change in emissions, after controlling for other firm-level characteristics. This is done using a panel regression with fixed effects over the sample period from 2008 to 2020.13 We use the price-to-book ratio as our dependent variable because investors commonly use it to assess a firm’s potential for growth.14 Specifically, the price-to-book ratio can be broken down into two parts:

• The numerator represents the market price of a firm’s stock. The market price is a forward-looking measure that captures equity investors’ expectations about the discounted present value of a firm’s future cash flows. Thus, stock prices should incorporate the anticipated impact of climate policies on firms’ future income and how risky or uncertain this future income could be (through the discount rate).
• The denominator represents the book value per share of a firm’s equity. It is a proxy for the dollar value that each share of equity would receive if the firm were liquidated today, after all assets are sold off and debts are paid. Book values are mostly backward-looking (i.e., based on historical accounting costs) and do not generally consider expectations of future income.

Our regression specification allows for the estimated impact of additional GhG emissions on the market pricing of a firm to change over time and based on its level of emissions (i.e., the impact can be different for large emitters than it is for small ones). The Appendix specifies the exact regression we estimated.15

Chart 4 illustrates this impact for three types of firms based on their level of emissions in each year: the firms at the 10^th, 50^th (median) and 90^th percentile in terms of emissions.

Chart 4: Equity investors are increasingly taking emissions into account

Overall, our results suggest that equity investors may be becoming increasingly concerned with the future income potential of firms that emit GhGs, particularly for those with emissions above the median. Equity investors started applying a discount to these firms’ price-to-book ratio in 2009 and for firms with relatively low emissions (i.e., those in the 10^th percentile of emissions) a few years later. The discount has been growing since.16

As well, the discount for high emitters is more sensitive to further increases in emissions. This finding suggests that climate policies will have a greater impact on higher emitting firms than firms that emit less. This result is intuitive. The stock market price of a firm captures the present discounted value of future profits. Investors may perceive large emitters as less profitable because climate-pricing schemes are more likely to affect firms in higher polluting industries (Bolton and Kacperczyk 2021).

However, we assess the discount to be relatively small. For example, if a firm with the median level of emissions in 2019 were to cut its emissions by half (which would likely be costly), it would be rewarded with an increase in its price-to-book ratio of only 4.4%.17 This discount appears small when compared with usual movements in a firm’s price-to-book ratio. For instance, the average matched firm in our dataset has an annual standard deviation in its price-to-book ratio of around 40%. Thus, the reward for cutting emissions in half is roughly one-tenth the size of the typical movement in the price-to-book ratio.18 The discount investors apply therefore does not appear to be economically significant.

Several reasons could explain why investors only marginally discount firms’ production emissions:

• Investors may believe that:
  • companies are credible when they pledge to reduce their dependence on fossil fuels and that this will have minimal impact on profitability
  • companies can avoid climate policies by shifting their production (when possible) and the recording of emissions to locations with less stringent climate policies
• In contrast, investors may have difficulty assessing the future profitability of firms that generate GhGs due to:19
  • the high level of uncertainty associated with the transition to a low-carbon global economy, including transition paths the economy could follow and the full set of policies that will be required
  • the lack of information on the exposure to climate-related risks and transition plans for most firms

In the 2022 Financial System Review, the Bank of Canada raised a concern about the possibility that assets exposed to climate-related risks could experience a sudden repricing. This repricing would reflect changes in investors’ beliefs due to shifts in global climate policies or better information on firm-level exposures to climate risks, for example. Our finding that the discount applied to GhG emissions by equity investors in Canada is economically small appears to support this concern.

Conclusion

This study is part of a large body of work by the Bank to better understand the risks that climate change poses to financial stability. In this study, we combined information from several sources to create a new dataset of publicly traded firms incorporated in Canada that report GhG emissions. From these data, we find that:

• a small number of firms generate the majority of emissions
• emitting firms rely primarily on equity financing
• equity investors are increasingly discounting stock prices due to emissions
• equity investors apply greater discounts to firms with large emissions

Nevertheless, the impact of emissions on equity prices appears small, suggesting that assets exposed to climate change remain at risk of a sudden repricing.

However, it is important to highlight a few caveats to our analysis:

• The data on emissions in our estimation end in 2020. Since then, the global financial community’s awareness about climate considerations has continued to grow. Therefore, whether our results would hold when assessed using more recent data is unclear.
• Significant data gaps remain, which complicates the assessment of the exposure of the non-financial sector to climate change, despite our best efforts. The main source of data gaps reflects the lack of standardized and systematic reporting by firms of their GhG emissions. Domestic and international initiatives to reduce these gaps are underway, which would help improve future assessments of firms’ vulnerabilities to climate change.
• We do not consider the exposure of firms to physical climate change. Firms are not required to provide data on their exposure to physical risks, such as whether their production facilities or headquarters are in areas prone to floods or wildfires.

Given these caveats, our results are suggestive rather than definitive.

Appendix: Specification of the regression and estimation results

We estimate the following equation:

\(\displaystyle log⁡({price-to-book\ ratio}_{i,t})\) \(\displaystyle=\, \beta_0\) \(\displaystyle+\,\beta_1\ {log(emissions}_{i,t-1})\) \(\displaystyle+\,\beta_2\ {(log⁡(emissions}_{i,t-1}))^2\) \(\displaystyle+\,\beta_3\ (\delta_t\ ×\ log⁡(emissions_{i,t-1}))\) \(\displaystyle+\,\beta_4 X_{i,t-1}\) \(\displaystyle+\,\alpha_i\) \(\displaystyle+\,\delta_t\) \(\displaystyle+\,\epsilon_{i,t}\)

Our regression includes three variables related to a firm’s emissions:

• \(\displaystyle{log(emissions}_{i,t-1})\) captures the impact of the (log) level of emissions on the firm’s price-to-book ratio
• \(\displaystyle{(log⁡(emissions}_{i,t-1}))^2\) allows for the possibility that the sensitivity of the price-to-book ratio to GhG emissions could be greater for firms that generate relatively large amounts of GhG emissions
• \(\displaystyle{\delta_t}\ ×\ log⁡(emissions_{i,t-1})\)—the product of a time dummy variable and the log of total emissions—allows for the possibility that the impact of emissions on the firm’s price-to-book ratio could vary over time. This could reflect the adoption of new climate policies or changes in investor sentiment relative to GhG emissions.

We also include firm- and time-fixed effects \(\displaystyle{(\alpha_i,{\delta_t})}\) as well as several controls \(\displaystyle(X_{i,t-1})\), which include:

• the log of firm assets
• the annual growth rate of firm sales
• the ratio of total debt over total assets

We lag emissions by one period to reflect the delay in their reporting time.

Table A-1 reports our results across different specifications. The first column presents our main results, while the remaining columns show results when we exclude certain variables. Numbers in parenthesis are standard errors associated with our parameter estimates.

References

Bank of Canada. 2022. “Climate change considerations.” Financial System Review–2022.

Bolton, P. and M. Kacperczyk. 2021. “Do Investors Care About Carbon Risk?” Journal of Financial Economics 142 (2): 517–549.

Frankel, K., M. Shakdwipee and L. Nishikawa. 2015. “Carbon Footprinting 101: A Partial Guide to Understanding and Applying Carbon Metrics.” New York: Morgan Stanley Capital International.

Government of Canada. 2021. “Timeline: Canada’s Domestic and International Climate Change Milestones.”

Government of Canada. 2022. “Canada’s Climate Plans and Targets.”

Government of Canada. 2022. “Greenhouse Gas Reporting Program (GHGRP)—Facility Greenhouse Gas (GHG) Data.”

Iwata, H. and K. Okada. 2011. “How Does Environmental Performance Affect Financial Performance? Evidence from Japanese Manufacturing Firms.” Ecological Economics 70 (9): 1691–1700.

Karniol-Tambour, K., D. Hochman, J. Ng and L. Pinasco. 2022. “Where Do Greenhouse Gas Emissions Come From, and What Does That Mean for Investors?” Bridgewater Associates LP.

Konar, S. and M. A. Cohen. 2001. “Does the Market Value Environmental Performance?” Review of Economics and Statistics 83 (2): 281–289.

Refinitiv. 2021. “Refinitiv ESG Carbon Data and Estimate Models.”

World Business Council for Sustainable Development and World Resources Institute. 2004. “The greenhouse gas protocol.” In A corporate accounting and reporting standard, Rev. ed. Washington, DC, Conches-Geneva.

Acknowledgements

We thank Marie-Christine Tremblay, Russell Barnett and Louis Morel for helpful comments.

Endnotes