ESG Shareholder Engagement and Downside Risk

European Commission Conference ‘Promoting Sustainable Finance’, Brussels, Jan 9th 2019

Presented previously at AFA 2018, IAF, MFS, Pictet, Q-Group, SOAS, SSE, and UCD

Andreas G. F. Hoepner Smurfit Graduate Business School, University College Dublin

Ioannis Oikonomou ICMA Centre, Henley Business School, University of Reading

Zacharias Sautner Frankfurt School of Finance & Management

Laura T. Starks McCombs School of Business, University of Texas at Austin

Xiao Y. Zhou Smith School of Environment, University of Oxford

1

ESG Engagement by Institutional Investors

• Institutional investors increasingly engage to improve firms’ ESG profiles, often through private engagements (McCahery, Sautner & Starks, 2016)

• A goal is often the reduction of downside risks • Negative ESG exposure can imply substantial legal, reputational, operational, and

financial risks

• BP’s Deepwater Horizon oil spill showed the importance of robust E policies (Dyck, Lins, Roth & Wagner, 2018)

• A number of large investors engage firms on E&S as well as G (Dimson, Karakas & Li, 2016) • Firms with better E&S performed better during the crisis (Lins, Servaes & Tamayo, 2017)

• Investors also collaborate on their ESG engagements (Dimson, Karakas & Li, 2018)

2

“Vanguard investors collectively own about 5% of every publicly traded company in the United States and about 1% of nearly every public company outside of the U.S… At Vanguard, we’ve been on a journey toward increased engagement over the past decade or so. Our peers in the mutual fund industry have as well.”

Getting to know you: Sharing practical governance viewpoints By F. William McNabb III Vanguard Chairman and CEO

Speech at University of Delaware. John Weinberg Center for Corporate Governance.

3

• ESG engagement data from a specialised shareholder engagement agent

• Represented institutional investors with more than $200bn assets under advise during our sample period (now > $1 trillion)

• Investor provided full access to its engagement database, including action reports, engagement activities, and measures of success

• 682 engagements targeting 296 firms from 2005-2014 [currently update in progress]

Engagement Data

4

• Top 3 engagement concerns: Board Structure, Remuneration & Climate Change

• ESG engagement themes (with example issue)

• Environmental: Climate Change, Carbon Intensity

• Governance: Board Structure, Remuneration

• Social and Ethical: Health and Safety, Human Rights

• Strategy and Risk: Capital Structure, Risk Management

Engagement Themes

5

0

20

40

60

80

100

120

140

160

Geographic Distribution of Engagements

6

Engagement process

Milestone 1 (completed by Agent): Concern raised with target company management

Milestone 2 (completed by Target Company): Issue acknowledged by target company management

Milestone 3 (completed by Target Company): Action/strategy taken by management to solve the issue

Milestone 4 (completed by Target Company): Action/strategy successfully completed

Engagement Actions

• Actions with the intention to enhance ESG characteristics • 1778 meetings

• 606 conference calls

• 204 emails

• 203 letters

• Contacts within target companies • 1004 contacts with senior executives

• 805 contacts with members of the boards of directors

• 471 contacts with the chairman of the board

[Some engagements were in progress at the end of our sample period. We are currently updating the dataset until end of 2017. This will provide us with more statistical power to dive deeper into sub-themes, actions and contacts.]

8

Measures of Downside Risk

• Lower partial moments (below 0%) • Second order (square root of semi-variance below 0%) - Main Risk Measure

• Markowitz (1959)

• Third order (cube root of the semi-variance below 0%) • Bawa (1975), Fishburn (1977)

• Value at risk (at 5% percentile) • Worst historical loss over the post-engagement period

• Duffie and Pan (1997), Jorion (2002)

9

Two Empirical Approaches

I. Endogenous treatment-effects models (with annual data)

II. Stock return analysis (with weekly data)

In both approaches: Matched Sample • Matching based on country, industry and size within FTSE All-World

index • Matching Approach:

• 1:N in Empirical Approach I • 1:1 in Empirical Approach II

10

Endogenous Treatment-Effects Models

• Accounts for selection bias in the engagement decisions, to extract causal effects from observational data (Wooldridge, 2010)

• Outcome and selection equation, for targets and control firms: 𝐷𝑜𝑤𝑛𝑠𝑖𝑑𝑒 𝑅𝑖𝑠𝑘𝑖,𝑡+2 = 𝛼1 + 𝛿 𝐸𝑛𝑔𝑎𝑔𝑒𝑚𝑒𝑛𝑡 𝑇𝑎𝑟𝑔𝑒𝑡𝑖,𝑡 + 𝑥𝑖,𝑡𝛽 + 휀𝑖,𝑡

𝐸𝑛𝑔𝑎𝑔𝑒𝑚𝑒𝑛𝑡 𝑇𝑎𝑟𝑔𝑒𝑡𝑖,𝑡 = 𝛼2 + 𝑧𝑖,𝑡−1𝛾 + 𝑢𝑖,𝑡,

• Downside risk measured over 24 months period post engagement

11

Effect of ESG Engagement on Downside Risk

12

13

Effect of ESG Engagement on Downside Risk

Average treatment effect (ATE) of engagement on LPM is -6.9% in Column (2) This equals 31% of the variable’s mean

Role of Engagement Success

Engagement Success: Low: The target company does not acknowledge the concern of the leading shareholder activist. Views do NOT align and remain opposed. Milestone 2 has not been achieved High: The target company acknowledges the concern of the leading shareholder activist. Their views commence to align. Milestone 2 has been achieved

14

Engagement Success: Low: The target company does not acknowledge the concern of the leading shareholder activist. Views do NOT align and remain opposed. Milestone 2 has not been achieved High: The target company acknowledges the concern of the leading shareholder activist. Their views commence to align. Milestone 2 has been achieved

15

Role of Engagement Success

Role of Engagement Topics

• Risk-reduction effects vary across engagement themes

• Stronger effects when governance or strategy topics are addressed

• Effects also stronger for social engagements, but only when such engagements are combined with governance engagements

• Supports the view that changing a firm’s social agenda without addressing governance is unlikely to yield a risk-reduction effect

16

Stock Return Analysis

• Measure the change in targets’ return sensitivity to a Downside Risk Factor (DOWN; Highest 30% minus Lowest 30%) • Similar to Ang et al.’s (2009) estimations of sensitivity to idiosyncratic volatility

• Weekly returns, two-year period around initial engagement • Measure a Post vs. Pre effects of engagement

• Two dependent variables

• Weekly target excess returns [Time-difference analysis] • Weekly target excess returns minus matched peer excess returns [Differences-in-differences analysis]

17

Time-Difference Analysis

18

Time-Difference Analysis

19

Difference-in-Differences Analysis

20

Conclusions

• ESG engagements can create value through a reduction in a firm’s downside risk • Risk reduction effects are stronger for more successful engagements

• Effects also stronger when governance and strategy topics are addressed, and if social and governance engagement is combined

• Evidence from two complementary approaches

• Our analysis contributes new insights into understanding the channel through which ESG engagement can create value for investors

21

To Do List • Update sample into 2018

• More on effects of E&S, by industry, country, etc. • Expand the first evidence we have that G matters for S

• Introduce entropy balancing

• DiD Analysis for treatment-effects model, preferably weekly

• Analysis of voting data

• ….

22

Thank you!

Looking forward to your feedback!

23