The ESG investing background of water activist investing

• Water consumption disclosure is part of ESG metrics (in the Environmental pillar) Markowitz (1952).

• Water consumption & disposal risk (World Economic Forum, 2022) can be attended through water activist investing.

• Environmental Social and Governance investing (ESG investing) and water management disclosure are related issues for companies and investors. The former has seen ESG practices (including water management ones) as a way of engaging in a better relationship between reputation and profits (Dawkins y Fraas, 2010; Fombrun, 1990; Hasseldine et al., 2005; Lee et al., 2016; Tetlock, 2011). Also, some firms engage in fundamental ESG practices to better impact their environment and stakeholders. This practice significantly improved their total or market risk profile and debt or equity rising costs (Delgado-García et al., 2013; Jo y Na, 2012; Kölbel et al., 2017; Oikonomou et al., 2012; Orlitzky y Benjamin, 2016; Salama et al., 2011).

• ESG investing manages a portfolio with stock selected in a ESG screening process.

  • Positive screening: includes stocks of companies with ESG standards, given an ESG threeshold.
  • Negative screening: excludes stocks with low ESG standards (also given a threeshold).

ESG investing as investment style

ESG investing has evolved from its origins in religious practices to a modern portfolio management or investment style of its own.

Berry and Junkus (2013) and Chatzitheodorou et al. (2019) suggested ESG investment refers to a type of investment style in which the investor prefers to invest in socially responsible companies in her or his managed portfolio.

The companies of interest must fulfill ESG criteria that the investor considers appropriate for her or his long-term financial performance and sustainability level.

The Friedman vs. Freeman debate

The debate that an ESG company sacrifices its company performance (profitability) due to costs incurred with ESG practices.

That is, sometimes a company is profitable if it has a negative social or environmental impact. To comply with sustainability rules, it must invest in more expensive production processes or sacrifice production to be ESG compliant.

• Friedman (2007): “financial performance is the only sustainability task a company must comply with”.
• Freeman (1984; 1994): “The benefits of ESG practices are in the long-term because the company improves its customer or employee relations, reduces legal, reputational, or social risks, and, as a result, increases its productivity”.

Water activist investing definition

Similar to ESG investing’s historical origins in which investors made a sort of “activism” to push companies to a better behavior in Environmental, Social or Governance, water activist investing could be defined as

“The series of actions in which investors push companies to be more responsible in their water management practices.”

For the optics of this research, water activis investing means to perform a positive screening of stocks with good water consumption standards.

Water activist investing measurement

Water consumption items are part of the Environmental ESG pillar and is part of the ones of interest of several ESG grading companies such as MSCI-Eiris, Robbeco-Bloomberg, Morningstar sustainalytics, S&P ESG scores, or Refinitiv ESG scores

The score methodology of interest is the one of Refinitiv (2022). It grades more than 400 items and organize them in 10 factors and three pillars:

Refinitiv ESG score methodology. Source: Refinitiv (2019)

Among the 19 resource use items we found the water consumption (in cubic meters) per revenues item (\(WTR_{i,t}\)). Therefore, water activist investing will operate through \(WTR_{i,t}\) positive screening.

That is, the research simulated a theoretical agent that invested only in companies with high \(WTR_{i,t}\) from **January 2nd of 1998 to April 29th of 2022).

\[ WTR_{i,t}=\frac{M^3\text{ of water consumed}_{i,t}}{\text{Revenues}_{i,t}} \]

Theoretical conception of the working hypothesis

This theoretical figure shows the Classical Financial Economics position (the one tested herein in which the Research has an opposite position).

Working hypothesis

Given the portfolio \(P\) invested mainly in companies with high \(WTR_{i,t}\) values it is expected that the simulated portfolio \(P\), invested in stocks of the four main Latin American stock markets (Argentina, Brazil, Chile and Mexico), according to the WFE (2022), will be more or at least equally mean-variance efficient than these four stock markets portfolio (\(M\)):

\(H_0:\) “The simulated water activist portfolio is more or, at least, equally mean-variance efficient than the market one.”

Practical implications of the hypothesis test

1. If the results prove as true, then an individual or institutional investor could invest in water responsible companies.
2. Companies could feel compelled to enhance their water consumption profile, given the higher demand of water responsible stocks (the shunned-stock hypothesis of Derwall (2011, 2019)), their cost of capital could be lower.
3. Water activism could reduce the impact of water disposal.

Papers related to ESG investing

There are only works that tested the performance of ESG portfolios v.s. a non-ESG (even sinful or “bad practices” portfolio). Some of the most recent are focused in:

• ESG-ROE relationship:
  • U.S. stocks (Blasi et al., 2018; Brogi y Lagasio, 2019): Positive relationship as conclusion.
  • Australia, France and Korea (Crifo et al., 2016; Galbreath, 2013; Lee et al., 2016; Sethi et al., 2017): ESG harm profits.
  • Mexico (Alonso-Almeida et al., 2009; García-Santos y Zavaleta-Vázquez, 2019; Godínez-Reyes et al., 2021): ESG practices hav eno financial cost and a positive relationship with ROE.
• U.S. companies’ ESG score vs. market risk (Botha, 2015; Jo y Na, 2012; Salama et al., 2011): The higher the ESG practices the lower the market risk in a given company.
• Pension funds (Amalric, 2006; De la Torre-Torres et al., 2018; Hongbo et al., 2006; Sethi, 2005): ESG screened stock portfolios enhance the performance of pension funds.

Papers related to water intvesting

Only the work of Zeng et al. (2020) relates closely with the present research. These authors found no negative ESG-market risk in chinese stocks.

No previous works tested the performance of a water screened portfolio vs. a market portfolio. This is a gap the research intends to fill for the Latin American case in two perspectives:

• A USD based investor.
• A local currency based portfolio to control the impact of FX rates movements.

Investment universe (the set of stocks used)

To perform the simulations the research used historical weekly data (price levels) of the stock members of the Refinitiv Latin America price return index that is a set of the Refinitiv stock indexes mentioned in Table 1. The full investment universe (set of stock used) is appendix A.

The stocks in this set, are the investment universe \(\mathbf{\Phi}\).

Data gathering and processing for the market portfolio \(M\)

With the investment universe of Table 1 and Appendix A the research used the historical price of each stock each day \(P_t\) to estimate the continous time return \(r_t\):

\[r_{i,t}=\Delta\%P_t=ln(P_{i,t})-ln(P_{i,t-1})\] Also, the historical free-float market capitalization or \(MC_{i,t}\) (in USD and local currency) value of each stock was downloaded and used to estimate the investment level \(\omega_{i,M.t}\) of each stock at \(t\), given the cardinality of the investement universe \(N=\#(\Phi)\):

\[ \omega_{i,M,t}= \frac{MC_{i,t}}{\sum_{i=1}^N MC_{i,t}} \]

With the investment level \(\omega_{i.M,t}\) a weighted average market portfolio return was estimated:

\[r_{M,t}=\sum_{i=1}^N \omega_{i,M,t}\cdot r_{i,t}\]

\(r_{M,t}\) was used to estimate a base 100 value (at January 1998), and as a benchmark to compare the water portfolio’s \(P\) performance.

Data gathering and processing for the water portfolio \(P\)

To simulate the water screened portfolio (water activist portfolio) the research used the investment levels in each stock in the market portfolio \(\omega_{i,M,t}\) and estimated an investment level cap value \(ic_{i,t}\). This cap or maximum investment level is estimated with the water-to-revenues \(WTR_{i,t}\) value in each stock as follows:

\[\begin{equation} ic_{i,t}= \begin{cases} 1-\frac{WTR_{i,t}}{\sum_{i=1}^N WTR_{i,t}} & \text{ if }WTR_{i,t}\neq0\\ 0 & \text{ if }WTR_{i,t}=0 \end{cases} \end{equation}\]

The rationale of the cap investment level is “The lower the water consumption level related to revenues, the higher the investment level in the water portfolio \(P\)”.

This led to the investment levels in the water portfolio \(P\):

\[\begin{equation} \omega_{i,P,t}=\omega_{i,M,t}\cdot ic_{i,t} \end{equation}\]

Similar to the market portfolio \(M\), the historical return of the water portfolio was estimated as a weighted return average of each stock in the portfolio:

\[r_{P,t}=\sum_{i=1}^N \omega_{i,P,t}\cdot r_{i,t}\]

Historical portfolio parameters in the simulations

Given an \(N\times N\) variance covariance matrix \(\mathbf{C}=[\sigma{i,j}]\) of the stocks in the portfolio, and a \(N\times 1\) expected return vector of each asset \(e=[\bar{r}_{i,t}]\), along with the \(N\times 1\) vector, either market portfolio investment levels \(\mathbf{w}_M=[\omega_{i,M,t}]\), or water portfolio \(\mathbf{w}_P=[\omega_{i,P,t}]\) ones, and the expected return vectors (\(\mathbf{e}_M\), \(\mathbf{e}_M\)), the portfolio expectd returns (\(E_M\),\(E_P\)) , and portfolio risk exposure (\(\sigma_M\), \(\sigma_;P\)) were estimated as follows:

\[E_M=\mathbf{w}_M'\mathbf{e}_M \text{, }E_P=\mathbf{w}_P'\mathbf{e}_P\] \[\sigma_M=\sqrt{\mathbf{w}_M'C\mathbf{w}_M} \text{, } \sigma_P=\sqrt{\mathbf{w}_P'C\mathbf{w}_P}\]

Historical ex-post and ex-ante Sharpe ratios in the simulations

With these portfolios parameters, the ex-ante and ex-post parameters were estimated as follows

\[Sharpe_{ex-ante,M}=\frac{E_M-rf}{\sigma_M}\text{, }Sharpe_{ex-ante,P}=\frac{E_P-rf}{\sigma_P}\]

\[Sharpe_{ex-post,M}=\frac{r_{M,t}-rf}{\sigma_M}\text{, }Sharpe_{ex-post,P}=\frac{r_{P,t}-rf}{\sigma_P}\]

Hipotesis test

We made a visual time series and box-plot comparison, along with a one-way ANOVA and Kruskal-Wallis test of the historicas portfolio returns and Sharpe ratios.

\(H_0\) is true if the p-value of the ANOVA and Kruskal-Wallis test are higher than 5% or the time series and box-plot suggest \(r_{P,t}>r_{M,t}\) and \(Sharpe_{ex-post,P}>Sharpe_{ex-post,M}\)

Historical time series (portfolios performance)

Portfolio returns Box-Plot

Ex-post Sharpe ratios Box-Plot

Portfolio returns ANOVA test

Portfolio Sharpe ratios

Historical time series (portfolios performance)

Portfolio returns Box-Plot

Ex-post Sharpe ratios Box-Plot

Portfolio returns ANOVA test

Portfolio Sharpe ratios

Highlights

1. The water-to-revenues (WTR) portfolio had a more stable (mean-variance efficient) performance than the market one.
2. In terms of accumulated return, the WTR paid a better return than the market one, but this outperformance holds only in the short term.
3. From the previous results, it is crucial to highlight that even if the performance of the WTR portfolio is not as spectacular as the market one in some periods, the WTR is a relatively stable portfolio. Given the CAPM, cointegration, and ANOVA tests, the WTR portfolio had a statistically equal performance to the market in the short term. In the long-term, the WTR portfolio had no significant alpha (Jensen’s alpha) generation (extra return) and a significant, low systematic risk (beta of 0.26), and no cointegration with the latter. This result aligns with the literature (Jo & Na, 2012; Salama et al., 2011; Zeng et al., 2020) that found evidence that disclosing water management practices reduces total and systematic (market) risk.
4. The results are similar to the USD scenario if the simulations control the impact of currency price movements. The only difference is that both portfolios paid a positive accumulated return in the simulation period
5. If a given investor wants to control currency impact, she or he could safely manage a Latin America WTR portfolio by hedging with currency futures or options (a practice suggested for further research).

Main conclusion

The simulation results suggest that, despite the performance of the USD WTR portfolio being worst than the market in some periods, its mean-variance efficiency is statistically equal and has a low systematic risk (beta of 0.262). This result aligns with previous works that test water management disclosure and risk relationship (Jo & Na, 2012; Salama et al., 2011; Zeng et al., 2020).

Guidelines for further research

1. Due to space restrictions, the authors didn’t show the results of an Argentinian, Brazilian, Mexican or Chilean investor perspective.
2. The best approach in this paper are the results in local currency. Also, the results didn’t test the performance with hedging strategies.
3. Extending the current simulations to other countries and regions, such as North America, Europe, Asia-Pacific, the Middle East, or Africa, could be a natural and noteworthy extension.
4. Other water management items, such as the percentage of recycled water used by a company or the quality of the water management policy, could be another exciting perspective.