Sector Biases When Building an ESG Portfolio

We hope you enjoyed your Fourth of July with your family while we took the week off to do the same. Continuing our “Summer of Sustainability” series, we’ll begin discussion on how to create a true ESG (Environmental, Social, & Governance) portfolio. This week, we’ll dig into performance trends in our aggregate ESG portfolio to determine how sector biases can affect portfolio construction.

First, here’s a market and factor update ranging back to our most recent Factor Spotlight:

‍US Market (7/2/19 - 7/11/19)

• The major indices continued their push into record territory, with the Dow and S&P 500 reaching new highs on Friday. Most of this momentum can be attributed to remarks from Fed Chairman Powell that fortified investor expectations for a rate cut at the end of this month.
• Chinese exports and imports both fell in May, while China’s trade surplus with the US rose to $29.92 billion in June (from $26.9 billion in May).
• The EU reported a sizable uptick in industrial production in May: 0.9% higher than in April - suggesting a measure of stability in the eurozone’s long-suffering manufacturing sector.
• Next week, banks will kick off the Q2 earnings season (starting with Citi on Monday). Many investors are wary that general 2H guidance may be lower than expected due to slowing global growth and uncertainty around trade

Here’s how factors have moved over the past two weeks in both US and global models, using our normalized return indicator:

US Model

‍Global Model

• Market Sensitivity and Volatility were both the biggest winners in each model, with US Volatility crossing over into Overbought territory at +1.13 standard deviations above the mean. Global beta is on course to become Overbought in the near term as well at +0.91 SD above the mean.
• Profitability in the US appears to have reached a peak at +1.38 SD above the mean on 6/28 and has been flattening out since then.
• Growth has continued to fall in both models, with US Growth nearing Oversold space, and global Growth dropping 0.84 standard deviations in the past two weeks.
• The rapid decline in Momentum continued as it now sits around Neutral in both models while the trend suggests more room on the downside.
• US total risk (using the Russell 3000 as proxy) and global total risk (using the ACWI as proxy) both ticked down 23 bps in lock step.

‍ESG Performance Attribution Analysis

‍In our last Factor Spotlight, we introduced our new ESG factor framework, with factors provided by our partner OWL Analytics. Today, we’re going to decompose performance at the Sector (GICS I) and Industry Group (GICS II) level in two versions of our aggregate ESG portfolio - Base and Sector Neutral. Our ultimate goal is to understand the differences between these two approaches to see which one is better suited for ESG portfolio construction. We’ll focus on the last 5 years (2014-2019), since we know sustainable investing has been increasingly on the rise during that time.

As a reminder, our methodology was to:

• Take OWL Analytics’ 12 main KPIs (hierarchy below) and create equal weighted portfolios that are long the top 200 and short the bottom 200 companies in the Russell 1000, ranked by their exposure to those KPIs.

• Aggregate all 12 portfolios into one super-portfolio called “Composite ESG - Base.”
• Create a second sector neutral version of Composite ESG - “Composite ESG - Sector Neutral.”
• Decompose performance for both portfolios over the past 5 years at the Sector (GICS I) and Industry Group (GICS II) level and then rank by average exposure.

Performance: Composite ESG - Base (2014 - 2019)

Performance: Composite ESG - Sector Neutral (2014 - 2019)

Performance Decomposition by Sector (GICS I)

Here, we can see how much wider the range between sector exposures is in the Base portfolio (-36% to +11%) vs. the Sector Neutral version (-8% to 4%). While it’s expected that the Base portfolio would be taking larger sector bets, in quantifying it we’re able to see just how sizable the gap is. The Sector Neutral portfolio is much less biased, with the sectors themselves not contributing a dramatic amount of P&L in either direction.

Interestingly, while one might think that Energy would be the most reviled sector in an ESG KPI-driven portfolio, it’s actually the Financials group that earns that dubious distinction, and by a huge margin. Much of this can be explained by poor scores on the Governance KPIs among Financials, which we will examine when we analyze the Governance component later this summer.

Performance Decomposition by Industry Group (GICS II)

In similar fashion, we again see a much wider range of exposure when we break out the Industry Groups, with the Base portfolio ranging from -15% to 11%, and the Sector Neutral exposures ranging from -7% to +4%. Funnily enough, the group that ended up being most overweight in the Sector Neutral portfolio was Energy. But again the impact to P&L from any of these Industry Groups is pretty muted in the Sector Neutral version.

‍Conclusions

‍If we’re seeking to isolate the ESG factors to create a truly sustainable portfolio, it’s not helpful to make huge implied bets on the sectors themselves. Especially when these can be accomplished through the Sector SPDR ETF’s (ex: going long the XLK and shorting the XLF). If we’d like to overlay these types of sector bets on top of an ESG portfolio, we can, but our goal here is to emphasize sustainability by extracting the value of pure ESG. To this end, we’ve just determined that investing in a sector neutral ESG portfolio is a superior way to more closely isolate the effect of the ESG investing phenomenon while controlling for sector bias. In future analyses, we’ll be primarily focusing on the sector neutral versions of these portfolios.

Regards,
Omer