Selecting stocks through a modular investment process

Between ESG and company/investment performance we see not a causality but a relatively stable correlation. Our rules-based and quantitative strategy includes fundamental analysis, portfolio optimisation and risk management, as well as our sustainability process. The strategy has daily liquidity and does not use leverage, shorting, derivatives or stock lending, and can be offered as segregated accounts. The three-stage investment process consists of:

• a sustainability process;
• fundamental security analysis;
• quantitative risk and investment technology.

Sustainability process

Applied: Quarterly

Our sustainability process prepares the investment universe. It has four steps:

• forensic accounting (which identifies companies with aggressive accounting and governance practices);
• UN Global Compact compliance;
• proprietary ESG scores;
• preference-based screens such as industry or company exclusions.

We apply the process to the 77,000 stocks in our database, reducing the weights of the stocks that do not pass to zero, resulting in an eligible investment universe of about 1,200 stocks. When calculating our proprietary ESG scores, we analyse more than 200 ESG criteria from four data providers. This ESG information is then mapped into twelve sub-categories: five for environmental issues, four for social and three for governance. To calculate best-in-class scores for companies, we identify the most material issues per sector and apply a proprietary weighting to each. The resulting ESG score is a combination of qualitative research and quantitative and statistical analysis, to avoid spurious correlations and to aid implementation and constant quality checks.

Fundamental security analysis

Applied: quarterly

For each of the roughly 1,200 stocks in the eligible investment universe, we calculate three scores based on financial data. Standard concepts such as g-score and f-score (identifying companies with good growth prospects and strong balance sheets) are applied together with our proprietary earnings pressure measure, based on sell-side broker research. We then apply to companies’ earnings estimates a statistical model driven by behavioural economics (again based on sell-side broker research).

Quantitative risk and investment technology

Applied: daily

The quantitative investment model selects up to 100 stocks from the approximately 1,200 in the eligible investment universe. It consists of:

• an asset allocation model, which is driven by the aggregate momentum of each stock;
• a CVaR-Portfolio optimisation that integrates into its target function g-score, f-score, earnings pressure and the individual prices of all stocks in the investment universe. Applying the analysis daily enables us to protect the assets of the investor in times of volatility or larger stock market drawdowns.

Impact on stock selection

Our process is a much more thorough and nuanced approach than simply selecting investments based on individual environmental, social or governance factors. For example, US solar power developer SunEdison could be deemed an attractive ESG proposition based solely on its environmental credentials, but the stock price has collapsed (as of early April 2016, the company is preparing to file for bankruptcy) after word came out that both the Securities and Exchange Commission and the Department of Justice were to investigate the company primarily over alleged overstating of cash positions.

When passing through our Sustainability Process, SunEdison had always scored poorly on accounting and governance risk (AGR), indicating an increased likelihood of running into litigation because of aggressive accounting practices. Consequently, the stock never passed the sustainability process and its weight was reduced to zero, meaning despite its environmental credentials it was not in our portfolios when the stock price collapsed.