From Talking to Doing

A Framework for ESG Portfolio Construction

ESG is not merely a trend in the investment world. It is becoming core, making its way to the heart of the investment process globally. If you are in the investment business, ignore this force at your peril. Large inflows into ESG-focused equity funds, and even larger outflows from traditional equity funds indicate that a profound sea change is underway.

If you are in the investment business, ignore this force at your peril.

Senior figures in the industry and firms, large and small, have embraced this fact and are putting in place multi-year initiatives to embed ESG thinking into their governance, business plans and communications policy. All leading index providers offer ESG-branded index suites, and the number of ESG-themed funds and ETFs continues to rise. While the focus on this important topic is good news, a side-effect of it is a proliferation of ESG scores and measures produced by a variety of sources including index providers, asset managers and specialised firms. After having had to master the factor alphabet, investors and practitioners now have to deal with yet another set of metrics and considerations.

For those building investment strategies, a key question is how to integrate ESG considerations into investable portfolios using a process that is robust, repeatable, verifiable and adaptable to the inevitable changes that will happen to ESG scores and their methodologies going forward. In other words, how to create a robust and evolutionarily stable allocation process that naturally incorporates ESG measures, keeping in mind they will likely evolve over time.


For simplicity’s sake, we represent the portfolio construction and testing process as requiring three main stages. We begin with a starting investment universe; for example, this may be composed of developed markets large caps. We then refine this further using:

• A corresponding data source, including Market Data and Fundamental Data, and crucially, ESG Data including individual scores for all of the constituents in the universe going back in time;

• An allocation methodology/Objective Function (potentially with ESG considerations embedded) & an optional set of Constraints to further refine the allocation.

• An iterative process which typically takes place before a strategy is deployed based on the evidence gathered, as shown by the arrow between Performance Analysis and Universe Selection. Such loops are typical in evidence-based product design and quantitative investing.

When an optimal solution has been identified, the investment strategy enters the Deployment stage. Further refinement comes through a process of continual iteration.

The headings below show how to think of the different possible choices within this framework using a simple car engine analogy. In all cases, we begin by selecting our universe of assets; for example, the constituents of some parent index, e.g.the S&P 500 index. From here the approaches follow different paths.


1. Simply track, or optimize without taking ESG into account. This is not a long-term viable proposition in the ESG age.

2. Analysis focused on traditional risk and performance metrics (ex-post and ex-ante).


1. Filter the universe based upon ESG factors (e.g., select companies above certain scores).

2. Optimize the portfolio based on objective functions such as tracking error to parent index, minimal variance, or some other methodology.

3. Analyze (i) risk and performance metrics and (ii) portfolio-level ESG score and trace.


1. Select which ESG metrics to optimize for, e.g., composite ESG score, female board representation, or carbon emissions.

2. Optimize allocation for an objective function that includes the selected portfolio-level ESG scores in its formulation.
Examples include: Maximise portfolio-level ESG score subject to tracking error to the parent index, minimize the ratio of the portfolio variance to the portfolio-level ESG score, maximize the product of portfolio-level ESG score and the portfolio’s diversification ratio, minimize tracking error to a specific factor time series designed based on selected ESG scores, etc. 

3. Analyze (i) risk and performance metrics and (ii) portfolio-level ESG score and trace.

Case study

We consider a case study to illustrate how this framework can be applied in practice. Vastly more complex cases can be envisaged, but it is useful to start with a simple example.

This simple case study examines the results of the combustion, hybrid and electric approaches on a US mega-cap equity strategy using the S&P 100 index as a starting point. For each, the ESG score of the overall strategy is plotted over time, creating what we call an ESG trace.

ESG trace comparison graph (since 03/03/2014).

As we might expect, the electric approach has the highest overall ESG trace, followed by hybrid and then finally combustion. Historical data suggests that incorporating ESG considerations into the construction process improved risk-adjusted performance for the universe considered, all the more so when ESG considerations were also integrated into the optimization logic itself.

Performance comparison since inception (03/03/2014).

This is because within US large caps, there is a clear divergence in historical performance between high-ESG scoring and low-ESG scoring companies, especially since mid-2017.

ESG portfolio return dispersion: Comparison between returns since inception (03/03/2014) of portfolios formed from the top ESG performers from the S&P 100 Index (OEX) vs the bottom ESG performers.

The usual disclaimers regarding future versus past performance apply. There is an ongoing debate as to whether ESG is a reliable source of performance, with cynics pointing to limited empirical evidence, inconsistent data, large differences between different providers, ESG-induced factor biases, and a contradiction in terms between capitalism and ESG considerations, to name but the most common arguments. The duration and vigor of this debate will depend on evidence, with 2019 lending credence to the ESG side.

If capital allocators increasingly include ESG considerations in their decision-making process in the coming years, as we hope, companies with poor ESG scores will likely be at a competitive disadvantage to access capital, which in turn may hurt their financial performance. Furthermore, we believe that consumers will increasingly seek to be associated with companies that are mindful of their environmental impact, act responsibly towards their broader stakeholder communities, and have in place ethical governance frameworks.

More information is available in our white paper. It contains a description of the methodology applied for this simple use case, and examines its impact on risks, returns, allocations and other metrics.

ESG trace comparison graph (since 03/03/2014).

ESG trace comparison graph (since 03/03/2014).

Performance comparison since inception (03/03/2014).

Performance comparison since inception (03/03/2014).

ESG portfolio return dispersion: Comparison between returns since inception (03/03/2014) of portfolios formed from the top ESG performers from the S&P 100 Index (OEX) vs the bottom ESG performers.

ESG portfolio return dispersion: Comparison between returns since inception (03/03/2014) of portfolios formed from the top ESG performers from the S&P 100 Index (OEX) vs the bottom ESG performers.

As the combustion engine is inevitably being replaced by electric and hybrid alternatives in transportation, so will traditional indices and portfolios across the investment value chain. ESG-based benchmarks will continue to increase in importance and may eventually replace traditional benchmarks, leading to a new normal. Investors will increasingly ask that ESG scores be known not only for individual companies, but also for portfolios and other investment solutions, both today and historically. Portfolio-level ESG scores and their historical traces will become standard metrics in factsheets, marketing documents and their digital equivalents. Much as in the world of credit ratings, it is likely that a small number of ESG scoring methodologies become leading industry standards, each with their own merit, strengths and weaknesses. Given the urgency of the ESG imperative, practitioners and investors cannot afford to wait. In the old world, a choice had to be made between doing well and doing good. Today, the two are aligning.

This new world requires new thinking and new tools:

New thinking

Those building and applying investment strategies need to re-think the portfolio construction process. Traditional risk and return metrics are no longer sufficient as ESG considerations must also be natively incorporated in the process. This, of course, requires that ESG metrics are available for the universe considered. We expect the availability, quality and granularity of metrics to improve over time as ESG requirements take hold in the coming years across the industry. The passage of time will naturally contribute to adding historical depth to them.

New Tools

This new paradigm also requires new tools. Using its modular platform, ALPIMA has developed a robust, consistent, repeatable framework to help professional investors build personalized strategies while integrating ESG metrics into the heart of the construction process.

We hope this adaptable framework, coupled with the flexibility of the ALPIMA platform will help practitioners and investors fully embrace ESG considerations and incorporate them into their investment decisions going forward.

For the full white paper, please contact us at