Integrating Alternative Data (Also Known as ESG Data) in Investment Decision Making

ABSTRACT

What is environmental, social, and governance (ESG) data and how do we evaluate its quality and effectiveness? This form of evaluation is important, as it is a precondition for investors trying to integrate ESG in investment decisions. Previous literature describes intrinsic properties of ESG data (e.g. multifaceted-ness and context dependence) and highlights a trade-off between the validity and reliability of ESG data, which is often tied to the lack of theoretical foundations and scarcity of high-quality ESG data. Encouragingly, new data technologies have improved the accessibility, availability, and transparency of ESG data, but an agreed theoretical framework to evaluate ESG data quality is still lacking. This paper seeks to fill that theoretical gap by proposing a ‘user-oriented’ approach to evaluate ESG data. In this framework, we consider ESG data to be a ‘continuous concept with limitless boundaries’ and characterise it in terms of its width and depth. The bearing of width and depth on ESG data quality is ultimately a function of the investment decisions in which such data is used: the approach we endorse is therefore user-centric. This study then shows how ESG data, when it is of high quality, maps onto the investment decision-making processes.

KEYWORDS:

• Environmental social and governance (ESG) integration
• environmental sustainability
• data innovation
• alternative data
• ESG data
• data mobilisation and integration
• investment decision making
• sustainable finance

JEL CLASSIFICATIONS:

• Q51
• Q55
• Q56

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Soh Young In http://orcid.org/0000-0002-5824-9816

Notes

1 Eccles, Kastrapeli, and Potter (2017) found, by conducting a global survey with 582 institutional investors, that sixty percent of respondents identified the lack of common standards for measuring ESG performance as their dominant concern.

2 Wood (2010, 1995) used the term corporate social performance (CSP)–she also listed its sister concepts such as corporate social responsibility, corporate social responsiveness, corporate citizenship–as a broader term that contains a firm’s social and environmental impact.

3 Some studies explain difference between single- and multi-dimensional approach from stakeholder perspectives (see Delmas & Blass [2010] for instance). The performance is defined as the extent to which companies achieve their principals’ target. If one believes there are multiple environmental stakeholders of a firm, the construct to defines its EP should be multiple as well.

4 Carbon intensity is defined by Trucost as the absolute GHG emissions divided by one million USD of revenue. It is thus estimated in metric units of (tCO2e/$million), allowing us to compare firm-level carbon efficiency in all firm sizes and across industries.

5 For instance, a dataset that is only 80% inaccurate may be unacceptable to many investors. Yet, to a fund running a high-frequency trading strategy, such a dataset may be deemed ‘high quality’ if it has sufficiently low latency (i.e., it can be accessed within a very short window after it is generated).

6 ‘Freshness’ subsumes a variety of temporal concepts. Perhaps the idea closest to freshness is latency: the lag between the occurrence of an event and the time-point of availability for data on that event. Data that is low-latency is thus inherently fresh. Somewhat confusingly, the notion of ‘high-frequency’ data is often treated synonymously with low-latency data in finance. The frequency or regularity with which specific data-points become available to investors is surely important in some cases. Yet it is unclear that the regularity with which data is generated should be a priori more important than the speed with which generated data becomes available for use, regardless of its periodicity. Freshness, therefore, seems a more universally applicable parent concept.