Are algorithms really the best way to assess diversity and inclusion in business? - Vicki Robinson
Data plays an essential role in assessing a company's ethics today. Investors are increasingly incorporating environmental, social and governance (ESG) factors into the way they analyse businesses, in order to identify growth opportunities. Diversity and inclusion indexes are gaining influence, and software is advancing rapidly. But is such an approach really the best way to look at such issues? And are the tools themselves politicised?
A major influence is Refinitiv’s Diversity and Inclusion Index. One of the largest providers of data and infrastructure for financial markets, it boasts $6.25 billion in revenue, over 40,000 customers and 400,000 end users across 190 countries. The index identifies the top 100 publicly traded businesses, and ranks over 12,000 companies globally.
Ethics and finance are explicitly linked, with Refinitiv stating that the list ‘is designed on the hypothesis that companies tracking, reporting and achieving on measures of diversity, inclusion and people development will offer better performance over time than those achieving lower scores, or not tracking these measures’.
Refinitiv assesses businesses using 24 metrics across four key pillars: diversity, inclusion, people development and controversies. The controversy pillar in particular requires closer analysis.
According to the methodology, businesses are scored according to the number of controversies published in the media relating to workforce diversity and opportunity, wages or wage disputes, and relations with employees. It is not clear if or how Refinitiv’s technology assesses the veracity of such reporting, particularly in relation to social media campaigns conducted by activist groups. Algorithms are not known for perceiving nuance.
Diversity assessment leans heavily on women, with the percentage of such employees taking up five of the eight diversity metrics. It is not clear whether this refers to sex or gender. Ethnicity is not addressed directly, though the percentage of board members having a cultural background different from the location of the corporate headquarters takes up one diversity metric. Disabilities and special needs take up one of five inclusion metrics.
It is relatively simple to calculate a business’s carbon footprint. Diversity is much more complex because it is tricky to define, and goes far beyond sex and ethnicity. A wide range of class and educational backgrounds, political beliefs and values benefit a business, for example. Social class should surely play a central role. And what about national concerns? OECD research shows that regional inequality is a priority in the UK compared to other advanced economies. There are questions to be raised about whether a one-size-fits-all model is appropriate.
The technology itself warrants attention. ESG software automates the collection and organisation of ESG data, enabling businesses to monitor their progress and compare their results with other companies. It is a relatively new, evolving category, and so there are different tools available, with varying emphases.
Such software is becoming essential. Reporting can be complex as ESG contains many data points, requiring information on areas including modern slavery, energy consumption and gender pay. The level of detail has increased dramatically since the EU announced numerous regulatory developments. Outsourcing to a provider, therefore, makes sense for many companies.
A leading provider is Workiva, a global business whose main selling point is bringing together corporate reporting, ESG, audit, and risk. It describes diversity, equity and belonging as its core values and, according to its 2020–2021 Environmental, Social and Governance Executive Global Impact Summary, aims to integrate DE&I into all aspects of working life at the company. Self-identification is also a priority: US staff identify their ethnicity at the time of hire; Workiva also collects ‘global binary gender demographics’ based on how people perceive themselves.
This focus on equity is an immediate concern. As The Equiano Project’s D&I guide explains, equity pertains to equality of outcome rather than equality of opportunity. It is an approach, not an undisputed fact. And how do Workiva’s values translate into their software? Are beliefs such as self-ID part of the programming? More clarity is needed.
Equity is gaining traction, to the point where large corporations are conducting racial equity audits. JPMorgan Chase conducted one recently as part of its $30 billion Racial Equity Commitment. Starbucks and Facebook have carried out civil rights audits/assessments with strong focuses on equity. Amazon agreed to one in 2022. According to Racial Equity Audits: A New ESG Initiative published on the Harvard Law School Forum on Corporate Governance, this has been driven by a push for racial equity from shareholders in recent years.
With these audits, it would be interesting to know which software is used and what assumptions are baked into it. But the wider issue is why equity is dominating. It fits neatly into a data-heavy method, which may explain its appeal to investors and the corporate world. But it does not take into account the wide range of factors behind variations in achievement. Family background and attitude to education are both central to social progress, for example. Improving people’s chances goes beyond creating quotas and ticking boxes.
And herein lies the rub. Improving social mobility is a political issue, and should be decided at the democratic level, not by corporations. Business initiatives to open up opportunities are laudable and essential, but they must be done openly and transparently, without imposing fashionable theories. Software should be as neutral as possible. And the driving force must be morality, not simply pleasing a set of algorithms.
Vicki Robinson
@storiesopinions