Zooming Out and Zooming In on Behaviour

Three weeks in January 2023 and three different stories illustrate the complex and often contradictory picture of how financial services firms are seeking to understand and address culture and behaviour.

One is the imminent demise of the UK’s Financial Services Culture Board (the FSCB), originally set up as the Banking Standards Board (BSB) in 2015. As a not-for-profit and voluntary membership body it appeared that an insufficient number of member banks felt they gained appropriate value out of it, or perhaps otherwise concluded they had the “culture” topic under control, or that it was yesterday’s news. I cannot comment on the perceived value-for-money point, although my experience in dealing with the FSCB was always positive. However, it is clear that the journey to better understand, measure and mitigate behavioural risks in banks remains far from mature, and the loss of an expert industry body is disappointing. As both the FSCB and the FCA noted: “the job of developing and maintaining good workplace cultures is not done and indeed will never be complete”.

Reinforcing why corporate culture remains critical is the example of Revolut, who are currently attempting to secure a UK banking licence. An article in the Guardian newspaper highlighted the creation of a new internal team, including psychology and behavioural science experts, as they seek to tackle criticism over an aggressive corporate culture. The company have denied that the formation of this group was directly part of the work to gain the banking licence, but rather a recognition of the need to “shift and change” following continued growth and staff feedback. However, as the FCA mulls a licence application they will doubtless be interested in past controversies over Revolut’s working environment, high staff turnover and what actions are being taken to mitigate these.

At the start of January, the FT ran a piece that detailed the increasing focus of investors on company culture, employee motivation and the “Human Capital Factor” (HCF) as a driver of performance. This is no ephemeral academic theory, with research by JP Morgan highlighting the long-term outperformance of companies with the highest HCF results, and the creation in 2022 of two exchange traded funds that track an index comprised of companies with strong HCF scores. If some in Financial Services doubt the importance of their own corporate culture in risk mitigation and value creation, then they should take time to speak to their own research departments and those tasked with understanding new sources of alpha in stock selection.

Synthesising these recent stories there are two trends for 2023 and beyond that will help inform the approach to cultural and behavioural insight in firms, and by inference will also be critical to their understanding of the drivers of both conduct and performance.

One requires zooming out and understanding the promises and threats of big data analytics; and the other requires zooming in and building internal behavioural science informed muscle. In this edition of the COO Magazine, we will focus on the former, and we will return to the latter in Q2.

Zooming Out – Big Data Analytics

The volume of data created, captured, copied, and consumed globally was forecast to be 97 zettabytes (ZB) in 2022 (a ZB being 1 with 21 zeros after it) with 90 % of the data in the world being generated over the last two years. We can safely say that these numbers are guestimates and that a lot of that data is duplicated rather than new. Nonetheless the growth of data is exponential and the analytics to sift and interpret the information are also developing at pace. Within this sea of material is a subset of data that is externally available, and from which interested parties are gaining new insight into the traits, culture and behaviours of companies.

Many people will already have used the Culture 500 website, developed through MIT Sloan in conjunction with Glassdoor, to peek into the perceptions of their own firms, or to help them make decisions on where to move next. But as the above January article on HCF highlights, investors are increasingly using similar information. The two newly issued ETFs were developed in partnership with Irrational Capital, an investment research firm employing advanced data science to investigate the relationship between corporate culture and stock price returns.

In a similar vein, Alliance Bernstein have also noted that “the onset of data science has added an important dimension to assessing companies’ human capital and corporate culture—with a data-driven lens. With data science, investors can track trends in companies’ culture and values ratings over time and compare it with peer firms”. One example is the Unobtrusive Corporate Culture Analysis Tool (UCCAT) which is a scientifically tested methodology, developed by academics at the London School of Economics, for analysing and benchmarking corporate culture. UCCAT analyses publicly available data indicative of a company’s cultural footprint.

Another illustration is provided by Sparkline Capital, who describe themselves as

“…an investment management firm applying state-of-the-art machine learning and computing to uncover alpha in large, unstructured data sets. We are passionate about helping investors navigate the wave of technological innovation transforming financial markets”.

When looking at key indicators of innovation, they noted that, as with culture more generally, they could not simply rely on firms to tell them what was happening. They therefore tracked other markers of investment in a skilled workforce to provide an unfiltered truth.

Inevitably it is not just investors and employees who are exploring the data to gain insight. Regulators, research houses, ratings agencies and activists are all leveraging the approach, and this is radically changing the dynamic on how much of their cultural foibles a firm can realistically keep hidden and away from the public glare.

In 2021 the Bank of England issued a work-paper on the use of multiple, unobtrusive sources of data to gain a deep insight into bank culture, and which also found robust evidence that poor culture leads to substantially higher risk, demonstrating the importance for prudential outcomes. With many other regulators embracing the digital and big data age, and with their access to additional information not in the public domain, their level of detailed insight into cultural and behavioural outliers and risks will rapidly grow.

A further driver will be the demand for accurate and unvarnished Environmental, Social and Governance information (ESG) information which has exploded in recent years, and with which the cultural and behavioural landscape of a company is inextricably intertwined. The data sources used to populate ESG ratings models include public, quasi-public, and private data. Traditionally the public data has included company reported filings, sustainability reports, press, newswires, and media reports. However, data and ratings providers have developed increasingly advanced approaches to mining information.

Examples, including IHS Markit’s Research Signals and Refinitiv’s MarketPsych Analytics sentiment scores, already collect data from thousands of news and social media sites to create actionable intelligence.

The interest in interrogating these data sets from other societal stakeholders and activists should also not be under-estimated, with latent anger from perceived cultural failings in the Financial Crises of 2008 and the myriad of other conduct failings; to a continued focus on whether a bank’s espoused values, in where and how it does business, align with the reality of their business practice. As the global response to the 2022 Russian invasion of Ukraine showed, the spotlight can quickly shine on the behaviour, actions or inactions of companies

Conclusion

Management need to ask themselves and their organizations the following three questions:

• When an investor, regulator, ratings agency or activist asks a detailed question on a behavioural or cultural concern they have identified, through their analysis of big data, how well prepared and informed will your company be?
• To what extent are you using similar techniques to understand what the available public data is revealing, and do you also use parallel big data and AI approaches and insights internally on your own in-house information, to better understand the cultural landscape and to identify outliers?
• If you do identify cultural issues and outliers what do you then do, how do you explore them in more detail and using what expertise?

Answering the third question will be the focus of the next article, highlighting what leading banks and other organizations are doing to dive deeper and to apply scientific rigour.

In a 2022 speech James Hennessy of the Federal Reserve Bank of New York highlighted the key emerging factors he saw as strengthening the study and analysis of culture. This included the increasing use of behavioural science; and the use of big data, in conjunction with artificial intelligence and network analysis. Nonetheless, he also advocated for the continued need for the “human touch” in cultural diagnosis.

Similarly, in his best-selling book “Everybody Lies” Seth Stephens Davidowitz said:

“I am predicting a revolution based on the revelations of big data. But this does not mean we can just throw data at any question. And big data does not eliminate the need for all the other ways humans have developed over the millennia to understand the world. They complement each other. “

David Grosse is an advisor to Galaxy Sciences and Armstrong Wolfe, and runs his own behavioural science consultancy focussing on the challenges of conduct, culture and behavioural risk in the Financial Services industry.

He has held senior roles in London, Hong Kong and Singapore, including the Asia Pacific COO at CLSA and Global Head of Operational Risk at Barclays Capital. In his most recent banking role at HSBC he formed a behavioural risk team in the first line of defence. He also works with regulators, industry bodies and academia in helping develop more advanced approaches to risk management, through the application of behavioural and data science.