Management guru Peter Drucker is most well-known for his maxim 'culture eats strategy for breakfast'. The gist being that strategy without a compatible culture is doomed. This is not management consultancy snake oil: the evidence backs him up. Research by psychologists and management scholars increasingly finds a positive relationship between organizational culture and organizational effectiveness.
Culture acts as a brake on firm-level choices that can’t be regulated in advance, and as the world becomes increasingly characterised by volatility, complexity and uncertainty, culture will become more important.
Consider the implications and ethics of AI-led decision-making, and the challenges faced by regulators who cannot write rules for every foreseeable situation where AI is used. Instead, we rely on culture to provide the social control for guiding staff on how to respond to change, threats, or the absence of detailed rules for every situation. Equally, high level regulatory principles, like ‘Treating Customers Fairly’ only work if there is a healthy culture.
So culture matters. Firms and regulators get this. But this is not enough. Firms, and potentially regulators too, also need to be able to understand, diagnose and change cultures. As Peter Drucker says in his other famous maxim: 'What gets measured, gets managed'.
If culture is so crucial, then it needs to be managed. If it needs to be managed, then it needs to be measured.
And there’s the rub.
Culture is famously nebulous. There are over 176 definitions of culture, from Schein’s set of values that guide and constrain people in organisations, to stories used to garner resources, to Trevino's system of informal and formal structures. Culture is both a cognitive construction and an ‘objective’ experience. The science is fragmented across different academic silos: business ethics, anthropology, organisational behaviour, management, psychology and sociology.
Perhaps unsurprisingly, regulators have steered away from measuring culture. Unlike capital ratios and compensation which are quantifiable and tangible variables, culture is variously seen as something fluffy, mystical and intangible.
An impossible task?
So, can you measure culture at all? The idea that culture can be measured is much debated. In the sceptics corner, Dr Allen Zimbler, Formerly Chief Integration Officer of Investec and psychologist nicely summarises the challenge here: 'Culture might be possible to understand, but is very difficult to measure – how does one measure something as intangible as a pattern of assumptions emerging within a group, and shifting, dynamically, as the group begins to change, within a field that is constantly changing?'
Others argue precise measurement is both achievable and essential, mainly through the use of sophisticated employee surveys. Take the aviation industry as an example. Here, a poor culture can lead to one of the worst outcomes of all: airline crashes. Reliable and valid questionnaire tools have been developed by companies themselves and/or in partnership with the regulator. Within the air traffic management industry, employees working at Air Navigation Service Providers (ANSPs) are routinely surveyed and then interviewed on their 'safety culture' or the values and practices relating to risk. These include error reporting, balancing safety and production, and communication about safety issues.
Through a partnership between organisations, regulators, and academics, these data are used to identify cultural strengths that can be shared with other organizations, such as effective methods of communication, or areas that require development, for example capturing and learning from safety incidents. A focus on safety culture is widely credited, along with improved technology and crew training, in reducing fatal accidents down to 0.5 fatalities per million flight departures in Europe alone. Could measuring and changing culture lead to similar improvements for the financial industry?
One of the challenges of culture measurement using surveys, particularly in relation to ethical practices or risk taking, is social desirability, subjectivity, sampling errors, and limited longitudinal analyses. That is, people can say what they think you want to hear, and you can’t see what is happening over time. Companies, or employees, who are engaging in problematic practices are unlikely to participate, and assessors can be influenced by factors totally independent of a firm’s culture – like the prestige or size of a company. Plus, where judgements on culture are made within and across assessors, there is a major challenge in standardising how observations are combined.
Arguably, measuring culture solely through staff surveys and interviews is somewhat dated, using methodologies developed in the 1980s. Can advances in theories from organisational psychology, computing power and tech enable us to measure culture unobtrusively, without ever entering a company?
The authors of this article think it can.
The tools for the job
Since May 2016, Dr Tom Reader and Dr Alex Gillespie from the London School of Economics (LSE) have led an academic research project to develop a new approach to measuring corporate culture. Funded by the AKO Foundation, a UK charity, this research has produced the Unobtrusive Corporate Culture Analysis Tool (UCCAT).
UCCAT is a theoretically based and scientifically tested methodology for analysing and benchmarking corporate culture. Uniquely, rather than gathering employee interviews and questionnaires to assess culture, UCCAT analyses publicly available data, such as annual reports, financial records, press releases and databases, that are indicative of a company’s cultural ‘footprint’. These data points are termed 'unobtrusive indicators of organisational culture' (UICs).
Each UIC represents an observable and measurable aspect of organisational activity indicative of culture (for example, research spending or customer engagement), and can be understood as a ‘dipstick’ into a company’s culture. UCCAT aggregates over one hundred UICs in ‘cultural dimensions’, thus providing a broad-based and naturalistic measure of corporate culture.
In organisational psychology, ‘dimensions’ are like personality traits, describing characteristics consistent between and within organisations, that can be scaled from weak to strong. Organisational culture research has identified ‘cultural dimensions’ associated with institutional success and avoiding failure.
For example, the dimension ‘employee focus’ has been unobtrusively measured using reviews left by current and former employees on Glassdoor as well as the way in which management talk about employees in the annual report and earnings calls. Other UICs examine discrepancies between what the employees say compared to management.
Another aspect of corporate culture suggested by the literature to be a feature of healthy companies is transparency. Here, a UIC was developed from the adequacy of responses given to analysts’ questions during an earnings call on the organisation’s performance, combined with a Flesch reading score for investor-facing communications. Analyses of data at the LSE, on companies within the MSCI Europe Index, shows the cultural profile of companies varies to some extent according to industry sector, albeit not significantly. These unobtrusive measures of culture are associated with company performance: as measured by Return on Capital Employed (ROCE).
It remains an open question as to what core dimensions are particularly relevant for a healthy culture in financial services (e.g., integrity, transparency, customer focus), but clearly there is scope for these to be identified and made the focus of measuring a finance firm’s cultural health.
Whilst unobtrusive measurements can complement qualitative methods and surveys, they can also, in tandem, be used to provide new insights. For example, divergences between employee views of company rules provided in a survey combined with unobtrusive indicators of what is happening in practice, might reveal a disjunction between attitudes and behaviour not recognised within a firm.
Using unobtrusive data to study culture is inexpensive, and because data can be gathered across the entire a sector, a form of ‘culture observatory’ can be built for mapping practices across the sector, for example on risk. It may be able to predict key outcomes, such as compliance or risk assurance. And it could be used to track culture over time, for example before and after a crisis. Moreover, if new UICs are developed, they can be applied retrospectively to all past data – something that is impossible with surveys.
This potentially leads to significant benefits for regulators and governments across all consumer and public markets. One such use could be measuring the distribution of firms’ cultural health within and across markets, and being able to better understand the relationship between culture, employee behaviours, and organizational outcomes – both negative and positive.
This could be followed up with more ‘up close’ methods like surveys, ethnography and other qualitative methods to learn from the best and prioritise resources on the poorest.
So, to the question 'Can culture be measured?' We think the answer is very much, 'Yes it can.'
However, there is still a lot of work to be done to determine which aspects of culture relate to performance in which sectors.
The next and perhaps equally challenging question is whether firms themselves, auditors or investment analysts should take on this task of their own volition, or whether legislators and regulators need to take the lead?
By Alex Chesterfield (FCA), Dr Tom Reader (London School of Economics) and Dr Alex Gillespie (London School of Economics)