This article was written by Nick Wainwright, Head of Data Science, FSCB
At the FSCB we define a good culture as one that produces positive outcomes for customers, clients, and society as a whole. Recognising that organisations differ in so many ways and that there is no ‘one size fits all’ optimum culture, what we do is measure how far organisations can demonstrate that they possess nine characteristics that our extensive research has indicated would underpin a good culture: honesty, respect, openness, accountability, competence, reliability, resilience, responsiveness, and shared purpose. We assess culture in this way because we believe that it is a way of predicting outcomes (i.e., a firm that scores more highly in our Survey is more likely to generate tangible benefits to its stakeholders than one that has lower scores). This, however, is a very difficult relationship to robustly demonstrate as it requires a lot of observations, both in terms of organisations taking part in the Survey over many years and subsequent outcomes data. Now with our dataset having grown – the 2022 FSCB Survey is our seventh annual exercise assessing culture at financial institutions in the UK – we can begin to explore the link between results in our Survey and outcomes that truly matter to firms and society at large.
Measuring ‘good’ outcomes
Thinking back to when we at the FSCB (then known as the Banking Standards Board) first began to systematically assess culture, we identified different types of outcomes as being desirable for different stakeholders. For customers or clients, a key question is whether the products they are offered are appropriate to their needs and circumstances, and clearly explained. For society – and especially one in which the aftermath of the 2008 financial crisis continues to reverberate – a positive outcome might be described as a sustainable and trustworthy financial services sector in which systemic risk is appropriately recognised and managed.
No single metric can perfectly capture the full range of outcomes, particularly not ones that are available for a wide range of organisation (e.g., universal banks, large and small building societies, investment banks, private banks, and challengers). In this initial exercise we have taken two approaches.
- For customers, we have focused on conduct risk using publicly reported complaints data from banking and credit card accounts collected by the FCA. The hypothesis here is that a firm which is meeting customers’ individual needs will, all other things being equal, attract fewer complaints than one that is not.
- For society, we have focused on prudential risk by looking at Z-scores that can be extracted or derived from data reported by large firms. The greater the Z-score, the further the firm is from default, and therefore, it is hypothesised, the lower the level of prudential risk.
Data matching exercises are often messy and each of these outcome measures have their limitations.
- Both metrics are, for example, more readily available in public datasets for larger organisations and often unavailable for smaller or medium-sized firms.
- For complaints data there is an inherent information asymmetry which means that it may not be apparent to customers if a product is not in their best interests, or it can take quite some time for such doubts to materialise.
- In the case of Z-scores, some of these were available based on firm’s international entities, but the FSCB dataset on experiences or perceptions of employees that this is matched with is based on employees in the UK.
- There is also the question of time lag: a particular action that a firm takes now may lead to wider impacts at a future point in time, whether those be positive or negative, but the time it takes for these effects to be realised is difficult to predict.
Initial findings, partnerships and next steps
The initial analysis conducted was based predominantly on linking the FSCB Survey scores with publicly available Z-scores and complaints data as outlined above. This data linking was possible for all large, systemically important institutions in the FSCB dataset, but not for several of the small or medium sized organisations where outcomes data was unavailable. Correlation analysis and regressions, which allowed for the size of organisations to be controlled for, were carried out.
The results from this work indicate that the FSCB question set as a whole correlates positively with lower levels of prudential risk and better customer outcomes. In particular:
- employee perceptions on risk orientation (e.g., respect for Risk and Compliance, following the spirit of rules and not just the words) and customer centricity (e.g. responding effectively to customer feedback and innovating in the best interests of customers) link with lower levels of prudential risk; and
- Environments where leaders are seen as authentic (e.g. senior leaders meaning what they say), where employees believe that the organisation is customer-centric, and where there are higher levels of psychological safety (e.g. not worrying about negative consequences when raising concern) are associated with lower levels of customer complaints.
While these results are an encouraging first step in demonstrating a link between culture assessment and outcomes, they should be considered in the context of the caveats outlined. We will continue to build on this work and look to complete this analysis in a way that encompasses a wider range of firms and potentially other outcome measures. In this context, we have also begun a partnership with researchers at the Bank of England to measure cultural cohesion or fragmentation in firms to establish how this is associated with prudential and business outcomes. This work will draw not only on the FSCB dataset that captures employees’ perceptions of their organisation’s culture but also on other external data sources. We look forward to sharing results of our research in this area as it becomes available. if you’d like to know more or have ideas on this subject, please get in touch.
Joining the FSCB (or working with us as a Working Culture client, if outside the financial services sector) can help you assess, understand and improve your firm’s culture.
 Z-scores are a measure of the number of standard deviations on a firm’s return on assets is required to offset the amount of regulatory capital it holds. They are widely used as a proxy for the probability of default.