Staff turnover, regulation among five factors holding banks back

Over the past decade the software industry has been revolutionised by a set of new technologies. They keep close company – where you find one you’ll often find several – and the firms that have played the largest role in their creation collectively define the new economy of the 21st century.

These millennial technologies range from usable voice recognition and specialised storage for vast structured and unstructured datasets, to cloud computing, natural language processing, artificial intelligence and machine learning, and the rise of ubiquitous computing – meaning apps, mobile and the Internet of Things.

And yet, walking the IT floors of the large investment banks, you could easily miss the phenomenon completely. On its face, this seems bizarre.

Prior to the era of Google and Facebook, wholesale banking stood ahead of every other industry sector in its reliance on proprietary software engineering and invested the highest proportion of its revenues on IT. But while there have been numerous fintech successes in retail banking – in areas such as payments and money transfers, mobile, credit and lending – equivalent successes in the more arcane world of wholesale banking are much rarer. Boston Consulting Group estimates that of around $96 billion in venture capital funding in fintech during this millennium, only about $4 billion has been directed towards capital markets.

The investments banks are, of course, aware of the potential revolutionary nature of the millennial technologies: they have variously appointed heads of innovation and left coast board advisors; set up fintech labs and funded incubators and accelerators; directed principal investment towards fintech firms; and spawned pilot projects to experiment with at least some of the technologies such as distributed ledgers. Nonetheless, they’re not actually using the technologies on a meaningful scale.

Here are five reasons for the lag.

1. The software and data estate

Given the vast scale of wholesale banks and their IT groups and the endlessly energetic pre-crisis efforts of the banks to bring new services to clients in territories across the globe, it is no surprise now to find a legacy of thousands of applications written in an unthinkable array of programming languages deployed on a huge range of platforms at each bank. The same picture of fragmentation by accretion is true of bank data. Indeed, one of our clients that tried to classify its software assets found it had more data schemas than staff at the bank and well over 10 million lines of code for which the programming language could not even be easily determined.

While banks have evolved to be able to service such complex infrastructure, overhauling and re-platforming it onto a newer generation of technology is another matter.

On a more optimistic note, some banks have taken an entrepreneurial approach to their software estate. Given that investment in strategic systems by the large banks significantly outstrips that of even the largest software companies, a few of them have taken the sensible step of trying to monetise it. The motivations are various: cementing client relationships; writing down assets on the balance sheet and transferring intellectual property to a software company; funding continued development in the face of intense budget pressure; reducing run-the-bank costs; establishing software and data standards.

If deals such as these can succeed – and, while hoped for, that remains to be seen – they will not only restore institutional pride but open a door to new multilateral models for investing in excellent software that make sense in increasingly standardised markets.

2. Business organisation versus IT organisation

Banks have historically enjoyed some signal successes in the pioneering use of technology to provide enhanced services to clients. Examples include single-dealer platforms such as Autobahn and Barx over which Deutsche Bank and Barclays – along with several peer platforms – offer a wide range of research, analysis and trade execution capabilities. We might also add Credit Suisse’s Advanced Execution Services (AES), which rode and accelerated the wave towards agency execution, and, looking further back, JP Morgan’s RiskMetrics, which became a standard-bearer for value-at-risk measurement and was successfully spun out into a standalone company.

A defining characteristic of each of these successes was close partnership between IT and the business responsible for the relevant service. The business heads had an exceptionally high degree of technical knowledge in comparison to their peers – for example, Tim Cartledge, who ran the Barx business before going on to run global spot foreign exchange trading for Barclays, was a computer science graduate. Conversely, the developers were often pulled out of the main IT areas to be seated within the business unit and sometimes given extra incentives from desk profits. In such teams, all worked together on designing the client service, framing the marketing proposition and optimising the implementation detail.

This is not standard operating procedure. For every Autobahn or AES there are literally thousands of other applications that do not receive public acclaim and are developed by programming teams who are organisationally distant from their user population. It is as rare for the users to be able to name more than a couple of members of the development team as it is for programmers to know the current P&L of the businesses they support.

This division between programmers and users has been exacerbated by the trend over the past several years to shift development to offshore ‘centres of excellence’. Most chief financial officers will know the relative staff costs of employees in high-cost versus low-cost locations; few will know their relative code quality, even though perfectly good tools and methodologies exist to quantify it.

Without a deep unity of vision between business and IT leadership it is virtually impossible to formulate and implement a strategy to leverage radical new technologies. Even where there is consensus on the value of these technologies, IT and business heads may have different ideas about to benefit. For some on the finance side of the large banks, making an equity investment in selected start-ups is the most natural and direct mode of participation in the fintech scene. This does not necessarily promote the use of fintech software within the investing bank; indeed, it risks the unintended consequence of distortion in the adoption of the software when the interests of the bank as an investor conflict with the interests of the bank as a client.

3. Regulation

The tsunami of regulations following the 2008 crash has affected IT just as much as every other department of the large investment banks. Firstly, there is the direct impact on the appetite for IT investment of enormous fines and compliance costs coincident with a halving of fixed income revenues. More even than this, the need to comply with regulations such as the Fundamental review of the trading book and Europe’s revised Markets in Financial Instruments Directive (Mifid II) creates massive new work programmes for IT. These regulations have changed the structure of several markets, imposed many new kinds of reporting requirements, mandated wide-scale technical surveillance of practitioners and raised explicit new demands on technology and data quality. In many cases, strict timeframes for compliance have been set down by the regulators in advance of a workable specification.

Following the crisis, many banks announced ambitious targets for IT cost cutting but this was rarely achieved. Instead, IT budgets tended to flatten out at the start of this decade and all available IT resource was channelled to regulatory initiatives, virtually eliminating the capacity to do anything other than keep the lights on and comply with new laws. Most commentators who estimate IT spending now report rises over the past couple of years that are set to continue; this spending remains dominated by conformance to regulatory initiatives and associated market changes.

In principle, some of the millennial technologies appear suitable for use in meeting some of the new regulatory requirements. However, the precarious nature of the banks’ software and data estate, the lack of time and money to incorporate research and development into the delivery pipeline, and the combination of tight deadlines and uncertain regulatory specifications all conspire against thoughtful IT strategising.

4. Partner engagement

Given that all banks need to transform the way they operate at the same time to comply with the same regulatory changes, an obvious idea presents itself: rather than each bank bearing the full cost of their necessary change programmes, couldn’t they mutualise the cost through collaboration initiatives and/or increased use of third-party software? While we believe this will ultimately come to pass, it has not happened yet on a significant scale.

To understand why, we need to comprehend how banks engage with “vendors”, how banks engage with each other and how banks, in IT terms, think of themselves.

Taking the last point first, we have seen that IT heads find themselves responsible for software and data estates of boggling complexity, with no time to research new technologies or collaboration plays that may improve active project outcomes and with forbidding compliance deadlines set down in law. It is, perhaps, a fact of psychology that the more deeply you feel yourself to be lost in a byzantine wormhole, the harder it is to recognise it as a shared plight that may be addressable by social strategies. Accordingly, senior technologists with accountability for meeting, say, a key Mifid II date are reluctant to let a day pass in which they do not take a step forward from their particular form of darkness into a state that may be adequately light. In this context, collaboration initiatives can feel like an irrelevance or even a distraction.

Historically, the largest banks – as opposed, say, to regional banks – have had a patchy experience with third-party software. In part, the difficulty stems from the same issue: the first priority of the bank is to husband its proprietary tech stack and the adoption of alien standards, interfaces or design patterns comes a poor second. Even when banks do license third-party software, their integration and deployment processes – the purpose of which is to keep the bank’s current Jenga tower intact – frequently impair its proper use.

This aside, it is notoriously tough for a software company to negotiate with an entity whose IT, business and procurement departments are often individually powerful, organisationally distant and misaligned. With few exceptions, software companies often simply give up and look for more attainable revenues from smaller banks and buy-side firms that are more proficient at licensing software.

It is worth noting that millennial technology is often available as open source software (OSS). This has not normally been true of software from the vendors that serve banks. The ensuing lack of transparency has tended to make implementations harder and engender mistrust. Unfortunately, banks do not generally play a full and healthy role in OSS communities. While all developers use OSS, policy and legal restrictions and security concerns at banks make it harder to download and evaluate, and also block normal code contribution practices. The consequence is that banks tend to have something along the lines of a “Don’t ask, don’t tell” policy to OSS in which they make hundreds of thousands of downloads but then need to maintain their own versions of it all rather than checking changes back into the public code line. As well as impairing the natural role of talented developers as good OSS citizens, this is costly and quality-reducing. If this can be fixed, it will enable banks to develop higher quality, more functional relationships with alternative software suppliers and potentially with each other.

The relationship of the banks to each other may ultimately be the most subtle and important factor that affects technology transformation. We still encounter senior staff inside and outside IT who hang onto the idea of technology differentiation and believe their anti-money laundering system or their special flavour of data lake will confer a competitive edge. Much more often these days we find staff who are almost overwhelmed by the forbidding difficulty of delivering core projects that are either critically underfunded or, more rarely, staffed to an unmanageable degree out of regulatory panic. Between the two there are occasional calm heads creatively seeking a path of rational cooperation with peer banks and other external parties. For these people, the twin challenge is to find fruitful collaboration paths outside their firms and to establish traction within. If initiatives with, say, blockchain or cybersecurity or fraud detection achieve significant success over the coming years it will be down to them.

5. Staff turnover

The single greatest obstacle to collaboration initiatives and the alignment of software and business goals is turnover among decision-makers at the banks. The visionaries who might develop, nurture and promote great new practices don’t stand much chance when they and their management are frequently reshuffled. Heads of innovation come and go even more often than chief information officers: in a spot check of nine heads of innovation for this article, only one had been in place in 2015.

Whether this indicates dissatisfaction of the staff themselves or of their employers or of both is hard to gauge; most probably it reflects a sense of mutual frustration arising from the chasm between the blue-sky tech strategists and the developers who run production systems.

For strategic progress to be sustained the pace of turnover of critical staff needs to slow down. Tenure in roles is required for the more thoughtful staff to find and connect with each other, win mindshare with colleagues and start to change things.


A version of this article was published by Risk in March 2017.

IT budgets tended to flatten out at the start of this decade and all available IT resource was channelled to regulatory initiatives, virtually eliminating the capacity to do anything other than keep the lights on.

We still encounter senior staff inside and outside IT who believe their anti-money laundering system or their special flavour of data lake will confer a competitive edge.