Automation in Banking, will there still be new hires for your job in 2025?

Tolga Çelebi, Jens de Jong, Benjamin Oosterom

November 2018

Rise of the machines

Automation is the technology by which a process or procedure is performed with minimum human assistance. In recent years, start-ups and the high-tech industry have become the main supplier of automation technologies. One after another is coming up with new systems that are based on artificial intelligence (AI). One of the most publicly known projects is the DARPA (Defence Advanced Research Projects Agency) Grand Challenge in 2005, in which self-driving cars had to navigate themselves through a 132-mile course with narrow tunnels and sharp turns in the dessert. Due to these and similar public projects, admiration for this new technology was followed up by fear for the technology. If a car can drive itself, we don’t need a driver. If a car can drive by itself, then it would be a matter of time before a system replaces my daily routine tasks, right? However, we - human people - are not good in imagining the impact of a new technology at the future. Who could have predicted, when the Internet went commercial in 1990, what impact it would have on our daily activities? Even the most important people in the computer business could not have imagined the impact of the PC. Thomas J. Watson, then president of IBM, is widely quoted saying in 1943: “I think there is a world market for maybe five computers” (Sejnowski, T. J., 2018). At this moment, we try to imagine what the impact of AI technology might be on automation in the future. Will it really replace your job? Or will it help you to get better results?

Lessons from the past

When we contemplate about the future of banking jobs, it is relevant to look back to previous innovations in banking. One of the first big automations was the introduction of the ATM, or automated teller machine as it was once introduced as. The first ATMs originate from the 1960s and a society without them seems unimaginable nowadays. Before ATMs, all cash deposits, withdrawals and transfers were handled manually by bank tellers. Today most people make use of an ATM or an online application for such activities. However, ATMs did not have the impact on jobs as one might expect. The opposite of what you are probably expecting is even true: the number of bank teller jobs in the US increased 450.000 in 1999 to over 500.000 in 2016 (Bureau of Labour Statistics, n.d. A). Even though a decrease of 8% is expected next decade, this is anything but the impact on the number of jobs one would expect.

Another famous example is the changing role of brokers in the financial system. Before the introduction and widespread implementation of electronic trading, one had to call a broker to make a trade. This broker would call this request onto the floor of the stock exchange, where it would be matched manually with another order. If the trade regarded an over-the-counter (OTC) stock, the broker would call around, again manually, market makers who quoted different prices for the stock. Nowadays, many of these activities are automated and electronic trading has allowed for retail investors to observe the same prices as the brokers they used to call. As the work required to be an intermediary in a trade is reduced, one would expect the number of stock brokers in financial markets to reduce too. Again, you’re wrong. There are still over 375.000 financial intermediaries employed within the US and this number is expected to increase by 6% in the next decade (Bureau of Labour Statistics, n.d. B). In both examples the results of automation did not lead to the loss of jobs that was expected. So, what happened?

The human-machine cooperation

As it turns out, automation is proven to be most efficient in combination with human interaction. Automated systems most often seem to be not replacing humans, but rather complement them. This has mainly three reasons; O\one of them is the fact that automated systems require maintenance. To perform this maintenance, the sustainer has to understand the system. Another reason is the fact that a company can change its course, to which the system has to addept. This is another form of maintenance - and again - the sustainer has to understand the system. The third and maybe the most important reason is the fact that human and systems get better results together rather than working individually (PwC, 2017).

A specific example of reinforcing collaboration between humans and machines is the research that has been done on the capability of skin cancer recognition. A machine used to recognize skin cancer on photographes has been used with- and without the help of a doctor. The results were interesting: the doctors had a success rate of 92%, the machine had a success rate of 95%. However, when the doctor used the machine to confirm his findings, the overall success rate was 98% (Sejnowski, T. J., 2018). Despite that this research is not focussed on jobs in banking, it brings a crucial message. Machines and human complete each other, instead of cancelling each other out of the process.

Let us take a look at the role of a trader. These days, algorithmic trading is responsible for the lion’s share of bids and offers in the market (Hendershott, T. et al., 2011). Trading algorithms have to be built by persons that actually understand the task of a trader and understanding of the market in general, combined with the knowledge of how to set up an algorithm. A trader could change his or her strategy at a certain point. This can be due to an event in the market or just because the trader thinks that the current state of the market needs a different approach. By changing the algorithm, the trader adapts his algorithm to a new situation. In this case, the algorithm has to be ‘helped’ by the trader to change its course. Meanwhile, the algorithm performs trades much faster than that the trader could have achieved manually.

The consultancy firm McKinsey & Company (Berruti, F. et al., 2017) estimates machines take over 25% of the banking tasks currently performed manually over the next decade. Within banking the biggest impact automation will have over the next decade will be the result of robotic process automation (RPA) and machine learning (ML). These technologies will have a major impact on the service operations, IT, finance, risk management and human resource departments. Work such as handling IT service tickets, account closures and performing valuation processes all have the potential of being automated. Within risk ML may play a big role in analysing and defending against cyber-attacks and improve fraud detection. For human resource management AI even has the ability to identify and screen promising job candidates, predict when employees are likely to leave and also to find and analyse potential drivers for making valuable employees stay. On the other hand, product development, marketing and sales activities are much harder to automate and are not expected to endure any significant changes in terms of automation in the near future. These aspects of could be described as the creative and human aspects of banking. For the departments subject to automation this will mean that many repetitive jobs will be redundant and that a shift in activities will take place. The most important activities will be managing the technology and development and interpreting the results of automated processes. This is the important interaction between automation and humans as described above. This is likely to make the average banking jobs more challenging, but also more enjoyable. According to research by consultancy firm Accenture (2016), 84% of managers from all levels within banking expect that the implementation of the earlier mentioned automation techniques will make them more effective and their work more interesting.

The activities performed by departments subject to automation will not vanish, but the jobs will change. The bank teller profession still exists on a large scale, but the profession became more complex. The job entails no longer solely performing basic tasks, bus focuses more on relationship-building efforts and complex non-routine activities (Bessen, J., 2015). Likewise, the broker profession changed but it has not vanished. The broker profession got more intertwined with IT, which requires more skilled workers. Moreover, fees from OTC trades have become more important due to automation, hence the focus on relationship building with large pension funds and mutual funds increased. Moreover, the automation process will also require many high skilled who have understanding of the development and maintenance of automated systems.

The successful banker of tomorrow

To conclude, the implementation of automation technologies carries both risk and opportunities within banks. Most of the risk actually falls not with the bank itself, but rather with its employees. Among the employees, low-skilled employees are most at risk. These employees are often engaged in administrative and repetitive tasks and have a high risk of being being replaced by automated systems. On the other hand, automation will free up a lot of time for both higher-skilled employees and employees who are engaged with direct client contact. These employees are involved with strategic work that requires creativity, contextual understanding, and managing technology and other employees. Other opportunities lay with new departments being developed and enhanced; especially the jobs that require coding knowledge within the IT departments are expected to increase. For banks it is important to realize the opportunities of automation and shift their workforces towards higher-skilled activities. The successful banks of tomorrow will be those that manage to embrace automation efficiently. Therefore, the successful banker of tomorrow will be the one that knows how to add value in an automated environment. Your job will probably not be gone in 2025, but your daily activities might be very different and hopefully more enjoyable.




Tolga Çelebi

MSc Econometrics & Operations Research at VU Amsterdam

Jens de Jong

MSc Finance at VU Amsterdam

Benjamin Oosterom

MSc Econometrics & Operations Research at VU Amsterdam