I started a career in asset management in 1985 as a new graduate consultant with no experience or knowledge of the financial services industry. As the business was all new to me, I asked lots of questions, listened carefully to the answers, and started to acquire the knowledge I needed to be able to test system components against very detailed specifications. I then moved on to have the ability to independently write those specifications for various business functions.
The level of automation in asset management at that time was low. Since then, various processes have become automated and those automated areas have then joined up to decrease the need for human intervention, increase processing capacity, decrease errors, and reduce costs.
The availability of “off the shelf” solutions and outsourcing of functions means that asset management organisations can now implement automation without needing in-house knowledge to fully verify that the underlying processing of data is 100% accurate—as that has already been done by the vendor/service provider and existing users of the platform.
In recent assignments, I have seen that the number of operational and IT staff with the knowledge and ability to manually calculate or verify investment accounting data has significantly decreased. To give a few examples:
A team responsible for booking derivatives trades and reconciling to the clearing broker was unaware that their open positions and unrealised P/L were incorrect. Once it was pointed out, they were unable to determine why by looking at the trades in a single portfolio and future contract.
Fund managers were accounting for client-initiated transactions incorrectly which risked breaching investment restrictions.
Custodian reconciliation teams using the wrong transaction types, as they did not understand the business. This adversely affected client reporting and performance calculations.
The common issue in these instances was that these teams have always trusted the system to get it right. They simply kept on with the existing process, without knowing how to verify that it was correct. This has become an unfortunate side-effect of increased automation and lowered dependency on people—we reduce the required skillsets and knowledge needed to do the job on a day-to-day basis.
This may be fine while everything is running smoothly, but at some point, something will go wrong, or a new requirement will arise that necessitates changes. To change the system, it is necessary to understand what it does now and how it needs to be altered to fix problems or meet ever-changing requirements.
This could also lead to passive fatigue, “the inability to intervene in a timely manner when an automated task starts to go awry”, which David Glenn covered in a recent blog. This will only be exacerbated when an individual does not have the knowledge to spot where something does not look right or how to react when things go awry.
This malaise is not limited to asset management. Speaking to friends and colleagues in other industries including health and retail, they all report the same—increased automation is de-skilling jobs to the point of over-reliance on a system vs using a system as a tool to support the knowledge and abilities of the human brain.
Automation, STP, and increased efficiency all bring benefits to the business, but there is a potential dark side if the broader needs for vigilance and the ability to change and adapt systems are ignored. Access to that deep understanding of data and detailed processing requirements is essential for long-term, effective business management, either by retaining it in house or via external expertise from consultancies like Citisoft.