The Economics of Complexity

The cost nobody budgets

Most financial institutions know what they spend on people, systems, premises, service providers and data. They are less likely to know how much of that spending is driven by complexity.

The cost is already in the budget. It is simply scattered across it: extra steps, reconciliations, recurring exceptions, duplicated data, manual controls and senior people spending time on problems that should not require senior people.

Complexity compounds

One additional process may seem inexpensive. The problem is interaction.

A new product needs another data source. The data source needs another reconciliation. The reconciliation creates another exception process. The exception process needs a control. The control needs reporting. The reporting goes to a committee.

Each step may have a reasonable explanation. Together, they create an operating cost that nobody assessed as a whole.

At no point did anybody approve “complexity” as an investment. Yet the institution is now paying for it.

The organisational tax

Complexity slows change because every improvement touches more interfaces. It increases key-person dependency because somebody needs to remember why a particular exception exists. It weakens service when problems cross organisational boundaries without clear ownership. And it creates opportunities for error as the same information is repeatedly transformed between systems and teams.

This is why complexity behaves like an organisational tax: individually small, collectively persistent and surprisingly difficult to remove.

Its cost is also measured in what does not happen: the product launch delayed, the client request left unresolved, the improvement repeatedly postponed because the people needed to deliver it are busy keeping existing processes running.

Measure before you simplify

Some complexity is necessary. Products, client needs and regulatory obligations can require it. The useful question is whether each layer still serves a purpose that justifies its cost.

Where are the manual hand-offs? Which processes rely on competing versions of the same data? Which exceptions recur often enough that they are no longer exceptions? Which tasks exist only because two systems do not speak to each other?

Start with a recurring process. Count the people involved, the systems touched, the time spent waiting and the work repeated. A precise monetary figure may be difficult to establish. A credible estimate is already more useful than treating the cost as zero.

For client-specific requirements, assess that cost against the value of the relationship and the reason for offering the service. A deliberate investment in a client is a commercial choice. An unexamined commitment can become a permanent subsidy.

Simplification begins with seeing the cost. Management then has to decide which complexity is worth paying for - and who is responsible for removing the rest.

This article reflects the author's views and is intended for general information and discussion. It does not constitute legal, regulatory or investment advice.