In a turbulent and globalized economy, the focus is no longer on growth but on stability. The stability of a business is of paramount importance to investors, shareholders, banks, insurance companies, etc. Banks in particular have great interest in monitoring their customers in order to appreciate their state of health. Complexity allows to establish such a mechanism and in an innovative fashion. Based on each customer’s transactions with his bank, not on a Balance Sheet, it is possible to measure the customer’s complexity and to track its variation over time. In particular, the rate of change of complexity establishes a unique holistic index of the stability of a business. The computation of the index on a monthly basis allows a bank to quickly rank and classify its customers. In other words,
Bank-client transactions data may be analyzed to determine the bank’s state of health
Sudden changes in the complexity of a business points to “traumas”. Independently of whether these changes are of endogenous or exogenous nature, they point to potential customer-retention problems: either a company is defaulting or it is changing banks. In both cases, the bank must be aware of the situation. However, a company has the means to occult its true state for many months while still applying for and obtaining credit. While a stable business will point to a constant value of complexity, sudden changes reflect some sort of rupture with the status-quo. It is precisely this approach that is used as a pre-alarm system alerting a bank that a particular customer requires special attention. The scheme is illustrated below.
Ontonix has developed a specific pe-alarm system which performs time-domain complexity analyses of portfolios (batches) of companies/clients. It measures their stability and signals those that are becoming unstable, i.e. those requiring attention. The stability index of a client ranges from 0 to 100% and reflects the rate of change of his complexity. The system requires monthly data on each customer’s transactions, such as balance, credits, No. operations, assets in stocks, etc.
The system accesses a single customer-transactions data-base and processes them in a sequential fashion. The output is a simple text file in which the stability index of each customer is indicated. Customers with critical stability values are flagged. An example of a client portfolio of approximately 220 companies is illustrated below, where the so-called Stability Profile is shown. The red vertical lines represent a stability limit, below which a particular client is investigated in greater detail.
The computation is very fast. Two years of transactions of 500000 clients can be processed in just under 60 hours on one processor, in one hour on a 64 CPU cluster. This means that the bank has a very clear picture of its own state of health on a monthly basis.