Towards a systemic understanding of the dynamic interest income performance of a retail bank

Abstract

The strategic management of financial services firms has become an increasingly complex task over the last three decades. The banking industry, until the end of the 1970's, was heavily protected and characterised as a staid and stable industry in the age of '3-6-3' banking (borrow at 3%, lend at 6%, be on the golf course by 3 p.m.. This age came to an abrupt end and banks were forced to become competitive and fast [Matten: 1996]. Today, the banking industry is characterised by volatile markets, increasing consumerism, globalisation and competition, the rapid introduction of new technology, a changing regulatory environment and an increasingly inter-coupled business environment. Under conditions of increasing complexity and interconnectedness, it has become increasingly difficult to manage a modern financial services firm by intuition alone. The aim of this dissertation is to demonstrate how a system dynamics simulation model can expedite learning, insight and foresight for a typical retail bank involved in lending and deposit-taking activities in a retail banking market. The model represents the interest income generating system of a typical retail bank embedded in its environment. The exogenous environmental inputs can be initialised with a wide range of different characteristics; a stable interest rate environment versus a volatile environment, a mature market with low loans growth versus a fast growing market, and so on. Similarly the characteristics of the retail bank may be initialised with a wide range of characteristics; a start-up bank versus an established bank with a large loan and/or deposit book, aggressive versus conservative credit policies, etc. In particular, we develop a dynamic hypothesis of the complex structure of the interest income generating system of a retail bank. This hypothesis is translated into a quantitative system dynamics simulation model that serves to provide an experimental tool for testing strategic responses of a typical bank under different environmental scenarios enabling us to infer how performance arises. The model demonstrates how performance arises ex-ante from the dynamic structure of the bank's resource system. The experimental results provides a dynamic view on the performance characteristics of the income generating system of a retail bank described in the language of the Dynamic Resource Systems View of Warren and Morecroft [Warren: 2000]. - Firstly, the results show the scale of the resources developed; the size of the loans book, the number non-performing loans, the size of the depositor book book, and the consequent financial performance at any cross-sectional view in time. - Secondly, the experimental results indicate at what rate the resources have developed; the speed at which the advances book has developed, the rate with which the depositor book is generated under a particular set of exogenous circumstances and endogenous policies. - Thirdly, the results show that the observed performance of a number of similar banks can be described by a single, generic systemic feedback structure. Performance differences arise due to different resource endowments and decision policies of decision agents with different dominant logics about how they should structure their business. - Fourthly, the results show that the time evolution of financial performance of the interest income generating system is path dependent, that is, the time path of performance is constrained by its current resource endowments, policies and capabilities. The central thesis of this dissertation is that the dynamic behaviour of interest income is a result of an endogenous systemic structure in response to external exogenous factors such as interest rate movements. It is our contention that the complex interrelationships between issues such as capital management, interest rate management, liquidity management, cost management, strategic objectives, etc, justifies the use of a more complex systemic and dynamic framework for analysis that may produce new insights into the behaviour and management of the interest income system for a typical retail bank. The purpose of this research is to demonstrate how the dynamic behaviour of interest income generation of a typical retail bank arises as much from the interaction of decision agents and tangible and intangible resource flows within the bank's system as from any exogenous inputs into the system. The model developed here is a highly aggregate model of limited scope encompassing the traditional core issues facing a typical retail bank and sheds some light on the issues addressed by the model. Although being of limited scope, the model is deemed to be sufficiently comprehensive to demonstrate how the dynamics arise. The work presented here is not fully complete, it represents a plausible hypothesis of system under study but has not been sufficiently validated to be considered scientifically complete. Much further work is required. However, this dissertation represents work that has been undertaken to date in an attempt to understand the system under study from a systemic perspective. Further extensions may be undertaken to provide insight into some of the current issues facing the modern retail bank executive. Understanding the dynamic structure of a retail bank sheds some light on the potential responses available to improve performance. The contribution of this research lies in the systematic and systemic abstraction of the interest income generating structure of a typical retail bank, the development of a quantitative model thereof and the insights obtained from this. The model developed may be considered to be a generic model applicable to any bank with a similar structure but with different objectives, parameter values, resource levels, policy parameters and exogenous inputs and provides different performance time paths with the same structure. The insights about solutions are specific and time dependent. This provides an insight that traditional static models do not provide by identifying specific solutions applicable at specific time frames and valid for specific durations thus coupling a time dimension to all solutions.