Internal versus external complexity: How organizations react

Abstract

This paper investigates the effects of environmental complexity on organizational performance by means of computational modeling. There are few works that prepare a bottom-up, mechanism-based account of organization-environment interactions. Rather, a lot of top-down, statistical models have been developed which fall short of explaining how an organization can handle uncertainties. The theoretical foundation of this paper is based on the concept of learning, which is a running thread linking organization science with the agent-based modeling paradigm. The definition of environmental complexity is based on Shannon's famous information entropy which represents a good conceptualization of complexity. The definition contains appropriately all facets of environmental uncertainty including degree of uncertainty, number of decision elements, and interdependence or interrelatedness among decision elements. In this paper, the process conception of learning is considered, because it is consistent with behavioral theory of organizational learning and can be modeled as a learning algorithm. Organizational learning is modeled as a Reinforcement Learning (RL) algorithm that operates as a trial-and-error search with delayed reward. The appropriateness of RL comes from the fact that the overall objective of managers is to align organizations towards environmental requirements and thus the environment feedback plays a key role in the decision making process within any organization. Among various conceptualizations of organizations, they are viewed problem-solving entities in this paper. The question considered is how environmental complexity affects organizational performance. To examine the effects of environmental complexity, an organization is exposed to different levels of environmental complexity, represented as decision rules. The requirements to accommodate environmental complexity have some counterintuitive effects on productivity. The results of the model show that the complexity of environment requirements has positive effects on organizational performance. In other words, if the organization can learn from the complexity of the environment, it can improve its performance. The computational model developed can be used to investigate other questions like the effects of internal complexity or organizational memory on performance.