The complex and interconnected world in which organizations operate presents many challenges to the traditional neo-classical view of research and management and associated research techniques. Fundamental to the operation of financial capital markets, investor confidence relies on accurate investment analyst earnings forecasts. We propose agent-based modeling (ABM) as a viable tool to account for the interaction of local and environmental factors to determine organizational success. In an illustrative case study of Frontier Airlines, we develop and execute an ABM of Frontier’s consumer airline market to derive market share for the upcoming year. In the model, Frontier is impacted by internal policies, competitors, and environmental factors of fuel costs, federal regulation, and credit availability. We conclude with a discussion on how ABM can be effectively incorporated into future research activities and decision-making situations.
Investor confidence. What is it and what affects it? State Street is one of the world’s leading providers of financial services to institutional investors. The company tracks and records the State Street Investor Confidence Index that “provides an objective, quantitative measure of global risk tolerance of the world’s sophisticated investors” [30]. Investors rely on investment analyst forecasts as a gauge of performance for individual firms, industries, national economies, and the global financial market as a whole and evidence shows that firms meeting or exceeding analysts’ earnings forecasts (i.e. expected revenues) receive an equity price premium (i.e. better stock price return) [25]. Underlying this phenomenon is the belief or confidence in the accuracy of the analyst forecasts [15]. A key component in an analyst’s forecast of earnings for a particular firm is that firm’s anticipated percentage of sales in the industry or industries in which the firm operates relative to its competitors – commonly referred to as expected or projected market share.
How can a single human being reliably account for all, or even just the most pertinent, factors that may impact the future financial performance of a firm? Research has shown that the human mind struggles to capture and process excessive amounts of information and have difficulty “connecting the dots” of cause and effect relationships when numerous factors come into play in a decision-making scenario [19]. Many simulation approaches such as the Kim et al. [16] reinforcement learning algorithm designed to find the optimal direct marketing strategy focus on static, direct relationships. In an increasingly connected, global business environment, though, more participants and variables than ever exist and they interact and affect each other both directly and indirectly [3]. These indirect effects represent complexity. Social, political, and environmental events on the other side of the world may affect the business success here in the US. For instance, Airlines often engage in ‘code-sharing’ where the host airline operating the flight will allow key strategic partners to sell seats on the flight. This is particularly common for international travel where one trip from the United States to India may use multiple carriers for different legs of the trip. A pilot strike, change in government regime or travel policies, spike in oil prices, steep discounts offered by a local competitor on a single leg, or a natural event such as the recent eruption of volcanic ash in Iceland may affect the number of flights and routes available between the US and India. These individual events, individually or in combination, directly impact the airlines servicing the various legs as well as all the partners in the trip and, of course, the travelers. Song et al. [28] state traditional capital markets research for the airline industry, particularly when examining international travel, has difficulty controlling and accounting for major external changes in the environment. Computing technology through the use of agent-based modeling (ABM) simulation, however, offers a way of modeling the complex, interactive effects of many agents acting and reacting to the actions of each other and the environment thus allowing the modeler to view potential system states given a certain set of parameters [23].
ABM, with direct historical roots in complex adaptive systems (CAS), represents autonomous entities referred to as agents – each with their own dynamic behavior and heterogeneous characteristics – that interact with each other and their environment, resulting in emergent outcomes at the macroscale that can be used to quantitatively analyze complex systems [13]. Intelligent agents and ABM simulation have been used in a wide variety of domains such as planning military operations [14], understanding Web services management [9], monitoring workflow [33], etc. Most of this simulation activity and research focuses on a narrow scope with little to no macro-level environment interaction and offers no guidance on how others can integrate ABM into their own research and/or decision-making toolset. Through an illustrative case study, this paper guides the reader through the development of an ABM designed specifically to account for macro-level factors affecting one specific agent’s future business prospects by simulating the market and developing an expected market share for the next year. First, we present background on ABM, in particular emphasizing when ABM is appropriate and the general steps in the ABM creation process. The subsequent section describes the complex business scenario to be modeled followed by a detailed explanation of the modeling process, ABM construction, test design, and simulation results. The final section discusses how ABM can be incorporated into future research activities and complex decisions.
