© 2018 authors. Published by the American Physical Society. Although active learning is supported by strong evidence of efficacy in undergraduate science instruction, institutions of higher education have yet to embrace comprehensive change. Costs of transforming instruction are regularly cited as a key factor in not adopting active-learning instructional practices. Some cite that alternative methods to stadium-style, lecture-based education are not financially viable to an academic department. This paper examines that argument by presenting an ingredients approach to estimating costs of two instructional methods used in introductory university physics courses at a large public U.S. university. We use a metric common in educational economics, cost effectiveness (CE), which is the total cost per student passing the class. We then compare the CE of traditional, passive-learning lecture courses to those of a well-studied, active-learning curriculum (Modeling Instruction) as a way of evaluating the claim that active learning is cost prohibitive. Our findings are that the Modeling Instruction approach has a higher cost per passing student (MI=$1,030/passing student vs Trad=$790/passing student). These results are discussed from perspectives of university administrators, students, and taxpayers. We consider how MI would need to adapt in order to make the benefits of active learning (particularly higher pass rates and gains on multiple measured student outcomes) available in a cost-neutral setting. This approach aims to provide a methodology to better inform decision makers balancing financial, personnel, and curricular considerations.