Structural Equation Modeling in Archival Accounting Research: An Application to Disclosure and Cost of Capital

Structural equation modeling (SEM), a methodology currently underutilized by archival accounting researchers, enables examination of paths linking constructs in complex models. In this paper, we first discuss SEM, including basics, underlying assumptions, advantages and disadvantages, properties of the estimates, and best practices. We then apply SEM to an accounting topic by using common factor analysis to create latent variables (constructs or factors) for earnings quality, voluntary disclosure, information asymmetry, and cost of equity capital and then examining the paths among those constructs. We find that higher earnings and voluntary disclosure quality are negatively associated with cost of capital both directly and indirectly through information asymmetry. Further, the paths from voluntary disclosure quality to cost of capital are significant even after controlling for the paths from earnings quality to cost of capital. SEM allows us to perform our analysis, which would not be possible with other techniques, and potentially offers fruitful avenues for future research in accounting.