Explaining Implausible Results in Shadow Economy Estimation Using MIMIC Models

Abstract

For decades, economists have been trying to estimate the magnitude of the shadow economy (SE), which is not directly observable. This paper explores how the MIMIC (Multiple Indicator Multiple Cause) model can yield estimates of the SE/GDP (the proportion of the SE to the gross domestic product) below 0%, above 100%, and other implausible results. The focus is on the new calibration methods by Dell’Anno (2022) and data on the Czech Republic (1993–2021). The paper concludes that one of the leading causes of implausible results is the misalignment between the SE definition implied by the MIMIC model and that used for the exogenous estimates applied for calibration. Therefore, the authors propose testing the alignment between the SE definitions, such as assessing trends in the latent variable (first-stage scores) resulting from the MIMIC model and the exogenous estimates or applying regression or correlation analysis.