The application of time-series data and analysis to study the effects of changes in unemployment rates on mortality rates has been a controversial issue in health-unemployment research for many years. This article presents new criticism against previous aggregate time-series regression models and concludes that these models are misspecified in functional form, and the t-ratios used in significance tests are grossly overstated. Future empirical analysis of the Economic Change Model of Mortality, i.e. the aggregate, time-series relationship between mortality rates and economic variables must pay more attention to the salient characteristics of time-series data and implications for regression results.