This paper asks whether consolidated government in Canada is too large in relation to its effect on economic performance. Rather than impose a parametric shape, nonparametric methods are used to motivate the nonlinear form that best describes this relationship. Using data from 1929 through 2014 and controlling for the different time-series characteristics of the variables, the bias in standard errors that can be expected from correlations arising across time, and the presence of Wagner’s-law-type endogeneity in the relationship running between government size and private performance, we find evidence consistent with government size peaking in its effect on private output at roughly 32 percent of GDP.