Grootendorst thanks Marie Dean and seminar participants at the University of Toronto, the 2006 meetings of the Canadian Association for Health Services and Policy Research and the 2007 International Health Economics Association meetings for helpful comments on an earlier version of this manuscript. [...] Lichtenberg estimates the contribution of both medical technology (measured using the number of NMEs approved for use in the US each year) and the amounts of medical care and other inputs on LE at birth, whereas the others estimate the contribution of productive inputs alone. [...] The health care spending inputs considered in these models are not the homogenous inputs of the textbook production function because they embody – to varying degrees, depending on the input – the current technology. [...] This assumption is perhaps less defensible for the case of the distribution of alcohol, cigarette consumption and other lifestyle factors – the distribution of these factors could evolve differently in the different provinces. [...] Finally, we considered models with and without year dummies included to assess the robustness of the estimates to the presence of time-varying confounders.