Neoclassical, semi-endogenous or endogenous growth theory? Evidence based on new structural change tests

摘要: One of the prevalent topics in economic growth literature is debate between neoclassical, semi-endogenous and endogenous theory for model that best describes data. An important part this discussion can be summarized three mutually exclusive hypotheses: \constant trend", \level shift" \slope hypothesis. The objective paper to classify countries according each these hypotheses analyze which theories seems favored. We approach problem two-steps: rst, number timing trend breaks are determined using Nunes Sobreira (2010); second, conditional on estimated breaks, break dates coecients identied, a statistical framework introduced test general linear restrictions disjoint broken model. Here, we prove result that, under certain conditions, standard F-statistic additional parameter restrictions, given rst step partition, converges asymptotically distribution usual chi-square distribution. further show how aforementioned formulated as parameters breaking apply methodology per capita output an extensive list countries. All our tests robust whether data I(0) or I(1) surpassing technical methodological concerns previous empirical evidence.