Abstract:
Because of convergence, initial income is an important explanatory variable in many growth regressions. Its omission can lead to misspecification bias that may seem to – but still does not – contradict the common albeit somewhat controversial empirical finding that natural resources, if not well managed, tend to depress incomes and growth. We illustrate the bias using simple algebra and cross-country regression analysis with two different data sets: old and new. We distinguish between natural resource abundance and dependence and propose a method for identifying the level of national income at which the net effects of natural resources shift from adverse to advantageous.
