Much evidence has been presented that legal institutions, and particularly property rights, cause economic growth. Yet, precisely for those countries with the most urgent need for development, we know less about the direction of causation and the functional form of the association between the strength of legal property rights and per-capita incomes. A seminal work in this literature, by Acemoglu, Johnson and Robinson (2000), provides optimistic outlooks, whereby also the Least Developed Countries could grow out of poverty if they implemented strong legal property rights.

In this paper, we challenge two of Acemoglu et al’s findings. First, we derive a range of micro-mechanisms suggesting that the relationship between the strength of legal property rights and output may not be linear as asserted by Acemoglu et al and a number of other scholars. Instead, our mechanisms imply the existence of a weak relationship between these variables in the lowest stages of development and a structural break after early growth. Second, we suggest why the direction of causation may turn around after early development.

We confirm these predictions of our small models with non-parametric, instrumented regressions. Our results hold not only for Acemoglu’s dataset but across a range of measures for property rights protection. Property law is shown to matter greatly for Mid-Income Countries but it is important not to overestimate its potential for the Least Developed Countries. We must not stop searching for the causes of growth in the earliest stages of development; property reform alone does not resolve the poorest nations’ poverty traps.