Reexaming the link between instability and growth in Latin America: A dynamic panel data estimation using K-median clusters

CECILIA BERMUDEZ; CARLOS DABUS; GONZÁLEZ, GERMÁN HÉCTOR

LATIN AMERICAN JOURNAL OF ECONOMICS

Pontificia Universidad Católica de Chile

Lugar: Santiago; Año: 2015 vol. 52 p. 1 - 23

0719-0425

We estimate a dynamic panel data model to empirically assess the relationship between instability -measured by GDP growth volatility and inflation rate- and growth in Latin America over the last fifty years. The presence of aberrant observations seriously distorts the results of classical estimations and leads to spurious regimes. Thus, we use the k-median clustering algorithm as a pre-estimation technique in order to mitigate the outlier problem and to properly identify different ?scenarios? for the instability variables. Our main findings are that while inflation has a significant and negative effect on economic growth only over an average annual rate of 57%, our volatility measure also has a significant and negative effect on growth but without any differential effect among the various clustering techniques applied to the data.