Analysis of bilateral air passenger flows: A non-parametric multivariate adaptive regression spline approach

Gravity models have long served as a framework for analyzing bilateral flows, trade and many other activities. To estimate a gravity model, (parametric) linear regression techniques have been commonly employed to develop the relationship between passenger flows and factors that can significantly influence these flows. This study explores the application of an alternative method, the non-parametric multivariate adaptive regression spline (MARS) technique, to identify the determinants for air passenger flows between pairs of countries. The data of 2006 and 2007 air passenger flows between pairs of countries in the APEC region were collected to develop the MARS models. Results indicate that distance between the countries, annual import value, national per capita income, unemployment rate and consumer price index of the origin country, as well as GDP, annual import value, and consumer price index of the destination country are significant determinants for bilateral air passenger flows.