Debating the assumptions of the Thirlwall Model: A VECM analysis of the Balance of Payments for Argentina, Brazil, Colombia, and Mexico

Danilo Spinola


This article challenges the main assumptions of the Balance of Payments Constrained Model (BPCM, aka Thirlwall model) related to the long-run: (1) equilibrium of the trade balance. (2) Stability of price-effects. (3) Foreign income growth positively affecting domestic income. Some authors raise the argument that the BP is rarely observed in equilibrium (Alonso & Garcimartín, 1998); price effects, through the real exchange rate, do affect the long-run (Rodrik, 2008); and foreign income has no effect (or negative) on domestic income (Razmi, 2016). The BPCM is based on its assumptions to defend the existence of a long-run growth rate compatible with a stable growth of the balance of payments, in which the effective growth rate converges to avoid external constrains (McCombie & Thirlwall, 1994; Thirlwall, 1979). In order to challenge the assumptions of the BPCM, we apply a time series co-integration Vector Error Correction Model (VECM) using the BPCM related variables to Argentina, Brazil, Colombia, and Mexico, the larger countries in Latin America. The data source is the Penn World Tables (PWT) for 1950-2014. We apply impulse-response and permanent shocks in selected variables, observing their effects on Real Exchange Rate, GDP, and Trade Balance. The results are compared to the assumptions raised in this research; we empirically find that the BPCM assumptions are not empirically robust for the selected countries. This offers an invitation to more empirical work that can strength the arguments of the BPCM model.

Keywords: Balance of Payments Constrains, Latin America, Economic Development

JEL Classification: O11, F41, E12

Download the working paper