Camilo González-Sabogala, This article focuses on the conditional forecasting of Colombia's current account by identifying its main macroeconomic determinants. To this end, it reviews the factors that, according to the literature, explain current account dynamics and groups them into categories that allow for their systematic use in forecasting models. The methodology combines different econometric approaches to generate multiple specifications and selects those with the best performance and strongest consistency with the theoretical channels. The forecast evaluation shows that the strategy presented in this study performs well relative to market analysts' surveys, thereby offering a complementary tool for external balance analysis.
The most recent
Approach
To obtain timely and reliable forecasts of Colombia’s current account, this paper proposes a methodology that combines information on its main macroeconomic determinants to generate a broad set of forecasts and select those with the best performance. To this end, the most relevant factors are organized into groups, allowing for a systematic comparison of different ways of explaining current account dynamics and producing forecasts that are consistent with both the statistical evidence and economic intuition. The analysis uses quarterly data for the 2000-2024 period and evaluates the predictive performance of the methodology over 2021-2024.
Contribution
The paper’s main contribution is to provide a practical and systematic strategy for forecasting the current account of the balance of payments based on its key macroeconomic fundamentals. Rather than relying on a single specification, the methodology makes it possible to assess a large number of alternative models simultaneously and select those with the strongest predictive performance and greatest economic coherence. From a policy perspective, this approach complements the construction of macroeconomic scenarios for decision-making by generating current account projections that are consistent with international statistical standards and with the variables typically monitored by analysts. In addition, the tool provides an empirical input for general equilibrium models and, because it relies on projections generated by those same exercises, preserves consistency with the broader macroeconomic forecasting framework.
To obtain timely and reliable forecasts of Colombia’s current account, this paper proposes a methodology that makes it possible to assess a large number of alternative models simultaneously and select those with the strongest predictive performance and greatest economic coherence. The results indicate that this strategy produces current account forecasts that are more robust and better aligned with their underlying macroeconomic determinants.
Results
The results indicate that this strategy produces current account forecasts that are more robust and better aligned with their underlying macroeconomic determinants. In particular, by comparing a wide range of specifications and retaining only the best-performing ones, the approach reduces forecast dispersion, discards extreme results, and concentrates the analysis on a more precise set of projections. For Colombia, the empirical evidence suggests that this methodology matches or outperforms market analysts’ survey forecasts both at short-term horizons and in annual projections, supporting its usefulness as a complementary tool for monitoring the external sector balance. Moreover, in 2022, the methodology successfully anticipated the widening of the current account deficit and linked it primarily to strong domestic demand growth and higher import prices, effects that were partially offset by an improvement in the terms of trade.