Inflation and Inflation Uncertainty in Growth Model of Barro: An Application of Random Forest Model

Abstract

One of the major problems of the empirical economists while building an economic model is the selection of variables which should be included in the true regression model. Conventional econometrics use several model selection criteria to determine the variables. Recent years' developments in Machine Learning (ML) approaches introduced an alternative way to select variables. In this paper, we have an application of ML to select variables to include for a nonlinear relationship between inflation and economic growth. Among ML methodologies, Random Forest approach is one of the most powerful to capture nonlinear relationships. Therefore, we applied RF and found that both high and low inflation can be the cause of low economic growth which is a major contribution of the paper to economic literature. Moreover, in the paper, as an outcome of RF there are other variables effecting economic growth with an order of importance.

Keywords

• Growth
• Inflation
• Machine learning
• Random Forest

References

1. Barro R.J. (1991) Economic Growth in a Cross Section of Countries. The Quarterly Journal of Economics 106(2):407-443.
2. Breiman L. (2001) Random Forests. Machine Learning 45(1):5-32.
3. Sala-i-Martin X.X. (1997) I Just Run Two Million Regressions. The American Economic Re- view 87(2):178-183.