On the link between current account and fiscal imbalances in the presence of structural breaks: Empirical evidence from Egypt

Highlights

• Study the link between fiscal and current account imbalances in the presence of breaks.

• Endogenous determination of break points.

• Use ARDL model and bootstrap causality test to study the link.

• Reject twin deficit hypothesis over the long run and find support for Ricardian equivalence hypothesis.

• Find support for twin deficit hypothesis in the short run.

Abstract

This study provides an empirical analysis for Egypt of the link between fiscal and current account imbalances, which is also known in the literature as the twin deficit hypothesis. We assess the link between the two deficits in the presence of structural breaks using the Autoregressive Distributed Lag (ARDL) model and the bootstrap causality test. Our empirical results suggest that we reject the twin deficit hypothesis over the long-run. Instead, we find support for the Ricardian equivalence hypothesis. The absence of any long-run relationship between the two variables is backed by our bootstrap causality test and our FM-OLS (Fully Modified Ordinary Least Squares) regression results. Still, we find support for the existence of a link between the two deficits in the short-run. The results from the ARDL model with breaks contrast with those in earlier studies without breaks. The latter rejected both the twin deficit and the Ricardian equivalence hypotheses.

Introduction

Most economists believe that uncontrolled large budget deficits can lead to serious economic problems, such as inflation, low growth, and a worsening of the current account balance. The economic importance of current account imbalances has been subject to several debates in academic and policy circles. These discussions mainly focus on whether external imbalances are pure trade competitiveness phenomena, or otherwise generated by government macroeconomic policies. In the last two decades, several papers examined the relationship between government and current account deficits in different countries around the world. This relationship is known in the literature as the twin deficit hypothesis. The Egyptian economy is not an exception to this trend. The recent economic dynamics in Egypt, characterized by recurrent current account deficits, has rekindled the debate over the relationship between the two variables in the country. El-Baz (2014); Helmy and Zaki (2017), and Marinheiro (2008) are examples of studies that have looked into the nexus between the two variables in Egypt.

In an international context where several economies are compelled to reduce their growing current account deficits, clarifying the nature of the link between the two deficits in Egypt could prove useful for policymakers. These insights would provide policymakers with a better understanding of the relationship between the budget deficit and the current account deficit, which is still controversial; and thereby help inform the design of appropriate policy responses.

In this paper, we aim to provide additional insights on the dynamic relationship between budget and current account deficits in Egypt based on data from 1962 to 2017. The main rationale for a new study on this topic in Egypt relates to the methodological limitations of existing studies that could lead to misleading conclusions. These existing studies on the twin deficits in Egypt, including Marinheiro (2008), El-Baz (2014), Nazier and Essam (2012) and Helmy and Zaki (2017) fail to consider the possibility of structural breaks in the data. These studies reject the twin deficits hypothesis and even find a reverse causality going from the external imbalance to the fiscal disequilibrium. This finding, however, is not the case for Nazier and Essam (2012) who find evidence for the twin divergence hypothesis. Yet, studies such as Perron (1989) and Hatemi-J and Shukur (2002) have cautioned about the potential implications of ignoring structural breaks when using time-series data. For example, ignoring structural breaks in the augmented Dickey-Fuller test can produce misleading results regarding the stationarity of the time-series or lead to a false acceptance of the null hypothesis. Moreover, Perron (1989) has shown that standard tests of the unit root hypothesis against trend stationary alternatives fail to reject the unit root null, if the true data generating mechanism is that of stationary fluctuations around a trend function containing a break.

Numerous factors peculiar to Egypt point to the existence of structural breaks in the macroeconomic time-series of the country. For example, during the study period (1962–2017), Egypt experienced several political and economic changes through Structural Adjustment Programs (SAP) requested by the International Monetary Fund (IMF) and the World Bank.

In light of these potential issues, we offer a re-examination of the empirical relationship between the two deficits that accounts for structural breaks in the data. We test for the existence of structural breaks in Egypt’s time-series data by using the methodology suggested by Bai and Perron (1998); J. Bai & Perron, 2003. The most important advantage of the methodology proposed in the latter papers is the endogenous determination of the breakpoints rather than an exogenous choice determined by the researcher.

Furthermore, in contrast to the existing studies on Egypt, we use the Autoregressive-Distributed Lag (ARDL) approach to assess the long-run relationship between the budget deficit and the current account deficit. We also test for the existence of any causality between the two variables using the bootstrap causality test proposed by Hacker and Hatemi-J (2006). The latter test is based on the Toda and Yamamoto (1995) test of causality, which is valid even if the variables are not integrated with the same order. The Hacker and Hatemi-J (2006) bootstrap causality test is adequate for small sample sizes and has the additional advantage of generating more accurate critical values that are robust to non-normality and time-varying volatility. In addition to the Hacker and Hatemi-J (2006) causality test, we perform another robustness test of the results using the FM-OLS (Fully Modified Ordinary Least Squares) method introduced by Phillips and Hansen (1990)¹ . The latter approach is adequate for a robustness check in the context of this study for several reasons. First, it allows testing for the existence of one or multiple co-integrating vectors among variables. Second, similar to the ARDL method, it allows for the combination of stationary and non-stationary variables; and lastly, it accounts for both serial correlation effects and endogeneity of the regressors.

The remainder of the paper is structured as follows. We provide in the next section a brief examination of the budget and current account deficits in Egypt. The third section presents a cursory review of the relevant literature on the twin deficit hypothesis. In the fourth section, we present the data sources, and discuss the empirical results in the fifth section. We conclude in the last section.