Dynamics of capital account and current account in India: Evidence from threshold cointegration with asymmetric error correction

ABSTRACT

This study investigates the nexus between India’s current and capital accounts using threshold cointegration and asymmetric error-correction approaches. The results validate the threshold cointegration between the current account (CA) and India’s capital account (KA). The asymmetric error correction-based results support the unidirectional Granger causality from KA to CA. Further, results reveal that KA’s cumulative sum of positive and negative deviation substantially impacts the CA. Finally, results of equilibrium adjustment path asymmetric effect from CA to KA suggest that KA reverts to equilibrium in the short run when CA decreases.

KEYWORDS:

• Current account
• capital account
• threshold cointegration
• asymmetric adjustment
• causality

JEL CLASSIFICATION:

• F41
• F32
• B22
• F21
• B41

Acknowledgments

We would like to thank David Peel, Editor of the journal, for giving us a chance to revise this paper. We are thankful to both anonymous referees for their valuable comments and insightful suggestions, which enabled us to improve the quality of this paper to a large extent.

The study has not received any funding from anybody. A usual caveat is applied.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ Following prior literature (Garg and Prabheesh 2015) and to indicate the country’s level of international competitiveness, we have taken the CA and KA in real terms and scaled-down by GDP instead of taking the CA and KA in nominal terms.

² The factors like exchange rate and interest rate are the prominent factors, which can affect the nexus between CA and KA in India. However, the present study explicitly focuses on the one-to-one nexus between CA and KA using the familiar threshold cointegration and asymmetric error-correction approach. The primary motivation of the paper is to examine the asymmetric transmission mechanism between CA and KA, which could provide clear-cut ideas on how both variables asymmetrically adjusted to each other. It could be interesting to examine if we introduce some prominent factors into the CA and KA model and explore how the non-linear adjustment is possible between variables. We have preserved this dimension of research for the future.

³ The four asymmetric models are threshold autoregressive (TAR), momentum TAR (MTAR), consistent TAR (CTAR), and consistent MTAR (CMTAR). For more detailed discussion, see Enders and Siklos (2001).

⁴ The positive and negative changes in KA indicate the (quarter) months. 1/-0.499 = 2${\rho }_1$2 quarters, 1/-0.233 = 4.29${\rho }_1$4 quarters.

⁵ The C-MTAR model is more dynamic and superior than the CTAR model (Enders and Siklos 2001).

⁶ See Mallick, Behera, and Dash (2020) and (2021); Mallick and Behera (2020) for a detailed discussion of different types of hypotheses.

⁷ The study rejects the null hypothesis of no evidence of Granger causality (H₀₁: α_i⁺ = α₁^‒ = 0 for all lags), indicating the unidirectional causality from KA to CA.

⁸ The reported F statistic is insignificant for the CA equation, indicating that CA does not Granger cause KA.

⁹ In the CA specification, also the estimated coefficients of ${\rho }_1$ and ${\rho }_1$ are −0.158 and −0.089, respectively, and are not significantly different from zero.