Asymmetric Cointegration between Services Exports and Economic Growth in Sri Lanka: Based on Nonlinear ARDL Model

Abstract

This study examines the asymmetric cointegration between services exports and economic growth in Sri Lanka. A nonlinear ARDL model is employed to analyze on the annual time series data over the period from 1984-2013.The results show that(1) ignorance of asymmetries between services exports and economic growth may lead towards misleading conclusions. (2) there is statistically significant difference in the response of economic growth to positive and negative shocks to the services exports both in long-and short-run. (3) the long-run coefficient of decreasing services exports is-0.65434 and it is six time lager than the coefficient (0.10917) of increasing services exports. Hence, economic growth is more sensitive to the downturn of services exports than upturn. This empirical evidence of asymmetries could be of important in advocating services exports policies in achieving long-run economic growth in Sri Lanka.

Innocent devils: The varying impacts of trade, renewable energy and financial development on environmental damage: Nonlinearly exploring the disparity between developed and developing nations

“…The study has engaged three co-integrating techniques non-linear autoregressive distributed lag (NARDL), Fully Modified Ordinary Least Squares (FMOLS), and Dynamic Ordinary Least Squares (DOLS) methods. The NARDL technique is best employed when there is a non-linear relationship among the variables in empirical analysis (Priyankara & Li, 2018), it has also been argued that the impact of the majority of the variables is non-linear (Esteve & Tamarit, 2012). Just like the traditional autoregressive distributed lag (ARDL), the NARDL can be employed regardless data series are purely co-integrated at levels or first difference, or as a mixture of both levels (Nkoro & Uko, 2016).…”

“…Also, it is observed that most of these studies engaged the bound cointegration test and the traditional Johansen co-integration test; which all assume a linear relationship among the variables of the regression model. However, most economic variables do not exhibit linear relationships instead, they are nonlinearly correlated (Priyankara & Li, 2018).…”

“…Hence, the choice of linear co-integration technique in the possibility of non-linear correlation aggregates spurious results (Gaies et al, 2022;Haug & Ucal, 2019;Priyankara & Li, 2018). To bypass this problem, this study is engaging the non-linear autoregressive distributed lag (NARDL) analysis by decomposing variables into positive and negative partial aggregates to allow for the asymmetric impressions of exogenous variables (Priyankara & Li, 2018;. Based on the discussion, the following are the two main objectives of this study:…”