The relationship between LNG price, LNG revenue, non-LNG revenue and government spending in China: An empirical analysis based on the ARDL and SVAR model

Abstract

The present paper examines the dynamic relationship between liquefied natural gas (LNG) price, LNG revenue, non-LNG revenue and government spending (GOVS) in China using autoregressive distributed lag (ARDL) and structural vector auto-regressive (SVAR) model. The goal of carrying out ARDL and SVAR together is to consolidate and strengthen the consistency of the results obtained from both approaches. ARDL results show that a positive influence relationship between both short-run and long-run LNG prices, LNG revenue, non-LNG revenue and GOVS, but there was no significant relationship between LNG price and GOVS. The SVAR also substantiates the results of ARDL test and provides further insight which shows that long-run fiscal synchronization hypothesis is evidenced between the LNG revenue and GOVS, while spend-tax hypothesis exists in the long-run between GOVS and non-LNG revenue. It is also evidenced that there is a complementary relationship between LNG revenue and non-LNG revenue, but this complementary role is stronger than the substitution role. Since non-LNG revenue has a greater impact on GOVS in the short-run, and the impact of LNG prices and LNG revenue on GOVS in the long-run increases over time, thus, GOVS mitigates the direct impact of non-LNG revenue to some extent, and that an appropriate allocation of spending in the non-LNG industry will have a positive impact on the development of the market economy supporting the Keynesian and spend-tax hypothesis.

Study on the Dynamic Relationship between Chinese Residents’ Individual Characteristics and Commercial Health Insurance Demand

“…However, the VAR model does not set any constraint conditions and simply reflects the dynamic relationship between endogenous variables. It does not explicitly explain the meaning of economic structure among endogenous variables in the model, that is, it ignores the implied economic structure among endogenous variables and the influence relationship of endogenous variables over the same period to a certain extent [ 25 ]. Based on these defects of the VAR model, in this paper we constructed a 6-variable SVAR(P) model to analyze the dynamic influence mechanism of residents’ age characteristics, residents’ household registration characteristics, residents’ gender characteristics, residents’ education characteristics and residents’ marriage characteristics on the demand for commercial health insurance.…”

“…First of all, since the SVAR model was adopted in this paper (the verifications of the basic assumptions of the SVAR model are shown in Table A2 and Table A3 in Appendix A of this paper), which reflects the economic structure relationship between variables [ 25 ], the Granger causality test was not necessary. Based on the research purpose and empirical research, this paper recognizes the following conclusions: “age”, “sex”, “married”, “urban” and “edu” are the Granger causes of “lnhealthins”.…”

The dynamic relationship between trade openness, foreign direct investment, capital formation, and industrial economic growth in China: new evidence from ARDL bounds testing approach

“…Compared to the traditional co-integration model, the ARDL cointegration model has a number of advantages. Firstly, ARDL does not require the implementation of the same integrating sequence for all variables in the model (we can use the 0 order or I (0) integration of the data variable and the 1 order or I (1) integration for the analysis) (Mohammed and Ruslee, 2015;Hao, 2021). Secondly, the ARDL co-integration method is more suitable for sequence data with small sample size, and the processing and interpretation of the data in the analysis process is relatively simple.…”

“…According to Pesaran et al (2001), whether or not to reject the null hypothesis is determined by comparing the F-statistic of the correlation co-efficient in the F-test with the critical value of the ARDL co-integration co-efficient of the critical value of the maximum asymptotic spread of the F-statistic (Hao, 2021). However, as the selected sample size is rather small, we compare the value of the model F-statistic with the threshold value of the asymptotic distribution of the F-statistic suggested by Narayan (2005).…”

“…According to the relevant definition, if the ARDL (p, q, r, m, n) co-integration model passes the criterium test of the F-statistic assay of the asymptotic dispersion, then a cointegrating or longrun link exists between the variables (Hao, 2021). Following Odhiambo (2009), the short-run economic activity parameters are derived by estimating an error-correction model with respect to the long-run estimation, and there is at least one causality amongst the variables in the direction determined by the Fstatistic of the long-run estimation and the lag of the ECM.…”