Does the China-U.S. Trade Imbalance Stem from Difference in the Comparative Advantages of Service Trade? Empirical Analysis Based on SVAR Model

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

“…Based on the time series data adopted in this paper, it was necessary to conduct a stationarity test for each time series variable. The stability test method of variable data used in this paper is the ADF unit root test [ 26 , 27 , 28 , 29 , 30 , 31 ], and the test results of each variable are shown in Table 2 . It can be seen from Table 2 that none of the variables passed the stationarity test, that is, the original time series is unstable.…”

“…Therefore, it is very necessary to determine the optimal number of lag periods of the SVAR model when both are taken into account. In this paper, four information criteria (FPE, AIC, HQIC and SBIC) were used to select the optimal number of lag periods of the SVAR model [ 26 , 27 , 28 , 31 , 33 ]. According to the evaluation criterion that the smaller the value of an information criterion is, the better the result will be, the optimal lag period was finally determined to be 2 (asterisk “*” in the table), as shown in Table 4 .…”

“…In order to identify the “AB model” of SVAR, constraints need to be imposed on the matrices “A” and “B”. Chen (2014) [ 34 ] and Cheng et al (2022) [ 26 ] pointed out that for the “AB model” with M endogenous variables, [2M 2 − M(M + 1)/2] constraints need to be imposed on matrix “A” and “B” to identify the AB model. Therefore, 51 constraints need to be set in this paper (See Table A4 in Appendix A of this paper for the parameter estimation results of model).…”

“…The constraint method is to follow the idea of Cholesky decomposition: Matrix “A” is set as the lower triangular matrix and all the main diagonal elements are 1. Matrix “B” is set as a diagonal matrix [ 26 , 34 , 35 ]. So the constraint matrix “A” and “B” are defined as follows: …”

“…The sufficient and necessary condition for the stability of the SVAR model is that all the characteristic roots of the characteristic equation are in the unit circle [ 26 , 34 , 36 ], that is, the reciprocal of the roots of the polynomial equation with autoregressive coefficient are in the unit circle. In this part of the paper, model stability is tested (see Figure 2 for details).…”