4.1. The Test Results, Stationarity and Cointegration
Before estimating the parameters, stationarity and cointegration tests were performed to show that the nonlinear panel approach ARDL is appropriate for the data. The unit root test is a popular method for stationary tests for both annual time series and panel data. The stationarity test is conducted in both “individual intercept” and “individual intercept and trend” in test equations. There are many types of unit root test for panel data such as Levin, Lin and Chu t (LLC) and Breitung t-stat with common unit root process; I’m, Pesaran and Shin W-stat (IPS), ADF—Fisher Chi-square (ADF), and PP—Fisher Chi-square (PP) with individual unit root process. The panel data in this study are balanced so that both hypotheses can be applied. The LLC test is chosen for the hypothesis “common unit root process” and the hypothesis “individual unit root process” is chosen for the IPS test. The results of panel unit root tests for logarithms of variables are summarized in Table 3
According to Table 3
, most of the series are non-stationary at level, but stationary at first difference, except for lnVA in LLC test of intercept and trend; lnTC in LLC test of intercept; and lnRF in LLC, IPS and ADF of intercept. Based on the majority of the results, it can be seen that the series are non-stationary at level but stationary at first difference, so a cointegration test should be performed to consider the long-term relationship between variables.
To analyze the cointegration relationship between variables in the panel data model, this study chooses the Pedroni and Kao tests because they are more comprehensive and universal. Cointegration tests are conducted for both “individual intercepts” and “individual intercept and individual trends” in the Pedroni test. By contrast, it is only conducted in the case of individual intercepts in the Kao test. The Pedroni test used seven test statistics (four tests for within-dimension and three tests for between-dimension). The Schwarz Information Criterion (SIC) automatically chooses the lag length with Newey-West automatic bandwidth selection and Bartlett kernel. Table 4
below presents the results of panel cointegration analysis.
According to the results of the Pedroni test in Table 4
, 4/7 tests are significant at the 0.01 level for both “individual intercept” and “individual trend and individual intercept”. This means that most cointegration tests in the Pedroni test result in the cointegration series. However, the Kao test gives the opposite result, meaning that the Kao test result does not give cointegration series at the level of 0.05, so is not compelling evidence to conclude clearly that series shows cointegration. Because of lnVA, lnTC, lnHR, and lnRF containing both I(0) and I(1), and when the existence of long-run associations is unclear, the ARDL technique is the most appropriate.
4.2. Estimated Results
This study uses the Pooled Mean Group (PMG) estimator to estimate the impact of investment in tourism infrastructure development on attracting international visitors to Vietnam. The PMG estimator is a well-known technique used in the estimation of a dynamic heterogeneous panel data model. Furthermore, by design, in addition to the panel regression results, the PMG also generates results for the individual units (Blackburne and Frank 2007
). Thus, computing the impact of tourism infrastructure development on attracting international visitors can assess both long-run and short-run responses for the general sample and each sample (each source market). First, the parameters are estimated by the PMG estimator for the general sample (panel data) with Automatic selection in three maximum lags, Akaike info criterion (AIC) in the Model selection method, and Linear trend in trend specification. Table 5
below summarizes the regression results by the PMG estimator for the general sample for both long-run and short-run.
As shown in Table 5
, the Log-Likelihood is large; Standard error of regression, Sum squared residual, and Akaike info criterion, Schwarz criterion, and Hannan-Quinn criterion statistics are relatively small, so the model is appropriate and fits with the data. For the long-run equation, all variables of interest are significant at the 0.01 level, so they are accepted. The estimated coefficients have the same sign as the initial expectation. Investment in tourism infrastructure such as transport and communications infrastructure, the hotel and restaurants industry, and recreation facilities, all positively impact attracting international visitors to Vietnam. Meanwhile, uncertainty factors have been negatively affected.
In the short-term equation, the coefficient of cointegrating equation has a negative sign (−0.4743) and is significant at the 0.01 level. This means that the variables converge to the long-run equilibrium, and the convergence rate is 47.43%. The lnTC and Dummy are not significant at the 0.05 level for all lags. By contrast, the variable lnHR is significant at the level, the first difference, the second difference, and lnRF at first difference and second difference, to be more specific, the sign of the coefficients of the negative lnHR and the sign of the positive lnRF coefficients. These findings imply that no significant impact of investment in transport and communications infrastructure has been found on attracting international visitors to Vietnam in the short-term. In comparison, there is a positive effect of investment in recreation facilities, while investment in the hotel and restaurant industry has the opposite effect in the short-run.
in Appendix A.1
provides short-run coefficients across cross-sections of the 10 source countries. Accordingly, there are nine source markets moving towards long-run equilibrium, except the US (where the Cointegrating Equation is positive). Additionally, there is at least one coefficient at one level in the short-run of significance at 0.05 or 0.01 for the variables of interest in each source country, except lnTC in the Korean source market. These coefficients indicate the different short-run roles of investments in tourism infrastructure in attracting international visitors to different source markets. At lag 3, the coefficients of lnTC, lnHR, and lnRF are significant in most source markets. Considering this lag, investment in transport and communications infrastructure has different positive and negative roles for each source market in the short-run. To be more specific, investment in transport and communications infrastructure has an active role in source markets in descending order, Germany, the US, Japan, and China. The source markets with a negative role in ascending order are Australia, the UK, France, Malaysia, and Singapore. As for the role of investment in transport and communications infrastructure, investment in the hotel and restaurant industry also has different positive and negative roles for each source market in the short-run. The source markets where it has an active role in descending order are the US, Germany, Japan, respectively. The source markets where it has a negative role in ascending order are Australia, the UK, France, Malaysia, China, and Singapore, respectively. Meanwhile, investment in recreation facilities plays an active role in all source markets. In descending order, these are China, France, Germany, Japan, Korea, the UK, Australia, and the US, respectively. The coefficients of dummy variables with different signs in source markets indicate the short-run impact of different uncertainties on source markets. Positive effects were found in the short-run in China, Korea, Malaysia, Australia, the UK, Singapore, and France. In contrast, the negative effects were found only in Japan, the US, and Germany.
4.3. Diagnostic Test and Robustness Check
To further consider the reliability and validity of the model estimate, diagnostic tests are considered. There are two critical diagnostic tests for the panel PMG/ARDL method in Eview: coefficient diagnosis and residual diagnostic. However, according to Wooldridge
), based on the asymptotic theory, when there is a sufficient number of observations, it is not necessary to test the normal distribution of the residuals. With 275 observations, this study omits the residual diagnostic and only performs the coefficient diagnostic by coefficient confidence intervals and the Wald test, with the Null Hypothesis that the coefficients are all equal to 0. The results of the diagnostic coefficients are presented in Table 6
provides the values of the coefficients at the 95% and 99% confidence intervals. Accordingly, the maximum and minimum values of lnTC, lnHR and ln RF are all greater than 0. In contrast, the values of Dummy are all less than 0. The Wald test gives significance at 0.01 level for both F and Chi-squared statistics. Therefore, the null hypothesis is rejected and the alternative hypothesis is accepted, meaning that the estimated coefficients in the model are all non-zero, and they are all necessary for the model. This evidence lends support to the reliability and validity of the estimated model.
Next, the robustness check is performed by comparing the estimated results among PMG/ARDL, cointegration regression and OLS for panel data (assuming the cointegration series from the Pedroni test result). In the OLS method, Random Effects Model (REM) is selected from the Pooled OLS model, Fixed Effect Models (FEM) and REM. In the cointegration regression, the FMOLS estimator is chosen because there is a quite large difference in the long-term coefficient of variance in lnVA (Table 2
). The estimated results by FMOLS and OLS methods are detailed in Table A2
in Appendix A.2
. The coefficients estimated by PMG/ARDL, FMOLS and OLS methods are compared in Table 7
According to Table 7
, although the methods produce different estimation results, the signs of the coefficients are similar. To be more detailed, lnTC has quite similar results (bias of no more than 10%), lnHR has a maximum bias of 27.4% and Dummy variable has a bias of no more than 40%. Particularly, lnRF estimated by FMOLS and REM do not reach significance at the 0.05 level. Despite certain differences, it is believed that the results from the PMG/ARDL are more appropriate because of the advantage of PMG/ARDL discussed above, and the cointegration series is still in doubt.
The above findings indicate that investment in tourism infrastructure components positively impacts attracting international tourists to Vietnam. In the long-run, increasing 1% of investment capital in transport and communications infrastructure, the hotel and restaurant industry, and recreation facilities will increase international visitors to Vietnam by 0.7836%, 0.7503%, and 0.4026%, respectively. This indicates that capital investment in transport and communications infrastructure and the hotel and restaurant industry plays a crucial role in attracting international visitors. This evidence lends support to the view that investments in transportation and hotels have played an important role in attracting international tourism, as many earlier studies have found (Khadaroo and Seetanah 2007a
; Prideaux 2000
; Seetanah et al. 2011
). In this study, the role of investment in transport and communications infrastructure (coefficient 0.7836) and investment in the hotel and restaurant industry (coefficient 0.7503) in Vietnam is higher in some areas such as in Mauritius, where the coefficient is found to be 0.36 for investment in transport infrastructure and 0.56 for the investment and hotel industry (Khadaroo and Seetanah 2007b
) or 0.32 for investment capital in transport infrastructure and 0.54 for investment and the hotel industry (Seetanah et al. 2011
); in 26 island economies, the results are 0.064, 0.16, 0.074 and 0.28 for investment in road, air, communications, and the hotel and restaurant industry, respectively (Khadaroo and Seetanah 2007a
); and in 28 countries representing Europe, Asia, America, and Africa, these are 0.13, 0.18, 0.06 and 0.22, respectively, for investment in road, air, port and hotel (Khadaroo and Seetanah 2008
). The impact coefficient of the hotel and restaurant industry in this study is lower than that of the hotel accommodation infrastructure in Singapore, from 0.839 to 0.855 in the study by Lim et al.
). However, it must also be seen that the different roles of the hotel and restaurant industry depend not only on each country, but also on how the variable that represents it is measured. This role is appropriate because Vietnam is a developing country with great tourism potential and scenic beauty. However, the terrain is difficult, and transportation infrastructure and hotel availability are still limited. With the efforts of the government and the community, the transport and communications infrastructure, as well as the hotel and restaurant facilities in Vietnam, have been significantly improved, creating a favorable environment for tourists, and strongly enticing international visitors to Vietnam. The research results also show that the government and private sector investors cannot expect to see a fast Return on Investment. Their investment in transport and communications infrastructure and hotel and restaurant facilities will only be evident in the long-run. This can be explained by the long lead-in time required by infrastructure works and hotel developments. The impact, therefore, takes time to be fully demonstrated. However, it should be noted that transport and communications infrastructure investment attract visitors and develops other areas of the economy and society, including the hotel and restaurant industry and recreation facilities.
Cross-section short-run coefficients show that, in the short term, the role of investment in the hotel and restaurant industry is decreasing, in this order source markets: the US, Germany, Japan, Australia, the UK, France, Malaysia, China, and Singapore. Meanwhile, the order for investment in transport and communications infrastructure is as follows: Germany, the US, Japan, China, Australia, the UK, France, Malaysia, and Singapore, respectively. This is consistent with the idea that inhabitants of developed countries are accustomed to modern, high-quality transport infrastructure and high-quality restaurants and hotels. Consequently, they prefer to find similar infrastructure in other countries. In contrast, tourists from less developed countries tend to be less demanding of these infrastructures.
Research results also show that investment in recreation facilities is also important to attract international arrivals to Vietnam. Although its role in the long-term is not equal to that of the other two areas of tourism infrastructure in this study, it is effective in both the long-run and short-run. Investment in recreation facilities will directly make destinations more attractive. Formica
) states that without attractions, tourism destinations could not exist; attractions are the basis for visitation. These findings are consistent with Vengesayi et al.
), suggesting that attractions are the main reason people visit specific destinations and not others. The role of investment in recreation facilities in attracting international visitors in this study is empirical evidence supporting the tourism infrastructure model of Mandić et al.
). Accordingly, recreational facilities with hotels and other forms of accommodation, spas, and restaurants form the main tourism infrastructure.
Usually, investment in modern amusement parks will require considerable investment capital. In contrast, investment in developing conservation and ecological tourist areas may require a smaller amount of capital if considered per unit area. The above order of roles of investment in recreation facilities in the short-run implies that in general, visitors want to improve recreation facilities in Vietnam, but visitors from China, France, Germany, and Japan require much more improvement than visitors from Korea, the UK, Australia, and the US. This finding is indicative of visitor preferences from source markets.