The impact of FDI restrictions on tourism growth

Abstract

This study investigates the impact of FDI regulatory restrictions on tourism growth. Using system generalized method of moments (GMM) estimation of the dynamic panel data model, this study finds that FDI restrictions have significant negative effects on tourism growth, implying that liberalizing FDI policies could promote tourism growth.

Introduction

With the growing importance of foreign direct investment (FDI) in tourism (Endo, 2006; Nunkoo and Seetanah, 2018), a number of empirical studies examine the relationship between FDI and tourism development, and show that the causality runs from the former to the latter (Haley and Haley, 1997; Sanford and Dong, 2000; Tang et al., 2007; Selvanathan et al., 2012; Fauzel, 2020). However, little attention has been paid to the implications of FDI regulatory restrictions on tourism growth. To fill this gap in the literature, this study investigates the impact of FDI restrictions in the tourism industry on its growth.

The extent of tourism FDI restrictions in developed and developing countries has decreased significantly over time. However, many developing countries still maintain some FDI restrictions in the tourism industry. The obvious effects of FDI restrictions are impeding inward FDI, ceteris paribus. Mistura and Roulet (2019) finds that restrictive FDI policy incurs an opportunity costs in terms of forgone FDI inflows. However, the effects of FDI restrictions go beyond discouraging inward FDI; restrictive FDI policy may also result in less competition, diversity, and innovation in the tourism industry. Thus, FDI restrictions in tourism may harm its growth. Using system generalized method of moments (GMM) estimation of the dynamic panel data model, this paper finds that FDI restrictions have significant negative effects on tourism growth, implying that reforms liberalizing FDI restrictions could promote tourism growth.

Empirical methodology

The tourism literature shows that FDI has a positive spillover effect on the productivity of domestic firms by facilitating transfer of technology, skills and knowledge to local tourism enterprises, and upgrading the tourism related facilities and services (Nunkoo and Seetanah, 2018), thereby enhancing total factor productivity of host economies. Based on these spillover effects of FDI, we assume that FDI and FDI policy have an important impact on total factor productivity in tourism. Thus, we include the variable of FDI regulatory restrictions in the empirical model to examine the effect of FDI policy on tourism growth.

The basic empirical model is specified as follows:

$y_{it}-y_{it-\mathit{1}}=\left(\alpha -\mathit{1}\right)y_{it-\mathit{1}}+{\beta FR}_{it}+\gamma x_{it}+{\xi }_i+{\lambda }_t+u_{it}$

(1)where y _it denotes the logarithm of tourism output or value added in country i in year t; FR is FDI restrictiveness; x _it is a vector of control variables such as labor and capital stock; ξ _i is the unobserved country-specific effects; λ _t is year fixed effects; and u _it denotes the error term.

Equation (1) can also be written as follows:

$y_{it}={\alpha y}_{it-\mathit{1}}+{\beta FR}_{it}+{\gamma x}_{it}+{\xi }_i+{\lambda }_t+u_{it}$

(2)

It is well known that in the presence of the country-specific effects ξ _i , the estimate of the coefficient on the lagged dependent variable by the ordinary least squares (OLS) is biased upward because the lagged dependent variable is positively correlated with ξ _i (Blundell and Bond, 1998). Arellano and Bond (1991) propose the first difference GMM estimator to eliminate the country-specific effects:

$∆y_{it}=\alpha ∆y_{it-1}+{\mathrm{\beta }∆FR}_{it}+{\mathrm{\gamma }∆x}_{it}+{∆\lambda }_t+{∆u}_{it}$

(3)

The first difference GMM estimators control for unobserved country-specific effects ξ _i by taking the first differences. However, as demonstrated by Blundell and Bond (1998) and Arellano and Bover (1995), large finite-sample biases can be generated, mainly because lagged levels are only weakly correlated with subsequent first differences. To address this issue, Arellano and Bover (1995) and Blundell and Bond (1998) propose system GMM estimator in dynamic panel data models, which combines two sets of moment conditions, namely, lagged first differences as instruments for the level equations and lagged levels as instruments for the first differences equations. Thus, system GMM estimation corrects for the endogeneity of the lagged dependent variable and other potentially endogenous explanatory variables; it also controls for time-invariant country-specific effects (Blundell and Bond, 1998).

Data on output, value added and labor in the tourism industry are from OECD Inter-Country Input Output Table (ICIO). ¹ Meanwhile, FDI restrictiveness is measured by the FDI Regulatory Restrictiveness Index (OECD). This index gauges the level of restrictiveness of an economy’s statutory measures on FDI (Kalinova et al., 2010). The FDI restrictiveness index measures four types of regulatory restrictions on foreign direct investment: (i) equity restrictions, (ii) screening and prior approval requirements, (iii) restrictions on the employment of foreigners, and (iv) other restrictions on the operation of foreign enterprises. This index takes values between 0 and 1, with 1 being the most restrictive. We use the overall index for the tourism industry obtained by adding the scores for all four types of measures.

The aggregate capital stocks provided by Penn World Table 9.2 are divided into the sectoral initial capital stocks in 2005 using the output shares split method. Moreover, the capital stocks in the tourism industry from 2006 to 2015 are calculated by using the Perpetual Inventory Method, with investment measured by gross capital formation from OECD ICIO. FDI restrictiveness index is provided for 1997, 2003, 2006, and 2010–2018, and the OECD ICIO for 2005–2015. Therefore, the analysis in this paper is limited to 2006, and 2010–2015. ²

Estimation results

Before moving on to estimates of (2), we examine the impact of FDI restrictions on output and value-added level by using the OLS with country and year fixed effects. The estimation results are shown in Table 1. The first two columns show the effects of FDI restrictions on output, whereas the next two columns show the effects on value-added. The capital stock data may be inappropriately constructed, resulting in biased estimates of the other coefficients. Therefore, we estimate the effects of FDI restrictions on output and value-added level by excluding or including capital stocks. Column (1)–(4) show that regardless of including or excluding capital stock, the coefficients of the FDI restrictiveness are negative and statistically significant. These results imply that regulatory restrictions on FDI in the tourism industry are negatively associated with output and value-added level.

Table 1.

Impact of FDI restrictions on output and value-added.

Dep. Variable	ln(output)		ln(value-added)
ln(capital)		2.892** (1.123)		1.156 (1.185)
ln(labor)	0.910*** (0.111)	0.943*** (0.110)	1.024*** (0.115)	1.038*** (0.116)
FDI Restriction	−1.315** (0.550)	−1.458*** (0.547)	−1.412** (0.561)	−1.469** (0.577)
No. observations	323	323	323	323
Adj. R2	0.86	0.94	0.84	0.93
Country fixed effect	Yes	Yes	Yes	Yes
Year fixed effect	Yes	Yes	Yes	Yes

Note: Robust-cluster standard errors are in parentheses. Intercept is included but not reported. ** and *** indicate that the estimated coefficients are statistically significant at 5%, and 1%, respectively.

Next, using system GMM we estimate the impact of FDI restrictions on tourism growth. The estimation results are shown in Table 2. We check the robustness and goodness of fit of GMM estimates using the Arellano-Bond autocorrelation test and the overidentification test. Column (1)–(4) show that the assumption of serially uncorrelated errors is not violated and that the null hypothesis of the “instruments” validity cannot be rejected. The FDI restrictiveness coefficients are negative and statistically significant, implying that a restrictive FDI policy results in lower growth rates. ³

Table 2.

FDI restrictions and tourism growth.

Dep. Variable	ln(output)		ln(value added)
Dep. Variable(-1)	0.730*** (0.012)	0.702*** (0.019)	0.667*** (0.010)	0.556*** (0.017)
ln(capital)		0.052 (0.034)		0.164*** (0.050)
ln(labor)	0.220*** (0.013)	0.206*** (0.017)	0.277*** (0.016)	0.216*** (0.027
FDI Restriction	−0.697*** (0.055)	−0.624*** (0.070)	−1.082*** (0.045)	−0.804*** (0.064)
No. observations	233	233	233	233
Year fixed effect	Yes	Yes	Yes	Yes
AR1 (p-value)	0.01	0.01	0.03	0.04
AR2 (p-value)	0.18	0.18	0.17	0.16
Sargan test (p-value)	0.13	0.15	0.20	0.24

Note: Robust-cluster standard errors are in parentheses. Intercept is included but not reported. ** and *** indicate that the estimated coefficients are statistically significant at 5%, and 1%, respectively. Two year lagged dependent variable and year dummies, and the lagged independent variables are used as instruments for differenced equations. Moreover, the differences of the lagged dependent variables are used as instruments for level equations.

Conclusions

This paper investigates the impact of FDI restrictions on output and value-added level, and growth in the tourism industry. We find that these restrictions have a negative effect on output and value-added level, and growth of the tourism industry. These results suggest that liberalizing the FDI restrictions regime could spur tourism growth.