Openness and economic growth: empirical evidence on the relationship between output, inward FDI, and trade.

Abstract

The relationship between openness and economic growth in developed and developing countries has been of continuing interest in both the theoretical and empirical literature. In this paper, we employ a vector autoregressive (VAR) model and error correction techniques to test for the existence and nature of the causal relationship between output level, inward FDI and exports across a cross-section of both developed and developing countries using data from 1960-2001. Our main objective is to analyze the extent and sources of international linkages between openness and economic performance. The evidence supports bi-directional causality between exports growth and economic growth; the economic growth and FDI relationship has mixed results.

Introduction

The relationship between export growth, foreign direct investment (FDI), and economic growth in both developed and developing countries is a question that continues to be of considerable theoretical and empirical interest. Cross-country trade and capital flows, and interpreting the importance of these activities towards economic growth lie at the heart of the debate on economic development policy since the early literature on import-substitution to the current literature on openness and economic growth.

Recent literature has highlighted the role of both exports and FDI on economic growth. On the one hand, the export led growth (ELG) hypothesis states that exports are the main determinants of overall growth. At the heart of the ELG model are beliefs that (a) the export sector generates positive externalities on non-export sectors in the economy through more efficient management and production techniques (Feder, 1983); (b) export expansion increases productivity by creating scale economies (Helpman and Krugman, 1985; Krugman 1997); (c) exports help to alleviate foreign exchange constraints and thus provide greater access to international markets (Esfahani, 1991). Endogenous growth theory extends this analysis by emphasizing the role of exports on technological innovation and dynamic learning (Romer, 1986; Lucas, 1988; Grossman and Helpmann, 1995; Alisana and Rodrick, 1999).

On the other hand, empirical evidence in the last few decades indicates that FDI flows have been growing at a pace far exceeding the volume of international trade. Between 1975 and 1995, the aggregate stock of FDI rose from 4.5% to 9.7% of world GDP, with sales of foreign affiliates of multinational enterprises substantially exceeding the value of world exports (Barrell and Pain, 1997). The effect of FDI on economic growth appears to have become quite explicit with multinational enterprises acting as the primary vehicle for the international transfer of technology (OECD, 1991). Blomstrom and Persson (1983) and Blomstrom (1986) find that FDI has created significant positive spillover effects on the labor productivity of domestic firms. It is argued that FDI plays a central role in the technological progress of recipient countries through the generation of productivity spillovers (Borensztein, De Gregorio, and Lee, 1998; Lim 2001).

However, empirical work from both the ELG literature and the FDI and growth literature when studied in isolation show mixed results. This is mainly, due to the omission of a relevant mechanism through which openness or the re-structuring of an economy promotes growth. Liberalization, in particular, is expected to increase not only trade but also FDI. If a complementary relationship between FDI and exports exists, then foreign investment may increase the volume of exports in specific and international trade in general. Direct investment may encourage export promotion, import substitution, or greater trade in intermediate inputs, especially between parent and affiliate producers (Goldberg and Klein, 1998). Along the same lines, Blomstrom, Globerman and Kokko (2000) argue that the beneficial impact of FDI is only enhanced in an environment characterized by an open trade and investment regime and macroeconomic stability. In this environment, FDI can play a key role in improving the capacity of the host country to respond to the opportunities offered by global economic integration. In the absence of such an environment, FDI may impede rather than promote growth by enhancing the private rate of return to investment for foreign firms while exerting little impact on social rates of return in the recipient economy (Balasubramanyam, Salisu and Sapsford, 1996).

Early studies supporting the ELG hypothesis such as those by Balassa (1978), Heller and Porter (1978) and Tyler (1981) examined the simple correlation coefficient between export growth and economic growth, and based their conclusions based upon the high degree of correlation between the two variables. Other studies, characterized by Voivades (1973), Feder (1983), Balassa (1985), Ram (1987), Sprout and Weaver (1993) and Ukpolo (1994) find support for ELG based upon growth and output regressions drawn from a growth accounting framework. These studies make the 'a priori' assumption that export growth causes output growth without considering the direction of the causal relationship. A third group of studies has emphasized the issue of causality between export growth and economic growth. In this approach, exemplified by Jung and Marshall (1985), Darrat (1987), and Serletis (1992), the Granger or Sims causality test is applied to growth and export data to test the ELG hypothesis. The causality tests are only valid if the original time series underlying the analysis are cointegrated.

For a complete study on economic growth, the focus has to be not only on ELG but FDI as well. Therefore, the objective of this paper is to investigate the causal relationship between export growth, inward FDI and economic growth (measured as output growth) in developed and developing countries using the cointegration and error-correction models. These techniques, as successfully applied in studies by Serletis (1992), Bahmani-Oskooee and Alse (1993), Dutt and Ghosh (1996), Rahman and Mustafa (1998), Islam (1998), Cuadros, Orts and Alguacil (2001) and Trevino, Daniels, Arbelaez, and Upadhyaya (2002), demonstrate their econometric robustness and their ability to root out spurious relationships.

So far, only a few studies have used this methodology to study the causality relation between export growth, economic growth, and FDI in both developed and developing countries. Given the small number of studies conducted using this methodology, it is expected that this paper will contribute to this expanding body of literature.

The rest of the paper is organized as follows. Section 2 explains the methodology of the cointegration and error-correction models and a description of the data sources. Section 3 contains the empirical results and comparison of our results with previous studies. Finally, Section 4 provides a discussion about the implication of the results and some summary conclusions.

Methodology and Data

Methodology

This paper uses the cointegration and error-correction models, to test the causal relationship between FDI, exports, and economic growth. We start by considering the three-variable vector autoregressive (VAR) model comprised of foreign direct investment (), exports (), and gross domestic product (), all expressed in natural logs. As shown in equation (1), all variables are systematically and endogenously considered at first.

(1) [[FDI.sub.t][EXP.sub.t][GDP.sub.t]] = [A.sub.0] + [A.sub.1] [[FDI.sub.t-1][EXP.sub.t-1] [GDP.sub.t-1]] + [A.sub.2] [[FDI.sub.t-2][EXP.sub.t-2][GDP.sub.t-2]] + .... + [A.sub.s] [[FDI.sub.t-s][EXP.sub.t-s][GDP.sub.t-s]] + [[epsilon].sub.t]

where [A.sub.0] is a vector of constant terms, are all matrices of parameters (i = 1, 2, ..., s), and [[epsilon].sub.t] ~ IN (0,1).

In order to analyze the causal relationship it is necessary to first check whether the variables are stationary. According to Granger (1988), standard tests for causality are valid only if there exits cointegration. Therefore, a necessary precondition to causality testing is to check the cointegrating properties of the variables under consideration. The cointegration and error-correction methodology is briefly outlined below.

Testing for cointegration among the three variables, real FDI, real exports, and real GDP (expressed in logarithmic form), is accomplished in two steps. First, following Engle and Granger (1987), the time series properties of each variable are examined by unit root tests. In this step, it is tested whether FDI, exports, and GDP are integrated of order zero, or in other words, that the three series are stationary. This is accomplished by performing the augmented Dickey-Fuller (ADF) test. The ADF test is based on the regression equation with the inclusion of a constant and a trend of the form

(2) [MATHEMATICAL EXPRESSIONS NOT REPRODUCIBLE IN ASCII]

where [DELTA][X.sub.t] = [X.sub.t] - [X.sub.t-1] and X is the variable under consideration, p is the number of lags in the dependent variable (chosen so as to induce a white noise term), and [[epsilon].sub.t] is the stochastic error term. The stationarity of the variable is tested using the null hypothesis of |[[theta].sub.1]| = 1 against the alternative hypothesis of |[[theta].sub.1]| < 1. If the null hypothesis cannot be rejected, it implies that the time series is non-stationary at that level and therefore it requires taking first or higher order differencing of the level data to establish stationarity. The optimum lag length (p) in the ADF regression is selected using the minimum final prediction error (FPE) criterion developed by Akaike and then the results were confirmed by the Schwarz criterion.