Import demand in heterogeneous panel setting

Abstract

To study the elasticities of import demand function, we build a heterogeneous panel with data of 40 countries and use panel unit root tests (Im, Pesaran and Shin, 1997) and panel cointegration tests (Pedroni, 2004). We test our model with two previously used activity variables: GDP and GDP minus Export for a performance comparison.To estimate our elasticities, we make use of two modified panel version of FMOLS and DOLS developed by Pedroni (1996Pedroni ( , 2000Pedroni ( , 2001. Our tests prove that GDP outperforms GDP minus Exports as an activity variable in the cointegration context. I INTRODUCTIONMany attempts have been made to estimate the Import Demand Function (IDF, hereafter) in different countries. The importance of this applied exercise stems from the effect of foreign trade and trade policy on the local economy. Also, devaluation in many countries is based on the negative effect it has on real exchange rate, which in turn discourages imports and improves trade balance. Thus, the value of import elasticity with respect to major macroeconomic factors reveals the degree to which the local economy is subject to foreign countries' disturbances and the effectiveness of a devaluation policy.Among the earliest papers in this field of applied research is Thursby and Thursby (1984) where the authors estimated different specifications of the IDF for five developed countries. They concluded that including lagged values of the dependent variables improves the model specification. Goldstein and Khan (1985) presented a detailed literature review on the Import and Export functions, their specifications, estimation methodologies and the problems arising from the choice of variables and simultaneity. Both papers however, are dated before the development of the cointegration literature. Cointegration technique is important in the case of IDF because of the presence of unit root in the related data series. Clarida (1994) used these econometric advances to estimate the US import elasticity of non-durable goods. Instead of an ad-hoc model, he estimated an IDF based on a simple rational expectations general equilibrium model. To tackle the problem of simultaneity, he applied a technique developed by Phillips and Loretan (1991), which consists of 4 including a lagged value of the deviation from the long run relationship. His results showed that US income and price elasticities of imports are 2.20 and -0.94 respectively. Reinhart (1995) estimated price and income elasticities of imports for 12 developing countries with 25 observations each. Her model suggested that the right scale variable is permanent income or a measure of wealth for which she used GDP as a proxy. She applied a dynamic estimator proposed by Stock and Watson (1990). Her estimates proved to be sensible. Moreover, she found evidence of Houthakker and Magee (1969) results; that is, the developing countries' income elasticity of imports is lower than developed countries' (which in her model are equal to the exports of the developing countries). Howev...

Untitled

“…This magnitude is consistent with the findings for developed countries of Hong (1999) and Marquez (2002). In addition, Harb (2005) records similar values that are at odds with the standard theory. As far as the relative price is concerned, the estimated coefficient is negative but different from zero only at the 10 level.…”

“…Several years later, Goldstein and Khan (1985) publish a comprehensive survey on import elasticities, but all previous findings have been dismissed mostly because they derive from possibly spurious regression, as cointegration analysis was not yet introduced. Nonetheless, more recent long-run estimates are still in the range of those reported by previous authors (Clarida 1994;Reinhardt 1995;Senhadji 1998;Marquez 2002;Harb 2005). This literature addresses different specifications and employs heterogeneous econometric techniques, adding several explanatory variables apart from gdp and import prices.…”

“…As reported above, empirical findings are often very different. This may be due to either the over simplified theoretical model or the noise introduced by the proxies (Harb 2005).…”

Alternative estimators of cointegrating parameters in models with nonstationary data: an application to US export demand

“…Examples of its use can be found in Masih and Masih (1996), Agarwal (2001), Konno and Fukushige (2003), Narayan (2004) and Narayan and Narayan (2005). Recently it has also been used extensively in the estimation of equilibrium parameters in dynamic panel models as in Auteri and Constantini (2004) and Harb (2005Harb ( , 2007.…”

Türkiye nin İhracata Dayalı Büyümesinin Ekonometrik Analizi

“…Geliri temsil etmek üzere genellikle reel GSYH kullanılmaktadır (Harb (2005), Chang vd. (2005, Konno ve Fukushige (2002), Hye (2008), Camarero ve Tamarit (2004) …”