MERCHANDISE EXPORT DEMAND FUNCTION FOR EGYPT: A PANEL DATA ANALYSIS IBRAHIM, Mohamed Abbas 1 Abstract This study empirically estimates the critical parameters of merchandise export demand function for Egypt by using annual time series-cross section data (1990-2008) and by applying fixed effects model. The empirical results confirm that there exists a significant relationship among the real value of merchandise exports for Egypt and trade partner's real income, relative export price, and trade partner's real exchange rates. Our estimation results show that all variables have a significant relation and its theoretical expected sign (positive for both real income and real exchange rate but negative for relative price), and the elasticities of real income, relative price and real exchange rate are smaller than unity. JEL Classification: C23, F13, F14. Keywords: Export demand, merchandise trade, Fixed Effects Model, Trade policy. 1. Introduction Egypt is classified as a lower-middle income country with a real per capita Gross Domestic Product (constant 2000 US$) in 2008 of $1836. It is one of Middle East and North Africa (MENA) countries that boost a population of 81.5 million, as of 2008. However, over the past decade, Egypt achieved major economic progress. Its real GDP grew at about 5.7 % during 2000-2008; foreign direct investment inflows amounted by 9.49 billion dollar, which represents 5.83% of GDP in 2008 (http://data.worldbank.org/indicator/). This performance was accompanied by increased trade openness, export promotion policies. Exports play an important role in the economic growth and development of many countries. In this respect, measuring income and price elasticities of export demand has received much attention because of the implications on trade policy and balance of payments issues. The higher the foreign income elasticity of export demand, the more significant exports will be as an engine to economic growth. In addition, the higher the export price elasticity, the more competitive is the international market for exports of a particular country, and thus a real devaluation will be more successful in promoting export earnings. As such, an aggregate export demand estimate linking exports with a measure of foreign income and relative prices is important in many conventional trade models. Export demand elasticities are also important for meaningful export forecasts, planning, and policy formulation. Because of the importance of the trade sector for Egypt's economic growth and development and the ensuing implications on the balance of payments, the central aim of this paper is to estimate the determinants of Egypt's merchandise exports during the period 1990-2008. As shown in table 1, real total trade has increased from 38.34 billion dollar to 106.27 billion dollar during the period 1990-2008, with average growth rate about 5.83%. This accompanied with a rising trade/gdp ratio from 59.5% to 72% at the same period. 1 Mohamed Abbas Ibrahim, Assist. Professor of Economics, College of Administrative Sciences and Humanities, Majmaah University, Majmaah, Saudi Arabia.E-mail: m.a.ibrahim@mu.edu.sa and albarsi2020@gmail.com.
This means that, Egypt's economy during this period directed to be more trade openness economy. Table 1:Some key trade indicators of Egypt in constant prices (2005=100) (1990-2008) Value (billion dollar) % of GDP* 1990 2008 1990 2008 Merchandise Exports 5.2 23.8 8 16.1 Services Exports 8.92 22.59 13.8 15.3 Total Exports 14.12 46.39 21.9 31.4 Merchandise Imports 18.55 43.88 28.8 29.7 Services Imports 5.67 16 8.8 10.8 Total Imports 24.22 59.88 37.6 40.5 Total Trade 38.34 106.27 59.5 72 Source:http://data.worldbank.org/indicator/. *GDP at constant prices (2005=100) calculated by the author. Although, as shown in table 2, trade account was suffering from increasingly deficit along the period 1990-2008, the deficit/gdp ratio was decreasing from 15.7 to 9.1 at the same period. This means that, Egypt's economy is directing toward diminishing the deficit, and at the point in the future, this deficit will finish. Also, as it can be seen from table 2, the real value merchandise exports achieved an average growth rate 8.82% during the period 1990-2008. On the other hand, the average growth rate of real merchandise imports that have been achieved during the same period was 4.9%. This means that, the gap between the merchandise exports and imports has diminishing. Thus, the merchandise trade can achieve a surplus in the future. Table 2: Trade account in constant prices (2005=100) (1990-2008) Average Annual % GDP Value (billion dollar) Growth rate (%) 1990 2008 1990-2008 1990 2008 Merchandise Exports 5.2 23.8 8.82 Merchandise Imports 18.55 43.88 4.9 Merchandise Account Balance -13.35-20.08 2.29 20.7 13.6 Services Exports 8.92 22.59 5.3 Services Imports 5.67 16 5.93 Services Account Balance 3.25 6.59 4.01 5 4.5 Trade Account Balance -10.1-13.49 1.62 15.7 9.1 Source: http://data.worldbank.org/indicator/. * calculated by the author. In analyzing the Egypt's real merchandise export performance, the structure of real merchandise exports has to be analyzed. Table 3 presents the structure of real merchandise exports in 1990 and 2008; the average growth rates of real merchandise exports components during the period and its shares of total real merchandise exports. We can observe that ores and metals achieved the highest yearly average growth rate during the period 1990-2008, which enabled it to achieve the highest share of real merchandise exports in 2008. Food and manufacturers exports also achieved high average growth rates during the same period although the manufacture's share had declined. On the other hand, the share of Agricultural raw materials exports is the only sector that witnessed a negative yearly average growth rate that about 1.12%. 108
Ibrahim, M.A. Merchandise Export Demand Function for Egypt: A Panel Data Analysis Table 3: The structure of real value of merchandise exports in constant prices (2005=100) and its yearly average growth rates 1990-2008 Yearly Average % of Merchandise Value (billion dollar) Growth Rate* Merchandise Exports Export* (%) 1990 2008 1990-2008 1990 2008 Manufactures 2.2 8.7 7.94 42.5 36.6 Agricultural raw materials 0.49 0.4-1.12 9.5 1.7 Food 0.51 2.55 9.35 9.8 10.7 Ores and metals 0.46 10.47 18.96 8.9 44 Fuel 1.52 1.66 0.49 29.3 7 Total 5.2 23.8 8.82 100 100 Source: http://data.worldbank.org/indicator/. * calculated by the author. Figure 1 illustrates the changes of the structure of merchandise exports during the period 1990-2008. Fuel and agricultural raw materials exports shares witnessed an obviously decline at the same period although the positive low average growth rates during the period. Figure 1: The structure of merchandise exports 1990-2008 Source: Drawn by the author from table 3. 109
Table 4 presents Egypt s main trading partners of real value of merchandise exports in 1990 and 2008. As it can be seen from table, Italy has the highest share in real merchandise exports in 1990 and 2008. The relative importance of some countries in total real value of merchandise exports has been decreased such as Germany, Singapore and United States of America. In the other hand, the shares of some countries at the same period obviously increased, such as India and Spain. In the other countries, we didn't observe big changes between 1990 and 2008. Figure 2 shows also the changes of real value of merchandise exports to Egypt s main trading partners. Figure 2: Egypt's trade partners shares of real value of merchandise exports (1990-2008) Source: Drawn by the author from table 4. 110
Ibrahim, M.A. Merchandise Export Demand Function for Egypt: A Panel Data Analysis Table 4: Egypt's real merchandise exports and Yearly average growth rates by country in constant prices (2005=100) (1990-2008) To Exports (Million Dollar) Yearly Average Growth Rate (%) Relative Importance* (%) 1990 2008 1990-2008 1990 2008 France 154.52 713.74 8.87 2.97 3 Germany 214.2 402.23 3.56 4.12 1.69 Greece 58 295.37 9.46 1.12 1.24 India 2.81 1496.64 41.73 0.05 6.29 Italy 471.36 2449.05 9.59 9.07 10.3 Japan 104.88 793.3 11.9 2.02 3.33 Netherlands 244.27 1351.5 9.97 4.7 5.68 Saudi Arabia 114.45 1133.82 13.59 2.2 4.77 Singapore 177.3 259.35 2.14 3.41 1.09 Spain 53.94 1342.4 19.55 1.04 5.64 Syria 17.46 505.97 20.57 0.34 2.13 Turkey 24 708.12 20.69 0.46 2.98 United 115.43 842.86 11.68 2.22 3.54 Kingdom United States 331.16 1161.034 2.14 6.37 4.88 Study Group 2083.78 13455.38 10.92 40.1 56.57 Other 3113.203 10331.08 6.89 59.9 43.43 Egypt 5196.983 23786.46 8.82 100 100 (Total)** Source: http://comtrade.un.org/db/ * calculated by the author. ** http://data.worldbank.org/indicator/. 2. Recent empirical studies A large volume of literature exists on the study of export demand functions for developing and developed countries. However, from the empirical literature surveyed, no recent study was found that estimates the determinants of the merchandise exports demand function for Egypt. Abdul Samad, A. (2009) investigated the economic determinants of Medium Density Fiberboard (MDF) exports with the export price, exchange rate and average world GDP for a panel of twenty eight countries over the sample period 1996-2005. The twenty eight countries considered in this study are known as exporting countries on MDF product. The results indicate that the economics determinants of MDF export are export prices and domestic exchange rate. Of which the export price are most significant impact on export demand. However, the income variable in this study shown not significantly influences the export of MDF product. Husein, J. (2008) estimated the critical parameters of the export demand function for Jordan during the period (1970-2004). The empirical results obtained show that foreign income is a significant variable in explaining the demand for exports and that foreign income elasticity is much larger than unity and Export demand for Jordanian goods is unitary price elastic. Khedhiri and Bouazizi (2007) estimated demand elasticities for Tunisian exports to the major European trading partners by using panel data, they found a significant relationship between real exchange rate index, foreign income and Tunisian exports. Khedhiri and Bouazizi found foreign income to be elastic (2.95), while real exchange 111
rate to be inelastic (-0.162). They conclude that devaluation of Tunisian currency will slightly improve export demand but may be very costly since it will have a depressing effect on domestic output. Reinhart (1995) found that relative prices are a significant determinant of demand for exports in developing countries; however, the elasticity tends to be low, suggesting that large relative price swings are required to have an appreciable impact on trade patterns. Senhadji and Montenegro (1998, 1999) obtained results on relative price elasticities of exports demand similar to those of Reinhart (1995). Guisan, M. and Cancelo, M. (2002) considered supply side determinants in addition to the traditional demand factors when estimating the determinants of exports for 25 OECD countries. Using data for the period 1960-1997, their econometric model included supply side variables such as domestic GDP, domestic private consumption and a measurement of human capital (proxied by educational levels of the population) in addition to foreign income and relative export price. Their results, reveal that both external demand (foreign income), domestic GDP, and average years of schooling have a positive and significant impact on exports, while both consumption and relative prices have a significant and negative impact on exports. 3. The model and the methods To estimate the major trade partner's demand for Egypt's merchandise exports, we will use the traditional specification of merchandise export demand function that relates the value of a country s merchandise exports to trade partners buying power (foreign income represented in gross domestic product of trade partners), the ratio of the price of its exports to the trade partner's export price and trade partner's real exchange rate. This study hypothesized that the explanatory variables have a linear relationship with the merchandise exports of Egypt. To reach the objective of the study, we will use the balance panel data. There are several advantages in using this type of data. First, the use of annual data avoids the problems due to seasonality. Second, by using the different origin countries as observational units, an increase in the range of variation of the variables is considered. Finally, the utilization of a pooled time-series/crosssectional data set enables us to have more degrees of freedom than, and reduce the problem of multicollinearity, hence improving the accuracy of parameter estimates (Garin-Munoz and Martin Montero, 2007; Hsiao, 2003). We estimate a fixed effects model to explain the demand for merchandise exports by using data from fourteen countries from major destination countries for Egypt's merchandise exports which represented in 2008 about 56.57 % of total Egypt's real merchandise exports (as it shown in table 5). These fourteen countries are: France, Germany, Greece, India, Italy, Japan, Netherlands, Saudi Arabia, Singapore, Spain, Syria, United Kingdom and United States of America. The data set pointed out the period between 1990 and 2008 (t=1990,, 2008). Accordingly, the estimated demand function for real merchandise exports in Egypt involves the following variables; RVEX i,t = ƒ( RGDP i,t,tcpi i,t, REX i,t ) (1) where RVEXi,t is the real value of merchandise exports from Egypt to country i during year t in constant prices (2005=100), RGDPi,t is the real gross domestic product in each 112
Ibrahim, M.A. Merchandise Export Demand Function for Egypt: A Panel Data Analysis of the importing country i during year t in constant prices (2005=100); The relative price variable TCPI i,t is given by the indicative ratio of Egypt's consumer price index (2005=100) to consumer price index (2005=100) of the importing country i during year t and REX i,t is the real exchange rate of importing country i in the year t. TCPI i,t = (CPIEgypt,t/ CPI i,t ) (2) Where, CPIEgypt,t is the consumer price index of Egypt in year t, CPIi,t is the consumer price index of importing country i in year t. There are several functional forms that can be used to determine the demand for merchandise exports. In this study, the model to be estimated would be: log RVEX i,t = β0 + β1 log RGDP i,t+ β2 log TCPI i,t+ β3 log REX i, t + ξi,t. (3) We expect β1 to be positive, an indication that as trade partner real income rises, their demand for goods and services increases, including those of Egypt; β2 to be negative reflecting the fact that as relative price level of Egypt to trade partner rises, as Egyptian merchandise exports become more expensive than trade partner's domestic merchandise, which make the foreign consumer transfers to foreign goods. Moreover, we expect β3 to be positive, an indication that as trade partner real exchange rate rises as Egyptian merchandise become cheaper than trade partner domestic merchandise. The loglinear form is chosen, since it is found to be the most appropriate function form for demand functions in many empirical studies (Khan and Ross (1977), Boylan et. al.(1980), Emran and Shilpi(1996). It also has the added advantage of reducing hetroskedasticity (Maddala 1992). 4. Data and variables This study will use the annual data from 1990 to 2008 for the 14 countries as stated earlier in Table 4. The real value of merchandise exports to these countries was proxied by dividing Exports at current prices in dollar (obtained from COMTRADE (http://comtrade.un.org/db/) by the Consumer price index of USA (2005=100) that obtained from (http://data.worldbank.org/indicator/ )). The real income at constant prices (2005=100) for each country is calculated by dividing gross domestic product (in dollars) of these countries at current prices by the Consumer price index of USA (2005=100), both of them obtained from World Bank Development Indicator (http://data.worldbank.org/indicator/ ). The trade partner's exchange rate is taken from international macroeconomic data set published by the Economic Research Service (ERS) in the US Department of Agriculture (USDA). Since the exchange rate is expressed as domestic exporting countries' currencies per US dollar, a decline in exchange rate indicates a real depreciation of the US dollar. We selected the domestic currencies per US dollar exchange rate because typically the merchandise exports and GDP have quoted in US dollar. Finally, it is noted that, since all variables are converted to natural logarithms, the estimated coefficients can interpreted as elasticities. 113
5. Empirical results For the estimation of equation (3) we have used E-views econometric software to obtain the fixed effects panel estimates of the model by SUR Method. Table 5 shows the results from the estimation. The results of Table 6 show that the model performs satisfactorily. In Table 5, we see the results with the fixed effects estimator. The explanatory power is very high (Adjusted R 2 =0.999). The explanatory variables are significant at 1% level with expected sign (Log(RGDP), Log(TCPI) and Log(REX). Table 5: Estimation results for the fixed effects model (1990-2008) Dependent Variable: LOG(RVEX_?) Method: Pooled EGLS (Cross-section SUR). Sample (adjusted): 1990 2008 Included observations: 19 after adjustments. Cross-sections included: 14 Total pool (balanced) observations: 266 Convergence not achieved after 500 weight iterations Variable Coefficient Std. Error t-statistic Prob. C -5.399146 0.688117-7.846267 0.0000 LOG(RGDP_?) 0.771933 0.050617 15.25060 0.0000 LOG(TCPI_?) -0.658806 0.007107-92.70014 0.0000 LOG(REX_?) 0.337210 0.032941 10.23676 0.0000 Fixed Effects (Cross SA--C 0.659331 NEI--C 0.943852 USA--C -0.862150 ITA--C 0.904245 UK--C -0.531012 GER--C -0.946357 GRE--C 0.908763 JAP--C -3.639265 SPA--C 0.274618 SIN--C 1.112446 FR--C -0.380623 IND--C -1.180753 SYR--C 1.128996 TUR--C 1.607910 Effects Specification. Cross-section fixed (dummy variables). Weighted Statistics R-squared 0.999661 Mean dependent var 7.604396 Adjusted R-squared 0.999639 S.D. dependent var 54.38981 S.E. of regression 1.033572 Akaike info criterion 0.038796 Sum squared resid 265.9997 Schwarz criterion 0.267817 Log likelihood 11.84012 F-statistic 45849.21 Durbin-Watson stat 1.962138 Prob(F-statistic) 0.000000 Unweighted Statistics R-squared 0.445601 Mean dependent var 5.262028 Sum squared resid 132.0914 Durbin-Watson stat 0.493685 114
Ibrahim, M.A. Merchandise Export Demand Function for Egypt: A Panel Data Analysis From estimation results as it shown in table 4, trade partner's real income, relative price, and trade partner's real exchange rate elasticities carry their theoretical expected signs. The elasticity coefficient of trade partner's real income (Log(RGDP)) is positive, significant and smaller than unity. This means that changes in trade partner's real income, ceteris paribus, will lead to nearly small changes in real merchandise export earnings. For instance, a one percent increase in foreign real income induces a 0.77 percent increase in real export earnings, all else unchanged. The estimated relative price elasticity is negative, significant, and is smaller than unity (-0.66), which means that changes in relative price, ceteris paribus, will lead to nearly small changes in real export earnings. The elasticity coefficient of trade partner's real exchange rate is positive and is significant and smaller than unity. This implies that as trade partner's real exchange rate increase, Egypt will capture a relatively small portion of real value of merchandise exports. For instance, a one percent increase in trade partner real exchange rate induces a 0.34 percent increase in real export earnings, all else unchanged. References Abdul Samad, A. (2009), "Determinants of MDF Exports: A Panel Data Analysis", International Business Research, vol. 2, no. 3, July, Canadian Center of Science and Education, Toronto, Canada. Arize, A. (2001) Traditional Export Demand Relation and Parameter Instability an Empirical Investigation, Journal of Economic Studies, 28, 378-96. Bahmani-Oskoee, M. and Niroomand, F. (1998) Long-run PiceElasticities and the Marshall-Lerner Condition Revisited, Economics Letters, Vol. 61, 101-109. Boylanet. al. (1980),"The Functional Form of TheAggregate Demand Equation", Journal of International Economics, vol. 10, 561-66. Emran, M.,Shilpi, P. (1996), "Foreign Exchange Rationing and The Aggregate Demand Function", Economics Letters, Vol. 51, 315-22. Garín-Muñoz T. and L.F. Montero-Martín (2007),"Tourism in the Balearic Islands: a Dynamic Model for International Demand Using Panel Data",Tourism Management, 28(5), 851-865. Guisan, M.C., and Cancelo, M.T. (2002), Econometric Models of Foreign Trade in OECD Countries, Applied Econometrics and International Development, Vol. 2-2, 65-81. Hsiao, C. (2003),Analysis of Panel Data, Second edition, Cambridge, Cambridge University Press, U.K. 115
Husein, J. (2008), "Traditional Export Demand Relation: A Cointegration and Parameter Consistency Analysis", International Journal of Applied Econometrics and Quantitative Studies, Vol. 5-2, USC, Spain. Khan, M. S. and Ross,K. (1977),"The Functional Form of TheAggregate Demand Equation",Journal of International Economics, Vol. 7, 149-60. Khedhiri, S. &Bouazizi, T. (2007), "An Empirical Analysis Of The Demand Elasticity For Tunisian Exports," Applied Econometrics and International Development, Euro- American Association of Economic Development, vol. 7(1). Maddala, G. S. (1992),Introduction to Econometrics, 2nd edition, Maqcmillan Press, New York. Reinhart, C. M. 1995, Devaluation, Relative Prices, and International Trade: Evidence from Developing Countries,IMF Staff Papers, 42, no. 2: 290 312. Senhadji, A. and Montenegro, Claudio E. (1998), Time Series Analysis of Export Demand Equations: A Cross-Country Analysis,IMF Working Paper,No. WP/98/149, Washington D.C., International Monetary Fund. (1999),"Time Series Analysis of Export Demand Equations: A Cross-Country Analysis,IMF Staff Papers, 46, no. 3: 259 73. United Nations,United Nations Commodity Trade Statistics Database (COMTRADE), (http://comtrade.un.org/db/). US Department of Agriculture (USDA), Economic Research Service (ERS), Washington, USA. World Bank, World Bank Indicator, http://data.worldbank.org/indicator/. http://www.usc.es/economet/eaat.htmjournal Publisher by the EAAEDS: 116