DEM_2019_5to27 © 2019 Nicolaus Copernicus University. All rights reserved. http://www.dem.umk.pl/dem D Y N A M I C E C O N O M E T R I C M O D E L S DOI: http://dx.doi.org/10.12775/DEM.2019.001 Vol. 19 (2019) 5−27 Submitted March 27, 2019 ISSN (online) 2450-7067 Accepted September 22, 2019 ISSN (print) 1234-3862 Sharif Hossain, Kanon Kumar Sen*, Thasinul Abedin, Muhammad Shafiur Rahman Chowduhury Revisiting the Import Demand Function: A Comparative Analysis A b s t r a c t. This study attempts to revisit import demand function across three panels of fron- tier, emerging, and developed economy from 1980 to 2016. Long-run relationship exists among import demand, relative price, exchange rate, and real GDP in economy. Due to increase in real GDP, import demand responds positively across economies. It responds in same direction in short-run in frontier and emerging economies with relative price unlike that of long-run in same economies. However, it responds in same direction with relative price in developed economy. It moves in opposite direction with respect to movement in exchange rate of frontier economy unlike that of developed economy. Next, the behavior of import demand in short-run due to change in exchange rate varies from that of long-run in emerging economy. This study will help to predict the dynamics of import due to change in income level, relative price, and exchange rate at national and international level. Keywords: Exchange Rate; Import Demand Function; Real GDP; Relative Price JEL Classification: C01; C23; F13; F31 * Correspondence to: Kanon Kumar Sen, Lecturer, Faculty of Business Studies, Bangla- desh University of Professionals, Dhaka-1216, Bangladesh, e-mail: kanon- kumardu@gmail.com; Dr. Md. Sharif Hossain, Professor of Econometrics, Faculty of Busi- ness Studies, University of Dhaka, Dhaka-1000, Bangladesh, email: sharif_hossain0465@ya- hoo.com; Md. Thasinul Abedin, Assistant professor, Department of Accounting, University of Chittagong, Bangladesh email: abedin@cu.ac.bd; Muhammad Shafiur Rahman Chowduhury, Assistant professor, Department of Accounting, University of Chittagong, Bangladesh email: shafiur_eco_cu@yahoo.com. S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 6 Introduction Open economy with confined production base due to the trend of globali- zation stimulates foreign trade, resulting high interdependent linkage among different economies. The economies of scale in a particular product is given priority to few countries to become dominant in the international market, in- creasing the import level of that particular product around the world. Besides, the World Trade Organization (WTO) paves the way to shrink the trade bar- riers and implement rules and regulations among, especially, frontier and emerging economies, rapidly raising the import level. The estimation of import demand function has drawn a considerable at- tention among the researchers and policy makers due to the movement in im- port demand across different countries with respect to the movement in a few key macroeconomic variables namely exchange rate, relative prices, and in- come level etc. Moreover, macroeconomic policies designed by the policy- makers affect the trend of import due to the development in foreign trade in frontier and emerging economies. The elasticity of import with respect to in- come and relative price is convenient for the policy makes to design commer- cial policies and the elasticity of import with respect to exchange rate is es- sential to ensure good position in international trade. In this regard, two nota- ble studies conducted by Orcutt (1950) and Kreinin (1967) can be mentioned. They attempted to explain import elasticity with respect to changes in price level for a number of countries. Later econometric analysis revealed that apart from the price level, income level plays a role in explaining the import demand especially in developing economy. Houthakker and Magee (1969) validated the situation in the way that improvement or decline in trade balance in some countries was affected by differences in income elasticities of their demand for imports. With the collapse of Bretton Woods Systems, fixed exchange rate regime has been disappeared. During the existence of Bretton Woods Systems, policy makers and researchers were not familiar with the changes in imports due to movement in exchange rate. It is however believed that the extent and timing of foreign exchange rate effect on import reflected the reaction of rel- ative prices. A few notable studies which tried to find out the impact of relative prices and exchange rate on imports are Junz and Rhomberg (1973), Wilson and Takacs (1979), Warner and Kreinin (1983), Bahmani-Oskooee (1986), Bahmani-Oskooee and Niroomand (1998), and Bahmani-Oskooee and Kara (2003, 2005). The main objective of this paper is to estimate the import demand function across three panels of frontier, emerging, and developed economies. More specifically this paper attempts to reinvestigate how the import changes with Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 7 respect to changes in gross income level (gross domestic product), relative price, and average exchange rate across frontier, emerging, and developed economies. Finally, this study will recommend some policies to policy makers and economists at national and international level to forecast the dynamics of important demand due to change in income level, relative price, exchange rate for all economies. In the following part, this study starts with literature review, after which a few models for estimating import demand function are briefly explained. Next, this study specifies the data and model. Subsequently, this study pre- sents econometric methodology, estimated results, and discussion. Finally, this study concludes the findings with policy implications. 1. Literature Review A growing curve of import, pushed by the air of globalization, is stimu- lated by the desire to gain rapid speed of economic growth with maximum utilities from foreign trade. Hence, considering the significance of interna- tional trade, the large number of empirical studies are found investigating the import demand function in numerous economies. For example, investigating the aggregate import demand function of Pakistan, Rehman (2007) revealed that in the long run movements in import price and real income significantly affect import. Studying in detailed the import demand function of developing countries, Rana (1983) deduced that the increase in exchange rate had a sig- nificant negative impact on import demand. Inspecting the import demand of Thailand, Sinha (1997) revealed that import depends on the import price, do- mestic price, and gross domestic product. Estimating import demand of Italy over the period 1970 to 1986, Giovannetti (1989) revealed that the compo- nents of different expenditures have a significant impact on the import de- mand. Erlat and Erlat (1991) studied that international reserve plays the key role to explain import demand. However, relative price has no significant im- pact on import demand. Kotan and Saygili (1999) found that domestic income is the most significant variable in explaining import demand. Mwega (1993) found insignificant impact of relative prices and domestic income on import demand by using annual data for the period 1964–91 for Kenya. Investigating the import demand of Pakistan for the period 1960–1999, Afzal (2001) iden- tified that relative price has insignificant negative impact on import demand unlike domestic income. Anyemedu (1995) argued that policies to liberalize trade will increase the aggregate import demand. Egwaikhide (1999) found that previous import had insignificant impact on current import and lagged foreign exchange rate, relative price, and real income have significant impact S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 8 on import in Nigeria. Therefore, he deduced that foreign exchange rate, rela- tive price, and real income play important role in influencing import behavior of Nigeria. Estimating import demand function for the UK, Abbott and Seddighi (1996) revealed that followed by investment and export expendi- tures, consumption expenditure has the highest impact on import demand. Es- timating South Korea’s import demand over the 1963–1998 period, Min et al. (2002) found long-run significant elastic impact of final consumption ex- penditure on import demand, significant inelastic impact of export expenditure on import demand, the insignificant negative impact of investment expendi- ture on import demand, and the significant negative impact of relative price on import demand. Dutta and Ahmed (2004) estimated the Indian import de- mand function and revealed that real GDP explains import demand to a sig- nificant extent and import demand is less sensitive to changes in import price. Tang (2004) revealed through estimation of import demand function of ASEAN 5 that exchange rate policy such as devaluation can be used to im- prove trade balance in Malaysia, Singapore, Philipine, and Thailand but not in Indonesia. Islam and Hassan (2004) revealed that the impact of income on import is positive and significant where the impact of relative price is signifi- cantly negative in Bangladesh import demand function. Hye and Mashkoor (2010) found the significant positive impact of income and significant nega- tive impact of relative price on import in Bangladesh. Ghorbani and Motallebi (2009) found significant positive impact of gross domestic income on import. They reported that import is elastic with respect to income. Yin and Hamori (2011) revealed that price plays a significant role in explaining the import de- mand unlike the exchange rate in the long run. Assessing import demand func- tion of Malaysia for 1970 to 1998, Mohammed and Tang (2000) revealed a negative inelastic relationship between relative price and import demand. They also revealed that all expenditures have an inelastic effect on import de- mand in the long-run. However, investment expenditure is highly correlated with import demand. Assessing the import demand function of South Fiji, Na- rayan and Narayan (2005) found that consumption expenditure, investment expenditure, and exports have positive impact on import demand unlike that of relative price on import. Estimating China import demand function, Tang (2003) found that export expenditure has the highest correlation with import. Investigating US data from 1948 to 2007, Katsimi and Moutous (2011) found evidence of a long-run cointegration of a standard import demand function including income inequality (see also, Adam et al., 2008). Analyzing Ja- maica’s import demand function with US and UK, Hibbert et al. (2012) found that real GDP, relative price, foreign reserves, and volatility of exchange rate Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 9 have positive impact on the import demand. Ahad et al. (2017) found that fi- nancial development increases the import demand unlike economic growth and relative price in Pakistan. A negative and significant relationship between exchange rate and import but a positive and significant relationship between the value of merchandise import and gross domestic product have been re- vealed by Ibrahim (2017) (see also Ibrahim, 2015). Arize and Malindretos (2012) found the significant positive impact of foreign exchange reserve and domestic income on import demand both in the short run and in the long run. Mishra and Mohanty (2017) found the evidence of negative impact of relative price on import demand unlike increase of domestic income and foreign ex- change reserve in India in the long run. This paper attempts to estimate whether the import demand function varies across frontier, emerging, and developed economies by using Bahmani-Os- kooee (1986) model. Even of this issue is studied extensively, it requires in depth analysis due to variation in results across different countries and econ- omies. Therefore, the principle objective is to revisit the import demand func- tion across three panels of frontier, emerging, and developed economies by using econometric tools and techniques. Due to use of different econometric tools and variation of sample size in a few previous studies, the conclusion on the results of import demand function are very mixed. This study will also uncover reasons of variation in import demand function across frontier, emerging, and developed economies. 2. Models of Import Demand Function A few models are suggested by the researchers to estimate the import de- mand function. For example, Khan (1974) suggested the following model: (1) The logarithmic transformation of equation (1) is given below: (2) Where is the value of merchandise import of country i imported from country j. is the real GNP (Gross National Product) of country i different from country j. is the ratio of world price to domestic price. and are 1 2 1 ijt ijt ijt ijtX A P Y ewq q= 0 1 2ln ln lnijt ijt ijt ijtX P Yq q q w= + + + ijtX ijtY ijtP 1q 2q S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 10 the elasticities of import with respect to relative price and real GNP. de- notes the random error term and denotes time period. Warner and Kreinin (1983) suggested the following model: (3) The logarithmic transformation of equation (3) is given below: (4) Where is the value of merchandise import of country i imported from country j. is the price level of goods in foreign country needs to be im- ported in country i different from country j, denotes price level in domes- tic country i different from country j, and denotes real GNP (Gross Na- tional Product) of country i different from country j. and denote elastic- ities of import with respect to price level in foreign country and domestic country respectively. denotes elasticity of import with respect to real GNP of country i different from country j. denotes random error term and t de- notes time period. Bahmani-Oskooee (1986) suggested the following model: (5) The logarithmic transformation of equation (5) is given below: (6) denotes the value of merchandise imports to country i from j, denotes the real GDP (Gross Domestic Product) of country i different from country j, denotes relative price measured as the ratio of foreign over domestic price index of country i different from country j, and denotes the average ex- change rate of country i different from country j. , , denote import elasticity with respect to real GDP, relative price, and exchange rate (DC/USD). denotes random error term and t denotes time period. Bah- mani-Oskooee and Niroomand (1988) suggested the following model specifi- cation for a study period of 1960–1992: ijtw t 31 2 , 1( ) ( ) ijtF D ij t ijt ijt ijtX A P P Y eudd d= 0 1 2 3ln ln ln lnF D ijt ijt ijt ijt ijtX P P Yd d d d u= + + + + ijtX F ijtP D ijtP ijtY 1d 2d 3d ijtu 31 2 0 ijt ijt ijt ijt ijtX A Y P E ehll l= 0 1 2 3ln ln ln lnijt ijt ijt ijt ijtX Y P El l l l h= + + + + ijtX ijtY ijtP ijtE 1l 2l 3l ijth Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 11 (7) The logarithmic transformation of equation (7) is given below: (8) denotes the value of merchandise imports to country i from j, denotes the real GDP of country i different from country j, denotes relative price measured as the ratio of foreign over domestic price index of country i differ- ent from country j. and denote import elasticity with respect to real GDP and relative prices. denotes random error term and t denotes time period. 3. Data, Variables, and Descriptive Statistics This study uses data of three panels of 8 frontier countries, 8 emerging countries, and 10 developed countries from 1980 to 2016. Economies have been classified as per the annual market classification by MSCI1 (MSCI, 2019). Frontier economy includes Bangladesh, Pakistan, Kenya, Nigeria, Jor- dan, Trinidad, Sri Lanka, and Cameroon. Emerging economy includes China, India, South Africa, Mexico, Thailand, Malaysia, South Korea, and Philip- pines. Developed economy includes United States, United Kingdom, Canada, Japan, Switzerland, Denmark, Australia, New Zeeland, Sweden, and Norway. However, Pakistan, included in frontier market up to May 2017, currently be- longs to emerging market (MSCI, 2019). Selected frontier countries represents African, Middle East, and Asian countries. Next, selected emerging countries represent Americas, African, and Asian countries. Finally, developed coun- tries represent Americas, United Kingdom and European, and pacific Coun- tries. The characteristics of economies have been provided in Table 1 and the definitions of key variables have been provided in Table 2, and descriptive statistics of the data set have been provided in Table 3. 1 MSCI Inc., which formerly Morgan Stanely Capital International and MSCI Barra, is an global provider of equity, fixed income, hedge fund stock market indexes, and muti-asset portfolio analysis tools. It prepares MSCI BRIC, MSCI World and MSCI EAFE indexes. 1 2 2 ijt ijt ijt ijtX A Y P exg g= 0 1 2ln ln lnijt ijt ijt ijtX Y Pg g g x= + + + ijtX ijtY ijtP 1g 2g ijtx S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 12 Table 1. Characteristics of Economies Economies Definition Frontier Economy (Characteristics) § Earlier stages of development. § Lot of scope of improvement in GDP. § Very young population. § Low level of urbanization. § Embracing economic growth. § Growing middle class as income rises. § Low level of internal and foreign debt. § Government pursuing policies of liberalization and reform. § High political risk. § Per capita income is less than $4,035. Emerging Economy (Characteristics) § Investing in more productive capacity. § Moving away from traditional economy that have relied on agriculture and the export of raw materials. § Rapidly industrializing and adopting a free market or mixed economy. § Lower per capita income than the average world per capita income (as per the World Bank). § Rapid economic growth. § Per capita income is greater than $4,035 and less than $12,236 as per the World Bank. Developed Economy (Characteristics) § High level of security. § High per capita income is above $12, 236 as per the World Bank. § High human development measured by Human Development Index (HDI) (greater than 0.8). § High level of industrialization. § Supremacy of capital. § Large scale production. § Human efforts are directed towards earning more and more income. Table 2. Definition of Key Variables Variables Definition Data Sources Import de- mand (IMP) The import value of each country of different economies in USD. World Bank Development Indicators Real GDP (GDP) The real value of final goods and services in particular financial year adjusted for inflation and deflation. World Bank Development Indicators and International Monetary Fund Relative prices (RP) World Price Index /Consumer Price Index. World Bank Development Indicators, and UNTACD statistics Exchange rate (ER) Domestic Currency (DC) of each country per USD (DC/USD). World Bank Development Indicators, and UNTACD statistics Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 13 Table 3. Descriptive Statistics Frontier Economy Variables Minimum Maximum Mean Standard Deviaton JB Statistic Import (mn USD) 1,420 88,378 10,708 13,292 1139.07*** (0.00) Real GDP (mn USD) 4,601 207,172 38,284 40,065 267.46*** (0.00) Relative Price 0.63 50.58 3.34 7.35 8377.21*** (0.00) Exchange Rate 0.3 733.03 94.21 151.81 569.31*** (0.00) Emerging Economy Variables Minimum Maximum Mean Standard Deviaton JB Statistic Import (mn USD) 11,010 2,261,248 181,892 309,789 6772.05*** (0.00) Real GDP (mn USD) 45,773 8,909,478 741,768 1,244,385 5381.71*** (0.00) Relative Price 0.5 266.2 4.4 23.68 97342.64*** (0.00) Exchange Rate 0.02 1401.44 134.7 320.74 565.06*** (0.00) Developed Economy Variables Minimum Maximum Mean Standard Deviaton JB Statistic Import (mn USD) 702 2,814,841 329,657 510,871 1634.48*** (0.00) Real GDP (mn USD) 982 16,597,446 2,147,298 3,571,335 732.77*** (0.00) Relative Price 0.46 11.13 1.47 1.01 17381*** (0.00) Exchange Rate 0.43 249.08 16.71 42.5 2798.82*** (0.00) Note: ***Significant at 1% level, **Significant at 5% level, *Significant at 10% level. From descriptive statistics, we have observed that developed economy with the lowest mean relative price and exchange rate has the highest mean import and mean real GDP than those of frontier and emerging economies. The high level of import in developed economy is due to the high level of industrialization and supremacy of capital. However, if we see the mean im- port to mean GDP ratio, frontier economy has the highest mean import to mean GDP ratio (28%) relative to that of emerging economy (25%) and de- veloped economy (15%). Since emerging and frontier economies have the scope for more industrialization and economic growth, still high dependency on import exists in those economies. The developed economy’s real GDP and import are 56 times and 31 times higher than those of frontier economies and the emerging economy’s real GDP and import are 19 times and 16 times S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 14 higher than those of frontier economies. All the variables are not normal sug- gested by the significant JB statistic. Therefore, natural logarithm of all vari- ables is considered to ensure the normalization. 4. Model Specification Following Bahmani-Oskooee (1986), the following model is specified to estimate the long-run elasticities of import with respect to gross domestic product, relative prices, and exchange rate: , (9) denotes the value of merchandise imports to country i from j at time t, denotes the real gross domestic product of country i different from country j at time t, denotes relative price measured as the ratio of foreign over domestic price index of country i different from country j at time t, and denotes the average exchange rate of country i different from country j at time t. Relative price is measured by the world price index divided by the consumer price index of a particular country. 5. Econometric Methodology, Estimated Results, and Discussion 5.1. Unit Root Test Before estimation of the long-run equation (9), at first step, we need to ensure that whether each variable contains unit root problem or not. In this regard, Im, Peasaran, and Shin (2003), Choi (2006), and Hadri (2000) tests have been applied. Under Im, Peasaran, and Shin (2003) and Choi (2006) tests, the null hypothesis represents the variable under investigation is non-station- ary and under Hadri (2000) test, the null hypothesis represents the variable under investigation is stationary. Several tests have been applied to reach at an overwhelming conclusion. Sometimes a few tests may not give the same conclusion. It is merely due to the power of the tests. The appropriate lag length for each unit root test is selected by AIC and SBIC criteria. The tests results have been highlighted in Table 4, Table 5, and Table 6. 31 2 0. . . . ijt ijt ijt ijt ijtIMP A GDP RP ER ehll l= ijtIMP ijtGDP ijtRP ijtER Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 15 Table 4. Unit Root Test Results Summary for Frontier Economies Model with Constant and Trend Term [Level Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value –1.4820* 0.069 –0.9770 0.164 6.4289*** 0.000 0.1508 0.559 0.6623 0.746 3.4681*** 0.000 –224.59*** 0.000 –5.138*** 0.000 7.6544*** 0.000 1.7592 0.960 1.6461 0.950 6.8065*** 0.000 Model with Constant Term [Level Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value 6.0043 1.000 5.6546 1.000 8.6388*** 0.000 1.8279 0.966 1.8306 0.966 5.5959*** 0.000 –230.60*** 0.000 –5.083*** 0.000 8.8289*** 0.000 9.3871 1.000 8.5589 1.000 9.5433*** 0.000 Model with Constant Term [Difference Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value –12.541*** 0.000 –10.66*** 0.000 0.7721 0.211 –8.5375*** 0.000 –8.124*** 0.000 –0.7008 0.758 –79.671*** 0.000 –6.468*** 0.000 0.7621 0.223 –6.5192*** 0.000 –6.261*** 0.000 0.7105 0.259 Note: ***, **, * denote significance at 1%, 5% and 10% level respectively. Table 5. Unit Root Test Results Summary for Emerging Economies Model with Constant and Trend Term [Level Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value –0.0165 0.4934 0.1011 0.5403 5.1054*** 0.0000 –1.2681 0.1024 –0.5948 0.2760 4.8657*** 0.0000 0.6615 0.7485 0.7029 0.7589 6.0481*** 0.0000 0.8914 0.8136 1.1927 0.8835 6.0000*** 0.0000 Model with Constant Term [Level Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value 3.0899 0.9990 3.0726 0.9989 9.7945*** 0.0000 –1.0222 0.1533 –0.8869 0.1876 5.0464*** 0.0000 –4.06*** 0.0000 49.987*** 0.0000 8.0623*** 0.0000 4.7299 1.0000 4.2505 1.0000 10.116*** 0.0000 Model with Constant Term [Difference Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value –10.8*** 0.0000 –9.594*** 0.0000 0.0387 0.4846 –7.69*** 0.0000 –7.403*** 0.0000 –0.0118 0.5074 –8.27*** 0.0000 –7.709*** 0.0000 0.7411 0.2233 –8.93*** 0.0000 –8.128*** 0.0000 1.0156 0.1549 Note:***, **, * denote significance at 1%, 5% and 10% level respectively. ln IMP ln RP ln ER lnGDP ln IMP ln RP ln ER lnGDP ln IMPD ln RPD ln ERD lnGDPD ln IMP ln RP ln ER lnGDP ln IMP ln RP ln ER lnGDP ln IMPD ln RPD ln ERD lnGDPD S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 16 Table 6: Unit Root Test Results Summary for Developed Economies Model with Constant and Trend Term [Level Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value 0.0070 0.5028 –0.0254 0.4899 4.6409*** 0.0000 –2.7329*** 0.0031 –2.788*** 0.0026 5.4084*** 0.0000 –2.7232*** 0.0032 –2.840*** 0.0023 4.3680*** 0.0000 1.4253 0.9230 1.4443 0.9257 5.4194*** 0.0000 Model with Constant Term [Level Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value 2.7557 0.9971 2.8139 0.9976 11.123*** 0.0000 2.4554 0.9930 2.5089 0.9939 9.4156*** 0.0000 –4.0277*** 0.0000 –4.095*** 0.0000 2.8644*** 0.0021 –0.1334 0.4469 –0.0863 0.4656 11.218*** 0.0000 Model with Constant Term [Difference Form] Variables IPS Test Choi Test Hadri Test Statistic P-value Statistic P-value Statistic P-value –12.019*** 0.0000 –10.50*** 0.0000 0.5951 0.2759 –12.242*** 0.0000 –10.79*** 0.0000 0.4064 0.3422 –7.6854*** 0.0000 –7.341*** 0.0000 0.1940 0.4231 –8.9506*** 0.0000 –8.391*** 0.0000 0.8064 0.2100 Note: ***, **, * denote significance at 1%, 5% and 10% level respectively. From Table 4, Table 5, Table 6, it can be concluded that under Im, Peasa- ran, and Shin (2003) and Choi (2006) tests a few variables are stationary at level form. After making first difference, all variables become more station- ary. However, to reach at an overwhelming conclusion, Hadri (2000) test has been applied. This test reveals that no variable is stationary at level form. All variables are integrated of order one (I(1)). Since each test suggests all varia- bles are stationary at integrated of order one, this study doesn’t need to go for higher order integration. This study gives emphasis on the result of Hadri (2000) test. This test works well in panel data models with fixed effects, indi- vidual deterministic trends, and heterogeneous errors across cross‐sections (Hadri, 2000). Besides, the test can be used for the more general case of seri- ally correlated disturbance terms (Hadri, 2000). 5.2. Panel Cointegration Test At the second step, we need to ensure that whether there exists any coin- tegrating relationship or not. Since all variables are integrated of order one (I(1)), therefore, Johansen Fisher Panel Cointegration test (Johansen, 1995) and Kao Residual Based Cointegration test (Kao, 1999) have been applied. Johansen Fisher Panel Cointegration test (Johansen, 1995, 1988) can only be ln IMP ln RP ln ER lnGDP ln IMP ln RP ln ER lnGDP ln IMPD ln RPD ln ERD lnGDPD Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 17 applied if all variables are integrated or order one (I(1)). The test results are provided in Table 7. The appropriate lag length of Johansen and Fisher Panel Cointegration test (Johansen, 1995, 1988) is selected by the AIC and SBIC criteria. Kao Residual Based Panel Cointegration test (Kao, 1999) is applied to make sure the existence of cointegrating relationship suggested by Johansen Fisher Panel Cointegration test (Johansen, 1995, 1988). The test results are provided in Table 8. Table 7. Johansen and Fisher Panel Cointegration Test Results Test Results for Frontier Economies Case-1: Intercept (no trend) in CE and VAR Case-2: Intercept and Trend in CE and No Trend in VAR CEs Trace Test Max-Eigen Test Trace Test Max-Eigen Test Statistic p-value Statistic p-value Statistic p-value Statistic p-value None 84.34*** 0.000 77.43*** 0.000 86.58*** 0.000 68.10*** 0.000 At most 1 26.42** 0.048 28.34** 0.028 33.07*** 0.007 33.10*** 0.007 At most 2 9.24 0.903 7.08 0.972 11.47 0.779 12.67 0.697 At most 3 20.23 0.209 20.23 0.209 6.84 0.976 6.84 0.976 Test Results for Emerging Economies CEs Trace Test Max-Eigen Test Trace Test Max-Eigen Test Statistic p-value Statistic p-value Statistic p-value Statistic p-value None 140.7*** 0.000 114.5*** 0.000 132.0*** 0.000 161.5*** 0.000 At most 1 47.20*** 0.000 36.41*** 0.002 31.82** 0.011 30.84** 0.014 At most 2 25.15* 0.067 16.25 0.435 12.46 0.712 9.85 0.874 At most 3 33.84*** 0.005 33.84*** 0.005 11.30 0.790 11.30 0.790 Test Results for Developed Economies CEs Trace Test Max-Eigen Test Trace Test Max-Eigen Test Statistic p-value Statistic p-value Statistic p-value Statistic p-value None 96.55*** 0.000 66.78*** 0.000 160.8*** 0.000 124.1*** 0.000 At most 1 44.77*** 0.000 33.01** 0.016 61.61*** 0.000 44.59*** 0.001 At most 2 26.39* 0.091 22.82 0.197 31.15** 0.027 24.89 0.128 At most 3 26.42* 0.091 26.42* 0.091 19.08 0.387 19.08 0.387 Note: ***, **, * denote significance at 1%, 5% and 10% level respectively. Appropriate lag length for this test has been selected by AIC and SBIC. Case-1 and Case-2 suggest one cointegrating equation. Table 8. Kao Residual Based Panel Cointegration Test Results Frontier Economies Emerging Economies Developed Economies ADF Test –4.9828*** (0.000) –4.6298*** (0.000) –4.1689*** (0.000) Note: ***, **, * denote significance at 1%, 5% and 10% level respectively. The results from both tests suggests that there exists cointegrating rela- tionship among the variables. Johansen Fisher Panel Cointegration test (Jo- hansen, 1988, 1995) is applied for two cases. At first this test is applied by incorporating only intercept in cointegrating equation and VAR (Vector Auto S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 18 Regressive) model and later by incorporating intercept and trend in cointegrat- ing equation and only intercept in VAR (Vector Auto Regressive) model. In both cases we have found the existence of cointegrating relationship among the variables for frontier, emerging, and developed economies. Therefore, in the long-run, import, relative price, real GDP, and exchange rate will move together in frontier, emerging, and developed economies. The same conclu- sion has been drawn from Kao Residual Based Panel Cointegration test (Kao, 1999). 5.3. Granger Causality Here all variables are integrated of order one (I(1)). Hence, F-test under a multivariate VECM framework (Engel and Granger, 1987) is used to iden- tify the direction of causal relationship among the variables. To account for the long-run causality, an error correction term (ECM) is added in the VAR system. The multivariate VECM framework to check the direction of causality is given below: (10) The parameters – , , and are to be estimated. The one period lagged error term, , is derived from long-run equation (11). are random error terms being serially independent with mean zero and finite var- iance-covariance matrix. The causality analysis results have been provided in Table 9. From the estimated results, it can be said that there is a short-run unidirec- tional causality from import to relative price in case of frontier economy, bidirectional causality between import and relative price in case of emerging economy, and unidirectional cau- sality from relative price to import in case of devel- oped economy. Next, it is found that there is a short-run bidirectional causality between import and real GDP in case of frontier econ- omy, short-run unidirectional causality from real GDP to import 11 12 13 14 21 22 23 24 31 32 33 34 41 42 43 44 ln ln ln ln ln ln ln ln ijt ijt kk k k k ijt ijt kk k k k k k k kijt ijt k k k k kijt ijt k K IMP IMP RP RP ER ER GDP GDP q q q q q q q q q q q q q q q q - - - - = D Dé ù éé ùé ù ê ú êê úê úD Dê ú êê úê ú= +ê ú êê úê úD D ê ú êê úê ú ê ú êD Dë û ë ûë û ë å M 1 2 3 4 1 C C C C 1 2 1 3 4 ijt ijt ijt ijt ijt ECM l l l l µ µ µ µ - ù é ùé ù ú ê úê ú ú ê úê ú+ú ê úê ú ú ê úê ú ú ê úë ûû ë û 1 2 3 4 + 'C s 'sq 'sl 1ijtECM - ijtµ ( ln ln )IMP RPD Þ D ( ln ln )IMP RPD Û D ( ln ln )RP IMPD Þ D ( ln ln )IMP GDPD Û D Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 19 Table 9. Granger Causality Results Frontier Economies [t-statistic] 0.4004 (0.5274) 0.0327 (0.8566) 20.0404*** (0.0000) –3.0632*** (0.0024) 3.4822* (0.0631) 1.1838 (0.1768) 3.4256* (0.0653) –3.1559*** (0.0018) 0.0060 (0.9383) 0.5769 (0.4482) 3.5163* (0.0618) –1.2876 (0.1989) 5.2600** (0.0226) 4.2718** (0.0397) 2.7622* (0.0977) 2.0019** (0.0463) Emerging Economies [t-statistic] 6.0414*** (0.0027) 5.2807*** (0.0057) 3.5862** (0.0291) –3.8686*** (0.0000) 3.2367** (0.0409) 22.2954*** (0.0000) 4.0832** (0.0179) –2.7320*** (0.0067) 0.6587 (0.5184) 4.0616** (0.0184) 0.3600 (0.6980) 0.0420 (0.9665) 0.4153 (0.6606) 5.4684*** (0.0047) 5.4327*** (0.0049) –3.6400*** (0.0000) Developed Economies [t-statistic] 2.3797* (0.0703) 4.1415*** (0.0069) 1.1445 (0.3318) –2.2492** (0.0254) 0.5974 (0.6173) 7.5760*** (0.0001) 3.2305** (0.0231) –0.9746 (0.3307) 0.3635 (0.7794) 0.6554 (0.5803) 0.5817 (0.6275) 0.0628 (0.9499) 2.5610* (0.0556) 1.4997 (0.2152) 2.3786* (0.0704) –1.1614 (0.1078) Note:***, **, * denote significance at 1%, 5% and 10% level respectively. In the parentheses (), the p-values of the Wald statistics is presented. Here, the represents the unidirectional causality and represents the bidirectional causality. in case of emerging economy, and short-run unidirec- tional causality from import to real GDP in case of developed economy. Further, it is found that there is a short-run bidirectional causality between relative price and exchange rate in case of emerging economy and short-run unidirectional causality from exchange rate to relative price in case of developed economy. In addition, it is found that there is a short-run bidirectional causality between ln IMPD ln RPD ln ERD lnGDPD 1( 1)ECM - ln IMPD ln RPD ln ERD lnGDPD ln IMPD ln RPD ln ERD lnGDPD 2 ( 1)ECM - ln IMPD ln RPD ln ERD lnGDPD ln IMPD ln RPD ln ERD lnGDPD 3( 1)ECM - ln IMPD ln RPD ln ERD lnGDPD Þ Û ( ln ln )GDP IMPD ÞD ( ln ln )IMP GDPD ÞD ( ln ln )RP ERD Û D ( ln ln )ER RPD ÞD S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 20 real GDP and relative price in case of both frontier and emerging economies, and short-run unidirectional causality from real GDP to relative price in case of developed economy. Again, it is found that there is a short-run unidirectional causality from ex- change rate to import in case of both emerging and developed economies. Finally, it is found that there is a short-run bidirectional causality between exchange rate and real GDP in case of frontier economy and short-run unidirectional causality from exchange rate to real GDP in case of both emerging and developed economies. Besides, The significance of , , and confirmed the existence of long-run causality. Therefore, ex- change rate, relative price, and real GDP cause import demand in the long- run. 5.4. Long-run Elasticities The long-run import elasticities with respect to real gross domestic prod- uct, relative prices, and exchange rate are estimated from the following equa- tion: (11) The equation 11 is estimated by Panel Dynamic Ordinary Least Square (DOLS) approach (Stock and Watson, 1993). The appropriate lead and lag differences of independent variables have been used to control the endogenous feedback. The lead and lag length ( ) are selected by the AIC and SBIC criteria. The panel DOLS models of all economies are free from serial correlation. The long-run elasticities have been provided in Table 10. ( ln ln )GDP RPD Û D ( ln ln )GDP RPD ÞD ( ln ln )ER IMPD Þ D ( ln ln )ER GDPD Û D ( ln ln )ER GDPD ÞD 1( 1)ECM - 2 ( 1)ECM - 3( 1)ECM - 0 1 2 3 1 2ln ln ln ln ln ln p q ijt ijt ijt ijt k ijt k k ijt k k p k q IMP GDP RP ER GDP RPl l l l q q- - =- =- = + + + + D + Då å 3 ln r k ijt k ijt k r ERq h- =- + D +å [ ] [ ] [ ], , , , ,p p q q r r- + - + - + Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 21 Table 10. Long-run Elasticities [Dependent Variable – ] Frontier Economy Emerging Economy Developed Economy 0.9295*** (0.0000) 0.9078*** (0.0000) 0.9022*** (0.0000) –0.3636*** (0.0000) –0.2999*** (0.0000) 0.6415*** (0.0000) –0.1137*** (0.0000) 0.0093*** (0.0095) –0.0467*** (0.0000) Constant 1.0122 (0.2429) 1.0470 (0.2468) 0.9085*** (0.0043) Note: ***, **, * denote significance at 1%, 5% and 10% level respectively. From the above analysis, it can be concluded that real GDP has a significant positive impact on import demand across frontier, emerging, and developed economies in the long-run (see also Dutta and Ahmed, 1999; Tang, 2003; Na- rayan and Narayan, 2005; Hye and Mashkoor, 2010). Irrespective of the nature of the economy, increase in real GDP always boosts the import level of an economy. Again, the relative price has a significant negative impact on the import demand in case of frontier (Hye and Mashkoor, 2010; Ahad et al., 2017) and emerging economies (Mishra and Mohanty, 2017) but has a signif- icant positive impact on developed economy in the long-run (see also Hibbert et al., 2012; Tang, 2003). Therefore, import in developed economy is less sen- sitive to the increase in price level. This indicates the existence of differential human characteristics between frontier and developed economies. However, the economic theory argues that raise in relative price of import dampens the level of import (see also Narayan and Narayan, 2005; Chang et al., 2005; Thaver and Ekanayake, 2010). Hence, in case of emerging and frontier econ- omies, the increase in relative price diminishes the demand of high price goods from abroad. However, the customers of the developed economy have differ- ent attitude towards foreign goods with respect to price, quality, brand value, excellence, safety etc., resulting a high demand in mentioned feature products even in high relative prices. Finally, the exchange rate (DC/USD) has a sig- nificant negative impact on import in case of frontier economy (Omotor, 2010) and developed economy but has a significant positive impact on import de- mand in case of emerging economy in the long-run. Therefore, in case of de- veloped and frontier economies, the level of import decreases due to the raise in import cost, originated from the depreciation of the domestic currency. However, the growing demand for raw materials and capital assets in emerg- ing economy is not affected by exchange rate movement, increasing the import level even in the depreciation of exchange rate. ln IMP lnGDP ln RP ln ER S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 22 5.5. Short-run Analysis To find out the short-run elasticities of imports demand, the following er- ror correction model is estimated: , (12) To check the existence of cross sectional dependence, Presaran Cross Sec- tional Dependency test is applied. After detecting cross sectional dependence, the cross sectional seemingly unrelated regression (SUR) is applied. When cross sectional SUR is used, the estimation become devoid of cross sectional dependence and other diagnostic problems such as autocorrelation, heterosce- dasticity, and functional misspecification. Next denotes one period lagged error term derived from long run cointegration equation. denotes the speed of adjustment with the expected negative value and magnitude less than or equal to one ( ) towards the long run equilibrium if there is any shock in the import demand due to change in real GDP, relative price, and exchange rate. Table 11. Short-run Results [Dependent Variable – ] Frontier Economies Emerging Economies Developed Economies 0.7715*** (0.0020) 1.6852*** (0.0000) 0.6771*** (0.0000) 0.3071*** (0.0000) 0.0738** (0.0301) 0.4559*** (0.0000) –0.1820** (0.0139) –0.1550*** (0.0005) 0.0682* (0.0674) –0.0368* (0.0625) –0.0260*** (0.0008) –0.0222*** (0.0072) 0.0156 (0.4561) 0.0235 (0.3569) 0.0187 (0.8523) Note: Values in () represent p-value. ***, **, * denote significance at 1%, 5% and 10% level respectively. For panel equation estimation, panel autocorrelation and heteroscedasticity consistent estimation has been used and only cross sectional dependence test statistic has been reported. CD represents cross sectional depend- ence. From the above analysis (Table 11), it can be concluded that real GDP has a significant positive impact on import in case of frontier, emerging, and de- veloped economies in the short-run (see also Dutta and Ahmed, 1999; Tang, 2003; Narayan and Narayan, 2005; Hye and Mashkoor, 2010). Moreover, rel- ative price has a significant positive impact on imports regardless of frontier, emerging, and developed economies in the short-run. Therefore, in the short- 1 2 3 1ln ln ln lnijt ijt ijt ijt ijt ijtIMP GDP RP ER ECMf f f l w-D = D + D + D + + 1ijtECM - l 1l £ ln IMPD lnGDPD ln RPD ln ERD ( 1)ECM - 2 CDc Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 23 run import is less sensitive with respect to increase in relative price in all econ- omies. More specifically, it can be said that excessive dependency on the im- port in short-run in all economies cannot even decrease import level due to increase in price level. Moreover, increase in relative prices stimulates the im- ports of foreign products in all economies considering foreign products as su- perior in reference to quality, safety, features and so on. Next, it can be said that exchange rate (DC/USD) has a significant negative impact on import de- mand in case of frontier and emerging economies and a significant positive impact on import demand in case of developed economy in the short-run. It may be occurred that high exchange rate or depreciation in the domestic currency may lessen the import level in short run in case of both frontier and emerging economies but those economies require the high level of import in long run due to the large scale demand in raw materials, capital goods and so on. However, the developed economy is less affected by exchange rate in the short-run due to the existence of fascination over foreign products. The coef- ficient of error correction term , is significant with an expected neg- ative sign and magnitude. If there is any shock to import demand due to change in real GDP, relative price, and exchange rate, it will adjust by 3.68%, 2.6%, and 2.22% in the first year in case of frontier, emerging, and developed econ- omies respectively. The entire convergence process will take approximately 27.17 years, 38.46 years, and 45.45 years in case of frontier, emerging, and developed economies respectively to approach into the long-run equilibrium, if there is any shock in the import demand. Conclusions and Policy Implications This study revisits the import demand function across frontier, emerging and developed economies using the panel variables- import, real GDP, relative price, and exchange rate from 1980–2016. The panel cointegration test sug- gests that long-run relationship exists among import demand, real GDP, rela- tive price, and exchange rate across all the economies. The long-run and short- run estimation results reveal the impact of real GDP, relative prices, and ex- change rate on import demand. For example, real GDP has a significant posi- tive impact on import demand irrespective of the nature of the economy in the long-run and in the short-run. Therefore, from the estimated results, it can be said that increase in GDP in the long run accelerates the import demand in all economies. The relative price has a significant negative impact on the import demand in case of frontier and emerging economies in the long run but has a significant positive impact in the short run whereas relative price has a sig- nificant positive impact on import demand in case of developed economy in ( 1)ECM - S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 24 the long-run and in the short run. The exchange rate (DC/USD) has a signifi- cant negative impact on import demand in case of frontier economy both in short-run and in the long-run but has a significant positive impact on import demand in case of developed economy both in the short-run and in the long- run. However, the exchange rate has a significant positive impact on import demand in the long-run and a significant negative influence in the short run in case of emerging economy. Thus it can be said at the early stage of economic development the relative price and exchange rate have significant negative impacts on import demand. Due the development of the economy, the human characteristics will be changed as a result the nature of important demand function will also be changed. Form the policy perspectives, exchange rate of the frontier and emerging economies should be regulated to increase the import level. However, to con- trol the loopholes of this policy, import of luxury goods rather than capital goods and raw materials should be controlled from getting benefits of lower exchange rate. The relative prices of the frontier and emerging economies should be kept in a certain level to strengthen the competitive position of the domestic products whereas the relative prices of developed economies should be kept at a certain level to protect the invasion of the foreign products which in turn will make the domestic products more popular. This research is very much deductive in nature. We have tested an estab- lished import demand function (Bahmani-Oskooee, 1986) at different context. However, an inclusion of another macroeconomic variables like financial de- velopment and foreign exchange reserve in the existing import demand func- tion can make the findings more trustworthy and draw a sheer attention of the economists and policy makers. The level of relative price making the domestic products of frontier and emerging economy more competitive is a matter of further investigation. Finally it can be concluded that this study plays significant role for policy makers at national and international level to forecast the dynamics of im- portant demand due to change in income level, relative price, exchange rate for all economies. References Abbott, A.J. Seddighi, H.R. (1996). Aggregate Imports and Expenditure Components in the UK: An Empirical Analysis. Applied Economics, 28, 1119–1125. http://dx.doi.org/10.1080/000368496327967 Adam, A., Katsimi, M., Moutos, T. (2008). Inequality and the Import Demand Function CE- Sifo. Working Paper No.2196, Category 6: Monetary Policy and International Finance January 2008. Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 25 Afzal, M. (2001). Exports in a Simultaneous Model: The Case of Pakistan. Government College, Economic Journal, 34(1&2), 53–66. Ahad, M., Afza, T., Shahbaz, M. (2017). Financial Development and Estimation of Import De- mand Function in Pakistan: Evidence from Combined Cointegration and Causality Tests. Global Business Review, 18(1), 1–14. http://dx.doi.org/10.1177/0972150916666909 Anyemedu, K. (1995). The Effects of Trade Liberalization on Industry and Labour in Ghana, the Friedrich Ebert Foundation. Accra, Ghana. Arize, A. C., Malindretos, J. (2012). Foreign Exchange Reserve in Asia and its Impact on Im- port Demand. International Journal of Economics and Finance, 4(3), 21–32. http://dx.doi.org/10.5539/ijef.v4n3p21 Bahmani-Oskooee, M. (1986). Determinants of International Trade Flows: The Case of Devel- oping Countries. Journal of Development Economics, 20(1), 107–123. Bahmani-Oskooee, M., Niroomand, F. (1998). Long-run Price Elasticities and the Marshall- Lerner Condition Revisited. Economics Letters, 61(1), 101–109. http://dx.doi.org/10.1016/S0165-1765(98)00147-5 Bahmani-Oskooee, M., Kara, O. (2003). Relative Responsiveness of Trade Flows to a Change in Price and Exchange Rate. International Review of Applied Economics, 17(3), 293– –308. http://dx.doi.org/10.1080/0269217032000090496 Bahmani-Oskooee, M., Kara, O. (2005). Income and Price Elasticities of Trade: Some New Estimates, The International Trade Journal, 19(2), 165–178. http://dx.doi.org/10.1080/08853900590933125 Chang, T., Ho, Y.H., Huang, C. (2005). A Reexamination of South Korea’s Aggregate Import Demand Functions: The Bounds Test Analysis. Journal of Economic Development, 30(1), 119–128. Choi, I. (2006). Nonstationary Panels. Chapter 13 in T.C. Mills and K. Patterson (eds.), Pal- grave Handbooks of Econometrics, 1, 511–539, Palgrave, Macmillan. Dutta, D., Ahmed, N. (1999). An Aggregate Import Demand Function for Bangladesh: A Coin- tegration Approach. Applied Economics, 31, 465–72. http://dx.doi.org/10.1080/000368499324174 Dutta, D., Ahmed, N. (2004). An Aggregate Import Demand Function for India: A Cointegra- tion Analysis. Applied Economics Letters, 11, 607–13. http://dx.doi.org/10.1080/1350455042000271134 Egwaikhide, F.O. (1999). Determinants of Imports in Nigeria: A Dynamic Specification. De- partment of Economics, University of Ibadan, Nigeria. AERC Research Paper 91. Engle, R.F., Granger, C.W.J. (1987). Cointegration and Error Correction: Representation, Esti- mation and Testing. Econometrica, 55, 251–276. Erlat, G., Erlat, H. (1991). An Empirical Study of Turkish Export and Import Functions. CBRT and METU. Ghorbani, M., Motallebi, M. (2009). Application of Pesaran and Shin Method for Estimating Irans’ Import Demand Function. Journal of Applied Sciencs, 9(6), 1175–1179. http://dx.doi.org/10.3923/jas.2009.1175.1179 Giovannetti, G. (1989). Aggregated Imports and Expenditure Components in Italy: An Econo- metric Analysis, Applied Economics, 21, 303–313. Hardi, K. (2000)., Testing for Stationarity in Heterogeneous Panel data. The Econometrics Journal, 3(2), 148–161, https://doi.org/10.1111/1368-423X.00043 Hibbert, K., Thaver, R., Hutchinson, M. (2012). An Econometric Analysis of Jamaica’s Import Demand Functions with the US and UK. The International Journal of Business and Fi- nance Research, 6(1), 109–120. S. Hossain, K. K. Sen, T. Abedin, M. S. R. Chowduhury DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 26 Houthakker, H.S., Magee, S.P. (1969). Income and Price Elasticities in World Trade. Review of Economics and Statistics, 21(2), 111–125. Hye, Q.M., Mashkoor, M. (2010). Import Demand Functions for Bangladesh: A Rolling Win- dow Analysis. African Journal of Business Management, 4(10), 2150–2156. Ibrahim, M. A. (2015). Merchandise Import Demand Function in Saudi Arabia. Applied Eco- nomics and Finance, 2(1), 55–65. http://dx.doi.org/10.11114/aef.v2i1.626 Ibrahim, M.A. (2017). An Examination of Merchandise Imports Demand Function for Egypt. Applied Economics and Finance, 4(2), 101–112. http://dx.doi.org/10.11114/aef.v4i2.1969 Im, K.S., Peasaran, M.H., Shin, Y. (2003). Testing for Unit Roots in Heterogeneous Panels. Journal of Econometrics, Vol. 115, pp. 53–74. http://dx.doi.org/10.1016/S0304-4076(03)00092-7 Islam, A.M., Hassan, M.K. (2004). An Econometric Estimation of the Aggregate Import De- mand Function for Bangladesh: Some Further Results. Applied Economics Letters, 11, 575–580. http://dx.doi.org/10.1080/1350485042000217990 Johansen, S. (1988). Statistical Analysis of Cointegration Vectors. Journal of Economic and Control, 12, 231‐254. http://dx.doi.org/10.1016/0165-1889(88)90041-3 Johansen, S. (1995). Likelihood–Based Inference in Cointegrated Vector Autoregressive Mod- els. Oxford, Oxford University Press. http://dx.doi.org/10.1093/0198774508.001.0001 Junz, H.B., Rhomberg, R.R. (1973). Price Competitiveness in Export Trade Among Industrial Countries. The American Economic Review, 63, 412–418. Kao, C. (1999). Spurious Regression and Residual-Based Tests for Cointegration in Panel Data. Journal of Econometrics, 90,1–44. http://dx.doi.org/10.1016/S0304-4076(98)00023-2 Katsimi, M., Moutos, T. (2011). Inequility and the US Import Demand Function. Journal of International Money and Finance, 30(11), 492–506. Kotan Z., Saygılı M. (1999). Estimating an Import Demand Functions for Turkey. The Central Bank of the Republic of Turkey Research Department, Discussion Paper No: 9909. Kreinin, M.E. (1973). Disaggregated Import Demand Functions – Further Results. Southern Economic Journal, 40, 19–25. Min, B.S., Mohammad, H.A., Tang, T.C. (2002). An Analysis of South Korea’s Import De- mand. Journal of Asia-Pacific Affairs, 4, 1–17. Mishra, B.R., Mohanty, A. (2017). An Empirical Analysis of Aggregate Import Demand Func- tion for India. Global Economy Journal, 17(4), 1–11. http://dx.doi.org/10.1515/gej-2017-0049 Mohammad, H.A., Tang, T.C. (2000). Aggregate Imports and Expenditure Components in Ma- laysia: A Cointegration and Error Correction Analysis. ASEAN Economic Bulletin, 17, 257–269. MSCI (2019). MSCI 2019 Annual Market Classification Review. Retrieved from https://www.msci.com/market–classification. Mwega, F.M. (1993). Import Demand Elasticities and Stability during Trade Liberalization: A case Study of Kenya. Journal of African Economies, 2, 381–416. http://dx.doi.org/10.1093/oxfordjournals.jae.a036790 Narayan, S., Narayan, K. (2005). An Empirical Analysis of Fiji's Import Demand Functions. Journal of Economic Studies, 32(2), 158–168. http://dx.doi.org/10.1108/01443580510600931 Omotor, D. (2010). An Aggregate Import Demand Functions for Nigeria. Economic Research, 23(1), 1–14. http://dx.doi.org/10.1080/1331677X.2010.11517397 Revisiting the Import Demand Function: A Comparative Analysis DYNAMIC ECONOMETRIC MODELS 19 (2019) 5–27 27 Orcutt, G.H. (1950). Measurement of Price Elasticities in International Trade. Review of Eco- nomics and Statistics, 32(2),117–132. http://dx.doi.org/10.2307/1927649 Rana, P. (1983). The Impact of Generalized Floating on Trade Flows and Reserve Needs: Se- lected Asian Developing Countries. New York: Garland Publishers. Rehman, H. U. (2007). An Econometric Estimation of Traditional Import Demand Function for Pakistan. Pakistan Economic and Social Review, 45 (2), 245–256. Sinha, D. (1997). Determinants of Import Demand in Thailand. International Economic Jour- nal, 11, 73–83. http://dx.doi.org/10.1080/10168739700000027 Stock, J.H., Watson, M.W. (1993). A Simple Estimator of Cointegrating Vectors in Higher Order Integrated Systems. Econometrica, 61, 783–820. http://dx.doi.org/10.2307/2951763 Tang, T.C. (2003). An Empirical Analysis of China's Aggregate Import Demand Functions. China Economic Review, 14(2), 142–163. http://dx.doi.org/10.1016/S1043-951X(03)00021-X Tang, T.C. (2005). A Reassessment of Aggregate Import Demand Function in the ASEAN-5: A Cointegration Analysis. The International Trade Journal, 18, 239–268. Thaver, R., Ekanayake, E.M. (2010). The Impact of Apartheid and International Sanctuions on South Africa's Import Demand Fuction: An Emipirical Analysis. International Journal of Business and Finance Research, 4(4), 11–22. Warner, D., Kreinin, M.E. (1983). Determinants of International Trade Flows. The Review of Economics and Statistics, 65, 96–104. Wilson, J. Takacs, W. (1979). Differential Responses to Price and Exchange Rate Influences in the Foreign Trade of Selected Industrial Countries. The Review of Economics and Sta- tistics, 61, 267–279. http://dx.doi.org/10.2307/1924595 Yin, F., Hamori, S. (2011). Estimating Import Demand Function in the Autoregressive Distrib- uted Lag Framework: The Case of China. Economics Bulletin, 31(2), 1576–1591.