BANK PARIKRAMA A Journal of Banking & Finance Volume September 2020 - June 2021 XLV (3&4) & XLVI (1&2) - Mirza Azizul Islam, Ph.D. The Role and Functions of a Central Bank - Atiur Rahman, Ph.D. Ethics in Banking: Interfacing Financial Inclusion for Socially Responsible Development - Mousumi Choudhury, Ph.D. - Ranjit Singh, Ph.D. Measuring Customers' Satisfaction in Bancassurance Channel: An Investigation Using Gap Model - Md. Rostam Ali, Ph.D. Examining Cost and Accessibility to Institutional Trade Credit by SMEs of Bangladesh in an Environment of Financial Crisis - Md. Thasinul Abedin - Kanon Kumar Sen - Muhammad Shafiur Rahman Chowduhury - Sharmin Akter Revisiting the Investment-Remittance Nexus in Bangladesh: Do Interest Rate, Exchange Rate, and Per Capita GDP Matter? - Mohammad Monirul Islam Sarker Determinants of Bangladesh’s Trade: A Gravity Model Approach - Jubair Ahmad - Tahamina Akter Effectiveness of Monetary Transmission Mechanism in Bangladesh: A Price-based Approach - Mohammad Mohidul Islam Determining the Monetary Dynamics of Inflation in Bangladesh www.bibm.org.bd journal.bibm.org.bd QUARTERLY JOURNAL OF BANGLADESH INSTITUTE OF BANK MANAGEMENT BANK PARIKRAMA A Journal of Banking & Finance Volume XLV (3 & 4) & XLVI (1 & 2) September 2020 - June 2021 BANGLADESH INSTITUTE OF BANK MANAGEMENT ISSN 1019-7044 Mohammed Farashuddin, Ph.D. Qazi Kholiquzzaman Ahmad, Ph.D. Former Governor, Bangladesh Bank and Chief Advisor and Founder Vice Chancellor East West University Chairman, Palli Karma-Sahayak Foundation and Chairman, Dhaka School of Economics Mustafa K. Mujeri, Ph.D. Executive Director, Institute of Inclusive Finance and Development; and Former Director General, Bangladesh Institute of Development Studies Fahad Khalil, Ph.D. Professor, Department of Economics University of Washington United States of America Motiur Rahman, Ph.D. JP Morgan-Chase Endowed Professor of Finance and MBA Director, McNeese State University United States of America Barkat-e-Khuda, Ph.D. Dr. Muzaffer Ahmad Chair Professor, BIBM and Former Professor of Economics, University of Dhaka Professor Mustafizur Rahman Distinguished Fellow, Centre for Policy Dialogue Chairman Md. Akhtaruzzaman, Ph.D. Director General, Bangladesh Institute of Bank Management (BIBM) Members Shah Md. Ahsan Habib, Ph.D. Professor (Selection Grade), BIBM Md. Nehal Ahmed Professor (Selection Grade) and Director (Dhaka School of Bank Management), BIBM Prashanta Kumar Banerjee, Ph.D. Professor (Selection Grade), BIBM Ashraf Al Mamun, Ph.D. Associate Professor and Director (Research, Development & Consultancy), BIBM Ashraf Al Mamun, Ph.D. Associate Professor and Director (Research, Development & Consultancy), BIBM Md. Sirajul Islam Faculty Member (on Deputation), BIBM and General Manager, Bangladesh Bank Md. Mosharref Hossain, Ph.D. Associate Professor, BIBM Md. Shahid Ullah, Ph.D. Associate Professor, BIBM Md. Mahabbat Hossain, Ph.D. Assistant Professor, BIBM Papon Tabassum Research Officer, BIBM Bank Parikrama Volume XLV (3 & 4) & XLVI (1 & 2), September 2020 - June 2021 (pp. 89-120) Revisiting the Investment-Remittance Nexus in Bangladesh: Do Interest Rate, Exchange Rate, and Per Capita GDP Matter? - Md. Thasinul Abedin - Kanon Kumar Sen* - Muhammad Shafiur Rahman Chowduhury* - Sharmin Akter* Abstract Due to the existence of intriguing association and mixed conclusions from the past studies, this study reexamines the linkage between domestic investment and remittances in Bangladesh by using time series data from 1976 to 2016. Employing Fully Modified Ordinary Least Square (FMOLS), this study reveals that remittances unlike domestic interest rate have a significant positive impact on domestic investment in the long-run. This study also exposes the significant positive impact of both foreign exchange rate and per capita GDP on remittances in the long-run. Hence over time, depreciation in Bangladeshi currency and an increase in per capita GDP increase remittances which afterwards increase domestic investment. Significant positive impact of foreign exchange rate and per capita GDP on remittances and significant positive impact of remittances on domestic investment in the long-run claim the dominance of investment motive of remittances. This study suggests that to boost up remittance, depreciation of domestic currency should be in a convenient level for the greater interest of the economy. Besides the double-digit interest rate affects negatively to domestic investment and our domestic lending interest rate is not lower enough to increase domestic investment. Keywords: Bangladesh, Domestic Investment, Domestic Interest Rate, Foreign Exchange Rate, Per Capita GDP, Remittances. JEL Classifications: C01, C02, C31 1. Introduction The remittances usually inflowing through the formal channels stimulate the macroeconomic scenarios of a country. For example, boosting savings and investment, shrinking poverty, and building up capital (Akter, 2016). Increasing private expenditures and incurring more household consumption, remittances The authors are Assistant Professor, Department of Accounting, University of Chittagong, Chittagong; Lecturer, Department of Business Administration in Accounting and Information Systems, Bangladesh University of Professionals, Dhaka; Assistant Professor, Department of Accounting, University of Chittagong, Chittagong; and Lecturer, Department of Accounting and Information Systems, Jashore University of Science and Technology, Jashore; Bangladesh, respectively. The views expressed in this paper are the authors’ own. 90 Bank Parikrama fuel the grass root development of a country (Connell and Conway, 2000). The large extent of foreign reserves built from remittances makes the domestic currency stronger (Taylor and Wyatt, 1996). Moreover, remittances improve the balance of payment deficits, sanction and education, infrastructure development, and quality of life (Matuzeviciute and Butkus, 2016). However, the effect of remittances swells inflation in an economy (Chami et al., 2003; Faini, 2002; Narayan et al., 2011; Pradhan et al., 2008). Hence, remittances have both positive and negative impact on the economy. Previous empirical studies conclude that the influence of remittances highly depends on country specific factors such as government policies, restriction in the flow of remittances, investment environment, politics, and corruption. 75% of worldwide remittances come into developing countries (MRF, 2016). In 2017, estimated remittances to developing countries are 465.7 billion USD; 3.7% greater than those of 2016, whereas remittances into South Asia are 129.3 billion USD; 4.6% greater than those of 2016. Bangladesh has placed fifth in the stocks of emigrants in 2013 (MRF, 2016) and among the top remittances receiving countries, Bangladesh has placed tenth in 2015. Moreover, Bangladesh has among the lowest cost corridors for sending remittances from Singapore to Bangladesh and from Saudi Arabia to Bangladesh; increasing the remittances to the economy (MRF, 2016). In 1976, the amount of remittances in Bangladesh was only USD 24 million (Barua et al., 2007) whereas in 2016, the amount of remittances was USD 13.6 Billion (source: Bangladesh Bank). This high growth rate of remittances contributes into the domestic investment to change the economic structure of Bangladesh (Barua et al., 2007). The trends of domestic investment, remittances, foreign exchange rate (BDT/USD), per capita GDP, and the domestic interest rate are given below: Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus Figure 1: Remittances Figure 2: Domestic Investment % of GDP % of GDP 15 10 5 0 1960 1980 2000 1976-2016 40 20 1980 2000 1976-2016 2020 Source: WBDI Figure 3: Lending Interest Rate Figure 4: Per Capita GDP 1500 In USD In % 60 0 1960 2020 Source: WBDI1 20.0000 15.0000 10.0000 5.0000 0.0000 1960 91 1980 2000 1976-2016 Source: WBDI 2020 1000 500 0 1960 1980 2000 1976-2016 2020 Source: WBDI BDT/USD Figure 5: Foreign Exchange Rate (BDT/USD) 100 80 60 40 20 0 1960 1980 2000 1976-2016 2020 Source: WBDI From above figures, it has been observed that remittances in Bangladesh have an increasing trend from 1976 to 2016. Domestic investment and per capita GDP have an upward trend from 1976 to 2016. An increase in domestic investment can be supported by an increase in remittances. For example, as the local currency has been depreciated from 1976 to 2016, it is seemingly expected that due to depreciation in Bangladeshi currency, more remittances inflow into 1 World Bank Development Indicators 92 Bank Parikrama the Bangladesh economy and as a result domestic investment goes up due to increase in the households’ savings. However, the relationship is not that much straight forward rather it is more inter-winded. For example, due to depreciation in Bangladeshi currency, households’ expenditure on import goes up due to increase in import costs. The relative increase in import cost increases the inflation leading to a subsequent increase in the interest rate. However, in Bangladesh, both inflation and lending interest rates are decreasing over years. Even real lending interest rate is also in decreasing trend. Therefore, among numerous factors affecting lending interest rate of an economy, increased import cost, theoretically a contributing factor, might affect inflation rate and thereby to the lending interest rate of Bangladesh. The high interest rate negatively affects the domestic investment. The domestic investment is also affected by the fall households’ savings due to increase in households’ consumption expenditure. It is well known in the economic literature that a country’s part of domestic investment is usually backed by the country’s domestic savings. Even though remittances increase as a result of the depreciation in foreign exchange rate, the households do not change the consumption behavior overnight. In fact, the households do not significantly increase their consumption expeditiously. As a result, the increase in the households’ savings due to increase in remittances increases the domestic investment. Hence, the domestic investment increases due to depreciation in foreign exchange rate, if and only if the impact of households’ savings can outweigh the impact of domestic interest rate. The empirical literatures (Hall, 1977; Osundina and Osundina, 2016; Wuhan et al., 2015) have explained the negative relationship between domestic investment and interest rate. Moreover, domestic investment and per capita GDP have a common upward trend from 1976 to 2016. Therefore, it will not be unwise to expect a positive relationship between domestic investment and per capita GDP. There is dearth of literature that per capita GDP indicating the economic condition of a country has a positive impact on remittances. Most of the studies considered remittances as determinant of per capita GDP. However, per capita GDP may have either positive or negative impact on remittances inflow. Now the question is how? It depends on three motives: altruistic motive, insurance motive, and investment motive. Migrants use income level to judge the economic Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 93 condition of home country with that of the host country. For example, due to decrease in income level, migrants send more remittances from altruistic motive of remittances. If the altruistic motive dominates the remitting behavior, the expected sign of this variable is negative. For example, when the income of the family in the home country decreases, the migrant will send more money in order to assure the same level of the utility for his family. Moreover, the migrants transfer remittances for the welfare (health and education, etc.) of wives and children. However, given the dominance of altruistic motive, the depreciation in local currency avails less remittances in dollars that ensure same level of national currency. When the income of the family in the home country increases, the migrant will send more money for financial investments or for inheritance reasons, because his potential of inheritance will increase. This is known as investment motive. For example, if a country like Bangladesh is on the development stage, the people of Bangladesh working abroad may send money to Bangladesh to invest into real estate (buying home) and other sectors offering better returns. Since Bangladesh is on the development stage, it is trying to attract more remittances to improve its socioeconomic structure like construction and development of schools, colleges, health centers, water supply and sanitation, and rural electrification etc. Hence, to stimulate the investment motive, the economic policies designed by the officials in Bangladesh may offer several facilities to the senders of remittances. For example, banks are crediting the proceeds of remittances to the beneficiary's account promptly within maximum by 3(three) days and foreign banks, exchange houses, and subsidiaries or overseas branches of Bangladeshi scheduled banks are drawing remittance facilities with different banks in Bangladesh. If the investment motive dominates, the depreciation in local currency avails more remittances for financial investments or for inheritance reasons, because his potential of inheritance increases. Therefore, it is expected to decrease the amount of remittances in the case of an appreciation of the local currency if investment motive dominates. In case of insurance motivation, a decrease in the income of the family in the home country also decreases the remittances. The migrants take time to remit for offsetting the impact of the appreciation of the local currency (because he must send more money in the foreign currency). Insurance motive is also based on 94 Bank Parikrama Intra-familial arrangements against income volatility. It is a contractual arrangement between the migrant and his family. In the rural areas of most developing counties, where financial and assurance markets are incomplete, the revenues are subject to risks such as drought and price fluctuations, etc. To diversify the risk of rural income volatility, families can decide to allocate some members to urban or foreign migration. Although urban and foreign jobs are also subject to risks, these risks are independent from the agricultural income variations. At the beginning of the contract, family pays the migration costs in exchange of future remittances. In the case of these types of family contracts, remittances flow to the family in case of agricultural income drops and to the migrant in case of unemployment (Rapoport and Docquier, 2006). Due to the existence of a strong relationship between the per capita GDP and foreign exchange rate, the impacts have been drawn from separate equations. The growth in per capita GDP reflects the increased productivity and economic growth, increasing domestic investment. To retain economic growth, a significant portion of GDP is reinvested (Demers et al., 2003), particularly in frontier and emerging economies. The foreign exchange rate (BDT/USD) also has significant impact on domestic investment, depending on the country specific factors (Darby et al., 2000) such as trade openness, cost of capital goods, higher capital mobility, and external debt. Again, depreciation in domestic currency has a positive impact on domestic investment when the import quantity of capital products is larger relative to the extent of capital mobility (Serven, 1990). The study contributes into the literature in three ways. Firstly, the study critically examines the investment, altruistic, and insurance motives unlike the existing studies (Barua et al., 2007and Hasan, 2008) in Bangladesh. Secondly, they study proves the dominance of investment motive in Bangladesh employing large sample size and controlling endogenous feedback. Finally, the study recommends that lowering the interest rate and encouraging more remittances boost up domestic investment. Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 95 2. Literature Review and Hypotheses Development Several studies have been found inspecting the linkage among remittances, domestic investment, domestic interest rate, foreign exchange rate, and per capita GDP. The findings of these studies have been outlined below. 2.1 Domestic Investment and Remittances Giuliano and Ruiz-Arranz (2006) have found that remittances work as catalyst to increase domestic investment in a country, particularly countries, which have a weak financial sector. Zeisemer (2006) has found that remittances in a country build up a large extent of saving, decrease the rate of interest, and increase the level of investment. Adams (2006) has found that remittance receiving households spend more on investment goods and invest more on entrepreneurial activities than other households. Mundaca (2009) has found that remittances have a positive impact on investment (see also Gupta et al., 2007). Still very few studies have been found about the impact of remittances on domestic investment in Bangladesh. H1: It is expected that remittances have a positive impact on domestic investment in the short-run and in the long-run. 2.2 Domestic Investment and Domestic Interest Rate Wang and Yu (2007) have identified the impact of interest rate on the domestic investment, which is significant and negative (see also Aysan et al., 2005; Hall, 1977; Osundina and Osundina, 2016; Wuhan et al., 2015). However, Larsen (2004) has found positive impact on interest rate on real estate investment, and Greene and Villanueva (1990) have found a positive relationship between investment and the interest rate in Least Developed Countries (LDCs). H2: It is expected that domestic interest rate has either positive or negative impact on domestic investment in the short-run and in the long-run. 2.3 Remittances and Foreign Exchanges Rate Barua et al. (2007) have found that depreciation in local currency positively affects remittances in Bangladesh. Bouhga-Hagbe (2004) has stated that exchange rate positively affects the remittances flow due to the wealth effect 96 Bank Parikrama (investment motive). Pant and Budha (2016) have found that foreign exchange rate has a significant positive impact on remittances (see also Barajas et al., 2010; Faini, 1994; Hasan, 2008; Rajan and Subramanian, 2005). Chamon et al., (2005) have found that sometimes transient workers take loans to send desired remittances to take advantage of favorable foreign exchange rate. However, Swamy (1981) and Straubbaar (1986) have found insignificant positive effect of exchange rate on remittances and Elbadawi and Rocha (1992) and Freund and Spatafora (2005) have found a negative and significant effect of exchange rate on remittances. H3: It is expected that foreign exchange rate has either positive or negative impact on remittances in the short-run and in the long-run. 2.4 Remittances and Economic Growth Matuzeviciute and Butkus (2016) have found that remittances have a significant positive impact on economic growth (see also Imai et al., 2014; Nwaogu and Ryan, 2015; Tahir et al., 2015; Zizi, 2014). However, Lim and Simmons (2015) have found that there is no significant relationship between economic growth and remittances and Feeny et al. (2014) have found that remittances have no impact on economic growth (see also Jouini 2015). Therefore, it can be said that remittances have either positive impact or no impact on economic growth. However, those papers have avoided to examine the impact of economic growth on remittances. Therefore, this study fills out the gap by addressing the impact of per capita GDP on remittances. It is also notable that Barua et al. (2007) have used ratio of foreign GDP to domestic GDP as a determinant of remittances. The study varies in this regard by using solely per capita GDP as a determinant of remittances. In this regard, the study considers the altruistic motive of remittances. If altruistic motive dominates, high income negatively affects remittances. In absence of altruistic motive (presence of investment motive), high income positively affects remittances. Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 97 H4: It is expected that per capita GDP (usually used as a proxy of economic growth) has either positive or negative impact on remittances in the short-run and in the long-run. 2.5 Domestic Investment and Economic Growth Khatib et al. (2012) have found a significant positive impact of per capita GDP on domestic investment in the long-run (see also Acosta and Loza, 2004; Ajide and Lawanson, 2012; Harupara, 1998; Karagoz, 2010; Nghifenwa, 2009). However, Sohail et al. (2014) have found that domestic investment in Pakistan increased real GDP growth. H5: It is expected that per capita GDP (usually used as a proxy of economic growth) has positive impact on Domestic Investment both in the long-run and in the short-run. 2.6 Domestic Investment and Foreign Exchange Rate Buffie (1986) has found that the exchange rate is an important determinant of domestic investment (see also Chavali, 2014). He has found that permanent domestic currency appreciation has a positive impact on domestic investment in the long-run. On the other hand, Campa and Goldberg (1999) have found that domestic currency depreciation has a significant positive impact on the domestic investment in the United States and Japan. Moreover, Serven (1990) has stated that the depreciation in the domestic currency has a positive impact on domestic investment when import quantity of capital products is larger relative to the extent of capital mobility. However, Darby et al. (2000) have found that the impact of exchange rates on domestic investment vary across countries. H6: It is expected that foreign exchange rate has either positive or negative impact on domestic investment in the short-run and in the long-run. Most of the past studies in abroad have not critically examined the three motives of remittances – investment, altruistic, and insurance and proved the existence of dominance of a specific or two or all the motives. In fact, in Bangladesh, all the studies (Barua et al., 2007 and Hasan, 2008) are less thought provoking without considering any motive of remittances. This study fills the gap in this regard. This study has critically examined all the three motives of 98 Bank Parikrama remittances and claimed the dominance of investment motive. Moreover, no study in Bangladesh has analyzed the linkage between domestic investment and remittances where the foreign exchange rate, per capita GDP, and domestic interest rate matter. The study suggests that depreciation of local currency and lower domestic interest rate are important to boost up domestic investment which subsequently contributes to the economic growth. 3. Data, Methodology, and Model Formation The study uses remittances data, domestic investment, domestic interest rate, per capita GDP, and foreign exchange rates from 1976 to 2016 of Bangladesh from the World Bank Development Indicators of the World Bank. The definitions of the key variables used in this study has been described in Table-1. Table 1: Defining Key Variables Variables DINV REM FOREX INT Explanation Domestic Investments (DINV) represent the total investment in a fiscal year within the border of the Bangladesh. Remittances (REM) show the amount foreign currency sent by the migrant Bangladeshis and through outsourcing. Foreign Exchange Rates (FOREX) determines the value Bangladeshi currency with respect to USD. Domestic Interest Rate (INT) represents the average domestic lending rate fixed by banks and financial institutions. This study assumes that the interest rate, per capita GDP, and foreign exchange rate are strictly exogenous in estimating the long-run equation. However, a few empirical studies have found that due to flush in remittances, domestic investment goes up and as a result interest rate is affected. The study assumes that only interest rate affects the domestic investment. Hence, considering the domestic investment, interest rate is also strictly exogenous in long-run equation. Again, the study assumes that remittance is not the strictly exogenous factor in determining the level of domestic investment as the effect of foreign exchange rate and economic growth is usually channeled to the domestic investment through remittances. Hence, the underlying assumption of this study can be delineated in the way that foreign exchange rate and economic growth affect remittances, which affects domestic investment. However, it is notable that Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 99 the inflow of remittance in an economy might be explained by the other socioeconomic factors for example, political turmoil, tax regime, remittance cost, rate of return on real estate, wage rate, inflation, the efficiency of the banking system, income inequality, and working agents (Russell, 1986; Wahba, 1991; ElSakka and McNabb, 1999; Schiopu and Seigfried, 2006). To find out the linkage among per capita GDP (PGDP), Domestic Investment (DINV), Foreign Exchange Rate (FOREX), Domestic Interest Rate (INT), and Remittances (REM), the following two models have been used: DINVt = A0 .REM t1 .INTt2 .et (1) REMt = A1.FOREX t1 .PGDPt2 .et (2) The logarithmic transformation of these equations (Equation-1 and Equation-2) is given below: ln DINVt = 0 + 1 ln REM t + 2 ln INTt + t (3) ln REM t = 0 + 1 ln FOREX t + 2 ln PGDPt + t (4) Let, ln A0 = 0 and ln A1 = 0 . 2016. t represents the time period from 1976 to 1 and 2 represent the elasticity of Domestic Investment (DINV) with respect to Remittances (REM) and Domestic Interest Rate (INT). 1 and 2 represent the elasticity of Remittances (REM) with respect to Foreign Exchange Rate (FOREX) (BDT/USD) and Per capita GDP (PGDP). t and t represent the random error terms. Equation-4 will check whether Foreign Exchange Rate (FOREX) and Per capita GDP (PGDP) have positive or negative impact on remittances. Equation-3 will check whether Remittances (REM) and Domestic Interest Rate (INT) have a positive or negative impact on Domestic Investment (DINV). If Foreign Exchange Rate (FOREX) and Per capita GDP (PGDP) have a positive impact on Remittances (REM) and Remittances (REM) have a positive impact on domestic investment, then it can be concluded that both Foreign Exchange Rate (FOREX) and Per capita GDP (PGDP) have a positive impact on 100 Bank Parikrama Domestic Investment (DINV). The negative (negative impact) can be true and would thus have a negative impact on Domestic Investment (DINV). To ensure whether the Foreign Exchange Rate (FOREX) and Per capita GDP (PGDP) have positive or negative impact on domestic Investment (DINV), the following two models have been used: DINVt = Q0 .PGDPt1 .INTt2 .et ........................................... (5) DINVt = R0 .FOREX t1 .INTt2 .et ........................................(6) The logarithmic transformation of these equations (Equation-5 and Equation-6) is given below: ln DINVt = 0 + 1 ln PGDPt + 2 ln INTt + t ...................(7) ln DINVt = 0 + 1 ln FOREX t + 2 ln INTt + t ...............(8) Let, ln Q0 = 0 and ln R0 = 0 . 2016. t represents the time period from 1976 to 1 and 2 represent the elasticity of Domestic Investment (DINV) with respect to Per capita GDP (PGDP) and Domestic Interest Rate (INT). 1 and 2 represent the elasticity of Domestic Investment (DINV) with respect to Foreign Exchange Rate (FOREX) (BDT/USD) and Domestic Interest Rate (INT). t and t represent the random error terms. It is notable that, due to high positive correlation (multicollinearity problem), per capita GDP (PGDP) and Foreign Exchange Rate (FOREX) are used separately in two equations instead of one equation. Equation-3, Equation-4, Equation-7, and Equation-8 have been estimated by the Feasible Modified Ordinary Least Square (FMOLS), which fixes out the endogeneity and auto-correlation problems if exists (Philips and Hansen, 1990). The appropriate lead and lag length for FMOLS has been selected by SBIC. Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 101 3.1 Unit Root Tests Before estimation of Equation-3, Equation-4, Equation-7, and Equation-8, unit root problem has been checked. In this regard, the ADF test has been applied. The form of this test is given below: m Zt = K0 + K1t + Zt −1 + j Zt − j + ut .........(9) [With constant and trend terms] j =1 m Zt = K0 + Zt −1 + j Zt − j + ut ..........(10) [With constant term only] j =1 Here, Z is the variable under investigation. The variable is of ( I (1) ) if 𝛿 = 0. Appropriate lag length of equation (9) and (10) would be selected by the AIC and SBIC criteria. Apart from ADF test (Dickey and Wayne, 1979), PP test (Phillips and Perron, 1987) has been applied to get an overwhelming conclusion (Phillips, 1986). The unit root test results have been provided in Table-3. The test results suggest all variables are integrated of order one ( I (1) ). 3.2 Cointegration Tests Since all variables are integrated of order one (I(1)), at this stage cointegration relationship among the variables across different equations has been checked. To check cointegrating relationship, the Johansen and Juselius (1990) test has been applied. A brief description of this test is given below: p X t = B0 + X t − p + BX t − j + t ......... (11) j =1 Where, X t represents the vector of endogenous I (1) variables, B0 represents a vector of constant terms, B is the matrix of co-efficients, t is the vector of residuals, and p denotes the lag length. All variables in Equation (11) seem to be endogenous. The long-run relationship among X t determined by the rank of If r = 0 , the variables in the level form do not have cointegrating relationship and equation (11) can be transformed to VAR-model of pth order. (say r ). 102 Bank Parikrama If 0 < r < n, then there are (n r ) matrices of and such that: = . The strength of co-integration relationship is measured by . is called the cointegration vector and X t is of I(0) even if X t is of I (1) . Johansen and Juselius (1990) test suggests the existence of cointegration relationship among the variables in Equation-3, Equation-4, Equation-7, and Equation-8. The results of these tests have been provided in Appendix Table-2. 3.3 Causality Analysis The cointegration test results only says the existence of a long-run relationship among the variables, but it does not say the direction of causal relationship among the variables. To know the direction of causal relationship among the variables, Engel and Granger (1987) F-test has been applied on the first difference forms of the variables in VAR framework. In order to capture the long-run relationships, an error correction term has been included in the VAR system. The augmented form of the Granger causality test in a multivariate VECM framework is presented below for Equation-3, Equation-4, Equation-7, and Equation-8: X1t C1 m 11k 12k 13k X1t −k 1 1t 21k 22k 23k X 2t −k + 2 ECM t −1 + 2t (12) X 2t = C2 + X 3t C3 k =1 31k 32k 33k X 3t −k 3 3t The C ' s , ' s , ' s are the parameters to be estimated. ECM t −1 is the one period lagged error terms derived from the Equation-3, Equation-4, Equation-7, and Equation-8. ' s are serially independent with mean zero and finite covariance matrix. The appropriate lag length for Equation-12 has been selected by the AIC and SBIC. 3.4. Fully Modified Ordinary Least Square (FMOLS) FMOLS considers endogeneity and serial correlation problems. Most of the macroeconomic variables cause each other even if we assume that interest rate, per capita GDP, and foreign exchange rate are strictly exogenous. Moreover, we assume that only interest rate causes domestic investment. Sometimes, domestic investment causes interest rate. Hence by default, the simultaneity might exist. Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 103 This problem is also known as reverse causality. We could have identified it by designing equation in simultaneous form and using formal test. Simulation experience also suggests that FMOLS estimator performs very well in relation to other methods of estimating cointegrating relations (Cappuccio and Lubian, 1992; Hansen and Phillips, 1991; Hargreaves, 1993; Phillips and Loretan, 1991; and Rau, 1992). 4. Results and Interpretation 4.1 Descriptive Statistics A few descriptive statistics of PGDP, DINV, REM, INT, and FOREX are given in Table-2. Table 2: Descriptive Statistics Name of variables Mean Std. Dev. CV (%) REM (%) 4.5211 2.9762 65.83 DINV (%) 22.8317 5.1408 22.52 FOREX (BDT/USD) 45.3676 21.0215 46.33 INT (%) 13.0318 1.6155 12.39 PGDP (in USD) 424.267 262.9987 62.11 JB Statistics 16.3165*** (0.0003) 12.9315*** (0.0016) 2.6636 (0.2640) 2.0562 (0.3577) 15.8728*** (0.0004) Note: ***P<0.01 denotes significant at 1% level, **P<0.05 denotes significant at 5% level, *P<0.10 denotes significant at 10% level. The distributions of FOREX and INT are normal unlike those of REM, DINV, and PGDP. Remittances have the highest variability. 4.2 Causality Results From the estimated results of VEC models, short-run unidirectional causality exists from domestic investment to domestic interest rates (ln DINV ln INT ) , from remittances to foreign exchange rate (BDT/USD) (ln REM ln FOREX ) , from foreign exchange rate (BDT/USD) to per capita GDP (ln FOREX ln PGDP) , from domestic investment to per capita GDP (ln DINV ln PGDP) , from foreign exchange rate (BDT/USD) to domestic investment (ln FOREX ln DINV ) , and from foreign exchange rate (BDT/USD) to 104 Bank Parikrama domestic interest rate (ln FOREX ln INT ) . The significance of ECM (−1) is each case confirms the existence of long-run causality among the variables. In the long- run, remittances along with domestic interest rate, per capita GDP along with foreign exchange rate, domestic interest rate along with per capita GDP, and domestic interest rate long with foreign exchange rate cause domestic investment, remittances, domestic investment, and domestic investment respectively. The results of Granger Causality Test have been provided in Table-5. 4.3 Long-run Analysis From the estimated long-run coefficients of Equation-3, it can be said that remittances have a significant positive impact on domestic investment (Gupta et al., 2007; Adams, 2006; Mundaca, 2009) and domestic interest rate has a significant negative impact on domestic investment (Hall, 1977; Greene and Villanueva, 1990; Aysan et al., 2005; Wang and Yu, 2007; Osundina and Osundina, 2014; Wuhan et al., 2015). From the estimated long-run coefficients of Equation-4, it can also be said that both the foreign exchange rate (BDT/USD) and per capita GDP have a significant positive impact on remittances (Elbadawi and Rocha, 1992; Faini, 1994; Freund and Spatafora, 2005; Rajan and Subramanian, 2005; Schiopu and Seigfried, 2006; Barua et al., 2007; Hasan, 2008; Barajas et al., 2010; Pant and Budha, 2016). From the estimated long-run coefficients of Equation-7, it has been found that per capita GDP has a significant positive impact on domestic investment (Acosta and Loza, 2004; Nghifenwa, 2009; Karagoz, 2010; Ajide and Lawanson, 2012; Khatib et al., 2012; Sohail et al., 2014) and domestic interest rate has significant negative impact on domestic investment. From estimated long-run coefficients of Equation-8, foreign exchange rate has a significant positive impact on domestic investment (Campa and Goldberg, 1999; Chvali, 2014) and interest rate has a significant negative impact on domestic investment. The results have been provided in Table-6. Since per capita GDP and foreign exchange rate have a significant positive impact on remittances, here the investment motive is dominant. When the income of the family in the Bangladesh increases and Bangladeshi currency depreciates, the migrant sends more money for financial investments or for inheritance reasons, because Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 105 his potential of inheritance will increase. In the case of insurance motivation, depreciation in Bangladesh currency will also decrease the remittances, as the migrant prefers to take time to remit for offsetting the impact of the appreciation of the local currency. If altruistic motive dominates, and income of the family in the Bangladesh decreases, the migrant sends more money in order to assure the same level of the utility for his family and the depreciation in Bangladeshi currency avail less remittances since it will ensure same level of national currency with relatively small amount of foreign currency. Hence, the study has not found any dominance altruistic and insurance motives of remittances. The investment motive is dominant has been also proved by a significant positive impact of remittances on domestic investment. It can also be said that remittances change the savings pattern of the households. Hence, a part of the households’ savings is going to the domestic investment. Therefore, over time depreciation in Bangladeshi currency and an increase in per capita GDP increase remittances, which subsequently increase domestic investment. It can also be said that to support the faster economic growth, it is inevitable to increase the level of domestic investment. Over time, an increase in the domestic interest rate squeezes down the level of domestic investment as the higher cost of funding makes investors more risk-averse ultimately lowering the level of domestic investment. The foreign exchange rate has a significant positive impact on domestic investment, depending on the country specific factors (Darby et al., 2000) such as trade openness, cost of capital goods, higher capital mobility, and external debt apart from the investment motive. For example, depreciation in domestic currency has a positive impact on domestic investment as the import quantity of capital products is larger relative to the extent of capital mobility (Serven, 1990). 4.4 Short-run Analysis Finally, the short-run equations have been estimated in the following formsp q r j =1 l =0 k =0 ln DINVt = 0 + j ln DINVt − j + 1l ln REM t −l + 2 k ln INTt −k + 1ECM1,t −1 + t .....(13) m n s g =1 h =0 i =0 ln REM t = 0 + g ln REM t − g + 1h ln FOREX t −h + 2i ln PGDPt −i + 2 ECM 2,t −1 + t ..... (14) 106 Bank Parikrama u v w j =1 m=0 n =0 ln DINVt = 0 + j ln DINVt − j + 1m ln PGDPt −m + 2 n ln INTt −n + 3 ECM 3,t −1 + t ...(15) q r s j =1 k =0 l =0 ln DINVt = 0 + j ln DINVt − j + 1k ln FOREX t −k + 2l ln INTt −l + 4 ECM 4,t −1 + t ...(16) Equation-13, Equation-14, Equation-15, and Equation-16 have been estimated by OLS. 0 , 0 , 0 , 0 , ' s , ' s , ' s , ' s 1 's , 1 ' s , 1 ' s , 1 's , 2 ' s , 2 's , 2 's , and 2 ' s are the parameters to be estimated. 1 , 2 , 3 , and 4 denote the speed of adjustments to approach into the long-run equilibrium if there is any shock in the systems. From the estimated results of short-run Equation-13, it can be said that remittances have significant positive impact on domestic investment and the domestic interest rate has a significant negative impact on domestic investment. Therefore, in the short-run, an increase in the domestic interest rate slows down the vehicle of domestic investment and an increase in remittances keeps the vehicle of domestic investment going. The coefficient of error correction term for Equation-13 is significant with an expected negative sign. If there is any shock to domestic investment due changes in remittances and domestic interest rates, it will adjust 74.55% in the first year. The entire convergence process will take approximately 1.3 years to approach into the long-run equilibrium, if there is any shock to the domestic investment. Therefore, the convergence process is significantly faster, if there is any shock to domestic investment. From the estimated results of short-run Equation-14, it can be said that both foreign exchange rate (BDT/USD) and per capita GDP have a positive impact on remittances, even the impact of per capita GDP is insignificant. Therefore, in the short-run the increase in foreign exchange rate and per capita GDP boosts up remittances. To generalize, an appreciation in the foreign exchange rate (BDT/USD) and more per capita GDP scales up remittances, which subsequently positively contributes to domestic investment in the short-run. The coefficient error correction term for Equation-14 is significant with the expected negative sign. If there is any shock to the remittances, due to changes in foreign exchange rate and per capita GDP, it will adjust 24.02% in the first year. The entire Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 107 convergence process will take approximately 4.16 years to approach the long-run equilibrium, if there is any shock to the remittances due to changes in the foreign exchange rate (BDT/USD) and per capita GDP. From the estimated results of short-run Equation-15, it has been found that per capita GDP has an insignificant positive impact on domestic investment, and the domestic interest rate has a significant negative impact on domestic investment. The error correction term for Equation-15 is significant with expected negative sign. If there is any shock to domestic investment due to changes in per capita GDP and the domestic interest rate, it will adjust 97.35% in the first year. The entire convergence process will take approximately 1 year to approach equilibrium if there is any shock to domestic investment due to changes in per capita GDP and the domestic interest rate. From the estimated results of short-run Equation-16, it has been found that the foreign exchange rate (BDT/USD) has an insignificant positive impact on domestic investment and the domestic interest rate has an insignificant negative impact on domestic investment. The error correction term of Equation-16 is significant with the expected negative sign. If there is any shock to domestic investment due to changes in the foreign exchange rate (BDT/USD) and domestic interest rate, it will adjust 48% in the first year. The entire convergence process will take approximately 2 years to approach the equilibrium level if there is any shock to domestic investment due to changes in the foreign exchange rate (BDT/USD) and domestic interest rate. The results have been provided in Appendix Table-4. 4.6 Stability of the Parameters Due to the use of FMOLS to estimate the long-run equations (Equation-3, Equation-4, Equation-7, and Equation-8), stability checking of the long-run parameters has been avoided. Stability of the short-run parameters has been checked by CUSUM test and CUSUMSQ test (Borensztein, De Gregorio, and Lee., 1998). The short-run parameters are stable for all equations (Equation-1e, Equation-14, Equation-15, and Equation-16) since all parameter values lie within the critical bounds during the estimation period. Therefore, short-run parameters along with long-run parameters can be used for conclusions and policy implications. The results have been provided in Appendix Figure-1. 108 Bank Parikrama 5. Conclusion and Policy Implication The study exposes a significant positive impact of remittances unlike domestic interest rate on domestic investment and a significant positive impact of the foreign exchange rate (BDT/USD) and per capita GDP on remittances in the long-run. In addition, in the long-run, per capita GDP and the foreign exchange rate (BDT/USD) have a significant positive impact on domestic investment. Over time, depreciation of Bangladeshi currency and increase in per capita GDP increase remittances, which consequently increase domestic investment. Hence the investment motive is dominant in the long-run. Again, per capita GDP has an insignificant positive impact on domestic investment, and foreign exchange rate has a significant positive impact on remittances. It can be claimed that even in the short-run, investment motive of remittances is dominant due to significant positive impact of remittances on domestic investment, significant impact of foreign exchange rate on remittances, and insignificant positive impact of per capita GDP on remittances. Hence, this study suggests that to boost up remittance, depreciation of domestic currency should be in a convenient level for the greater interest of the economy. Although movement of domestic currency is in floating regime, embargoing on the import of luxury items, restricting money laundering, and promoting foreign investment, the increased demand of foreign currency can be restricted. Besides, remittances have a significant positive impact on domestic investment and domestic interest rate has a significant negative impact on domestic investment in the short-run. Since, the main objective of remittances is to change the socioeconomic structure (developing living standard, construction and development of schools, colleges, health centers, water supply and sanitation, and rural electrification, etc.), the level of domestic investment will be increased by remittances. Contribution of domestic investment is not satisfactory, even if inflation and real lending interest rate are in decreasing trend. Given many factors affecting lending interest rate of an economy, our motive is to tell that increased import cost, theoretically a contributing factor, might affect inflation rate and thereby to the lending interest rate of Bangladesh. Therefore, the double-digit interest rate affects negatively to domestic investment and our lending interest Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 109 rate is not lower enough to increase domestic investment. Hence, boosting up domestic investment via remittances would contribute into economic growth. If there is any shock to domestic investment due to changes in remittances and domestic interest rates, it will correct 74.55% in the first year. The entire convergence process will take approximately 1.3 years to approach into long-run equilibrium. If there is any shock to the remittances due to changes in the foreign exchange rate and per capita GDP, it will adjust 24.02% in the first year. The entire convergence process will take approximately 4.16 years to approach into long-run equilibrium. If there is any shock to domestic investment due to changes in per capita GDP and domestic interest rates, it will adjust 97.35% in the first year. The entire convergence process will take approximately 1 year to approach into long-run equilibrium. If there is any shock to domestic investment due to changes in foreign exchange rates (BDT/USD) and domestic interest rates, it will adjust 48% in the first year. The entire convergence process will take approximately 2 years to approach into long-run equilibrium. All short-run parameters are stable suggested by CUSUM test and CUSUMSQ test. Therefore, all the long-run and short-run coefficients can be used for policy implications. Finding the convenient level of the exchange rate and domestic lending interest rate for the greater interest of the economy could be the further research opportunity. Besides, the time variant parametric model could be used to consider structural change of the Bangladeshi economy. End Notes a) Per capita GDP (PGDP) and foreign exchange rate (FOREX) are used separately in two equations instead of one equation, due to the multicollinearity problem (See: Equation-5 and Equation-6). b) Interest rate, domestic lending interest rate, and domestic interest rate have been used interchangeably in this study. c) All short-run and long-run parameters have carefully been estimated in Eviews and RATS. d) All short-run parameters are stable suggested by CUSUM test and CUSUMSQ test. 110 Bank Parikrama e) Both the foreign exchange rate (BDT/USD) and per capita GDP have a significant positive impact on remittances in the long-run, which will positively contribute to domestic investment (Remittances have a significant positive impact on domestic investment). Therefore, it can be concluded that foreign exchange rates and per capita GDP have indirect positive impacts on domestic investment. f) All short-run error correction models are free from auto-regressive conditional heteroscedasticity, autocorrelation, and functional misspecification. The distribution of error terms in all error correction models are normal. g) AIC stands for Akaike information criterion and SBIC stands for Schwartz Bayesian information criterion. References Acosta, P. & Loza, A. (2004). Short and long-run determinants of private investment in argentina. Journal of Applied Economics, 8(2), 389-406. Adams, R. J. (2006). International remittances and the household: analysis and review of global evidence. Journal of African Economies, 15(2), 396- 425. Ajide, K. B. & Lawanson, O. (2012). Modeling the long-run determinants of domestic private investment in nigeria. Asian Social Science, 8(13), 139-152. Akter, S. (2016). Remittance inflows and its contribution to the economic growth of bangladesh. 現代社会文化研究 , 62, 215-245. Aysan, A., Gaobo, P. & Marie-Ange, V. (2005). How to boost private investment in the mena countries: the role of economic reforms. Topics in middle eastern and north african economics. MEEA, Online Journal, 7, 1-15, Retrieved April 25, 2017, from http://www.luc.edu/orgs/meea/volume7 /Aysan.pdf. Barajas, A., Chami, R., Hakura, D. & Montiel, P. (2010). Workers’ remittances and the equilibrium real exchange rate: theory and evidence. Economia Journal of the Latin American and Caribbean Economic Association, Spring, 20, 45-99. Barua, S., Majumder, M. A. & Akhtaruzzaman, D. M. (2007). Determinants of workers' remittances in bangladesh: an empirical study. MPRA Paper No. 15080. Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 111 Borensztein, E., Gregorio, J. De & Lee, J.W. (1998). How does fdi affect economic growth?. Journal of International Economics, 45(1), 115-135. Bouhga-Hagbe, J. (2004). A theory of workers remittances with an application to morocco. International Monetary Fund, Working paper series WP/04/194. Buffie, E. (1986). Devaluation, investment and growth in ldcs. Journal of Developed Economics, 20, 361-380. Campa, J. & Goldberg, L. S. G. (1999). Investment, pass-through, and exchange rates: a cross-country comparison. International Economic Review, 40, 287-314. Cappuccio, N. & Lubian, D. (1992). The relationships among some estimators of the cointegrating coefficient: theory and monte carlo evidence. University of Padova. Chami, R., Fullenkamp, C. & Jahjah, S. (2003). Are immigrant remittance flows a source of capital for development?. International Monetary Fund, Working Paper series WP/03/189. Chamon, M., Semblat, R. & Morant, A. (2005). Samoa: selected issues and statistical appendix. International Monetary Fund, Country Report No. 05/221. Chavali, S. A. (2014). Investment, exchange rates and relative prices: evidence from emerging economies. Doctoral Dissertation:University of Glasgow, UK. Connell, J. & Conway, D. (2000). Migration and remittances in island microstates: a comparative perspective on the south pacific and the caribbean. International Journal of Urban and Regional Research, 24, 52–78. Darby, J., Hallett, A. H., Ireland, J. & Piscitelli, L. (2000). Exchange rate uncertainty and business sector investment. Contributed Papers-0600. Econometric Society World Congress 2000 Demers, C., Giroux, N. & Chreim, S. (2003). Merger and acquisition announcements as corporate wedding narratives. Journal of Organizational Change Management, 16(2), 223-242. Dickey, D. A & Wayne, A. F. (1979). Distribution of estimators for autoregressive time series with a unit root. Journal of American Statistics Association, 74, 427-431. 112 Bank Parikrama Elbadawi, I. A. & Rocha, R. (1992). Determinants of expatriate workers’ remittances in north africa and europe. Working Papers 1038, 1-56, World Bank. El-Sakka, M. I. T. & Mcnabb, R. (1999). the macroeconomic determinants of emigrant remittances. World Development, 27(8), 1493-1502. Engle, R. F. & Granger, C. W. J. (1987). Cointegration and error correction: representation, estimation and testing. Econometrica, 55(2), 251-276. Faini, R. (1994). Workers remittances and the real exchange rate: a quantitative framework. Journal of Population Economics, 7(2), 235-245. Faini, R. (2002). Migration, remittances, and growth. Development Economics and Policy, Springer: New York, NY, USA. 171–187. Feeny, S., Iamsiraroj, S. & Mcgillivray, M. (2014), Remittances and economic growth. larger impact in smaller countries?. Journal of Development Studies, 50, 1055–1066. Freund, C. & Spatafora, N. (2005). Remittances: transaction costs, determinants and informal flows. Working Paper No. WPS 3704, International Monetary Fund. Giuliano, P. & Ruiz-Arranz, M. (2006). Remittances, financial development and growth. Journal of Development Economics, 90(1), 144-152. Greene, J. & Villanueva, D. (1990). Determinants of private investment in ldcs. Finance and Development, 27(4), 40, December. Washington D.C. Gupta, S., Catherine, P. & Smita, W. (2007). Impact of remittances on poverty and financial development in sub-saharan africa. Working paper series WP/07/38, International Monetary Fund. Harupara, G. E. (1998). Macroeconomic determinants of private investment in namibia. Master Dissertation, Addis Ababa University, Addis Ababa. Hasan, M. M. (2008). The macroeconomic determinants of remittances in bangladesh. MPRA Paper No. 27744, Retrieved on June 16, 2017, Available at: http://mpra.ub.uni-muenchen.de/27744. Hall, E. R. (1977). Investment, interest rates and the effects of stabilization policies. MIT. Brookings Papers on Economic Activity, 1, 61-103. Hansen, B.E. & Phillips, P. C. B. (1991). Estimation and inference in models of cointegration: a simulation study. Advances in Econometrics, 8, 225-248. Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 113 Hargreaves, C.H. (1993). A review of methods of estimating cointegrating relationships. Oxford University Press: Oxford. Imai, K. S., Gaiha, R., Ali, A. & Kaicker, N. (2014). Remittances, growth and poverty: new evidence from asian countries. Journal of Policy Modelling, 36, 524–538. Jouini, J. (2015). Economic growth and remittances in tunisia: bi-directional causal links. Journal of Policy Modelling, 37, 355–373. Johansen S. & Juselius, K. (1990). Maximum likelihood estimation and inference on cointegration-with application to the demand for money. Oxford Bulletin of Economics and Statistics, 52, 169-210. Karagoz, K. (2010). Determining factors of private investment: an empirical analysis for turkey. Sosyoekonomi, 11, 1-20. Khatib, H. B., Altaleb, G. S. & Alokor, S. M. (2012). Economical determinants of domestic invetment. European Scientific Journal, 8(7), 1-17. Larsen, E. J. (2004). The impact of loan rates on direct real estate investment holding period return. Financial Services Review, 13, 111-121. Lim, S. & Simmons, W. O. (2015). Do remittances promote economic growth in the caribbean community and common market?. Journal of Economics and Business, 77, 42–59. Matuzeviciute, K. & Butkus, M. (2016). Remittances, development level, and long-run economic growth. Economics, 4(4), 2-20. MRF (2016). Migrations and remittances factbook. 3rd Edition. World Bank. Retrieved on April 25, 2017, Available at: https://openknowledge. worldbank. org/bitstream/handle/10986/23743/9781464803192.pdf Mundaca, B. G. (2009). Remittances, financial markets development and economic growth: the case of latin amercia and carribean. Review of Development Economics, 13(2), 288-303. Narayan, P.K., Narayan, S. & Mishra, S. (2011). Do remittances induce inflation? fresh evidence from developing countries. Southern Economic Journal, 77, 914– 933. 114 Bank Parikrama Nghifenwa, N. F. (2009). Factors of influen cing investment: a case study of the namibian economy. Master Thesis: University of Namibia, Namibia. Nwaogu, U.G. & Ryan, R. J. (2015). FDI, foreign aid, remittance and economic growth in developing countries. Review of Development Economics, Vol. 19, pp. 100–115. Osundina, A. J. & Osundina, K. C. (2016). Stock market development and welfare in nigeria: a vecm approach. Research Journal of Accounting and Finance, Vol. 7 No. 16, pp. 145-150. Pant, B. & Budha, B. B. (2016). Remittances and exchange rate linkages experiences of nepal. Working Paper No. 33, Nepal Rastra Bank. Phillips, P. (1986). Understanding spurious regressions in econometrics. Journal of Econometrics, 33(3), 311-340. Phillips, P.C.B. & Hansen, B. (1990). Statistical Inference in Instrumental Variables Regression with I (1) Processes. The Review of Economic Studies, 57, 99-125. Phillips, P.C.B. & Loretan, M. (1991). Estimating long-run equlibria. Review of Economic Studies, 59, 407-436 Phillips P.C.B. & Perron, P. (1987). Testing for a unit root in time series regression. Biometrika, 75, 335-346. Pradhan, G., Upadhyay, M. & Upadhyaya, K. (2008). Remittances and economic growth in developing countries. European Journal of Devlopment, 20, 497–506. Rajan, R. & Subramanian, A. (2005). What undermines aid’s impact on growth?”. Working Paper 1167, National Bureau of Economic Research, 1050 Massachusetts Avenue. Cambridge, MA 02138. Rapoport, H. & Docquier, F. (2006). The economics of migrants remittances. Handbook on the Economics of Reciprocity. Giving and Altruism, Elsevier. Rau, H.H. (1992). Estimation and inference of linear regression models with conintegrated regressors. Yale University Doctoral Dissertation. Russell, S. S. (1986). Remittances from international migration: a review in perspective. World Development, 14, 677-96. Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus 115 Schiopu, L. & Siegfried, N. (2006). Determinan ts of workers’ remittances: evidence from the european neighboring region. European Central Bank Staff Working Paper 688, European Central Bank. Serven, L. (1990). Anticipated real exchange rate changes and the dynamics of investment. Working Paper No. 562, World Bank Policy Research. Sohail, A., Rehman, U. & Azeem, M. (2014). Economic determinants of domestic investment: a case of pakistan. Global Journal of Management and Business Research: B. Economics and Commerce, 14(7), 41-44. Straubbarr, T. (1986). The determinants of workers remittances: issues and prospects. World Bank Staff Working Paper 481, Washington: World Bank. Swamy, G. (1981). International migrant workers’ remittances: issues and prospects. World Bank Staff Working Papers 481, Washington: World Bank. Tahir, M., Khan, I. & Shah, A. M. (2015). Foreign remittances, foreign direct investment. Foreign Imports, and Economic Growth in Pakistan: A Time Series Analysis. Arab Economics and Business Journal, 10, 82–89. Taylor, J.E. & Wyatt, T. J. (1996). The shadow value of migrant remittances, income, and inequality in a household-farm economy. Journal of Development Studies, 32, 899–912. Wahba, S. (1991). What determines workers’ remittances?. Finance and Development, 28(4), 41-44. Wang, D. H & Yu, H. (2007). The Role of Interest Rate in Investment Decisions: A Fuzzy Logic Framework. Global Business and Economic Review, 9(4), 448457. Wuhan, Suyuan, L. & Khurshid, A. (2015). The effect of interest rate on investment: empirical evidence from jiangshu province, china. Journal of International Studies, 8(1), 81-90. Ziesemer, T. (2006). Worker remittances and growth: the physical and human capital channel. Merit Working Paper Series 20, United Nations University Zizi, G. (2014). Remittances as an economic development factor. Empirical Evidence from the CEE Countries. Procedia Economics and Finance, 10, 54-60. 116 Bank Parikrama Appendices Appendix Table 1: Unit Root Tests Results Variables ln DINV Model with constant term [Level Form] ADF test P-value PP test -0.1829 0.9321 -0.7991 P-value 0.8082 ln REM -1.7274 0.4078 -2.8441* 0.0614 ln FOREX ln INT ln PGDP -1.8672 0.3438 -1.7814 0.3839 -2.2136 0.2050 -1.8681 0.3434 0.6453 0.9892 0.6000 0.9880 Variables ln DINV ln REM ln FOREX ln INT ln PGDP Model with constant and trend term [Level Form] ADF test P-value PP test -1.5691 0.7856 -3.0125 P-value 0.1419 -2.4400 -1.8550 0.3532 0.6578 3.0797 -1.0302 0.9992 0.9278 -2.1203 0.5184 -1.7659 0.7018 -1.0362 0.9268 -1.0362 0.9268 Variables Model with constant term [Difference Form] ADF test P-value PP test -9.1055*** 0.0000 -9.1165*** P-value 0.0000 ln REM -8.4449*** ln DINV ln FOREX ln INT ln PGDP Variables ln DINV ln REM ln FOREX ln INT ln PGDP ***P<0.01 0.0000 -8.4449*** 0.0000 *** -4.2548 -4.1663*** 0.0019 0.0023 *** -3.7157 -3.9613*** 0.0077 0.0040 -5.3707*** 0.0001 -5.3707*** 0.0001 Model with constant and trend term [Difference Form] ADF test P-value PP test -9.0654*** 0.0000 -9.0654*** -8.1156*** ** -3.9958 -4.2945*** -5.3108*** P-value 0.0000 0.0000 -8.2149*** 0.0000 0.0178 0.0083 *** -5.6653 -4.0969** 0.0000 0.0135 0.0005 -5.3108*** **P<0.05 Note: denotes significant at 1% level, denotes significant at 5% level, significant at 10% level. Appropriate lag length for these tests has been selected by SBIC. Source: 0.0005 *P<0.10 denotes CEs None* At most 1 At most 2 Case-1: Intercept (no trend) in CE and VAR Equation (3) Equation (4) Equation (7) Trace Max-Eigen Trace Max-Eigen Trace Statistic Max-Eigen Statistic Statistic Statistic Statistic Statistic 30.4262** 21.5510** 33.1539** 25.0450** 40.3123*** 36.0620*** [29.7971] [21.1316] [29.7971] [21.1316] [29.7971] [21.1316] (0.0423) (0.0437) (0.0198) (0.0133) (0.0022) (0.0002) 8.8751 8.7452 8.1089 6.5262 4.2503 4.2495 [15.4947] [14.2626] [15.4947] [14.2646] [15.4947] [14.2646] (0.3771) (0.3080) (0.4538) (0.5466) (0.8823) (0.8321) 0.1299 0.1299 1.5827 1.5827 0.0008 0.0008 [3.8415] [3.8415] [3.8415] [3.8415] [3.8415] [3.8415] (0.7185) (0.7185) (0.2084) (0.2084) (0.9784) (0.9784) Case-2: Intercept and trend in CE and no intercept in VAR Equation (3) Equation (4) Equation (7) Trace Max-Eigen Trace Statistic Max-Eigen Trace Max-Eigen Statistic Statistic Statistic Statistic Statistic Equation (8) Trace Max-Eigen Statistic Statistic 26.0982 19.2480* [29.7971] [21.1316] (0.1258) (0.0899) 6.8501 6.7527 [15.4947] [14.2646] (0.5951) (0.5188) 0.0974 0.0974 [3.8415] [3.8415] (0.7549) (0.7549) Equation (8) Trace MaxStatistic Eigen Statistic None* 44.5900** 21.1496 56.2666*** 29.0742** 47.9449** 36.3956*** 39.9019* 21.1010 [42.9153] [25.8232] [42.9153] [25.8232] [42.9153] [25.8232] [42.9152] [25.8232] (0.0336) (0.1421) (0.0014) (0.0180) (0.0145) (0.0014) (0.0970) (0.1861) At 22.4404 17.2434* 27.1923** 22.1210** 11.5493 8.7520 18.8009 12.7137 most [25.8721] [19.3870] [25.8721] [19.3870] [25.8721] [19.3870] [25.8721] [19.3870] 1* (0.1261) (0.0997) (0.0341) (0.0195) (0.8420) (0.7497) (0.2927) (0.3517) At 5.1970 5.1970 5.0714 5.0714 2.7973 2.7973 6.0871 6.0871 most 2 [12.5180] [12.5180] [12.5180] [12.5180] [12.5180] [12.5180] [12.5179] [12.5179] (0.5683) (0.5683) (0.5862) (0.5862) (0.8997) (0.8997) (0.4493) (0.4493) Note: Value under [ ] represents 5% critical value and value under ( ) represents p-value. ***P<0.01 denotes significant at 1% level, **P<0.05 denotes significant at 5% level, *P<0.10 denotes significant at 10% level. Appropriate lag length for this test has been selected by AIC and SBIC. Case-1 suggests one cointegrating equation and Case-2 suggests two cointegrating equations. CEs Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus Appendix Table 2: Cointegration Test Results 117 118 Bank Parikrama Appendix Table 3: Causality Analysis Results Equation (3) Causalities: Short-run ln DINV ln REM ln INT ln DINV 0.000007 (0.9979) 5.1705** (0.0296) ln REM ECM (−1) [ t-statistic] 0.4791 (0.4937) - 0.3856 (0.5389) 0.1402 (0.7104) - -1.8656* (0.0710) -0.7599 (0.4527) -2.8728*** (0.0071) 0.2662 (0.6093) Equation (4) Causalities: Short-run ln REM ln FOREX ln PGDP ln REM ln FOREX - 0.8576 (0.5052) - 4.4442*** (0.0093) 0.7334 (0.5794) 4.7784*** (0.0067) 0.1785 (0.9470) 0.4159 (0.7953) - ln DNIV ln PGDP ln INT ln DNIV - ln PGDP 3.4123* (0.0737) 2.4315 (0.1285) 0.0775 (0.7824) - 0.0367 (0.8492) 1.3365 (0.2560) - 0.4139 (0.5244) Equation (8) Causalities: Short-run ln DNIV ln FOREX ln INT ln DNIV ln FOREX ln INT - 6.8132** (0.0135) - 0.1454 (0.7054) 1.2142 (0.2785) - 0.0742 (0.7870) 6.3758** (0.0166) 6.4679** (0.0159) Causality: Long-run ln PGDP ECM (−1) [ t-statistic] Equation (7) Causalities: Short-run ln INT Causality: Long-run ln INT -1.7572* (0.0935) -1.3985 (0.1766) 2.4263** (0.0243) Causality: Long-run ECM (−1) [ t-statistic] -3.7947*** (0.0006) -3.3465*** (0.0021) -1.9711* (0.0571) Causality: Long-run ECM (−1) [ t-statistic] -4.4095*** (0.0000) 1.6344 (0.1117) -3.3686*** (0.0019) Note: ***P<0.01 denotes significant at 1% level, **P<0.05 denotes significant at 5% level, *P<0.10 denotes significant at 10% level. Significance of ECM (−1) has ensured the existence of long-run causality among the variables. Variables Equation (3) ln REM ln INT 0.1211*** (0.0000) -0.6540*** (0.0001) - Long-run Coefficients Equation Equation (4) (7) Equation (13) Equation (16) - - - - - - - -0.6120*** (0.0000) - -0.9495*** (0.0001) 0.3557*** (0.0000) - - - - - - - - - - - - 0.2253* (0.0705) -0.6258** (0.0492) - - - - - -0.4142* (0.0992) - -0.4520 (0.1789) 0.2004 (0.6335) - ln PGDP - ln REM - ln INT - ln FOREX 0.3322*** (0.0000) - - - - - - - - ln PGDP - - - - - ECM(-1) - - - - Constant 2.2231*** (0.0000) - 7.1157*** (0.0000) - 2.7257*** (0.0000) - 4.2042*** (0.0000) - ARCH Test - - - - Normality Test - - - - Misspecification Test - - - - -0.7455*** (0.0007) -0.0087 (0.7465) 0.1898 (0.8281) 0.5096 (0.7668) 3.3282 (0.1894) 0.0884 (0.7680) Auto-correlation Test Short-run Coefficients Equation Equation (14) (15) - 2.1843*** (0.0000) 1.6203*** (0.0000) - ln FOREX Equation (8) 1.3697** (0.0337) 0.5983 (0.3077) -0.2402** (0.0370) -0.0292 (0.6996) 0.2623 (0.7710) 1.2543 (0.3067) 3.3133 (0.1899) 1.3106 (0.2605) 0.0751 (0.7254) -0.9735*** (0.0000) 0.0102 (0.5830) 0.5546 (0.5797) 0.1382 (0.9365) 0.3535 (0.8380) 1.7748 (0.1919) -0.4800*** (0.0099) 0.0242 (0.3308) 0.8052 (0.4558) 0.3986 (0.7548) 5.7729* (0.0558) 0.1781 (0.6758) Abedin, Sen, Chowduhury and Akter: Revisiting the Investment-Remittance Nexus Appendix Table 4: Long-run and Short-run Coefficients Note: ***P<0.01 denotes significant at 1% level, **P<0.05 denotes significant at 5% level, *P<0.10 denotes significant at 10% level. The Short-run Equation has been estimated by OLS and Long-run Equation has been estimated by FMOLS. The lead and lag in FMOLS have been selected by SBIC. 119 120 Appendix Figure 1: Stability of Short-run Parameters Figure 1(a): For Equation (13) Figure 1(b): For Equation (14) 20 1.4 20 1.4 15 1.2 15 1.2 1.0 10 0.8 0.8 5 1.0 10 5 0.6 0.6 0 0 0.4 0.4 -5 -5 0.2 0.2 -10 0.0 -15 -0.2 -20 -0.4 84 86 88 90 92 94 96 98 00 CUSUM 02 04 06 08 10 12 14 84 86 88 5% Significance 90 92 94 96 98 00 CUSUM of Squares 02 04 06 08 10 12 -10 0.0 -15 -0.2 -20 -0.4 86 14 88 90 92 94 96 98 CUSUM 5% Significance Figure 1(c): For Equation (15) 00 02 04 06 08 10 12 86 14 92 94 96 98 00 02 04 06 08 10 12 14 5% Significance Figure 1(d): For Equation (16) 1.4 20 1.4 15 1.2 15 1.2 10 1.0 1.0 10 0.8 0.8 5 0.6 0.6 0 90 CUSUM of Squares 20 5 88 5% Significance 0 0.4 0.4 -5 -5 0.2 0.2 -10 0.0 -15 -0.2 -20 1985 1990 1995 CUSUM 2000 2005 5% Significance 2010 2015 -10 0.0 -15 -0.2 -0.4 -20 -0.4 1985 1990 1995 CUSUM of Squares 2000 2005 5% Significance 2010 2015 1985 1990 1995 CUSUM 2000 2005 5% Significance 2010 2015 1985 1990 1995 CUSUM of Squares 2000 2005 2010 2015 5% Significance Bank Parikrama ISSN 1019-7044 sh na de ge m ent Bangla In sti tute of Bank M a Price: BDT 400.00 USD 20.00 Published by the Bangladesh Institute of Bank Management (BIBM) Main Road No. 1 (South), Section-2, Mirpur, Dhaka-1216, Bangladesh.