DOES A DEVALUATION OF THE REAL EXCHANGE RATE
3.3 Model Specification, Data and Variables
As a small open economy, import prices of Bangladesh are given in world market and the prices are independent of the volume of imports. Thus, the demand for imports is influenced basically by domestic real income and the real exchange rate (RER).On the other hand, demand for exports depends mainly on: (i) relative prices of competing goods from the competing countries and (ii) aggregate demand of importers. For example, if theRER increases, the exports commodities become cheaper to importers which may contribute to increasing the demand for export. Therefore, it can be anticipated that an increase in the real exchange rate would improve the balance of trade. The following Figure 3.3 demonstrates the likely effects of change in the real exchange rate and real income on exports, imports thereby the balance of trade in a small open economy. It is worth mentioning that standard small country assumption suggests an elastic export demand function for developing countries; however, since most least developed countries (LDCs) specialize in exporting few specific commodities in which they attain some market power, the export demand curve is rather a downward sloping curve for LDCs (see, for example, Sehadji, 1998; Faini, Clavijo and Senhadji, 1992). Since about 80 percent Bangladesh‘s exports is explained by only the RMG industry, we assume a downward sloping export demand curve for Bangladesh.
Figure 3.3: Impact of the exchange rate and domestic income on import and export demand of a small open economy.
(a) Increase in import prices due to increase in the REER which reduces import demand.
(b) Import demand rises due to the increase in the real domestic income.
PX SX0
SX1
DX
QX0 QX1 Export Demand
(c) Increase in export demand due to increase in REER
The study attempts to empirically estimate the ‗two-country model‘ of trade by Rose and Yellen (1989) and Rose (1991) which are applied by empirical literature in both the developing (see, for example, Singh, 2002; Arora, Bahmani-Oskooee and Goswami, 2003; Hsing, 2008) and developed (see, for example, Rose and Yellen, 1989; Rose 1991; Bahmani-Oskooee and Brooks, 1999) countries context. Rose (1991) is a popular model for estimation of the trade balance model because it includes all basic determents of trade
D1 D0 SX1 SX0 DM Price SM DM1 Import Demand D0 D1 Price DM0 Import Demand
balance in a simple framework. Most importantly, as our objective is to find the impact of the exchange rate on trade balance and as the exchange rate variable is one of the determinants in Rose and Yellen (1989) and Rose (1991) models, we find this model best suit our purpose.
The theoretical basis of our empirical model can be given as follows:
The quantity demand for domestic import basically depends on domestic income and relative prices of import.
RP Y
f Md m, (3.3) where, * * * * * RP RER P P P P E P P E RP X X m , M is quantity demand for d
foreign goods, RP relative prices of import, m Y domestic income, E nominal exchange
rate, *
x
P foreign price for domestic imports, RER real exchange rate, RP foreign relative *
prices, and Pand P are domestic and foreign prices, respectively. *
Similarly, the demand for export can be defined as:
* *
* f RP ,Y Md m (3.4) where, RER RP RP Xm* , Md* is foreign demand for imports, RPm* relative prices of imports
abroad, Y foreign income and *
x
RP domestic relative prices of export.
Hence, the quantity of supply of export mainly depends on relative prices of exports.
X
S f RP X (3.5)
* * X S f RP X (3.6)The equilibrium is determined by demand and supply, that is, Md Xs*, and Md* Xs.
Thus, real balance of trade in local currency terms can be written as:
x d
d
x M RER RP M
RP
B * * (3.7)
where,Bis explained as the domestic net export in local currency. Equation (3.7) can be express as the following functional form:
RER, YY, *
f
B (3.8)
It is worth noting that relative prices of export and import are implied in the real exchange rate
RER .
A log-linear time series specification of the model (the testable model for this study) can be stated as follows:
t t t
t
t InRER InY InY
InB *
3 2
1
0 (3.9)
where,lnBt, lnXt, lnMt imply logarithm of balance of trade (lnXt lnMt), export and
import at time t , respectively. lnRERt, lnYt, and lnYtare the logarithms of the real
effective exchange rate, industrial product index of Bangladesh and trade weighted real GDP of trade partners. Exports, imports, industrial production index (proxy for domestic income/production) data come from the ‗World Development Indicators (Edition December 2010)‘ of the World Bank. The study constructs the trade weighted foreign income variable. We use the same trade weights for foreign income variable that we have used to construct the real effective exchange rate variable. The GDP data of trade partners‘ are also collected from the ‗World Development Indicators (Edition December 2010)‘ of the World Bank. All data for Bangladesh are in local currency term. The study uses annual data from 1976-2009 for Bangladesh. It is worth mentioning that we use annual data because quarterly or monthly data of some relevant variables are not available in the
existing data sources. The study uses industrial production index as a proxy of domestic income (production) because the ‗Trade Policy Review Report‘ (2006) of Bangladesh government (August 2006), the ‗Financial Sector Review (2006)‘, Bayes, Hossain and Rahman (1995), and Hossain and Alauddin (2005) suggest that since the 1980s the trade and the exchange rate policies of Bangladesh were driven mainly by the ‗export-led growth‘ and the ‗imports substitution‘ targets. Bangladesh imports mainly industrial products. Hence, by employing the industrial production index as one of the determinants of trade balance, we examine, whether industrial production significantly influence the trade balance of Bangladesh. It is worth noting that if a significant level of imports substitutions occurs, we may expect a positive sign for the coefficient of industrial production index in the trade balance model. This would suggest that imports substitutions significantly reduce imports demand of the country.
It is also worth noting that Bangladesh has been receiving substantial amount of ‗aid‘ and ‗remittance‘ from abroad. Hence, there may arise a question that why then this study has not included the aid and the remittance variables as the determinants of trade balance. This is because, firstly, ‗aid‘ is mainly a special or climactic issue for Bangladesh. When there is some large natural calamities (say, prolonged floods in 1988, 1998, 2007) which hit the economy substantially, the country receives major amounts of aid and concessionary finance from its donors, which comes as a large amount on the specific year (of disaster) and partly it is paid step-by-step later. Secondly, the aid and the remittance variables are found to be stationary at levels, i.e., I(0) series. Hence, we cannot include them with other non-stationary variables (i.e., I(1) series) in the cointegration framework.
As mentioned earlier, the theoretical notion suggests that exports and imports increase as the real income of the trade partners and domestic income rises respectively,
and vice versa. Hence, we could expect β2<0 and β3>0. On the contrary, imports may
decline as income increases if real income rises due to an increase in the production of import-substitute goods and in that case we would expect β2>0. The effect of movements
in the real effective exchange rate on the balance of trade
1 is however ambiguous. 1is the focus of this study and it could be positive or negative. Generally, if real devaluation takes place, exports increase, imports fall as a consequence and thus it improves the trade balance. In that case we can expect 1 0. However, if the J-curve pattern holds in the data, we would find that 1 0 in the short-run and1 0 in the long-run.