METHODOLOGY OF THE STUDY

DATA SOURCE

This study used various data sources. The main source of trade data is the United Nations
Commodity Trade Statistics Database (UN com-trade). World Development Indicators compact
disc read-only memory (CD-ROM), International Coffee Organization (ICO), Agriculture
Organization Statistics (UN-FAOSTAT) and national bank of Ethiopia relevant to the study. The
study covers the period from 1990 to 2021.

MODEL SPECIFICATION

The theoretical framework of this paper is following the work of Goldstein and Khan (1985).
The main characteristics of the imperfect substitute’s model can be summarized as follows.
Conventional demand theory postulates that, given a budgetary limitation, the customer will
maximize utility. The quantity demanded as a function of the level of (money) income in the
importing region, the price of the imported commodity itself, and the price of domestic
substitutes is thus represented by the resulting demand functions for imports and exports. Export
supply has traditionally been influenced by real export prices, real exchange rates, and
productive capacity, according to the imperfect substitute’s model. As a result, the export supply
function can be represented generally as follows:

EX= f(PC, REER, REP)………………………………………………………..1

Where Ex represents the volume of exports, REP is a real export price, REER is real exchange
rate, and PC is the capacity of production.

The purpose of the study is to carefully investigate international coffee market dynamics on the
export performance of Ethiopian coffee. Because it fits the model based on having the greatest
number of relevant variables impacting coffee export which is similar to Diriba (2021) in
assessing export performance sector in Ethiopia a log-linear function is utilized to determine
Ethiopia's total coffee export performance. Squeezing the estimated coefficients of each
regressor in the model, the total coffee export model used in Ethiopia captures the factors
influencing total coffee export by taking into account the impact of other stochastic factors not
controlled in the mode and by the model econometrically specified, as would be the case with
single equation 2 below.

XCt= α+ α1WPRt + α2RERt+ α3FDIt + α4DPDt+ α5WCPt+ α 6GDPt+ α7TOTt….2

Where:

XCt Is total coffee export at time t ;

WPRt Is annual average world coffee price at time t;

RERt Is real exchange rate at time t;

FDIt Is theFDI stock in Ethiopia at time t;

DPCt Is domestic production of coffee in year t;

WPCt Is world production in year t

GDPt Is the value of Ethiopia's GDP at current market prices at time t;

TOTt Is Term of trade at time t;

DEFINITION OF THE VARIABLES

1, Total coffee export (XCt): The study uses the amount of coffee exported or supplied to all
destinations measured in bags (60kg each) at time:

2, World coffee price (WPRt): The paper captured as the average annual value both Arabica
and Robusta ICO indicators measured by US cents per pound. As a result, the projected sign of
the World coffee price in this study is either positive or negative.

3, Real Exchange Rate (RERt): is Ethiopia's real bilateral exchange rate with her trading
partner at time t. Exchange rate appreciation/depreciation is thought to
have suppressed/encouraged exports. As a result, the projected sign of the exchange rate in this
study is either positive or negative.

4, Foreign Direct Investment (FDIt¿: FDI reflects the stock of foreign direct investment in
Ethiopia (in USD million). It is one method of increasing a country's economic capacity, as it
implies that attracting foreign direct investment in the future will result in increased exporting
capacity. As a result, the projected result is positive.

5, Total domestic production of coffee (DPCt): In this study, total domestic production of
coffee is the total amount of green coffee beans produced domestically. the projected sign of the
Total domestic production of coffee in this study is positive.

6, World coffee production (WPCt ): This can be the amount of coffee produced worldwide by
the top ten (81.23%) coffee producing and exporting countries measured in bags (60kg each). As
a result, the projected sign of the World coffee production in this study is either positive or
negative.

7, Growth Domestic Product (GDP): The primary criteria in explaining export are the size of
the exporting countries, as measured by GDP or population, because exports should be affected
by domestic income growth. The capacity to supply exporting commodities is said to be reflected
in the home economy's GDP. A high GDP level indicates a high level of output in the exporting
country. As a result, the variables are projected to be positive.

8, Terms of Trade (TOTt): Terms of trade show the ratio of export prices to import prices. An
improvement in terms of trade means that Ethiopia gets higher prices for its exports compared to
what it pays for imports, which improves its economic performance. As a result, the variables are
projected to be either positive or negative.

TECHNIQUES OF ESTIMATION AND SPECIFICATION

In order to identify approaches for estimating the model, it is necessary to understand the nature
of the data. To achieve this, we employ the robust econometric technique known as
Autoregressive Distributed Lag (ARDL), as developed by Pesaran et al. (2001), Pesaran (1997),
Pesaran & Shin (1995, 1998, 1999), and Pesaran et al. (1996). The ARDL method offers several
advantages over other conventional methods of cointegration, such as fully modified OLS,
Johansen, Maximum likelihood estimation, and inference on cointegration. One notable
advantage of the ARDL approach is its ability to distinguish between dependent and independent
variables. Unlike other models that strictly adhere to endogenous variables, the ARDL model can
handle both combinations of endogenous and exogenous variables. Additionally, it allows for
estimation even when the explanatory variables are endogenous (Pesaran and Shin, 1999;
Pesaran et al., 1996). Another advantage of the ARDL method is its applicability to variables that
are integrated or stationary at different levels, such as I(0) or I(1), or even fractionally
cointegrated (Pesaran and Pesaran, 1997). It is worth noting that the choice of estimation
procedure can significantly impact the empirical results, and various alternative options are
available (Pesaran and Smith, 1995). Furthermore, the ARDL method allows for the inclusion of
different numbers of lags for each variable, providing flexibility in the analysis. It is also
particularly suitable for studies with small sample sizes. Additionally, ARDL provides unbiased
long-run estimates (Odhiambo, 2008). Lastly, the model is well-suited for examining the short
and long-term effects of regressors on The Impact of International Coffee Market Dynamics on
Ethiopian Coffee Export .

Reference

Odhiambo, N.M. (2008), “Financial depth, savings and economic growth in Kenya: a dynamic
causal linkage”, Economic Modelling, Vol. 25 No. 4, pp. 704-713.

Pesaran, M.H. and Shin, Y. (1998), “An autoregressive distributed-lag modelling approach to
cointegration analysis”, Econometric Society Monographs, Vol. 31, pp. 371-413

Pesaran, M.H., & Pesaran, B., (1997) “Working with Microfit 4.0: Interactive Econometric
Analysis”, Oxford University Press.

Pesaran, M.H., & Shin, Y., (1995) “Autoregressive distributed lag modelling approach to
cointegration analysis”, DAE WP 9514, Department of Applied Economics, University of
Cambridge

Pesaran, M.H., & Shin, Y., (1999) “An autoregressive distributed lag modelling approaches to
cointegration analysis”, In Chapter 11 in Econometrics and Economic Theory in the 20th
Century: The Ragnar Frisch Centennial Symposium, Strom S. Cambridge University Press

Pesaran, M.H., (1997) “The role of economic theory in modelling the long-run”, the economic
journal 107, 178–191
Pesaran, M.H., Shin, Y. and Smith, R.J. (2001), “Bounds testing approaches to the analysis of
level relationships”, Journal of Applied Econometrics, Vol. 16 No. 3, pp. 289-326