One of the chief concerns in economic policy is to better the life criterions of the population. This can be achieved through faster economic growing. However, there are different positions sing the most efficient manner to better growing and one of them is through the publicity of investing. In bend, to advance investing a state must hold adequate resources ; otherwise, it will hold to borrow from other states. Indeed, this is one of the restraints faced by most developing states. They do non hold adequate resources, hence, they have to incur on external debt as a channel to spur economic growing.

The issue about external debt as a mechanism to advance economic growing creates a relevant argument among economic experts ( Ayadi 2008 ) . The chief concern is whether or non external adoption behaviors to faster economic growing in debitor states. This argument consequences in two chief positions for explicating the relationship between external debt and economic growing.

On the one manus, there are some theories such as the neoclassical and the endogenous growing theoretical account which advocate that external debt has a positive impact on economic growing. The justification is that external debt is one of the beginnings for funding capital formation. And, this funding in capital formation impacts positively on investing, hence, promotes economic growing. On the other manus, others such as Krugman ( 1988 ) contradict this position adverting to external debt as one of the factors haltering economic growing. Kalonji ( 2003 ) refers that big external debt load is considered by the non governmental organisations as a major cause of poorness through its effects on economic growing and human development internationally. Indeed, external debt has being mentioned by some writers as one of the factors difficulty economic growing in developing states. Ayadi ( 2008 ) has argued with this position saying that external debt load had dramatically limited developing countriesA? engagement in the universe economic system and the debt service duties continue to attest as an hindrance to economic growing and development.

Apart from this in 1980 more than a half of developing states ended up in debt crises. Harmonizing to Hjertholm ( 2001 ) this job was largely related to factors such as civil war and societal discord, external dazes, deficiency of accommodation and reforms, and besides creditors financing and refinancing policies.

Mozambique is one of these developing states and besides signifiers portion of the Heavily Indebted Poor Countries ( HIPC ) . Therefore, in 1980 it was besides affected by the debt crisis job. In this period the state was in a Reconstruction procedure after the war against the Lusitanian colonialism and at the same clip confronting a civil war. These factors added to others such as hapless public presentation in footings of societal and economic indexs, conducted to accretion of external debt.

Mozambique as other HIPC depends on external financess to advance economic growing. Until the present twenty-four hours this state does non hold sufficient resources and depends largely on external loans to finance its development undertakings. In this instance it is implicitly assumed that there is a positive relationship between external debts and economic growing.

Therefore, the purpose of this survey is to look into whether external funding is holding a positive impact on economic growing, for the Mozambican economic system. Specifically the survey has the aim of analyzing the relationship between these two variables. This will be done trough the analysis of cointegration, correlativity and causality of the variables.

To accomplish the above mentioned aim the paper is structured as follows after this brief debut, subdivision two presents the literature reappraisal. Then subdivision three introduces the reader to the methodological analysis used in informations analysis. Section four nowadayss background information of Mozambique with regard to the development of its external debt stock and economic public presentation since the 1980 ‘s. This is followed by subdivision five which presents the information analysis and eventually the presentation of reasoning comments in subdivision six.

## Literature Review

In this subdivision is presented different positions sing the expected relationship between external debt and economic growing. And besides the mechanisms through one variable can impact other.

As mentioned before, there are some theoreticians such as the neoclassical and the endogenous growing economic experts who advocate a positive relationship between external debt and economic growing. Harmonizing to Ricci et Al. ( 2002 ) in the neoclassical and endogenous theoretical accounts the positive impact on growing is through capital mobility or ability of the state to borrow and to impart. This means, when the state is able to pay the debt is it will heighten economic growing through capital accretion and productiveness growing.

Additionally, if the state is sing macroeconomic stableness, economic inducements and at the same clip utilizing the financess for productive investing, this good environment, should increase growing and let for timely debt refunds. And this positive relationship can be largely benefited by states in the early phase of development.

Contrary to this position, there are some oppositions who advocate that external debt impacts negatively on economic growing. Kalonji ( 2003 ) states that, the negative consequence on economic growing can be as follows. First, by deviating the financess meant to investing ; secondly by making uncertainness about authorities actions paying the debt, hence detering private sector-led investing and employment ; thirdly by taking to capital flight. Additionally, Arnone ( 2005 ) refers that big debt stocks lead to capital flight, high revenue enhancement rates and uninterrupted over-borrowing with a negative consequence on growing.

Krugman ( 1988 ) has argued that external debt can act upon negatively investing and accordingly economic growing through debt overhang ( DO ) . Harmonizing to him a state has a debt overhang job when the expected present value of possible hereafter resource transportations is less than its debt. There are different ways of explicating the mechanisms in which DO can impact investing and therefore economic growing.

First Claessens et Al. ( 1997 ) states that the consequence can be through revenue enhancement deterrence and macroeconomic sustainability. The revenue enhancement deterrence is justified by disheartenment of possible investors created by presuming that there would be revenue enhancements on future income in order to do debt refunds. And the macroeconomic instability relates to additions in financial shortage, uncertainness due to exceeding funding, exchange rate depreciation, possible pecuniary enlargement, and awaited rising prices.

Second, Kalonji ( 2003 ) argues that DO creates inducements to low public and private investing because a big portion of resources is transferred abroad for debt payment. It besides increases outlooks that debt service will be given to be financed by distortionary steps such as rising prices revenue enhancement or cuts in public investing. The same writer refers that normally the new flows of external resources can be curtailed in high liability states because international fiscal markets and givers consider investing in these states as being hazardous. These states are usually considered by these international administrations as exhibiting jobs of economic misdirection and bad administration.

Third, Ayadi ( 2008 ) refers that the negative impact is chiefly caused by the liquidness restraint. The liquidness restraint, implied by the debt-servicing demands, may switch the budget off from the societal sector or public investing.

Fourthly, Ricci ( 2002 ) argues that big debt service is expected to hold negative effects on economic public presentation because of the uncertainness sing the part of the debt that will be serviced with the state ‘s ain resources. This misallocation of investing in bend will take down the efficiency of overall capital accretion, therefore proposing that high degrees of debt and associated uncertainness might impact growing besides via investing efficiency and productiveness. The uncertainness is verified peculiarly in low income states where there is no certainty about the part of the debt will be serviced with the states ain resources and it may besides take to misallocated and poorer quality investing undertakings that slow productiveness growing.

Last, Fosu ( 1996 ) besides considers that an of import channel by which debt can impact growing is by cut downing the productiveness of investing. This may take to decrease in growing even when the degree of investing is non affected. The phenomenon might be peculiarly relevant for African states whose major beginning of investing is likely to be of the public type, and whose investing rate may non needfully diminish with debt, instead the type of investing can be affected.

The DO can be represented in a Debt Laffer Curve ( Figure 2.1 ) . This curve represents the consequence of debt on growing. Harmonizing to Ricci ( 2000:6 ) this curve represents the expected refunds as a map of the face value of debt. Along the left side of the curve the addition in the face value of debt service is related with additions in debt refund, while in the right side an addition in the face value reduces the expected refund. The curve besides reveals that there is a bound in exciting growing by spread outing debt. There is a maximal point after which addition in debt Begins to hold a negative impact on growing. Harmonizing to Ricci et Al. ( 2002 ) there is a scope after which the impact of debt on growing becomes negative. The same writer besides argues that for the net present value of debt to exports index such a scope is really near to the debt mark ( 150 per cent of exports ) of the current HIPC Initiative.

## Figure 2.1 Debt Laffer Curve

## Methodology

In this subdivision will be described the methodological analysis and all the stairss used in informations analysis. The paper is centred on analyzing the relationship between growing rate in external debt and growing rate in Gross Domestic Product, maintain all the other variables changeless ( ceteris paribus ) . To be more precise the focal point is on happening whether these two variables have a long tally relationship and besides analyse Granger causality and correlativity between them.

The chief beginnings of the informations were the World Bank and the Banco de Mocambique web sites. The information used is a clip series[ 1 ]traveling from 1980 to 2008 and the analysis will be done utilizing the Ordinary Least Squares ( OLS ) method.

The first measure in the analysis is to look into whether the variables are stationary since the variables used are clip series. The justification for this measure is the fact that the premise behind the usage of clip series informations is stationarity of the analysed variables. However, most of the clip series are non stationary, therefore, before utilizing it ; one has to look into whether the clip series are stationary. Harmonizing to Gujarati ( 2003: 798 ) a clip series is considered stationary when it ‘s average, discrepancy and autocovariance at assorted slowdowns do non alter over clip in other words the variable do non depend on clip. This belongings is of import to allow the usage of clip series variables for other periods instead than merely for one specific period, the same as to state that, stationary clip series allow the information to be more utile particular for the intent of prediction.

The trial for stationarity will be done utilizing two methods, viz. ocular secret plan or in writing method and unit root trial. The in writing method consists in pulling a additive graph and analyze the form taking into consideration the chief premise of stationary variables which are the non alteration over clip of the mean, discrepancy and autocovariance of the variable. And, proving the unit root belongings can be considered the same as to prove whether the variable is non stationary. In this trial it is used the first-difference variables arrested development where the variable in survey is regressed holding as independent variable it self lagged one period. From the consequences obtained it is tested whether the coefficient from the lagged variable is equal to zero. If that is the instance, it means the variable is non stationary. The Augment Dickey- Fuller ( ADF ) trial will be used to prove this hypothesis. This trial is an drawn-out version of Dickey- Fuller trial ( DF ) and it is called augment because the arrested development has been augmented with lagged alterations ( ) . The equation to be regressed is:

( 3.1 )

Where:

is consecutive uncorrelated mistake ;

can be either gross domestic merchandise ( GDP ) or external debt stock ( EDS ) .

Therefore the unit root trial whether I? = 0 which means. If I? is equal to zero one can reason that the Zt is a non stationary variable, while if its negative the variable is considered stationary. The determination regulation in this instance will be as follow:

t* ( calculated ) & gt ; ADF critical value – the void hypothesis is non rejected, so the variable is non stationary ;

t* ( calculated ) & lt ; ADF critical value – the void hypothesis is rejected, so the variable is stationary.

The 2nd measure in analyzing the information will be to transform the non stationary clip series since the purpose is to work with stationary variables. The method to be applied will depend on the features of the non stationary variables. In the instance where the clip series has a unit root it can be transformed to stationary merely by taking the first differences of the variables. The same as to regress:

( 3.2 )

Where:

Contrary, in the instance where the non stationary variable is non unity root the option can be to transform to stationary by regressing the dependant variable on clip. Therefore, the remainders from this arrested development will be stationary. The equation to regress will be:

( 3.3 )

Following this, the 3rd measure will be to analyze the relationship between the variables. However, one has to be cautious because running a arrested development of a non stationary clip series on the other non stationary clip series variable can ensue in a specious or non sense arrested development. Wooldridge ( 2009: 363 ) provinces that specious arrested development job is the phenomenon of happening a relationship between swerving variables merely because each is turning over clip. To avoid this job, the same as to state hold a meaningful arrested development one has to analyze whether the two variables are cointegrated.

Harmonizing to Gujarati ( 2003 ) two or more variables are cointegrated if they have a long tally or equilibrium relationship. Therefore to carry on this trial between growing rate of GDP and stock of external debt the Engle- Granger trial will be applied. The stairss in this trial are as follow: foremost, regress the one non stationary variable on the other non stationary variable. In this instance arrested development will be utilizing the invariable and the independent variable, and the equation will be as follow:

( 3.4 )

Where: represents the gross domestic merchandise, and

represents the growing rate of external debt stock.

Second, use the remainders from this arrested development to execute a unit root trial. As mentioned before if the remainders are found to be stationary, the variables are considered cointegrated. That is to state, the variables have a non specious or a meaningful relationship. Therefore, the undermentioned arrested development will be run:

( 3.5 )

In this instance for the determination regulation the criterion Dickey Fuller trial are no longer valid, hence new critical values are needed. Harmonizing to Harris ( 1995:54 ) the standard critical values are non valid for two chief grounds, foremost, because they will be given make reject the hypothesis of non cointegration and 2nd, because different critical values are needed as the sample size alterations. Therefore the new critical values are calculated harmonizing to the follow Mackinnon equation:

( 3.6 )

Where: represents the new critical value ;

T the figure of observations ; and

stand for the response surface for critical values of cointegration trial.

If the mistake term is stationary, the two variables are considered cointegrated and equation ( 3.4 ) represents the long tally relationship between them. However, if the mistake is non stationary there is no long tally relationship between so, and in this instance the relationship must be analysed utilizing the stationary variables.

The last measure in analyzing the information will be to carry on a causality trial. Gujarati ( 2003:696 ) provinces that “ the being of relationship between the variables does non turn out causality or the way of influence. ” Therefore must be clear between GDP and EDS which variable causes other and besides the way of the influence and in this instance the chief aim is to happen whether there is causality from EDS to GDP.

Although the old mentioned sing the fact that external debt has effects on economic growing, there are some theoreticians as Easterly ( 2001 ) which consider that the causality is in opposite way. It means this writer considered that growing lag contributes to the debt crises. Following this seemingly contention, one has to be clear about the causality between GDP and EDS in Mozambique. Indeed, rephrasing Ricci ( 2004:5 ) one of the cardinal issues in guaranting the hardiness of the consequences is to do certain that the consequence identified is from to debt to growing, and non the contrary causality. The analysis will be done utilizing Granger causality trial. This trial consists in estimation the follow arrested developments:

( 3.7 )

( 3.8 )

Equation 3.7 posits that current GDP is related to past values of itself and EDS, and equation 3.8 posits that current GEDS depends on the same variables. To do a decision sing causality two hypotheses must be tested. First, if a?‘a??i=0 and secondly, if a?‘I?j=0. These hypotheses will be tested utilizing F trial and the significance of the parametric quantities will find which variable ( Granger ) causes other. In other words, four possible instances can be distinguished in analyzing the causality, viz. unidirectional causality from debt to growing or frailty versa, bilateral causality and independency.

## Mozambique background Information

This subdivision has the aim to depict the development of the variables GDP and EDS since the twelvemonth ‘s 1980 in Mozambique.

From the period of its independency ( 1975 ) until 1984 Mozambique was following a cardinal planned economic system. During this period the debt was chiefly from the Eastern axis and Oil exporting states. These states were giving financess in order to assist the state care of the cardinal planned economic system.

As many other African states, in the 1980 ‘s Mozambique besides experienced a bad economic public presentation, with the growing rate of GDP uncovering negative values and the rising prices seemingly at low degrees. Dava ( 2005 ) stated that in the 1980 ‘s Mozambican accumulated debt achieved values ten times more than the exportations and was besides 50 per cent greater than the Gross Domestic Product. Harmonizing to GDM[ 2 ]( 2007 ) this addition in the degree of external debt resulted from internal and external factors such as lower degree of industrialisation, deficiency of nest eggs, societal and political instability, civil war, natural catastrophes, and impairment on footings of trade.

In 1984, the state became portion of the Bretton Woods establishments, viz. World Bank and The International Monetary Fund ( IMF ) and the economic system was changed from a planned into a market orientated system. In this period, these establishments became the chief creditors of the state and from so the state had entree to diversified beginnings of financess, such as Organization for Economic Cooperation and Development ( OECD ) , East European Countries.

Following the conditionalities of the Bretton Woods establishments, in 1987 Mozambique adopted a structural accommodation programme ( SAP ) . This programme included reforms such as liberalisation of the market, and this resulted in the betterment on the economic public presentation. From the tabular array 1 below can be seen that the GDP achieved positive growing rates from 1987 to 1991 with the highest value of 14.7 % in 1987. Additionally, the entire debt service reduced from degrees beyond 50 % of exports after the twelvemonth 1987.

One of the albatrosss in Mozambique was in the twelvemonth 1992 when the civil war ended. The table 1 reveals that this twelvemonth the GDP growing rate was negative, with the value of -5.1 % . However, after this twelvemonth the economic public presentation was improved with the GDP accomplishing positive values. Apart from this, the terminal of civil war implied an entryway into a recovery period and consequently it increased the demand for public goods and services. Zaqueu ( 2007 ) argued that the high demand for public services and substructures pushed up the demand for extra resources and this called for an augment in foreign aid, increasing all of a sudden the degree of public debt. Therefore, since this period the nominal stock of external debt increased and it reached the highest value in 1998 with a entire value of $ 6,056.00 million ( table1 ) . The same writer provinces that in this twelvemonth the state achieved an unsustainable degree of debt and was non able to carry through the debt service duties. Consequently Mozambique had to inquire for forgiveness of the international community.

In 1999, Mozambique received a debt alleviation and since so the stock of nominal external debt and debt service have reduced in Mozambique. The over twelvemonth decrease on debt was besides consequence of assorted enterprises created by the international establishments to supply debt alleviation to the most indebtedness states. Some the enterprises adopted by Mozambique are the to a great extent indebted hapless states initiative in 1999, many-sided debt alleviation enterprise ( MDRI ) in 2005 and debt buy-back operation in 2007.

The consequence of the enterprises can be visualised on table 1 from where can be seen that from 1999 the nominal stock of external debt and debt service in Mozambique decreased. The stock of external debt reduced from $ 8362.10 1000000s to $ 7219.30 million from 1998 to 1999 when HIPC was adopted. Further, the MDRI enterprise started in 2005 and was considered an extension of HIPC enterprise. One of the effects of this enterprise was the decrease on nominal debt stock from $ 4507.07 million to $ 2984.62 million in 2006. From them the entire external debt indexs continue to grounds a decrease on its degree over clip, thanks to the enterprises created by the international organisations.

## Table1. Mozambique external debt stock, GDP and rising prices

twelvemonth

External debt stocks, entire ( DOD, current US $ )

GDP ( current US $ )

Entire debt service ( % of exports of goods, services and income )

GDP growing ( one-year % )

Inflation, GDP deflator ( one-year % )

1980

## ..

3526.29

## ..

## ..

## ..

1981

## ..

3532.10

## ..

5.00

4.08

1982

## ..

3615.04

## ..

-6.90

17.46

1983

## ..

3237.80

## ..

-15.70

13.03

1984

1438.28

3372.75

7.83

-6.50

17.68

1985

2870.51

4458.20

34.46

1.00

33.15

1986

3491.71

5243.43

54.60

-2.30

12.71

1987

4125.44

2353.87

23.52

14.70

181.45

1988

4163.24

2093.39

27.08

8.20

48.33

1989

4362.79

2314.10

29.15

6.50

47.38

1990

4649.73

2463.24

26.19

1.00

34.06

1991

4718.40

2695.48

22.53

4.90

61.06

1992

5130.34

1968.90

22.85

-5.10

35.04

1993

5211.58

2027.65

32.92

8.65

45.92

1994

7271.67

2162.77

31.18

6.77

55.72

1995

7458.43

2246.79

34.51

2.70

51.17

1996

7565.88

3178.65

26.04

7.40

64.85

1997

7631.88

3751.83

19.17

10.24

8.97

1998

8362.10

4240.34

18.15

10.78

5.39

1999

7219.38

4448.02

16.99

8.12

4.38

2000

7248.53

4248.75

12.47

1.09

12.03

2001

4875.32

4075.06

8.32

11.90

14.88

2002

5033.79

4201.33

6.22

8.82

8.36

2003

3884.74

4666.19

5.83

6.02

5.22

2004

4811.24

5697.99

3.95

7.88

7.47

2005

4507.07

6578.52

3.78

8.39

8.78

2006

2984.62

7094.99

1.75

8.68

9.30

2007

3104.87

8010.52

1.32

7.02

7.40

2008

## ..

9735.33

## ..

6.46

6.54

Beginning: World Bank

## Datas analysis

This subdivision presents the information analysis sing the relationship between Domestic Product and External Debt entire stock in Mozambique. The first measure presented is the stationarity trial of the variables, followed by the cointegration trial and eventually the causality and correlativity between the variables.

## 5.1 Stationarity trial

## 5.1.1 Ocular secret plan method

Normally, the first measure in analysing clip series informations is to make a ocular secret plan where the purpose instance is to utilize this ocular secret plan to foretell the stationarity of the variables. The Figure 5.1 represents a secret plan of the informations from Mozambique External Debt noun phrase ( ruddy line ) stock and besides the Gross Domestic Product ( bluish line ) . A ocular review of this secret plan indicates a somewhat alteration over clip in both variables. From it, one can foretell that the variables mean, discrepancy and autocovariance alterations overtime. Therefore from the ocular review one can reason that the variables GDP and EDS are seemingly non stationary.

## Figure 5.1.1 External Debt Stock and Gross Domestic Product

## 5.1.2 Augment Dickey Fuller Test

This trial is conducted in order to corroborate the consequences from stationarity trial obtained utilizing the in writing method, the same as to state that, the ADF trial is to corroborate whether the variables are non stationary as predicted by the ocular secret plan. ADF is a left side tailed trial where the void hypothesis is the variable is non stationary and the alternate hypothesis is the variable is stationary.

## 5.1.2.1 Gross Domestic Product ADF trial

The consequences for Gross Domestic Product ADF trial are shown in the tabular array 5.1.2.1 below:

## Table 5.1.2.1.1 Gross Domestic Product ADF trial

Null Hypothesis: GDP has a unit root

Exogenous: Changeless

Lag Length: 0 ( Automatic based on SIC, MAXLAG=6 )

t-Statistic

Augmented Dickey-Fuller trial statistic

A 1.220977

Test critical values:

1 % degree

-3.689194

5 % degree

-2.971853

10 % degree

-2.625121

*MacKinnon ( 1996 ) nonreversible p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D ( GDP )

Method: Least Squares

Date: 12/17/09 Time: 06:52

Sample ( adjusted ) : 1981 2008

Included observations: 28 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

GDP ( -1 )

0.118570

0.097111

1.220977

C

-233.4912

401.9990

-0.580825

R-squared

0.054229

A A A A Mean dependant volt-ampere

Adjusted R-squared

0.017853

A A A A S.D. dependant volt-ampere

S.E. of arrested development

795.2269

A A A A Akaike info standard

Sum squared resid

16442030

A A A A Schwarz standard

Log likeliness

-225.6943

A A A A F-statistic

Durbin-Watson stat

1.779448

A A A A Prob ( F-statistic )

The ADF trial for the variable GDP evidences a t value ( 1.220 ) greater than the ADF critical values ( -3.689, -2.971, -2.625 at 1 % , 5 % and 10 % important degrees severally ) . Therefore, harmonizing to the determination regulation there are no groundss to reject the void hypothesis significance that GDP is non stationary variable. This consequence can besides be obtained by analysing the P value. This value represents the chance of rejecting the void hypothesis. For the instance of GDP unit root test the chance is about equal 0.99 which is greater than all the standard degrees of significance[ 3 ]. This high chance means that for all standard degrees of significance, one can non reject the hypothesis that GDP is a unit root or non stationary variable. This trial conducts to a similar decision obtained utilizing the ocular secret plan method which is, the GDP is non stationary variable.

Following this consequence, the non stationarity job must be solved. Since the variable is a unit root, it is expected to obtain a stationary variable by taking its first difference as mentioned in old chapter. Therefore, the Table 5.1.2.2 holla shows the unit root trial for GDP first difference variable. The comparing between T values reveals that the T calculated is less than the ADF critical values at all the degrees of significance, therefore, the void hypothesis of non stationarity is rejected. From it can be concluded that after taking the first difference, the variable GDP becomes stationary at all the standard degrees of significance.

In this instance the job of non stationarity was solved merely taking by the first difference. Therefore, there is no demand to regress the dependant variable on clip as suggested on the former chapter.

## Table 5.1.2.1.2 Gross Domestic Product first difference ADF trial

Null Hypothesis: D ( GDP ) has a unit root

Exogenous: Changeless

Lag Length: 0 ( Automatic based on SIC, MAXLAG=6 )

t-Statistic

Augmented Dickey-Fuller trial statistic

-3.765665

Test critical values:

1 % degree

-3.699871

5 % degree

-2.976263

10 % degree

-2.627420

*MacKinnon ( 1996 ) nonreversible p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D ( GDP,2 )

Method: Least Squares

Date: 12/17/09 Time: 06:59

Sample ( adjusted ) : 1982 2008

Included observations: 27 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

D ( GDP ( -1 ) )

-0.793428

0.210701

-3.765665

C

195.4412

161.1125

1.213073

R-squared

0.361923

A A A A Mean dependant volt-ampere

Adjusted R-squared

0.336400

A A A A S.D. dependant volt-ampere

S.E. of arrested development

817.1794

A A A A Akaike info standard

Sum squared resid

16694552

A A A A Schwarz standard

Log likeliness

-218.3305

A A A A F-statistic

Durbin-Watson stat

1.834911

A A A A Prob ( F-statistic )

The stationarity in GDP fisrt difference is besides depicted by the its line graph ( figure 5.2.1 ) . From it one can reason that the mean and discrepancy of this variable do non depend on clip, hence the variable can be considered stationary.

## Figure 5.1.2.1.1 Gross Domestic Product foremost difference

## 5.1.2.2 External Debt stock ADF trial

The tabular array 5.1.2.2.1 above indicates that chance of rejecting the void hypothesis ( p value ) is greater than the standard degrees of significance ( 1 % , 5 % and 10 % ) . Therefore there are no groundss to reject the void hypothesis of unit root. Consequently, one can reason that the variable EDS is a nonstationary variable. This consequence can besides be obtained by the comparing between the T values. In this instance the T calculated is greater than the ADF values for all the standard degrees of significance besides uncovering non stationarity job.

## Table 5.1.2.2.1 External Debt stock ADF trial

Null Hypothesis: EDS has a unit root

Exogenous: Changeless

Lag Length: 0 ( Automatic based on SIC, MAXLAG=5 )

t-Statistic

Augmented Dickey-Fuller trial statistic

-1.936440

Test critical values:

1 % degree

-3.752946

5 % degree

-2.998064

10 % degree

-2.638752

*MacKinnon ( 1996 ) nonreversible p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D ( EDS )

Method: Least Squares

Date: 12/17/09 Time: 07:09

Sample ( adjusted ) : 1985 2007

Included observations: 23 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

EDS ( -1 )

-0.201401

0.104006

-1.936440

C

1114.640

569.0027

1.958936

R-squared

0.151508

A A A A Mean dependant volt-ampere

Adjusted R-squared

0.111104

A A A A S.D. dependant volt-ampere

S.E. of arrested development

885.7616

A A A A Akaike info standard

Sum squared resid

16476045

A A A A Schwarz standard

Log likeliness

-187.6777

A A A A F-statistic

Durbin-Watson stat

1.863720

A A A A Prob ( F-statistic )

The non stationarity job is solved by taking the first difference. The table 5.2.2.2 represents the ADF trial for EDS first difference and the values reveals a T calculated smaller than the ADF critical values. As a effect the variable becomes stationary after taking the first difference. From the T values one can besides reason that External debt stock first difference is stationary because the DEDS P value is smaller than the critical values at all the standard degrees of significance. This conducts to rejection of unit root void hypothesis.

Figure 5.2.2 represents ocular secret plan of External debt stock variable. This ocular secret plan this conducts to the same consequence as ADF trial. From the graph one can reason that the mean and discrepancy of this variable no longer depends on clip, therefore, the variable is stationary.

## Table 5.1.2.2.2 External Debt stock foremost difference ADF trial

Null Hypothesis: D ( EDS ) has a unit root

Exogenous: Changeless

Lag Length: 0 ( Automatic based on SIC, MAXLAG=5 )

t-Statistic

Augmented Dickey-Fuller trial statistic

-4.834943

Test critical values:

1 % degree

-3.769597

5 % degree

-3.004861

10 % degree

-2.642242

*MacKinnon ( 1996 ) nonreversible p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D ( EDS,2 )

Method: Least Squares

Date: 12/17/09 Time: 07:12

Sample ( adjusted ) : 1986 2007

Included observations: 22 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

D ( EDS ( -1 ) )

-1.025594

0.212121

-4.834943

C

12.45168

199.8301

0.062311

R-squared

0.538923

A A A A Mean dependant volt-ampere

Adjusted R-squared

0.515869

A A A A S.D. dependant volt-ampere

S.E. of arrested development

934.6738

A A A A Akaike info standard

Sum squared resid

17472304

A A A A Schwarz standard

Log likeliness

-180.6526

A A A A F-statistic

Durbin-Watson stat

2.060931

A A A A Prob ( F-statistic )

## Figure 5.1.2.2.1 External Debt Stock foremost difference

## Cointegration trial

The tabular array 5.2.1 represents a arrested development which has as dependant variable the GDP and EDS as independent variable. Additionally, the tabular array 5.2.2 represents the unit root trial for the remainders of the arrested development 5.2.1. The unit root trial for these remainders must be done. Therefore if the remainders are stationary the variables are considered cointegrated or holding a long tally relationship.

For the instance of cointegration test the rejection regulation is applied utilizing the T calculate and the new critical values given by Mackinnon equation ( 3.6 ) . The new critical values for cointegration trial are presented in table 5.2.3 holla. Indeed, the comparing between the T values reveals that T calculated ( -0.45 ) is greater than the critical values for cointegration trial ( -4.8093, -4.0362, -3.6629 at 1 % , 5 % and 10 % important degrees severally ) .As a affair of fact, one can non reject the void hypothesis of unit root. This means that there is no cointegration or long tally relationship between GDP and EDS. In other words both variables are non stationary and the arrested development between them is specious arrested development.

## Table 5.2.1 GDP and EDS arrested development

Dependent Variable: GDP

Method: Least Squares

Date: 12/17/09 Time: 07:18

Sample ( adjusted ) : 1984 2007

Included observations: 24 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

C

5302.626

1021.945

5.188759

Explosive detection system

-0.275711

0.189493

-1.454992

R-squared

0.087781

A A A A Mean dependant volt-ampere

Adjusted R-squared

0.046316

A A A A S.D. dependant volt-ampere

S.E. of arrested development

1658.853

A A A A Akaike info standard

Sum squared resid

60539473

A A A A Schwarz standard

Log likeliness

-210.9436

A A A A F-statistic

Durbin-Watson stat

0.286772

A A A A Prob ( F-statistic )

## Table 5.2.2 Remainders ( GPD and EDS ) unit root trial

Null Hypothesis: RESID_1 has a unit root

Exogenous: Changeless

Lag Length: 0 ( Automatic based on SIC, MAXLAG=5 )

t-Statistic

Augmented Dickey-Fuller trial statistic

-0.451330

Test critical values:

1 % degree

-3.752946

5 % degree

-2.998064

10 % degree

-2.638752

*MacKinnon ( 1996 ) nonreversible p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D ( RESID_1 )

Method: Least Squares

Date: 12/17/09 Time: 07:23

Sample ( adjusted ) : 1985 2007

Included observations: 23 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

RESID_1 ( -1 )

-0.057425

0.127235

-0.451330

C

212.7221

183.4971

1.159266

R-squared

0.009607

A A A A Mean dependant volt-ampere

Adjusted R-squared

-0.037555

A A A A S.D. dependant volt-ampere

S.E. of arrested development

874.9270

A A A A Akaike info standard

Sum squared resid

16075442

A A A A Schwarz standard

Log likeliness

-187.3946

A A A A F-statistic

Durbin-Watson stat

1.621339

A A A A Prob ( F-statistic )

## Table 5.2.3 Critical values for Cointegration trial

## N

## Model

## % Point

## Critical Value

1

Changeless

1

-4.8093

no tendency

5

-4.0362

## A

## A

10

-3.6629

Since there is no cointegration or long tally relationship between the variables the following measure is to find the relationship between the variables utilizing stationary variables. The relationship will be determined utilizing the growing rate of the variables which are besides stationary. The tabular arraies 5.2.4 and 5.2.5 holla reveals that these two variables viz. GGDP and GEDS are stationary at 5 % degree of significance.

## Table 5.2.4 GGDP unit root trial

Null Hypothesis: GGDP has a unit root

Exogenous: Changeless

Lag Length: 0 ( Automatic based on SIC, MAXLAG=6 )

t-Statistic

Augmented Dickey-Fuller trial statistic

-3.013235

Test critical values:

1 % degree

-3.699871

5 % degree

-2.976263

10 % degree

-2.627420

*MacKinnon ( 1996 ) nonreversible p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D ( GGDP )

Method: Least Squares

Sample ( adjusted ) : 1982 2008

Included observations: 27 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

GGDP ( -1 )

-0.534404

0.177352

-3.013235

C

2.414558

1.418772

1.701864

R-squared

0.266423

A A A A Mean dependant volt-ampere

Adjusted R-squared

0.237080

A A A A S.D. dependant volt-ampere

S.E. of arrested development

6.146509

A A A A Akaike info standard

Sum squared resid

944.4893

A A A A Schwarz standard

Log likeliness

-86.30124

A A A A F-statistic

Durbin-Watson stat

1.889011

A A A A Prob ( F-statistic )

## Table 5.2.5 GEDS unit root trial

Null Hypothesis: GEDS has a unit root

Exogenous: Changeless

Lag Length: 0 ( Automatic based on SIC, MAXLAG=4 )

t-Statistic

Augmented Dickey-Fuller trial statistic

-6.409849

Test critical values:

1 % degree

-3.769597

5 % degree

-3.004861

10 % degree

-2.642242

*MacKinnon ( 1996 ) nonreversible p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D ( GEDS )

Method: Least Squares

Sample ( adjusted ) : 1986 2007

Included observations: 22 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

GEDS ( -1 )

-0.893670

0.139421

-6.409849

C

1.164670

3.767671

0.309122

R-squared

0.672594

A A A A Mean dependant volt-ampere

Adjusted R-squared

0.656223

A A A A S.D. dependant volt-ampere

S.E. of arrested development

17.20621

A A A A Akaike info standard

Sum squared resid

5921.073

A A A A Schwarz standard

Log likeliness

-92.76418

A A A A F-statistic

Durbin-Watson stat

2.397707

A A A A Prob ( F-statistic )

Since the two variables are stationary at 5 % degree of significance one can continue analyzing the relationship between them. Therefore the relationship between the two variables will be given by the arrested development:

## Table 5.2.6 GGDP and GEDS relationship

Dependent Variable: GGDP

Method: Least Squares

Date: 11/22/09 Time: 21:19

Sample ( adjusted ) : 1985 2007

Included observations: 23 after accommodations

Variable

Coefficient

Std. Mistake

t-Statistic

C

6.581158

0.961357

6.845699

GEDS

-0.057346

0.036356

-1.577351

R-squared

0.105928

A A A A Mean dependant volt-ampere

Adjusted R-squared

0.063353

A A A A S.D. dependant volt-ampere

S.E. of arrested development

4.487371

A A A A Akaike info standard

Sum squared resid

422.8665

A A A A Schwarz standard

Log likeliness

-66.11855

A A A A F-statistic

Durbin-Watson stat

2.268266

A A A A Prob ( F-statistic )

Table 5.2.6 has every bit dependent variable the non stationary GGDP and as independent variable the non stationary GDES. From this arrested development can be visualized that these two variables have a negative relationship. Indeed, an addition in 1 % in growing rate of EDS will ensue in a lessening in 0.05 % in the growing rate of GDP. Mentioning the debt Laffer curve one can state that Mozambique is on the right side of this curve. It is a scope where the impact of debt is negative significance that an addition in EDS behaviors to a decrease on GDP.

Although this relationship, the smallest R- squared value ( about 11 % ) evidences that this independent variable ( GEDS ) is non good plenty in explicating the fluctuations in the dependant variable ( GGDP ) . Additionally, the P value depicts that the variable GEDS is non statistical important at the standard degrees of significance. These features above mentioned, intend that this theoretical account is weak in explain the fluctuations on GGDP, hence, much of the fluctuations in GGDP are explained by variables non included in the theoretical account. These variables non included in the theoretical account are absorbed by the error term.

## 5.3 Correlation trial

The consequences obtained from the cointegration analysis are besides supported by the correlativity matrix ( table 5.3.1 ) . This matrix gives the correlativity between the variables considered in the survey. In this instance the matrix evidences that the GGDP and GEDS are decrepit and negatively correlated back uping the consequences found before.

## Table 5.3.1 Correlation Matrix

GGDP

GEDS

GGDP

A 1.000000

-0.325466

GEDS

-0.325466

A 1.000000

## 5.4 Causality trial

Although there is a correlativity between the variables, the Granger causality trial suggests that there is no causality between the two variables ( GGDP and GEDS ) in Mozambique. From the tabular array 5.4.1 one can see that for both void hypotheses tested the P value is greater than the standard degrees of significance, hence, there is no grounds to make non reject the hypothesis of no causality between the two variables. It means both variables are correlated but the fluctuation in both is explained by other factors. As a consequence these two variables are non relevant in explicating causality one in another. This consequence supports the consequence obtained by the cointegration trial which rejected the chief hypothesis from this paper and revealed that GDES is a weak variable to explicate GGDP.

The consequences suggest that growing in external debt stock do non explicate the growing rate on economic public presentation in Mozambique. Therefore in this instance it can be rejected the hypothesis that the hapless economic public presentation is related to the degree of external debt. Even cognizing that Mozambique is a HIPC, the high degree of debt is non the chief factor impacting the economic growing.

## Table 5.4.1 Granger Causality trial

Pairwise Granger Causality Trials

Date: 11/22/09 Time: 14:35

Sample: 1980 2008

Slowdowns: 2

A A Null Hypothesis:

Ob river

F-Statistic

A A GEDS does non Granger Cause GGDP

21

A 0.54403

A A GGDP does non Granger Cause GEDS

A 1.77664

## Decision

Knowing what have been advocated by some theoreticians that the high degree of external debt is one of the factors haltering economic growing in developing states, this survey had the purpose to analyze the relationship between Gross Domestic Product and external debt stock in Mozambique. This was done by analyzing the cointegration, correlativity and causality between these two variables. The processs were as follows, first analyse the stationarity of the variables, second the cointegration and correlativity and in conclusion the causality.

The stationarity trial revealed that both variables GDP and EDS are non stationary at the standard degrees of significance. This job of non stationarity of the variables was solved simple by taking the first difference of each variable. After taking the first difference the stationarity trial revealed that both variables became stationary.

Since the two variables were non stationary, the cointegration trial had to be done. This was made by making the stationary trial for the remainders of the arrested development incorporating GDP as dependent variable and EDS as independent variable. The stationary trial revealed that the remainders are non stationary. This consequence means that there is no cointegration or long tally relationship between the variables. In other words since the two variables are non stationary the arrested development between them is specious. In this instance the relationship between the variables had to be analysed utilizing stationary variables. The stationary trial for the variables GGDP and GEDS evidenced that both variables are stationary at 5 % degree of significance, therefore, these variables were used to find the relationship.

The theoretical account regressed had GGDP as dependent variable and GEDS as independent variable. From it, it was seen that there is a negative relationship between GGDP and GEDS significance that an addition in GEDS behaviors to diminish in GGDP. Put otherwise an addition in 1 % in growing rate of EDS consequences in a lessening in 0.05 % in the growing rate of GDP. Apart from this the consequences depicted that GDES is statistical non important in explicating the fluctuation in GGDP. Additionally, the lower R-squared ( about 11 % ) showed that this theoretical account is weak in explicating the fluctuations in GGDP.

The decision old mentioned was besides supported by the correlativity trial. This trial revealed a weak and negative correlativity between GGDP and GEDS. Sing the causality trial it depicted that there is no causality between the two variables. To be more precise, none of these two variables granger causes the other.

In sum the information suggests that there is a negative and weak relationship between the GGDP and GDES in Mozambique. Therefore, the economic public presentation in this state can non be justified by its degree of external debt.

Even so, one must be cautious in sing these consequences because there are some factors which can do them undependable. First, the consequences might be realistic given the little figure of observations used in the sample. The ideal would be to utilize quarterly informations in order to increase the figure of observations. Second, the trials used have some failing, for case the Dickey Fuller trial and Granger Causality test depend on the figure of lags chosen, therefore it can act upon the consequence obtained when utilizing it. Third, the paper considered a simple arrested development theoretical account with one dependant variable. This simple theoretical account increases the chance of correlativity between the error term and the explanatory variable. This is because one of the variables absorbed by the error term can be correlated with the independent variable. Therefore it can carry on to a biased and inconsistent calculator.

Despite the failing of the theoretical account, it can be used to give a general thought sing the consequence of external debt on economic growing in Mozambique. From it one can reason that it is non the degree of debt which hampers economic growing in Mozambique. Indeed, the state is sing socio-political stableness, good economic public presentation, attractive economic environment and has being portion of assorted enterprises of debt alleviation. Therefore, if the external financess are used in productive investings added to this good environment, it should ensue in an addition growing and let for debt refunds in clip.