Measuring the Impact of the General Budget Deficit on the Trade

Balance Deficit in Algeria for the Period 1990–2020

ABDELKARIM ELMOUMEN

Department of Economics,

University of Adrar,

ALGERIA

NAEIMAH FAHAD S ALMAWISHIR

Department of Business Management

College of Business,

Jouf University,

SAUDI ARABIA

HOUCINE BENLARIA

College of Business,

Jouf University,

SAUDI ARABIA

TAHA KHAIRY TAHA IBRAHIM

College of Business,

Jouf University,

SAUDI ARABIA

Abstract: - The study aims to describe, analyze and measure the impact of the general budget deficit on the

trade balance deficit in Algeria for the period 1990–2020, through the application of the VAR autoregressive

ray model. The study concluded that there is a direct relationship between the two variables, indicating that the

rise in the general budget deficit contributes to raising the trade balance deficit. The analysis of the response

and variance countries also showed that there is a positive significant effect of the general budget deficit on the

deficit of the Algerian trade balance in the short term, but no effect in the long term.

This study contributes to the current literature by using the VAR model to analyze the impact of the general

budget deficit on the trade balance deficit in Algeria for the period 1990–2020. It is measured through the

balance of the general budget and of the trade balance. The study shows that there is a strong positive impact

and statistical significance of the general budget deficit on the trade balance deficit.

Key-Words: - general budget, trade balance deficit, VAR, Algeria.

Received: April 12, 2022. Revised: September 20, 2022. Accepted: October 23, 2022. Published: November 23, 2022.

1 Introduction

The issue of the relationship between the general

budget and the trade balance is one of the most

important economic issues, not only at the level of

developing countries, but also at the level of

developed countries. The importance of this issue

emerged in the eighties in the United States, which

witnessed a deficit in both the general budget and

the trade balance, which was known at that time as

twin impotence or bilateral impotence. It can be said

that the analysis of the relationship between the

general budget and the trade balance shows the

extent of the mutual impact between both financial

and trade policies in any economy, meaning that any

change in the outcome of one of them may affect the

other in the same direction. In other words, the

financial policy tools represented in public revenues

and public expenditures from It can affect the trade

balance through the movement of exports and

imports, and vice versa. On tracking the

characteristics and features of the Algeria economy,

it is noted that it relies heavily on public spending to

finance development projects and bring about

development in general, as well as depending on

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1226

Volume 18, 2022

imports to provide consumer and investment goods

and collecting its revenues from foreign currency

through exports from the main source, oil.

Therefore, if we wish to study and analyze the

relationship between the general budget and the

trade balance in the Algerian economy, we must

study and analyze the relationship of the financial

sector expressed through the general budget, with

the commercial sector expressed in foreign trade

through the trade balance.

Successive governments in Algeria have

focused on the need to reduce dependence on oil

taxation. Efforts in this field have focused on

reducing the public budget deficit through

rationalizing expenditure, and also on reducing the

trade balance deficit by encouraging exports and

limiting imports, through so-called import licenses.

It is noted that efforts have been focused on one of

the two deficits without looking at its impact on the

other disability or the relationship between them.

Hence the problem of the study centers on

determining whether the general budget deficit and

the trade balance deficit constitute two separate

phenomena or whether they depend on each other.

This will contribute to assisting decision-makers in

addressing the two deficits.

Based on the foregoing, the problem of the

study is as follows: What is the impact of the

general budget deficit shock on the trade balance

deficit in Algeria for the period 1990–2020?

We summarize it in the following point: The

general budget deficit shock affects Algeria's trade

balance deficit more in the short term than in the

long term.

In our theoretical study, we relied on the

descriptive analytical approach for its relevance to

the nature of the subject, by describing the study

variables and analyzing the effects of the interpreted

variable on the dependent variable with the analysis

of the study results. In our applied study, we relied

on the standard method in order to conduct the

standard study, determining the optimum model to

explain the studied problem, and determining the

relationship and direction of the impact of the

general budget deficit on the trade balance deficit in

Algeria, using the statistical program Eviews10.

2 Literature Review

Many studies have dealt with this subject by

description and analysis and through the application

of different standard models. The most important

studies are discussed below.

The paper by Ogba (2014), [1], empirically

examines the impact of the budget deficit on the

trade balance in Nigeria. The general objective is to

examine the causality between budget deficit and

trade balance. The specific objective is to measure

the impact of the budget deficit on the trade balance

in Nigeria using annual data as a means of

determining the econometric relationship. In a time-

series context, modern econometric techniques were

used: the augmented Dickey–Fuller (ADF) unit root

tests for stationarity, Johansen and Juselius

cointegration for long-term relationship and Granger

causality tests were used to establish the direction of

causality in the model relationships. The ordinary

least squares method was used to measure the

impact of budget deficit on trade balance. The

findings of the study show that the Granger

causality test revealed a unidirectional relationship

between budget deficit and trade balance in Nigeria,

going from trade balance to budget deficit.

Similarly, budget deficit has a positive impact on

trade balance. The implication is that economic

policies that will minimize budget deficit will have

to be addressed through demand management such

as increases in tax and reduction in government

expenditure as a means of maintaining the trade

balance in Nigeria for the period under study.

The main objective of the paper by Abbasi

(2015), [2], is to present the theoretical argument of

the twin deficit hypothesis. In this study the author

evaluates the effect of the budget deficit on the

current account deficit in Iran for the period 1981–

2012. For this purpose, the authors use the

generalized method of moments approach. The

authors also use Keynesian and Ricardian theory on

budget deficit. They find that the coefficient of

budget deficit equals 0.09, which shows that a unit

of increase in budget deficit leads to a 0.09 unit

decrease in the current account balance. Indeed, a

one-unit increase in budget deficit leads to an

increase in the current account deficit. Also, the

results show that there is a positive and significant

relationship between oil revenue and the current

account balance. But the results show that the real

exchange rate does not significantly affect the

current account balance.

The paper by Wakeel (2013), [3], attempts to

analyze the impacts of budget deficit on

macroeconomic aspects in Pakistan. In fact, the

ways through which the budget deficit is financed

can affect money supply, output, exchange rate and

then foreign trade. Annual data for the period 1970–

2010 was taken for analysis. The ADF test was used

for stationarity, and the 3-Stage least squares

method was adopted for estimation using STATA-

10 software. The study revealed that the output

changes are positively related to BCP and

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1227

Volume 18, 2022

government expenditures but negatively to interest

rates. Money supply is positively related to GBD,

BCP and foreign reserves (R). So the money supply

does increase whenever we try to finance the budget

deficit through government, private or external

borrowing. On the other hand, changes in exports

and imports depend on changes in ER and their

relative prices respectively which are affected by

money supply. But the changes in imports are

bigger than changes in exports, pushing the balance

of trade towards deficit. The study also measured

the negative relationship between balance of trade

and output. The study concludes that when the

government tries to use government expenditures to

get higher output, a deficit may come into existence,

and then financing the budget deficit results in

inflation, trade deficit and afterwards affects output.

Budget deficit is a change in the public sector

borrowing requirement. It occurs when a

government's expenditure exceeds its revenue in a

given year. The world has seen an increase in

budget deficits, which has been attributed to the

Great Recession as well as other factors. Some

countries have a greater propensity for incurring

more frequent deficits owing to their reliance on oil

revenues or banking sectors for economic growth,

[4]. With the increase in budget deficits, economists

have ascribed two possible implications: 1) An

increase in government spending will result in an

increase in aggregate demand and a rise in prices;

and 2) The deficit will diminish economic growth.

In order to understand the effects of budget deficit

on output, employment and inflation, it is necessary

to establish a relationship between budget deficit

and GDP growth. GDP growth is expected to rise as

a country's debt-to-GDP ratio rises. This can occur

because of an increase in government spending or

because of a decline in the money supply, thus

leading to lower interest rates, [5]. However, there

are exceptions to the rule. This can occur because of

a rise in the demand for goods and services. This is

because a country's debt-to-GDP ratio is a function

of its interest rate and inflation rate. A country's

debt-to-GDP ratio will rise owing to an increase in

interest rates, which could be caused by another

economic shock or the desire to maintain the

currency's value in international markets. An

increase in inflation will also result in a higher debt-

to-GDP ratio as it generally translates into higher

interest rates, [6].

The effects of budget deficit on output and

inflation depend largely on national debts. When the

budget deficits are high and when the debt-to-GDP

ratio is low, government borrowing will have little

effect on interest rates. On the contrary, when the

budget deficit is low and when the debt-to-GDP

ratio is high, government borrowing is expected to

increase interest rates as there would be an increased

demand for credit, [7].

Our study contributes to the current literature by

using the VAR model to analyze the impact of the

general budget deficit on the trade balance deficit in

Algeria for the period 1990–2020. It is measured

through the balance of the general budget and the

balance of the trade balance.

3 Research Methodology

The autoregressive ray model (VAR) is written in its

generic form with a variable K and a number of

delays P as follows, [8]:

        

 󰇛󰇜

So that Xt=(x1t, x2t ,….,xkt )t, which are stable

variables.

Xt: ray after it (K,1)

A: array of features with a dimension (k,k)

A0: the ray of constant values beyond (k,1)

µt: The whitejamming ray which has a

dimension of (k,1) and fulfills the following

hypotheses:

 E(Ut)=0

 E(Ut,Ùt)=Ω

 E(Ut ,Ùs)=0,S≠t

Using the delay coefficient L, the model can be

written as follows [9]:

 XI=A0+A1 LXt+A2 L2

Xt+….+ApLpXt+µt

 Ø(L) Xt=A0+µt (2)

In which Ø(L) =lk-A1 L -A2 L2 -…-ApLp.

The VAR model studies the short-term

relationship between several variables, as it does not

depend on economic theory and does not distinguish

between dependent and independent variables, as it

considers them all internal or dependent variables.

This model is applied in two cases, [10]:

-The chains must be stable to the same degree, with

no cointegration and causation.

-The presence of different series in the degrees of

stability at level I(1) and the first difference I(1) or

in the first difference I(1) and the second difference

I(2).

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1228

Volume 18, 2022

3.1 The Unit Root Test

In a study conducted by Nelson and Plosser (1982),

[11], it was found that most of the macroeconomic

variables are not static at the level, but rather in the

first difference I(0) or the second difference, in

order to avoid the emergence of the problem of false

regression (spurious regression), which may not

give a real dimension or a meaningful economic

explanation. The first step in analyzing the data will

be to test the inactivity of the time series, which is

used to determine the "non-stationary" properties of

time series using the augmented Dickey–Fuller test.

However, although this test is widely used, it suffers

from the problem of not taking into account the

absence of the problem of variance and the test of a

normal distribution. Therefore, an additional test is

used, the Phillips–Perron (PP) test, which is better

and more accurate than the Dickey–Fuller test,

especially when the sample size is small, and in all

tests we rely on MacKinnon’s, [12], values, [13].

The time-series inactivity test depends on the

significance of the parameter (), by comparing the

calculated (t) with the tabular () (tau-statistique). If

the calculated value is greater than the tabular value

(in absolute values), this means that the time series

is stationary at the level. The time series is not

stationary if the calculated value is less than the

tabular (in absolute values), in which case the first

difference is required.

3.2 Cointegration Test

Cointegration is defined as the association between

two time series X_t, Y_t or more, such that

fluctuations in one cancel out fluctuations in the

other in a way that makes the ratio between its

values constant over time. The time-series data may

be unstable if taken separately, but it is stable as a

group. Such a long-term relationship between the set

of variables is useful in predicting the values of the

dependent variable in terms of a set of independent

variables, [14]. Achieving cointegration requires

that the two series X_t and Y_t be complementary

of the same order, and that the residuals resulting

from the estimation of the relationship between

them are complementary of rank zero, [15].

3.3 Autoregressive Ray Model Estimation

VAR

To estimate the autoregressive ray model the

following steps are taken:

Determining the degree of delay (p) for the

autoregressive ray path and determining the number

of appropriate time lags for the VAR(P) model

depend on the criteria of each of the Akaike

criterion (AIC), the Schwarz criterion (SIC), the

final prediction error criterion (FPE) and the

Hannan–Quinn (HQ) criterion. We chose the period

in which the lowest observed values are for these

criteria.

3.3.1 The study of Causation

The study of causation is of paramount importance

for understanding and explaining economic

phenomena. In order to establish appropriate and

correct economic policies, the direction of causation

also allows relationships to be established between

economic variables. We say that there is a causal

relationship from Y_t〖 to X〗_t with the Granger

concept, if the prediction results for Y_t values

based on the knowledge of the past of the variables

Y_t and X_t are better than the prediction results

based on the knowledge of the past of the variable

Y_t only. Granger considered that the future cannot

affect the present, [16]. If phenomenon B occurred

after phenomenon A, then phenomenon B could not

affect or cause phenomenon A, [17]. However, Sims

had a slight difference with Granger regarding

causation, as he considered that the future can affect

the present, [18].

3.3.2 Autoregressive Ray Trajectory Estimation

VAR models are estimated by estimating each

equation separately by the ordinary least squares

method (OLS) or by using the greatest probability

method.

Let the stable autoregressive ray model VAR(p):

          

Estimating the VAR(p) model requires

estimating the radius of the parameters B, and the

OLS method allows us to choose an estimator that

minimizes the sum of the square of the residuals

where: 〖B=(A_0 〖,A 〗_1 〖...A 〗_p) 〗

_((k,kp+1)).

3.4 VAR Autoregressive Ray Model Test

Using this test, we discuss the validity of the

estimated model, the analysis of response functions

and the analysis of variance.

3.4.1 VAR Model Stability

The path VAR(P) is stable if the following

conditions are met:

󰇛󰇜  󰇛󰇜 

 󰇛󰇜 󰇟󰇛 󰇜󰇛  󰇜󰇠

  

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1229

Volume 18, 2022

When the model is stable, the roots of the

polynomial computed by the determinant Det(Ik-A.Z

- A2.Z2-...-Ap.Zp) = 0 are inside the unit circle (all

unit roots are less than one). It must also be ensured

that the model is free of standard problems such as

the autocorrelation of errors and the stability of error

variance and that the residuals follow a normal

distribution.

3.4.2 Response Function Analysis and Variance

Analysis

VAR models allow for analysis of the effects of

economic policies by simulating random shocks and

analyzing the error variance with the assumption

that the external environment is constant. Random

shock analysis depends on measuring the effect of a

random shock on the model variables. This effect

can persist over a period of time of t + h. We can

represent this effect as follows, [19]:

Let us have an autoregressive ray model

VAR of degree p=1 and k=2:

X1t=A0+A1 X 1t-1+A2 X 2t-1+e1t

X2t=B0+B1 X 2t-1+B2 X 2t-1+e2t

The method of response functions for

calculating dynamic multipliers is characterized by

taking into account the sum of the relationships that

exist between the model variables and showing their

reaction to the occurrence of random shocks in

errors. In addition to the analysis of response

functions, the analysis of the variance of the

prediction error is one of the basic concepts in

autoregressive ray models, as it allows us to know

the extent to which each random shock contributes

and the weight of each random shock to the variance

of the error. It is also possible to know the

contribution of each variable to this variance, by

dividing the variance of the error of this variable by

the variance of the total error. If we return to the

previous model (1) VAR with two variables, the

prediction error for the variable X_(1t+h) can be

written as follows, [20]:



󰇛󰇜 



󰇛󰇜 

󰇛󰇜 

 

󰇛  󰇜  



󰇛󰇜

 

󰇛󰇜   

󰇛  󰇜

Where mn is the effect matrix.

As for the horizons of period h, dividing the

variance into ratios for renewals of X1t over X1t is

given by the following relationship:



󰇟

󰇛󰇜 

󰇛󰇜   

󰇛  󰇜󰇠



󰇛󰇜

The variance for renewals of X_1t over X_2t can

also be divided into ratios on the horizons of period

h by the following relationship:



󰇟

󰇛󰇜 

󰇛󰇜   

󰇛  󰇜󰇠



󰇛󰇜

The analysis of the results obtained is very

important, as it is presented as follows, [21]:

 If a random shock occurs in e1t and does

not affect the error variance of the variable

X_2t whatever the prediction period, we

say that the variable X2t is exogenous

because its evolution is independent of e1t.

 But if the random shock in e1t significantly

affects the error variance of the variable

X2t, then this variable is considered

dependent.

4 Analysis of the Results of the

Standard Study

In order to determine the relationship between the

general budget deficit and the trade balance in

Algeria, annual data for the two variables during the

period 1990–2020 were used. The statistics of the

balance of the general budget and the balance of the

trade balance were obtained from the reports of the

Bank of Algeria, which are in billions of US dollars.

The balance of the general budget (RB) is expressed

as the difference between public revenues and

public expenditures, and the balance of trade

balance (BC) as the difference between total exports

and total imports.

4.1 Time Series Stability

We use the unit root test for both time series using

the augmented Dickey–Fuller (ADF) test and the

Phillips–Perron (PP) test, which is better and more

accurate than the Dickey–Fuller test, especially

when the sample size is small. In all tests we rely on

MacKinnon's values, [22], [23]. The results of the

test are recorded in Table 1.

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1230

Volume 18, 2022

Table 1. Static test for model variables data

UNIT ROOT TEST TABLE (PP)

At Level

RB

BC

With Constant

t-Statistic

-1.3304

-1.5725

Prob.

0.6022

0.4839

With Constant & Trend

t-Statistic

-2.1449

-1.5845

Prob.

0.5013

0.7752

Without Constant &

Trend

t-Statistic

-0.6111

-1.5260

Prob.

0.4442

0.1171

At First Difference

d(RB)

d(BC)

With Constant

t-Statistic

-5.7188

-5.5253

Prob.

0.0001

With Constant & Trend

t-Statistic

-5.6253

-5.6933

Prob.

0.0004

Without Constant &

Trend

t-Statistic

-5.7384

-5.5823

Prob.

0.0000

UNIT ROOT TEST TABLE (ADF)

At Level

RB

BC

With Constant

t-Statistic

-1.3537

-1.5725

Prob.

0.5912

0.4839

With Constant & Trend

t-Statistic

-3.0047

-1.5845

Prob.

0.1498

0.7752

Without Constant &

Trend

t-Statistic

-0.6433

-1.5260

Prob.

0.4301

0.1171

At First Difference

d(RB)

d(BC)

With Constant

t-Statistic

-5.6392

-5.4683

Prob.

0.0001

With Constant & Trend

t-Statistic

-5.5353

-5.4726

Prob.

0.0005

0.0006

Without Constant &

Trend

t-Statistic

-5.6671

-5.5497

Prob.

0.0000

Source: Prepared by the researcher using Eviews10.

We note that the absolute value of the estimated t-

statistic is greater than the absolute value of the

tabulated values (Mackinnon) at 5% in both the

ADF and PP tests at the first difference, [24]. This

means that it is statistically significant, and therefore

we reject the hypothesis, that is, the two series are

stable at the first difference, meaning that all the

variables are integrated of the first degree.

4.2 Johansen's Cointegration Test

Johansen's simultaneous integration model states

that there is an equilibrium relationship between

economic variables in the long run which diverges

from the equilibrium in the short run. This is

corrected by economic forces that restore

equilibrium in the long run, [25]. Since the two

variables RB and BC are complementary of the

same degree, we will test the possibility of a

simultaneous integration relationship between them

at a level of significance of 5%. The results are

presented in Table 2.

Table 2. Johansen's cointegration test

Date: 12/21/21 Time: 22:55

Sample (adjusted): 1997 2020

Included observations: 24 after adjustments

Trend assumption: Linear deterministic trend

Series: DBC DRB

Lags interval (in first differences): 1 to 5

Unrestricted Cointegration Rank Test (Trace)

Hypothesized

Trace

0.05

No. of CE(s)

Eigenvalue

Statistic

Critical Value

Prob.**

None *

0.489463

15.43865

21.49471

0.0056

At most 1 *

0.198272

3.303655

5.841466

0.0023

Trace test indicates 2 cointegrating eqn(s) at the 0.05 level.

*Denotes rejection of the hypothesis at the 0.05 level.

**MacKinnon-Haug-Michelis p-values, [26].

Source: Prepared by the researcher using Eviews10.

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1231

Volume 18, 2022

From Table 2, it is clear that the hypothesis that

simultaneous integration exists at the 5% level of

significance is rejected because the effect values are

smaller than the critical values. This means that

there is no long-term equilibrium relationship

between the general budget deficit and the trade

balance in Algeria. As long as this test is not met, it

is possible to pass to the VAR model.

4.3 Determining the Optimum Delay Degree

for the VAR(P) Model

Determining the appropriate number of time lags for

a VAR(P) model depends on the criteria of Akaike

(AIC), Schwarz (SC), final prediction error (FPE)

and Hannan–Quinn (HQ). We chose the period in

which the lowest observed values are for these

parameters. The test results are recorded in Table 3.

Table 3. Determining the optimum delay degree for

the VAR(P) path

VAR Lag Order Selection Criteria

Endogenous variables: DBC DRB

Exogenous variables: C

Date: 12/21/21 Time: 22:50

Sample: 1990 2020

Included observations: 25

Lag

LogL

LR

FPE

AIC

SC

HQ

0

-284.9171

NA

3188663

2*

22.9533

7

23.05088*

23.08041

*

1

-283.0509

3.28455

1

3790669

8

23.1240

7

23.41660

23.20520

2

-281.5908

2.33602

7

4684913

7

23.3272

7

23.81482

23.46249

3

-280.4360

1.66296

8

5999994

0

23.5548

8

24.23745

23.74420

4

-269.8102

13.60103

*

3663830

6

23.0248

2

23.90241

23.26822

5

-264.0719

6.42692

2

3389976

5

22.88575

*

23.95836

23.18325

Source: Prepared by the researcher using Eviews10.

From Table 3, it is clear that the optimum degree of

delay and agreement with the AIC criteria is P=5.

4.4 Study of Causation

Table 4 shows the results according to the degree of

delay appropriate to the VAR model (5).

Table 4. Results of Granger's causality test between

the general budget deficit and the trade balance

Pairwise Granger Causality Tests

Date: 12/28/21 Time: 19:41

Sample: 1990 2020

Lags: 5

Null Hypothesis:

Obs

F-

Statistic

Prob.

DRB does not Granger cause DBC

25

5.92093

0.0459

DBC does not Granger cause DRB

0.08615

0.4979

Source: Prepared by the researcher using Eviews10.

Since the probability value (pro = 0.04) is less than

the 5% level of significance, we conclude that there

is a one-way causal relationship that runs from the

general budget deficit towards the trade balance

deficit at the 5% level of significance.

4.5 Model Estimation of Autoregressive Ray

Technique VAR(5)

The VAR ray model is estimated by applying the

least squares method in the event that we estimate

each equation separately. But if we estimate all the

equations once, we use the method of great

reasonableness. The results of the estimation are

shown in Table 5.

Table 5. VAR model estimation results (5)

Vector Autoregression Estimates

Date: 12/21/21 Time: 22:52

Sample (adjusted): 1996 2020

Included observations: 25 after adjustments

Standard errors in ( ) & t-statistics in [ ]

DBC

DRB

DBC(-1)

-0.232894

-8.873769

(0.00221)

(11.6142)

[-0.78891]

[-0.76404]

DBC(-2)

-0.219733

-4.358966

(0.00274)

(11.4776)

[-0.75318]

[-0.37978]

DBC(-3)

-0.062214

5.764418

(0.00265)

(10.4451)

[-0.23433]

[ 0.55188]

DBC(-4)

-0.022113

-21.25495

(0.00243)

(9.81314)

[-0.08865]

[-2.16597]

DBC(-5)

0.210546

19.07256

(0.00448)

(11.5854)

[ 0.71498]

[ 1.64626]

DRB(-1)

0.004535

0.167144

(0.00406)

(0.27767)

[ 0.64249]

[ 0.60195]

DRB(-2)

0.051060

-0.183038

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1232

Volume 18, 2022

(0.00494)

(0.19426)

[ 1.04502]

[-0.94225]

DRB(-3)

0.060803

0.115839

(0.00513)

(0.20173)

[ 1.18625]

[ 0.57423]

DRB(-4)

0.08432

0.429225

(0.00645)

(0.25377)

[ 1.30723]

[ 1.69142]

DRB(-5)

0.070392

-0.200320

(0.00806)

(0.31702)

[ 0.88009]

[-0.63188]

C

2.573207

-41.32861

(2.66170)

(104.717)

[ 0.96675]

[-0.39467]

R-squared

0.523881

0.625192

Adj. R-squared

0.244776

0.357472

Sum sq. resids

1764.083

2730441.

S.E. equation

11.22524

441.6237

F-statistic

3.528058

2.335245

Source: Prepared by the researcher using Eviews10.

We note from the previous table that the coefficient

of determination is R = 0.52, which indicates the

good explanatory power of the model, that is, 52%

of the changes in the Algerian trade balance deficit

are explained by their previous values and the

values of the general budget deficit for the past year,

and the rest are due to other variables that were not

included in the model. Also, Tstat statistics indicate

the significance of the parameters constituting the

autoregressive ray. Fisher's statistics Fcal = 3.52 are

greater than the tabular FT = 2.24 and indicate the

overall significance of the model. Hence, the

characteristic equation for the inflation rate is

considered statistically acceptable at the 5% level of

significance.

We also note that the trade balance deficit in

year (t) is explained and influenced by the trade

balance deficit and the general budget deficit for the

previous year (t-1), and the trade balance deficit in

year (t) is inversely related to the trade balance

deficit for the previous year, where the sign of

transactions was negative. That is, if the trade

balance deficit rises this year by 1%, it will decrease

in the next year. This is explained by the state

intervention to limit the increase in the trade balance

deficit. The trade balance deficit in the year (t) is

directly related to the general budget deficit in the

year (t-1), so that the transactions sign is positive.

This means that if the general budget deficit in the

year (t) increases by 1%, the trade balance deficit

will be expected to rise in the year (t+1), meaning

that the effect of the trade balance deficit by

increasing the public budget deficit continues from

year to year.

4.6 Validity Test of the Studied Model

In order to ensure the validity and validity of the

model, the following tests must be conducted:

4.6.1 Model Stability Test

The results of this are presented in Fig 1 and Table

6.

Fig. 1: Single circuit of the model

Source: Prepared by the researcher using Eviews10.

Table 6. Unit roots of the model

Roots of Characteristic Polynomial

Endogenous variables: DBC DRB

Exogenous variables: C

Lag specification: 1 5

Date: 12/21/21 Time: 23:00

Root

Modulus

-0.150627 - 0.938179i

0.950194

-0.150627 + 0.938179i

0.950194

-0.863904

0.863904

0.178906 - 0.839391i

0.858245

0.178906 + 0.839391i

0.858245

-0.664407 - 0.476780i

0.817775

-0.664407 + 0.476780i

0.817775

0.652962 - 0.421573i

0.777228

0.652962 + 0.421573i

0.777228

0.764487

No root lies outside the unit circle.

VAR satisfies the stability condition.

Source: Prepared by the researcher using Eviews10.

From the results, it can be seen that the values of the

unit roots are less than one, i.e. they are located

within the monolithic circle, which is evidence of

the stability of the estimated model.

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

Inverse Roots of AR Characteristic Polynomial

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1233

Volume 18, 2022

4.6.2 Test for the Normal Distribution of

Residuals

The results of this test are presented in Table 7.

Table 7. Test for the normal distribution of residuals

VAR Residual Normality Tests

Orthogonalization: Cholesky (Lutkepohl)

Null Hypothesis: Residuals are multivariate normal

Date: 12/21/21 Time: 23:03

Sample: 1990 2020

Included observations: 25

Component

Jarque–Bera

df

Prob.

1

11.40728

2

0.0033

2

0.866761

2

0.6483

Joint

12.27404

4

0.0154

*Approximate p-values do not account for coefficient

estimation

Source: Prepared by the researcher using Eviews10.

From these results, the Jarque–Bera statistic for all

residuals (11.4) and (0.86) is less than the tabulated

value (12.27), from which we accept the null

hypothesis, meaning that the residuals follow a

normal distribution at the 5% level of significance.

4.6.3 Error Autocorrelation Test

Fig 2 presents the results of the error autocorrelation

test.

Fig. 2: Autocorrelation of errors

Source: Prepared by the researcher using Eviews10.

Table 8. LM test

VAR Residual Cross-Correlations

Ordered by lags

Date: 12/21/21 Time: 19:25

Sample: 1990 2020

Included observations: 25

DBC

DRB

DBC

1.000000

-0.578588

DRB

-0.578588

1.000000

DBC(-1)

-0.078318

0.218708

DRB(-1)

-0.091490

0.208177

DBC(-2)

-0.057815

0.106584

DRB(-2)

-0.060939

-0.012876

DBC(-3)

-0.081758

-0.007436

DRB(-3)

0.038300

-0.091785

DBC(-4)

-0.047537

-0.062550

DRB(-4)

0.170940

-0.153192

DBC(-5)

0.026953

-0.087395

DRB(-5)

-0.060054

-0.144321

DBC(-6)

0.148049

-0.419480

DRB(-6)

0.078024

-0.065509

DBC(-7)

-0.003296

-0.153753

DRB(-7)

0.101042

0.004147

DBC(-8)

0.029384

0.039105

DRB(-8)

0.109994

0.041072

DBC(-9)

-0.163795

0.202986

DRB(-9)

0.021415

0.047022

DBC(-10)

-0.122225

0.334737

DRB(-10)

0.249505

-0.320336

DBC(-11)

0.259932

-0.116602

DRB(-11)

-0.002239

-0.069407

DBC(-12)

-0.059673

0.092803

DRB(-12)

0.052050

-0.022544

Asymptotic standard error (unadjusted) for lag > 0:

0.200000

Source: Prepared by the researcher using Eviews10.

From the results above, it can be seen that the test

probabilities of the various delays are greater than

the 5% level of significance, and therefore the null

hypothesis is acceptable, which means that the

errors are independent. Also, all points are located

within the confidence field in the form of (02),

which confirms that there is no autocorrelation

between errors of the estimated model.

4.6.4 Test for the Heterogeneity of Variance of

the Model

Table 9 presents the results for the heterogeneity of

variance of the model.

-.6

-.4

-.2

.0

.2

.4

.6

12345678910 11 12

Cor(DBC,DBC(-i))

-.6

-.4

-.2

.0

.2

.4

.6

12345678910 11 12

Cor(DBC,DRB(-i))

-.6

-.4

-.2

.0

.2

.4

.6

12345678910 11 12

Cor(DRB,DBC(-i))

-.6

-.4

-.2

.0

.2

.4

.6

12345678910 11 12

Cor(DRB,DRB(-i))

Autocorrelations with Approximate 2 Std.Err. Bounds

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1234

Volume 18, 2022

Table 9. White test for consistency of error variance

VAR Residual Heteroskedasticity Tests (Levels and Squares)

Date: 12/21/21 Time: 23:17

Sample: 1990 2020

Included observations: 25

Joint test:

Chi-sq

df

Prob.

64.21787

60

0.3311

Individual components:

Dependent

R-squared

F(20,4)

Prob.

Chi-sq(20)

Prob.

res1*res1

0.907449

1.960960

0.2709

22.68621

0.3045

res2*res2

0.975552

7.980630

0.0284

24.38880

0.2258

res2*res1

0.924114

2.435543

0.2011

23.10286

0.2838

Source: Prepared by the researcher using Eviews10.

From Table 8, we note that the probability (prob =

0.33) is greater than the 5% level of significance,

which indicates the stability of the error variance of

the estimated model.

4.7. Dynamic Study of the Estimated

Autoregressive Ray Model

In this step, we will discover the response of the

trade balance deficit to the public budget deficit

shocks by analyzing the following:

4.7.1 Impulse Response Functions Analysis

Table 10 and Fig 3 show the results of the impulse

response functions analysis.

Fig. 3: Analysis of response functions

Source: Prepared by the researcher using Eviews10.

Table 10. Analysis of response functions

Response of DBC:

Period

DBC

DRB

1

11.22524

0.000000

(1.58749)

(0.00000)

2

-3.772979

1.633367

(2.97678)

(2.55272)

3

-3.551665

1.751195

(3.15077)

(1.76160)

4

-1.296571

1.415776

(3.28553)

(2.01384)

5

-1.411887

2.207313

(3.32907)

(2.21473)

6

-0.668096

2.384169

(3.41402)

(2.30157)

7

-1.278758

-0.519496

(3.13213)

(1.92108)

8

0.453854

-0.655493

(3.15773)

(1.80156)

9

-0.290456

0.826020

(2.84196)

(1.68776)

10

-1.029081

0.810556

(2.62445)

(1.69185)

Source: Prepared by the researcher using Eviews10.

When a positive random shock of one standard

deviation occurs in the general budget deficit, it will

have a positive response in the trade balance deficit

after one year, i.e. in time (t+1). This positive effect

continues increasingly from the second year until

the end of the sixth year. Then it turns into an

increasing negative response until the end of the

eighth year, and then turns into a weak positive

response until the end of the tenth year. This result

confirms that the positive impact of the public

budget deficit on the trade balance deficit is in the

short term, while this effect weakens and fades in

the long run.

Also, the occurrence of a positive random shock

of the amount of one standard deviation in the trade

balance deficit has a positive response in the trade

balance deficit after one year, i.e. in time (t+1), then

it turns into a decreasing negative response until the

end of the seventh year. Then it turns into a weak

positive response during the eighth year, and then

turns into a decreasing negative response until the

end of the tenth year. This result shows that the

Algerian trade balance deficit is weakly affected by

the level of the trade balance deficit of previous

years. From the results obtained, we can say that

there is a positive moral effect of the general budget

deficit on the deficit of the Algerian trade balance in

the short term, and there is no effect in the long

term.

-5

0

5

10

1 2 3 4 5 6 7 8 9 10

Response of DBC to DBC

-5

0

5

10

1 2 3 4 5 6 7 8 9 10

Response of DBC to DRB

-600

-400

-200

0

200

400

1 2 3 4 5 6 7 8 9 10

Response of DRB to DBC

-600

-400

-200

0

200

400

1 2 3 4 5 6 7 8 9 10

Response of DRB to DRB

Response to Cholesky One S.D. (d.f. adjusted) Innovations ± 2 S.E.

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1235

Volume 18, 2022

4.7.2 Analysis of Variance

Table 11 shows the analysis of variance for the

estimated model.

Table 11. Bilateral covariance analysis between the

trade balance deficit and the public budget deficit

Variance Decomposition of DBC:

Period

S.E.

DBC

DRB

1

11.22524

100.0000

0.000000

(0.00000)

2

11.95446

98.13316

1.866842

(6.26946)

3

12.59326

96.38403

3.615970

(6.31787)

4

12.73875

95.23096

4.769042

(7.13437)

5

13.00544

92.54397

7.456027

(8.43985)

6

13.23903

89.56166

10.43834

(10.7352)

7

13.31079

89.52158

10.47842

(11.2716)

8

13.33464

89.31739

10.68261

(11.6842)

9

13.36336

88.98118

11.01882

(12.8840)

10

13.42741

88.72165

11.27835

(13.5373)

Source: Prepared by the researcher using Eviews10.

From results of the analysis of variance, which are

shown in the previous table, we can see that in the

short term (the future second year), 98.13% of the

variance of the forecast error of the trade balance

deficit is due to its own shocks, while the general

budget deficit contributes to 1.86% in explaining the

variance of error forecasting. The percentage of

explanation of the error variance in relation to the

general budget deficit rises to 4.76% during the

fourth year. In the medium term (the future fifth

year), the trade balance deficit contributes about

95.23% to the interpretation of its forecast, while the

general budget deficit contributes 7.45% to

explaining the forecast error for the trade balance

deficit. The ratio of explaining the error variation in

relation to the general budget deficit rises to 10.43%

during the sixth year, and then to 11.27% during the

tenth year, to prove these rates in the coming years,

i.e. in the long term. We also conclude that the

results of the variance segmentation analysis are

almost compatible with the role that the general

budget deficit plays in increasing the trade balance

deficit, by increasing government spending, which

leads to an increase in income. An increase in

income leads to an increase in demand for imported

goods, which is reflected in an increase in the

volume of imports compared to exports, and thus an

increase in the trade balance deficit, which is

consistent with the results of the batch response

functions analysis.

5 Conclusion

The analysis in this study focused on measuring the

impact of the general budget deficit on the Algerian

trade balance deficit for the period 1990–2020,

within the framework of the basic postulates of the

Algerian economy and economic theory, using VAR

models. The study was limited in terms of variables

to budget balance statistics, the difference between

public revenue and public expenditure, and the

statistics of the balance of trade balance as the

difference between total exports and total imports,

estimated in billions of US dollars. The study

obtained a set of results presented as follows:

- The general budget witnessed a deficit in most

of the study period (1990–2020), owing to the rise

in public expenditure resulting from the investments

and development programs carried out by the

Algerian government during this period.

- The trade balance during the 1990s fluctuated

between deficit and recovery, but during the period

2000–2014 it recorded a continuous surplus thanks

to the rise in oil prices in the global market.

However, with the drop in prices at the end of 2014,

the trade balance recorded a deficit that continued

until the end of the study period, when the oil price

shock was already so severe that the trade balance

recorded a deficit after several years of successive

surpluses.

- By examining the stability of the time series

using the ADF and PP tests, we found that the two

series are complementary to the same degree (the

first difference). To ensure that there is an

equilibrium relationship between the trade balance

deficit and the public budget deficit in the long term,

we conducted the Johansen test, the results of which

confirmed that there is no relationship between

them.

- It was found through the Granger test of

causality between the trade balance deficit and the

general budget deficit that there is a one-way causal

relationship that runs from the general budget deficit

towards the trade balance deficit.

- By estimating the autoregressive ray model,

we discovered that there is a positive impact of the

deficit of the balance of the public budget on the

deficit of the trade balance. That is, the greater the

deficit of the balance of the public budget, the

greater the deficit of the trade balance.

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1236

Volume 18, 2022

- The occurrence of a random shock in the

general budget deficit by one standard deviation will

have a positive effect on the trade balance deficit.

This positive effect will continue to increase

incrementally from the second year until the end of

the sixth year, when it turns into a weak positive

response until the end of the tenth year, when the

reaction decreases as we move away from the period

of the shock, to fade and become almost non-

existent. The strength of the immediate response

function and the analysis of variance ratios in the

short term is greater than in the long term, and the

overall effect of a positive random shock on the

public budget deficit is the rise in the trade balance

deficit in the short term.

- The strength of the immediate response

function and the analysis of variance ratios in the

short term is greater than in the long term, and the

overall effect of a positive random shock on the

public budget deficit is the rise in the trade balance

deficit in the short term.

Accordingly, and from these results, we accept

the main hypothesis, which is that a positive shock

in the public budget deficit affects the trade balance

deficit in Algeria more in the short term than in the

long term. In light of these results, we recommend

the following suggestions:

 The need to diversify the Algerian economy and

revitalize all its sectors to bring in hard currency

and develop foreign reserves, diversifying

sources of income and not relying entirely on oil

as a main source of income.

 Restoring the information systems entrusted

with measuring and estimating macroeconomic

indicators and providing an information base on

which scientific research and studies are based

in order to derive solutions that serve the

Algerian economy.

 The necessity to adopt a financial and

commercial policy based on the comprehensive

goals of the state and on solid rules and

foundations that enable it to face the

repercussions of internal and external

influences, and to achieve internal and external

economic balance.

 The necessity to rationalize public spending and

limit wasteful public spending in order to reduce

the public budget deficit and to limit the

increase in imports and thus reduce the trade

balance deficit.

References:

[1] Ogba, L., The Impact of Budget Deficit on

Trade Balance in Nigeria: An Empirical

Analysis, 1980–2011, 2014.

[2] Abbasi, E., The Effect of Budget Deficit on

Current Account Deficit: Evidence from

Iran, 2015.

[3] Wakeel, I., Impacts of Budget Deficit on

Output, Inflation and Balance of Trade: A

Simultaneous Equation Model Approach,

2013.

[4] Buiter, W. H., Persson, T., & Minford, P., A

Guide to Public Sector Debt and Deficits,

Economic Policy, Vol.1, No.1, 1985, pp. 13.

https://doi.org/10.2307/1344612

[5] Brender, A., & Drazen, A., How Do Budget

Deficits and Economic Growth Affect

Reelection Prospects? Evidence from a

Large Cross-Section of Countries, 2005.

https://doi.org/10.3386/w11862

[6] Stiglitz, J. E., More Instruments and

Broader Goals: Moving Toward the Post-

Washington Consensus. In Wider

Perspectives on Global Development (pp.

16–48), Palgrave Macmillan, 2005.

[7] Musyoka, H. M., Relationship Between

Budget Deficit and Economic Growth in

Kenya (Doctoral dissertation, University of

Nairobi), 2013.

[8] Allybokus, M., A Vector Autoregressive

(VAR) Approach to the Credit Channel for

the Monetary Transmission Mechanism in

Mauritius, University of Mauritius Research

Journal, Vol.16, 2010, pp. 169–195, p. 175.

[9] Nymoen, R., The VAR and Econometric

Models of the VAR, Department of

Economics University of Oslo, 2011, p. 20.

[10] Canova, F., Estimating Multi-country VAR

Models, Working Paper Series (603), 2006,

pp. 04–30, p. 19.

[11] Nelson, C. R., & Plosser, C. R. (1982).

Trends and random walks in macroeconmic

time series: some evidence and implications.

Journal of monetary economics, 10(2), 139-

162.

[12] MacKinnon J (1991) 'Critical values for

cointegration tests'. in Engle R and Granger

C (eds) Long Run Economic Relatiorchips,

Oxford: Oxford U P.

[13] Patterson, K., An Introduction to Applied

Econometrics: A Time Series Approach,

Palgrave Macmillan, 2002, p. 265.

[14] Borbonnais, R. Économétrie (9th ed.),

Dunod, 2015, p. 279.

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1237

Volume 18, 2022

[15] Faik, B., Stationarity and Cointegration

Tests: Comparison of Engle-Granger and

Johansen Methodologies, Journal of Faculty

of Economics and Administrative Sciences,

Vol.13, 1998, pp. 131–141, p. 136.

[16] Granger, C.W.J. (1969). Investigating

causal relations by econometric models and

cross-spectral models. Econometrica, 37,

424-438.

[17] Stock, J. H., & Watson, M. W., Vector

Autoregressions, Journal of Economic

Perspectives, Vol.15, No.04, 2001, pp. 101–

115, p. 103.

[18] Sims, C. A. (1980). Macroeconomics and

reality. Econometrica: journal of the

Econometric Society, 1-48.

[19] Athanasopoulos, G., Model Selection,

Estimation and Forecasting in VAR Models

with Short-run and Long-run Restrictions,

Working Paper Series (205), 2010, pp. 1–

49, p. 08.

[20] Hafner, C. M., Testing for Linear

Autoregressive Dynamics under

Heteroskedasticity, The Econometrics

Journal, Vol.3, 2000, pp. 177–197, p. 180.

[21] Hossain, A., Vector Autoregressive (VAR)

Modeling and Projection of DSE, Chinese

Business Review, Vol.06, June 02, 2015, pp.

273–289, p. 280.

[22] MacKinnon 1991 MacKinnon J (1991)

'Critical values for cointegration tests'. in

Engle R and Granger C (eds) Long Run

Economic Relatiorchips, Oxford: Oxford U

P.

[23] Otero, J., & Baum, C. F., Unit-root Tests

Based on Forward and Reverse Dickey–

Fuller Regressions, The Stata Journal,

Vol.18, No.01, 2018, pp. 22–28, p. 24.

[24] MacKinnon J (1991) 'Critical values for

cointegration tests'. in Engle R and Granger

C (eds) Long Run Economic Relatiorchips,

Oxford: Oxford U P.

[25] Hjalmarsson, P. B., & Österholm, P.,

Testing for Cointegration Using the

Johansen Methodology When Variables Are

Near Integrated, IMF Working Paper,

Vol.141, No.07, 2007, p. 08.

[26] MacKinnon, J. G., Haug, A. A., &

Michelis, L. (1999). Numerical distribution

functions of likelihood ratio tests for

cointegration. Journal of Applied

Econometrics, 14(5), 563–577.

Creative Commons Attribution License 4.0

(Attribution 4.0 International, CC BY 4.0)

This article is published under the terms of the

Creative Commons Attribution License 4.0

https://creativecommons.org/licenses/by/4.0/deed.en

_US

WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT

DOI: 10.37394/232015.2022.18.115

Abdelkarim Elmoumen,

Naeimah Fahad S Almawishir,

Houcine Benlaria, Taha Khairy Taha Ibrahim

E-ISSN: 2224-3496

1238

Volume 18, 2022