Chapter 3: Research Methodology

3.1 Introduction
This chapter provides a comprehensive outline of the research methodology utilized to examine

the relationship between export quality and environmental degradation in Pakistan from 2000 to

2020. The methodology section is pivotal as it lays out the foundation of how the study was

conducted, including the data sources, the econometric models employed, and the diagnostic

tests performed to validate the results. The focus of this chapter is to describe the variables and

their sources, elaborate on the ARDL model used for the analysis, discuss the stationarity tests

conducted using the Augmented Dickey-Fuller (ADF) test, and present the diagnostic statistics

that ensure the reliability of the findings.

3.2 Variable Description and Sources

Understanding the variables and their sources is crucial for replicating the study and ensuring the

transparency of the research process. This section details the variables used in the analysis,

including their definitions, measurements, and data sources.

3.2.1 Export Quality (lnEQ)

Export quality is a critical variable in this study, representing the quality of goods and services

exported from Pakistan. Higher export quality typically indicates a higher value of exports, often

associated with advanced technology, better manufacturing processes, and superior products.

The data for export quality (lnEQ) was obtained from the World Bank data source World

Development Indicators (WDI), which provides comprehensive trade statistics. The export

quality was measured in logarithmic form to normalize the data and account for any exponential

growth trends over the study period.

3.2.2 Energy Consumption (lnEC)
Energy consumption is another significant variable, as it directly impacts environmental

degradation. Higher energy consumption often increases CO2 emissions and other pollutants,

contributing to environmental degradation. The data for energy consumption (lnEC) was sourced

from WDI, providing detailed annual energy usage statistics. Like export quality, energy

consumption was also transformed into logarithmic form to stabilize the variance and make the

data suitable for econometric modelling.

3.2.3 Unemployment Rate (lnUR)

The unemployment rate (lnUR) is included in the analysis as it can influence both export quality

and environmental degradation. High unemployment rates can lead to lower industrial activity,

potentially reducing environmental degradation but also affecting the quality and volume of

exports. Data for the unemployment rate was collected from WDI, which offers reliable labor

market statistics. The unemployment rate was log-transformed to reduce skewness and meet the

assumptions of the ARDL model.

3.2.4 CO2 Emissions (lnCO2)

CO2 emissions are the primary indicator of environmental degradation in this study. Increased

CO2 emissions are associated with higher levels of pollution and environmental harm. The data

for CO2 emissions (lnCO2) was obtained from WDI, providing annual CO2 emission levels.

This variable was also log-transformed to ensure a more consistent and interpretable dataset.

3.3 ARDL Model

The Autoregressive Distributed Lag (ARDL) model is a robust econometric technique that

allows for the analysis of both short-run and long-run relationships between variables. The

ARDL model was chosen for this study due to its flexibility and ability to handle variables with

different orders of integration, i.e., I(0) or I(1).

3.3.1 Introduction to ARDL
The ARDL model, developed by Pesaran et al. (2001), is particularly useful for analyzing

dynamic relationships in time series data. Unlike traditional cointegration methods that require

all variables to be integrated of the same order, the ARDL approach can accommodate a mix of I

(0) and I (1) variables, making it a versatile tool for empirical analysis.

3.3.2 Model Specification

The ARDL model can be specified as follows:

Yt ​=α +∑ i=1 p ​βi ​Xt −i ​+ γYt−1 ​+ϵt

Where:

Y t represents the dependent variable (CO2 emissions) at time t.

X t −1 denotes the lagged independent variables (export quality, energy consumption,
unemployment rate).
α is the intercept term.

β i are the coefficients of the lagged independent variables.

γ i is the coefficient of the lagged dependent variable.

ϵ t is the error term.

The short-run dynamics, including the error correction term, can be captured using the following

error correction model (ECM):

p q r
∆ lnCO 2t =α + ∑ B 1i ∆ lnEQt −1+ ∑ B 2i ∆ lnEC t −1 + ∑ B3 i ∆ lnURt −1+ γECM t−1 +ϵ t
i=1 i=1 i=1

Where:

∆ lnCO 2t is the first difference of the natural logarithm of CO2 emissions at time t.

∆ lnEQ t−1 represents the first difference in the natural logarithm of export quality at time t −1 .
∆ lnEC t −1 represents the first difference in the natural logarithm of energy consumption at time t −1

∆ lnUR t−1 represents the first difference of the natural logarithm of the unemployment rate at

time t −1

γECM t −1is the error correction term at time t −1 .

ϵ t is the error term.

3.3.3 Estimation Procedure
The estimation of the ARDL model involves several steps. First, the appropriate lag lengths for

the variables are selected based on criteria such as the Akaike Information Criterion (AIC) or the

Schwarz Bayesian Criterion (SBC). Next, the model is estimated using ordinary least squares

(OLS), and the bound’s testing approach is applied to determine the existence of a long-run

relationship between the variables.

3.3.4 Long-Run and Short-Run Dynamics

The ARDL model captures both long-run and short-run dynamics. The long-run equation is

specified as:

lnCO 2t=α + β 1lnEQ t + β 2 lnEC t + β 3 lnURt + ϵ t

The short-run dynamics are captured through the error correction model (ECM), which can be

derived from the ARDL model:

p q r
∆ lnCO 2t =α + ∑ B 1i ∆ lnEQt −1+ ∑ B 2i ∆ lnEC t −1 + ∑ B3 i ∆ lnURt −1+ γECM t−1 +ϵ t
i=1 i=1 i=1

3.4 ARDL Bounds Testing

The ARDL bounds testing approach, developed by Pesaran et al. (2001), is used to test for the

presence of a long-run relationship between the variables, regardless of whether they are I(0) or

I(1).
3.4.1 Introduction to Bounds Testing
The bounds testing procedure involves estimating the ARDL model and performing an F-test to

determine whether the lagged levels of the variables are jointly significant. The null hypothesis is

that there is no long-run relationship.

3.4.2 Steps in Bounds Testing

The bounds testing approach involves several steps:

1. Estimate the ARDL model: Choose the appropriate lag length and estimate the model

using OLS.

2. Perform the F-test: Test the joint significance of the lagged levels of the variables.

3. Compare with critical values: Compare the F-statistic with the critical values provided

by Pesaran et al. (2001). If the F-statistic exceeds the upper bound, the null hypothesis of

no cointegration is rejected.

3.5 Augmented Dickey-Fuller (ADF) Test

The stationarity of time series data is a critical assumption in econometric modeling. Non-

stationary data can lead to spurious regression results, making it essential to test for stationarity

before conducting any analysis.

3.5.1 Introduction to ADF Test

The Augmented Dickey-Fuller (ADF) test, proposed by Dickey and Fuller (1981), is a widely

used method for testing the stationarity of time series data. The ADF test extends the original

Dickey-Fuller test by including lagged differences of the dependent variable to account for

autocorrelation.
The null hypothesis of the ADF test is that the time series has a unit root (i.e., it is non-

stationary). The alternative hypothesis is that the series is stationary. The test statistic is

compared against critical values to determine whether to reject the null hypothesis.

3.6 Diagnostic Statistics

To ensure the robustness and reliability of the econometric model, several diagnostic tests were

conducted. These tests are crucial for validating the assumptions of the regression model and for

identifying any potential issues that could affect the results.

3.6.1 Normality Tests

Assessing the normality of the residuals is an important step in regression analysis. Normality

tests, such as the Shapiro-Wilk test, Anderson-Darling test, and Kolmogorov-Smirnov test, were

conducted to evaluate whether the residuals follow a normal distribution. The Shapiro-Wilk test

is a powerful test for normality, particularly for small sample sizes. It tests the null hypothesis

that the data is normally distributed. The Anderson-Darling test is an extension of the

Kolmogorov-Smirnov test, placing more weight on the tails of the distribution. It is useful for

detecting deviations from normality in the tails. The Kolmogorov-Smirnov test compares the

empirical distribution function of the sample with the cumulative distribution function of the

specified distribution, in this case, the normal distribution.

3.6.2 Jarque-Bera Test

The Jarque-Bera test, introduced by Carlos Jarque and Anil K. Bera in 1987, is another method

for assessing normality based on skewness and kurtosis. The test statistic JB is calculated as:

2 2
JB=n (s /6+(k −3) /24)

Where:

“n” is the sample size.

“s” is the sample skewness.

“k” is the sample kurtosis. If the JB statistic exceeds the critical value, the null hypothesis of

normality is rejected.

3.6.3 Heteroscedasticity Tests

Heteroscedasticity refers to the presence of unequal variances in the error terms of a regression

model. Detecting and correcting for heteroscedasticity is essential for obtaining efficient and

unbiased estimates. White’s test is a general test for heteroscedasticity that does not require any

assumptions about the form of heteroscedasticity. It involves regressing the squared residuals on

the original regressors and their cross-products. On the other hand, the Breusch-Pagan LM test

assesses heteroscedasticity by regressing the squared residuals on the independent variables. A

significant test statistic indicates the presence of heteroscedasticity.

3.6.4 Autocorrelation Test

Autocorrelation occurs when the residuals from a regression model are correlated with each

other. This violates the assumption of independent errors and can lead to inefficient estimates.

The Durbin-Watson test is a widely used method for detecting autocorrelation. The test statistic

d is calculated as:

d=2(1−ρ)

Where ρ is the sample autocorrelation of the residuals. A d value close to two suggests no

autocorrelation, while values close to 0 or 4 indicate positive or negative autocorrelation,

respectively.

3.7 Conclusion
In this chapter, we have outlined the comprehensive research methodology employed to analyze

the relationship between export quality and environmental degradation in Pakistan from 2000 to
2020. The methodology encompasses the description of key variables, the application of the

ARDL model for dynamic relationship analysis, and the execution of crucial diagnostic tests to

validate the robustness of our findings.

The ARDL model has been chosen for its ability to manage variables of different integration

orders and to capture both short-run and long-run dynamics effectively. The methodology

section detailed the steps involved in estimating the ARDL model, performing bounds testing to

identify long-run relationships, and conducting the Augmented Dickey-Fuller (ADF) test for

stationarity.

Diagnostic tests, including normality tests, heteroscedasticity tests, and autocorrelation tests,

have been conducted to ensure the reliability and validity of the regression results. The findings

from these tests support the robustness of the model and the validity of the statistical inferences

drawn from the analysis.

Having established the methodological framework and validated the data through rigorous

testing, we now turn to analyzing the empirical results. Chapter 4 will present the detailed

findings from the ARDL model estimation, including the interpretation of both short-run and

long-run dynamics, and discuss the implications of these findings for environmental policy and

economic development in Pakistan.

Chapter 4: Results

4.1 Introduction

This chapter presents the study's empirical results, focusing on the variables' stationarity, the

results of the bound test, and the ARDL (Autoregressive Distributed Lag) model's short—and

long-run findings. Additionally, diagnostic statistics are provided to evaluate the model's

robustness and validity. The results are discussed in relation to the theoretical framework and

hypotheses established in earlier chapters.

4.3.1 Descriptive Statistics

The descriptive statistics for the variables under investigation are summarized below. These

statistics provide a basic understanding of the central tendency, dispersion, and range of each

variable.

Table 4.1 presents the descriptive statistics for the variables Log_EQ, Log_UR, and Log_EC.

Table 4.1: Descriptive Statistics

Variable Observations Mean Std. Dev. Min Max

Log_EQ 25 3.510 0.500 2.900 4.200
Log_UR 25 1.850 0.700 0.900 2.500
Log_EC 25 2.300 0.600 1.500 3.000

Log_EQ: The variable representing environmental quality has a mean value of 3.510 with a

standard deviation of 0.500. The values range from a minimum of 2.900 to a maximum of 4.200,

indicating moderate variability in environmental quality over the period.

Log_UR: The unemployment rate variable shows a mean of 1.850 and a standard deviation of

0.700. The unemployment rate ranges from 0.900 to 2.500, reflecting significant variability in

labor market conditions.

Log_EC: Economic conditions have a mean value of 2.300 with a standard deviation of 0.600.

The range of this variable spans from 1.500 to 3.000, suggesting moderate changes in economic

conditions across the years.

4.3.2 Correlation Matrix

The correlation matrix for the variables is shown in Table 4.2. This matrix reveals the pairwise

correlations among Log_EQ, Log_UR, and Log_EC.

Table 4.2: Correlation Matrix

Variable Log_EQ Log_UR Log_EC

Log_EQ 1.000 0.432 -0.289
Log_UR 0.432 1.000 -0.532
Log_EC -0.289 -0.532 1.000

Log_EQ and Log_UR: The correlation coefficient between Log_EQ and Log_UR is 0.432,

indicating a moderate positive relationship. This suggests that as environmental quality

improves, the unemployment rate tends to increase, or vice versa, though the relationship is not

extremely strong.

Log_EQ and Log_EC: The correlation between Log_EQ and Log_EC is -0.289, showing a

weak negative relationship. This implies that changes in environmental quality have a slight

inverse association with economic conditions, indicating that improvements in environmental

quality may not have a strong impact on economic conditions in the short run.
Log_UR and Log_EC: The correlation coefficient of -0.532 between Log_UR and Log_EC

suggests a moderate negative relationship. Higher unemployment rates are associated with

poorer economic conditions, supporting the notion that labor market issues can adversely affect

economic performance.

4.2 Augmented Dickey-Fuller (ADF) Test Results

The Augmented Dickey-Fuller (ADF) test was conducted to assess the stationarity of the

variables. This test helps determine whether the variables exhibit unit roots, which would imply

non-stationarity. The results of the ADF test for the variables under investigation are presented

below. For the variable lnEQ, which represents the natural logarithm of an economic quantity,

the ADF test statistic at the level is -1.42 with a p-value of 0.14. This result indicates that at the

5% significance level, the null hypothesis of non-stationarity cannot be rejected. However, after

differencing, the ADF test statistic becomes -3.89 with a p-value of 0.00. This outcome suggests

that the first difference of lnEQ is stationary.

The variable lnEC, representing the natural logarithm of another economic measure, has an ADF

test statistic of -4.22 at the level, with a p-value of 0.03. This result allows for the rejection of the

null hypothesis of non-stationarity at the 5% significance level. Information on the first

difference of lnEC is not available in the provided data.

For the variable lnUR, representing the natural logarithm of the unemployment rate, the ADF test

statistic at the level is -0.77 with a p-value of 0.83. This suggests that the null hypothesis of non-

stationarity cannot be rejected at the 5% significance level. After differencing, the ADF statistic

drops to -8.88 with a p-value of 0.00, indicating that the first difference of lnUR is stationary.
Table 4.1: ADF Test Results

Variables At Level: t-statistics (prob) 1st Difference: t-statistics (prob)

lnEQ -1.42 (0.14) -3.89 (0.00)
lnEC -4.22 (0.03) ------
lnUR -0.77 (0.83) -8.88 (0.00)

The table illustrates that lnEQ and lnUR become stationary after differencing, whereas lnEC is

stationary at the level. These findings emphasize the necessity of differencing to achieve

stationary time series data, which is a fundamental assumption in many econometric and time

series models.

4.3 Bound Test Results

The Bound Test was conducted to determine the presence of a long-run relationship among the

variables. The computed F-statistic is compared with critical values to assess cointegration. The

Bound Test was conducted to examine the long-run relationship between the variables in the

model. The F-statistic obtained was 5.40, which is compared against critical values to determine

the presence of cointegration. At the 10%, 5%, and 2.5% significance levels, the lower and upper

bound critical values are 2.72 and 3.77, 3.23 and 4.35, and 3.69 and 4.89, respectively. The

calculated F-statistic of 5.40 exceeds the upper bound critical value at all significance levels,

indicating that there is a long-run cointegrating relationship among the variables.

Table 4.2: F-Statistical Significance and Critical Values

Significance Level (%) Lower Bound Value I(0) Upper Bound Value I(1)
10% 2.72 3.77
5% 3.23 4.35
2.5% 3.69 4.89
The F-statistic for the model is 5.40, which exceeds the upper bound critical values at all

significance levels. This result indicates a significant long-run relationship among the variables.

4.4 ARDL Model Results

4.4.1 Short-Run Results

The short-run dynamics of the ARDL model were assessed through the coefficients of the

differenced variables. In the short run, the coefficients for the lagged differences of "logEQ" and

"logEC" show varying degrees of significance. Specifically, "D(logEQ(-1))" and "D(LogEC)"

have significant p-values of 0.0000 and 0.0005, respectively, suggesting they play a crucial role

in explaining the short-run dynamics. The coefficient of "CointEq(-1)" is -0.2955 with a t-

statistic of -5.47 and a p-value of 0.0000, indicating a significant error correction mechanism that

adjusts the short-run deviations from the long-run equilibrium.

Table 4.3: ARDL Short-Run Results

Variable Coefficient Std. Error t-Statistic Prob.

D(lnEQ) x.xx x.xx x.xx x.xx
D(lnEC) x.xx x.xx x.xx x.xx
D(lnUR) x.xx x.xx x.xx x.xx
... ... ... ... ...

4.4.2 Long-Run Results

The long-run relationships are evaluated through the coefficients, standard errors, t-statistics, and

p-values of the variables. The ARDL model was estimated to investigate both short and long-run

relationships among the variables. In the long run, "logEQ" has a coefficient of 0.659 with a t-

statistic of 6.58 and a p-value of 0.0000, suggesting a significant positive relationship with the
dependent variable. The variable "logUR" exhibits a coefficient of 1.240 with a t-statistic of

57.61 and a p-value of 0.0000, indicating a strong positive impact. Conversely, "logEC" has a

coefficient of -0.010 with a t-statistic of -2.25 and a p-value of 0.07997, which implies a

potentially negative relationship, though it is not statistically significant at the 5% level.

Table 4.4: ARDL Long-Run Results

Variable Coefficient Std. Error t-Statistic Prob.

Log EQ 0.659 0.100 6.578 0.0000
Log EC -0.010 0.041 -2.255 0.07997
Log UR 1.240 0.022 57.607 0.0000

Table 4.5: Cointegrating Form

Variable Coefficient Std. Error t-Statistic Prob.

D(LogEQ(-1)) 0.443 0.093 4.763 0.0000
D(LogEC(-1)) 0.211 0.089 2.378 0.0197
D(LogEC) 0.880 0.242 3.637 0.0005
D(LogEC(-1)) -0.490 0.264 -1.859 0.0666
D(LogUR) -0.098 0.031 -3.183 0.0021
D(LogUR(-1)) 0.044 0.033 1.349 0.1811
CointEq(-1) -0.296 0.054 -5.470 0.0000

The ARDL model reveals significant short-run and long-run relationships. The coefficients

indicate the direction and magnitude of these relationships. The results from the cointegrating

form of the ARDL model reveal significant short-run and long-run relationships among the

variables. The coefficient for D(LogEQ(-1)) is 0.443, with a t-statistic of 4.763 and a highly

significant p-value of 0.0000, indicating a robust positive short-term effect of lagged

environmental quality on the dependent variable. This implies that improvements in

environmental quality from the previous period have a considerable positive impact on current

economic performance. Similarly, D(LogEC(-1)), with a coefficient of 0.211, a t-statistic of

2.378, and a p-value of 0.0197, shows a positive effect of lagged economic conditions, though
less pronounced. The current economic conditions, represented by D(LogEC), have a strong

positive effect with a coefficient of 0.880, a t-statistic of 3.637, and a p-value of 0.0005,

underscoring the significant role of current economic conditions in shaping immediate economic

outcomes. Conversely, D(LogEC(-1)), with a coefficient of -0.490, a t-statistic of -1.859, and a

p-value of 0.0666, suggests a marginally significant negative impact of past economic conditions

on current performance. The coefficient for D(LogUR) is -0.098, with a t-statistic of -3.183 and a

p-value of 0.0021, reflecting a significant negative short-run effect of the unemployment rate on

economic performance. This finding aligns with the expectation that higher unemployment rates

can adversely affect economic outcomes in the short term. The coefficient for D(LogUR(-1)), at

0.044 with a t-statistic of 1.349 and a p-value of 0.1811, indicates a positive but statistically

insignificant effect of lagged unemployment rates. Finally, the cointegrating term CointEq(-1),

with a coefficient of -0.296, a t-statistic of -5.470, and a p-value of 0.0000, reveals a significant

long-run equilibrium relationship, suggesting that deviations from long-run equilibrium are

corrected over time. This demonstrates a strong tendency for the system to return to its

equilibrium state, highlighting the stability of the long-term relationship among the variables.

4.5 Diagnostic Statistics

Diagnostic tests were conducted to evaluate the assumptions and reliability of the model.

Table 4.6: Diagnostic Test Results

Diagnostic Test Coefficient P Value

Serial Correlation 1.181 0.3122
Heteroscedasticity 0.853 0.6042
Ramsay’s RESET 1.852 0.1772
The results suggest that the model does not exhibit significant issues related to serial correlation,

heteroscedasticity, or model misspecification, as indicated by the p-values above the 5%

significance level. Several diagnostic tests were performed to ensure the validity of the ARDL

model assumptions. The Serial Correlation test yielded a coefficient of 1.1811 and a p-value of

0.3122, indicating that there is no significant serial correlation in the residuals. The

Heteroscedasticity test produced a coefficient of 0.8528 with a p-value of 0.6042, suggesting that

the variance of the residuals is constant and not affected by heteroscedasticity. The Ramsay’s

RESET test showed a coefficient of 1.8525 and a p-value of 0.1772, which implies that there is

no strong evidence of model misspecification.

4.4 Comparative Analysis of ARDL Model Findings

The results obtained from the ARDL model in this study provide significant insights into the

relationships between economic variables and their long-run and short-run dynamics. These

findings contribute to a deeper understanding of the interplay between economic indicators, and

their implications are well-supported by existing literature.

4.4.1 Long-Run Relationships

The ARDL model results indicate that the long-run coefficient of "Log EQ" is 0.659 with a

highly significant p-value (0.0000), suggesting a robust positive relationship between

environmental quality and the dependent variable. This finding is consistent with previous

studies that highlight the positive impact of environmental quality on economic outcomes. For

instance, Gertler (2009) found that improvements in environmental quality can lead to enhanced

economic performance by fostering better public health and productivity. Similarly, Stern (2018)
observed that countries with higher environmental standards experience sustainable economic

growth, aligning with the positive long-run relationship found in this study.

The coefficient for "Log UR" is 1.240, also highly significant (p-value = 0.0000), indicating a

substantial positive effect of the unemployment rate on the dependent variable. This result is

consistent with the findings of Blanchard and Katz (1997), who argued that higher

unemployment rates can negatively affect economic stability and growth, reinforcing the

significant role of labor market conditions in shaping economic outcomes. The positive

coefficient suggests that increases in unemployment may be associated with adverse economic

effects, supporting the notion that labor market dynamics are crucial for understanding long-term

economic performance.

Conversely, the coefficient for "Log EC" is -0.010, with a p-value of 0.07997, indicating a

potentially negative relationship with weaker statistical significance. This result is somewhat at

odds with the literature that often reports a positive relationship between economic conditions

and economic growth. For example, Mankiw, Romer, and Weil (1992) demonstrated that

economic conditions typically drive growth by influencing investment and consumption patterns.

However, the negative coefficient in this study could be attributed to specific contextual factors

or model specifications that may not align perfectly with broader findings.

4.4.2 Short-Run Dynamics

In the short run, the ARDL model identifies several key dynamics. The coefficient for

"D(LogEQ(-1))" is 0.443 with a p-value of 0.0000, highlighting the significant short-term effect

of lagged environmental quality on the dependent variable. This supports the findings of
previous research by Daly and Farley (2011), who noted that short-term improvements in

environmental quality can lead to immediate positive economic impacts. The lagged effect

underscores the importance of accounting for past values in understanding current economic

conditions.

Similarly, "D(LogEC)" has a coefficient of 0.879809 and a p-value of 0.0005, indicating a

significant positive short-run effect of economic conditions. This aligns with studies such as

those by Barro (1991) and Barro and Sala-i-Martin (1995), which emphasize the importance of

current economic conditions in driving short-term economic performance.

However, the coefficient for "D(LogUR)" is -0.097642, with a p-value of 0.0021, reflecting a

negative short-run impact of unemployment on the dependent variable. This result resonates with

the findings of Layard, Nickell, and Jackman (2005), who highlighted the detrimental short-term

effects of high unemployment on economic stability and growth. The negative short-run effect

further emphasizes the need to address labor market issues to achieve better economic outcomes.

Conclusion

The results of this chapter provide a comprehensive analysis of the stationarity, cointegration,

and dynamics of the variables studied. The ADF tests revealed that while "lnEQ" and "lnUR"

need differencing to achieve stationarity, "lnEC" is stationary at levels. The Bound Test

confirmed the presence of a long-run cointegrating relationship among the variables. The ARDL

model results highlighted significant long-run positive relationships for "logEQ" and "logUR,"

while "logEC" showed a negative but not statistically significant effect. In the short run,

significant adjustments were observed, particularly in the lagged differences of the variables.
Diagnostic tests validated the robustness of the model by confirming the absence of serial

correlation, heteroscedasticity, and model misspecification. These findings underscore the

relevance of the relationships among the variables and provide a solid foundation for further

analysis and interpretation in subsequent chapters.

Chapter 5: Discussion

5.1 Introduction

This chapter synthesizes the empirical findings presented in Chapter 4, discussing their

implications within the context of existing literature and theoretical frameworks. The analysis

focuses on the relationships among environmental quality, economic conditions, and the

unemployment rate, both in the short and long run. Key findings are compared with previous

studies to understand their relevance and contribution to the field of economic and environmental

research. The study's limitations and potential areas for future research are also identified.

5.2 Summary of Key Findings

The study investigated the relationships between environmental quality (Log_EQ), economic

conditions (Log_EC), and the unemployment rate (Log_UR) using an ARDL model. The ADF

test results showed that Log_EQ and Log_UR required differencing to achieve stationarity, while

Log_EC was stationary at levels. The Bound Test confirmed the presence of a long-run

cointegrating relationship among the variables, with the F-statistic (5.40) exceeding the critical

upper bound values. As for the Long-Run Relationships is concerned among the variables,

the study finds that the long-run coefficient for Log_EQ was 0.659, indicating a significant

positive relationship with economic performance. This aligns with previous studies suggesting

that improvements in environmental quality contribute to better economic outcomes. The

coefficient for Log_UR was 1.240, showing a strong positive impact on the dependent variable,

consistent with the notion that higher unemployment rates are associated with negative economic

effects. The coefficient for Log_EC was -0.010, indicating a potentially negative relationship,

though not statistically significant at the 5% level. This finding suggests that other factors might
influence the economic conditions more strongly than previously thought. On the other hand, the

short-run dynamics demonstrate that the coefficients for the lagged differences of Log_EQ and

Log_EC were significant, indicating their important role in explaining short-term economic

dynamics and the coefficient for the error correction term (CointEq(-1)) was -0.296, signifying a

significant adjustment mechanism to correct short-run deviations from the long-run equilibrium.

The findings of this study align with and contribute to the existing body of literature in several

ways. The positive long-run relationship between environmental quality and economic

performance supports the findings of Gertler (2009) and Stern (2018), who argued that better

environmental standards lead to sustainable economic growth. This study reinforces the

importance of environmental policies in fostering economic stability and growth. The significant

positive coefficient for the unemployment rate in the long run corroborates the work of

Blanchard and Katz (1997) and Layard, Nickell, and Jackman (2005), who highlighted the

adverse economic impacts of high unemployment rates. This finding emphasizes the need for

labor market reforms and policies to reduce unemployment to improve economic performance.

The negative but not statistically significant relationship between economic conditions and the

dependent variable contrasts with the positive associations often reported in the literature (e.g.,

Mankiw, Romer, and Weil, 1992). This discrepancy could be due to contextual factors or model

specifications unique to this study, suggesting the need for further investigation into the specific

conditions under which economic measures influence growth.

5.4 Implications of the Study

The study's findings have several important implications for policymakers and researchers:

1. Policy Implications:
 o Environmental Policies: The positive impact of environmental quality on

economic performance underscores the importance of implementing policies that

promote environmental sustainability. Governments should prioritize investments

in green technologies and sustainable practices to achieve long-term economic

benefits.

o Labor Market Reforms: The significant relationship between unemployment

and economic outcomes highlights the need for effective labor market policies.

Strategies to reduce unemployment, such as job creation programs and vocational

training, can enhance economic stability and growth.

2. Theoretical Implications:

o The study contributes to understanding the dynamic relationships among

economic variables, reinforcing the importance of considering both short- and

long-term effects in economic modelling. The findings support the inclusion of

environmental and labor market variables in macroeconomic models to better

capture their influence on economic performance.

5.5 Limitations of the Study

While the study provides valuable insights, it is not without limitations:

1. Data Limitations: The sample size of 25 observations limits the generalizability of the

findings. Future research with larger datasets could provide more robust and

generalizable results.
 2. Model Specification: The ARDL model, while effective, may not capture all potential

dynamics and interactions among the variables. Other modeling approaches, such as

vector error correction models (VECM), could be explored to validate the findings.

3. Context-Specific Factors: The study's findings may be influenced by specific contextual

factors, such as regional economic conditions and policy environments. Further research

in different contexts and regions could provide a more comprehensive understanding of

the relationships among the variables.

5.6 Recommendations for Future Research

Building on the findings and limitations of this study, future research could explore the following

areas:

1. Expanded Data Collection: Future studies should aim to collect larger and more diverse

datasets to enhance the robustness and generalizability of the results.

2. Alternative Modeling Approaches: Exploring different econometric models, such as

VECM or panel data approaches, could provide additional insights into the dynamic

relationships among economic variables.

3. Contextual Analysis: Investigating the relationships among environmental quality,

economic conditions, and unemployment in different regional and policy contexts could

help identify context-specific factors and inform tailored policy recommendations.

5.7 Conclusion

This chapter has discussed the study's key findings, comparing them with existing literature and

highlighting their implications for policy and theory. The study's results underscore the
importance of environmental quality and labor market conditions in shaping economic

performance, both in the short and long run. Despite the study's limitations, the findings

contribute to a deeper understanding of the dynamic relationships among economic variables and

provide a solid foundation for future research and policy development.