Assignment on Econometrics

30/09/2024

Done by
23-PCO-028
23-PCO-029
 Assignment on Econometrics

Introduction:
Econometrics plays a vital role in analysing economic relationships and understanding the
factors that influence various outcomes. One of the essential tools in econometrics is the
linear regression model, which quantifies the relationship between a dependent variable and
one or more independent variables. This assignment focuses on estimating the impact of
education and work experience on salary, specifically within the Indian context.
The linear regression model can be expressed as:
Y=β0+β1X1+β2X2+ϵ
In this equation:
● Y is the dependent variable, representing salary (in Indian Rupees).
● β0 is the intercept, which indicates the expected salary when education and
experience are both zero.
● β1 is the coefficient for the first independent variable (years of education), reflecting
the change in salary for each additional year of education.
● β2 is the coefficient for the second independent variable (years of work experience),
representing the change in salary for each additional year of work experience.
● ϵ is the error term, capturing the variations in salary not explained by education and
experience.
This model provides valuable insights into how education and experience influence salary
levels in India, which is critical for individuals making career decisions and for policymakers
aiming to enhance workforce productivity

The Linear Regression Model:

The model can be formally stated as:
Salary=β0+β1(Years of Education) +β2(Years of Experience)+ϵ
Where:
● Salary: The dependent variable, measured in Indian Rupees (INR).
● Years of Education (X1): The number of completed years of formal education.
● Years of Experience (X2): The number of years worked in a professional capacity.
● β0: Intercept (baseline salary).
● β1: Coefficient for years of education.
 ● β2: Coefficient for years of work experience.
● ϵ: Error term (captures other unobserved factors affecting salary).

Example:
Let’s consider a hypothetical analysis using data from a survey of 150 employees across
various sectors in India. After conducting a linear regression analysis, we arrive at the
following estimated model:
Salary (INR)=300000+20000(Years of Education)+15000(Years of Experience)+ ϵ

Interpretation of Coefficients:
● Intercept (β0=300000): The base salary for an individual with zero years of
education and zero years of experience is ₹300,000.
● Coefficient for Years of Education (β1=20000): For each additional year of
education, the salary increases by ₹20,000, holding work experience constant. This
suggests that higher educational attainment is linked to higher salaries.
● Coefficient for Years of Experience (β2=15000): For each additional year of work
experience, the salary increases by ₹15,000, holding education constant. This
indicates the value of experience in the job market.

Analysis of Results:
After running the regression analysis, additional statistics to include might be:
1. R-squared Value: This value indicates how much of the variability in salary can be
explained by education and experience. A value close to 1 suggests a good fit.
2. Statistical Significance: T-tests for each coefficient should be performed to
determine if the relationships are statistically significant (commonly p < 0.05).
For instance, if the output shows that both coefficients are statistically significant, it would
imply that both education and work experience significantly affect salary in the Indian
context.

Application of the Salary Model: Autocorrelation, Multicollinearity, and

Heteroskedasticity
1. Autocorrelation
Autocorrelation occurs when the residuals of the regression model are correlated with one
another, often seen in time series data.
Application in Salary Model:
Context: If you are analysing the salary data collected over multiple years for the same
individuals, it’s crucial to check for autocorrelation, as the salary of an individual in one year
may influence their salary in the following years due to various factors (e.g., promotions,
inflation).
Testing for Autocorrelation:
Durbin-Watson Test: This is a common statistical test for detecting autocorrelation in the
residuals of a regression analysis. The test statistic is computed based on the differences
between consecutive residuals.
Interpretation:
A value around 2 indicates no autocorrelation.
A value significantly less than 2 suggests positive autocorrelation, which is often a sign that
the model is missing a relevant variable or that the relationship is not adequately modelled.
Example:
Suppose you run the Durbin-Watson test and obtain a statistic of 1.2. This would indicate
positive autocorrelation, suggesting that past salary influences current salary. In this case,
consider including lagged salary values in your model or employing an autoregressive model
to account for this relationship.
Implications:
If autocorrelation is present, the estimated standard errors of the coefficients may be biased,
leading to unreliable hypothesis tests. It may lead to an underestimation of the significance of
variables.

2. Multicollinearity
Multicollinearity occurs when independent variables in a regression model are highly
correlated, making it difficult to determine the individual effect of each variable.
Application in Salary Model:
Context: In the salary model, if "Years of Education" and "Years of Experience" are highly
correlated (e.g., individuals with higher education levels tend to have more years of
experience), it could distort the estimation of their effects on salary.
Testing for Multicollinearity:
Variance Inflation Factor (VIF): Calculate VIF for each independent variable. A VIF value
greater than 10 indicates a high degree of multicollinearity.
Example:
If the VIF for "Years of Education" is 15 and for "Years of Experience" is 12, this suggests
that both variables are highly collinear.
How to Address:
Consider removing one of the variables or combining them into a single index (e.g., a
composite measure of educational attainment and experience).
Implications:
High multicollinearity can lead to inflated standard errors, making it challenging to assess the
significance of the individual predictors. This could obscure the understanding of how each
factor uniquely contributes to salary.

3. Heteroskedasticity
Heteroskedasticity occurs when the variance of the residuals is not constant across all levels
of the independent variables.
Application in Salary Model:
Context: In the salary model, it is plausible that as education and experience increase, the
variance in salaries may also increase (e.g., highly educated individuals may have a broader
range of salaries).
Testing for Heteroskedasticity:
Breusch-Pagan Test: This test regresses the squared residuals on the independent variables
to check for non-constant variance.
Example:
If the Breusch-Pagan test results in a significant p-value (e.g., p < 0.05), this indicates the
presence of heteroskedasticity.
How to Address:
Employ robust standard errors that adjust for heteroskedasticity, or consider transforming the
dependent variable (e.g., using a logarithmic transformation) to stabilize the variance.
Implications:
If heteroskedasticity is present and unaddressed, it can lead to inefficient estimates and biased
standard errors, affecting hypothesis testing and confidence intervals. This can result in
misleading conclusions about the impact of education and experience on salary.
Conclusion:
This econometric analysis has demonstrated the significant impact of education and work
experience on salary levels in India. By employing a linear regression model, we estimated
the relationship between these independent variables and the dependent variable (salary). The
results indicated that each additional year of education correlates with an increase in salary,
and similarly, years of work experience also positively influence earning potential.
Through the examination of key assumptions, including checks for autocorrelation,
multicollinearity, and heteroskedasticity, we ensured the robustness of our model. Addressing
these issues is crucial, as they can distort the reliability of the estimated coefficients and the
overall validity of the analysis. The application of appropriate statistical tests, such as the
Durbin-Watson test for autocorrelation, the Variance Inflation Factor (VIF) for
multicollinearity, and the Breusch-Pagan test for heteroskedasticity, highlighted the
importance of model diagnostics in econometric studies.
Ultimately, the findings from this analysis underscore the value of education and experience
in enhancing salary prospects, offering insights for individuals seeking to invest in their
careers. Furthermore, these results can inform policymakers about the necessity of promoting
educational initiatives and vocational training programs to foster economic growth and
improve workforce productivity.
By acknowledging the limitations of this study and suggesting areas for future research, such
as exploring additional factors influencing salary and expanding the dataset, we pave the way
for a more comprehensive understanding of the dynamics at play in the labour market.