Deakin University: Applied Econometrics Assignment - NBA Players

This assignment solution focuses on applying econometric techniques to analyze data related to NBA players. The analysis begins with an Ordinary Least Squares (OLS) regression model to examine the relationship between points per game and variables such as experience, age, and college years, incorporating a quadratic term for experience. The solution explores the impact of experience on points, determining the experience level at which additional experience reduces points. Further models are developed to assess the influence of age, college years, and player positions (guard, forward, center) on points and wage, including interaction variables. The analysis incorporates multiple regression models, evaluating statistical significance, adjusted R-squared values, and the impact of different variables on player performance and wages. The document also includes a discussion on model selection and interpretation, along with relevant references and bibliography.

Running head: APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Applied Econometrics for Economics and Finance
Name of the Student:
Name of the University:
Author Note:

1APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Table of Contents
Answer 1. A...............................................................................................................................2
Answer 1.b.................................................................................................................................3
Answer 1.c..................................................................................................................................3
Answer 1.d.................................................................................................................................3
Answer 1. E................................................................................................................................4
Answer 1. F................................................................................................................................5
Answer 1. G...............................................................................................................................6
Answer 1. H...............................................................................................................................7
Answer 1. I.................................................................................................................................7
Answer 1. J.................................................................................................................................7
Answer 1. K...............................................................................................................................8
Answer 1. L................................................................................................................................9
Answer 2..................................................................................................................................10
Answer 3. A.............................................................................................................................11
Answer 3. B..............................................................................................................................12
Answer 3. C..............................................................................................................................12
Answer 3. D.............................................................................................................................13
Reference and Bibliography.....................................................................................................14

2APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Answer 1. A
Table 1: OLS regression result for points using quadratic form of experience
The coefficient estimate for experience, squared experience, age and years spent in
college are 2.364, -0.077, -1.07 and -1.286 (rounded up to 4 decimal) respectively. For
example, it can be interpreted as if there is one unit change in experience will increase the
points by 2.364 units.
Standard error of coefficient shows variability of the corresponding variable and the
lower the SE of a variable to its coefficient better the prediction is. For an example, the
standard error of experience is 0.405 implies the prediction will be better as SE is low.
The independent variables that are used in the model are statistically significant at 5%
significance level as the p-values are less than 0.05.
The higher sample size implies lower standard error and narrow confidence interval.
The sample size used in the above regression is 269.

3APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
The adjusted R2 0.1282 which implies the model can explain dependent variable with
12.82% accuracy in the presence of used variables.
Answer 1.b
The following condition needs to be satisfied to reduce the points:
2.364∗exp−0.077∗exp2 <0
2.364∗exp<0.077∗exp2
exp> 2.364
0.077
exp>30.701
After 30 years of experience the points will reduce per game.
Answer 1.c
Coll presents the years of played in college which has a significant negative impact on
the points. This is because the maximum NBA players were drafted from the college that
means most of the NBA players did not complete their college.
Answer 1.d
Controlling the experience and the years spent in college, age and squared have no
significant impact at 5% significance level on points as the p-value is greater than 0.05.
Table 2: OLS regression result for points using quadratic form of experience and age

4APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Answer 1. E
The coefficient estimate for points, experience, squared experience, age and years
spent in college are 0.078, 0.218, -0.007, -0.048 and -0.040 (rounded up to 4 decimal)
respectively. For example, it can be interpreted as if there is one unit change in age, there will
be a reduction in the points by 0.048 units.
The standard error of age and squared age is 0.035 and 0.053 respectively which
implies the prediction will be worse by both age variables as the SE is very high to the
corresponding coefficient.
Both the age variables are not statistically significant at 5% significance level as the
p-values are greater than 0.05.
The adjusted R2 0.4781 which implies the model can explain dependent variable with
47.81% accuracy in the presence of used variables.
Table 3: OLS regression result for lwage using points and quadratic form of experience

5APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Answer 1. F
There is neither joint significance nor separate significance of age and years spent in
college on the log of wage. The p-values are greater than 0.05 for the 3 variables which is
evidence that there is no effect of these variables on the dependent variable.
Table 4: OLS regression result for lwage using interaction variable

6APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Answer 1. G
The coefficient estimate for experience, squared experience, guard and forward are
1.281, -0.072, 2.315 and 1.545 (rounded up to 4 decimal) respectively. For example, it can be
interpreted as if there is one unit change in forward, there will be an increment in the points
by 1.545 units.
The standard error of guard and forward is 1 and 1.002 respectively which implies the
prediction will be worse by both the variables as the SE is very high to the corresponding
coefficient.
Forward variable is not statistically significant at 5% significance level as the p-values
are greater than 0.05 while guard is significant at 5% significance level.
The adjusted R2 0.0772 which implies the model can explain dependent variable with
7.72% accuracy in the presence of used variables.
Table 5: OLS regression result for points using dummy variable considering centers as base
group

7APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Answer 1. H
The reason behind not including all the dummy variable is to avoid linear dependency
among the dummy variables. One more thing is that there may arise heteroskedasticity.
Answer 1. I
The guard scores more than a center as the p-value of the mean coefficient is
significant at 5% significance level. The difference between the points of guard and center is
2.315.
Answer 1. J
There is no significant effect of marital status on the productivity as the p-value is
much higher than 0.05 and the SE is 0.74 which is much higher than the corresponding
coefficient.
Table 6: OLS regression result for points with addition of marital status

8APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Answer 1. K
Both the interaction variables have p-value greater than 0.05 but less than 0.1 which
implies the variables are not significant at 5% significance level but significant at 10%
significance level. So, it can be said that there is no strong relation between interaction
variable and dependent variable.
Table 7: OLS regression result for points using interaction variable of marital status with both
experience variable

9APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Answer 1. L
The adjusted R2 is better in this model than the previous model this implies that the
independent variables used in both models can better explain assists than points (Singh 2016).
Table 8: Regression result for assists using OLS

10APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
Answer 2
The third model is better than the other models in predicting the wage of NBA
players. The OLS 1 and OLS 2 are the linear regression and the other two models are
quadratic model. The adjusted R2 is higher for the OLS 3 model as the R2 value is 0.524107
and the second highest adjusted R2 is for OLS 4 that equals to 0.523578.
Table 9: Results for different models
OLS 1 OLS 2 OLS 3 OLS 4
Variable Coefficie
nt
Std.
Erro
r
Coefficie
nt
Std.
Erro
r
Coefficie
nt
Std.
Erro
r
Coefficie
nt
Std.
Erro
r
Dependent Lwage Lwage Lwage Lwage
C 5.910 0.13
9 5.908 0.13
9 6.022 0.13
4 5.984 0.13
0
AVGMIN 0.021 0.00
8 0.021 0.00
8 0.019 0.00
8 0.019 0.00
8
DRAFT -0.012 0.00
2 -0.012 0.00
2 -0.012 0.00
2 -0.012 0.00
2
EXPER 0.145 0.03
7 0.142 0.03
7 0.093 0.01
6 0.099 0.01
5
EXPERSQ -0.006 0.00
3 -0.006 0.00
3

11APPLIED ECONOMETRICS FOR ECONOMICS AND FINANCE
MARR -0.033 0.07
9 -0.225 0.20
0
POINTS 0.032 0.01
3 0.032 0.01
3 0.033 0.01
3 0.034 0.01
3
EXPER*MARR 0.131 0.05
8 0.074 0.02
8
EXPERSQ*MA
RR -0.013 0.00
4 -0.010 0.00
3
Adjusted R2 0.506 0.508 0.524 0.524
Answer 3. A
The coefficient estimate for inf is 0.641 (rounded up to 4 decimal). It can be interpreted as if
there is one unit change in inf, there will be an increment in i3 by 0.641 units (Wooldridge
2015).
Standard error of coefficient shows variability of the corresponding variable and the
higher the SE of a variable to its coefficient worsen the prediction. The standard error of inf is
0.09 implies the prediction will be better as SE is low.
The independent variable that is used in the model is statistically significant at 5%
significance level as the p-values are less than 0.05.
The higher sample size implies lower standard error and narrow confidence interval.
The sample size used in the regression model is 56.
The adjusted R2 0.4511 which implies the model can explain dependent variable with
45.11% accuracy in the presence of used variables.
Table 10: OLS regression result to determine i3 using inft