HAWASSA UNIVERSITY

AWADA BUSINESS AND ECONOMICS COLLAGE

DEPARTMENT OF ECONOMICS

GROUP ASSIGMENT

Name. ID. No

Dabi Girma.............................................................0135/15

Fetena Mekonin................................................... 1024/15

Fikadu Zeleke.......................................................1025/15

Hawas Gemade..................................................1037/15

Abdurham Gelato.................................................1002/15

Nasru working.....................................................

Ahemad Nur..............................................

SUBMISSION DATE

SUBMITED TO.Mr. Henok S.

I
 Econometrics Group Assignment
TABLE CONTENTS. page

Abstract................................................................................................................................................4
Key words............................................................................................................................................5
Introduction..........................................................................................................................................6
ASSIGNMENT (THESIS) ANALYSIS.............................................................................................7
Marginal Effects And it's interpretation............................................................................................10
Descriptive analysis...........................................................................................................................12
Distriboution of WTP over each only key variables.........................................................................12
Conclusion........................................................................................................................................15
References.........................................................................................................................................16

2
 Econometrics Group Assignment
Acknowledgment
Firstly , We would like to thank our God for his considerable help in everything. Then , We would
like to express our sincere gratitude to our instructor , Mr. Henok Sisay , for his invaluable
guidance , support, and expertise throughout our class periods and labs, which enabling us for the
completion of this assignment . His constructive feedback and encouragement were instrumental
in shaping the direction and quality of this assignment and for our understanding of the subject .
Finally, We thankful to our peers and group members for their collaboration and insightful
discussions , which greatly enriched the analysis.

3
 Econometrics Group Assignment

Abstract
This assignment delve in to the determinants of consumer willingness to pay (WTP) for vehicle
attributes using a Probit regression model. Data from 48 consumers were analyzed, with variables
including cost, income, age, and gender. We perform some excel applications in order to use
descriptive analysis to examine the relationship between dependent variable and key variables
along with probit model. We apply regression diagnostic test for data and confirmed model
robustness (no heteroskedasticity/multicollinearity; residuals normally distributed) after doing
nolinear probability model (LMP). Policy recommendations target gender- and age-specific
marketing strategies for vehicle sellers.

Additionally we examine the results of data through linking our different method of analysis.
According to finding we interpret the output of given data. This study (assignment) suggests likely
summary and recommendations depending on the only results thesis analyzations.

4
Econometrics Group Assignment
Key words
Age

Analysis

Gender

Paradoxical trend

Probit model

Introduction

5
 Econometrics Group Assignment
This report provides a detailed analysis of the willingness to pay (WTP) for the Sahara Water
Project using probit regression and linear regression models.Consumer willingness to pay (WTP)
for vehicle attributes is a critical factor in shaping marketing strategies and pricing policies in the
automotive industry.This study investigates the key factors influencing WTP, focusing on cost,
income, age, and gender, using a Probit regression model.The analysis is based on data from 48
consumers, examining binary WTP responses alongside continuous and dummy explanatory
variables.The Probit model reveals that gender and age are the most significant predictors of WTP.
Females exhibit a 60.6% higher probability of being willing to pay compared to males, while each
additional year of age increases WTP by 1.27%. These findings align with descriptive analysis,
which shows higher WTP among middle-aged and female consumers. Surprisingly, income and
cost, though theoretically relevant, are statistically insignificant, suggesting that non-monetary
factors like preferences and perceived value may dominate WTP decisions.

Diagnostic tests confirm the model's robustness, with no evidence of heteroskedasticity or

multicollinearity, and residuals approximating normality. However, the paradoxical positive
relationship between cost and WTP warrants further qualitative investigation, possibly indicating
perceived quality or small sample limitations.The assignment finished with actionable
recommendations, emphasizing gender- and age-specific marketing strategies, such as highlighting
safety and durability features. Future research should expand the sample size and explore additional
variables to deepen understanding of WTP dynamics.

ASSIGNMENT (THESIS) ANALYSIS

Descriptive analysis

6
 Econometrics Group Assignment
The variables included in our studying.

i. wtp: Binary dependent variable (1=willing to pay, 0=not willing to pay)

ii. cost: Continuous variable .
iii. gender: Binary variable (1=female, 0=male) Independent
iv. income: Continuous variable variables
v. Age: Continuous variable

Facts about variables (our observations

 Cost: it ranges from 12 to 40

 Income:it is largely varied,I.e with min =19043 and max 189876
 Age: cover wide range from young to old
 Gender: include 43.75% female and 56.25% male.

Continuous Statistics
Variables
Min Max Average S.deviation

Income 19043 189,876 34116.27 13,456

Age 17 90 39.833 20.11

Cost 12 40 22.75 5.03

There is high variations income as we have seen from standard deviation. The average annual
income is $34,116.30, with significant variability (standard deviation = $16,890.42). The
maximum income ($189,876) is an outlier compared to the rest of the data.

Gender: The sample is slightly skewed toward males (56.25% male, 43.75% female). Gender is
coded as 1 for female and 0 for male.

7
 Econometrics Group Assignment
Age: The average age of respondents is 39.83 years, with a wide range (17 to 90 years),I.e it
includes from young consumers to old consumers.The standard deviation (20.13) indicates
considerable variability in age even if it not highly variation as income.

Parcentage of Each variables for willingness

Independent Categories WTp (=1) % Not WTP(=0) Total numbers in

variables each categories

Gender Female 15 (71.43%) 6 (28.57 21

Male 3 (11.1%) 24 ( 88.9%) 27

Age <30 0 (0%) 23 (100%) 23

30-60 15 (88.24%) 2(11.76%) 17

>60 3 (37.5%) 5 ( 62.5%) 8

Income <23,000 O (0%) 7 (100%) 7

23000- 29000 4 (25%) 12 (75%) 16

29000-35000 8 (50%) 8 (50%) 16

>3500 8 (88.9%) 1 (11.1%) 9

 Gender category
Males (0): 11.1% (3 out of 27) are willing to pay (WTP), while 88.9% (22 out of 27) are not.
Females (1): 71.4% (15 out of 21) are willing to pay, showing a much higher willingness
compared to males.
 Age Groups
<30 years 0% (0 out of 23) are willing to pay, suggesting strong reluctance among younger
individuals.
30-60 years: 88.24% (15 out of 21) are willing, the highest among all age groups.

8
 Econometrics Group Assignment
>60 years: 37.5% (3 out of 8)are willing, relatively good next to the middle-aged group.
 Income Levels
<23,000: 0% (0 out of 7)WTP, indicating affordability is a major issue. 23,000-29,000: 25% (4
out of 16) WTP, still low but higher than the poorest group.
29,000-35,000: 50% (8 out of 16) WTP, showing a significant increase.
 Key Findings
Females are nearly 4 times more likely to pay than males (71.4% vs. 18.5%).
No one under 30 is willing to pay, while two-thirds of those 30+ are willing.
Income is as a predictor of—WTP jumps from 0% to 88.9% across income
brackets

Probit regression model Analysis

Probit regression Number of obs = 48

LR chi2(4) = 33.59

Prob > chi2 = 0.0000

Log likelihood = -14.957993 Pseudo R2 = 0.5290

wtp Coef. Std. Err. z P>|z| [95% Conf. Interval]

9
Econometrics Group Assignment
cost .1304383 .0795283 1.64 0.101 -.0254344 .286311

gender 1.85319 .5457842 3.40. 0.001 .783473 2.922908

income 2.10e-06 .0000101 0.21 0.836 -.0000178 .000022

age .0364918 .0124031 2.94 0.003 .0121823 .0608014

_cons -5.830308 2.028939 -2.87 0.004 -9.806954 -1.853661

1 . mfx

Marginal effects after probit y = Pr(wtp) (predict) = .29914851

variable dy/dx Std. Err. z P>|z| [ 95% C.I. ] X

cost .0452941 .02798 1.62 0.106. -.009555 .100143 22.75

gender* .6064123 .14528 4.17 0.000 .321668 .891156 .4375

income 7.29e-07 .00000 0.21 0.836. -6.2e-06 7.6e-06 34116.3

age .0126716 .00428 2.96 0.003 .004289 .021054 39.8333

(*) dy/dx is for discrete change of dummy variable from 0 to 1

2 . regress wtp cost gender income age

Source SS df MS Number of obs = 48

Model 6.28699303 4 1.57174826 F(4, 43) = 13.62

Residua. 4.96300697 43 .115418767 Prob > F = 0.0000

Total 11.25 47 .239361702 R-squared = 0.558

Adj R-squared = 0.5178

Root MSE = .33973

wtp Coef. Std. Err. t P>|t| [95% Conf. Interval]

cost .0213233 .011375 1.87 0.068 -.0016171 .0442637

10
Econometrics Group Assignment

gender .4696194 .1044128 4.50 0.000 .2590508 .6801879

income 2.27e-07 2.06e-06 0.11 0.913 -3.93e-06 4.38e-06

age .008607 .0024307 3.54 0.001 .0037049 .0135091

_cons -.6661529. .255286 -2.61 0.012 -1.180988 -.1513183

3 . hettest

Breusch-Pagan / Cook-Weisberg test for heteroskedasticity

Ho: Constant variance Variables: fitted values of wtp

chi2(1) = 0.16 Prob > chi2 = 0.6848

4 . vif

Variable VIF 1/VIF

cost 1.18 0.847082

income 1.12 0.894134

gender 1.12 0.896245

age 1.06 0.939024

Mean VIF 1.12

11
 Econometrics Group Assignment

swilk r

Shapiro-Wilk W test for normal data

Variable Obs W V z Prob>z

r 48 0.95704 1.957 1.428 0.07663

5 . sktest r

Skewness/Kurtosis tests for Normality

Obs. Pr(Skewness) Pr(Kurtosis) adj chi2(2) Prob>chi2

Variable
r 48. 0.2429 0.2631 2.77 0.2501

6 . ovtest

Ramsey RESET test using powers of the fitted values of wtp

Ho: model has no omitted variables

F(3, 40) = 2.73 Prob > F = 0.0567

Diagnostics test

1. Heteroskedasticity Test (hettest):

Through Breusch-Pagan test ( p-value = 0.6848), No evidence of heteroskedasticity

12
Econometrics Group Assignment
2. Multicollinearity (VIF):

Since, All VIF values < 1.18 No multicollinearity concerns

3. Normality of Residuals:

-Shapiro-Wilk test p-value = 0.0766 (borderline)

Skewness/Kurtosis test p-value = 0.2501

looks like normality but likely some suspension!

4. Omitted Variable Test (ovtest):

Ramsey RESET test p-value = 0.0567, Slight no indication of possible omitted variables

5.Good fitness Model: a substantial portion of variance (Pseudo R²=0.529), suggesting good
explanatory power.

Marginal Effects And it's interpretation

From Probit result we observe an expected results,for instance, though not statistically
significant (p=0.101) cost is expected to having storage Positive coefficient (0.130) relation
with dependent variable,the same is true for income but unlikely it doesn't seems that.Thus the
following can be the relationship between key Explanatory variables and explained variable in
the probit regression.

Gender: Highly significant (p=0.001) with positive coefficient , indicating females are more
likely to pay.

Age: Significant positive coefficient 0.036 with (p=) ,suggesting older individuals are more
willing to pay.

From marginal effect in the probit model the following is an interpretation of significant
variables:

13
 Econometrics Group Assignment
Gender: Being female increases probability of WTP by 60.6 percentage points (highly
significant),holding all other factors constant.

Age: The probability of willingness to pay for vehicle is increases by 12.7%, if the Age of the
consumer increases by 1%,holding all other factors constant.Each additional year increases
WTP probability by 1.27 percentage points

Paradoxical trend: from data provide to us and results we face discrepancies between reality
on ground and we get from probit model. This may unsatisfactory for us to say cost is
insignificant to determine the willingness to pay as well for income.

Thus We guess ,and give some possible:

Cost-WTP relationship is counterintuitive, Requires qualitative investigation (e.g., perceived

value) and (we'd expect higher costs to reduce WTP).

Higher cost tiers correlate with higher WTP. Higher costs may signal better quality/service.
Small sample size amplifies noise.

There may be the Limitations for this Paradox

1. Sample Size: With only 48 observations, the analysis may lack power for some effects.

2. Cost Interpretation: The positive cost coefficient requires further investigation.

3. Possibility of Omitted Variables: even if it is acceptable, The RESET test suggests possible
missing variables.

Linking probit result with descriptive analysis ( only for higher significant variables)

Distriboution of WTP over each only key variables

2.1 WTP in case of age group( classifiy in quartiles)

Age group Willing (WTP=1) Not willing (=0)

<30 0 (0%) 23 (100%)

14
 Econometrics Group Assignment

30-60 15 (88.24%) 2(11.76%)

>60 3 (37.5%) 5 ( 62.5%)

Interpretation of finding

 Its more likely that Middle –aged shows the highest from people WTP.Again older
consumers also shows moderate WTP.

Distriboution of WTP by gender

Gender Willing (=1) Not willing (=0) Total

Male 3 24 27

Female 15 6 21

Variables Coefficient S.d error P-value

Gender 1.85319 .5457842 0.001

Age .0364918 .0124031 0.003

Females show significantly higher willingness to pay for the water project compared to males, and
it take approximately 66%,it align with a positive marginal effect of probit mode. Middle-aged and
in somehow older age consumers are more willing to pay than that younger respondents in data.
The probit result with (p=0.003) confirmed this positive relationship as age increases the
willingness also increases.

15
 Econometrics Group Assignment
The insignificance of income in the probit regression model is may be associated with higher
outlier and in turn high variability.Average annual income is $34,116.30, with significant
variability (standard deviation = $16,890.42). The maximum income ($189,876) dis an outlier
compared to the rest of the data. This may be one sourse of above Paradox and insignificance this
variable.

Conclusion
The Probit regression and descriptive analysis reveal that gender and age are the most significant
determinants of consumers' willingness to pay (WTP) for vehicle attributes. The model's
robustness was confirmed through diagnostic tests, including absence of heteroskedasticity and
multicollinearity, with normally distributed residuals.With females showing a 60.6% higher
probability of WTP compared to males (p=0.001) and each additional year of age increasing WTP
by 1.27% (p=0.003). Specifically, female consumers are more likely to express WTP compared to
males, suggesting that marketing strategies should prioritize female preferences, such as safety and
comfort features. Additionally, each year increase in age raises WTP , indicating that older
consumers place higher value on long-term vehicle utility.

Contrary to expectations, income and cost of attributes had statistically insignificant effects
(p>0.10), suggesting that pricing strategies and income levels may not be primary drivers of
WTP.therefore in this contextc The insignificance of income suggests that WTP is driven by

16
 Econometrics Group Assignment
preferences rather than purchasing power, implying that automakers should focus on feature-
based differentiation rather than price adjustments. The positive but insignificant effect of cost (p
= 0.101) further supports this conclusion, as consumers appear less sensitive to price changes in
this market segment.

Recommendations

The findings of this study highlight key factors influencing consumers' willingness to pay (WTP)
for vehicle attributes. Based on the results, the following recommendations are proposed users and
particularly based on willingness of payment Sahara water project.

To enhance market penetration, vehicle sellers should prioritize gender-specific marketing

campaigns, particularly targeting female consumers. Since female consumers exhibit a significantly
higher WTP, marketing efforts should emphasize features that appeal to women, such as safety,
comfort ,and the like. Advertisements could showcase advanced safety technologies (e.g.,
collision avoidance systems) and ergonomic interiors to align with female preferences.
On the other hand Older consumers show increased WTP, suggesting a focus on durability, low
maintenance costs, and resale value in messaging. Dealerships could offer extended warranties or
service packages to attract this demographic.and tailor offerings to middle-aged buyers (30–50
years), who exhibited the highest WTP (53.3%). Further research should investigate the unexpected
insignificance of cost and income by expanding the sample size and incorporating additional
17
 Econometrics Group Assignment
variables (e.g., education, vehicle type). Experimentation with premium pricing for female-oriented
features could also be tested, given the strong gender effect observed. Policymakers and
manufacturers should use these insights to refine product positioning and customer segmentation
strategies. Given that cost is not a significant deterrent, automakers can bundle high-value features
(e.g., infotainment systems, free servicing) instead of competing on price alone.

References

 Econometrics II (Econ 3062) MODULE Prepared By: Kasahun Tilahun

(MSc).( Arba minch university)
 Introduction to Econometrics: Theory and Practice with Stata”Prepared
by:Tesfaye Etensa (Ambo university)
 John H. Aldrich and Forrest D. Nelson.
 Basic econometrics / Damodar N. Gujarati, Dawn C. Porter. — 5th ed.
 Frank, C. R., Jr., Statistics and Econometrics, Holt, Rinehart and Winston,
 Huang, D. S., Regression and Econometric Methods, John Wiley & So ns
 Nachane, Dilip M., Econometrics: Theoretical Foundationsand Empirical Perspectives,
Oxford University Press, New Delhi, 2006

18