Office Use Only

Semester Two 2016

Examination Period

Faculty of Business & Economics

EXAM CODES: ETC1000 / ETW1000

TITLE OF PAPER: Business and Economic Statistics - PAPER 1 of 1

EXAM DURATION: 3 hours writing time

READING TIME: 10 minutes

THIS PAPER IS FOR STUDENTS STUDYING AT:( tick where applicable)

Berwick x Clayton x Malaysia Off Campus Learning Open Learning
Caulfield Gippsland Peninsula Enhancement Studies Sth Africa
Parkville Other (specify)

During an exam, you must not have in your possession any item/material that has not been authorised
for your exam. This includes books, notes, paper, electronic device/s, mobile phone, smart
watch/device, calculator, pencil case, or writing on any part of your body. Any authorised items are
listed below. Items/materials on your desk, chair, in your clothing or otherwise on your person will be
deemed to be in your possession.

No examination materials are to be removed from the room. This includes retaining, copying,
memorising or noting down content of exam material for personal use or to share with any other
person by any means following your exam.
Failure to comply with the above instructions, or attempting to cheat or cheating in an exam is a
discipline offence under Part 7 of the Monash University (Council) Regulations.

AUTHORISED MATERIALS

OPEN BOOK YES x NO

CALCULATORS YES x NO

SPECIFICALLY PERMITTED ITEMS YES x NO

Candidates must complete this section if required to write answers within this paper

STUDENT ID: __ __ __ __ __ __ __ __ DESK NUMBER: __ __ __ __ __

Page 1 of 11
INSTRUCTIONS TO CANDIDATES:

• Answer ALL questions in this examination paper.

• Paper is out of 100 marks.

• Time: The aim of this exam is to test your ability to give correct and carefully presented
answers to the questions without the pressure of time. It is expected that most students
ought to be able to complete the paper in well under the allocated three hours.

• Numerical Calculations: Where you are asked to perform calculations, you should write
out the solution as an equation containing the appropriate numerical values from within
the question. Except for simple calculations, you do not need to calculate exact values
in order to receive full marks for that part of the question.

Introduction

This exam will consider subjective poverty among households in Timor-Leste. The data and
analysis are all taken from the Timor Survey of Living Standard from 2007.

Subjective poverty is a measure of poverty based on two simple questions that are asked of
households:

• What is the value of food and other goods your household would need each month
in order to consider themselves “not poor”?
We will call this variable Monthly Need

• What is the value of food and other goods your household actually consumed in the
past month?
We will call this variable Monthly Actual

If the household’s actual consumption is less than what they say is needed, they are classified
as subjectively poor. Otherwise, they are not poor.

Page 2 of 11
Question 1: Monthly Need (35 marks)

a. Consider the following descriptive statistics and histogram of the data on Monthly
Need. Note outliers in the data set have already been removed.

Monthly Need

Mean 166.8189
Standard Error 1.755141
Median 150
Mode 100
Standard Deviation 117.3322
Sample Variance 13766.84
Kurtosis 7.041353
Skewness 1.938552
Range 1000
Minimum 0
Maximum 1000
Sum 745513.5
Count 4469

i. The histogram has several faults in design and presentation. List at least three things
that need to be improved in the histogram.
(3 marks)

ii. Comment on the wide range of values of Monthly Need, by interpreting the range,
minimum, maximum and standard deviation, relative to the mean of the data.
(2 marks)

iii. Can you suggest two possible reasons for why some households might report needing
so much more than other households?
(2 marks)

b. Presumably a household would need more food and other goods if they have more
people. So we next divide Monthly Need by another variable called Household Size.
This gives us Monthly Need per Person (MNPP).

Here are the descriptive statistics for MNPP.

Monthly Need per person

Mean 35.02128
Standard Error 0.438996
Median 27.77778
Mode 50
Standard Deviation 29.34716
Sample Variance 861.2556
Kurtosis 40.01307
Skewness 3.880275
Range 600
Minimum 0
Maximum 600
Sum 156510.1
Count 4469

Page 3 of 11
Comparing the standard deviation of MNPP with that for Monthly Need, we see MNPP seems
to have much less variability. But why would this be an incorrect conclusion to make? In fact,
it is better to say that MNPP has more relative variability than Monthly Need. Explain how we
conclude this.
(3 marks)

c. Now consider the following two regression equations that have been estimated:

Regression 1:
Y = Monthly Need
X = Household Size

Regression 2:
Y = Monthly Need per Person
X = Household Size

Dependent Variable: Monthly Need

Regression Statistics
Multiple R 0.3193803
R Square 0.1020038
Adjusted R Square 0.1018028
Standard Error 111.19954
Observations 4469

Coefficients Standard Error t Stat P-value Lower 95%Upper 95%

Intercept 84.279524 4.024109475 20.94365 5.53E-93 76.39028 92.16877
Household Size 14.822322 0.658017379 22.52573 1.6E-106 13.53228 16.11236

Dependent Variable: Monthly Need per Person

Regression Statistics
Multiple R 0.3834801
R Square 0.147057
Adjusted R Square 0.1468661
Standard Error 27.106574
Observations 4469

Coefficients Standard Error t Stat P-value Lower 95%Upper 95%

Intercept 59.809454 0.980937679 60.97172 0 57.88633 61.73258
Household Size -4.451433 0.16040171 -27.7518 1.6E-156 -4.7659 -4.13697

Interpret the coefficient of Household Size in these two regressions. What do they tell us about
the relationship between need and the number of people in a household?
(4 marks)

Page 4 of 11
d. Some people say that living in a large city makes the cost of living more expensive. In
Timor-Leste there is only one large city, Dili, which is the capital city. In the following
regression we investigate whether the need is higher in Dili than in the rest of the
country.

Y = Monthly Need per Person

X = Dili, which equals 1 if the household is in Dili, 0 otherwise

Dependent Variable: Monthly Need per Person

Regression Statistics
Multiple R 0.025023
R Square 0.000626
Adjusted R Square 0.000402
Standard Error 29.34125
Observations 4469

Coefficients
Standard Error t Stat P-value Lower 95%Upper 95%
Intercept 34.73799 0.47044 73.84157 0 33.8157 35.66029
Dili 2.186512 1.306982 1.672947 0.094408 -0.37582 4.748844

i. Using this regression output, estimate the mean MNPP for households in Dili, and for
households not in Dili.
(2 marks)

ii. Using the p-value approach, test the hypothesis that the households in Dili have greater
average needs than others.
(4 marks)

e. Now consider the following regression:

Y = monthly need per person

X1 = Dili
X2 = Household Size

Dependent Variable: Monthly Need per Person

Regression Statistics
Multiple R 0.393475
R Square 0.154822
Adjusted R Square 0.154444
Standard Error 26.98592
Observations 4469

Coefficients
Standard Error t Stat P-value Lower 95%Upper 95%
Intercept 59.73088 0.976648 61.15903 0 57.81616 61.64559
Dili 7.802528 1.218061 6.405698 1.65E-10 5.414526 10.19053
Household Size -4.61886 0.161813 -28.5445 9.2E-165 -4.93609 -4.30162

i. How would your interpretation of the coefficient of the variable Dili differ in this model
compared to that in part d?
(2 marks)

Page 5 of 11
ii. The coefficient of Dili is much larger in part e than in part d. What can you conclude
from this about the relationship between Household Size, living in Dili and Monthly
Need per Person?
(3 marks)

f. Write a short summary of what we have learned throughout this question about the
needs of households. Your target audience for this summary is a government official
who wants to understand the financial and food needs of households in their country.
(10 Marks)

Page 6 of 11
Question 2: Subjective Poverty (20 marks)

a. Consider the following summary statistics for Monthly Need and Monthly Actual.

Monthly Need Monthly Actual

Mean 166.8189 Mean 113.619

Standard Error 1.755141 Standard Error 2.263517
Median 150 Median 75
Mode 100 Mode 100
Standard Deviation 117.3322 Standard Deviation 151.3174
Sample Variance 13766.84 Sample Variance 22896.96
Kurtosis 7.041353 Kurtosis 258.9019
Skewness 1.938552 Skewness 10.70966
Range 1000 Range 4500
Minimum 0 Minimum 0
Maximum 1000 Maximum 4500
Sum 745513.5 Sum 507763.1
Count 4469 Count 4469

The following variables can also be calculated from the data on Monthly Need and
Monthly Actual:

Surplus = Monthly Actual – Monthly Need

Poor = 1 if Surplus < 0, and 0 otherwise

i. Use some simple mathematics based on the formula for the mean, to show that the
mean of Surplus can be calculated as follows:

Surplus = ℎ − ℎ

Reminder: the formula for the mean is:

= ∑ .
(3 marks)

ii. Estimate the average amount by which households’ monthly need exceeds their actual
income.
(2 marks)

iii. What does the variable Poor tell us?

(1 mark)

b. We now estimate the following regression:

Y = Poor
X = 1 for all households

Page 7 of 11
SUMMARY OUTPUT
Dependent Variable: Poor
Regression Statistics
Multiple R 0.854352
R Square 0.729917
Adjusted R Square 0.729693
Standard Error 0.444052
Observations 4469

Coefficients
Standard Error t Stat P-value Lower 95%Upper 95%
Intercept 0 #N/A #N/A #N/A #N/A #N/A
constant 0.729917 0.006642 109.8866 0 0.716895 0.74294

i. Based on this regression results, estimate the proportion of households who are poor.
(1 mark)

ii. State a 95% confidence interval for the proportion of households who are poor. Give a
technical interpretation of this interval, using the concept of repeated samples.
(3 marks)

c. Consider the following regression:

Y = Poor
X = Dili

Dependent Variable: Poor

Regression Statistics
Multiple R 0.146504
R Square 0.021463
Adjusted R 0.021244
Square
Standard Error
0.43931
Observations 4469

Coefficients
Standard Error t Stat P-value Lower 95%Upper 95%
Intercept 0.755013 0.007044 107.191 0 0.741204 0.768822
Dili -0.1937 0.019569 -9.89847 7.26E-23 -0.23206 -0.15534

i. Estimate the proportion of households who are poor in Dili and outside Dili.
(2 marks)

ii. In Question 1 we showed that Dili residents have higher needs than others, and here we
have found fewer households are poor in Dili than other parts of the country. These
two facts appear to contradict each other: what is the likely explanation?
(2 marks)

d. Write a paragraph telling the Government what we learn through the analysis in this
question about the prevalence of subjective poverty in Timor-Leste.
(6 marks)

Page 8 of 11
Question 3: Food Shortages (30 marks)

One of the most common ways poverty reveals itself is with periods of food shortages. In the
survey households are asked which months of the past year they experienced shortages of food.

a. Households experienced food shortages for an average of 3.3 months. Of those who
experienced one or more months of shortages, the average number of months of
shortages was 4.4. What proportion of households had zero months of food shortages?
(2 marks)

b. Below is the results for the following regression

Dependent Variable: Number of Months of Food Shortages (over the past year)

Explanatory Variables:
Gifts = the total value of food this household has given to needy neighbours and
family over the past year
Cultivate = 1 if the household cultivates crops as one of their main sources of
food, 0 otherwise
Monthly Income = the monthly income of this household
Dili = 1 of the household is based in the capital city Dili, 0 otherwise

Dependent Variable: Months of Food Shortages

Regression Statistics
Multiple R 0.423054
R Square 0.178975
Adjusted R Square 0.17815
Standard Error 1.982349
Observations 3987

Coefficients
Standard Error t Stat
Intercept 3.321419 0.122751 27.05821
Gifts -0.00049 8.06E-05 -6.04798
Cultivate 0.847747 0.108162 7.837773
Monthly Income -0.01005 0.001302 -7.71959
Dili -1.54268 0.11145 -13.8419
P-value Lower 95%Upper 95%
3.6E-148 3.080758 3.562079
1.6E-09 -0.00065 -0.00033
5.84E-15 0.63569 1.059805
1.47E-14 -0.0126 -0.0075
1.36E-42 -1.76119 -1.32418

i. Interpret the coefficient of the intercept in this regression. Does this value make sense?
(2 marks)

ii. Consider the signs of the coefficients of each of the other four variables in this model.
For each variable, give a possible explanation for the direction of effect of that variable
suggested in the estimated coefficients.
(8 marks)

iii. It is claimed that households who give more food away to others are going to
experience food shortages more often. Perform a hypothesis test to consider this claim.
(4 marks)

Page 9 of 11
c. The following table shows the number of households who experienced food shortages
in two months of the year – January and May.

January is a period when food is often scarce, while May is around harvest time for
several crops, so food shortages are much less likely.

Food Consumption in May

More than enough food Just enough food Food shortage Total
Food More than enough food 81 29 4 114
Consumption Just enough food 272 1130 59 1461
in January Food shortage 875 1911 108 2894
Total 1228 3070 171 4469

Based on the values in the table, perform the following calculations.

i. What is the probability that a household experiences a food shortage in January?

(1 mark)

ii. What is the probability that a household experiences a food shortage in May?
(1 mark)

iii. What is the probability that a household will be better off in food consumption in May
than they were in January?
(2 marks)

iv. Of those who experienced food shortage in January, what is the probability of also
experiencing a shortage in May?
(2 marks)

e. Write a paragraph summarising what we learn throughout Question 3 about food

shortages among households.
(8 marks)

Page 10 of 11
Question 4: Improvements over time (15 marks)

A local organization has collected data every year for the past 14 years on poverty and
population distribution in a region of the country.

a. Based on this data they can obtain a model as follows:

Poor = 0.753 – 0.015 time – 0.2 urban + error

where
time measures years, and time = 0 in 2002, 1 in 2003, etc
urban = 1 if the household is based in an urban area, and 0 if rural.

i. Interpret the estimated value of the intercept for this model.

(2 marks)

ii. What does the coefficient of the variable time tell us about the trend in poverty rates.
(2 marks)

iii. Use the model to predict the poverty rate for 2030 for a household in an urban area, and
a household in a rural area.
(3 marks)

b. The coefficient on urban in the model suggests poverty is much less likely in urban
areas. For this reason, a growing number of people are migrating from rural areas into
towns and cities. Based on data collected over the same period as used in the model
above, an analyst has estimated a model of the proportion of the population who live in
urban areas.

UrbanP = 0.25 + 0.01 time + error

where
UrbanP is the proportion of the population who are in urban areas
time is defined the same way as the model in part a.

i. Combine predictions from these two models to predict what the overall national poverty
rate would be in 2015, and in 2030.
(4 marks)

ii. There are several reasons to doubt the poverty rate prediction we have obtained for
2030 using the above models. Can you think of two possible problems with the
analysis? Explain carefully what you think the problem is, and how you think it might
affect the predictions.
(4 marks)

End of Exam Paper

Page 11 of 11