You are on page 1of 8

ETW1000 Business and Economic Statistics

Week 2

Learning Objectives
In this case study we will cover issues such as:
Categorical data Scatter plots Index numbers
Pivot tables Normalization of data
Descriptive statistics

We are going to look at two case studies that emphasise the importance of standardising data
in order to make more meaningful comparisons.

The first Case Study looks at comparing across entities (in this case, schools) at a point in
time. Entities we might compare across might be: countries, schools, companies, individuals,
regions, families, staff groups, etc, etc.

The second Case Study looks at comparing values (in this case, housing affordability) across
time. We might do this in evaluating changes in and economy, or identifying company
performance trends, following financial markets, measuring changes in consumer tastes, etc
etc.

Case Study 1: Reconstruction after the 2004 Indian Ocean Tsunami

Background

On December 26, 2004, and earthquake under the ocean near Sumatra created a
Tsunami that caused widespread
devastation and loss of life in a
number of countries, especially
Indonesia, Thailand, Sri Lanka, the
Maldives and India. More than
230,000 people died.

After the Tsunami, there was a need


for a long process of rebuilding and
reconstruction, including replacing
houses, rebuilding schools and roads,
re-establishing businesses, etc.

Who decides the priorities for


spending?

Is the money allocated in a fair way? Do political or ethnic factors influence how money
is spent?

The Data

1
We have a spreadsheet produced by the Department of Education in Sri Lanka that shows
all schools damaged by the Tsunami, and some features of each school location,
number of students, extent of damage, funds allocated for rebuilding, funding body, etc.

With some careful analysis of this data, we can address some important questions about
how the funding is allocated.

1. Overall Costs

Lets look at total cost of repair / rebuild for a school.

The average expenditure per school is 57.8 million Rupee (approximately A$530,000).

Costs range from 37 million to 181 million.


How do we choose what
BUT: Does it really make sense to analyse total cost
measure to use when
of repair? School sizes vary enormously, and costs
standardising?
should be proportional to size. There could be
There is no right measure.
several options for measuring size of a school. Eg
The criterion is to standardise in
Number or students, number of staff, number of
a way that means we are
classrooms, total school budget, etc etc.
comparing like with like.
Lets look at costs per student. In this case, it is our
only available variable of those suggested here.

The mean cost per student is LKR270,000, which To get a quick impression of a set
equates to about A$2500 per student. of data, we would often use:
-the mean (where data is centred)
Note the values range from LKR30,000 to LKR1.23 the range (how data varies
million. around the mean).
Maybe this reflects the differences in the extent of
damage to each school? Or other factors?

2. Political Factors

2
Sri Lanka has endured a long civil war, with internal conflict along ethnic and regional
lines. Until recently the Tamil ethnic group dominated the North-East Districts, with the
rebel group the LTTE virtually running parts of the North-East of the country as a
separate nation.

At the heart of this civil war was a long-


standing view that Tamil-dominated
areas were given much less resources
than the rest of the country.
Infrastructure (roads, electricity, etc) was
vastly inferior, due to political bias from
the ruling Sinhalese majority. Economic
decisions can have big political and
social implications!

We can examine whether any of this


alleged ethnic bias exists in the money
spent on reconstructing schools after the
Tsunami. How? First, we categorise the
district each school is located in by
whether it is in the North-East area
(Tamil-dominated) or not.

Next we construct a pivot table


examining average costs per student
separately for these two categories of
district. Heres what the pivot table
looks like.

Average of Cost per


Student
Not North-East Total
0 0.296287163
1 0.20872212
Grand Total 0.271130355
A pivot table allows us to
compare values of one
Note the average cost / amount spent is actually variable of interest (eg average
higher in the North-Eastern districts. There is no cost per student) across
evidence of discrimination against the Tamils here. different categories of the
other variable (eg location).

3. School Location

After the Tsunami the Government introduced a Law establishing a 100m Exclusion
Zone. Under this law, no building was to take place within 100m of the shore line. This
was to reduce the potential damage should a future Tsunami come.

This Law was very controversial, leading to many protests, as it meant many people
owned land that they previously lived on, but that had now become worthless.

3
For schools, 95 of the 183 schools had to be relocated, either because they were in the
exclusion zone, or for other reasons. One might expect that the cost of rebuilding a
relocated school would be higher. Lets look and see if that is the case:

Average of Cost per Student School Relocate?


North-East region? Not relocated Relocated Grand Total
North-East Region 0.301 0.292 0.296
Other Location 0.191 0.220 0.209
Grand Total 0.275 0.267 0.271

Now we have average costs for 4 different combinations whether the school relocated
or not, and whether the school is in the North-East of the country or not.

4. Economies of Scale? We use a scatter diagram here to


compare across the two variables
Finally, lets look at the relationship between size of because now the variables take
school and total cost of rebuilding. Well do this with a many values (Numerical
scatter diagram (more on this in upcoming lectures). variables). Pivot Tables are used
for Categorical variables.

Est Cost (ms Rs) What do we learn


from this graph?
200
180
Costs
160
increase as the
140
number of students
120
increase
100
80
There seems
60
to be a minimum
40
cost of rebuilding of
20
around LKR40
0
million
0 500 1000 1500 2000 2500 3000 3500

In this next chart, we compare cost per student with number of students.

4
Cost per Student

1.4
1.2
1
0.8
0.6
0.4
0.2
0
0 500 1000 1500 2000 2500 3000 3500

Should schools be rebuilt to the same size as they were before the Tsunami? It would be
better financially to evaluate the ideal school size, and build schools accordingly.

Of course, the choice of school size is not just determined by cost considerations. There
are other issues like access (travel time to school), sense of community ownership of
the school, etc. But this graph above clearly suggests that school below about 500
students are very uneconomical. On the other hand, there are relatively small benefits in
terms of further cost reductions from increasing size much beyond 500 students.

5
Case Study 2: Malaysian Housing Affordability Many important business or
economic questions are about
An overview how things are changing over
A common refrain in Malaysia is that housing time. Eg Are sales improving
is ridiculously unaffordable. Lets investigate! like they should? Are we seeing
To do this we will mainly use data from the a decline in complaints from
Department of Statistics: customers? It is necessary to
o It is the provider of official statistics in standardise before comparing
Malaysia. That is, it is a National values across time, in order to be
Statistical Office (NSO). able to compare like with like.
o Lets have a look at housing
affordability focusing on Malaysia.
_____________________________________________________________________
A first looknominal house prices
o Lets simply look at how house prices have changed over time.
Prices rose 128% since 1995.

Nominal Malaysia House Price


350,000.00

300,000.00

250,000.00

200,000.00
RM

150,000.00

100,000.00

50,000.00

0.00

Year

_____________________________________________________________________

A second lookreal house prices


o Lets look at how real house prices have changed over time. That is
house prices divided by the CPI/100 (Note: CPI=Consumer Price
Index).
Base year of CPI is 2010 (i.e. the year where CPI = 100)
Real prices rose by 42% from 1995.

6
Real and Nominal House Prices
350,000
300,000
250,000
200,000
RM

150,000
100,000
50,000
0

Nominal Malaysia House Price Real Malaysia House Price

A third and final lookmortgage payments to income


o Lets look at how mortgage payments on a house have changed over
time relative to income.
To do this we need to calculate the mortgage payments for a
household buying a home in each city. The Excel formula PMT()
is useful here.
E.g. for a 30 year mortgage, and 90% housing loan,
annual_mortgage_payment =
PMT(interest_rate/100,30,0.9*house_price)
We assume income levels to be at the median level.
We do this for different age groups, for the 4 housing types in
Malaysia: Terrace, high rise, semi-detached and detached.

7
_____________________________________________________________________
What do we conclude?
o The concept of housing affordability is somewhat difficult to pin
down.
o House prices have certainly risen significantly.
o The housing affordability situation is rather acute in Malaysia, with
affordability pressures being the greatest for the youngest and oldest age
cohorts.

You might also like