A THESIS
By
DEPARTEMENT OF ECONOMICS
INTERNATIONAL PROGRAM
FACULTY OF ECONOMICS
UNIVERSITAS ISLAM INDONESIA
YOGYAKARTA
2006
TABLE OF CONTENTS
Page of title i
Approval page ii
Acknowledgements iv
Devotion page vi
Motto vii
Lists of table ix
List of graph x
List of appendices xi
CHAPTER 1: INTRODUCTION
viii
CHAPTER 3: THEORITICAL BACKGROUND AND HYPOTHESIS
4.7.1.1 Multicollinearity 50
4.7.1.2 Heteroscedasticity 51
4.7.1.3 Autocorrelation 52
ix
4.7.2 Time Series Regression 53
4.7.3 T- Statistic Testing 54
4.7.4 F- Statistic Testing 55
4.7.5 PAM regression 55
4.7.6 The Multiple Regression Coefficient of Determination (R²) 55
x
LIST OF TABLE
Page
Data
Calculation of data change to Ln
Data with Ln
Heteroscedasticity
Autocorrelation
Multicollinearity
Linear
Log- linear
MWD multiple regression Linear
MWD multiple regression Log- linear
PAM linear
PAM log- linear
Unit root test
Dummy to prove negative exchange rate
ABSTRACT
Since long time, the main destination of Indonesian tobacco has been were
The United States, Japan, Malaysia, Belgium and Luxemburg, Netherlands,
Germany, France and Spain. But the main destination is to The United States and
the volume of tobacco export to USA is more stable than to another countries.
Indonesia has several kinds of tobacco export, they are: tobacco not stripped
Virginia type flue crude, tobacco not stripped, tobacco wholly/ partly stripped
Virginia type flue crude, and tobacco wholly/ partly stripped. All the fact above
interest the researcher, in holding the research to recognize the flows of export
tobacco Indonesia to The United States. Hence, the title of this thesis is: THE
ANALYSIS OF FACTORS AFFECTING INDONESIA’S TOBACCO
EXPORTS TO THE UNITED STATES: A PARTIAL ADJUSTMENT
MODEL (1981- 2001)
The results of this research are: Price of tobacco has a significant and
positive effect on the volume of tobacco export. The United States real total GDP
has a significant and positive effect on the volume of tobacco export. Exchange
rate has a significant and negative effect on the volume of tobacco export.
Tobacco export at previous year has a significant and positive effect on the
volume of tobacco export. From he short- run and long- run PAM regression
model the researcher conclude that Indonesia’s tobacco is less important to the
USA in the future because in the short- run Indonesia’s tobacco is elastic and
becoming more inelastic in the future. Beside that, the decision of buying
Indonesia’s tobacco is more induced by income decision. In the short- run the
commodity is necessity, however in the long- run it becomes non- necessity. From
both result the researcher can conclude that Indonesia cannot rely on their income
from tobacco. So Indonesia should make product diversification or country
diversification.
xii
CHAPTER I
INTRODUCTION
2003, even though it less than before the crisis (i.e. 7%). It is a good
condition, because after the crisis hit this country our per capita income is
very low. With the increasing of economic growth the investment to the
After the economic crisis hit Indonesia and many other Asian
country is able to purchase goods abroad more cheaply than it can produce
1
2
3. The exchanging and expanding the goods and services can be pursue the
country. Hence, trade between two countries can benefit both countries.
gas), it is known in 1980’s era Indonesia make lot of income from this
sector. At that time agriculture was bigger than the oil- gas sector. The
growth of agricultural sector (non oil and gas) is seen from the growth of
Table 1.1
good prospective market for The United States tobacco. Tobacco export
from Indonesia has a good market in United States market, because they
like the kind of tobacco from Indonesia. Indonesia has several kind of
During 1997 Indonesia also exported cigarettes worth US$ 127 million.
of tobacco and cigarettes in 1997 was US$ 588 million, with a positive
which has helped Indonesia’s exports. The simultaneous import and export
of tobacco is explained by the need for blending various tobaccos for the
is in Central and East Java (in a 200 Km radius). Important cities in this
region are Surabaya, Bojonegoro, Kudus, Kediri, and Malang, those cities
centers for tobacco manufacturing and trading. They have been marked red
5
Island, and the city of Medan is important center for tobacco trading.
Size of Firm
As for the "Size of the Firm" is varying, the market for tobacco can
Large Enterprises
pieces. Its brands are Bentley and Marlboro. Its market share is
6.96%.
having 2.93 % market share. Its popular brands are Dunhill, Kansas,
subsidiary.
Viet Nam and Myanmar. The tobacco sourcing for those factories is
Small Enterprise
There are 135 small size cigarette manufacturers; they have a total
26.8% market share. Their main raw material is local tobacco. However,
they also use some imported tobacco, primarily for blending. There are
they are:
and pesticides, the production will be lower. There is too much rain, the
crop won’t be good, and production is less than average. On the other
hand, the demand for cigarettes is increasing at the rate of 4.50% per
annum and hence the demand for imported tobacco will increase.
for tobacco. The need for blending of various types of tobacco for good
In the short (one year), medium (two to three years), and long term
(four to seven years) Indonesia will be important market for tobacco. The
Indonesian economy is now on the way to recovery. Interest rates has come
down from 40% to 20%, currency has stabilized within the IMF prescribed
limits, and inflation has come down from above 60% to less than 15%. The
the business confidence and there are strong indications that the economy
8
wrote:
Netherlands, Germany, France and Spain, but the most tobacco are
DESTINATION
LUXEMBOURG
lots of tobacco. To fulfill its needs of tobacco, USA imports tobacco from
tobacco exporting countries every year among them are from Indonesia.
First that USA with its technology can heap their tobacco reserve, second
USA like importing their tobacco reserve from Indonesia because the
10
USA import today is affected by previous years import, this research uses
PAM.
I. 2 PROBLEM IDENTIFICATION
This research will focus on what factors that affect the quantity
demand of Indonesian tobacco export to the United States. The factors that
will be discussed are price of good, income of the country and exchange
volume export of the previous year. The writer wants to find out when the
example if there were an increase in price of good, will this change affect
the question of what kind of factors that will affect on quantity demand of
identify the main problem of this research as the basis framework to write
I. 3 PROBLEM FORMULATION
tobacco export and its impact to the import condition, the writer
To know how far is the total real GDP of The United States will
To know how far is the exchange rate Rupiah to US Dollar affect the
tobacco export.
12
I. 5 RESEARCH OBJECTIVES
2. To analyze the effect of the United States total real GDP toward export
United States.
13
For the writer, this research is the opportunity to apply the knowledge and
tobacco export to The United States. The writer wants to search on what
price, the United States income (real total GDP), exchange rate Rp/US$,
Tobacco Control Paper no. 25. They also included time series data from
1970- 2001 and monthly data from January 1996- June 2001, to estimate
Indonesia. The study talks lot about tobacco, cigarettes, kreteks, demand,
Most of young people are very addicted with cigarette and lot of people
being passive smokers. This is Indonesian habit and they are hard to quit.
14
15
smoking. However the size of warning label is to small and the warning is
not clear, not specific and not strong. An empirical study shows that the
still allows the cigarette advertising, but the commercials should not show
that the audience cannot read and absorb it (Media Indonesia, No. 328)
excise revenue and predicts that an increase in the tax level of 10, 50 or
16
100 percent would increase total tax revenue by 9%, 43%, 82%
The study also showed the increased of cigarette taxes would result
predicts that if tax rate were increased government revenue from excise
cigarette demand.
instrument that the government can use to influence the price of tobacco
products, and hence their consumption. This study found that even very
large increases in the excise tax rate would increase total excise tax
Increasing the cigarette excise tax would increase the retail price of
consumption).
government adopt stronger tobacco control measures. They point out that
17
this happens, the increase in excise tax revenue and the reduction in
economy. The tobacco products industry has sizeable labor force, and
the transfer of the labor force and help vulnerable groups (such as farmers
Catherine Reynolds
December 1998 edition), with the title: “Worshipping Cancer Sticks: The
Ordinary Lives”.
revenue after oil, gas and timber, reliable internal revenue, unlikely to
legislation.
before. A 1985 Jakarta study found that 49% of boys and 9% of girls’ aged
10 - 14 were daily smokers. The dangers of smoking are not well known in
Indonesia. The earlier people start to smoke, the more likely they are to
maintain the habit throughout life. Unless people start smoking by the time
they are twenty, they usually never do. As more than 50% of the
population is under 24 years old, the potential market (to use the current
this 'market' because they must recruit new smokers in order to maintain
The teenage years are the time when people are more focused on
advertising. As people get older and become more secure in their identity,
the advertised 'attributes' of cigarettes are no longer such a lure. Habit and
Indonesia, but the four major producers, Gudang Garam, Djarum Kudus,
worth US$4.9 billion in November 1997, after the drop in the rupiah;
2. Budi Hartono and family (Djarum) were worth US$1 billion with 20%
is not litigations in nature, and therefore the industry here does not expect
counterparts'.
step in the right direction, but it is not enough. This move by the
huge mortality rate from tobacco. Certainly industry analysts predict that
the large tobacco companies will only benefit from this increase, with
and it is doubtful that the political and legislative actions being taken
medical and educational needs combined, and twice as much as they spent
on festivals. The poorest households spent more on tobacco than they did
smoked 535 billion cigarettes in the past year. More than 100,000
Cigarette companies had been spending over $200 million per year
on radio and television advertising, and since the ban, almost all of this
people from about 750,000 families receive $1.4 billion annually for the
collection by the Federal government behind the major giants, e.g., income
manufacturers excise taxes, alcohol taxes, and estate and gift taxes. In
were to decline.
Ross Hammond
citizen, until the American cigarette producer should import the tobacco
from foreign countries. It’s make tobacco expand globally. The title is:
tobacco. Between 1990 and 1993, tobacco leaf imports to the U.S. more
than doubled to over 1 billion pounds, and between 1995 and 1996 rose
over 21 percent. According to the USDA, "The main reason for this surge
was the rising popularity in the United States and abroad for low and mid-
importer tobacco, one of them is Philip Morris who own the HM.
1. Philip Morris
and hawks the world's most popular brand, Marlboro, which accounts
terms of profit and sales. Since 1990, the company's cigarette sales
have risen by only 4.7 percent in the United States, but 80 percent
percent since 1993. In 1997, the company sold 235 billion cigarettes in
24
the United States for a profit of $3.3 billion, while selling over 711
billion cigarettes abroad for a profit of $4.6 billion, marking the first
time that the company's international sales made more profit than
domestic ones. By the year 2000, predicts CEO Geoffrey Bible, the
Slims brands. The acquisition will also help the company's efforts to
Cigatam has over the past few years been exploring strategies to
Marlboro, L&M and Bond Street cigarettes. As Chairman Bible puts it,
"We are still in the foothills when it comes to exploring the full
Tobacco Germany, Souza Cruz (Brazil) and Brown & Williamson all
Mexico, South Korea and the United States, which are handled at
in Tekel, the Turkish state cigarette monopoly. The purchase gives the
Turks consume nearly 100 billion cigarettes a year. BAT will invest
Mexico's biggest cigarette maker, for $1.7 billion (see Mexico case
study). This was one of the largest foreign investments ever made in
Mexico and BAT's most expensive purchase ever. CLM produces both
increase its ability to boost exports to the United States and Asia. BAT
3. RJ Reynolds
company, has recently fallen further behind Philip Morris and BAT.
Saddled with a huge debt load (the result of a failed takeover bid in the
late 1980's made famous in the book Barbarians at the Gate), RJR saw
analysts that the company may enter into an international alliance with
tobacco divisions.
tobacco sales.
people. Cigarettes and promotional items are often given away at these
brand names:
people to smoke. What they are simply doing, they say, is competing
smoke more."
CHAPTER III
important to notice three things about this concept (Lipsey, 1996: 63)
that consumers wish to purchase that the price of the other product is
that people are willing to buy, given the price they must pay for the
and the quantity that will demand are related negatively, other
thing being equal. That is, the lower the price, the higher the
quantity demanded, and the higher the price, the lower the quantity
30
31
increasing.
3. Other Price
curve shift to the right. In this case the price of gasoline is the
otherwise.
4. Taste
right. In the case of demand for electricity we don’t talk about the
32
5. Population
that quantity demand is the amount of good that buyers are willing
and for the inferior goods increasing to the income will lend
Demand theory has two subjects, which are change in demand and
elasticity of demand.
other things being equal. That is, the lower the price, the higher
the quantity demanded, and the higher the price, the lower the
the prices of the product remain the same. For example, in the
3. Other Price
curve shift to the right. For example, when the price of oil is the
4. Taste
5. Population
Graph 3.1
Change in Demand
DI D DII
From the graph above it can be seen, for the example if the
decrease in the demand and a fall in the price of the good produces
the movement along the demand curve. If some other factors on the
buy, there is a shift in the demand curve to the right (from D to DII)
change, which reduces the quantity that people plan to buy that
goods, there is a shift in the demand curve to the left (from D to DI)
merely whether the quantity and price each rise or fall; it is also
important to know how much the change. This is what the concept
of elasticity does.
variable (Y).
% change in price
income.
% Change in income
or sold and also determine the relative price of the goods. Demand in
1994: 8).
These two methods lead to the same value for GDP for the
40
with the same total output1 (Case and Fair, 1999: 136).
1
Suppose the economy is made up of just one firm and the total firm's output this year sells for $1
million. Because the total amount spent on output this year is $1 million, this year's GDP is $1
million. Remember: The expenditure approach calculates GDP on the basis of total expenditures
for final goods and services in the economy. But every one of the million dollars of GDP is either
paid to someone or remains with the owners of the firms as profit. Using the income approach, we
add up the wages paid to employees of the firm, the interest paid to those who lent money to the
firm, and the rents paid to those who leased land, buildings, or equipment to the firm. What is left
over is profit, which is, of course, income to the owners of the firm. If we add up the incomes of
all the factors of production, including profits to the owners, we get a GDP of $1 million.
41
the dollar in the United States, the Pound sterling in Britain, the
332).
Foreign exchange rates, for the most part, are not fixed
over time. Instead, like any other price, they vary from week to
company in the economy and the lower the possibility for the
are willing and be able to buy at alternative prices in a given time period:
the sum of individual demands. The market demand curve expresses the
varies with the price of good. The downward slope of the demand curve
curve as indicating how the willingness of buyer to pay varies with the
demand curve indicates that the smaller the amount of goods and services
raises, other things constant. But other things tend to change, and these
2
Economics (1995) Paul A. Samuelson and William D. Nordhaus.15th edition.P.442
3
the Microeconomy Today. Bradley R. Schiller. 3rd edition. Random Hoouse. P.88
44
applied in monetary policy- steps by which the central bank cab affect the
supply of money and other financial conditions- and fiscal policy- taxes
commodities it can produce more cheaply than other countries, and then
research affect Indonesian tobacco export to the United States. The hypotheses
that according to the theory, when the price of tobacco increases the
cheaper.
that according to the theory when the United States GDP increases, it
This tobacco can be positive if superior good for the United States
In this thesis we use the United States real total GDP because the
price of tobacco its relatively expensive and tobacco not for individual
because those countries are hold the world economic and to avoid
c. Exchange rate
current tobacco export volume to USA. The influence can be: positive if
the needs of tobacco in the USA is not limited, negative: if the needs of
RESEARCH METHOD
which is the demand of Indonesian tobacco export by the United States. The
Where
X2 = total real GDP of the United States (Bill of US$) with 1990 as
47
48
tobacco exported from Indonesia. The research sought what variables that
This section analyze why the United States import tobacco from
Indonesia. Some reason why the United States import Indonesia’s tobacco
because it tobacco has good quality especially the flue cured Virginian
type, even though the quality is still less than the Pakistan. Tobacco is the
basic material for cigarette and the demand for cigar in the United States
places good position in trade. Most of Indonesian tobacco uses for big
Trading between Indonesia and the United States will create advantages in
both countries.
The data used in this research analysis were the data taken from
Based on the data used in this research variables in this thesis are
Dependent variable
Independent variable
they are:
conducted by using time series data from 1981 until 2001. Since this
independent variable.
lnYt-1
50
Where:
The types of data used in empirical analysis are time series data.
Empirical work based on time series data assumes that the underlying time
series is stationary.
4.7.1.1 Multicollinearity
conditions below:
collinear).
close +1 or -1).
4.7.1.2 Heteroscedasticity
ε²i = ά0+άo
1X1+ά0X2+ά0X²1+ά0X²2+ά0X1X2+Ui. (2)
heteroscedasticity.
4.7.1.3 Autocorrelation
will decline.
53
time series and cross sectional) data. Data used in this research is
and annually.
54
variables individually.
Ho: αi = 0
Ho: αi < 0
variable individually.
Ho: αi = 0
Ho: αi > 0
Ho: αi = 0
Ho: αi # 0
55
lag in the model), since in the short run the dependent variable may
between 0 - 1
CHAPTER V
Where:
X1 = price of tobacco
X3 = exchange rate
Regression model that is used is PAM of log linier (double log regression)
Where:
56
57
testing data collected from many resources to know the factors affecting
export to the United States (Q) measured in tons, Price of tobacco (P)
measured in dollar, the United States total real GDP (GDP) measured in
The factors used in this research are price of tobacco, the United
States total real GDP as the demand country, and exchange rate between
Indonesia and the United States. Analysis descriptions are based on the
This thesis uses the United States total real GDP because we are
private.
real GDP G5 minus USA (Japan/ Yen, UK/ Pound sterling, Germany/
Deutsche Mark, and France/ Franc) because those countries are hold the
Exchange rate
1. Fixed exchange rate, is the system that government set the fixed rate
of exchange
2. Flexible exchange rate, is the system that the rate determine by the
market
3. Managed exchange rate, is the system that the rate determined by the
supply.
59
The writer use exchange rate Rupiah to US Dollar because Rupiah kurs
will affect the US Dollar purchasing power that will be lead to increase or
The first step to analyze the data is by regress the data with the
representative with the research. The writer uses Eviews 3.0 computer
program in order to make the data estimation easier. Beside that Eviews
3.0 computer program helped the writer in avoiding the computation error.
The writer use PAM (Partial Adjusted Model) because this model
is gives better estimation result and has the biggest R² result. Besides that
the writer also run the MWD- test (McKinnon, White, Davidson, (1983))
to choose the best model for this research. The MWD test says that PAM
in log- linier model is the best way and as long as DW is greater than R²
The model used in this research is proper, shown by the size of R²=
5.4.2 F- Test
total real GDP, volume tobacco export at previous year) on the model have
T- table = tα df (n-k)
Where:
df = 16
Testing uses one tail or two tail (the confidence- interval approach)
expectation2.
1
See Gujarati, Damodar. Basic Econometrics 4th edition. P.128
2
Ibid. P.127
62
Graph 5.1
2%
1%
0,2% 0.43%%
3
See Gujarati, Damodar. Essentials of Econometrics.2nd Edition. 168
63
Graph 5.2
25%
25%
Rejected Ha
20% 20%
10%
10%
9,8% 9,8%
Accepted Ha Accepted Ha
25%
20%
Rejected Ha
9.3% 10%
Accepted Ha
Graph 5.3
Conclusion: ha is rejected, exchange rates have no effect on the
Graph 5.4
2% 2%
1%
1%
0.2% Rejected Ha
Accepted Ha 0.2%
0.18% 0.18%
The result of the PAM regression model can be seen in the following:
0.664566LnYt-1
R² = 0.944567
Adj R² = 0.929785
DW = 1.878440
However, the results of PAM regression above just for short run
estimation. For the long run we need to calculate the coefficient of long- run PAM
model:
1-δ = 1- 0.664566
= 0.335434
From the result above (in absolute value), the long- run coefficient
is bigger than the short- run; it means that the development of tobacco
• α1 = 2.726410
• α2 = 0.543760
• α3 = -0.248775
• α4 = 0.664566
• α1 = 8.128007298
• α2 = 1.621064054
1.621064054
• α3 = -0.741651114
• α4 = 1.981212399
Independent
Short- run Long- run
variable
Ln X1 2.726410 8.128007298
Ln X2 0.543760 1.621064054
both short- run and long- runs. These result shows that the
short- run and long- runs. But, in short- run the elasticity
year is in the positive sign for both short- run and long- run;
it means that the more export in the current year the more
5.4.5.1 Multicollinearity.
program the writer can search the value of each r and the result is
Table 5.3
From the table above it can conclude that the values of the
5.4.5.2 Heteroscedasticity
Table 5.4
Test Equation:
Dependent Variable: RESID^2
Method: Least Squares
Date: 01/12/06 Time: 08:28
Sample: 1982 2001
Included observations: 20
Variable Coefficien Std. Error t-Statistic Prob.
t
C 3.966290 1.892556 2.095732 0.0600
LNPRICE -3.381133 1.468246 -2.302837 0.0418
LNPRICE^2 1.430770 0.621453 2.302297 0.0419
LNGDP -1.287365 0.655899 -1.962749 0.0755
LNGDP^2 0.297219 0.154941 1.918266 0.0814
LNEXC 0.251681 0.392308 0.641540 0.5343
LNEXC^2 -0.016913 0.023774 -0.711380 0.4917
LNVOLUME(-1) -0.797962 0.314324 -2.538664 0.0275
LNVOLUME(-1)^2 0.099970 0.038216 2.615927 0.0240
R-squared 0.670712 Mean dependent var 0.027901
Adjusted R-squared 0.431231 S.D. dependent var 0.026536
S.E. of regression 0.020013 Akaike info criterion -4.682709
Sum squared resid 0.004406 Schwarz criterion -4.234629
Log likelihood 55.82709 F-statistic 2.800682
Durbin-Watson stat 3.094654 Prob(F-statistic) 0.058235
5.4.5.3 Autocorrelation
Table 5.5
Test Equation:
Dependent Variable: RESID
Method: Least Squares
Date: 01/12/06 Time: 08:34
Presample missing value lagged residuals set to zero.
Variable Coefficient Std. Error t-Statistic Prob.
C -0.369626 0.662890 -0.557598 0.5866
LNPRICE 0.048395 0.693455 0.069788 0.9454
LNGDP 0.319741 0.298327 1.071779 0.3033
LNEXC -0.095441 0.117774 -0.810377 0.4323
LNVOLUME(-1) 0.101971 0.113848 0.895672 0.3867
RESID(-1) -0.073171 0.261308 -0.280019 0.7839
RESID(-2) -0.729249 0.249677 -2.920766 0.0119
R-squared 0.396798 Mean dependent var 3.67E-16
Adjusted R-squared 0.118397 S.D. dependent var 0.171375
S.E. of regression 0.160910 Akaike info -0.546724
criterion
Sum squared resid 0.336597 Schwarz criterion -0.198217
Log likelihood 12.46724 F-statistic 1.425277
Durbin-Watson stat 1.693879 Prob(F-statistic) 0.277482
This test uses the level of degree (χ²), Ho expressing that there
contrary.
73
Table 5.6
Autocorrelation test with LM method
2. Hypothesis is proven. The United States real total GDP (X2) has
3. Hypothesis is not proven. The exchange rate (X2) has significant and
that might affect the effect of exchange rate on export. This data period
is from 1981 until 2001 and there are many events that might be the
but it does not effect the sign so exchange rate has no effect on
test on the appendices to prove that exchange rate data in this thesis
dependent variable.
CHAPTER VI
6.1 Conclusion
1. From the regression of time series data from 1981- 2001 with
dependent variable (Ln Y) and price of tobacco (Ln X1), the United
States real total GDP (Ln X2), Foreign Exchange (Ln X3) and
United States.
volume of tobacco export, both in short- run and long- run, price
75
76
5. The United States real total GDP (X2) has a significant and positive
6.2 Implication
From the short- run and long- run coefficient of PAM regression model we
can conclude:
that Indonesia’s tobacco is less important to the USA and has more
United States.
one.
6. Indonesia has to face the fact that Indonesia cannot rely on the
Where:
The United States real total GDP is weighted by average real total GDP
HETEROSCEDASTICITY
Test Equation:
Dependent Variable: RESID^2
Method: Least Squares
Date: 01/12/06 Time: 08:28
Sample: 1982 2001
Included observations: 20
Variable Coefficien Std. Error t-Statistic Prob.
t
C 3.966290 1.892556 2.095732 0.0600
LNPRICE -3.381133 1.468246 -2.302837 0.0418
LNPRICE^2 1.430770 0.621453 2.302297 0.0419
LNGDP -1.287365 0.655899 -1.962749 0.0755
LNGDP^2 0.297219 0.154941 1.918266 0.0814
LNEXC 0.251681 0.392308 0.641540 0.5343
LNEXC^2 -0.016913 0.023774 -0.711380 0.4917
LNVOLUME(-1) -0.797962 0.314324 -2.538664 0.0275
LNVOLUME(-1)^2 0.099970 0.038216 2.615927 0.0240
R-squared 0.670712 Mean dependent var 0.027901
Adjusted R-squared 0.431231 S.D. dependent var 0.026536
S.E. of regression 0.020013 Akaike info criterion -
4.682709
Sum squared resid 0.004406 Schwarz criterion -
4.234629
Log likelihood 55.82709 F-statistic 2.800682
Durbin-Watson stat 3.094654 Prob(F-statistic) 0.058235
83
AUTOCORRELATION
Test Equation:
Dependent Variable: RESID
Method: Least Squares
Date: 01/12/06 Time: 08:34
Presample missing value lagged residuals set to zero.
Variable Coefficien Std. Error t-Statistic Prob.
t
C -0.369626 0.662890 -0.557598 0.5866
LNPRICE 0.048395 0.693455 0.069788 0.9454
LNGDP 0.319741 0.298327 1.071779 0.3033
LNEXC -0.095441 0.117774 -0.810377 0.4323
LNVOLUME(-1) 0.101971 0.113848 0.895672 0.3867
RESID(-1) -0.073171 0.261308 -0.280019 0.7839
RESID(-2) -0.729249 0.249677 -2.920766 0.0119
R-squared 0.396798 Mean dependent var 3.67E-16
Adjusted R-squared 0.118397 S.D. dependent var 0.171375
S.E. of regression 0.160910 Akaike info criterion -
0.546724
Sum squared resid 0.336597 Schwarz criterion -
0.198217
Log likelihood 12.46724 F-statistic 1.425277
Durbin-Watson stat 1.693879 Prob(F-statistic) 0.277482
84
MULTICOLLINEARITY
1981- 2001
year
VOLUME EXPROTS
Year Y LnY
1981 23.56000 3.159550
1982 24.10000 3.182212
1983 24.20000 3.186353
1984 20.50000 3.020425
1985 20.90000 3.039749
1986 24.10000 3.182212
1987 22.90000 3.131137
1988 29.10000 3.370738
1989 49.90000 3.910021
1990 68.80000 4.231204
1991 75.60000 4.325456
1992 109.9000 4.699571
1993 125.2800 4.830551
1994 97.30000 4.577799
1995 77.00000 4.343805
1996 119.4000 4.782479
1997 136.6000 4.917057
1998 84.50000 4.436752
1999 110.3000 4.703204
2000 109.8000 4.698661
2001 126.7000 4.841822
Source: Statistical Year Book, BPS
86
TOBACCO PRICE
Year X1 LnX1
1981 2.372000 0.863733
1982 2.515000 0.922273
1983 2.771000 1.019208
1984 2.736000 1.006497
1985 2.990000 1.095273
1986 2.994000 1.096610
1987 3.257000 1.180807
1988 3.514000 1.256755
1989 3.726000 1.315335
1990 3.910000 1.363537
1991 3.904000 1.362002
1992 3.918000 1.365581
1993 3.867000 1.352479
1994 3.876000 1.354804
1995 4.012000 1.389290
1996 4.149000 1.422867
1997 3.973000 1.379521
1998 4.030000 1.393766
1999 4.030000 1.393766
2000 4.211000 1.437700
2001 4.233000 1.442911
Source: IFS
87
EXCHANGE RATE
Year X3 LnX3
1981 644.0000 6.467699
1982 692.5000 6.540308
1983 994.0000 6.901737
1984 1074.000 6.979145
1985 1125.000 7.025538
1986 1641.000 7.403061
1987 1650.000 7.408531
1988 1731.000 7.456455
1989 1797.000 7.493874
1990 1901.000 7.550135
1991 1992.000 7.596894
1992 2062.000 7.631432
1993 2110.000 7.654443
1994 2200.000 7.696213
1995 2308.000 7.744137
1996 2383.000 7.776115
1997 4650.000 8.444622
1998 8025.000 8.990317
1999 7085.000 8.865735
2000 9595.000 9.168997
2001 10400.00 9.249561
Source: IFS
91
LINEAR
Year Y X1 X2 X3
1981 23.56000 2.372000 7.285650 644.0000
1982 24.10000 2.515000 7.860180 692.5000
1983 24.20000 2.771000 9.212230 994.0000
1984 20.50000 2.736000 11.32877 1074.000
1985 20.90000 2.990000 8.957510 1125.000
1986 24.10000 2.994000 7.585120 1641.000
1987 22.90000 3.257000 6.212350 1650.000
1988 29.10000 3.514000 6.859390 1731.000
1989 49.90000 3.726000 6.850570 1797.000
1990 68.80000 3.910000 5.860370 1901.000
1991 75.60000 3.904000 5.791550 1992.000
1992 109.9000 3.918000 6.446600 2062.000
1993 125.2800 3.867000 7.006880 2110.000
1994 97.30000 3.876000 6.483230 2200.000
1995 77.00000 4.012000 5.735880 2308.000
1996 119.4000 4.149000 6.099210 2383.000
1997 136.6000 3.973000 6.987490 4650.000
1998 84.50000 4.030000 6.740120 8025.000
1999 110.3000 4.030000 11.87073 7085.000
2000 109.8000 4.211000 8.526080 9595.000
2001 126.7000 4.233000 8.854390 10400.00
Source: IFS
92
LOG LINEAR
MWD TEST
PAM-Linier
Dependent Variable: VOLUME
Method: Least Squares
Date: 01/17/06 Time: 11:36
Sample(adjusted): 1982 2001
Included observations: 20 after adjusting endpoints
Variable Coefficien Std. Error t-Statistic Prob.
t
C -176.5663 0.014190 -12443.10 0.0000
PRICE 49.23844 0.003389 14527.06 0.0000
GDP 5.619510 0.000755 7442.714 0.0000
EXC -0.003210 5.62E-07 -5707.691 0.0000
VOLUME(-1) 0.570713 3.99E-05 14287.21 0.0000
Z1 1.000189 6.06E-05 16506.68 0.0000
R-squared 1.000000 Mean dependent var 72.84400
Adjusted R-squared 1.000000 S.D. dependent var 42.48216
S.E. of regression 0.004132 Akaike info criterion -7.896945
Sum squared resid 0.000239 Schwarz criterion -7.598225
Log likelihood 84.96945 F-statistic 4.02E+08
Durbin-Watson stat 1.824467 Prob(F-statistic) 0.000000
94
MWD MULTIPLE-REGRESSION
Linier Regression
Dependent Variable: VOLUME
Method: Least Squares
Date: 01/17/06 Time: 11:51
Sample(adjusted): 1982 2001
Included observations: 20 after adjusting endpoints
Variable Coefficien Std. Error t-Statistic Prob.
t
C -312.8729 80.25034 -3.898712 0.0014
PRICE 93.78873 17.64639 5.314896 0.0001
GDP 7.401650 3.976989 1.861119 0.0824
EXC -0.002862 0.002855 -1.002259 0.3321
Z01 -45.51280 21.53078 -2.113849 0.0517
R-squared 0.817193 Mean dependent var 72.84400
Adjusted R-squared 0.768444 S.D. dependent var 42.48216
S.E. of regression 20.44252 Akaike info criterion 9.085429
Sum squared resid 6268.452 Schwarz criterion 9.334363
Log likelihood -85.85429 F-statistic 16.76341
Durbin-Watson stat 1.326571 Prob(F-statistic) 0.000021
Log-Lin Regression
Dependent Variable: LNVOLUME
Method: Least Squares
Date: 01/17/06 Time: 11:52
Sample: 1981 2001
Included observations: 21
Variable Coefficien Std. Error t-Statistic Prob.
t
C -3.601563 1.116952 -3.224455 0.0053
LNPRICE 7.427042 1.320272 5.625389 0.0000
LNGDP 1.157118 0.495392 2.335763 0.0328
LNEXC -0.488840 0.224132 -2.181038 0.0444
Z02 0.061895 0.016056 3.854932 0.0014
R-squared 0.893213 Mean dependent var 4.027179
Adjusted R-squared 0.866516 S.D. dependent var 0.736784
S.E. of regression 0.269187 Akaike info criterion 0.417436
Sum squared resid 1.159387 Schwarz criterion 0.666132
Log likelihood 0.616918 F-statistic 33.45776
Durbin-Watson stat 1.330902 Prob(F-statistic) 0.000000
96
TEST ON LNVOLUME
TEST ON LNPRICE
TEST ON LNGDP
TEST ON LNEXC