2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
The Impact of Tariff Reductions on Real Imports in
Malaysia from 1980-2010
Written by Juita Mohamad
Graduate School of Asia Pacific Studies, Waseda University,
Tokyo, JAPAN
Phone number: 00819091054007
Email: mjuita@suou.waseda.jp
June 27-28, 2012
Cambridge, UK
1
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
The Impact of Tariff Reductions on Real Imports in Malaysia from 1980-2010
Abstract
This paper investigates the long run relationship of drastic tariff reductions on the real
imports from 1980 - 2010 in Malaysia using the Johansen cointegration analysis. Time series
data for eight sectors according to the Standard International Trade Classification were
compiled. A dynamic error correction model is used to overcome the limited number of
observations for each of the sectors. A log-linear regression is run for each sector to test the
effects of income, domestic price, import price, tariff rates on real imports in each of these
sectors. With the independent variables selected, it is expected that as imports increase,
income and domestic price increase, while import price and tariff rates decrease. The results
from the regression exercise are mixed. It is observed that those sectors conforming to the
hypothesis are sectors concerning with basic necessities of the Malaysian consumers and
producers. The import demand for sectors with basic necessity goods are more sensitive with
the changes in tariff rates compared to sectors with non-basic necessity goods. For the latter
group, even when tariff rates are increasing, import demand still increases as these products
are mostly intermediary goods needed for production and processing activities. This
interpretation is appropriate for the case of Malaysia, whereby manufacturing activity is the
main driver of its domestic economy since the early 1990s. This study is beneficial to
determine the behaviour pattern of each sector in the wake of reduced import prices and
drastic tariff cuts.
Key words: tariff reductions, real imports, Stolper-Samuelson theorem, domestic price,
import price, Malaysia
June 27-28, 2012
Cambridge, UK
2
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Introduction
As the world economy shifts into a more globalized era, presently every developing nation is
harnessing their resources in aneffort to take part in a more free trade regime.Globalization
promotes the practice of free trade and it is believed to promote a more levelled playing field
in the world market, as tariff rates are driven down to near 0%. In other words, protectionism
policy is at the brink of extinction with the rise of globalization. Globalisation in the long run,
rewards efficient producers the competitive advantage that ensures its position in the global
market.Even with the promise of more wealth and ensuring increased welfare for all, the
issue of increased competition domestically in the midst oftrade liberalization for both
developed and developing nations, are widely discussed. In his book Making Globalization
Work(2006), Stiglitz emphasized that with globalization, "Everyone was supposed to be a
winner - those in both developed and thedeveloping world. Globalization was to bring
unprecedented prosperity to all."
Even Adam Smith, 1776 promoted the idea of free tradei. Here was how he put it at that time:
It is the maxim of every prudent master of a family, never to attempt to make at home what it
will cost him more to make than to buy.. . . If a foreign country can supply us with a
commodity cheaper than we ourselves can make it, better buy it of them with some part of the
produce of our own industry, employed in a way in which we have some advantage.
With all of the advantages of free trade, some economists are sceptical about the downside of
globalisation which includes increased competition in the domestic economy which can lead
to the absolution or demise of local businesses. With the world as one big market, local
producers are forced to compete with cheaper imported goods or higher quality products. The
survival of the fittest is being tested in today’s economic environment, with more pressure
being put not only onto business, but nations as well.
June 27-28, 2012
Cambridge, UK
3
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Malaysia, a small open economy, in the South East Asian region, is taking this threat and
turning it into an opportunity. The ASEAN bloc has become the hub for intratrade activity.
Intra- regional trade has been growing. According to Otsuki, 2011, intra-regional trade for
manufactured goods within the ASEAN region has increased to 150 billion USD in 2008 if
compared to 70 billion USD in year 2000ii. What makes this region a part of the world engine
of economic growth nowadays, is that as we increasingly trade intermediary goods among
ourselves in the assembly process of commodities, we also import and re-export the goods,
outside the region to China, the US and Europe as end users for our products. This activity is
how we are known as the entreportcenter.
Due to the importance of this intra trade activity as the driver of the nation’s economy, it is
important to examine the effects of the reduction of tariff rates on real imports. It is also
important to see how changes in income, domestic pricing, import pricing effect real imports.
The study will take a look at the behaviour of these variables in different sectors according to
the segregation of Standard Industrial Trade Classification, Revision 3 at one digit level iii.The
paper examines how reduction in import duties affect the demand for real imports in selected
sectors in a small, open, developing economy of Malaysia using time series data from 1980 to
2010.
Background on Tariff Reduction through AFTA and WTO Initiatives
On the 8th of August 1967, a simple yet clear agreement was drawn up between the six
Founding Fathers of the six main Association of Southeast Asian Nations (ASEAN)
members, which were Indonesia, Malaysia, the Philippines, Singapore and Thailand.
According to the document, ASEAN is seen as representing ―the collective will of the
June 27-28, 2012
Cambridge, UK
4
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
nations of Southeast Asia to bind themselves together in friendship and cooperation and,
through joint efforts and sacrifices, secure for their peoples and for posterity the blessings of
peace, freedom and prosperity. It is with that basic drive of cooperation and joint efforts,
ASEAN Free Trade Area (AFTA) came about in year 1992. When the AFTA agreement was
signed, the primary goals were to ―increase ASEAN's competitive edge as a production base
in the world market through the elimination, within ASEAN, and to attract more foreign
direct investment to ASEAN. Looking back since its inception, AFTA has come a long way
in changing trade patterns in Malaysia and the South East Asian region. As from year 2002, it
is now in full swing, ― aiming to promote the region‘s competitive advantage as a single
production unit. Even though specific rules and priorities are given for the elimination of
tariff mainly for manufacturing and agriculture products in member countries, the non-tariff
barrier is also expected to promote greater economic efficiency, productivity, and
competitiveness. To observe how far the region has come in promoting a leveled playing
field for its member countries, as of 1 January 2005, ― tariffs on almost 99 percent of the
products in the Inclusion List of the ASEAN-6 have been reduced to no more than 5 percent.
More than 60 percent of these products have zero tariffs.
The average tariff for ASEAN-6 has been brought down from more than 12 percent when
AFTA started to 2 percent today. The implementation of the Common Effective Preferential
Tariff –AFTA Scheme (CEPT-AFTA) was even accelerated further in January 2004, when
Malaysia decided to reduce tariffs for Completely Built Up (CBUs) and Completely Knocked
Down (CKDs) automotive units to smoothly meet its CEPT commitment one year earlier
than schedule. This proved to be a big commitment on behalf of Malaysia, as Proton, its
national car industry have been shielded under heavy protectionist policy since year 1985.
When full trade liberalization is achieved, the area could not only be beneficial for the 10
member countries but also holds attractiveness for trading partners all around the worldiv.
June 27-28, 2012
Cambridge, UK
5
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Malaysia is a founding Member of the WTO by virtue of its membership in the GATT since
1957. Through active participations in WTO negotiations, Malaysia continues to ensure that
trade regulations and trade measures that are negotiated are fair and provide the flexibility for
Malaysia to continue its development policyv. To date, under the commitment of
GATT/WTO, the average MFN applied tariff rates for Malaysian goods are down to 8.8% in
year 2010 compared to 11.3% in year 1995vi. This shows the continuous effort and
commitment of Malaysia as a WTO member and how tariff rate reductions are not only
induced by the ASEAN bloc agreement regionally but also the commitments of the
GATT/WTO at an international level.
Literature Review
There are many empirical studies on demand of imports. For example in the case of
Malaysia, the studies which are conducted to examine the behaviour of import demand are
Mohammad (1980), Semudram (1982), Awang (1988), and the Malaysian Institute of
Economic Research (MIER) Annual Model (1990). These studies estimated a traditional
import demand function, where the dependant variable is the volume of imports, where the
independent variables are real income and relative prices. For these studies, the assumption is
that data are stationary. These studies were done when cointegration analysis and error
correction model were not standard practice for time series analysis. Hence they used OLS
regression models or partial adjustment approaches to estimate the import demand function.
Granger and Newbold (1974) emphasized that if the stationary assumption is not satisfied,
this can lead to spurious regression. Due to this, the OLS results would be unreliable.
From previous literature review, the traditional import demand function has branched out, to
include other independent variables which can affect real import such as tariff rates. Caesar
June 27-28, 2012
Cambridge, UK
6
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
(2011) have used tariff and exchange rate as an extension of the general import demand
function to study the determinants of import demand in Zambia. Egwaikhide (1999) have
incorporated tariff rates into its import prices, in the empirical study for Nigeria. Karacaovali
(2011) has also incorporated tariff rates in its model.
The choice between a linear and log-linear import demand equation is important because the
influence of explanatory variables on demand is affected by the functional form. Kmenta
(1971) stated that misspecification of the functional form can result in misspecification of the
error term and violation of the classical assumptions of the error term. This leads the
estimates to be biased.
Previous studies by Khan and Ross (1977), Boylanet. al (1980) and Doroodian et. al (1994)
has proven that specification of a log-linear form is preferable when estimating import
demand functions.
Data
The variables used for this study areReal Imports, Gross National Income, Domestic Price
Index, Import Price Index and Simple Average Tariff Rates. All of the annual data from year
1980-2010 were supplied by the Department of Statistics Malaysia except for the tariff rates,
which were obtained from the WITS website. The real imports variable is based on import
value and not quantity as it is more comparable among the goods in a certain sector. The
sectors are divided into 9 Standard of Industrial Trade Classification at 1- digit level. The
author constructed an index for the real imports for each sector.For a more accurate analysis,
the author chose the log form, so that the result of the coefficients of the variables can be
interpreted as degree of elasticity. Due to lack of data, the author has focused her analysis on
June 27-28, 2012
Cambridge, UK
7
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
only 8 sectors instead of 9 sectors according to the SITC Revision 3 segregation. These
sectors are: 0. Food and Live Animals, 1. Beverages and Tobacco, 2. Crude Materials
Inedible Except Fuels, 3. Mineral fuels, Lubricants and Related Materials, 4. Animal and
Vegetable Oils, Fats and Waxes, 5. Chemicals and Related Products, 6. Manufactured Goods
Classified Chiefly by Material, Machinery and Transport Equipment, 8. Miscellaneous
Manufactured Articles. The author did not include the last sector, 9. Commodities and
Transactions Not Classified Elsewhere in the SITC in this study due to data unavailability.
Methodology and Findings
This methodology is replicated from the study by Tang T.C. and Mohammad H. A. (2000),
which looked at the demand of aggregate import as a whole for Malaysia. Instead of looking
into aggregate imports, the author uses the same analytical steps in observing the different
behavior of demand of imports for 8 different sectors.
The traditional formula for an import demand function can be specified that relates the
quantity of imports demanded to income, the price of imports and the price of the domestic
substitute. Import demand at time t is written as below:
Mit = ƒ (Ydt, Pmit, Pdit)
(1)
where Mit is the quantity of imports demanded for commodity class i (to the ithat one digit
level of the Standard International Trade Classification) at period t, Ydt is domestic income,
Pmit is the price level for the import commodity class i and , Pdit is the price level for the
domestic good i at time t.
The theory of demand suggests that ordinary demand functions are homogenous of degree
zero in prices and income. This implies that the absence of money illusion and allows the
June 27-28, 2012
Cambridge, UK
8
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
demand for imports to be expressed as a function of real income and relative prices
(Siddique, 1997). Therefore the traditional import demand function can be rewritten as:
Mit = g (Yt, Rit)
(2)
Where Ytequals toYdt/Pdtand represents real domestic income and Ritequals to Pmit/Pdit, which
is the price of the ith digit import relative to the domestic price. This equation has the
assumption that the effect of 2 price variables on demand will be equal but in opposite
directions. Gafar, 1988 explained that the two most common functional forms used in the
literature are in the linear and log-linear formulations. In linear terms, the empirical import
demand function can be written as:
Mit = δ + α0Yt+β0Rit +εit
(3)
Where δ is the constant term, α0 is the marginal propensity to import,β0 is the import
coefficient of relative prices andεit, is the random error term. From economic theory point of
view, it is expected that α0>0. However Goldstein and Khan, 1976 explained that if imports
represent the difference between domestic consumption and domestic production of imported
goods, production may rise faster (slower) than consumption in response to rise in real
income. Due to this, imports could fall (rise) as real income increases, resulting in negative
(positive) sign for the coefficient α0.
The author however has constructed an adjusted import demand function, to include the tariff
rate variable, T as an extension from the traditional function. Additionally in this new
equation, the import price level and domestic price level are disaggregated to relieve the
restrictions imposed on the traditional demand function.The author has chosen to
disaggregate the log import price and log domestic price, due to the assumption that these
prices do not move in the same degree, and also in opposite directions. For this kind of study
June 27-28, 2012
Cambridge, UK
9
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
both the linear and log linear specifications have been used as highlighted in the literature
review section. The author has chosen the latter due to the fact that it is more commonly
used. At the end of the day, the selection between these 2 forms of specifications is intuitional
based on previous literatures.
With thedisaggregation of the two prices, the behaviour of the import prices and domestic
prices can be observed, for specific commodity groupings. As log-linear form is mostly
adopted for this empirical study, the newly adjusted import demand function can be written
as:
lnMit = δ + φ0lnYt-β0lnIit +ѲlnDit- σlnTit +εit
(4)
whereln is the natural logarithm, Iitis the import price level of the ith commodity group, Dit is
the domestic price level of the ith commodity group andTit, is the tariff rate for the ith
commodity group according to the SITC segregation. Εit is the random error term.
For this study, Error Correction Model (ECM) is being used due to limited annual
observations for each sectors from year 1980-2010 (only 31 observations). ECM is the most
appropriate for limited observations in time series. OLS is not appropriate for this time series
study, as the outcome will be highly unreliable as mentioned in the literature review section.
Before the ECM analysis could take place, the author tested the time series data for each
sector for multicollinearity problems. For each sector, Augmented Dickey Fuller (ADF)
testing was undergone. The regression equation for ADF test (Dickey&Fuller, 1979) is stated
as follows:
∆𝑌𝑡 = 𝑎 + 𝑏𝑡 + 𝑐𝑌𝑡 − 1 + ∑𝑘𝑖=1 𝑑∆𝑌𝑡 − 1 + 𝑒𝑡
(5)
where∆ is the first difference operator, t refers to time trend, and k is additional terms in the
first differences for the Augmented Dickey-Fuller test, et is the regression error assumed to
June 27-28, 2012
Cambridge, UK
10
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
be stationary with zero mean and constant variance. The test were carried out to test the null
hypothesis of a unit root (c=0). The results are presented in Table 1 below. The table
highlighted that all variables real imports, GNI, Domestic Price, Import Price and Simple
Average Tariff are integrated in order one, I(1). This means that they are stationary in their
first difference.
Insert Table 1a, 1b and 1c here
After the ADF test were undergone, the author went on to the vector error correction model.
Before the modelling could be done, for the Johansen method, it is crucial to specify an
appropriate lag length for the VAR. For all of the eight subsectors, 2-year lag length was the
most appropriate. As for the trace test, for each different sectors the cointegration ranking
differs. Table 2 below, shows the Johansen test for cointegration and the appropriate lags
chosen for each of the eight sub sectors. Table 3 presents the normalised cointegrating
equation estimate for all of the subsectors.
Insert Table 2a-h, here
Insert Table 3a-h, here
In the next step, error correction model was estimated. The rank is set to 1, which is the
default number of the error correction terms. An error correction model was estimated to
examine the long run behaviour of Malaysian imports (according to their 8sectors). The
lagged residual from the Johansen Cointegration Equation was included the dynamic general
ECM. The general equation for ECM with 2 lag length is stated as below:
∆ln Mt = b0 + b1i∆ lnM t-2 + ∑𝑛𝑖=0 𝑏2𝑖 ∆ ln Yt-2+ ∑𝑛𝑖=0 𝑏3𝑖 ∆ lnIt-2 + ∑𝑛𝑖=0 𝑏4𝑖 ∆ lnDt-2
𝑛
+∑𝑖=0 𝑏5𝑖
∆ lnTt-2 + b6 ECt-2 + error term
June 27-28, 2012
Cambridge, UK
(6)
11
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
where EC is residual error derived from the cointegrating vector. This dynamic general
equation is used separately for all the 8 sectors and presented in the next section.
Analysis of Findings
The coefficients of income, domestic prices, import prices and simple average tariff rates
shows both expected and unexpected signs and are both significant and not significant,
depending on the sector being analysed. In this section a detailed breakdown of the findings
will be presented.
Let us look at the normalised cointegrating coefficients results for each of the 8 sectors in
Table 3 below.
Insert Table 3 here – result for vec for all sectors- make table
Let us take a look at the results of the coefficients, sector by sector.
A very interesting finding from this analysis is that the coefficient for GNI as a proxy of
income, α0 is negative in relations to import demand. From economic theory point of view, it
is expected that α0>0. However Goldstein and Khan, 1976 explained that if imports
represent the difference between domestic consumption and domestic production of imported
goods, production may rise faster (slower) than consumption in response to rise in real
income. Due to this, imports could fall (rise) as real income increases, resulting in negative
(positive) sign for the coefficient α0.
For Sector 0, Food and Live Animals, all the signs exhibited the expected signs. In this case,
even when income decreases by 1 point, real imports will still increase by 0.61 point. In this
June 27-28, 2012
Cambridge, UK
12
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
sector, tariff rates do affect real imports. When average tariff decreases by1 point, real
imports will increase by 0.17 point.
The second sector which exhibits the expected signs for its variables is Sector 6.
Manufactured Goods Classified Chiefly by Material. For this sector, in the long run, domestic
and import prices play important roles in affecting real imports. The long run elasticities of
import demand with respect to import price and domestic price are -27 and 30. It is not
surprising then that the tariff coefficient sign is positive. For this sector in the long run, even
when tariff increases by 1 point, real imports will still increase by 1.18 points. It is clear that
for manufactured goods, tariff rates do not affect the import demands for this sector.
A 1 point increase in import price, will decrease real imports by 27 points, while a 1 point
decrease in domestic price, will decrease imports by 31 point. With our re-exporting activity,
the more we export the goods, the more imports are needed as intermediary goods to support
this activity.This is not surprising as Malaysia’s imports and exports of manufactures from
overall merchandise trading account for 70% as of year 2010. This is further highlighted by
Graph 1 in the Appendix.
The last sector with expected signs for its variables is Sector 8, Miscellaneous Manufactured
Articles. In this sector, footwear, furniture and articles of apparel and clothing accessories
which are basic necessity goods are included. The implied long run elasticties of import
demand with respect to import price and average tariff are -1.97 and -1.60. Here tariff rate
does play a role in affecting real imports.
For all the other sectors, the variable signs are not as expected. However, due to the fact that
this paper focuses on the effect of trade reform on imports, it is important to see how changes
in tariff rates do or do not, influence the changes in import demands for different goods in
June 27-28, 2012
Cambridge, UK
13
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
different sectors. Table 4 below present the results for the signs and significance of the tariff
rate variable.
Insert Table 4 here – tariff rate
As can be observed in the Table above, only 3 sectors are presented to have negative signs for
their tariff variable in relations to real import. These sectors are Food and Live Animals,
Animal and Vegetable Oils, Fat and Waxes and Miscellaneous Manufactured Articles. Their
coefficients are -0.17, -3.38 and -1.60 respectively with all being highly significant at 1%
level. What all of them have in common is that, these are basic necessity goods. The rest of
the sectors have positive signs for their tariff variables in relations to real import. These
sectors are 1. Beverages and Tobacco, 2.Crude Materials Inedible Except Fuels, 3.Mineral
fuels, Lubricants and Related Materials, 5.Chemicals and Related Products, 6. Manufactured
Goods Classified Chiefly by Material and 7. Machinery and Transport Equipment. Their
coefficients are 0.11, 0.67, 1.30, 0.49, 1.18 and 0.11 respectively, with all being highly
significant at 1% and 5% level except for Sector 7. For this sector, tariff rate is insignificant
in affecting import demand. These are all non- basic necessity goods except for beverages
and tobacco sector. Beverages and tobacco sector belongs to a special group. Higher taxes
have always been imposed on tobacco and alcoholic beverages in Malaysia, making it more
expensive and therefore, unavailable for youths and children. Due to this even when tariff
rates are higher, due to preferences or lack of choice, these goods are still high in demand for
Malaysian consumers.
Examining the results of the analysis, it can be concluded that the import demand for sectors
with basic necessity goods are more sensitive by the changes in tariff rates compared to
sectors with non-basic necessity goods. For the latter group, even when tariff rates are
increasing, import demand still increases as these products are mostly intermediary goods
June 27-28, 2012
Cambridge, UK
14
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
needed for production and processing activities. This interpretation is appropriate for the case
of Malaysia, whereby manufacturing activity is the main driver of its domestic economy.
Conclusion
A number of conclusions can be drawn from this study. Firstly, if researchers are to obtain
robust results it is important they choose the right methodology, appropriate for its time series
data limitation. As this data set, has only 31 observations for each of its 8 sectors, the author
has chosen the dynamic error correction model to estimate the long run behaviour of
Malaysian imports according to sectors from year 1980-2010.
Secondly, the negative sign presented for the income coefficient suggests that in Malaysia,
according toGoldstein and Khan, 1976 imports represent the difference between domestic
consumption and domestic production of imported goods, production may rise faster (slower)
than consumption in response to rise in real income. Due to this, imports could fall (rise) as
real income increases, resulting in negative (positive) sign for the coefficient α0.
Thirdly, the empirical results suggest that only 3 sectors which include basic necessity goods
for end users and producers, have all the expected signs for their variables. These sectors are
Food and Live Animals, Manufactured Goods Classified Chiefly by Material and
Miscellaneous Manufactured Articles.
Fourthly, as this paper focuses on the effect of trade reform on imports, it is important to see
how changes in tariff rates do or do not, influence the changes in import demands for
different goods in different sectors. The import demand for sectors with basic necessity goods
are more sensitive by the changes in tariff rates compared to sectors with non-basic necessity
goods. For the latter group, even when tariff rates are increasing, import demand still
June 27-28, 2012
Cambridge, UK
15
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
increases as these products are mostly intermediary goods needed for production and
processing activities. This interpretation is appropriate for the case of Malaysia, whereby
manufacturing activity is the main driver of its domestic economy since the early 1990s.
i
Taken from the Concise Encyclopedia of Economics, http://www.econlib.org/library/Enc/FreeTrade.html
Statistics were taken from the paper by Otsuki, 2011 at http://www.osipp.osakau.ac.jp/archives/DP/2011/DP2011E006.pdf
iii
For a more information please visit http://unstats.un.org/unsd/cr/registry/regcst.asp?cl=14
iv
Information obtained from the ASEAN Secretariat website
v
Taken from the MITI website,
http://www.miti.gov.my/cms/content.jsp?id=com.tms.cms.section.Section_f5694606-c0a81573-78d578d5759be8c9
vi
Tariff rates obtained by the World Trade Indicators Report 2010
athttp://info.worldbank.org/etools/wti/docs/Malaysia_taag.pdf
ii
June 27-28, 2012
Cambridge, UK
16
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
References
Awang, A. H. (1988) An evaluation of the structural adjustment policies in Malaysia. In
Proceedings of the Eighth Pacific Basin Central Bank Conference on Economic Modelling,
Bank Negara Malaysia, Kuala Lumpur, November 11-15
Boylan et al. (1980) The functional form of aggregate import demand equation: a comparison
of three European economies, Journal of International Economics 10, 561-566
Cheelo, C. (2011) Determinants of Import Demands Zambia, University of Zambia,
Published on the Internet by the SAP - Project at http://www.fiuc.org/iaup/sap/
Doroodian et al. (1994) An examination of the traditional aggregate import demand function
for Saudi Arabia, Applied Economics 26, 909-915
Dickey, D. A., & Fuller, W.A (1979) Distribution of the estimators for autoregressive time
series with a unit root, Journal of the American Statistical Association, 74, 427-431
Egwaikhide, F. (1999) Determinants of Imports in Nigeria: A Dynamic Specification, AERC
Research Paper 91, African Economic Research Consortium, Nairobi
Gafar, J.S. (1988) The determinants of import demand in Trinidad and Tobago: 1967-1984,
Applied Economics, 20, 303-13
Goldstein, M. and Khan, M.S (1976) Large versus small price changes and the demand for
imports, Journal of International Economics 7, 149-160
Granger, C. W. J. and Newbold, P. (1974) Spurious regressions in econometrics, Journal of
Econometrics, 2, 111-20
J. Kmenta, Elements of Econometrics, Macmillan, New York (1971)
Karacaovali, B. (2011) Trade policy and Trade Reform in a Developing Country
Khan, M.S. and Ross, K.Z, The functional form of the aggregate import demand equation,
Journal ofInternational Economy 7, 200-225.
Malaysian Institute of Economic Research (1990) MIER Annual Model, In Imaoka, H.,
Semudram, M., Meyanathan. S. and Kevin Chew (eds). Models of the Malaysian Economy: A
survey. Kuala Lumpur, MIER
Mohammad H. A. (1980), A demand equation for West Malaysian Imports, Journal Ekonomi
Malaysia, 1, 109-20.
Otsuki (2011) Quantifying the benefits of Trade Facilitation in ASEAN, OSIPP Discussion
Paper : DP-2011-E-006
June 27-28, 2012
Cambridge, UK
17
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Semudram, M. (1982), A macro-model of the Malaysian economy, 1957-1977, The
Developing Economies, June
Siddique, M.A.B (1997), Demand for Machinery and Manufactured Goods in Malaysia,
Mathematics and Computers in Simulation43, 481-486
Tang T.C. and Mohammad H. A. (2000) An aggregate import demand function for Malaysia:
a cointegrating and error correction analysis, Utara Management Review, 1, 43-57
June 27-28, 2012
Cambridge, UK
18
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Appendix
Table 1a
Sector/Variable
0
ADF
level
1
FD
ADF
level
2
FD
ADF level
3
FD
ADF level
*
-3.15
*
-2.48
**
-2.17
3.84
3.84
0.41
lngni
-0.81
**
-0.81
**
-0.81
**
-0.81
3.43
3.43
3.43
lndomprice
-2.14
**
-2.23
*
-1.56
***
-1.43
3.41
3.41
4.17
lnimpprice
-4.33
***
-2.44
-2.22
***
-1.91
4.65
4.65
4.49
lnsimpleavtariff
-2.26
***
-2.58
***
-2.50
**
-1.72
4.30
4.30
3.70
*,**,*** denoted rejection of a unit root hypothesis based on MacKinnon’s critical value at 1 percent, 5
percent and 10 percent.
Note:
1) Constant and trend were included in level, and only constant in first difference (refer to Baghestani&
Mott, 1997). In common practice, an augmentation of one or two, generally appears to be sufficient
to secure lack of autocorrelation of the error terms (Ghatak, Milner &Utkulu, 1997) One augmented
lag was used due to limitation of annual data (refer to Doroodian, Koshal& Al- Muhanna, 1994:912)
ln imports
-2.83
FD
2.75
3.43
3.25
2.67
3.20
*
**
***
***
**
Table 1b
Sector/Variable
4
ADF
level
ln imports
-4.174
5
FD
-7.13
***
-4.174
FD
-7.13
ADF
level
***
-1.456
7
FD
2.926
3.425
3.566
3.458
3.602
ADF
level
*
FD
-0.665
**
-0.805
**
-0.802
3.425
lndomprice
-2.428
***
-2.428
**
-2.587
**
-4.052
3.361
lnimpprice
-3.952
***
-3.952
***
-3.007
**
-1.887
4.258
lnsimpleavtariff
-2.247 -3.97 ***
-2.247
**
-1.672
**
-2.35
3.396
*,**,*** denoted rejection of a unit root hypothesis based on MacKinnon’s critical value at 1 percent, 5
percent and 10 percent.
Note:
1) Constant and trend were included in level, and only constant in first difference (refer to Baghestani&
Mott, 1997). In common practice, an augmentation of one or two, generally appears to be sufficient
to secure lack of autocorrelation of the error terms (Ghatak, Milner &Utkulu, 1997) One augmented
lag was used due to limitation of annual data (refer to Doroodian, Koshal& Al- Muhanna, 1994:912)
lngni
June 27-28, 2012
Cambridge, UK
-0.805
3.425
7.214
3.828
ADF
level
6
**
-0.805
19
2.459
**
3.596
***
5.649
***
4.447
**
3.375
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Table 1c
Sector/Variable
8
ADF
level
ln imports
-0.154
lngni
-0.802
lndomprice
-0.925
lnimpprice
-3.88
-2.336
lnsimpleavtariff
FD
*
2.823
**
3.596
**
3.545
**
3.649
***
3.968
*,**,*** denoted rejection of a unit root hypothesis based on MacKinnon’s critical value at 1 percent, 5
percent and 10 percent.
Note:
1) Constant and trend were included in level, and only constant in first difference (refer to Baghestani&
Mott, 1997). In common practice, an augmentation of one or two, generally appears to be sufficient
to secure lack of autocorrelation of the error terms (Ghatak, Milner &Utkulu, 1997) One augmented
lag was used due to limitation of annual data (refer to Doroodian, Koshal& Al- Muhanna, 1994:912)
Results for Trace Test for Cointegrating Vector
Table 2a
Sector 0
Johansen tests for cointegration
Trend: constant
Sample: 1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
June 27-28, 2012
Cambridge, UK
LL
356.99761
395.24529
408.28555
416.56421
423.32156
424.33671
Number of obs =
Lags =
eigenvalue
.
0.92848
0.59316
0.43501
0.37251
0.06762
29
2
5%
trace
critical
statistic
value
134.6782
68.52
58.1829
47.21
32.1023
29.68
15.5450
15.41
2.0303*
3.76
20
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Table 2b
Sector 1
Johansen tests for cointegration
Trend: constant
Sample: 1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
Number of obs =
Lags =
LL
281.49667
310.20511
321.79697
329.80182
333.67385
335.07264
eigenvalue
.
0.86192
0.55042
0.42424
0.23435
0.09196
29
2
5%
trace
critical
statistic
value
107.1519
68.52
49.7351
47.21
26.5513*
29.68
10.5416
15.41
2.7976
3.76
Table 2c
Sector 2
Johansen tests for cointegration
Trend: constant
Sample: 1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
Number of obs =
Lags =
LL
307.75654
329.31429
339.38405
344.35754
347.24741
348.17139
eigenvalue
.
0.77389
0.50066
0.29036
0.18070
0.06173
29
2
5%
trace
critical
statistic
value
80.8297
68.52
37.7142*
47.21
17.5747
29.68
7.6277
15.41
1.8479
3.76
Table 2d
Sector 3
Johansen tests for cointegration
Trend: constant
Sample: 1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
June 27-28, 2012
Cambridge, UK
LL
276.54163
295.00419
310.9101
320.86889
323.97805
325.4728
Number of obs =
Lags =
eigenvalue
.
0.72009
0.66612
0.49682
0.19299
0.09795
29
2
5%
trace
critical
statistic
value
97.8623
68.52
60.9372
47.21
29.1254*
29.68
9.2078
15.41
2.9895
3.76
21
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Table 2e
Sector 4
Johansen tests for cointegration
Trend: constant
Sample:
1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
Number of obs =
Lags =
LL
247.68692
264.54345
278.44208
287.94386
291.8668
295.59271
29
2
5%
trace
critical
statistic
value
95.8116
68.52
62.0985
47.21
34.3013
29.68
15.2977*
15.41
7.4518
3.76
eigenvalue
.
0.68730
0.61654
0.48071
0.23704
0.22660
Table 2f
Sector 5
Johansen tests for cointegration
Trend: constant
Sample:
1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
Number of obs =
Lags =
LL
383.87484
402.83103
412.56721
418.09083
421.32271
421.5472
eigenvalue
.
0.72946
0.48904
0.31678
0.19980
0.01536
29
2
5%
trace
critical
statistic
value
75.3447
68.52
37.4323*
47.21
17.9600
29.68
6.9127
15.41
0.4490
3.76
Table 2g
Sector 6
Johansen tests for cointegration
Trend: constant
Sample: 1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
Number of obs =
Lags =
LL
337.85647
358.2123
368.71177
375.65743
379.94405
382.28019
eigenvalue
.
0.75435
0.51524
0.38060
0.25594
0.14880
29
2
5%
trace
critical
statistic
value
88.8474
68.52
48.1358
47.21
27.1368*
29.68
13.2455
15.41
4.6723
3.76
Table 2h
Sector 7
Johansen tests for cointegration
Trend: constant
Sample:
1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
June 27-28, 2012
Cambridge, UK
LL
370.51508
382.64727
393.51361
398.83227
402.22917
403.62435
Number of obs =
Lags =
eigenvalue
.
0.56686
0.52735
0.30705
0.20885
0.09174
29
2
5%
trace
critical
statistic
value
66.2185*
68.52
41.9542
47.21
20.2215
29.68
9.5842
15.41
2.7904
3.76
22
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Table 2i
Sector 8
Johansen tests for cointegration
Trend: constant
Sample:
1982 - 2010
maximum
rank
0
1
2
3
4
5
parms
30
39
46
51
54
55
Number of obs =
Lags =
LL
399.7938
410.13805
419.18136
424.82861
427.32992
429.25134
eigenvalue
.
0.51002
0.46403
0.32258
0.15845
0.12411
29
2
5%
trace
critical
statistic
value
58.9151*
68.52
38.2266
47.21
20.1400
29.68
8.8455
15.41
3.8428
3.76
Results of Normalized Cointegrating Coefficients
Table 3a
Sector 0
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-.6060141
.8931411
-1.993184
-.1693067
1.810574
Std. Err.
.
.0592761
.1284432
.1447514
.0414817
.
z
.
-10.22
6.95
-13.77
-4.08
.
P>|z|
.
0.000
0.000
0.000
0.000
.
[95% Conf. Interval]
.
-.7221931
.641397
-2.276891
-.2506093
.
.
-.4898351
1.144885
-1.709476
-.088004
.
Table 3b
Sector 1
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-1.131859
-2.085972
2.856678
.1094989
.261836
Std. Err.
.
.1309922
.3314365
.1577977
.0112684
.
z
.
-8.64
-6.29
18.10
9.72
.
P>|z|
.
0.000
0.000
0.000
0.000
.
[95% Conf. Interval]
.
-1.388599
-2.735576
2.5474
.0874133
.
.
-.8751192
-1.436369
3.165955
.1315846
.
Table 3c
Sector 2
June 27-28, 2012
Cambridge, UK
23
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-.9276013
-1.932647
2.936185
.6654426
-1.549599
Std. Err.
.
.0876075
.2743563
.4216616
.0850619
.
z
P>|z|
.
-10.59
-7.04
6.96
7.82
.
.
0.000
0.000
0.000
0.000
.
[95% Conf. Interval]
.
-1.099309
-2.470376
2.109743
.4987243
.
.
-.7558938
-1.394919
3.762626
.8321609
.
Table 3d
Sector 3
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-.8279099
-1.237617
4.937531
1.296926
-8.292437
Std. Err.
.
.4080996
1.578216
1.463931
.3905323
.
z
P>|z|
.
-2.03
-0.78
3.37
3.32
.
.
0.042
0.433
0.001
0.001
.
[95% Conf. Interval]
.
-1.627771
-4.330864
2.068279
.5314964
.
.
-.0280493
1.85563
7.806782
2.062355
.
Table 3e
Sector 4
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-.4917603
-7.264583
-6.746218
-3.38468
30.73145
Std. Err.
.
2.016952
.9989268
5.327957
1.320966
.
z
P>|z|
.
-0.24
-7.27
-1.27
-2.56
.
.
0.807
0.000
0.205
0.010
.
[95% Conf. Interval]
.
-4.444913
-9.222443
-17.18882
-5.973726
.
.
3.461393
-5.306722
3.696385
-.7956346
.
Table 3f
Sector 5
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-3.087823
-7.801147
12.77593
.4926158
-4.475497
Std. Err.
.
.2341412
1.645763
2.420679
.19913
.
z
.
-13.19
-4.74
5.28
2.47
.
P>|z|
.
0.000
0.000
0.000
0.013
.
[95% Conf. Interval]
.
-3.546731
-11.02678
8.031487
.1023282
.
.
-2.628915
-4.575511
17.52038
.8829035
.
Table 3g
Sector 6
June 27-28, 2012
Cambridge, UK
24
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-3.101338
30.65089
-26.91914
1.180727
-3.287551
Std. Err.
z
.
.5490705
4.659498
4.642102
.2482223
.
P>|z|
.
-5.65
6.58
-5.80
4.76
.
.
0.000
0.000
0.000
0.000
.
[95% Conf. Interval]
.
-4.177496
21.51844
-36.01749
.6942204
.
.
-2.025179
39.78334
-17.82079
1.667234
.
Table 3h
Sector 7
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-3.239978
3.712937
1.654194
.107072
-5.643254
Std. Err.
.
.2745702
1.938091
.897383
.2681926
.
z
P>|z|
.
-11.80
1.92
1.84
0.40
.
.
0.000
0.055
0.065
0.690
.
[95% Conf. Interval]
.
-3.778126
-.0856527
-.1046447
-.4185758
.
.
-2.70183
7.511526
3.413032
.6327198
.
Table 3i
Sector 8
Johansen normalization restriction imposed
beta
Coef.
_ce1
realimports
gniconstbill
domprice
impprice
simpleavta~f
_cons
1
-2.854489
.6313831
-1.974583
-1.60409
9.299292
June 27-28, 2012
Cambridge, UK
Std. Err.
.
.5562836
1.47894
.6130574
.2347441
.
z
.
-5.13
0.43
-3.22
-6.83
.
P>|z|
.
0.000
0.669
0.001
0.000
.
[95% Conf. Interval]
.
-3.944785
-2.267286
-3.176154
-2.06418
.
.
-1.764194
3.530052
-.7730128
-1.144
.
25
2012 Cambridge Business & Economics Conference
ISBN : 9780974211428
Table 4 : Coefficients of Tariff Rates for Each Sector
Coefficient of
Significance
tariff
level
0
-0.167 ***
1
0.11 ***
2
0.67 ***
3
1.3 ***
4
-3.38 ***
5
0.49 **
6
1.18 ***
7
0.11
8
-1.6 ***
Note: *,**,*** denote significance at 10 percent, 5 percent and 1 percent respectively.
Sector
Graph 1: Malaysian Manufactures Imports and Exports from 1990-2010
Source: World Development Indicator website
June 27-28, 2012
Cambridge, UK
26