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: [email protected]June 27-28, 2012 Cambridge, UK 1
38
Embed
The Impacts of Tariff Reductions on Real Imports in … Mohamad.docx · Web viewThe Impact of Tariff Reductions on Real Imports in Malaysia from 1980-2010 Abstract This paper investigates
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
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,
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+∑i=1
k
d ∆ Yt−1+et (5)
June 27-28, 2012Cambridge, UK 10
2012 Cambridge Business & Economics Conference ISBN : 9780974211428
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
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:
June 27-28, 2012Cambridge, UK 11
2012 Cambridge Business & Economics Conference ISBN : 9780974211428
∆ln Mt = b0 + b1i∆ lnM t-2 + ∑i=0
n
b 2i ∆ln Yt-2+ ∑i=0
n
b 3i ∆lnIt-2 + ∑i=0
n
b4 i ∆lnDt-2 +∑i=0
n
b5 i ∆lnTt-2
+ b6 ECt-2 + error term (6)
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.
June 27-28, 2012Cambridge, UK 12
2012 Cambridge Business & Economics Conference ISBN : 9780974211428
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
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
June 27-28, 2012Cambridge, UK 13
2012 Cambridge Business & Economics Conference ISBN : 9780974211428
in tariff rates do or do not, influence the changes in import demands for different goods in
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
June 27-28, 2012Cambridge, UK 14
2012 Cambridge Business & Economics Conference ISBN : 9780974211428
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.
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
June 27-28, 2012Cambridge, UK 15
2012 Cambridge Business & Economics Conference ISBN : 9780974211428
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.
June 27-28, 2012Cambridge, UK 16
iTaken from the Concise Encyclopedia of Economics, http://www.econlib.org/library/Enc/FreeTrade.htmliiStatistics were taken from the paper by Otsuki, 2011 at http://www.osipp.osaka-u.ac.jp/archives/DP/2011/DP2011E006.pdfiiiFor a more information please visit http://unstats.un.org/unsd/cr/registry/regcst.asp?cl=14ivInformation obtained from the ASEAN Secretariat websitevTaken from the MITI website, http://www.miti.gov.my/cms/content.jsp?id=com.tms.cms.section.Section_f5694606-c0a81573-78d578d5-759be8c9viTariff rates obtained by the World Trade Indicators Report 2010 athttp://info.worldbank.org/etools/wti/docs/Malaysia_taag.pdf
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
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
Appendix
Table 1a
Sector/Variable 0 1 2 3
ADF level
FD ADF level
FD ADF level
FD ADF level
FD
ln imports -2.83-
3.84
* -3.15-
3.84
* -2.48-
0.41
** -2.17-
2.75
*
lngni -0.81-
3.43
** -0.81-
3.43
** -0.81-
3.43
** -0.81-
3.43
**
lndomprice -2.14-
3.41
** -2.23-
3.41
* -1.56-
4.17
*** -1.43
-3.25
***
lnimpprice -4.33-
4.65
*** -2.44
-4.6
5 -2.22
-4.49
*** -1.91
-2.67
***
lnsimpleavtariff -2.26-
4.30
*** -2.58
-4.3
0
*** -2.50
-3.70
** -1.72-
3.20
**
*,**,*** 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)
*,**,*** 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)
Table 1c
Sector/Variable 8
ADF level
FD
ln imports -0.154-
2.823
*
lngni -0.802-
3.596
**
lndomprice -0.925-
3.545
**
lnimpprice -3.88-
3.649
**
lnsimpleavtariff -2.336-
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)