Price and income elasticities of consumer goods imports of Pakistan.
Sarmad, Khwaja ; Mahmood, Riaz
I. INTRODUCTION
Estimation of disaggregated import elasticities for developing
countries presents a formidable data-handling problem. The available
studies on the subject are concerned mostly with the estimation of
income and price elasticities of imports at a disaggregated level
corresponding to the one-digit level of the Standard International Trade
Classification (SITC), see, e.g., Khan [1], Melo and Vogt [4], Nguyen
and Bhuyan [5]. Consequently, they apply a common elasticity estimate to
all commodity sub-groups.
The lack of disaggregated estimates of import elasticities is a
serious constraint on the efforts to quantify the effects of policy
measures on the volume of imports and economic welfare in general. In
this study an attempt has been made to overcome this limitation by
estimating price and income elasticities of the consumer goods imports
of Pakistan at the three-digit level of the SITC. Consumer goods imports
have also been distinguished by functional classes, viz. consumer goods
and raw materials for consumer goods.
In this study the importance of changes in relative prices, customs
duties and an income variable as explanatory variables determining the
quantity of consumer goods imports of Pakistan, has been investigated
for the period from 1969-70 to 1979-80. During this time there was a
growing structural concentration of imports in favour of consumer goods,
which, by the end of the Seventies, accounted for around 60 percent of
the total imports. There is also evidence of concentration between
different products of the same industry, which suggests that some
benefit can be gained by studying imports on a disaggregative basis, in
particular by using information at the three-digit SITC level as the
basic data. The fact that this is also the decision level for tariff policy provides additional cause for a disaggregative approach.
The required data on quantity and value of imports were generated
by aggregating the six-digit series available in the Foreign Trade
Statistics [6]. Prices for domestic substitutes were obtained from
Government of Pakistan publications [8], [9], [10] and [11]. The
division of consumer imports into functional classes was made according
to the methodology outlined in [3]. The value of the selected
commodities in the categories of consumer goods and raw materials for
consumer goods accounts for 68 percent and 59 percent respectively of
the total imports in these categories. Together they account for 35
percent of the total imports of Pakistan during the period from 1969-70
to 1979-80.
II. THE IMPORT FUNCTION
The following import demand function is estimated:
log [M.sup.d.sub.it] = [e.sub.0i] + [e.sub.li] log (1 + [CD.sub.i])
[P.sub.im]/[P.sub.di] + [e.sub.2i] log [Y.sub.t] + [u.sub.it] ... (1)
where
[M.sup.d.sub.i] = quantity demanded of the ith import commodity;
[P.sub.im] = price of the ith import commodity;
[P.sub.di] = price of a domestic substitute;
[CD.sub.i] = custom duty on the ith commodity;
[Y.sub.t] = real income variable - the proxies used are real
consumption expenditure for consumer goods imports, and real value added in the manufacturing sector for imports of raw materials for consumer
goods; and
[u.sub.i] = random disturbance term.
The specification of the import function in logarithms allows the
import elasticities to be obtained directly from equation (1). The
underlying assumptions for estimating the import equations are as
follows: (i) importers are always in equilibrium ([M.sup.d.sub.i] =
[M.sub.i]); and (ii) real private consumption expenditure, real value
added in the manufacturing sector, the price of the imported commodity
P/m and that of the domestic substitute are exogenously determined. In
general, the adjusted relative price elasticities ([e.sub.li]s) are
expected to be less than zero and the income elasticities ([e.sub.zi]s)
greater than zero. Negative values for the income elasticity can also be
justified, when domestic production increases more than consumption with
a rise in the level of income. (1)
III. RESULTS
The estimates of the parameters of equation (1) were obtained for
various commodity groups and are reported in Table 1. The table also
reports the t-value in parentheses, the value of [[bar.R].sup.2], the
Durbin-Watson (D.W.) statistic, and the Standard Error (S.E.) of the
estimated equation.
For a number of cases, code-numbered 051 (Fresh Fruits), 074 (Tea),
211 (Waste and Used Leather), 231 (Crude Rubber), 422 (Palm Oil), 512
(Acids and Compounds) and 561 (Urea), there was evidence of first-order
autocorrelation in the residuals, [u.sub.t].
The results in Table 1 show that out of 26 cases the price
elasticity estimates are statistically significant at the 95-percent
level for thirteen groups of imports; for three cases the estimates are
statistically significant at the 90-percent level; and for another five
cases the estimates are statistically significant at the 80-percent
level. For almost half the cases the price elasticities have the
expected negative sign. (2)
IV. AGGREGATION BIAS
To obtain estimates of price and income elasticities for aggregate
imports on the basis of disaggregate estimates, Khan [1] has suggested
that the estimates be adjusted by the "distribution
elasticities" which, if not accounted for, would bias the aggregate
estimates. To investigate the aggregation bias we have calculated the
price and income elasticities for consumer goods imports at the
one-digit level of the SITC directly and on the basis of disaggregated
import elasticities.
The distribution elasticities are calculated from the following
equation and are reported in Table 2:
[DELTA]log [P.sub.it] = [[beta].sub.i][DELTA] log [P.sub.t] +
[v.sub.t] ... ... ... ... ... (2)
where
[P.sub.i] = [P.sub.im]/[P.sub.di] and
[P.sub.t] = [summation over (i)][P.sup..sub.i],
[[beta].sub.i]s are the distribution elasticities; and
[v.sub.t] is a random disturbance term.
On the basis of the distribution elasticities reported in Table 2,
the derived elasticities have been obtained, using average shares of
individual commodity groups in the total of the aggregate imports
(one-digit level) for the entire period from 1969-70 to 1979-80. The
directly estimated elasticities and the derived estimates are reported
in Table 3.
The results in Table 3 show that the elasticities derived from
disaggregated estimates diverge significantly from those obtained
directly from aggregate import equations reported in Table 2. The
direction of aggregation bias is upward for the price and income
elasticities of Food Products (0) and for the price elasticities of
Crude Materials (2) and Animal and Vegetable Oils (4). The other
commodity groups show a downward bias.
V. CONCLUSIONS
On the basis of the data for the period from 1969-70 to 1979-80 we
estimate import elasticities at a disaggregation level corresponding to
the three-digit level of the PSTC for nine different sub-groups in the
Consumer Goods category and thirteen different sub-groups in the Raw
Materials for Consumer Goods category. Price and income elasticities
were also obtained for aggregate imports, corresponding to the one-digit
level of the PSTC.
The results showed that relative price, adjusted for customs duties
and an income variable, were enough to explain a large proportion of the
variation in the imports of individual commodity groups as well as in
aggregate imports. For several sub-groups the price elasticities were
significant and had the expected signs. The income elasticities are on
the higher side. There were some cases in which the relative-price
variable turned out to be statistically insignificant. In a few other
instances the relative-price variable was significant but positive. In
these cases there is a need to specify additional factors which can
influence import demand.
The derived estimates of import elasticities were lower than those
obtained directly from the equations estimated for aggregate imports,
corresponding to the one-digit level of SITC, implying a downward
aggregation bias. The important exceptions were the price and income
elasticities for the imports of Animal and Vegetable Oils and Chemicals,
which were biased upward.
Comments on "Price and Income Elasticities of Consumer Goods
Imports of Pakistan" (1)
I share the authors' assertion that disaggregated import
demand functions are hard to come by for the Third World countries. They
should be commended for their efforts to fill this void. My comments are
intended to raise a few conceptual points and should not be construed as
a criticism of the authors' professional competence.
1. Data
Use of Pakistan's trade data on functional classification on
commodities into consumption goods and raw materials for consumption
goods begs for a reconciliation with a classification of commodities
based on the economic characteristics of a commodity. The apparent
arbitrariness in Pakistan's trade data is best removed in an
input-output framework where a clear distinction between intermediate
use and final demand use of commodities is maintained. I assume that the
authors had an access to PIDE's input-output tables for 1975-76. I
wonder why they did not modify the functional classification of
commodities into meaningful categories based on the economic
characteristics.
A related point is that of the use of prices of domestic
substitutes. What if there is no imaginable domestic substitute
available? What kind of prices would one use? Palm oil or crude rubber
are good examples of this issue. It may be that an industry in question
needs these items which are imported, and without whose availability
their outputs cannot be produced. It would be, in my opinion, desirable
to distinguish between competitive and non-competitive imports.
2. Model Specification
My comments refer to the choice of real income as an explanatory
variable. A correction is required in the explanation of the symbol
[y.sub.t]: real consumption expenditure for consumer goods should be
read as real private consumption expenditure, as it is clearly stated in
the assumptions listed by the authors.
Firstly, there are problems in using Pakistan's National
Accounts data on total consumption expenditure. These aggregate data in
the Accounts are obtained residually. Then, I assume, some other ad hoc basis is used to disaggregate it further into private and government
consumption expenditure.
Secondly, the choice of deflators to transform the current-prices
data into constant prices needs to be carefully examined. It is not made
explicit in the paper what the base-year is of the prices that have been
used to transform the data. If I recall the PIDE's
Macro-econometric Model, 1959-60 was the base year. I noticed from the
Model's series of private consumption that there were two or three
years when the observed consumption expenditure exceeded the observed
income. The unreliable nature of data on this variable is clearly
evident.
Thirdly, I would like to know the deflators used for determining
real value added in manufacturing. How does one deflate various
components of the value added such as wages, salaries and profits? Is it
fair to assume that 100 percent of the profit income earned by
industries is spent? It must be remembered that profits are not the only
component of the value added.
Suppose the choice of real private consumption expenditure as a
proxy for real income is a good one. This appears as an explanatory
variable in the equation. How can one digest the fact that while
consumption is usually a function of income it has been used here as a
dependent variable which is a function of itself. It is then used to
explain the behaviour of another dependent variable--in this case,
demand for an imported commodity.
3. Interpretation of Results
Invariably, the estimated price and income elasticities have been
compared with the results obtained from other studies--e.g., the one by
Khan for Venezuela. In the absence of any description of these other
studies with respect to the choice of variable, estimation methods used,
size of the sample period, functional forms, etc., it is rather
arbitrary to draw any comparative inferences.
In my opinion, such a comparison is unwarranted: for, as the
authors claim, many of these other studies are not as disaggregated as
their own. The authors further claim that increase in the magnitude of
income elasticities over time increases the degree of
"openness" of an economy. I do not quite understand this
point. Does it relate to import liberalization as a function of rising
income levels or what? Do the facts surrounding import policy as an
allocative instrument of resources in Pakistan justify this
prescription.
Dr Aftab A. Syed
Senior Analyst
Input-Output Division, Statistics Canada, Ottawa (Canada)
(1) Comments by the discussant are based on the paper distributed
and presented in the meeting and not on the revised version of the
paper. (Editor)
REFERENCES
[1.] Khan, M. S. "The Structure and Behaviour of Imports of
Venezuela". Review of Economics and Statistics. Vol. LXII, No. 2.
1975.
[2.] Kreinin, M. "Disaggregate Import Demand
Functions--Further results". Southern Economic Journal. Vol. 40,
No. 1. 1973.
[3.] Mahmood, Riaz. Growth and Structure of 1reports of Pakistan
(1959-60 to 1978-79). Islamabad: Pakistan Institute of Development
Economics. 1980. (Statistical Paper Series No. 3).
[4.] Melo, O., and Michael G. Vogt. "Determinants of the
Demand for Imports of Venezuela". Journal of Development Economics.
Vol. 14, No. 3. 1984.
[5.] Nguyen, D. T., and A. R. Bhuyan. "Elasticities of Export
and Import Demand in Some South Asian Countries. Some Estimates".
Bangladesh Development Studies. Vol. V, No. 2. 1977.
[6.] Pakistan. Central Statistical Office. Foreign Trade Statistics
of Pakistan. Karachi. (Various Issues)
[7.] Pakistan. Central Statistical Office. Pakistan Standard Trade
Classification. Revision 1, (PSTC). Karachi. July 1966.
[8.] Pakistan. Finance Division. Economic Adviser's Wing.
Pakistan: Basic Facts. Islamabad. (Various Issues)
[9.] Pakistan. Finance Division. Economic Adviser's Wing.
Pakistan Economic Survey. Islamabad. (Various Issues)
[10.] Pakistan. Ministry of Food, Agriculture and Cooperatives.
Food and Agriculture Division (Planning Unit). Agricultural Statistics
of Pakistan. Islamabad (Various Issues)
[11.] Pakistan. Statistics Division. Pakistan Statistical Yearbook
Karachi. (Various Issues)
[12.] Price, J. E., and J. B. Thornblade. "U.S. Import Demand
Functions Disaggregated by Country and Commodity". Southern
Economic Journal. Vol. 39, No. 1. 1972.
(1) Given that import demand is excess demand, very little can be
said in general about the properties of the import demand function other
than that it is homogeneous of degree zero in absolute prices. As such
there can be no a priori presumptions about the elasticities without
additional assumptions.
(2) Kreinin [2] and Price and Thornblade [12] have suggested that a
positive price elasticity may be due to the high level of
disaggregation.
KHWAJA SARMAD and RIAZ MAHMOOD *
* The authors are, respectively, Senior Research Economist and
Staff Economist at the Pakistan Institute of Development Economics
(PIDE), Islamabad. They are grateful to Prof. Syed Nawab Haider Naqvi,
Director, PIDE, for his constant encouragement in completing this study.
They also thank Mr Masood Ashfaque for computing assistance.
Table 1
Estimates of Import Equations
Commodity PSTC Price Income
Group Equivalent Elasticity Elasticity
(1) (2) (3) (4)
Consumer Goods
Food Products 0 1.040 1.217
(1.824) (1.319)
Milk and Cream 022 -1.408 4.209
(-4.592) (7.200)
Milk Food for 099 0.031 5.852
Infants (0.035) (5.733)
Milk, Cream and 002 + 099 -1.351 4.267
Milk Food (-4.354) (7.929)
Wheat, unmilled 041 1.191 0.422
(2.553) (0.436)
Fresh Fruits 051 -0.908 1.379
(-4.104) (3.667)
Tea 074 0.588 5.935
(0.164) (1.325)
Spices 075 0.452 1.596
(1.222) (1.592)
Petroleum 332 0.677 4.507
(0.377) (1.854)
Cleansing 554 -1.576 1.484
Preparations (-2.131) (2.702)
Raw Materials for
Consumer Goods
Crude Materials 2 0.508 1.745
(4.090) (6.795)
Waste and Used 211 0.114 1.610
Leather (0.616) (2.195)
Crude Rubber 231 0.351 0.319
(1.168) (0.416)
Wool 262 -0.489 -1.585
(-1.268) (-1.343)
Worn Clothings 267 0.411 3.432
(1.394) (6.439)
Crude Petroleum 331 -0.200 1.785
(-3.288) (5.009)
Animal and 4 -0.286 4.601
Vegetable Oils (-0.878) (3.812)
Animal Tallow 411 -0.661 2.017
(-1.969) (3.346)
Soya bean Oil 421 -0.788 3.591
(-2.194) (3.360)
Palm Oil 422 1.682 9.893
(1.426) (2.588)
Chemicals 5 -0.736 4.216
(-0.721) (1.762)
Acids and Compounds 512 -0.131 1.810
(-0.517) (4.748)
Dyes and Coal Tar 531 0.891 -5.140
(4.306) (-10.496)
Urea 561 0.603 2.970
(2.066) (1.830)
Plastic Moulding 581 -1.051 3.090
Powder (-3.265) (3.639)
Manufactures 6 0.976 4.104
(1.505) (3.686)
Commodity [[bar.R]
Group .sup.2] D.W. S.E.
(1) (5) (6) (7)
Consumer Goods
Food Products 0.407 2.149 0.527
Milk and Cream 0.861 1.559 0.366
Milk Food for 0.756 2.572 0.647
Infants
Milk, Cream and 0.878 1.684 0.338
Milk Food
Wheat, unmilled 0.500 2.282 0.503
Fresh Fruits 0.800 1.327 0.229
Tea 0.294 0.838 1.673
Spices 0.644 2.203 0.379
Petroleum 0.587 1.955 0.836
Cleansing 0.417 1.527 0.332
Preparations
Raw Materials for
Consumer Goods
Crude Materials 0.932 1.931 0.091
Waste and Used 0.392 1.175 0.238
Leather
Crude Rubber 0.524 1.170 0.142
Wool 0.372 1.756 0.419
Worn Clothings 0.814 1.652 0.229
Crude Petroleum 0.712 2.048 0.091
Animal and 0.680 1.900 0.347
Vegetable Oils
Animal Tallow 0.479 2.758 0.212
Soya bean Oil 0.484 2.536 0.374
Palm Oil 0.721 1.127 1.167
Chemicals 0.137 1.495 0.743
Acids and Compounds 0.676 1.290 0.161
Dyes and Coal Tar 0.924 1.788 0.211
Urea 0.380 1.179 0.699
Plastic Moulding 0.546 1.751 0.250
Powder
Manufactures 0.571 2.210 0.480
Table 2
Distribution Elasticities
PSTC Distribution
Commodity Group Equivalent Elasticity
Consumer Goods
Milk and Cream 022 0.400
Milk Food 099 0.203
Wheat, unmilled 041 1.070
Fresh Fruits 051 1.695
Tea, Black 074 0.069
Spices 075 0.906
Raw Materials for Consumer Goods
Waste Leather 211 1.396
Crude Rubber 231 0.417
Wool 262 1.075
Worn Clothing 267 0.566
Animal Tallow 411 0.469
Soyabean Oil 421 1.263
Palm Oil 422 1.100
Acids and Compounds 512 0.615
Dyes and Coal Tar 531 1.342
Urea 561 2.279
Plastic Moulding Powder 581 1.127
Table 3
Directly Estimated and Derived Elasticities for Imports
SITC Direct
Equivalent Estimate
Commodity
Group Price Income Price Income
Elas- Elas- Elas- Elas-
ticity ticity ticity ticity
Consumer Goods
Food Products 0 0 1.040 1.217
Raw Materials
for Consumer Goods
Crude Materials 2 2 0.508 1.745
Animal & Vegetable
Oils 4 4 -0.286 4.601
Chemicals 5 5 -0.736 4.216
Derived
Estimate
Commodity
Group Price Income
Elas- Elas-
ticity ticity
Consumer Goods
Food Products 0.696 0.378
Raw Materials
for Consumer Goods
Crude Materials 0.072 0.709
Animal & Vegetable
Oils 0.279 6.755
Chemicals 1.021 4.638