role adv agri tech

8/2/2019 Role Adv Agri Tech

1/125


2/125

Forman Christian College (A Chartered University) was founded in 1864 by

Dr. Charles W. Forman, a Presbyterian missionary from USA. In 1972 the

college was nationalized by the government of Pakistan and it was returned to

the present owners of the college on March 19, 2003. In March 2004, the

government of Pakistan granted university status to Forman Christian

College.

For submission of articles for publication and purchase of Forman Journal of

Economic Studies:

Contact

Editor

Forman Journal of Economic Studies

Department of EconomicsForman Christian College (A Chartered University)

Ferozepur Road, Lahore-54600, Pakistan

E mail:[email protected]

Ph: +92 42 99231581-8, Ext: 380

Fax: +92 42 99230703

www.fccollege.edu.pk

Subscription Rate

Inland

Students Rs.200General Rs.300

Overseas US $ 40

ISSN: 1990-391XAbbreviated Key Title: Forman j. econ. stud.

Recognized by: HEC (Ref. Letter No. DD/SS&H/JOUR/2011/78)Internationally Indexed by: EconLit, EBSCOhostTM, IBSS & UlrichsJournal Website: www.fccollege.edu.pk/academics/departments/academic-

departments/department-of-economics/research


3/125

i

Forman Journal o f Economic Stud ies

PatronPeter H. Armacost

Editor Associate Editors Managing EditorMuhammad Aslam Chaudhary Tanvir Ahmed Ghulam Shabbir

Muhammad Akbar

National Advisory BoardAsad Zaman International Islamic University, IslamabadEatzaz Ahmad Quaid-i-Azam University, IslamabadFazal Hussain PIDE, IslamabadImran Sharif Chaudhry Bahauddin Zakariya University, MultanKhair-uz-Zaman Gomal University D. I. KhanMichael Murphy Forman Christian College University, LahoreMuhammad Aslam Lahore School of Economics, Lahore

Muhammad Idrees Quaid-i-Azam University, IslamabadMumtaz Anwar Ch. University of the Punjab, LahoreMushtaq Ahmed Lahore University of Management Sciences, LahoreNaveed Ahmed Institute of Business Administration, KarachiRazaque H. Bhatti International Islamic University, IslamabadShah Nawaz Malik Bahauddin Zakariya University, Multan

International Advisory BoardDavid Graham Institute of Defense Analysis, Alexandria, VA, USAIsmail Cole University of California, PA, USAJames Fackler University of Kentucky, USAKiyoshi Abe Hanazono, Hanamigaaku, Chiba City, Japan

M. Arshad Chaudhary University of California, PA, USAMack Ott Gravitas International LLC, USAMarwan M. El Nasser Fredonia University, USAMuhammad Ahsan Academic Research Consultant / Adviser, UKNasim S. Sherazi Islamic Development Bank, Saudi ArabiaRoger Kormendi University of Michigan, USASarkar Amin Uddin Fredonia University, USASoma Ghosh Albright College, USAStephen Ferris Carleton University, CanadaSteve Margolis North Carolina State University, USAToseef Azid Taibah University Madinah, Saudi ArabiaThomas Zorn University of Nebraska, USA


4/125

ii

Declaration

The findings, interpretations and conclusions expressed in this journal are

entirely those of the authors and should not be attributed in any manner to the

FCC or Editorial Board. The journal does not guarantee accuracy of the data

included in this publication and accepts no responsibility for any consequence

of their use.


5/125

iii

FORMAN JOURNAL OF ECONOMIC STUDIES

Volume: 7 2011 January-December

Estimating Food Demand Elasticities in Pakistan: 1

An Application of Almost Ideal Demand SystemBabar Aziz, Khalil Mudassar,

Zahid Iqbal and Ijaz Hussain

Exchange Rate Exposure on the Automotive Industry: 25

Evidence from USA and Japan

Zeresh Mall, Saqib Jafarey,

Shabib Haider Syed and Ijaz Hussain

Rice Policy Reforms of the European Union and its 55

Impact on Rice Exports from Pakistan

Mohammad Aslam

Modeling Demand for Money in Pakistan: An ARDL Approach 75

Muhammad Asad, Shabib Haider Syed

and Ijaz Hussain

Role of Advance Agri-Technologies in Reducing the Rural 89

Poverty in Central Punjab, Pakistan

Hazoor Muhammad Sabir

and Safdar Hussain Tahir

Factors Influencing Student Achievement Scores: 99

Public vs. Private SchoolsShahnaz Rashid

Book Review 117

DEPARTMENT OF ECONOMICSFORMAN CHRISTIAN COLLEGE (A CHARTERED UNIVERSITY)

FEROZEPUR ROAD, LAHORE, PAKISTAN


6/125


Vol. 7, 2011 (JanuaryDecember) pp. 1-24

Estimating Food Demand Elasticities in Pakistan:

An Application of Almost Ideal Demand System

Babar Aziz, Khalil Mudassar, Zahid Iqbal and Ijaz Hussain1

Abstract

The main focus of the study is to estimate the rural-urban income and own

price elasticities across a range of consumption quintiles. The Linear

Approximate Almost Ideal Demand System (LAAIDS) is used to estimate the

parameters of aggregate food commodity groups. Due to the specific features

of the data, spatial variations in regional prices are estimated and used as

proxies for food prices (i.e. unit values) by using household survey data.

Regarding household specific elasticity estimates, households exhibit

increasing consumption of vegetables, fruits, milk and meats with higher

income. The expenditure elasticities are larger in rural areas compared to

urban areas and expenditures on most food groups increase at a decreasing

rate as income increases. Expenditure elasticities for all food groups were

positive and less than one, except for fruits, meats, and milk that have been

identified as luxuries. Cereals tend to have the lowest expenditure elasticity of

demand. The uncompensated own-price elasticities of demand for all food

groups are negative and their absolute amounts are lower than unity i.e.

demand reacts in-elastically to own-price changes, except for meats (elastic).

According to the values of the cross-price elasticities and on the level of all

selected food groups, only substitution relationships are observed. The high

price elasticities of demand for many food items stress the importance of food

price changes for households, and their reactions should be taken into

account in the development of comprehensive agricultural and food policies in

Pakistan.

Keywords: Consumer demand analysis; PIHS data; LAAIDS; price and

expenditure elasticities

JEL classification: D01, D12, C31

1The authors are Associate Professor of Economics at Forman Christian College (A

Chartered University) Lahore, Assistant Professors of Economics at Government SE College

Bahawalpur, Forman Christian College (A Chartered University) Lahore and Gomal

University D. I. Khan, respectively.


7/125

Babar, Khalil, Zahid and Ijaz

2

I. Introduction

Demand elasticities for a particular country provide valuable

information for policy analysts in understanding the pattern of growth of thenational food consumption. Specific country elasticities are influenced by both

the level of income attained and the quantities of food that are currently eaten

by the consumer. Estimation of complete demand functions is incrediblyuseful not only in obtaining price elasticities, but also in getting reliable

estimates of expenditure (income) elasticities. The measurement of these

elasticities is required for the design of many different policies; for example,

intelligent policy design for indirect taxation and subsidies requiresknowledge of these elasticities for taxable commodities and, in addition, in the

projections for future food consumption2.

Such knowledge would normally be obtained by the analysis of time-

series data on demand for commodities, prices, and incomes. For Pakistan aswell as for many developing countries, there is typically rather few time-series

data from which price elasticities can be inferred. As a result of this limitationand with the available cross-sectional data resulting from extensive surveys onhousehold expenditures, most studies in Pakistan concentrated on the

estimation of expenditure elasticities (Engel relationship) and overlooked the

price elasticities.

As the estimation of complete demand functions is incredibly usefulnot only in obtaining price elasticities, but also in getting reliable estimates of

expenditure (income) elasticities, so towards this end the study lays out the

estimated rural-urban income and own price elasticities, across a range ofconsumption quintiles, of aggregated food groups. Section II addresses the

issue that how price elasticities could be estimated from cross sectional data?

Section III is specified for model specification of LAAIDS, for the estimationof complete demand system along with the description of income and price

elasticity formulas. Section IV highlights the adopted estimation technique

along with description of the variables. The empirical findings are reported in

section V. Concluding remarks are presented in section VI along with policyimplications for Pakistan.

2

See e.g. Deaton (1986, 1987, 1988), for a meticulous discussion.


8/125

Food Demand Elasticities in Pakistan

3

II. Price Elasticities from HIES data

Deaton (1987) developed a methodology by using household surveydata to detect the spatial variation in prices and to estimate the price

elasticities by comparing spatial price variation to spatial demand patterns.

The household surveys contain information on the spatial distribution of

prices, and thus, by recovering this information in a useful form, there is apotential for estimating the impact of prices on quantity demanded. Since

prices for food products are not provided by the survey, the ratio of

expenditure to purchased quantity can be used as a proxy for prices. Theseprices should be corrected before being incorporated into the demand system

according to the causes of cross-sectional price variations.

Prais and Houthakker (1955, 1971) identify price variation due to

region, price discrimination, services purchased with the commodity, seasonaleffects, and quality differences caused by heterogeneous commodity

aggregates. When the structure of demand is relatively constant, price

variation can be attributed to changed supply conditions and can be used toidentify commodity demand curves. In order to interpret correctly the effects

of prices in the analysis of household budget data, the causes of cross-

sectional price variations must be identified and only supply related pricevariations should be used to estimate the demand functions.

In the survey data used by Deaton, there are variations in the cross-

sectional price data due to region, household characteristics (male, female, age

groups), seasonal effects, aggregation of the commodities, etc. Similar data forthe survey data used by Deaton are available for a wide range of developing

countries so that the technique should have wide applicability.

Keeping in line with the methodology of Deaton nine aggregated food

commodity groups were chosen for the analysis of this study: cereals (CR),pulses (PL), fruits (FR), edible oils and fats (EOF), sugar and gur (SG), meats

(MT), vegetables (VG), tea, coffee and soft drinks (TCS), and milk and milk

products (MMP). Each of selected food group is not a homogeneous good but

consists of a number of components. For example, in the data it is possible toseparate the cereals group into wheat, rice, and maize, but a category such as

rice does not encompass different kinds of rice, some of which are more

expensive than others. This food-grouping is to reduce the total number ofparameters in the model and then estimation demand system more


9/125


4

manageable3. Each food group includes those commodities that have the same

nutritional value and their prices are very likely to move in tandem and hencethere would be no serious aggregation problem.

The variation in food group prices is due to differences in consumed

items in each group and the variation in prices of each item across provinces.

The latter is due to regional market conditions. Therefore, the price of eachfood group is computed as a weighted average of prices on specific items. The

price obtained is effectively a value and quantity ratio, which is called a unit

value by Deaton (1988) and consequently could be used as a proxy of prices.

This unit value as defined by Deaton is used in this study after the name of

unit value of the aggregated commodity.

Using unit values as price proxies, as in this study, brings about

another specific concern. Unit values are not only affected by the actual pricesconsumers face, but also by the composition of the commodity group. When

separate goods are aggregated into a single commodity group, this leads to

variations in the average price, i.e. unit value of the aggregated commodity,changing with the quantities of the goods of which it is composed. This means

that quality choice in this context is not only a question of differentiated goods

but also quality choice is reflected in the quantity shares of the componentgoods.

The published data of the Pakistans HIES is aggregated at eight rural-

urban regions across four provinces. The ratio of expenditure to quantity, the

cost of the purchase, gives the cost of the commodities for four provinces.This information can be used as a proxy for the prices after calculating the

unit value of the aggregated commodity. Given, for example, different

cereals costs, and then, there will be spatial variation in the costs of this food

group across the regions. This variation can be used to obtain the priceinformation, which is missing in the household survey data. Thus, a complete

demand system can be estimated, and price and expenditure (income)

elasticities can be calculated as a result.

So, in continuation of the previous discussion and keeping in mind thespecific features of the data, the study has made use of spatial variation in

regional prices estimated using household survey data. The estimated spatial

variation in regional prices, as per methodology suggested by Deaton, is usedas proxies for food prices. They are incorporated into the complete food

3

See for instance Abdulai (2002, 2003); and Abdulai and Aubert (2004).


10/125


5

demand analysis, i.e. LAAIDS after calculating the unit value of the

aggregated commodity, to measure own and cross price elasticities for anassortment of food groups.

III. The LAAIDS

The LAAIDS has been chosen as the basic model for the complete

demand system estimation in this study due to its flexible functional form andnimbleness in estimation. In a short and snappy way the demand function of

LAAIDS in budget share form can be expressed as:*ln ln

ic i ij jc i c c

j

w p px P (1)

Where the commodities 1, ,9i and the consumption quintile 1, ,5c .

icw is the budget share of good i in the respective consumption quintile c ,

jcp is the price of good j in the respective quintile, cx is households total

food expenditure in the specific quintile c . *cP is the stones price index, and

i , i , and ij are the parameters that need to be estimated.

The demand elasticities are calculated as functions of the estimated

parameters, and they have standard implications. The specific form of

expenditure elasticity ( i ), which measures sensitivity of demand in response

to changes in consumption expenditure, is as:

1i i iw (2)

The uncompensated (Marshallian) own-price elasticity ( ii ) and cross-price

elasticity (ij ) measure how a change in the price of one product affects the

demand of this product and other products with the total expenditure and other

prices held constant. The specific form of uncompensated own and cross priceelasticities is as, respectively:

1ii ii i iw (3)

ij ii i i j iw w w (4)The compensated (Hicksian) price elasticities own and cross ( *ii and

*

ij ),

which measures the price effects on the demand assuming the real expenditure*

c cx P is constant, is described as:

* 1ii ii i iw w (5)


11/125


6

*ij ij i jw w (6)Also, the compensated price elasticity can be derived easily by using

i , ii ,

andij

, and the following relation:

*

ij ij i jw (7)

In particular, the sign of the calculated *ij indicates the substitutability or

complementarily between the destinations under consideration.

Using the LAAIDS model to estimate the two-stage budgeting demand

function presents several advantages. Probably the most important is that it is

a flexible functional form. The LAAIDS substitution pattern implies an

unconstrained pattern of conditional cross-price across products within sub-segments. This is an advantage, because competition is probably higher

among differentiated products within sub-groups. Another important

advantage of the LAAIDS model is the perfect aggregation over consumers,

without requiring linear Engle curves. This is very important in studies ofaggregate data. Finally, the demand function derived from this model crosses

the price axis, avoiding the presence of virtual prices.

IV. Data and Estimation Procedure

Data for this study is obtained from the Federal Bureau of Statistics(FBS) for the year of 2007-08. FBS provided an electronic copy of the data

sets for four provinces aggregated into five consumption percentiles. The cost

indices of the bundles of the aggregated food commodities are calculated fromthe given data set. The expenditure data are pooled across the four provinces

and five consumption percentiles in each province in the study. It is assumed

that cost indices of the bundles of the food commodities are only differentacross the provinces and for each consumption quintile, but not within the

province according to Deatons methodology. In simple words it is assumed

that households at different consumption percentiles have the same cost

indices for the aggregated food commodities within the same province. Thecost indices of these commodities in each Province are used as proxies for

prices and hence enabled us to estimate income and price elasticities across

these defined consumption quintiles.

No regional elasticities (rural and urban) are estimated keeping in linewith the assumption of no variation in the unit values within the same region.

Our study includes nine aggregated food commodity groups, as defined


12/125


7

earlier. The prices for these commodity aggregates will be proxied by the cost

of these commodity aggregates in each province across the quintiles.

A system of share equations based on first equation and subject to the

restrictions (adding-up, homogeneity, and symmetry) is estimated using

Iterative Seemingly Unrelated Regression (ISUR) method of Zellner. This

method is equivalent to Full Information Maximum Likelihood (FIML)estimation. The adding-up property of demand causes the error covariance

matrix of system to be singular, so one of the expenditure share equations is

dropped from the system to avoid singularity problems. The estimates areinvariant of which equation is deleted from the system. Homogeneity is

maintained by normalizing all of the prices (proxied by the aggregate cost

figures) by the price of others group (OT). The coefficients pertaining to theexpenditure share equation of others aggregate (OT), which is dropped from

the system in the estimation stage, are obtained by using the adding-up

property. Symmetry is imposed during the estimation of the system ofequations. Now, we present the results of our estimation. The above models

are initially estimated for the whole sample of households, regardless of theirincome and consumption levels. Later, households are split according to

consumption quintiles, and the models are estimated for each group.

V. Model Results

The above model in first equation was initially estimated for the whole

sample of households, regardless of their respective consumption quintiles.

Later, households were split according to their consumption patterns, and themodels were estimated for each group. Following Green and Alston (1990,

1991), we assume that the preference structure is such that, in the first stage,

consumers choose how to spend their income among groups of products, such

as food, housing, transportation, health services, education, etc. In the secondstage, the level of expenditure in each group, as determined in the first stage,

is allocated to the commodities in that group.

The empirical results for the specified model for demand functions

(LAAIDS) illustrate that all estimated coefficients agree with a prioritheoretical expectations. As a result of 2nd stage of the two-stage budgeting

process the estimates of the structural parameters for food groups of the

LAAIDS model for the whole sample of households are shown in Table 1.Following the same line of action, the parameters of LAAIDS for 1st quintiles

(low income households) and 5th

quintiles (high income households) quintiles

are reported in Table 2 and 3 respectively. The equation for milk and milk


13/125


8

products was excluded to avoid singularity, but its coefficients were later

recovered with the use of the homogeneity property. The parameters estimatessatisfy the adding-up restriction. Overall, it can also be seen from the

estimated results that a reasonable number of coefficients of the explanatory

variables are significant. Out of eighty one coefficients we have twenty five

ij's with significant t-statistics.

However of interest to researchers and policy makers is the knowledge

concerning elasticities of demand for food. According to value of the

expenditure elasticities, the selected food groups are classified as inferior

goods (i


14/125


Table: 1. Parameter Estimates of LAAIDS for Total Sample and for Aggregated Food Groups

Food

Groups i i 1i 2i 3i 4i 5i 6i 7i 8i 9i0.423 -0.078 0.060 0.103 -0.178 0.194 -0.201 0.087 -0.102 -0.063 -0.030Cereal

(CR)1.419 -1.709* 0.128 1.227 -2.134** 1.698* -2.957** 1.065 -2.745** -0.535 -1.046

0.048 -0.015 0.003 0.084 0.023 0.006 -0.069 0.006 -0.010 0.102 0.007Pulses

(PL)0.507 -1.000 0.254 3.132*** 0.879 0.152 -3.230*** 0.238 0.900 2.714 0.755

-0.039 0.009 -0.005 -0.020 0.008 0.066 -0.020 -0.015 0.001 0.006 0.011Fruits

(FR)-0.987 1.603* -0.743 -1.733* 0.699 4.256*** -2.257** -1.414 0.046 0.417 3.055***

-0.066 -0.007 0.020 0.011 -0.020 0.029 -0.007 -0.002 0.001 0.070 0.005Edible Oil

& Fats

(EOF)-1.269 -0.895 2.441** 0.806 -1.358 1.478 0.565 -0.163 0.029 3.378*** 0.862

0.235 -0.022 -0.017 0.023 -0.003 -0.032 0.020 -0.002 0.013 -0.067 0.006Sugar(SG)3.392*** -1.770* -1.490 1.048 -0.190 -1.096 1.090 -0.143 1.259 -2.120** 0.799

0.176 0.099 -0.045 -0.069 0.086 -0.023 0.041 0.015 0.029 -0.089 0.008Meats

(MT)0.586 2.170** -0.970 -0.801 1.029 -0.207 0.598 0.177 0.760 -0.742 0.285

0.091 -0.011 0.036 0.037 -0.043 -0.154 0.120 -0.060 0.022 0.162 -0.023Vegetables

(VG)0.478 -0.385 1.187 0.686 -0.799 -2.094** 2.724 -1.145 0.913 2.110 -1.206

0.132 -0.008 0.002 0.006 0.045 0.036 -0.026 -0.032 -0.001 0.007 -0.009Tea,

Coffee &

Soft

Drinks

(TCS)

2.487** 0.915 0.282 0.412 2.990** 1.700* -2.193** -2.207** -0.063 0.327 -1.877*

0.149 0.033 -0.056 -0.170 0.074 -0.121 0.141 -0.002 0.060 -0.129 0.022Milk &Milk

Products

(MMP)

0.941 1.399 -2.315** -3.774*** 1.660 -1.987* 3.912*** -0.057 3.015*** -2.020** 1.380

Note: 2nd line of each group describes the t-values, in smaller font size. * * * Indicates significant at one percent

level of significance, * * Indicates significant at five percent level of significance and * Indicates significant at ten

percent level of significance.


15/125


Table: 2. Parameter Estimates of LAAIDS for Quintile 1st

and for Aggregated Food Groups

Food

Groups i i 1i 2i 3i 4i 5i 6i 7i 8i 9i 0.396 -0.073 0.056 0.097 -0.167 0.182 -0.188 0.082 -0.096 -0.059 -0.028 0Cereal

(CR)1.326 -1.597* 0.119 1.147 -1.995* 1.588* -2.765** 0.995 -2.566** -0.500 -0.978

0.045 -0.014 0.003 0.078 0.021 0.005 -0.064 0.005 -0.010 0.096 0.006 0Pulses

(PL)0.474 -0.935 0.238 2.928** 0.821 0.142 3.020*** 0.223 0.842 2.537**8 0.706

-0.037 0.009 -0.004 -0.018 0.008 0.061 -0.018 -0.014 0.001 0.005 0.011 0Fruits

(FR)-0.922 1.498 -0.694 -1.620* 0.654 3.978*** -2.110** -1.322 0.043 0.390 2.856**

-0.061 -0.006 0.018 0.011 -0.018 0.027 -0.006 -0.002 0.001 0.066 0.004 0Edible Oil

& Fats

(EOF)-1.187 -0.836 2.282** 0.754 -1.270 1.381 0.528 -0.153 0.027 3.158*** 0.806

0.219 -0.020 -0.016 0.021 -0.003 -0.030 0.018 -0.002 0.012 -0.062 0.005 0Sugar(SG)3.171*** -1.654* -1.393 0.979 -0.177 -1.024 1.019 -0.133 1.177 -1.982* 0.747

0.164 0.092 -0.042 -0.064 0.081 -0.021 0.039 0.014 0.027 -0.083 0.008 0Meats

(MT)0.548 2.028** -0.907 -0.749 0.962 -0.193 0.559 0.166 0.711 -0.693 0.267

0.085 -0.011 0.033 0.034 -0.040 -0.144 0.112 -0.056 0.020 0.152 -0.021 0Vegetables

(VG)0.447 -0.360 1.109 0.642 -0.747 -1.957* 2.547** -1.071 0.854 1.973* -1.128

0.124 -0.008 0.002 0.005 0.042 0.033 -0.025 -0.030 -0.001 0.006 -0.009 0Tea,

Coffee &

Soft

Drinks

(TCS)

2.325** 0.856 0.263 0.385 2.795** 1.589* -2.050** -2.063** -0.059 0.305 -1.754*

0.140 0.031 -0.053 -0.159 0.069 -0.113 0.132 -0.002 0.056 -0.120 0.020 0Milk &Milk

Products

(MMP)

0.879 1.308 -2.16** -3.5*** 1.552* -1.857* 3.657*** -0.054 2.818** -1.889* 1.290


level of significance, * * Indicates significant at five percent level of significance and * Indicates significant

at ten percent level of significance.


16/125


Table: 3. Parameter Estimates of LAAIDS for Quintile 5th

and for Aggregated Food Groups

Food

Groups i i 1i 2i 3i 4i 5i 6i 7i 8i 9i0.444 -0.082 0.063 0.109 -0.187 0.204 -0.211 0.092 -0.107 -0.066 -0.031Cereal

(CR)1.490 -1.795* 0.134 1.289 -2.241** 1.784* -3.106*** 1.118 -2.883** -0.562 -1.099

0.051 -0.016 0.004 0.088 0.024 0.006 -0.072 0.006 -0.011 0.107 0.007Pulses

(PL)0.533 -1.051 0.267 3.290*** 0.923 0.159 -3.392*** 0.250 0.946 2.850** 0.793

-0.041 0.010 -0.005 -0.021 0.008 0.069 -0.021 -0.016 0.001 0.006 0.012Fruits

(FR)-1.036 1.683* -0.780 -1.820* 0.734 4.470*** -2.371** -1.485 0.048 0.438 3.209***

-0.069 -0.007 0.021 0.012 -0.021 0.030 -0.007 -0.002 0.001 0.074 0.005Edible Oil

& Fats

(EOF)-1.333 -0.940 2.564** 0.847 -1.426 1.552* 0.593 -0.171 0.030 3.548*** 0.906

0.246 -0.023 -0.018 0.024 -0.004 -0.034 0.021 -0.002 0.013 -0.070 0.006Sugar(SG)3.563*** -1.859* -1.565* 1.100 -0.199 -1.151 1.145 -0.150 1.322 -2.227** 0.839

0.185 0.104 -0.047 -0.072 0.091 -0.024 0.043 0.016 0.030 -0.093 0.008Meats

(MT)0.616 2.279** -1.019 -0.842 1.081 -0.217 0.628 0.186 0.798 -0.779 0.299

0.095 -0.012 0.037 0.039 -0.045 -0.162 0.126 -0.063 0.023 0.170 -0.024Vegetables

(VG)0.502 -0.405 1.246 0.721 -0.839 -2.199** 2.861** -1.203 0.959 2.216** -1.267

0.139 -0.008 0.002 0.006 0.047 0.037 -0.028 -0.034 -0.001 0.007 -0.010Tea,

Coffee &

Soft

Drinks

(TCS)

2.612** 0.961 0.296 0.432 3.140*** 1.785* -2.303** -2.317** -0.066 0.343 -1.971*

0.157 0.035 -0.059 -0.179 0.077 -0.127 0.149 -0.002 0.063 -0.135 0.023Milk &Milk

Products

(MMP)

0.988 1.470 -2.431** 3.963*** 1.744* -2.087** 4.108*** -0.060 3.166*** -2.122** 1.449


level of significance, * * Indicates significant at five percent level of significance and * Indicates significant at ten

percent level of significance.


17/125


12

amounts to 0.871 and for vegetables and sugar and gur it amounts to 0.764

and 0.664, respectively. The food groups such as fruits, meats, and milk and

its products have expenditure elasticities larger than unity (i

>1) which

identifies them as luxuries. It is expected that these food groups will

experience an increase in demand when consumers income increases in

tandem with the overall economic growth of the country. However, if realincome of households further decreases, in relative terms, less expenditures

will be allocated to these food commodities. This result indicates that as

households expenditures increase and households diversify their diets, they

tend to increase their consumption of non-staple foods rather than staple

foods.Table: 4. Expenditure (Income) and Marshallian Own-price

Elasticities for Total Sample

Food group Expenditure Own-price

Cereals 0.541 -0.582Pulses 0.871 -0.238

Fruit 1.327 -0.745Edible oils and fats 0.821 -0.247

Sugar and gur 0.664 -0.672

Meats 1.222 -1.053Vegetables 0.764 -0.290

Tea, coffee and soft drinks 0.833 -0.839

Milk and milk products 1.209 -0.898

Another interesting finding is that cereals tend to have the lowest expenditure

elasticity of demand. The consumption of this group is relatively little affectedby income changes and has already occupied a special position in the

Pakistanis diet, as it is a staple food among the population.

The LAAIDS model permits the calculation of elasticities for different

consumption quintiles, so in addition, expenditure elasticities has also beensurged out for the poor and rich households of Pakistan (i.e. for 1

st and 5th

quintile) 1st

quintile refers to the poor group and 5th

quintile is meant for the

upper class having high rate of consumption expenditure share. It is observedthat income elasticities for almost all of the included groups are higher for

lower class and lower for the rich class. Its as per the theoretical

consideration that income elasticities move down ward as income increasesand vice versa. So for poor high income elasticity is expected and the results

of Table 5 confirm it. Among the food groups fruits; meats; tea, coffee and

soft drinks; and milk and milk products with elasticities greater then one


18/125


13

seems to have a luxurious nature for the poor. In addition to these groups

pulses; edible oils and fats; and vegetables with the expenditure elasticityclose to one also conforming their existence very close to the luxurious items.

Table: 5. Expenditure (Income) and Marshallian Own-price

Elasticities for 1st

Quintile

Food group Expenditure Own-

price

Cereals 0.653 -0.694

Pulses 0.878 -0.245

Fruit 1.436 -0.854

Edible oils and fats 0.941 -0.367


Meats 1.350 -1.181

Vegetables 0.873 -0.456



Table 6 demonstrates the expenditure and own price elasticities for the upperclass (i.e. the consumers belonging to 5

thquintile). All the observed

expenditure elasticities are of

Table: 6. Expenditure (Income) and Marshallian Own-price

Elasticities for 5th

Quintile

Food group Expenditure Own-price

Cereals 0.429 -0.470

Pulses 0.864 -0.231

Fruit 1.218 -0.636

Edible oils and fats 0.701 -0.127


Meats 1.094 -0.925

Vegetables 0.653 -0.181



reasonable magnitude. The magnitude of the expenditure elasticities for this

upper class, as per theoretical consideration and prior assumption, is low as

compared to the poor class. Three groups reflect the tendency of being theluxury items like fruits, meats, and milk and milk products with expenditure

elasticities 1.218, 1.094, and 1.114 respectively. No group reveals the status of

Giffen commodity.


19/125


14

Cereal group for both of the income classes shows a behavior of basic

need for the people. The expenditure elasticity of this group is lower ascompared to all other included groups in both of the cases. It is overall 0.541,

and 0.653 for 1st

quintile and 0.429 for 5th

quintile. As a basic need cereal

group is les elastic towards the change in income as it has a certain fixed

proportion in the expenditure of the households.5.2. Uncompensated own-price elasticities

Uncompensated own-price elasticities of demand for all food groups

are negative and consistent with the a priori expectation. The absolute

amounts of these elasticities for all food groups are lower than unity exceptfor meats in total sample of households as displayed in Table 4. The demand

reacts in-elastically to own price changes. An exception is meat where the

elasticity amounts to -1.053 (elastic) thus price changes affect the demand formeat in a greater extent as compared to the other included groups.

The uncompensated own-price elasticities for most the selected food

groups, such as pulses, edible oils and fats, and vegetables are much lowerthan the total expenditure elasticities, implying that responsiveness of demandto own price changes of these aggregates is much lower than to variations in

total expenditure. The largest absolute value of uncompensated own-price

elasticity is calculated for the meats group (i.e. -1.053). This implies thatdemand reacts elastically to changes in the prices of these products. The own

price elasticities are lowest for pulses (-0.238), edible oils and fats (-0.247),

and cereals (-0.582) where demand reacts least to price changes.

Having a look on Table 5, it is observed that meats; tea, coffee and softdrinks, and milk groups showed a high elastic attitude towards the change in

own price, having own price elasticities -1.181, -1.081 and -0.993

respectively. While, on the other hand, pulses and edible oil groups depict alow magnitude of own price elasticities in absolute terms i.e. -0.245 and -

0.367, respectively.

Table 6 reveals the information about the uncompensated own price

elasticities for the rich class (5th

quintile). No own price elasticity is foundhere, which have a magnitude greater than one in absolute terms. However,

meat, and milk and milk products groups, with elasticities -0.925 and -0.803,

respectively, reflect highly responsive towards the change in own price as

compared to the other items pertaining to this aggregate food groups. On the


20/125


15

other side edible oil and fats, and pulses showed a very in-elastic behavior

with elasticity magnitudes -0.127 and -0.231, respectively.

5.3. Compensated own-price elasticities

As predicted by demand theory, the compensated own-price

elasticities are negative for all commodities (see table 8). For all commodity

groups, they are lower in absolute terms than the uncompensated ones.Especially for vegetables, meats, and milk and milk product group, the

compensated own-price elasticities are much smaller in absolute terms than

the uncompensated ones, suggesting that a rise or fall in the price of the

respective commodities would have considerable real expenditure effects.

5.4. Cross-price elasticities

The values of the cross-price elasticities are smaller - in absolute terms

- than those of the expenditure or own-price elasticities. This holds true for

uncompensated and compensated cross-price elasticities (see, Tables 7 and 8).The cross-price elasticities characterize pairs of goods as substitutes or

complements. On the level of all selected food commodity groups, there areonly substitution relationships and no complementary ones. As a matter of

fact, in Pakistan, many diets are based on a single food with small amountsfrom plant or animal products. They lack dietary diversity. The fact that all

food groups showed a substitution relation4

may be one reason explaining the

lack of diversity in the Pakistanis diet. It is important that a number ofdifferent food sources be consumed and efforts should be made to encourage a

wide variety of foods to improve the nutritional quality of the Pakistanis diet

and health of the population. Dietary diversity is one of the most importantways to ensure a balance of nutrients for people of all ages. However, one

would have expected a complementary relationship for cereal products with

vegetable products, where in Pakistan, cereal products are frequentlyconsumed jointly with vegetables (especially potatoes). This might result from

aggregation decisions of the composite commodities.

5.5. Results by consumption quintiles

The LAAIDS model permits the calculation of elasticities for different

consumption quintiles groups and HIES data materialized this happening. Inorder to do so, income and price elasticities for two extreme quintiles (1 st and

5th

) are estimated. It is obvious from table 5 to 6 and table 9 to 12 that poor

4In order to observe the cross price relationships among the food items, a more detailed

breakup of each food group (up to the individual commodity level) is needed.


21/125


16

people belonging to quintile 1st

exhibit higher income elasticities for fruits,

meats, milk and soft drinks groups as compared to the higher income groups(e.g. households belonging to 5th quintile). In other words, an increase in

income of poor households will lead to higher expenditure on these

commodity groups.

Table: 7. Uncompensated (Marshallian) Price Elasticities5 for Total Sample

Group6

CR PL FR EOF SG MT VG TCS MMP

CR -0.582 0.396 0.363 0.366 0.376 0.529 0.384 0.369 0.419

PL 0.768 -0.238 0.753 0.754 0.757 0.799 0.759 0.755 0.768

FR 0.279 0.295 -0.745 0.316 0.309 0.200 0.303 0.314 0.281

EOF 0.773 0.764 0.751 -0.247 0.757 0.816 0.760 0.754 0.772

SG 0.359 0.343 0.319 0.321 -0.672 0.441 0.334 0.323 0.357

MT 0.001 0.011 0.027 0.025 0.020 -1.053 0.017 0.024 0.001

VG 0.114 0.111 0.108 0.108 0.109 0.127 -0.290 0.108 0.114

TCS 0.179 0.171 0.159 0.160 0.164 0.219 0.167 -0.839 0.178

MMP 0.100 0.111 0.126 0.124 0.120 0.050 0.116 0.123 -0.898

Table: 8. Compensated (Hicksian) Price Elasticities7

for Total Sample

Group CR PL FR EOF SG MT VG TCS MMP

CR -0.502 0.449 0.377 0.385 0.406 0.739 0.423 0.390 0.492

PL 0.897 -0.153 0.777 0.786 0.847 1.097 0.817 0.802 0.891

FR 0.474 0.425 -0.710 0.361 0.382 0.715 0.399 0.366 0.368

EOF 0.894 0.845 0.773 -0.219 0.802 1.135 0.819 0.786 0.888

SG 0.456 0.407 0.336 0.344 -0.695 0.695 0.382 0.349 0.450

MT 0.180 0.131 0.059 0.067 0.088 -0.579 0.105 0.072 0.174

VG 0.165 0.145 0.116 0.120 0.128 0.261 -0.265 0.122 0.162

TCS 0.302 0.253 0.181 0.189 0.210 0.544 0.227 -0.806 0.296

MMP 0.278 0.229 0.157 0.165 0.186 0.519 0.203 0.170 -0.728

5Uncompensated (Marshallian) own-price elasticities are written in bold letters.

6cereals (CR), pulses (PL), fruits (FR), edible oils and fats (EOF), sugar and gur (SG), meats

(MT), vegetables (VG), tea, coffee and soft drinks (TCS), and milk and milk products

(MMP).7

Compensated (Marshallian) own-price elasticities are written in bold letters.


22/125


17

Table: 9. Uncompensated (Marshallian) Price Elasticities8

for 1st

Quintile

Group9


CR -0.694 0.284 0.251 0.254 0.264 0.417 0.272 0.257 0.307

PL 0.761 -0.245 0.746 0.747 0.750 0.792 0.752 0.748 0.761

FR 0.170 0.186 -0.854 0.207 0.200 0.091 0.194 0.205 0.172

EOF 0.653 0.644 0.631 -0.367 0.637 0.696 0.640 0.634 0.652SG 0.302 0.286 0.262 0.264 -0.729 0.384 0.277 0.266 0.300

MT -0.127 -0.117 -0.101 -0.103 -0.108 -1.181 -0.111 -0.104 -0.127

VG 0.550 -0.456 0.535 0.536 0.539 0.581 0.541 0.537 0.550

TCS -0.063 -0.071 -0.083 -0.082 -0.078 -0.023 -0.075 -1.081 -0.064

MMP 0.005 0.016 0.031 0.029 0.025 -0.045 0.021 0.028 -0.993


for 1st

Quintile


CR -0.614 0.337 0.265 0.273 0.294 0.627 0.311 0.278 0.380

PL 0.890 -0.160 0.770 0.779 0.840 1.090 0.810 0.795 0.884

FR 0.365 0.316 -0.819 0.252 0.273 0.606 0.290 0.257 0.259

EOF 0.774 0.725 0.653 -0.339 0.682 1.015 0.699 0.666 0.768

SG 0.399 0.350 0.279 0.287 -0.752 0.638 0.325 0.292 0.393

MT 0.052 0.003 -0.069 -0.061 -0.040 -0.707 -0.023 -0.056 0.046

VG 0.679 -0.371 0.559 0.568 0.629 0.879 0.599 0.584 0.673

TCS 0.060 0.011 -0.061 -0.053 -0.032 0.302 -0.015 -1.048 0.054

MMP 0.183 0.134 0.062 0.070 0.091 0.424 0.108 0.075 -0.823




(MMP).10



23/125


18

Table: 11. Uncompensated (Marshallian) Price Elasticities11

for 5th

Quintile

Group12


CR -0.470 0.508 0.475 0.478 0.488 0.641 0.496 0.481 0.531

PL 0.775 -0.231 0.760 0.761 0.764 0.806 0.766 0.762 0.775

FR 0.388 0.404 -0.636 0.425 0.418 0.309 0.412 0.423 0.390

EOF 0.893 0.884 0.871 -0.127 0.877 0.936 0.880 0.874 0.892SG 0.416 0.400 0.376 0.378 -0.615 0.498 0.391 0.380 0.414

MT 0.129 0.139 0.155 0.153 0.148 -0.925 0.145 0.152 0.129

VG 0.223 0.220 0.217 0.217 0.218 0.236 -0.181 0.217 0.223

TCS 0.421 0.413 0.401 0.402 0.406 0.461 0.409 -0.597 0.420

MMP 0.195 0.206 0.221 0.219 0.215 0.145 0.211 0.218 -0.803


for 5th

Quintile


CR -0.390 0.561 0.489 0.497 0.518 0.851 0.535 0.502 0.604

PL 0.904 -0.146 0.784 0.793 0.854 1.104 0.824 0.809 0.898

FR 0.583 0.534 -0.601 0.470 0.491 0.824 0.508 0.475 0.477

EOF 1.014 0.965 0.893 -0.099 0.922 1.255 0.939 0.906 1.008

SG 0.513 0.464 0.393 0.401 -0.638 0.752 0.439 0.406 0.507

MT 0.308 0.259 0.187 0.195 0.216 -0.451 0.233 0.200 0.302

VG 0.274 0.254 0.225 0.229 0.237 0.370 -0.156 0.231 0.271

TCS 0.544 0.495 0.423 0.431 0.452 0.786 0.469 -0.564 0.538

MMP 0.373 0.324 0.252 0.260 0.281 0.614 0.298 0.265 -0.633

VI. Conclusion and Policy Recommendations

Lack of dietary diversity is a particular problem among the people in

Pakistan, because their diets are predominantly based on starchy staples withlittle animal products and few fresh fruits and vegetables. It is observed that

the major sources of calories and proteins in Pakistan are plant products with

small amounts from animal products as a concentrated source of essentialprotein that are of high quality and highly digestible. In addition, the diets in

Pakistan are low in fat intake, since of all basic foodstuffs, fat is one of the




(MMP).13



24/125


19

most expensive. Therefore, in Pakistan, the consumers are still suffering from

malnutrition and unbalanced essential nutrients like caloric value, proteins,and fat content. Also, there is a marked difference between rural and urban

areas in food consumption patterns.

It is explored that the expenditure and price elasticities for selected

food groups are relatively high in Pakistan. As expected, the estimation resultsshow that expenditure elasticities for all food groups are positive and less than

one, except for fruits, meats, and milk; indicating that the selected food groups

are necessities. For food groups such as fruits, meats, and milk havingexpenditure elasticities larger than unity, identifying them as luxuries, it is

expected that these food groups will experience an increase in demand when

consumers income increases in tandem with the overall economic growth of

the country.

Another interesting finding is that cereals tend to have the lowest

expenditure elasticity of demand. This indicates that cereals have already

occupied a special position in the Pakistans diet, as it is the staple food of the

population. Uncompensated own-price elasticities of demand for all food

groups are negative and consistent with the theoretical expectation. The

absolute amounts of these elasticities for all commodity groups are lower thanunity and so the demand reacts in elastically to own price changes, except for

meats amounting to -1.053 (elastic). The uncompensated own-price elasticities

(in absolute value) for most food groups, such as pulses, oils and fats, and

vegetables than the total expenditure elasticities, implying that food demandreacts more elastically to expenditure changes than to own price changes. The

elasticities are lowest (in absolute value) for vegetables (-0.290), oils & fats (-

0.247), and cereals (-0.582) where demand reacts least to price changes.

For all commodity groups, the compensated own-price elasticities arelower - in absolute terms - than the uncompensated ones, suggesting that a rise

or fall in the price of the respective commodities would have considerable real

expenditure effects. According to the values of cross-price elasticities and onthe level of all selected food commodity groups, only substitution

relationships are observed. Many diets in Pakistan are based on a single of

food with small amounts from vegetables or animal products and lack dietarydiversity in the diet, which supports this result. However, one would have

expected a complementary relationship for cereal products with vegetables,

because in Pakistan, cereal products are frequently consumed jointly with


25/125


20

vegetables (especially potatoes). This might result from aggregation decisions

of the composite commodities.

The findings of the empirical analysis of price and expenditure

(income) elasticities for the selected food groups could be used in the

projections for future food consumption. Pakistan is expected to be getting

farther and farther away from being self-sufficient in its food production. Thisholds true particularly for food items exhibiting high expenditure elasticities

such as livestock products. The high price elasticities of demand for many

food items stress the importance of food price changes for Pakistanihouseholds, and their reactions should be taken into account in the

development of comprehensive agricultural and food policies in order to avoid

unattended effects harming consumers.

Due to the strong influence of diets on health, adequate foodconsumption is an important public health concern. In Pakistan, diets are

traditionally overly rich in calories due to high consumption of cereal products

and comparatively low consumption of healthy food such as fruits andlivestock products. It is important, therefore, that efforts undertaken to

encourage consumption of a wide variety of foods to improve the nutritional

quality of the diet and health of the population. Considering the relatively highexpenditure elasticities of demand for fruits and livestock products of all

households, income increases would exert a positive influence on the intake of

micronutrients that are delivered by fruits and livestock products. The results

of this study suggest that income oriented policies are important to achievebetter nutrition and reduce the problem of unbalanced diets in Pakistan. In

addition, complementing policies are necessary.

Since Pakistan has a high income inequality, it is expected that income

and price-elasticities are different between the richest and the poorest. Theresults supported this expectation, indicating that income-elasticities are

higher for the poorest for all staple food. Moreover, own-price elasticities are

higher for the poorest households in the case of cereals and pulses, the mostconsumed staple food commodities in Pakistan. These results are an important

step forward in understanding household consumption habits in Pakistan, and

highlight the consumption differences between poor and rich in the country.The elasticities calculated in this study are powerful instruments in helping

policymakers in devising policies targeted at poor people.

Food subsidies can be better targeted to the poor people by subsidizing

food items and distributing in villages and rural neighborhoods where the poor


26/125


21

are known to be concentrated. The total annual food subsidy resources could

be allocated to each region according to its contribution to total poverty. Thesubsidy system should re-establish subsidies on some of the healthy foods like

red meat and fish because these items are a relatively concentrated source of

essential protein of high quality and highly digestible. The best way for

Pakistan to improve its food distribution system is that the food subsidysystem should be changed from the commodities form to a cash subsidy

provided only to low-income households and reduces the benefits to the non-needy.

Increase in animal production must be focused, particularly small

ruminants and fisheries, aiming at increasing the per capita consumption of

animal protein in its various forms by means of raising productivity ofdomestic cattle of buffalo, cow and sheep using improved genetic techniques;

and by introducing high-yield genetics as a means to increase milking rate,

meats and eggs production. Increasing the quantities of animal products isexpected to have an effect on the prices as a whole and as a result may benefit

consumers. Decrease per capita consumption of cereals through redistributionof flour uses, raising the standard of living of the population and changing

food consumption patterns.

It is important that a number of different food sources be consumed

and efforts should be made to encourage a wide variety of foods to improve

the nutritional quality of the Pakistanis diet and health of the population.

Dietary diversity is one of the most important ways to ensure a balance ofnutrients for people of all ages. The results of this study suggest that income

oriented policies are important to achieve better nutrition and reduce the

problem of unbalanced diets in Pakistan.


27/125


22

References

Abdulai, A. (2002). Household demand for food in Switzerland: A quadraticalmost Ideal demand system. Swiss Journal of Economics andStatistics, Vol. 138(1), pp. 1-18.

Abdulai, A. (2003). Economies of scale and the demand for food in

Switzerland: Parametric and nonparametric analysis. Journal of

Agricultural Economics Society, Vol. 54(2), pp. 247-267.

Abdulai, A., & D. Aubert (2004). A cross-section analysis of household

demand for food and nutrients in Tanzania. Journal of Agricultural

Economics, Vol. 31, pp.1-13.Aziz, Babar (1997). Analysis of consumer demand systems using time series

data of Pakistan, M. Phil. Thesis submitted to Quaid-i-Azam

University, Islamabad.Aziz, Babar (2004). Demand for meat and structural changes in Pakistan: An

econometric analysis.Journal of Social Sciences and Humanities, Vol.

2(2), pp. 55-80.Aziz, Babar (2009). Analysis of consumer behaviours: Some practical

insinuations for Pakistan, PhD. Thesis submitted to Bahauddin

Zakariya University, Multan.Aziz, Babar & Shahnawaz, Malik (2010). Household consumption pattern in

Pakistan: A rural-urban analysis. Forman Journal of Economic

Studies, Vol. 6, pp. 1-26.

Barten, A. P. (1964). Consumer demand functions under conditions of almostadditive preferences.Econometrica, Vol. 32, pp. 1-38.

Barten, A. P. (1977). The system of consumer demand functions approach: A

review.Econometrica, Vol. 45, pp. 23-51.Brown, A. & A. Deaton (1972). Models of consumer behavior: A survey.

Economic Journal, Vol. 82, pp. 1145-1236.

Burki, Abid A. (1997). Estimating consumer preferences for food using timeseries data of Pakistan. Pakistan Development Review, Vol. 36(2), pp.

131-153.

Buse, Adolf (1994). Evaluating the linearized almost ideal demand system.

American Journal of Agricultural Economics, Vol. 76, pp. 781-793.Chaudhary, M.A., Eatzaz A., Abid A. Burki & Mushtaq A. Khan (1999).

Income and Price Elasticities of Agricultural, Industrial and Energy

products by sector and income groups for Pakistan, QUEC ResearchReport for Planning Commission, Government of Pakistan.


28/125


23

Deaton, A. (1986). Quality, Quantity and Spatial Variation of Price, Princeton

University, Princeton, NJ, processed.Deaton, A. (1987). Estimation of own-price and cross-price elasticities from

survey data.Journal of Econometrics, Vol. 36, pp. 7-30.

Deaton, A. (1988). Quality, quantity, and spatial variation of price. American

Economic Review, Vol. 78, pp. 418-430.Deaton, A. & J. Muellbauer (1980a). An almost ideal demand system

Economic Review, Vol. 70, pp. 312-326.Deaton, A. & J. Muellbauer (1980b). Economics and Consumer Behavior,

Cambridge University Press, Cambridge.

Green, R., & Julian M. Alston (1990). Elasticities in AIDS model.American

Journal of Agricultural Economics, Vol. 72, pp. 442-445.

Green, R. & Julian M. Alston (1991). Elasticities in AIDS models: A

clarification and extension. American Journal of Agriculture

Economics, Vol. 73, pp. 874-875.Hassan, R. N. & Babu, S. C. (1991). Measurement and determinants of rural

poverty: household consumption patterns and food poverty in ruralSudan. Food Policy, Vol. 16(6), pp. 451-460.Houthakker, H. S. (1985). Richard Stone and the Analysis of Consumer

Demand, Discussion Paper No. 1140, March 1985, Harvard Institute

of Economic Research, Harvard University, Cambridge,Massachusetts.

Malik, Shahnawaz & Babar Aziz (2006). Surmising consumer demand system

and structural changes using time series data for Pakistan. Pakistan

Economic and Social Review, Vol. 44(1), pp. 117-136.Moschini, G. (1998). The Semi-flexible almost ideal demand system.

European Economic Review, Vol. 42, pp. 349-364.

Pakistan, Government of (2005). Household Integrated Economic Surveys,Electronic data sets obtained from Federal Bureau of Statistics,

Statistics Division, Government of Pakistan.

Parks, R.W. (1969). System of demand equations: An empirical comparisonof alternative functional forms.Econometrica, Vol. 37. pp. 629-650.

Philips, L. (1983). Applied Consumption Analysis, North-Holland Publishing

Company, New York.Prais, S. J. & H. S. Houthakker (1955). The Analysis of Family Budgets,

Cambridge University Press, Cambridge.

Prais, S. J. & H. S. Houthakker (1971). The Analysis of Family Budgets,

Second Impression Abridged 1971, Cambridge University Press,Cambridge.


29/125


24

Raunikar, R., & C. H. L. Huang (1987). Food Demand Analysis, Problems,

Issues, and Empirical Evidence, Iowa University Press, Ames/Iowa.Schultz, T.W. (1957). The Theory and Measurement of Demand, University of

Chicago Press, Chicago.

Varian, Hal R. (1992). Microeconomic Analysis, 3rd

ed. W. W. Norton and

Company, New York.Yoshihara, K. (1969). Demand functions: An application to the Japanese

expenditure pattern.Econometrica, Vol. 37, pp. 257-274.


30/125


Vol. 7, 2011 (JanuaryDecember) pp. 25-54

Exchange Rate Exposure on the Automotive Industry:

Evidence from USA and Japan

Zeresh Mall, Saqib Jafarey, Shabib Haider Syed and Ijaz Hussain1

Abstract

This study analyses the impact of exchange rate shocks on firm value as well

as on the portfolio of automotive firms from U.S and Japan over a time period

of 1999-2007. The effect of intra industry competition on the relation between

exchange rate and firm value is also incorporated. The results indicate that

Japanese firms are more exposed to the dollar than U.S firms to yen and the

exposure to yen and dollar for the U.S and Japanese firms respectively is due

to the market share of Japanese firms in the U.S while the exposure to euro

for the Japanese firms is due to the market share of German firms in Japan as

well as Japanese firms in Germany.

Keywords: Exchange rate exposure; automotive industry; USA & Japan

JEL classification: F31, G30, G39

I. Introduction

Financial theory predicts that a change in an exchange rate shouldaffect the value of a firm or an industry. According to Eiteman et al. (1995)and Shapiro (1992), the exchange rate exposure is conventionally classified astransaction exposure and economic exposure. Transaction exposure is theeffect of exchange rate changes on committed cash flows such as accountsreceivables and is short term in nature. Economic exposure is the effect that

exchange rate changes have on a firms long-term cash flows and is long termin nature. Chow et al (1997) provide evidence that transaction exposure,economic exposure and the interest rates changes associated with exchangerate changes work together to determine the exchange rate exposure of stockreturns. A firm is subject to economic exposure if the firms value, asmeasured by the present value of its expected future cash flows, is sensitive to

1 The authors are Lecturer at Forman Christian College (A Chartered University) Lahore,Professor of Economics at City University London, Associate Professor at Forman ChristianCollege (A Chartered University) Lahore and Assistant Professor at Gomal University D. I.

Khan, respectively.


31/125

Zeresh, Saqib, Shabib and Ijaz

26

changes in exchange rates. For example, the value of an exporting firm islikely to fall if the domestic currency appreciates, while the value of animporting firm is likely to rise with that same appreciation. A change inexchange rate through its effect on the costs of inputs, outputs, and substitutegoods affects the competitive position of domestic companies with no directinternational involvement relative to foreign corporations. Exchange ratemovements can affect an individual investor who owns a portfolio consistingof securities in different currencies, multinational company (MNC) withsubsidiaries and branches in foreign locations, an exporter/importer whoconcentrates on international trade and even a firm that has no directinternational activities.

It is assumed that the automotive industry is competitive and thatcompetition acts as a proxy for the elasticity of demand for a product thereforecompetition that a firm faces in the domestic and foreign markets should be adeterminant of a firms exposure in that specific market. In this study the

impact of competition will be measured by examining the market share offirms in U.S of Japan and Germany, in Japan the market shares of firms ofU.S and Germany and in Germany the market shares of U.S and Japan toanalyse the impact of international sales on exposure. This is consistent withthe notion that the currency exposure of a firm is a function of the export salesand the competition faced in a specific market. Williamson (2001) resultsshow that domestic competition from foreign firms plays a vital role indetermining exposure.

This study will be investigating the effect of real exchange ratechanges on the value of firms in the automotive industry and the impact ofcompetition on exposure. Automobile manufacturers are often multinationals

as they have subsidiaries and manufacturing plants in many differentcountries. Automobile industry has strong international dependence for bothproduction inputs and exports of finished products and is likely to be sensitiveto foreign exchange rates. The automotive industry has progressed from oneof national competition, particularly in North America, to one of internationalcompetition, in which firms from the U.S. Japan etc export to foreign markets,to one of global competition, in which firms produce and sell in manycountries.


32/125

Exchange Rate Exposure on the Automotive Industry: USA and Japan

27

1.2. Exchange rate exposure and firm value

A company which manages its productions or delivers its services inmore than one country with export sales and costs in home currency shoulddepict exchange rate exposure. The exchange rate exposure tends to change

with the competition that the industry undergoes and also with the foreigncurrency position of the firms operations. If there is no competitor and theexporter have costs in the local currency while selling in a foreign market thecash flows will be affected by changes in exchange rates. The sensitivity of afirms cash flows to exchange rate changes is mainly a function of elasticity ofdemand for the firms product. If a firm has low elasticity of demand for thefirms product but high export sales it will face a low exposure as it canincrease prices in the local market when faced with depreciation of the localcurrency and this in turn lessens the impact on the home currency cash flows.

As the number of local competitors increases in the foreign market thesensitivity of the firms cash flows to exchange rates should increase. With theintroduction of competitors the ability to increase prices if the local currencydepreciates will be affected. Conclusively with the introduction of competitorsthe sensitivity of the firms cash flows to exchange rates will also increase.

Global industries undergo structural changes. This type of industryevolution has important implications for exposures for multinational firms andglobal competitors because it has a severe impact on their competitivemakeup. If there is a firm that exports to a foreign market and does notcompete directly with firms in that market, the firm's exposure will only be afunction of its foreign currency revenues because the firm may have nationalcompetition but little or no competition from foreign markets. If the foreign

firm then faces competition in the local market, exposure becomes a functionnot only of its foreign currency revenues but also of the elasticity of its ownand its competitor's product. The firm then becomes more competitiveworldwide and therefore have to be more concerned with local competitors ina foreign market or foreign competitors in the firm's domestic market. Thecomplexity of a firm's exchange rate exposure evolves as the industrybecomes more global, that is when firms begin to produce in various markets.Firms can be exposed to exchange-rate risk through various channels. Forinstance a firm with foreign sales is exposed to exchange-rate risk because thevalue of foreign sales in terms of domestic currency changes when the


33/125


28

exchange rate changes. The same firm may source inputs from abroad and thismay increase or decrease its exchange rate exposure depending on whether theimports and exports are in the same currency. Furthermore, this firm may alsohave assets and liabilities abroad; this can also increase or decrease a firmsexposure. Exchange-rate exposure is not limited to exporters, importers ormultinational firms. Even a domestic firm with no foreign activities may beexposed to exchange-rate risk, for example a local firm facing importcompetition.

This paper is structured in the following manner. Section reviewsthe literature available on exchange rate exposure. Section presents themodel specification and data set. Section IV describes the estimationprocedure and presents empirical results. The last section provides theconclusion and policy implications

II. Literature Review

The theoretical exchange rate exposure literature supports the commonbelief that exchange rate changes should impact firms that import fromforeign markets, export to foreign markets, or face foreign competition.Shapiro (1975) argues that a multinational firm with export sales andcompetition should exhibit exchange rate exposure and that the firmsexposure should be related to the proportion of export sales, the level offoreign competition, and the degree of substitutability between local andimported factors of production. Second Generation studies in contrast to FirstGeneration studies document exchange rate exposure as being significant. Toestimate the effect of an exchange rate shock on firm value, only those shocksshould be identified that are permanent and unanticipated.

2.1. First generation

For a set of US firms, Jorion (1990) shows insignificant exchange-rateexposure. Over the period from 1971 to 1987 only 15 out of 287internationally operating firms or 5.2% of the sample have a significantexchange rate exposure at the 5 percent significance level. This occurredbecause firms effectively managed their exposure which is quite similar to thestudy of Bodnar and Gentry (1993). Their study tested for exchange rateexposure at the industry level in the US, Japan and Canada and foundinsignificant exposure for countries which are less open and small. Bartov andBodnar (1994) provide an additional justification for finding insignificantexchange risk exposure. They suggest that firms that can respond to exchange


34/125


29

rate changes and overall international market conditions at low cost will tendto have insignificant exchange risk exposure. Choi and Prasad (1995) examineexchange-rate exposure of a sample of 409 multinational firms that haveforeign sales, profits and assets of at least 25 percent of their respective totals.

When they examined exchange risk exposures at the industry level bygrouping the firms into 20 portfolios they found limited support for theimportance of the exchange rate factor. This was explained by the fact thatalthough firms in a given industry are in the same primary line of business,they are still heterogeneous in terms of their operational and financialcharacteristics. Since industry groups include firms with positive and negativeexchange risk exposure, aggregating across such firms will result in finding aninsignificant exposure coefficient for the industry group.

2.2. Second generation

Some studies demonstrate that exchange rate movements can have aneconomically significant impact on firm value. A firm is said to exhibitexchange rate exposure if its share value is influenced by changes in currencyvalues. Priestley and Odegaard (2002) study uses data from the Norwegianequity market to investigate currency exposure. The Norwegian market isparticularly well suited for such an investigation as it is an open economy andtheir results provide comprehensive evidence that exchange rate exposure isstatistically significant and economically important. This study analyzes thecurrency exposure of industry stock returns. They show that when measuringcurrency exposure in regressions including the local stock market one has toaccount for the currency exposure of the local market itself in the estimates,account for possible regime changes by the monetary authorities in exposureestimations and use individual currencies of the major trading partners insteadof a currency basket. When these issues are accounted for, exposure estimatesare important in both an economic and statistical sense. Ligterink and Macrae(2006) examine the relationship between exchange-rate changes and stockreturns for a sample of Dutch firms over 19941998. They find that over 50percent of the firms are significantly exposed to exchange-rate risk. Withrespect to the determinants of exposure, they find total assets and the foreignsales ratio to be significantly and positively related to the firms exchange-rateexposure. In comparison with other economies in the world, the Netherlandshas a relatively open economy, which may be an explanation for the


35/125


30

prominent exchange-rate exposures for the Dutch firms. Williamson (2001)findings show that there is significant exchange rate exposure in theautomotive industry. He finds evidence supportive of the theoreticaldeterminants of foreign exchange rate exposures for firms in a globallycompetitive industry. His tests reveal that the ratio of foreign sales to totalsales and competition are major determinants of exchange rate exposure.Therefore firms in the automotive industry show a significant amount ofexchange rate exposure. Dominguez and Tesar (2006), Doidge, Griffin, andWilliamson (2006), Bartram and Bodnar(2007) Priestley and Odegaard (2007)findings are also consistent with the result that exchange rate exposure issignificant.

III. Methodology

3.1. Sample selection

This study incorporates information regarding automotive firms

headquartered in U.S.A and Japan. The automotive firms of both countriescompete in major markets and manage a large percentage of worldwidecontestable cash flows linked with the automotive industry. Contestable cashflows involve no significant barriers to foreign competition. JapaneseAutomotive industry is selected because it is one of the leading and prominentindustries of the world. The sample comprises of the big three of U.S.A andsix companies of Japan namely General Motors2, Ford Motor, Chrysler LLC3,

2 General Motors Corporation is an American automaker based in Detroit, U.S. For 77

consecutive years (1931- 2007), GM was the global sales leader. On June 1, 2009 GeneralMotors Corporation filed for bankruptcy under chapter 11 of the Bankruptcy Code. It is thethird largest bankruptcy filing of the world and the former General Motors Corporation is nowknown as Motors Liquidation Company. Therefore, the stock price taken to compute stockreturns is of Motors Liquidation Company.3 Chrysler Group, LLC is an American automobile manufacturer headquartered in Detroit. In1998 German based automobile manufacturer Daimler acquired Chrysler. From 1998 to 2007Chrysler was part of the German based Daimler Chrysler. On May 14, 2007, DaimlerChrysler announced the sale of 80.1% of Chrysler Group to an American private equity firm.Therefore to compute the stock returns for Chrysler I have used the stock price of Daimler asa proxy for Chrysler. The time period for this study is eight years. As Chrysler is a significantautomaker in U.S.A. I included it in my study; excluding it from the sample would haveaffected the results. Therefore, I had to truncate the time of the study to take account of

Chryslers sale.


36/125


31

Toyota Motor, Nissan Motor, Honda Motor, Suzuki Motor, Mitsubishi Motorand Subaru Motor respectively.

3.2. Data set

The information on exchange rates and market return is taken fromDataStream International. The data on stock returns, market return andexchange rates are with a monthly frequency. The data on stock price for U.Sfirms has been taken from Bloomberg while the data for stock price ofJapanese firms is from DataStream International. The stock return iscalculated by taking the logarithmic difference between stock prices of thecurrent and previous month.

Stock return = ( 1ln lnt tP P ) this is because returns are typically taken as

percentual return:

1 1 1% 100( / ) 100( ) / ~(ln )t t t t return X X X X X X

For the second regression equation the impact of sales in each marketas shown in figure 1, 2 and 3 depict market shares in each country. The salesfigures are taken for a period of eight years (1999-2007). The information forthe market shares is taken from the Japanese Automobile ManufacturersAssociation also known as JAMA and auto insight data by accessing GlobalInsight. To analyze the impact of competition on exposure market shares offirms in U.S.A, Japan and Germany are examined. The firms in Germany'ssample include Porsche, B.M.W and Daimler. Market share is evaluated bycalculating the portion of a firms sales to total sales in the particular countryand the figures are in percentage.

The firms stock return is used as a proxy for changes in firm value.To compute the stock returns I have taken the logarithmic difference betweenstock prices. A firm's stock price measures the value of its expected futurecash flows. These expected cash flows can follow many different patterns, andthe patterns can vary dramatically from firm to firm. There is a high positivecorrelation between low (high) firm value and low (high) stock price. Firmsthat have performed poorly (well) recently tend to have lower (higher) shareprices due to this poor performance.

Previous studies have been using the trade weighted exchange rate tomeasure the extent of a firms exposure to exchange rates but this can mute


37/125


32

the effect of an exchange rate shock on firm value. Therefore in this study toinvestigate a firms or a countrys automotive industry exposure individualcurrencies are used as well as real exchange rates. The real exchange rate canbe defined as the nominal exchange rate that takes the inflation differentialsamong the countries into account. Its importance stems from the fact that itcan be used as an indicator of competitiveness in the foreign trade of acountry. When there are nominal assets and liabilities present in the foreigncurrency a firm can be exposed to the nominal exchange rate because theseassets and liabilities should be interpreted at the nominal rate. In the absenceof foreign assets or liabilities a nominal rate change which is affected bychanging price levels across countries should have no effect on the real valueof the firm. Conclusively, the exchange rate change that should affect firmvalue is the real exchange rate.

The work of Dumas (1978) and Adler and Dumas (1980, 1984)suggest that exchange rate exposure can be quantified as the sensitivity of

stock returns to exchange rate movements. According to Adler and Dumas(1984) exchange rate exposure is the influence of exchange-rate changes onthe future cash flows of the firm. In their view firm value represents thepresent value of future cash flows and exchange-rate exposure is thesensitivity of firm value to exchange-rate changes. Under this assumption,exposure can be determined from the elasticity of firm value with respect tothe exchange rate. He defines the exposure elasticity as the change in themarket value of the firm resulting from a unit change in the exchange rate.With this approach the exposure elasticity of the firm can be obtained fromthe coefficient on the exchange rate variable in the following regression.Following Adler and Dumas empirical studies which have measured

exchange-rate exposure by the slope coefficient from a regression of stockreturns on exchange-rate changes we will be able to evaluate the effect ofexchange rate on firm value. To estimate the effect of exchange rate on thefirm value the following regression will be used.

m e

t mt t t r R S (1)

where rt is the monthly return of a firm and portfolio, is the intercept, St isthe change in real exchange rate, e measures the exposure of the country-specific industry portfolio, Rmt is the return on the country-specific marketportfolio, m is the market risk of firms and t is the error term. The exchange


38/125


33

rate exposure is tested for the specific firms as well as for the portfolio of bothJapanese and American automotive manufacturers.

The equation although very similar to Adler and Dumas (1984) isconsistent with Jorion (1990) as it includes the market factor. Market factor is

added to prevent misspecification of the model and control formacroeconomic factors. It is seen to be a significant component of the returnsgenerating process. This market portfolio addition controls for market-widefactors that represent macroeconomic effects correlated with the exchange rateand it changes the statistical properties and distribution of the exposureestimates. Because the market return explains a substantial amount of thetypical firms stock return variation, its inclusion in the exposure estimationmodel reduces the residual variance of the regression and improves theaccuracy of the exposure estimates. The market return has the additionalfeature of explicitly controlling for movements in the stock market.

It is widely accepted that, for some industries, competition betweencountries is economically important and this competition is strongly affectedby exchange rate changes. Economists around the world argue that some ofthe industries in their countries compete vigorously with the same industriesin other countries and that exchange rate shocks affect their competitiveness.In the U.S. it is routinely stated that some U.S. industries compete withJapanese industries and that an appreciation of the yen is good for these U.S.industries and bad for the competing Japanese industries. A firms exchangerate exposure is a function of foreign sales, the elasticity of demand in theforeign market and the elasticity of demand in the domestic market (Marston,1996) It is assumed that the automotive industry is competitive and thatcompetition acts as a proxy for the elasticity of demand for a product thereforecompetition that a firm faces in the domestic and foreign markets should be adeterminant of a firms exposure in that specific market. Therefore, a firm hasmore significant exposure to a particular currency not only if the firm hassubstantial sales in the foreign market but also if the firm faces competition inthe same market. This also holds for the domestic market. If the firm facescompetition from foreign firms, then the firm has exposure to the currency ofthat competitor. To evaluate the impact of competition on exchange rateexposure for U.S firms the market share of Japanese and German firms in USis analysed as well as the market share of U.S firms in Japan and Germany


39/125


34

and similarly for the Japanese firms. Germany is included as a third countryto take into account the export sales and competition encountered in aparticular market.To test the exposure of a firm to competition in the homeand foreign markets the following regression is used as follows:

1 ,t , 2 ,t , 1 , , 2 , ,

m

t mt jp us us jp dm t gr us dm t us gr t r R y S MS y S MS S MS S MS

(2)Where m is the market risk, Rmt is the return on the countrys market, 1 and1 are the exposure of the interaction between the exchange rate and portfoliomarket share in which MSA,B represents the market share of portfolio ofcountry A in country , Sk,t represents the rate of change of the real exchangerate in currency k at time t and rt is the monthly return.

IV. Empirical Results

4.1. USA firms

Table 1a shows the results for the USA portfolio and US firm-specific

exchange rate exposure. A negative exchange rate coefficient corresponds to adecrease in the firms stock returns when the home currency appreciates (aswould be the case for an exporter).The US portfolio shows an insignificantexposure to yen and euro along with a negative sign indicating that the U.Sportfolio loses values as yen and euro depreciate relative to the dollar. At thefirm specific level it can be seen that Ford loses value as yen and eurodepreciate relative to dollar, Chrysler loses (gains) value as yen (euro)depreciates and General Motors loses (gains) value as euro (yen) depreciaterelative to dollar. The results of the portfolio for yen/dollar are driven byChrysler and Ford as they both carry a negative sign while General motorsand Ford show a negative sign for the euro and therefore the negative sign ofthe portfolio for euro/dollar is driven through these two.

Table: 1a. USA portfolio and firm - specific exchange rate exposure

Firm Intercept market risk yen/dollar euro/dollar Adj.R (%)U.S. portfolio -0.0128 0.7375 -0.3191 -0.0128 14.8916

[-1.6783] [4.1255]*** [-1.0468] [-0.0418]G.M -0.0087 1.1482 0.3626 -0.2002 23.2602

[-0.9466] [5.3110]** [0.9832] [-0.5406]Ford Motor -0.0234 0.7961 -0.1943 -0.6380 8.2814

[-1.8881]* [2.7321]** [-0.3893] [-1.2839]Chrysler -0.0062 0.2703 -0.6829 0.3560 1.0961

[-0.6350] [1.1748] [-1.7401]* [0.9033]

*, **, ***denotes 10%, 5% and 1% significance level.


40/125


35

The full sample results of General Motors, Ford and Chrysler reveal that theyhave insignificant exposure to both yen and euro resulting in an insignificantexposure of the portfolio. Each firm and portfolio has been tested for thestandard tests of autocorrelation, multicollinearity, misspecification and

heteroskedasticity problems and there was none present in any firm. Theadjusted R is in percentage; the highest value for General Motors and thelowest value for Chrysler Motors.

4.2. Japanese firms

Table 1b shows the results for the Japanese portfolio and firm-specificexchange rate exposure. The Japanese portfolio shows a negative andsignificant exposure for dollar/yen while an insignificant and positiveexposure for euro/yen. This reveals that the Japanese portfolio loses value asdollar depreciates relative to yen and gains in value as euro depreciatesrelative to yen.

Table: 1b. Japanese portfolio and firm - specific exchange rate exposure

Firm Intercept market risk dollar/yen euro/yen Adj.R(%)

JapanPortfolio

0.0013 0.6439 -0.4272 0.2118 0.3389

T- statistic [0.2621] [6.9207] [-2.4926]** [1.2302]

Toyota 0.0067 0.7371 -0.6481 0.0070 31.880

T- statistic [1.0731] [6.3463]** [-3.0294]** [0.0328]

Nissan 0.0088 0.6420 -0.3437 0.1512 11.6529

T- statistic [0.9849] [3.8567]** [-1.1208] [0.4910]

Honda 0.0050 0.5793 -0.9650 0.0195 28.4657

T- statistic [3.6276]** [4.8831]** [4.4225]** [-0.9562]Suzuki 0.0046 0.7120 -0.6504 0.4230 25.3624

T- statistic [0.6561] [5.3897]** [-2.6728]** [1.7306]*

Mitsubishi -0.0114 1.0918 -0.1188 0.1490 11.5016

T- statistic [-0.7628] [3.9132]** [-0.2312] [0.2887]

Subaru -0.0059 0.093806 0.16895 0.5189 -0.2247

T- statistic [-0.6296] [0.5344] [0.5226] [1.5979]

*, **, ***denotes 10%, 5% and 1% significance level.


41/125


36

At the firm-specific level for dollar/yen Toyota, Honda, Suzuki, Mitsubishi,Nissan all have negative sensitivity to the dollar depicting that as dollardepreciates there is a loss in their firm value. The sign of the sensitivity isconsistent with the previous findings. Sensitivity to the dollar is drivenprimarily by Honda, Toyota and Suzuki as they have a higher percentage ofcars in the U.S compared to others and from the table it can be seen that thesethree firms have negative as well as significant exposures; therefore it can besaid that the results for the Japanese portfolio for dollar/yen is driven byToyota, Honda and Suzuki. Subaru shows a positive and insignificantexposure along with Mitsubishi and Nissan. At the firm specific level for euroall the firms in the portfolio have insignificant and positive exposure. The signof the coefficient discloses that all the firms along with the portfolio enhancetheir value as euro depreciates relative to dollar. Standard tests forautocorrelation, multicollinearity, misspecification and heteroskedasticitywere run to check for any of the problems present. The adjusted R is inpercentage; the highest value for the Japanese portfolio and the lowest valuefor Subaru.

4.3. Market shares of Japan, U.S and Germany

Before analysing the impact of competition on exchange rate exposuremarket shares should be taken into consideration. Figures 2a, 2b and 2c showthat the U.S automotive market has a higher market share of Japanese firmscompared to the number of U.S firms selling in Japan while Japan has a lower

Figure: 3a. USA Automobile Market Share


42/125

Exchange Rate Exposure on the Automotive Industry: USA and Jap

role adv agri tech

Documents