IMPACTS OF GLOBLISATION AND TRADE LIBERALISATION ON SMALL HOUSEHOLDS IN VIETNAM LIVESTOCK SECTOR Pham Thi Ngoc Linh A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy SCHOOL OF AGRICULTURAL AND RESOURCE ECONOMICS Faculty of Natural and Agricultural Sciences May 2011
230
Embed
IMPACTS OF GLOBLISATION AND TRADE LIBERALISATION ON … · Thang, Nguyen Thi Bich Ngoc, Dang Lan Huong, Phan Nguyen Trung Khanh, Tran Thanh Nhan, Bui Thu Ha, Tran Doc Lap, Nguyen
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
IMPACTS OF GLOBLISATION AND TRADE
LIBERALISATION ON SMALL HOUSEHOLDS IN
VIETNAM LIVESTOCK SECTOR
Pham Thi Ngoc Linh
A thesis submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
SCHOOL OF AGRICULTURAL AND RESOURCE ECONOMICS
Faculty of Natural and Agricultural Sciences
May 2011
ii
To my loved husband and son: Duc and Thanh
To my parents
iii
CERTIFICATION
The thesis is my own composition, all sources have been acknowledged and my
contribution is clearly identified in the thesis.
The thesis has been substantially completed during the course of enrolment in this
degree at UWA and has not previously been accepted for a degree at this or another
institution.
This thesis does not contain work that I have published, nor work under review for
publication.
Pham Thi Ngoc Linh
May 2011
iv
ACKNOWLEDGEMENTS
This thesis was completed with the valuable support of many people, to whom I am
indebted in one form or the other. However, all remaining shortcomings and mistakes in
this thesis are my own responsibility.
I would like to express my sincere thanks to Australian Centre for International
Agricultural Research (ACIAR) for granting me a John Allwright Scholarship to study
and complete my PhD at the University of Western Australia.
I would like to acknowledge my gratefulness to my principal supervisor, A. Prof. Dr.
Michael Burton, who has supported me in terms of knowledge, experience, material and
worthwhile comment. His tireless support has encouraged me to overcome difficulties
during my study. I remember his help in solving data problems and revising my papers
and thesis.
I take pleasure in offering my sincere appreciation to my co-supervisor, Dr. Greg
Hertzler, for his guidance and assistance. His critical comments and helpful suggestions
from the early stages of my study are acknowledged.
During the time of doing this thesis, I have also had external expert advice and
numerous supports from Dr. David Vanzetti from Crawford School of Economics and
Government, Australian National University. His excellent experience in dealing with
trade modelling helped me considerably in constructing and finalizing the trade part of
my thesis.
I whole-heartedly thank my sister, the best friend, Dr. Donna Brennan for gave me the
excellent opportunity of studying at UWA, and also for her continuous help and support
through my study time. Her experience and guidance helped me a lot, especially when
my supervisors were absent. Also, her sympathies and sharing both joy and sorrow with
me in the early stage of living far away from family was invaluable.
I would like to thank to School of Agricultural and Resource Economics, University of
Western Australia, for providing me the necessary facilities, excellent support and ideal
environment for undertaking research study. My thanks go to A. Prof. Dr. Ben White,
Sally Marsh, Helena Clayton, Jan Taylor and all teachers, staff and other colleges in
ARE, ACIAR scholarship officer Sharon Harvey, and AusAID officers at UWA:
v
Rhonda Haskell, Cathy Tang, and Christine Kerin who have been willing to help and
assisted me during my study at UWA. My thanks go to Dr. Christine Davies for her
English editing for the final thesis.
I wish to thank Dr. Dang Kim Son, my boss for both sending and encouraging me to
study. I sincerely appreciate his support in both my work and study time. My thanks
also go to my colleges: Nguyen Ngoc Que, Nguyen Do Anh Tuan, Tran Thi Quynh Chi,
Nguyen Le Hoa, Truong Thu Trang and others, for their continuous support and for
doing my work during my candidature.
I could not finish my hard study time without the support of my friends, Tran Cong
Thang, Nguyen Thi Bich Ngoc, Dang Lan Huong, Phan Nguyen Trung Khanh, Tran
Thanh Nhan, Bui Thu Ha, Tran Doc Lap, Nguyen Van Liem, Vu Le Tu, Pham Lan
Huong, Nguyen Thanh Hai and others. I would like to thank them for their friendship,
help and sharing, both physically and mentally, from small to big daily issues of life
during my time in Perth.
Especially, I would like to thank my family, my parents, and my mother in law, who
supported me by taking care my son and family, and encouraging me by giving me
comfortable conditions to study and finish my PhD. Thanks go to my dad and mum,
who always trust in and encourage me to follow up a research career. Last but not least,
I am indebted to my beloved husband Dinh Ngoc Duc, and my adorable son Dinh Duc
Thanh, who are always behind me, for their endless belief in me, for sharing with me all
sweet and hard things in those days. Without their boundless love, trust and
encouragement, I would not have made it this far.
Perth, May 2011
vi
ABSTRACT
Vietnam has negotiated a series of bilateral and multilateral trade agreements and has
made significant steps in integrating into the world economy. This integration is likely
to have both positive and negative effects on different stakeholders in the economy.
This study assesses the impacts of trade liberalisation on Vietnam’s small livestock
households on welfare, production and consumption behaviours, using a multi country
general equilibrium model (GTAP) at the macro level, linked with four, regionally
representative household models at the micro level.
A GTAP utility SplitCom is used to separate live pig and live poultry out of the general
livestock group prior to running several trade liberalisation scenarios. The household
model is linked to the trade model through changes in the prices of inputs and outputs
arising from different trade scenarios. Inside the household model, a recursive
household model is used, with a two-stage LES-AIDS model on the consumption side
and Cobb-Douglas functions on the production side.
The results of the simulations shows that Vietnam’s small households in the livestock
sector would benefit from almost all trade liberalisation scenarios considered here. The
largest benefit is generated by full liberalisation: by comparison there is almost no gain
from bilateral trade liberalisation with the USA. The voluntary trade liberalisation of
Vietnam in a unilateral liberalisation would generate some benefit for the country at the
national level without the need for negotiating with others, and also increase the welfare
of the households.
The changes in simulated prices lead to changes in the household’s welfare as a result of
changes in behaviour, including consumption, production (and hence profit), and also in
the time allocation decision of the household. In almost all trade liberalisation scenarios,
the decision to take more leisure as part of the utility maximising response contributes
about one quarter of the total increase in household’s welfare.
Assumptions that are made about how the labour market in structured, in both the
macro- and micro- models affect the impact on labour wages, and hence drive responses
of the household in the trade scenario simulations. The thesis reports the sensitivity of
the key findings of the research to these critical assumptions.
FAO Food and Agriculture Organisation of the United Nations
FAOStat FAO Statistics
FMD Foods Mouths Diseases
FSH Fishing
GATT General Agreement on Tariffs and Trade
GDP Gross Domestic Product
GEL General Exclusion List
GSO General Statistical Office
GTAP Global Trade Analysis Project
HS Harmonize System
IDN Indonesia
IFPRI International Food Policy Research Institute
IL Inclusion List
IND India
JPN Japan
KOR Korea
LA-AIDS Linear Approximation Almost Ideal Demand System
LAM Latin America
LDCs Least Developed Countries
LES Linear Expenditure System
MAN Manufactures
MARD Ministry of Agriculture and Rural Development
Mill. Million
MLK Milk and dairy products
MRD Mekong River Delta
MYS Malaysia
NCC North Central Coast
xvii
NE North East
NES North East South
NL Normal List
NTBs Non-Tariff Barriers
NW North West
OAP Live animals sector
OCR Other crops
ODV Other developed countries
OFD Processed food
OLS Ordinary Least Square
OMT Pork, poultry, and other meats
OSO Oilseed and vegetable oil
PHL the Philippines
RES Natural resources, and petroleum product
RIC Paddy and processed rice
ROW Rest of the World
RRD Red River Delta
RWGT Row users weights
SAMs Social Accounting Matrices
SCC South Central Coast
SL Sensitive List
SPS Sanitary and Phytosanitary
STEs State Trading Enterprises
SUREG Seemingly Unrelated Regression
SVC Services
TBT Technical Barriers to Trade
TCN Transport, communication
TEL Temporary Exclusion List
THA Thailand
xviii
TRIPs Trade-Related Aspects of Intellectual Property Rights
TWGT Trade users weights
TXT Textile and apparel
UN United Nations
UNCTAD United Nations Conference on Trade and Development
UNDP United Nations Development Programme
USA United States of America
USD USA Dollar
USDA United States Department of Agriculture
VF Vegetable and fruit
VHLSS Vietnam Household Living Standard Survey
VLSS Vietnam Living Standard Survey
VND Vietnam Dong
VNM Vietnam
WB World Bank
WITS World Integrated Trade Solutions
WTO World Trade Organisation
XEA Rest of East Asia
XSA Rest of South Asia
XSE Rest of South East Asia
XWGT Cross users weights
1
CHAPTER 1 : INTRODUCTION
1.1. Introduction
The objective of this study is to analyse the implications of trade liberalisation on
Vietnam’s small-scale livestock producers. The study examines how household
production, consumption, leisure, working time and welfare are affected when prices
change due to trade liberalisation. The specific objectives are as follows:
• Evaluate the impact of trade liberalisation on Vietnam’s smallholder livestock
producers following Vietnam’s integration process into the world economy, by
connecting a computable general equilibrium (GTAP) with a household
production model for small livestock producers, simulating trade liberalisation
scenarios (in GTAP) and analysing changes to identify welfare effects at the
household level
• Investigate how the effects of trade policy changes impact on the national
economy as well as the livestock sector and their consequent effects on social
welfare.
• Assess impacts of trade liberalisation to changes in livestock producers'
household welfare. Analyse the reaction of smallholders that is represented by
their responsiveness in the supply of labour and demand for commodities and
leisure. By constructing a household model, behaviour of the household in
reaction to external impacts is captured, not only in terms of production
decisions, but also in term of consumption and household time allocation.
• Discuss opportunities and threats from trade liberalisation for smallholder
livestock producers.
1.2. Background Information of Vietnam Agriculture
Vietnam is located in the east of the Indo-Chinese peninsula (see map). The total natural
land resource of the country is 325 360 km2, of which 20-25 percent is used for
2
agricultural production. Vietnam is divided into 64 provinces and 8 agro-ecological
regions.
With Gross Domestic Product (GDP) per capita of about 800 USD (2007) Vietnam is
still considered one of the world's poorest countries. Nevertheless, many observers note
the rapidly growing economy since the Doi Moi1 (renovation) in 1986 which
corresponds with the beginning of the transformation from a centrally planned economy
to one which is market oriented.
With 72.5 percent of the population
living in rural areas and 59 percent of
labour working in the agricultural sector
(year 2007), the agricultural sector is the
foundation of the economy. Over the
recent past, GDP growth of the
agricultural sector has been relatively
stable at around 4.3 percent per year.
Vietnam has also progressed from a
national chronic food shortage to one of
the world’s leading exporters of
agricultural products including rice,
coffee, tea, and rubber. Economic
growth, national food security,
improvement of income and lives of the
majority of people, and the reduction in
the number of people living under the
poverty line were largely made possible
by the Doi Moi reforms. The remarkable
economic growth that resulted from
these reforms was based largely on rural
households which have become the new
unit of agricultural production.
1 Refers to the decollectivisation and market orientation of the Vietnamese agriculture sector
VIETNAM AGRO-ECOLOGICAL REGION MAP Note: NE: North East, NW: North West, RRD: Red River Delta, NNC: North Central Coast, SCC: South Central Coast, CH: Central Highland, NES: North East South, MRD: Mekong River Delta
3
In the agricultural sector, livestock is considered as one of the important pillars of a
sustainable development strategy (IFPRI 2001). Given limited prospects in the growth
of rice production and changing patterns of demand both in Vietnam and in the world
market, the livestock sector can help to achieve higher and more stable rural incomes,
reduce the incentive for the flow of migrants from rural to urban areas, make the
farming system sustainable in the long run, and contribute to the alleviation of rural
poverty, especially among ethnic groups in mountainous areas.
For many parts of Vietnam, large animals (i.e. bull, cow, and buffalo) are important
primarily not for their meat and milk products, but for the draught power they provide
for crop production and local transportation. Livestock raising (especially swine and
cattle) are a major source of fertiliser. In addition, livestock is important for providing
high quality protein to millions of smallholders thus enhancing their nutrition.
After the Doi Moi, the livestock industry has developed with an average growth rate of
5.4 percent per year, higher than the crop and service2 sectors. However, there are many
issues which constrain producer profitability in the livestock sector such as high feed
price, low productivity, and poor veterinary system.
Livestock in Vietnam are predominantly raised in small-scale household production
units. At present, smallholder producers supply the majority of meat in the market, with
most households operating individually in the production and marketing of livestock
and livestock products. They are constrained by poor access to markets, very low scale
operations, poor access to improved genetics, and high-quality forage and concentrate
feed, and poor animal husbandry and nutrition. However, for most of these households,
raising livestock is an important source of income providing at least 50 percent of cash
income in the household (Lapar et al. 2003).
Since 1995, Vietnam officially joined the world economy by becoming a member of the
Association of South East Asia Nations (ASEAN), signed a bilateral trade agreement
(BTA) with USA in 2000, and became a member of World Trade Organisation (WTO)
in 2007, after 11 years of negotiation. Implementation of multilateral and bilateral trade
agreements has opened a door for Vietnam to integrate into the global economy
providing both opportunities and threats to the agriculture sector. In the case of the 2 Refers to service subsector in agriculture
4
livestock industry, there may be both supply and demand side effects. For example,
income growth may increase demand for meat, but the domestic industry may also have
to compete with imported produce. Reducing protection on domestic maize may make
feed prices decrease, but the opportunity cost of labour in livestock production may
increase. Thus it is not clear whether livestock households, especially smallholders, will
be worse or better off as a result of trade liberalisation.
1.3. General Methodology and Data for Study
In order to assess the impacts on small households as a consequence of trade
liberalisation, an international trade model is linked with a household model. In general,
the modelling process of the study includes three main steps:
Step 1: At the micro level, a household model is built where the household utility is
optimised in the context of a number of constraints such as budget constraint, time
constraints and production technology.
Step 2: At the macro level, a multi-country general equilibrium model with its focus on
worldwide trade policy - Global Trade Analysis Project Model (GTAP) - is used to
simulate trade liberalisation.
Step 3: The trade model and household model are linked together and the results of
changes in welfare, production and consumption behaviours, etc. of the household are
presented. Given the aim of investigating the reaction of the household in trade scenario
simulations, price changes for consumption commodities as well as production factors
including labour in the agricultural sector, which are derived from the GTAP
simulation, are incorporated into the household model as policy shocks. In this step, the
study only examines the one-way effects of trade liberalisation on households, and not
their influence on trade.
A variety of empirical methods are used in the study. For example, econometric models
such as Linear Approximation Almost Ideal Demand System (LA-AIDS) and Linear
Expenditure System (LES) are applied in step 1 to estimate parameters for functions of
consumption, production, and utility of the household. In step 2, the GTAP software
SplitCom is used to separate pig and poultry out of the aggregate group of livestock in
5
the standard GTAP aggregation. Other statistical analysis is also applied to estimate the
price and wage transmission among regions within the country, when the two models
are linked together in step 3.
The study uses data from the Vietnam Household Living Standard Survey (VHLSS
2004) as the main source for constructing the econometric and household models. The
total survey sample was 45 900 households, of which 36 720 households were surveyed
for income only, the other 9180 households for both income and expenditure. The
sample covers and represents 3063 communes, over eight ecological regions (Red River
Delta, North East, North West, North Central Coast, South Central Coast, Central
Highland, North East South, and Mekong River Delta), and almost all provinces in the
country. Since the study especially focused on small-scale households in the livestock
sector, around 7000 of the 9180 households are chosen for analysis. Data from a survey
on the Vietnam livestock sector of the International Food Policy Research Institute
(IFPRI 2001) is also used for additional information about small households who raise
livestock and other criteria of economic and farming systems in Vietnam.
In constructing the household model, the whole country is divided into four regions:
Red River Delta, Northern upland (includes North East and North West), Central region
(includes North Central Coast, South Central Coast and Central Highland), and South
(includes Mekong River Delta and North East South). Each region represents an
ecological area where agro-ecological and economy conditions are largely similar, so
one household model is constructed to represent each region.
The GTAP model version 6.2, with a database for 2001 is used in step 2. The database
includes 96 countries/regions and 57 individual sectors. The database includes all data
on trade, production, consumption, tariff and etc. for Vietnam individually; therefore
the Vietnamese economy with all its factor and activity flows is represented in the
model. However, it does not identify the pig and chicken sectors separately, which are
subsumed within “livestock”. Generating new sectors of pig and chicken to introduce
into the GTAP database are based on extracting data from sources of United Nations
(UN): Comtrade, International Statistics, World Integrated Trade Solutions (WITS),
Statistics of Food and Agriculture Organisation (FAOStat), and Social Accounting
Matrices (SAMs) of various countries.
6
In addition, the study uses other sources of primary and secondary data and information
including government institutions and organisations in Vietnam, such as the Ministry of
Agriculture and Rural Development (MARD), General Statistical Office (GSO),
Research Institute for Market and Prices. Information also comes from documents and
databases of the United Nation’s and international organisations and financial
institutions: United Nations Development Programme (UNDP), United Nations
Conference on Trade and Development (UNCTAD), United States Department of
Agriculture (USDA), World Bank (WB), Asian Development Bank (ADB), IFPRI, etc.,
and other non-government organisations in Vietnam.
1.4. Contribution to Knowledge
The major contribution to knowledge derived from this study is development of a link
between the trade liberalisation GTAP results and a household model to examine
changes in welfare as well as behaviours in production and consumption of small
livestock producers in Vietnam following trade liberalisation. GTAP has been used
since 1992 with some previous applications to Vietnam (Nguyen & Ezaki 2005, Daude
2004, Nin et al. 2003, Holst 2004, Vanzetti & Pham 2006). Household models have
been developed for around 30 years and applied to many countries (Barnum & Squire
GDP per capita (in the 1st axis) total meat per capita beef and buffalo pork poultry
Figure 2-9: Changes in Population, Income and Meat Consumption per capita
Source: GSO 2008. Meat per capita consumption is live weight meat per capita (kg)
24
Among livestock products, pork is consumed the most, with over 75 percent of total
meat. One reason is its cheap price and it is customary to eat pork compared with other
meat types. Poultry is the second most popular meat to consume. Figure 2-11 presents
pig and chicken consumption per capita over the regions.
In terms of quantity, pig consumption in urban areas is much higher than in rural areas,
but there is little difference for chicken. However, in terms of quality, consumer
requirements in urban areas are rather different to rural areas. About 36 percent of
households in urban areas prefer to buy high quality meat despite relatively higher
prices (Lapar et al. 2003). They are willing to pay a premium of about 6 to 16 percent
above the regular price if they can be sure of receiving high quality and safe products.
0
4
8
12
16
20
RRD NE NW NCC SCC CH NES MRD
region
cons
umpt
ion
per
capi
ta (
kg)
Pork Urban Pork Rural Chicken Urban Chicken Rural
Figure 2-10: Pig and Chicken Consumption per capita over Regions (kg)
Source: calculated from VHLSS 2006
Regarding consumption habits, Vietnamese consumers prefer fresh meat rather than
frozen or processed meat4, especially in rural areas. Processed meat consumption in
households is less than 5 percent of total meat volume5. For daily consumption,
consumers often buy fresh meat products from wet markets, where almost all livestock
products produced in Vietnam are sold.
Even though consumption per capita has increased over recent years, Vietnam’s
consumption is quite low compared with other countries. The most recent data shows
4 The processed meat is popularly in the form of canned meat such as slices, ham, and sausage. 5 Calculated from VHLSS 2006
25
that in 2003, when meat consumption per capita of Vietnam was about 26 kg, the
numbers for China, Malaysia, USA and Europe were 54 kg, 48 kg, 123 kg, and 91 kg,
respectively (FAOStat 2008). This gap suggests that the potential domestic market for
meat in Vietnam is quite high. Thus, focusing on the domestic market is still an efficient
strategy, especially since Vietnam does not have a high comparative advantage in
livestock production for export.
However, in order to meet the requirements of markets and have opportunities to
expand livestock production, livestock households need to overcome some difficulties
in trade and marketing.
Most trading activity takes place locally, with most small farmers selling at the farm
gate, as there is little access to organised markets and auctioning systems. Information
about markets, prices and other supporting information is limited. The lack of a
widespread system of organised live animal markets in Vietnam means that the majority
of marketing and distribution of live animals and animal products is undertaken through
a network of marketers operating in informal groupings and often undertaking
exchanges on a face-to-face basis. Table 2-4 reports the proportion of pig and poultry
sold to different stakeholders in the marketing channel, rather than to final consumers or
assemblers and distribution companies of livestock products.
Table 2-5: Share of Pig and Chicken for Sale by Household (percent)
Assemblers Retailers Slaughterhouses Wholesalers
Pig producers
Small-scale 23 12 60 5
Large scale 30 30 40 0
Chicken producers
Small-scale 50 25 0 25
Large scale 40 50 0 10
Source: IFPRI 2001
26
The marketing channels of livestock husbandry are complicated and involve many
stages. Stakeholders usually involved in the channel include producers, assemblers,
wholesalers, processors, retailers and consumers. There are often four middlemen:
traders, wholesalers, slaughterhouses/meat processors, and retailers. Live pigs and
piglets are primarily sold to traders, wholesalers and slaughterhouses while pig
carcasses and other meat products are usually sold to retailers or directly to consumers
in wet markets. Marketing channels in the chicken marketing system are similar to those
of the pig marketing system. Live chickens are usually sold to wholesalers and retailers,
while carcasses are sold to wholesalers, retailers or directly to customers. Through
marketing channels, profits are shared in many segments and final prices of the products
are also affected6.
Moreover, due to the predominantly small-scale operations and wide dispersion of
production units, costs of collection, selection, and classification of assemblers also
increase. In addition, the high risk of getting poor quality animals from various sources
with uncertain product quality, add to the costs in terms of moral hazard. Specifically,
when buying from uncertified sources, the uncertainty about product quality imposes a
tax on the value of the good, so that the buyer will not be motivated to pay the real
market price, but rather would offer to pay a lower price, thus generating a loss in
potential income to the producer.
2.3. Policy Environment for Livestock Husbandry Development
In the last sections, a review of livestock sector development and issues in development
were presented. For a clearer picture about the livestock husbandry situation in
Vietnam, the policy environment and strategy for the development of the sector in
recent years and the near future are now presented.
Since implementing economic reforms, a set of policies that support the development of
the agricultural sector in general was promulgated. In these, policies such as admitting
households as one unit in agricultural production, allocating production factors and land
6 According to research of IFPRI, the marketing margin (difference between farm gate price and retail price) of pig meat in 2001 varied from 40 percent of farm gate price in the North East to 76 percent in the Central Highlands
27
to each individual household, and opening the economy to external markets contributed
to remarkable growth of Vietnam’s agricultural sector.
For sustainable development of the agricultural sector, a diversification strategy was
adopted in which the livestock sector was considered a major component. However,
unlike the rice or sugar sectors, there was no specific set of government policies in place
for the livestock sector (Lapar et al. 2003) but the policies supporting the development
of the sector often form part of the framework in the general policy for agriculture.
For example, with resolution No.10 issued in 1988, households are recognised as
independent economic units and have the right to use their land over the long term (30
to 50 years), they are their own bosses and can decide how and what to produce and sell
and where to get their inputs7. This and other policies on lending to families and
agricultural extensions, connect two paramount resources, labour and land, in order to
develop agriculture to the utmost. This was an impetus for households to help their
productivity and sell products for more income. With limited prospects for high benefit
through rice production alone, and rapid demand increases for meat due to population
growth and urbanisation, livestock husbandry was chosen by households to develop in
response to market demand.
There were some other policies and programs such as supported programs from the
government for crop varieties and livestock breeds, hunger alleviation and poverty
reduction programs, etc., which also supported the development of the livestock sector.
These policies are often designed and combined into one general supporting policy. At
the national level, program 125 provides 10 to 13 billion VND every year for
supporting the cost of breeding of pigs, cows, and poultry for producers; Program 225
provides about 100 billion VND to upgrade research institutes which develop crop
seeds, and animal breeds, and to subsidize seed imports and promote seed
multiplication; the program for agricultural extension provides a subsidy of about 30 to
50 billion VND each year, to support the transfer of new technologies into agricultural
production. In addition to that, many programs are conducted at the provincial level to
7 Before then, under the collectivisation period, household members took part in production as workers and received salaries. The farmer-workers did not have any decision about cost-benefit in production. Since their salary was fixed and not affected by their works, they contributed as little as they could, this led to an unavoidable collapse of the collective system.
28
support agricultural production in general, therefore it is difficult to define the full
extent of support for the livestock sector (Barker et al. 2001).
The Vietnamese government continuously values livestock husbandry as an important
economic sector in agricultural production, contributing to the agricultural growth rate.
As an orientation for development of the livestock sector in the future, a Livestock
Sector Development Strategy to 2020 was developed at the end of 2007. The general
objectives were three-fold, that by 2020: (i) increase the share of livestock in
agricultural sector GDP to 42 percent (from an estimate of 32 percent in 2010); (ii)
increase the use of commercial feeds in total feeds to 71 percent; and (iii) monitor and
control a number of dangerous animal diseases such as Foot and Mouth Diseases
(FMD) and avian influenza (MARD 2007).
These objectives are being pursued through more rigorous planning from feed,
production, slaughtering, processing to marketing; using advanced science and
technology to improve productivity and quality of livestock products; supporting
livestock production through commercial farms and industrial farms at the same time
reorganising small household producers; and supporting other services in the livestock
sectors such as extension, market information, breeds and feed supply, etc.
2.4. Summary
Recently, the livestock sector has developed considerably, both in quantity and quality.
This development has been considered a major component in agriculture, contributing
to the agricultural diversification program not only in response to increased demand but
also as a source of higher, more stable incomes and better nutrition for people.
However, the development level of the Vietnam livestock sector is still low, in
comparison with other similar economies. In the development process, the sector also
faces many issues such as low productivity, low quality of livestock products, high
price and low quality of livestock feed, slow improvement, poor genetics, animal
diseases, etc. The complexity of these issues represents a challenge for the sustainable
development of the sector.
Livestock production in Vietnam is undertaken predominately at the household level, by
producers who are non-specialised. These households generally undertake cropping in
29
addition to raising livestock and they usually keep more than one type of livestock in a
mixed raising system, which often combines pigs and chickens. The traditional method
of raising livestock with local breeds and low cost feed products does not produce good
quality and high standard meat products with competitive prices.
In the future, confronted with opportunities to expand the livestock sector, the question
of how to develop the livestock production system effectively is important. The strategy
of the Vietnamese government is to develop the modern livestock system to large-scale,
intensive commercial farms to meet rising market requirements. In that context, how the
small livestock producers with traditional systems, which are the largest sector in the
rural areas, can cope with change is an important question. In particular, how they can
adjust to the challenges presented by trade liberalisation, is the question that this thesis
aims to address.
30
CHAPTER 3 : THEORETICAL FRAMEWORK OF
AGRICULTURAL HOUSEHOLD MODEL
The analysis of the household is an area of concern in all the social sciences:
anthropology, economics, sociology, etc. However, each discipline approaches the
household on the discipline's own terms. From the point of view of the economist, they
come to an analysis of the household from their earlier focus on the individual (as
consumer) and the firm (as producer) to the neglect of the family and have been
concerned primarily with defining the household in relation to production and
consumption.
A basic model with a conventional approach of maximising a utility function is useful
for understanding decision making of an agricultural household, since it can be used to
integrate information on their production and consumption activities. Such models have
been widely applied to households in developing countries (Braverman & Hammer
Sadoulet 1995). The main difference with this kind of model compared to applications
of consumer choice theory is that it provides a methodology for integrating an
agricultural household’s production and consumption decisions into a unified
theoretical and econometric framework. It means that, when consumers maximise their
utility, they are confronted with not only a limited budget, but also time and technology
constraints of the household.
The first model to integrate production and consumption decisions in the analysis of the
peasant household was proposed in the 1920s by Chayanov, a Russian agricultural
economist (Thorner et al. 1966), and later developed by Mellor in 1963, Sen in 1966,
and Nakajima in 1986 and others.
In the Chayanov model, the household seeks to maximize its utility, where utility is
derived from the consumption of goods produced on the farm, purchased goods, and
leisure. By combining utility maximization from consumption theory with the
production function from production theory, the Chayanov model provides a foundation
31
for the integrated models of household decision making. The conclusions of the
Chayanov model are linked to key assumptions regarding the interaction between the
household and markets. Most notable among these assumptions are that the household
lacks access to a wage labour market and that the household has unlimited access to
land. These assumptions lead to a prominent feature of the Chayanov model, namely,
the demographic cycle (or "life-cycle") hypothesis. Chayanov proposed a positive link
between the household's level of work effort and the consumer to worker ratio (c/w)
within the household. (Thorner et al. 1966).
The new Home-economics model is a refinement of the neoclassical theory of
consumption. In it, Becker provides an alternative model of time allocation within the
household (Becker 1965, 1991). Utility is redefined in the “new home economics
model” so that, rather than being based on purchased goods and services, the utility of
the household is based on home produced commodities. The household members
combine their time and human capital with purchased goods and services to create these
home produced commodities, otherwise known as “Z-goods”8. In general, most items
purchased by the household must be combined with household labour in order to
contribute to utility levels. In the model, the household, rather than the individual, is
clearly established as the unit of analysis. The model also has one advantage in
comparison with the previous one, as it provides a logical structure for exploring the
links between utility maximization and the allocation of time to productive activities.
More specifically, the model postulates that the time (or labour) of household members
is allocated according to the opportunity cost of each member's time. Finally, given that
the new home economics model is based on the assumption of full access to wage
labour markets, it means that the value of wage work, home production, and leisure are
all assigned an opportunity cost equal to the market wage rate.
In this chapter, a combined model of an agricultural household as both producer and
consumer is presented. The model was proposed by Barnum and Squire and further
developed in Singh, Squire, and Strauss (Barnum & Squire 1979a; Strauss 1984; Singh
et al. 1986). In the model, the demand side is determined by prices and income as
normal, however, income is partly determined by household production activities.
8 “Z-good” is produced for consumption by the household only. An example of a Z-good would be meals prepared at home, or higher-order goods such as “nutrition”, or “shelter”.
32
Therefore consumption behaviour is not independent of production behaviour. Under
some key assumptions this establishes a recursive property of the model: production
decisions are assumed to be made independently of consumption, but production profits
influence consumption. The framework of the model and relation between household
labour and labour market are presented in the next section, and detailed analytical
framework of the property is in Appendix A3.1.
3.1. The Basic Model
In general, an agricultural household is assumed to maximise its utility function. This is
specified as a function of market purchased goods, home produced goods and leisure
time9, and is written succinctly as:
),,,( iaMCLUU i =1, …., (3-1)
where
L = leisure
C = own-consumption of agricultural output
M = consumption of market purchased goods
ai = household characteristics (for example, number of dependents)
Clearly, L, C, and M could be vectors of commodities or leisure consumption for
different members of the household. However, in this particular application there is no
concern for intra-household distribution, so the household is assumed to maximise its
joint utility function as a whole.
This optimisation is subject to certain constraints. In the household model, the objective
function is constrained by three restrictions on household actions.
The first one is the technology constraint(s):
),( , AdDFF j j = 1,…,m, (3-2)
Where
9 The model is presented in this part is developed by Barnum, Squire, and Singh (Barnum & Squire 1979a, Singh et al. 1986)
33
F = total agricultural output
D = total labour inputs (both of household and hired) used in production of F
dj = other variable inputs, range of j from 1 to m
A = area of land used in F production
The production function of the household is assumed to be quasi-convex and increasing
in inputs, but marginal product is decreasing in inputs. The household can produce more
than one output, and hence have more than one technology constraint. However, total
land for cultivation activity is (here) assumed to be fixed.
The household has the opportunity of utilising its total endowment of time either
working on- or off-farm, or for leisure:
offf HHLT (3-3)
As mentioned above, total working time for on-farm work, D, includes both household
working labour and labour hired from outside (if needed).
hiredf HHD (3-4)
So if combining (3-3) and (3-4) together, we can rewrite the time constraint of the
household as follows:
hiredoff HHDLT (3-5)
where
T = total household time available for labour
L = leisure
D = total labour inputs (both of household and hired) used in production of F
Hf = level of household labour working on-farm
Hoff = level of household labour working off-farm
Hhired = level of hired labour working on-farm
34
The household maximises its utility subject to a budget constraint, which states that
total expenditure for physical commodities can not exceed the total money that the
household gets from work plus exogenous income. Assume that household labour and
hired labour are perfect substitutes and face the same wage rate:
jjhiredoff dwpFRHHwpCqM )( (3-6)
where
R = non-wage, non-farm net other income
q = price of M
p = price of C
w = wage-rate
Hoff = time working off- farm of household labour
Hhired = working time of labour hired in for farm
wj = prices of other variable factors
From this constraint, it can be seen that the role of the labour market is especially
emphasised in the model. In order to simplify the problem, those three constraints are
collapsed into a single constraint, namely the “full income” constraint as follows:
wTRwLpCqM (3-7)
where jj dwwDDpF )( is net profit from household’s agricultural
production. The left-hand side of the equation (3-7) is total expenditures of the
household and includes “expenditure” on leisure valued at the wage rate and the right-
hand side is an augmented version of Becker’s concept of “full income”, which is the
sum of any non-wage, non-farm net other income (R), a measure of the farm’s profits
(∏), and the value of the household’s stock of time (wT) (Becker 1965). Since land is
treated as a fixed factor, the rent payments or receipts, if any, are captured in the
definition of R.
This 'full income' constraint in particular distinguishes agricultural household models
from other approaches and highlights the interdependency between consumption and
35
production decisions made at the household level. Farm technology, quantities of fixed
inputs, and prices of variable inputs and outputs affect household consumption
decisions since they determine the size of the farm profit portion of the full income
constraint. Thus, this approach permits identification of linkages between farm
household production and consumption decisions.
3.2. Solving the General Model
By rearranging the full income constraint, now the problem of the household is
maximising its utility (3-1) with the constraint (3-7). The household can choose
quantities of the consumption for commodities and labour input for agricultural
production. Forming the Lagrangian, the household problem takes the following form:
)(),,( * wLpCqMYMCLU (3-8)
Where λ is the Lagrangian multiplier and Y* is the value of the full income that results
from profit maximising behaviour:
**** ),,( wDdwAdDpFRwTRwTY jjj (3-9)
where D* is labour input that the household chose for the farm’s agricultural production
to get maximum profit Π*, with land cultivation fixed A . Maximizing utility subject to
the full income constraint yields the following first-order conditions:
0
wL
U
L (3-10a)
0
qM
U
M (3-10b)
0
pC
U
C (3-10c)
0*
wLpCqMY
(3-10d)
The marginal conditions of the equations (3-10) can be solved to yield three demand
equations for choice variables C, M, L as follows:
36
),,,,( *iaYwpqCC (3-11a)
),,,,( *iaYwpqMM (3-11b)
),,,,( *iaYwpqLL (3-11c)
where ai are household characteristics. The demand system follows neoclassical theory,
where demand depends upon prices income and possibly household characteristics.
However, in the household model, full income, Y*, is determined by technological
production in the equation (3-9). Therefore changes in the factors that influence
production, profit and hence change in Y* will lead to changes in consumption
behaviour.
The model is also set up under some simplified assumptions, which help derive
consumer demand, output supply and variable input demand equations by modelling the
farm household decision-making process recursively as two separate stages, despite
their simultaneity in time. These assumptions briefly include: the household is a price-
taker in all markets and all markets exist; commodities are homogeneous, including the
labour market; decisions relating to the total stock of land and labour are treated as
given; intertemporal allocation and risk are omitted (Barnum & Squire 1979b).
3.3. Role of Labour Market
As mentioned in the last section, the assumption of the existence of a labour market,
and homogeneity in labour are particularly important in solving a recursive agricultural
household model. The dual role of an agricultural household as both producer and
consumer lets it make decisions at two disaggregated levels; that is one of an individual
farmer determining labour, and as the individual household deciding upon leisure or
supplying labour.
37
Figure 3-1: Separation of Household Decisions
The important implication of an active labour market is that household decisions about
the level of production and labour use can be made independently of household
decisions about consumption. In Figure 3-1 the relationship between household profit
and the amount of labour employed is represented by the curve π, which in turn is a
function of the physical production function and exogenous prices (P), at a particular
level of agricultural production technology. Given the objective of maximising profit,
and subject to prices of inputs, including labour cost, the household chooses the optimal
amount of labour to use on the farm at Hf = D* (where the marginal return to labour is
equal to the wage rate). The decision does not depend on household preference of
leisure, rather, it depends only on the production technology, and inputs and output
prices. Hence the difference between labour supply availability (DS) and labour demand
for farm (D*) is labour worked off-farm, Hoff. Since the household has access to a labour
market, and there is no restriction on the mix of household and hired labour, the
separation property allows separate estimation of consumer and producer sides of the
model.
However, this separation is very sensitive to violations of the assumptions of the
household model, of which the most important is an incomplete labour market10. Figures
10 There are a number of reasons why the household model could not be solved separately, such as preferences of farmers for working (on- or off-farm), family and hired labour are not perfect substitute, and the markets are incomplete. However, in the scope of this section, only the case of an imperfect labour market is considered to see the important role of taking part in the labour market
w
Hoff
U
DDS
$
0 Hf =D*
π {F(D,diA);P}
38
3-2 and 3-3 illustrate the role of a competitive labour market in decision making of an
agricultural household. Maintaining the assumption of perfect substitution of household
and hired labour, there are two cases which are most relevant in the “real world”:
“surplus labour” where off-farm opportunities are very limited, especially in a slack
season or after harvest time, and “labour shortage” in the peak season.
Figure 3-2: An Agricultural Household in Slack Season
Figure 3-2 shows a case of an agricultural household with limited opportunities to work
outside the farm11. Firstly, with the market wage w, the household would work on-farm
up to Hf, then supply the rest of the household labour Hoff12
to the labour market to get
the utility at U. Assume that there is a constraint on the amount of labour that can be
supplied off-farm H*off, which is less than Hoff. In that case, the income/labour frontier is
given by F’, which is production curve F(D,dj,A)+wH*off in the figure. Then, the
household will sell H*off of family labour to the market to have wH*
off off-farm income.
If applying the market wage for their farm, they only supply labour to farm work up to
the D* level, hence there is still some surplus labour in the household. In this case, the
11 The figure is drawn by the author, based on the discussion of Benjamin and Souficas (Benjamin 1992; Skoufias 1994), 12 Labour worked off-farm, Hoff, is the difference between total labour supply and labour working on farm, Hf.
U*
F(D,dj,A)
0
$
H*off
w
w*
U
F’=
F(D,dj,A)
+ wH*
off
wH*
off
D*D Hf Hoff Hf
*
39
household can increase utility by employing more labour on the farm till H*f, where the
opportunity cost of each day working is not the market wage w anymore, but w*<w
defined as the shadow wage. By supplying labour to H*f, the utility of the household
now is U*, lower than the utility U that the household could achieve if there was no
constraint on the labour market.
Figure 3-3: An Agricultural Household in Peak Season
Figure 3-3 shows the situation of an agricultural household in peak season, when there
is a shortage in labour. The figure includes two parts. Part A shows the standard
outcome where labour is hired in, without limit. The figure assumes that the labour
market is perfect, at the market wage rate w, as a producer, the household maximises
U
U*
w*
$
H*hired
w
F(D,dj,A)
- wH*
hired
D Hf Hf* D*
Hhired
F(D,dj,A)
- wHhired
U
w
$
D Hhired
0 Hf DD
F(D,dj,A)
- wHhired
0
Part A
Part B
40
their farm profit by producing at DD, using Hf household labour and Hhired13
labour
bought from the market, and get utility at U.
Part B shows the outcome when there is a limit on the amount of labour that can be
hired in. It is assumed that labour will be hired to the limit. The farm production
function is moved down a line with slope equal to the wage, so that the horizontal
displacement is equal to the limit. The intersection of this function with the vertical axis
indicates the profit that can be enjoyed from hired labour alone. The function then
indicates that increase in income earned as additional household labour is applied.
Utility maximisation occurs at the normal tangency between this function and the
highest indifference curve. In the first case, where the limit is exactly equal to the
amount of hired labour required - the constraint is not binding effectively. The
household with production function F, indicated by F(D,dj,A)- wHhired, chooses to
supply household labour at Hf, and pay wHhired to hire labour from market. By
construction, it can be seen that the optimal utility achieved is the same as if the market
were “unconstrained” at utility U. However, if the limit is genuinely constraining, only
H*hired labour can be hired from market, at the wage rate w* > w, the household supplies
the household labour to the farm till H*f, then a lower level of utility can be achieved, at
U*.
Both of the cases above illustrate the non-separation property of the household model if
the labour market is not perfect. Then, the decision of taking leisure and demanding
labour for the farm is a convolution, and depends not only on production technology,
prices of inputs, but also characteristics and composition of the household.
3.4. Profit Effects
As mentioned above, one of the major results of the agricultural model is the linkage
between production and consumption through profit. This section presents how the
consumption of a commodity changes when its price changes, if the household both
produces and consumes that commodity. In accordance with the previous section, the
commodity is C, with price p.
13 Labour hired from market is to cover the shortage between required labour DD and supply of household labour, Hf.
41
From the demand equation of C, the total change in quantity demanded, dC, where the
change in its price, dp, is determined as:
p
Y
Y
C
p
C
dp
dC
*
* (3-12)
The first term on the right hand side of the equation (3-12) is the standard substitution
effect. For normal goods, the change in the quantity consumed, given a change in its
own price, is negative. The second term captures the profit effect. A change in the price
of agricultural goods changes profit and hence full income through equation (3-9).
Then, in turn, full income changes the quantity demanded for C in equation (3-11a).
The second term can be expressed as:
dpDFp
dpp
Ydp
p
Y
Y
C)(**
**
*
(3-13)
Therefore equation (3-12) can be rewritten:
dpFp
C
dp
dC)(
As F(D) (the quantity of agricultural output) is always positive, the second term is
unambiguously positive, while, according to neoclassical demand theory, the first term
is negative for a normal good. Thus the total change in quantity demanded is the net
effect of two opposing substitution and profit effects.
Therefore, the “profit effect” is considered as a distinctive feature of the household
model. In applying the model, there are some policy implications different from those
that would be arrived at by traditional methods. Previous research results show the
advantages of the household model. Table 3-1 presents various studies in the literature
where notable differences in price elasticities with and without incorporation of the
“profit effect”.
42
Table 3-1: Selected Response Agricultural Price Elasticity With and Without
Profit Effect
Agricultural
commodity
Non-agricultural
commodity
Labour supply
Country/economy
A B A B A B
Taiwan -0.72 0.22 0.13 1.18 0.21 -1.59
Malaysia -0.04 0.38 -0.27 1.94 0.08 -0.57
Korea, Rep. of -0.18 0.01 -0.19 0.81 0.03 -0.13
Japan -0.87 -0.35 0.08 0.61 0.16 -1.00
Thailand -0.82 -0.37 0.06 0.51 0.18 -0.62
Sierra Leone -0.74 -0.66 -0.03 0.14 0.01 -0.09
Northern Nigeria -0.05 0.19 -0.14 0.57 0.03 -0.06
Source: Singh, Squire, and Strauss, 1986, Table 1-2, p.26.
A: Agricultural price elasticity without profit effect, B: Agricultural price elasticity with profit effect
However, the possible outcome of the total effect of price change on consumption is
highly dependent on what dominates: the substitution or profit effect, and varies across
each empirical study.
3.5. Summary
This chapter briefly presents an agricultural household model. As producers, the
household tries to maximise its net income with respect to levels of outputs and inputs,
subject to constraints determined by market prices, fixed factors and technology. As
consumers, the household maximises utility with respect to the quantity of goods
consumed, and leisure, subject to the constraints determined by the market prices
(includes wage), income, total time available, etc. However, in the household model,
additional flexibility regarding the consumption response to income change is allowed.
Through the “profit effect”, the two sides of production and consumption are linked
together in a recursive decision making.
43
The model of household behaviour presented above is a semi-commercial household
farm with a competitive labour market. As in other LDC countries, this type of
agricultural household is common in Vietnam, and lies on a continuum between wholly
commercialised farms only employing hired labour and marketing all output and a pure
subsistence farm using household labour and producing solely for home consumption.
This study applied this theoretical framework to build a representative household model
for Vietnamese small-scale livestock households. The results from econometric
estimation of the production and consumption sides of the model are presented in the
next chapter. Then the integration of these two sides allows for the construction of a full
simulation model of the household. The impacts on the household from trade policy
change can be simulated by integrating the trade and household model.
44
CHAPTER 4 : THE ECONOMETRIC MODELS
In this chapter, econometric models of production and consumption behaviour of the
agricultural household are presented. Developed as a recursive household model, the
production segment of the model is analysed employing a Cobb-Douglas (CD)
production function to estimate output, and hence profit from the farm. The
consumption side is specified using two stages: the Linear Expenditure System (LES)
for a broad grouping of goods and expenditures in the first stage, with the integration
between demand for commodities and the allocation of time for leisure and labour
supply. In the second stage, expenditure for each individual commodity in the main
food group is allocated using the Linear Approximation Almost Ideal Demand System
(LA-AIDS).
In this empirical study, econometric models are estimated from primary data of the
Vietnam Household Living Standard Survey 2004, a multi-purpose household survey.
The total survey sample was 45 900 households, in which 36 720 households report
income only, with 9180 households reporting both income and expenditure. The sample
covers and represents 3063 communes, over eight ecological regions14 and for almost
all provinces of the country. The survey included detailed questions on a household’s
characteristics and composition; income, expenditure, and saving; labour activities of
adults and children; health care and education access; housing and durable goods;
electronics; land holding and non-agricultural activities. The survey was carried out in
May and September 2004 (GSO 2004) with all questions relating to production and
consumption being framed for the previous 12 months. Therefore, even though the
avian influenza outbreak occurred in the winter of 2004, data collected from the survey
represents a normal year of the household.
Since this study pays specific attention to small-scale households in the livestock sector,
around 7000 of the 9180 households were chosen for research purposes15. In this
14 The ecological regions are Red River Delta, North East, North West, North Central Coast, South Central Coast, Central Highland, North East South, and Mekong River Delta 15 The classification of a household in the livestock sector is based on the criteria of gross value of production, includes value from raising livestock, cultivation production and other activities. This methodology is adopted from the research on Vietnam livestock sector of International Food Policy
45
research, the country is divided into four regions16: Red River Delta, the Northern
upland (includes North East and North West, in short NE+NW), the Central region
(includes North Central Coast, South Central Coast and Central Highland), and the
South (includes Mekong River Delta and North East South). Each region represents
ecological areas where agro-ecological and economy conditions are similar, and there is
a degree of homogeneity in farming system or activities of the household in production.
4.1. Production of Agricultural Household
All characterisations of the theoretical model seem to be suitable for the present study.
In Vietnam, most agricultural households take part in both livestock and other
agricultural crop production (mainly rice cultivation). Small-scale livestock production
accounts for as much as 80 percent of total livestock product in the country. Each
agricultural household often keeps 1–2 sows or less than 10 fatteners and/or some
dozens of chicken per year. For the purpose of the research, it was assumed that a
representative household produces only three commodities: paddy rice, pigs, and
chickens.
For the production segment of the household model, a direct estimate of the production
functions is employed, and the Cobb-Douglas functional form is chosen because of
analytical tractability.
Production functions take specific forms as follows:
rrrrrice VDAF 321
0 (4-1a)
ii
ii
iiii VDGF 321
0 i = pig, chicken (4-1b)
where A is land for rice production and is fixed, D is labour requirement, V are variable
inputs and G is feed for pigs or chickens. It is assumed that these production functions
can be estimated independently. It is also possible that there will be spatial
Research Institute (IFPRI 2001). Other criteria such as economic and farming systems are also considered for the purpose of household stratification. 16 In general, the researchers often divide Vietnam into 3 regions, the North, the Center, and the South based on the ecological conditions. However, in the realization that, RRD is in the delta, near by Ha Noi, the capital of the country, so the people in that area seem to have more opportunities to access to market and the other services rather than the people in the mountainous areas (NE and NW), therefore dividing the North into RRD and Up north mountainous areas (includes NW and NE) is necessary.
46
autocorrelation, given specific local conditions may affect spatially close farms.
However, detailed spatial information is not available on households and hence
consideration of any spatial correlation is limited to the fixed effects associated with
estimating regional models.
In the survey, only information on the total cost for each kind of input variable for
production, not the quantities, were available. In order to estimate input quantities for
production for the purpose of production function regression, the price of inputs are
required. Therefore price information was collected from the database of the
Vietnamese Institute for Market and Prices. This covers almost all agricultural and
consumption commodities’ retail prices over districts and provinces, weekly. The local
market price for each kind of variable input from the database was used as the price that
the household has to pay. According to Justino and Litchfield (2002), commune or local
prices often reflect accurately the prices faced by the households, since those
households as a commune usually have a single market where they purchase similar
goods at the same price.
Table 4-1 (next page) presents basic information on mean inputs and output levels of
rice, pig and chicken production of agricultural households in the four regions.
The results from ordinary least squares estimation of the CD production functions for
Vietnamese agricultural households are reported in Appendix A4.1. Here, the estimated
This section outlines the derivation and calculation of the production elasticity relevant
to the agricultural household model. Using the parameters estimates above, a table of
labour demand, output supply, and profit elasticity to price of output and price of labour
for each agricultural activity was constructed (Table 4-3). These elasticities are
interpreted for their significance in production decisions and are used later to determine
the direction and magnitude of the profit effect and its impact on household output price
responsiveness.
17 Production functions for other regions are calculated in similar ways; however, different neutral technological efficient parameters lead to shifting of production functions along output value.
50
Table 4-3: Elasticity of Output, Labour Demand and Profit with Respect to
where, D is labour requirement, V are variable inputs, and GR, GI are quantities of raw
and industrial feeds, and pR and pI are average price indices of raw and industrial feeds,
respectively. In the simulations, the proportion of raw and industrial feed in total feed is
constant.
5.1.2. The Consumption Model
On the consumption side, all parameters from the estimated econometric demand
functions are used in the household model. To estimate the demand of each food in the
19 In this study, retail prices of some main commodities are obtained from a weekly database representing markets in provinces of Vietnam constructed by the Institute for Market and Prices (under the Ministry of Finance) .
71
main food group using the LA-AIDS model, the functional form20 outlined in Chapter 4
is used:
**
* )ln()ln( iij
jijii P
Mpw
The constants (αi*) need to be adjusted to let the modelled demand shares equal the
observed ones of the household in the base year. Table 5-3 presents adjustments in
constants of the LA-AIDS model for the purpose of the models’ calibrations.
Table 5-3: Calibration of Constants in LA-AIDS
Regions α*rice α*pork α*chicken α*fish α*vegetable α*othermeat
Original 1.441 -0.369 -0.043 -0.028 0.174 -0.002 RRD
5.1.3. Opportunity Cost of Each Working Day and Exogenous Income in Base
Year
As mentioned in Chapter 4, in estimation of the LES model, defining proper prices for
consumption commodities and opportunity cost of each labour day is necessary to get
appropriate parameters. For a representative household, the opportunity cost of each
labour day in the base year is defined as the weighted average of wages dependent on
time working on-farm and off-farm, and is used as base year’s labour price.
20 Where wi is the budget share of a given food; pi is the price of commodity i; M is a measure of household welfare, typically per capita income or per capita expenditure for main food group; P is the price index, μi is random disturbances.
72
Note that the standard household model applies the recursive property, where the model
assumes factor markets are complete and the household is a price taker. In the model
simulations in Chapter 7, all prices including labour are considered as exogenous
variables.
As mentioned in Chapter 3, the household maximises its utility with a “full income”
constraint: wTRwLpCqM . This constraint implies total expenditure of
the household (left hand side of the equation) is not greater than total income (right
hand side). Therefore, in calibrating the model in the base year, the sum of any non-
wage, non-farm net other income (R) of the household is calculated as a subtraction
from total expenditure and net farm profit () and the value of the household’s stock of
time (wT).
For the purpose of simulating, it is assumed that the parameters defined in the base
year’s calibration remain constant over time.
5.2. The Interaction of Production and Consumption Decisions
According to the theoretical framework of an agricultural household model, the
household reacts to a change in an exogenous variable, such as the output price of a
commodity the household produces, through price and income effects on consumption
plus a change in production decisions. The change in production decisions is captured
through its influence on on-farm profits and referred to as the profit effect. Since the
model is designed to capture both consumption and production, it yields more realistic
results than an examination of each side individually. To get an indication of the
quantitative significance of the integrated model, the partial household response
elasticity is first derived under the assumption that net farm profit, П, is exogenous,
thereby ignoring the production side of the model. Then the total response elasticity of
the full model is derived by allowing profits to be determined endogenously and
incorporating the production side of the model.
The complete household model consists, on the consumption side, of the demand
functions derived from the first order conditions and the expenditure constraint (3-7).
And on the production side, of the profit function derived from the production functions
and first order profit maximising conditions. Or in terms of the estimated system, the
73
complete model consists of the demand functions derived directly from the equation
system (4-10).
Totally differentiating the full system of household equations and rewriting allows the
derivation of the set of total response elasticities, giving the proportional change in any
endogenous variable, Z, in response to a proportional change in any exogenous variable,
X. The total response elasticities are written in the form of a sum of component partial
elasticities:
X
XE
E
Z
E
E
Z
Z
X
X
Z
Z
X
X
Z**)constant()variable( (5-1)
or in other notation:
XEZEZXZX ..* (5-2)
where ZX represents the elasticity obtained if farm profits are allowed to change and
*ZX is the elasticity obtained if farm profits are fixed (Barnum & Squire 1979b). This is
the price elasticity of demand. The remaining three terms on the right hand side are
expenditure elasticity of demand, profit elasticity of expenditures and the exogenous
variable elasticity of profits. Note that the profit elasticity of expenditures
E simplifies to П/E from the 'full income' constraint (3-7)
Using the elasticities calculated for production and consumption in Chapter 4, Table 5-4
presents the elasticities of the RRD household with and without the profit effect.
Table 5-4 indicates the significance of integrating production and consumption
decisions by comparing ZX and *ZX . For some commodities, when the price changes it
affects both consumption and production (rice, pig, chicken, and wage of labour). When
farm profits are assumed to remain constant, the own-price effects reduce consumption
of all commodities. For example, when farm profit is assumed to remain constant,
increases in the price of rice, pork and chicken reduce consumption of the main food
group. But when farm profit is allowed to change, the response to price is reduced. This
may be explained by the fact that farm profits increase as these prices rise, and hence
households have more profit and income to cover expenditure.
74
Table 5-4: Household Response Elasticities with Farm Profit Alternatively
Exogenous and Endogenous in RRD
Elasticities with the consumption quantities/labour supply Variables
Main foods Other foods Industry and others Labour supply
P other foods 9E-17 -1 1E-16 -2.8E-17
P industrial goods -0.053 -0.109 -0.616 0.016
P main foods -0.645 -0.323 -0.180 0.049
P rice *ZX -0.656 -0.329 -0.183 0.049
ZX -0.264 0.476 0.264 -0.072
P pork *ZX -0.675 -0.338 -0.188 0.051
ZX -0.502 0.016 0.009 -0.002
P chicken *ZX -0.657 -0.329 -0.183 0.049
ZX -0.568 -0.147 -0.082 0.022
Labour wage *ZX -0.606 -1.243 -0.691 0.337
ZX 0.832 1.706 0.948 -0.106
Note: *ZX elasticity with exogenous farm profit, ZX elasticity with endogenous farm profit
At the same time, labour supply responds positively to increases in exogenous prices of
commodities. This may be due to increased prices reducing household purchasing
power and forces the household to sell more labour to cover expenditure. The last row
of Table 5-4 shows that an increase in wage results in an increase in consumption of all
foods, industrial commodities and leisure. These changes can be interpreted as more
household income due to wage increases resulting in increased consumption.
For commodities where price changes affect both consumption and production such as
rice, pork, chicken and wage, of the 16 pairs of elasticities shown in Table 5-4, 10 pairs
differ with respect to sign. Once again, this emphasises the need for assessing the
effects of exogenous variables to behaviours of the household in a single model, with
the integration of production and consumption decisions for more accurate results. The
response elasticities of the households in other regions are reported in Appendices A5.1,
A5.2, A5.3.
75
5.3. Summary
This chapter presents the process of constructing and calibrating representative
household models for each region. All parameters used in the model are taken from the
results of the econometric model in the previous chapter. In addition, some information
defined private characteristics in livestock production of households in each region are
used in the calibration.
By integrating two sides of the model - consumption and production - the response of
households to changes in exogenous variables through standard price and income
effects, and in production reactions, rather than changes in household behaviours of
consumption or production individually is also presented.
The models21 constructed here are considered representative of typical households for
each region, assuming that all households in the region have similar characteristics and
respond to a given change in the same way. A model for each region is constructed for
the base year. The model includes 2 elements: production of rice, pig, and chicken by
the household is represented by Cobb Douglas production functions; and consumption
by the household, represented by three categories:the main food group (which includes
rice, pork, chicken, beef, fish and shrimp, and vegetable), other food group, and
industrial group. the two sides of the model are linked together by a utility function.
Following any impact from the outside, such as price fluctuations, the household would
react by changing both production and consumption activities. On the production side:
the household will choose production level of outputs, and labour supply so that the
household maximises benefit. The benefit from agricultural production plus exogenous
income, known as total income of the household will be used to maximise utility of the
household via consumption of commodities.
In the next chapter, trade liberalisation scenarios and a global trade model will be
examined. Linking the trade and household models allows for analysis of how
households will be affected by changes in trade scenarios.
21 The structure of the model and detail of the model in equation form with the coefficients for each regions are presented in the Appendix A5.4
76
CHAPTER 6 : VIETNAM’S TRADE LIBERALISATION
AND SIMULATION MODEL
As part of the Doi Moi (renovation) restructuring process, Vietnam has started to
integrate into the world economy. Since then, a series of trade reforms in various forms
of unilateral, bilateral, regional and multilateral liberalisation have occurred. In its
integration process, Vietnam negotiated and signed with more than 100 trade partners.
Among them, a trade agreement with the European Union (EU) was signed in 1992, an
agreement to become an official member of the Association of South East Asia Nations
(ASEAN) in 1995 and a joint ASEAN Free Trade Area (AFTA) in 1996. In 2000
Vietnam entered into a bilateral trade agreement (BTA) with the USA. Becoming the
150th member of the World Trade Organisation (WTO) on 11th January 2007 is an
important milestone in the long process of efforts of Vietnam to integrate into
international markets (Nguyen 2004, Abbott et al. 2007).
Each time a major agreement was reached, Vietnam’s trade with that region expanded,
and these trade agreements were clearly an impetus to ongoing domestic economic
reforms in Vietnam to become a more open economy in the process of integration into
the global economy. Implementation of multilateral and bilateral trade agreements is
likely to provide benefits for the economy and increase welfare for society.
In order to investigate the implications of trade liberalisation on Vietnam, this chapter
applies a multi-country general equilibrium trade model using the Global Trade
Analysis Project (GTAP). Since the final objective of this study is to examine impacts
of trade liberalisation on small livestock households raising pig and chicken, separating
live pig and live chicken from the aggregate livestock sectors in GTAP database is
needed. This is achieved using SplitCom software, a necessary tool for splitting GTAP
commodities into homogeneous and differentiated sub-groups. Scenario simulations of
trade liberalisations are also presented.
77
6.1. Brief of Vietnam’s Trade Liberalisation Process and Commitments
In 1992 the ASEAN Free Trade Area (AFTA) was established in Singapore with
common targets of improving regional trade of ASEAN, absorbing foreign investment
through setting up one single market, and facilitating ASEAN countries to adapt to
increasingly changing international conditions, especially enhancing regional and
international trade negotiations. The major legal framework for the establishment of
AFTA is the Agreement on Common Effective Preferential Tariff (CEPT) which has
objectives of reducing tariffs and removing all non-tariff barriers. Officially
participating in ASEAN in 1995, Vietnam committed to reduce tariffs according to
CEPT, with the main commitment to reduce import tariffs to 0–5 percent for a wide
range of product groups (Vo 2005):
(a) The Inclusion List (IL) comprises products subject to immediate tariff reduction
ranging from 0–5 percent by 2006 (by 2003 for other ASEAN countries); those
products are neither sensitive nor require high levels of protection, and account for
about 96 percent of total tariff lines in Vietnam;
(b) The Temporary Exclusion List (TEL) comprises those items temporarily not subject
to immediate tariff reduction at the beginning, transferred gradually to the IL; and
(c) The Sensitive List (SL) comprises 89 unprocessed items of agricultural products,
such as eggs, vegetables, rice, prepared meat, sugar, etc. Tariff reduction for these
items, with the final rates ranging from 0–5 percent, started on 1 January 2004, to be
completed by 1 January 2013, except for sugar by 2010.
It also required the elimination of non-tariff barriers22 for most products originating
from other ASEAN members by 2006. However, AFTA allows member countries to
exclude some commodities from tariff reduction under CEPT in order to protect
national security, social morality, human lives and health. The commodities are put in a
group named General Exclusion List (GEL), with a requirement that these items are
also categorised as an exception under the General Agreement on Tariffs and Trade
22 According to the CEPT Agreement, non-tariff barriers that create quantitative restrictions on imports need to be removed after the tariff imposed on these imported products are reduced to less than 20 percent. Other non-tariff barriers applied to those products need to be removed within the next 5 years, but in any case not later than 2006.
78
(GATT) and other international commitments. Other items are excluded for different
reasons, such as guiding the consumers (tobacco and alcoholic drink), market control
(petroleum) or protection (un-manufactured tobacco), etc.
Since 2003 when CEPT/AFTA entered its final implementation phase. ASEAN leaders
decided to prioritise 11 sectors to speed up regional integration. Vietnam is actively
involved in some sectors including agriculture, fishery, textile, electronics, wood based
products and tourism (Vu & Nguyen 2005). In accordance with the program, in 2010 all
these sectors shall be liberalised for free flow of trade and investment throughout the
region. Beyond that target, these industries will become fully integrated both in terms of
trade, investment and technical aspects.
ASEAN agreed to expand the regional trade border by negotiating with other trade
partners. In November 2002, a new stage of cooperation between China in the post-
WTO accession era and dynamic ASEAN economies started by signing a framework
agreement on comprehensive economic cooperation between ASEAN and China. The
major goal of the agreement is to create an ASEAN-China Free Trade Area (ACFTA) in
2010, with the exception of Cambodia, Lao, Myanmar and Vietnam (CLMV) who will
join in 2015. The main content of the tariff negotiations under ACFTA covers: Early
Harvest Program (EHP), General Exclusion List (GEL), Sensitive List (SL) and Normal
List (NL). In particular, the EHP concerns all products in Chapters from 1 to 8 at the 8/9
digit level (Harmonize System - HS Code) which mainly covers agricultural products
and fish. Tariff reduction is scheduled for Vietnam from 2004 to 2008 with tariff rates
needing to be gradually reduced to 0 percent by 2008. Non-tariff barriers and safeguard
measures must comply with WTO disciplines (Vu & Nguyen 2005).
ASEAN also started negotiations with India and Japan in 2003, with Korea, Australia
and New Zealand in 2004 to establish free trade areas with these partners. ASEAN and
Japan agreed on the general pattern for Closer Economic Partnership (CEP) in 2012,
and 2015 for CLMV. However, unlike the ASEAN-China model of negotiation and
liberalisation commitment, this agreement is based on a bilateral Economic Partnership
Agreement (EPA) between Japan and each individual ASEAN country. Since then,
Japan has had five official talks with five ASEAN dialogue partners on EPA including
Thailand, the Philippines, Malaysia, Indonesia and Singapore. For the rest of ASEAN,
79
Japan has had limited consultations. ASEAN also started trade talks with Korea
covering three areas for liberalisation, namely trade in goods, trade in services and
investment. In 2004, the parties agreed on the objective to liberalise 80 percent of tariff
lines by 2009 (Vu & Nguyen 2005). Although these free trade agreements are still in the
process of being negotiated, and the scope of commitments have not yet been
determined fully, dialogues between ASEAN with other trade partners in the Pacific
Asia region are moving forward to the target of full liberalisation.
Simultaneously while taking part in the regional free trade agreements, Vietnam
negotiated and signed bilateral trade agreements (BTA) with other partners. The textile
trade agreement with the European Community in 1992 constituted one of its first trade
deals with a Western partner. The cooperation was followed by a broader cooperation
agreement in 1995, which granted Vietnam most favoured nation treatment in its trade
relations with the EU. The negotiation continued in 2004, when Vietnam was
negotiating to join the World Trade Organisation (WTO). With an agreement on market
access signed in December 2004, the European Union lifted all quantitative restrictions
for Vietnamese textiles. In return, Vietnam agreed to open its market further in a
number of areas of interest to the EU (Vo 2005).
After many rounds of negotiation, on 13 July 2000, Vietnam and the USA officially
signed a bilateral trade agreement, towards trade normalisation. The agreement was the
most comprehensive agreement ever signed by either Vietnam or the USA. The
agreement was prepared on the basis of WTO principles and regulations, especially
those relating to intellectual property and services trade23. In terms of trading goods,
USA committed to reduce the average rate of import tariffs on Vietnamese goods from
40 percent to 3 percent. In turn, Vietnam would also reduce import tariff rates on
agricultural and industrial products from the United States within two to seven years,
gradually removing all non-tariff measures, according to WTO standards. Clauses on
intellectual property and service trade were negotiated in BTA, based on the principles
and regulations under the WTO legal framework (Vo 2005).
23 The Agreement contains 63 articles and 14 appendixes with seven chapters, namely, (I) Trade in Goods; (II) Intellectual Property Rights; (III) Trade in Services; (IV) Development of Investment Relation; (V) Business Facilitation; (VI) Transparency-Related Provisions and Right to Appeal; and (VII) General Articles.
80
At the time of joining ASEAN in 1995, Vietnam formally began its WTO accession
talks. In 12 years of negotiation, Vietnam’s accession was undertaken through both
multilateral and bilateral channels. On the multilateral side, the working party
composed of 63 WTO member countries, organised dozens of meeting sessions to deal
with multilateral issues. Vietnam responded to more than 2000 questions from working
party members for transparency purposes relating to import licensing procedures,
sanitary and phytosanitary, technical trade barriers, intellectual property rights, etc
(Nguyen 2004). On the bilateral side, Vietnam negotiated with 20 countries/regions
including the USA, EU, China, Japan, India, Korea and Australia. The bilateral
negotiations mostly dealt with market access concession for goods and commitments in
the service sectors.
To be a member of WTO, Vietnam needs to conform to various WTO rules and
obligations, such as the Agreement on Agriculture24 (AoA), Sanitary and Phytosanitary
(SPS), Agreement on Technical Barriers to Trade (TBT), Agreement on Trade-Related
Aspects of Intellectual Property Rights (TRIPs), and other regulations related to state
trading enterprises (STEs), etc. The purpose of implementing these agreements is to
avoid distortion of trade and production.
Becoming the 150th member of WTO in January 2007, Vietnam has committed to
comply with WTO requirements since acceding. Particularly, Vietnam accepts as
binding its ceiling import tariff at an average rate of 17.4 percent for 10 600 tariff lines,
which decrease gradually to an average rate of 13.4 percent within 5 to 7 years. The
average of the current applied rate for agricultural products of 23.5 percent would
reduce to 20.9 percent within 5 to 7 years25. Regarding non-tariff measures, Vietnam
also commits to abolish export subsidies and domestic supports, except subsidies that
do not create obstacles to trade and is allowed by WTO. At the same time, all non-tariff
measures are transformed into tariffs or, in other words, tariffication. Other
commitments on service trading, STEs, TRIPs will also be implemented (MARD 2005).
24 The purpose of the agreement is to curb the policies that have, on a global level, created distortion in agriculture production and trade. This includes three categories namely market access restrictions, domestic support and export subsidies. The main contents of AoA are converting all non-tariff-barriers (NTBs) to trade into tariff equivalents, lowering import tariffs, gradually decreasing domestic support and export subsidy. 25 While the average committed agricultural tariffs of recently acceding members is only 19.8 percent, and of China, a recent WTO member, is 15 percent.
81
6.2. Trade Liberalisation in GTAP Model
6.2.1. Previous studies on Vietnam trade liberalisation
The previous section dealt with the process of trade liberalisation of Vietnam. It was
believed that joining multilateral and bilateral trade agreements would bring substantial
benefits to the Vietnam economy, by increasing efficient investment, technology
progress, and the ability to access cheaper and better quality goods and services.
However, it may affect the sensitive and infant sectors of the economy, since they have
to face fierce competition from foreign firms. To quantify the total effects to the
economy and stake holders in trade liberalisation, people often apply computable
general equilibrium (CGE) models.
There are number of such studies that assess the economy wide effects of Vietnam’s
trade liberalisation. Nguyen Chan and Tran Kim Dung (2001) used a CGE model to
evaluate the impact of trade liberalization for Vietnam as a small open/ price-taking
economy. By categorizing 97 sectors identified in the 1996 Vietnam I/O table into 33
production sectors (17 for domestic sale and 16 for export), the authors analysed the
effects of trade and tariff reforms on export.
Pham Lan Huong (2000) used a CGE model to study the effect of trade liberalization on
Vietnam’s economy. She found that removing tariffs eventually lead to a depreciation
of VND, which means an increase in competitiveness. Improvement in the
competitiveness will translate into an expansion in exports and consequently increased
output in the economy. She also estimated that under the context of tariff removal and
trade reforms, aggregate employment is expanded. And because of expansion of
employment, i.e. an increase in the demand for labour, household nominal income
increases although it may differ among household groups (Pham 2000).
Using a multi-sector, dynamic applied general equilibrium model and the Initial Offer
by Vietnam, Roland-Holst et al (2002) presented a set of assessments of the long-term
economic effects of Vietnam’s accession to the WTO. The study indicated that, while
WTO accession is essential to economic development and fuller participation in the
global economy, it is only a partial step towards realizing Vietnam’s economic
potential. They calibrated five different scenarios for Vietnam’s economy during the
82
period 2000-2020. The first scenario calibrates the Vietnam’s economy for this period
based upon the Business as Usual trends for productivity growth. This scenario is
viewed as the baseline for the dynamic counterfactuals of five different types. The other
four scenarios include the country joining the WTO and conforming to its commitments
but the domestic economy is not reformed; the country joins the WTO and implements
the agreed offers; the country signs the BTA with the US; and the capital market is
liberalized (Holst et al. 2002).
Another CGE model of trade, GTAP, which involves basic accounting relations that
track the value flows through the global data base, has been used with some previous
applications to Vietnam.
In 2001, Le Quoc Phuong used the GTAP to assess the economic impacts of the process
of Vietnam’s integration. Based on the arguments that Vietnam is integrating into the
world economy at different level, Phuong proposes four scenarios for his assessments.
These scenarios include: Vietnam’s unilateral integration; Vietnam joining AFTA;
Vietnam’s following the requirements of APEC; and global liberal trade. He found that
economic integration does have positive impacts on the Vietnam’s economy (Le 2001).
Nguyen Tien Dung and Misuo Ezaki (2005) link a global CGE model with a GTAP
2001 I-O database with a focus on international relations. The model includes 10
industries and 11 countries or regions. They found that regional economic integration
generally has positive economic impacts. The integration is not only welfare enhancing,
but also leads to a less unequal income-distribution. The research also predict that
household consumption and income rise significantly as a consequence of liberalization,
and the poor and rural groups benefit more than the rich. Moreover, the removal of
tariffs in trading partners provides Vietnam with greater market access, and exports rise
in all simulations (Nguyen & Ezaki 2005).
In 2005, Dimaranan et al. used GTAP to analyse liberalisation of tariffs and textile
export quotas. Dimaranan used the same GTAP I-O data as Nguyen and Ezaki (2005),
but aggregated into 22 sectors and 12 regions, and the research focused on industries
rather than households. (Dimaranan et al. 2005).
83
Recently, David Vanzetti and Pham Lan Huong (2006) provide the most recent CGE
impact assessment of the WTO accession on the Vietnamese economy. Six scenarios
are simulated in the GTAP 6 model: Unilateral, bilateral, harmonized, bilateral,
regional, multilateral liberalization, and free trade. Similar to Dimaranan et al. (2005),
they predict only limited gains in the agriculture and resource sectors, but large effects
on the textiles and apparel sectors. All scenarios, with the exception of harmonization,
lead to an increase in exports. The largest sectoral effects are in textiles and apparel. In
other sectors, chemicals show large percentage gains from a relatively low base. Growth
in these sectors pulls resources out of agriculture, and exports fall in several agricultural
sectors. The model also predicts that Vietnam can obtain most of the potential gains
from trade reform from unilateral liberalization (Vanzetti & Pham 2006).
To assess the impact of trade liberalisation on the livestock sector in Vietnam, Nin et al.
(2003) use a micro/macro approach that combines the GTAP general equilibrium model
with a simple micro model to measure expected impacts of trade liberalization on a
representative sample of Vietnam’s livestock producers. The results show that the
impact of trade liberalization on Vietnam’s livestock production tends to be small but in
general a more open Vietnamese economy would result in a deterioration of the trade
balance of livestock products. In spite of this, trade liberalization would benefit poor
livestock producers by increasing livestock prices relative to production costs, in
particular feed costs, and by increasing non-agricultural income (Nin et al. 2003).
Apart from using a CGE model, which provides a framework for economy-wide
analyses, and taking into account existing relations among the different sectors, factor
markets, households and the government, some researchers take a partial approach in
analysing impacts of trade liberalisation to specific sectors. It ignores feedback links
among markets and activities and relies instead on in-depth knowledge of specific
sectors and the economic actors who participate in them. There have been a number of
studies following this approach, including research on the impact of trade liberalisation
on some sectors such as rice (Oxfam 2001) or livestock in Vietnam (AusAid 2004).
84
6.2.2. GTAP database
In this research, GTAP, with its focus on worldwide trade policy is used, as it takes
advantage of the CGE model structure, in that it offers a consistent economy-wide
framework for analyse trade policies. The purpose is to analyse trade liberalisations,
both bilateral and multilateral trade agreements between Vietnam and other countries.
Since the latest version and most recent database include data for Vietnam, the
Vietnamese economy with all its factor and activity flows is represented in the model.
GTAP was initially developed in 1992 at Purdue University in the USA. It is a standard
Computable General Equilibrium (CGE) model based on the neoclassical theory of firm
and household behaviour assuming perfect competition, rational and utility optimising
behaviour. It is designed to be a multi-region, general equilibrium model with bilateral
trade flows between all regions and linkages between economies and between sectors
within an economy. The model uses the Armington approach where products are
differentiated by origin and assumed to substitute imperfectly for one another forming a
composite import aggregate that substitutes imperfectly for domestically produced
goods. Primary factors (land, unskilled labour, skilled labour, capital and natural
resources) are substitutable but are used as a composite in fixed proportions to
intermediate inputs (Hertel & Tsigas 1997).
Simulations are undertaken using the GTAP version 6.2 database, released in November
2003. The database has 96 countries and regions, 57 sectors, and includes tariffs, export
subsidies and taxes, subsidies on outputs and inputs such as capital, labour and land that
represented the world economy in 2001 (Dimaranan 2006). It is possible within the
GTAP model to define the required level of aggregation. For this study, in the
region/country aggregation, the objective is to split the ASEAN countries as much as
possible to distinguish the economic effects of trade liberalisation and highlight
important regional economic relations. Other important trade partners of Vietnam, such
as the USA, Japan, China, Korea, and Australia are detailed. Meanwhile, countries with
similar economic conditions such as European countries, or some developed countries
are aggregated together. African and Latin American countries are also grouped
together, since Vietnam has quite limited trade with them. Table 6-1 presents the final
aggregation for 20 countries and regions from the GTAP database for this study.
85
Table 6-1: GTAP Regional Concordance
No New region Old countries/regions
1 USA United States of America
2 EU25 (European Union 25)
Austria, Belgium, Denmark, Finland, France, Germany, United Kingdom, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, Cyprus, Czech Republic, Hungary, Malta, Poland, Slovakia, Slovenia, Estonia, Latvia, Lithuania
3 JPN Japan
4 CHN China, Hong Kong
5 VNM Vietnam
6 IDN Indonesia
7 MYS Malaysia
8 PHL Philippines
9 THA Thailand
10 KOR Korea
11 IND India
12 XEA (Rest of East Asia)
Taiwan, Rest of East Asia
13 XSE (Rest of South East Asia)
Cambodia, Singapore, Rest of Southeast Asia
14 XSA (Rest of South Asia)
Bangladesh, Pakistan, Sri Lanka, Rest of South Asia
15 AUS Australia
16 ODV (Other developed countries)
New Zealand, Canada, Rest of North America, Switzerland, Rest of EFTA
17 LAM (Latin America)
Mexico, Bolivia, Colombia, Ecuador, Peru, Venezuela, Argentina, Brazil, Chile, Paraguay, Uruguay, Rest of South America, Central America, Rest of Free Trade Area of Americas, Rest of the Caribbean
18 AFR (Africa) Egypt, Morocco, Tunisia, Rest of North Africa, Botswana, South Africa, Rest of South African Customs, Malawi, Mauritius, Mozambique, Tanzania, Zambia, Zimbabwe, Rest of Southern African Development Community, Madagascar, Nigeria, Senegal, Uganda, Rest of Sub-Saharan Africa
19 CEE (Central and East Europe)
Rest of Europe, Albania, Bulgaria, Croatia, Romania
20 ROW (Rest of the world)
Rest of Oceania, Russian Federation, Rest of Former Soviet Union, Turkey, Iran, Islamic Republic of, Rest of Middle East
Source: Aggregate from GTAP v 6.2
86
GTAP includes 57 sectors, covering all industries and commodities of the economy.
This can be classified into three broad sectors, namely, food and agriculture,
manufactures, and services. Food and agriculture includes 20 sectors, such as rice,
vegetables, fruit, cereal, livestock and meat products, etc., which together comprise 8
percent of world sales value. The 22 manufacturing sectors contribute 31 percent, and
the high-value services sectors comprise 61 percent of world sales value (Dimaranan
2006). For the purpose of this research, these 57 sectors are firstly aggregated into 18
commodity groups, with 7 in agricultural and food commodities and 11 in
manufacturing and services.
Since the study is interested in the impacts of trade liberalisation on small households
who raise pig and chicken as the main source of income, an examination of how trade
liberalisation affects the individual pig and chicken sectors at a national level is
necessary, and the impact of price changes of these two commodities to the household
production at the micro level is considered. Disaggregation of pork and poultry in the
meat products group is also needed, however separating pork and poultry is too
complicated since the required data on bilateral trade is not available in some countries,
so they remain aggregated in one group. The software SplitCom is used to separate live
pig and poultry sectors from the livestock group used in the standard GTAP framework,
the method of which is described below.
6.2.3. Using SplitCom and Introducing New Sectors into GTAP Database
Generating new sectors for live pig and live chicken to introduce into GTAP database is
based on extracting data from the sources of UN Comtrade, International Statistics,
WITS, FAOStat, and SAMs of countries, and using the SplitCom program, developed
by Mark Horridge, Centre of Policy Studies (Monash University) in 2005. The process
takes the following steps:
Step 1: Aggregating 18 sectors and 20 regions/countries from 57 individual sectors and
96 countries of GTAP version 6.2. The sectoral aggregation attempts to split sectors
with significant protection, such as textiles and apparel, manufactures, and electronics,
while grouping some sectors with similar characteristics in production and approximate
protection level together. In the initial aggregation, of the 11 sectors belonging to
87
manufacturing and services, four are process sectors that provide inputs from
agriculture.
Step 2: Applying SplitCom to disaggregate the live animals sector into three new
sectors: live pig, live poultry and live other. The program works with three sub-folders:
Input, Work, and Output. The original GTAP database and its associated files
(basedata.har, default.prn, and sets.har) are copied into an input folder. The files
SplitSec.har and UserWgt.har are created automatically by SplitCom, once the
commands of creating new split and new database are set up. However, in the primary
action, SplitCom weights the three new commodities equally, and the user needs to
supply the desired weights by adding new headers26 TWGT, RWGT, CWGT, and
XWGT to the userwgt.har file. To do this, data on bilateral trade between
countries/regions in 2001 are taken from UN Comtrade, International Statistics, and
WITS. Data on production and consumption of pig, chicken and other animals comes
from FAOStat, and Social Accounting Matrices (SAMs) of various countries27; and
assumptions are made that some countries with similar economic conditions have
similar production activity. Hence, SplitCom is used with updated userwgt.har file to
get the final split. Finally, the new expanded GTAP database is stored in the Output
folder of the SplitCom, ready for loading to run the GTAP model.
The database is now disaggregated into 20 commodity groups (Table 6-2).
26 These headers are found in the file named nuweght.har in the work folder of the SplitCom. These headers are originally designed to introduce user weights for sales, costs, self uses and trade data into SplitCom. 27 SAMs of countries are used from source of International Food Policy Research Institute (IFPRI)
88
Table 6-2: GTAP Sectoral Concordance
No New sector Old sectors
1 RIC (Paddy and processed rice) Paddy rice; Processed rice
2 VF (Vegetable and fruit) Vegetables, fruit, nuts
15 CRP (Chemicals, rubber, plastic) Chemicals, rubber and plastic products
16 TXT (Textile and apparel) Textiles; Wearing apparel; Leather products
17 MAN (Manufactures) Wood products; Paper products, publishing; Mineral products nec; Ferrous metals; Metals nec; Metal products; Motor vehicles and parts; Transport equipment nec; Machinery and equipment nec; Manufactures nec
18 ELE (Electronic) Electronic equipment
19 TCN (Transport, communication) Transport nec; Sea transport; Air transport; Communication
20 SVC (Services) Electricity; Gas manufacture, distribution; Water; Construction; Trade; Financial services nec; Insurance; Business services nec; Recreation and other services; PubAdmin/Defence/Health/Educat; Dwellings
Source: aggregate from GTAP v 6.2 and disaggregate new sectors using SplitCom
89
6.2.4. Trade Scenarios of Trade Liberalisation Simulation
In this study, several scenarios are explored using the GTAP model:
(1) Unilateral Vietnam trade liberalisation; Vietnam completely removes all of its trade
taxes.
(2) Vietnam and all other ASEAN countries fully eliminate all tariff and subsidies, and
apply a free trade area in ASEAN. The trade barriers among other countries stay the
same.
(3) Extension of AFTA by expanding free trade area to include Japan, Korea and China.
In this scenario, China is a competitor of many ASEAN economies, with its large, low-
cost labour force, and it may have some impacts for adjustment in the economies of
ASEAN in general and Vietnam in particular.
Bilateral trade agreements are relatively easy to negotiate but are of limited value if the
two economies are similar. For developing countries, agreements with large developed
countries were generally considered the most beneficial. Two agreement options were
investigated:
(4) Between Vietnam and the USA
(5) Between Vietnam and EU.
The reasons for choosing USA and EU is that both are large economies; the USA has
the potential to export maize and soybean to Vietnam which may affect the livestock
sector, and both USA and EU are large trade partners of Vietnam in apparel and textile
trading, which can absorb more labour from rural areas once they have opportunity to
develop.
Multilateral liberalisation refers to a potential WTO agreement. To simplify the analysis
the sixth scenario is:
(6) A 50 percent reduction in tariffs, exports subsidies and domestic support for all
regions.
90
(7) Full global liberalisation, without any trade barriers among countries that indicate
the potential gains from trade liberalisation and the opportunity cost of not liberalising
fully.
Even though the simulations in these scenarios do not fully describe real trade
liberalisation, they help to provide a closer representation of prospects that Vietnam has
to face in the near future, when all regulations of the agreements are confirmed.
Table 6-3: Alternative Trade Scenarios28
Scenarios Title Change in tariffs
(1) Unilateral Vietnam unilateral trade
liberalisation
VNM exempts 100% of import tax
for all countries
(2) AFTA Free trade area in ASEAN ASEAN countries exempt 100%
import tax to each others
(3) AFTA+3 Free trade area in ASEAN
plus China, Japan and Korea
ASEAN countries and JPN, KOR,
CHN exempt 100% import tax to
each others
(4) VNM-USA Bilateral trade between VNM
and USA
VNM and USA exempt 100% on
trade between two countries
(5) VNM-EU Bilateral trade between VNM
and EU
VNM and EU25 exempt 100% on
trade between two regions
(6) Multilateral Multilateral trade
liberalisation
All countries exempt 50% of import
tax to the others
(7) Global Free trade over the world No import tax over the world
6.3. Simulations with Different Modifications of GTAP Closures
The simulation model is based on the GTAP database version 6.2, aggregated into 20
regions and 20 commodities, with all macroeconomic, trade and protection data
referring to the common reference year of 2001, which is the most updated version at
the time of this study. The standard model is a comparative static model. Therefore, if
28 Implementing the detail and complexity of actual trade policy within a model requires a number of simplifying assumptions to be made i.e. rules of origin introduce considerable complexity, and representation of details of tariff schedules. However, the representation of trade policy used here is deemed sufficient to give an insight into the overall impact of liberalization.
91
all markets in the multi-region model are balanced (market-clearing condition), all firms
earn zero profit (zero-profit condition) and all households are on their budget constraint
line, then global savings must equal global investment and with this Walras’ Law is
satisfied. It implies that equilibrium is obtained for a system of n equations if n-1
equations are solved (Hertel & Tsigas 1997). When an exogenous shock is introduced,
the model works out a new equilibrium in all markets and determines new values for the
endogenous variables (Francois et al. 2003). The default solution method for the GTAP
model is Gragg’s method where the model is solved several times with an increasingly
fine grid until convergence is achieved.
Due to the introduction of new sectors of live pig, live poultry and other live animals to
the GTAP database, import tariffs for these commodities need to be adjusted to align
with the real situation. The changes to the tariffs contained in the database were
implemented as a 'pre-experiment' using 'Alter tax'29. Then, the updated post-simulation
database is used as a starting point for our subsequent policy experiments. Table 6-4
presents outputs and trade data of sectors of Vietnam in base year.
One of the questions that often arises when using a comparative static model is how to
“close” the model. Since the study tries to link a trade GTAP model at macro level with
a household model at a micro level, where labour allocation and wages are important
factors affecting welfare and behaviour of the household, the labour issue is especially
considered.
In the GTAP model, relating to labour market closure, the two most common
alternatives involve either assuming flexible wages and full employment or fixed real
wages and unemployment. However, GTAP also gives users a wide range of closure
options. In this study, the GTAP model is applied with some different closures for the
labour market.
29 Altertax runs a simulation that shocks tax rates to their desired value. It uses a special closure and a special parameter file to ensure that the rate-changing simulation leaves other cost and sales shares as little changed as possible.
92
Table 6-4: Vietnam’s Output and Trade Flows, 2001 (mill. USD)
Sector Output Export Import
Paddy and processed rice 6467 374 17
Vegetable and fruit 1902 257 71
Other crops 1541 810 225
Live Pig 881 2 5
LivePoultry 434 0 7
LiveOther 545 62 29
Pork, poultry, and other meats 168 33 20
Beef and sheep meats 22 0 7
Fishing 1541 49 6
Oilseed and vegetable oil 93 45 90
Processed food 2895 1365 374
Beverages and tobacco 1222 22 395
Milk and dairy products 241 2 239
Natural res, petroleum product 3703 2346 1692
Chemical, rubber, plastic 2938 495 2796
Textile and apparel 7994 4746 1848
Manufactures 10203 2313 6780
Electronic 528 446 1002
Transport, communication 2143 534 2546
Services 26763 1552 6997
Total 72223 15453 25145
Source: GTAP v.6.2
The first implication of the factor market clearing conditions (Closure A), an
assumption of an archetypal free market model is specified, with full employment and
93
full factor mobility in all factor markets. This is the standard closure, which is often
applied in the GTAP model.
Since there is an acknowledgement of unemployment in the world, the presumption of
full employment in all economies is questionable. The second alternative assumption on
the labour market (Closure B), in which excess supplies of unskilled labour in
developing countries exist, but other factors are still fully mobilised, is made. When
there is unemployment, the real wage is held constant and the supply of unskilled labour
adjusts following a policy shock.
The different assumptions on the employment situation will result in different values for
employment, wages, and prices, and thus different final welfare results, an assumption
on the level of unemployment needs to be considered if possible. The Closure C is made
based on an assumption that there is an unemployment situation that exists in Vietnam,
but the unskilled labour supply is not unlimited, and at some point real wages rise.
In the next section, the implications of alternative labour market clearing conditions are
investigated and the results of trade liberalisation scenarios presented.
6.3.1 Closure A - Standard GTAP Closure
Closure A is often used in simulations of the GTAP model. In the standard process,
prices, quantities of all non-endowment commodities, and regional incomes are
endogenous variables, conversely, policy variables, technical change variables, and
population are exogenous variables to the model.
In this closure, labour supply in all markets is fixed exogenously and, according to the
profit-maximising behaviour of the productive sector, labour demand is determined.
Since the fundamental property of a neoclassical general equilibrium model is the
concept of flow equilibrium in product and factor markets, labour demand is derived
endogenously as a function of output price and the wage, and is equalised to labour
supply. This is a fundamental requirement for market clearing. In the labour market the
variable responsible for equilibrating is the average wage rate which varies in order to
achieve economy-wide full employment. In that context, the labour market in Vietnam
can be illustrated as in Figure 6-1. The economy-wide supply of labour is fixed at L*.
Assume that there is a policy shock, which makes the demand for labour increase from
94
D1 to D2, hence the equilibrium in the labour market will move from L1 to L2, and the
price of labour will increase from P1 to P2.
Figure 6-1: Labour Market under GTAP Standard Closure - Closure A
The results of the GTAP simulations are presented in broad categories, such as
production, price, trade, and welfare effects, etc. Table 6-5 provides an overview of the
output effects of the various scenarios with simulations under Closure A.
Significant adjustments in production are observed following trade liberalisation. In
most scenarios, rice, pig and poultry outputs increase, or at least stay the same30.
Textile, electronic, and service sectors experience positive production effects31.
Meanwhile, manufacturing, meats and processed food sectors reduce production. Of
interest is the difference in the impacts of Unilateral and Regional or Multilateral
scenarios on production. In the Unilateral scenario there is no expansion in export
markets, as countries other than Vietnam do not reduce their tariffs. Most sectors
contract. With liberalisation in AFTA there is an increase in Vietnamese production of
oilseeds (OSO) whereas EU liberalisation led to an increase in Vietnamese production
of livestock. This contributes to a limited flow of labour into electronics and services.
30 Except paddy sector in Unilateral scenario, and live pig and live poultry in AFTA. 31 Except case of Electronic under the VNM-EU scenario.
D2
D1
L1
L2
S PL
QL
P2
P1
0 L*
95
Table 6-5: Initial Values (mill.USD) and Percentage Changes in Vietnamese Outputs under Alternative GTAP Scenarios* with Closure A
Transport, communication -0.25 0.1 0.32 0.14 0.82 0.18 0.59
Services 0.11 0.27 2.01 0.43 2.26 1.15 2.85
Source: GTAP simulations
103
6.3.3 Closure C - Modified Non-Standard Closure
In Closure C the wage of unskilled labour in all developing countries, except Vietnam,
is fixed. It is assumed that there is a limit on the maximum increase in unskilled labour
in Vietnam of R percent. If a trade liberalisation scenario results in a simulated increase
in labour of R percent, then the increase in labour is fixed at R percent and wages rise to
get the market in equilibrium. The reason for doing so is based on an assumption that
there is an unemployment situation in Vietnam and the unemployed are willing to work
at the current wage level, however, the number cannot increase over the total number of
unemployed of the society. Hence, the market for unskilled labour in Vietnam is
illustrated in Figure 6-3.
Figure 6-3: Unskilled Labour Market in Vietnam under Closure C
When labour demand is lower than Q1(1+R%), equilibrium of the labour market is at L1,
where Q1 unskilled labour is willing to work at wage level P*, as is the case for Closure
B. Labour would agree to work at the current market price until labour supply reaches
Q1(1+R%), after which, if demand for labour increases the new equilibrium on the
market would be L2, where the new wage is P2, higher than the current wage.
6.3.3.1. Estimation of R
Applying Closure B, fixing wages of unskilled labour in all developing countries, and
allowing labour demand in Vietnam to increase in all scenario simulations led to a
maximum increase in demand of approximately 25 percent, if global liberalisation
QL
D2
D1
L1
L2
P*
P2
PL
Q1(1+R%) 0
S
Q1
104
happens (Table 6-7). However, the issue is whether or not society can supply that much
labour at the current wage.
Statistics from Vietnam’s General Statistics Office show that in 200132, the
unemployment rate in Vietnam’s urban areas was 6.28 percent, and under-employment
in the rural areas was 25.74 percent (GSO 2008). However, to move to complete full
employment is impossible. “Full employment” has come to mean an unemployment rate
close to 6 percent; it would be difficult to reduce unemployment below this rate, even in
a highly developed economy (Baumol & Blinder 1988). Based on that argument, with
about 64 percent of the population working in rural areas (2001), increases in the labour
force supplied to the economy is about 12 percent33 at the current wage. Assuming that
the ideal solution is that about 12 percent of the population can find a job (limitations of
information accession, transportation, skill, etc., are no obstacle). With that assumption,
Closure C now fixes the maximum increase in labour supply of Vietnam at 12 percent.
When the demand for labour increases over that level, wages would increase since the
elasticity of labour supply is perfectly inelastic (Figure 6-3).
In this study, a rate of 12 percent more than the baseline labour supply is chosen as the
upper limit of the change in unskilled labour supply (R) in Vietnam at the current wage.
This means scenarios (1) (3) and (7) above need to be resolved using Closure C, as all
have increases in labour exceeding 12 percent (see Table 6-7). With closure C, fixing
the potential increase in unskilled labour supply in Vietnam at 112 percent compared to
the baseline, and allowing the wage of labour to vary34, changes in market prices under
trade scenario simulations are found in Table 6-10.
Scenarios (1), (3) and (7) have market wage increases of 0.31, 4.44, and 8.23 percent
respectively, corresponding with increases in labour demand over 12 percent in the
three scenarios with Closure B (see Table 6-7).
32 The study is based on reports of unemployment and underemployment in year 2001 to match with the economy situation of the base year in the GTAP model. 33 Assume that unemployment rate cannot go lower than 6 percent and unemployed people are prepared to work at the current wage level, the maximum labour can absorb into labour market = 100% - %labour in work - 6% limited = 0.64* (25.74% - 6%) + 0.36*(6.28% - 6%) = 12.73%. 34 For other developing countries and LDCs, wages of unskilled labour are fixed, as in Closure B
105
Table 6-10: Percentage Changes in Vietnamese Market Prices under Alternative
GTAP Scenarios with Closure C (R=12%)
Factors/ Commodities Uni-
lateral
(1)
AFTA
(2)
AFTA
+3
(3)
VNM-
USA
(4)
VNM-
EU
(5)
Multi-
lateral
(6)
Global
(7)
Land 2.26 9.36 20.5 0.61 5.62 8.81 15.7
Unskilled labour 0.31 -1.03 4.44 0.24 2 -0.23 8.23
Skilled labour 13.54 4.5 17.3 1.71 8.21 10.07 22.2
Transport, communication -0.11 0.1 0.47 0.14 0.82 0.18 0.88
Services 0.59 0.27 2.5 0.43 2.26 1.15 3.84
Source: GTAP simulations
108
6.3.4 Alternative Labour Assumptions and Real Wages
An interesting finding is the difference in the simulated changes in real wages for the
seven trade scenarios, using different assumptions about the possible expansion in
labour supply. In order to graphically present these differences, the GTAP model was
run with different closures based on the different levels of unskilled labour supply in
Vietnam (with the value of R% varying over 0, 8, 12 percent, to unlimited supply in
Closure B). Note that these are real wages, that is, the market wage deflated by the
consumer price index (CPI), and hence changes at a 12 percent change in unskilled
labour are different to changes in market wage reported in Table 6-10. The assumption
within GTAP under Closure C is that it is the real wage that remains constant.
0
4
8
12
16
20
0 4 8 12 16 20 24 28Maximum possible increase in labour supply (%)
perc
enta
ge c
hang
e of
rea
l wag
e co
mpa
re
with
bas
elin
e
Uni AFTA AFTA+3 VNM-USA VNM-EU Multi Glob
Figure 6-4: Changes in Real Wages under Scenario Simulations with Different Possible
Maximum Labour Supplies
Source: GTAP simulations
If there is zero expansion in possible labour (equivalent to Closure A) then Figure 6-4
shows increases in real wages ranging from 1.02 to 16.89 percent. If there are potential
increases in labour supply then the change in real wages fall but remains positive as
long as the labour constraint is binding (Closure C). Once the maximum possible labour
increase exceeds the increase in labour demand induced by a liberalisation scenario,
then the constraint is not binding, and real wages do not change (Closure B). The figure
shows the maximum change in labour supply at the point of transition from Closure C
to B for each scenario. At the maximum level of labour change used in this study (12
109
percent) only scenarios (1), (3) and (7) hit the constraint and have increased real wages.
One advantage of the figure is that it identifies the simulated increase in real wages for
any level of labour constraint, for each scenario.
6.3.5 Changes of Vietnam’s Welfare under Alternative Closures
The results of the GTAP simulations are presented in some broad categories. The
previous section presented changes in production, price, and trade of Vietnam through
scenario simulations under alternative closures. Another important indicator is the
welfare effect of trade liberalisation on the country at the macro level. In the GTAP
model, the welfare indicator summarises policy changes by incorporating changes in
consumption, production, price and trade flows, and is measured using equivalent
variation35 (EV) in income.
Figure 6-5 presents the changes of Vietnam’s welfare in trade simulation scenarios
under different closures. It shows significant improvement in welfare under different
closures, but with some variations.
0
1000
2000
3000
4000
Unilateral AFTA AFTA+3 VNM-USA VNM-EU Multilateral Global
Scenarios
mil.
US
D
closure A closure B closure C
Figure 6-5: Changes in Welfare under Alternative Trade Scenarios with Different
Closures
Source: GTAP simulations
With Closures B and C, creating more jobs and improving the employment situation or
unemployment threats in specific sectors under trade liberalisation scenarios enhance 35 EV represents the money-metric equivalent to the utility change brought about by a change in prices. It measures the amount of money needed to be taken away from the consumer before the price change to leave her as well off as she would be after the change in prices.
110
welfare gains of Vietnam. Different welfare changes under scenarios (1), (3) and (7)
between Closure B and C are due to the binding upper limit of unskilled labour supply.
Under scenarios (2), (4), (5), and (6), there are no differences in welfare, since demands
for unskilled labour do not reach the limiting 12 percent increase. This improvement in
welfare is attributed to a significant contribution of job creation for unskilled labourers
(Table 6-12). Further analysis of the basis for the differences in welfare is conducted in
the following chapter.
Table 6-12: Number of Jobs Created for Unskilled Labour and its Contribution to
Total Social Welfare in Vietnam under Trade Scenarios (percentage)
Number of jobs for unskilled labour
created compare with baseline
Contribution to total social
welfare
Scenarios
Closure B Closure C Closure B Closure C
Unilateral 17.97 12 20.37 13.79
AFTA 5.84 5.84 6.63 6.63
AFTA+3 18.27 12 20.76 13.78
VNM-USA 1.43 1.43 1.63 1.63
VNM-EU 6.06 6.06 6.89 6.89
Multilateral 10.13 10.13 11.5 11.5
Global 24.71 12 28.05 14.05
Source: GTAP simulations
Choosing the closure of the GTAP model has a significant effect on defining the scope
and changes of economies in general and of commodity/factor prices in particular.
However, it should be kept in mind that simulations are based on a database that
incorporates pre-existing distortions and these interact with policy shocks and produce
second-best effects. Another important point is to interpret numbers with an appropriate
attitude as certain numbers should not be taken literally, but they suggest certain
directions e.g. big returns to simulated policies (Daude 2004).
111
6.4. Summary
This chapter briefly presents progress in Vietnam and its main commitments in its
integration into the world economy. Even though some negotiations are still in progress,
prospects of near future trade liberalisation and its impacts on the economy are needed.
The computable general equilibrium model has been used widely to investigate the
implications of trade liberalisation in the world, and in some sectors in Vietnam,
previously. The current research uses the multi-country trade model GTAP to assess the
impacts of trade liberalisation at the national level and to household level. Trade
agreements between Vietnam and trade partners are very complicated, with variation
between fields. Presenting all agreements in a model simulation is not possible;
however, simulation scenarios are designed to represent real situations closely.
Since the study is interested in impacts of trade liberalisation on households raising pigs
and poultry, three new sectors of live pig, poultry and other animals were introduced to
the GTAP database using SplitCom software. This means price signals of these
commodities are better captured, hence reactions in household behaviours to the
changes are expected to be measured more accurately when linking trade and household
models together.
The labour market issue in the GTAP model is also considered. Acknowledgment of the
employment and underemployment situation in the process of liberalisation has an
important role in policy simulation, especially in the case of developing countries in
general and Vietnam in particular. This helps the model to be more useful in analysing
economic impacts of policy changes, since it presents a more accurate situation of
resource allocation and use of endowments. In this study, Closure C is chosen based on
the unemployment and under employment situation in Vietnam. With this closure, the
Vietnamese labour force can increase labour supply to the economy by 1.12 times in
comparison with the baseline, without causing any wage increase. After that point,
when the demand for labour increases, wages would increase according to the normal
rules of closure.
Regarding impacts of trade liberalisation on Vietnam at the macro level, the largest
benefit would be if full liberalisation happens. The voluntary trade liberalisation of
112
Vietnam in a unilateral liberalisation would benefit itself without negotiating with
others. However, market access benefits are limited because other countries do not open
their markets.
For the purpose of assessing impacts of trade liberalisation on the household, the next
chapter links GTAP results with household models to examine how small livestock
households react to price changes induced by trade liberalisation.
113
CHAPTER 7 : IMPACTS OF TRADE LIBERALISATION
ON SMALL LIVESTOCK HOUSEHOLDS - LINKAGES
BETWEEN TRADE AND HOUSEHOLD MODELS
The previous chapter presented the trade liberalisation process of Vietnam and
simulated impacts at the national level using the GTAP model. In this chapter, to
examine impacts of trade liberalisation on Vietnam’s small household livestock sector,
price changes of trade scenarios from simulations of the GTAP model are linked with
the household model from Chapter 5. Changes in household welfare and responses to
price signals in terms of substitution between commodities in consumption and
production, and labour allocation in the household, are reported. However, the analysis
only examines one-way effects of trade liberalisation on households, and not their
influence on trade. Therefore, feedback from households to the international system is
not considered. Policy conclusions on opportunities and threats from trade liberalisation
for small livestock producers under alternative assumed liberalisation scenarios and
labour market conditions are also presented.
7.1. Price Changes in the Trade Scenario Simulations
The impacts of trade liberalisation on Vietnam’s welfare, production, and trade at a
national level, were reported in the previous chapter. This chapter uses price changes for
both consumption and production as a result of trade liberalisation scenarios of the
GTAP macro model to assess impacts at the household level. To determine
liberalisation effects on consumption and production, as well as reactions in labour
allocation, certain assumptions are made to align different sectors or commodities of
GTAP with those in the household model.
Changes in consumption prices (pp) from trade scenario simulations in the GTAP
model are matched with the consumption side in the household model. Consumption
commodities at the household level are more detailed than aggregated sectors in the
GTAP macro level, however commodities in both models are aggregated into groups
which can be matched as best as possible. In the main food group, rice consumption
114
corresponds to the GTAP categories of processed rice and paddy rice (RIC),
consumption of fruits and vegetables are directly comparable with the fruit and
vegetable sector (VF), fish and shrimp consumption at the household level match the
fishing group (FSH), pork and chicken meat connect to OMT (pork and poultry meats)
and other meats correspond to CMT (group of beef, sheep, and other meats),
respectively in the GTAP. All other foods in the household model are summarised as
one common commodity namely “other food”, which matches various processed foods
in the GTAP model under oilseed and vegetable oil (OSO), processed food (OFD),
beverages and tobacco (B_T), and milk and dairy products (MLK). Household
consumption of all industrial commodities such as clothing, electronic equipment or
other expenditures on services, for example eating out, are grouped together to match
categories of textile and apparel (TXT), manufacture goods (MAN), electronic (ELE),
transport and communication (TCN) and services (SVC). Price changes in broad groups
such as “other food” or “industrial and other expenditures” in the household model are
therefore calculated as weighted changes in prices of corresponding commodities in the
GTAP model.
Similarly, inputs and outputs of a household’s production are suitably related to
subsumed sectors in the GTAP model. Changes in market price (pm) of commodities
from trade scenario simulations in the GTAP model are used in the household model.
For example, price changes for chemical fertiliser and pesticides for rice cultivation are
limited to average price changes of the chemical, rubber, and plastic (CRP) group in the
GTAP.
One of the main inputs of livestock production is feed. According to statistics, in
Vietnam feed often accounts for 65 to 70 percent of the total cost of raising livestock in
small households (IFPRI 2001, Lapar et al. 2003, Nguyen & Tran 2005). To
disaggregate industrial feed as an individual sector in the GTAP model is not possible,
so to estimate price changes in livestock feed due to trade liberalisation, an assumption
on feed contents is needed36. Hence, changes in the price of raw and industrial feeds for
36 Industrial feed for chicken often contains 60 percent maize, 20 percent soybean, 5 percent fish, and 15 percent other ingredients. Industrial feed for pig often includes 40 percent maize, 15 percent soybean, 7 percent meat and milled bone, 5 percent fish, 10 percent rice bran, and 23 percent other ingredients. Different kinds of feed have different ingredients; these numbers are an estimate based on author’s interview with feed livestock companies in Vietnam, in August 2007, during a CARD project (Collaboration for Agricultural and Rural Development Program) namely 'Developing a Strategy for
115
raising livestock are calculated based on changes in prices of feed ingredients such as
maize, soybean, milled bone, fish and others, all subgroups in GTAP. Table 7-1
presents the aggregation and/or splitting sector/commodities available in GTAP to be
matched with those of the household model.
Table 7-1: Matching between GTAP Sectors and Endowments in this Study and
their Concordance with Commodities and Goods in Vietnam’s Household Models
In household model Matched GTAP sectors and factors
Rice, Paddy, and Seeding RIC: Paddy and processed rice
Live pig Live Pig
Live chicken Live Poultry
Chemical fertiliser and
Pesticide
CRP: Chemical, rubber, plastic
Pork and chicken meat OMT: Pork and poultry meats
Fish FSH: Fishing
Vegetable and fruit VF: Vegetable and fruit
Other meats CMT: Beef, sheep, and other meats
Other foods OSO: Oilseed and vegetable oil, OFD: Processed food,
B_T: Beverages and tobacco, MLK: Milk and dairy products
In the household model, as mentioned in Chapter 3, total land for production is assumed
to be fixed, therefore changes in land prices due to trade liberalisation are assumed to
have no effect on production behaviour of the household. Other production factors such
as capital and natural resources are also assumed to have no impact on decision making
of the household in production. In order to simplify the model, labour in the household
is categorised as unskilled and assumed to be equivalent to unskilled labour in the Enhancing the Competitiveness of Rural Small and Medium Enterprises in the Agro-Food Chain: the Case of Animal Feed' 030/06VIE. Raw feed is assumed to include 50 percent rice bran and 50 percent vegetables.
116
GTAP model. In fact, this assumption is probably close to the truth: most agricultural
labour receives a low wage irrespective of education level.
7.2. Price and Wage Transmission
Four household models are constructed in this study to represent households in four
different regions: (1) Red River Delta, one of two main important deltas in the country
with favourable conditions for rice cultivation, however land cultivation is limited and
population density is the highest; (2) Northern Uplands (includes North East (NE) and
North West (NW)), where agricultural production is limited by ecological and economic
conditions; (3) the Centre, including North Central Coast, South Central Coast and
Central Highland; and (4) the South, including Mekong River Delta and North East
South, is similar to the Red River Delta, where urbanisation and industrialisation have
recently expanded very quickly. The South also has the highest commercialisation of
rice and livestock production. Different agro-ecological areas with different economic
conditions may affect the farming system and activities of the household in production
in different ways due to trade liberalisation.
Simulation results of trade liberalisation in the GTAP model only generate average
price changes of consumption commodities and production inputs/outputs at the
national level. However, it is possible that these national changes will translate into
differences in price changes of those commodities/inputs in different regions. This
section analyses differences in price transmission of some commodities/inputs between
each region from the simple average price at the national level.
Data used in this study include monthly prices of main commodities from 24 provinces
and cities in Vietnam, such as rice, paddy rice, live pig, live chicken, beef, maize,
soybean, cassava, orange, tomato and fertiliser. Data were collected by the Vietnamese
Institute for Market and Prices from 1997 to 2006 (120 observations) for all
commodities except rice, pig, beef and fertiliser which were collected from 2001 to
2004 (48 observations). The average price in the Red River Delta is calculated from five
provinces: Ninh Binh, Ha Tay, Hung Yen, Hai Duong and Ha Noi; Northern Uplands
(NE+NW) from six markets: Bac Giang, Thai Nguyen, Yen Bai, Tuyen Quang, Lai
Chau, and Son La; the Central region from six provinces: Thanh Hoa, Nghe An, Hue,
Da Nang, Quang Ngai, Dac Lac; and the South from seven markets including Dong Nai,
117
Ba Ria, Binh Duong, Tien Giang, Ben Tre, Can Tho, and Ho Chi Minh city. The
average price at national level is calculated from all 24 provinces.
A simple econometric method for measuring price integration is used. The functions
take the functional forms as follows:
iiiri Pp loglog
where
Pi is average price of ith commodity at national level
rip is price of ith commodity at region r, which includes RRD, NE+NW, the Central,
and the South
i include rice, paddy rice, live pig, live chicken, beef, maize, soybean, cassava, orange,
tomato and fertiliser.
This is the simplest way to measure spatial price relationships between two markets and
the transmission coefficients, βi are interpreted as a measure of how closely price
movements of a commodity at different markets are linked (Luu 2003).
The regression results show that the coefficients βi of most commodities in the four
regions fluctuate around 1 (range 0.712 to 1.33). However, testing the result shows 28
of the 44 coefficients are not significantly different from 1 at 95 percent confidence
level. Another five coefficients are different within 10 percent between regional and
national price (see Appendix A7.1). The results also show that there is no consistent
pattern of transmission of price for any region. In order to simplify the calculation, the
study assumes that if the price of a commodity changes by 1 percent at national level,
the price at regional market level also changes approximately by 1 percent.37
A similar question arises for changes in labour wages at the regional and national levels.
Since there is no annual labour and salary survey, time series data of labour wages in
Vietnam are not available. The main information sources are surveys on Vietnam living
standards - one in 1998 (VLSS 1998) and two other recent household living standard 37 The impacts of trade liberalization at the domestic level will obviously be affected by assumptions about the degree of price transmission. However, given the relatively weak statistical evidence for deviations from unity, that is applied here. Further research on identifying more precise estimates of price transmission in Vietnam are warranted.
118
surveys in 2002 and 2004 (VHLSS 2002 and 2004). Table 7-2 presents average wages
and annual growth rate of wages in the four regions.
Table 7-2: Average Wage of Labour at Current Price and Annual Growth Rate
Average wage ('000 VND) Annual growth rate of wage (%)Region
1998 2002 2004 1998–2002 2002–2004
RRD 411.58 652.96 773.70 1.122 1.089
NE+NW 367.82 642.90 728.37 1.150 1.064
Central 366.14 515.99 661.07 1.090 1.132
South 561.77 729.46 872.00 1.067 1.093
Whole country 481.07 666.97 791.79 1.085 1.090
Source: calculation based on data of VLSS1998, VHLSS 2002, and VHLSS2004
Table 7-2 shows similarities in annual growth rates of wages between each region and
national level over time. Assuming from 1998 to 2004 there were some macro policies
as well as changes in the economy that impacted on labour wage levels across the whole
country then it can be concluded that a change in labour wages will be approximately
the same in all regions of the country.
Therefore, changes in commodity prices and labour wages due to trade liberalisation
would be applied to all regional household models at the same percentage changes as
predicted by the GTAP national model.
7.3. Impacts of Trade Liberalisation on Households
Having generated price changes for trade liberalisation scenarios from the GTAP model,
and linking them to the household model, this section analyses how a household
responds to trade scenarios by changes in behaviour. Changes in commodity and factor
markets affect household welfare because adjustments are transmitted through changes
in production, consumption, and labour allocation, hence changes in utility.
Changes in prices are drawn from simulations of the GTAP model, using Closure C,
where the maximum increase in unskilled labour supply in Vietnam is fixed at 12
119
percent. Hence, a trade scenario can absorb labour up to a level equal to 112 percent of
the baseline without causing a wage increase, but wage increases occur when labour
demand goes beyond that point.
In calculating welfare impacts on the household using the household model, a measure
of compensating variation (CV) in income is applied. That is, the amount of money
which, when taken away from the household after price and income changes, leaves the
household with the same utility as before the change (Varian 1996):
000101 ,, upeupeYYCV
where: Y1 is income after the price change from p0 to p1, Y0 is baseline income and
expenditure function e(p,u) is the minimum income necessary to reach the level of
utility u at given price p.
Table 7-3 summarises the change in utility (measured by compensating variation) for
each region for each scenario. The results show that the biggest changes in welfare in
three of the regions occur under scenario 7, global liberalisation, although the NE+NW
are similarly affected by both this and scenario 3, an expanded AFTA arrangement.
Table 7-3: Welfare Changes in Households in Different Regions under Alternative
Liberalisations Compare with Baseline (percentage)
Welfare
change of
household
Uni-
lateral
(1)
AFTA
(2)
AFTA
+3
(3)
VNM-
USA
(4)
VNM-
EU
(5)
Multi-
lateral
(6)
Global
(7)
RRD 6.41 4.63 13.82 0.3 -0.12 3.26 19.13
NE+NW 5.90 -0.03 12.18 -- 4.76 5.62 11.56
Central 6.11 4.22 11.86 -- 3.36 3.50 15.80
South 6.85 7.29 19.56 -- 5.06 6.78 22.31
Source: Household model simulation
-- Negligible price changes due to trade liberalisation predicted, welfare change can be considered equal
to zero
120
The change in welfare of the household in the South in each scenario is always the
biggest, in (7), its welfare increases by 22 percent compared with the baseline, with a
CV of about 3 million VND.
Underlying each simulation is a large number of changes in specific commodity
production and demand. It is not possible to review, in detail, this information for all
regions, for all trade scenarios. Therefore, the South model is used as an example for a
more detailed analysis of the changes that occur in production, consumption and labour
as a result of the global liberalization. Later in the chapter, more general overviews of
the changes that occur in all regions are presented.
7.3.1. Impacts of Global Trade Liberalisation on Households in the South
On the production side, in response to increases in market output prices, the household
in the South increases agricultural production under the global trade scenario. Figure 7-
1 shows increases of about 7 and 10 percent in pig and chicken production, respectively,
compared with the base year, when live pig and chicken prices increase by 9 and 11.8
percent, respectively.
0
10
20
30
Rice Pig Chicken
quan
tity
(kg)
0%
4%
8%
12%
16%
Change in production quantity
Change in production compare w ith base year (%)
Change of output price compare w ith base year (%)
Figure 7-1: Changes in Agricultural Production of Household in the South under
Global Trade Scenario
Source: Household model simulation
Rice production, with the highest increase in quantity, has a very small increase in
percentage terms compared with the base year: 0.43 percent. This can be explained by
121
the limited possibility for expansion in land for rice cultivation, and productivity of rice
production may have already reached its upper bound.
0.5
1.0
1.5
2.0
2.5
3.0
Rice Pig Chicken
Rat
io o
f ch
ange
s in
out
put
pric
es
and
inpu
t pr
ices
-4%
0%
4%
8%
12%
Out
put
chan
ges
com
pare
with
ba
selin
e
Ratio o f changes o f P output/P raw feed Ratio o f changes of P output/P industrial feed
Ratio o f changes o f P output/wage Ratio o f changes of P output/P chemical inputs
Output change compare with baseline(%)
Figure 7-2: Changes in Relative Prices of Outputs and Inputs and Changes in
Agricultural Production of Household in the South under Global Trade Scenario
Source: Household model simulation
As stated in Chapter 6, the global trade scenario creates changes in market prices of all
commodities and factors. The Figure 7-2 illustrates the reaction of household
agricultural production activities to price changes. Since an increase in output price is
faster than that of input prices, using inputs are relatively cheaper compared with the
baseline. In livestock production, inputs such as feed are cheaper in comparison with
pre-liberalisation. For rice cultivation, using both chemical fertiliser and pesticides are
also less costly in the global scenario. This indicates why the household has incentives
to expand production of rice and livestock products under the simulation.
Thanks to the expansion in agricultural production as well as increases in output prices,
under the scenario, the household in the South increases their profit for rice and
livestock production by 14 and 30 percent, respectively (Figure 7-3).
122
Figure 7-3: Changes in Profit of Agricultural Production of Household in the South
under Global Trade Scenario
Source: Household model simulation
The increase in prices of all outputs and relative decrease in prices of inputs help the
household to expand agricultural production, hence increase farm profit. At the same
time, increased labour wages under the global scenario enrich the household in terms of
full income, and help cover the bigger cash expenditure they incur, and longer leisure
time.
-100
0
100
200
300
400
Food qty Other food qty Industry and others Leisure day
kg/u
nit/
day
-15%
-10%
-5%
0%
5%
10%
Ow n price effect Cross price effect
Income effect Total change in consumption
Price change compare w ith baseline (%)
Figure 7-4: Disaggregate Effects of Price and Income to Changes in Consumption of
Household in the South under Global Trade Scenario Compared with Baseline
(percentage)
Source: Household model simulation and calculation based on LES elasticities in Chapter 6
123
In the simulation, prices of groups of “main food” and “industrial commodities”
increase at 5.6 and 1.4 percent, and labour wage increases by 8.23 percent compared
with the pre-simulation period. As expected, own price effects make consumption of
these groups decrease. However, due to the availability of more income from the
expansion of agricultural production and the value of farm labour, income effects
dominate and help increase consumption of all broad groups of commodity under the
scenario.
Changes in consumption due to own-prices and income effects in the “main food” group
including rice, pig, chicken meat, fish, vegetable and other meat are shown in Table 7-4.
Conforming to the theory of demand, when prices of all individual commodities
increase under the simulation, consumption of each commodity decrease by 1.43, 0.82,
0.58, 0.66, 2.44 and 1.19 kg for rice, pork, chicken, fish, vegetable and other meats,
respectively. Nevertheless, increased cash income of the household allows allocation of
more funds for “main food” group consumption. This large income effect not only
prevents reduced consumption of each commodity, but also increases total quantity of
individual commodities. This explains why the household consumes more, even though
the commodities become more expensive in the trade liberalisation scenario.
Opportunities for consuming all commodities as well as leisure, at the same time
expanding agricultural production of the farm under the scenario of global trade,
significantly increase welfare of the household by 22 percent compared with the
baseline.
124
Table 7-4: Disaggregated Changes in Main Food Consumption Quantity of Household in the South due to Price and Expenditure Effects
Other meat -1.64 0.29 0.06 -0.14 -0.08 -1.19 11.79 9.09
Expenditure and price changes compared to baseline (%)
9.51
2.19
2.19
2.31
8.35
3.28
26.70
Source: calculation based on household model simulation and LA-AIDS elasticities in Chapter 6
125
7.3.2. Impacts of Global Trade Liberalisation on Households in Different Regions
The following section analyses the effects of each trade scenario on households in the
regions to compare changes in welfare and different reactions of households in
production and consumption within each trade liberalisation context.
With changes in prices of agricultural inputs and outputs under the global trade
scenario, households in the four regions increase farm profit. The household in the RRD
has the largest increase of more than 30 percent followed by the Central and the South.
The household in NE+NW has a limited change of 10 percent.
-20%
-10%
0%
10%
20%
30%
40%
RRD NE+NW the Central the South
Region
chan
ge c
ompa
re w
ith b
asel
ine
Rice Pig Chicken Farm profit
Figure 7-5: Changes in Output Production and Farm Profit of Households in Different
Regions under Global Trade Scenario
Source: Household model simulation
Rice outputs only increase at limited rates in the Central and South regions while in the
other regions, households tend to reduce rice production. Almost all households expand
livestock production by about 5 to 10 percent in pig and chicken output compared with
the baseline. In the NE+NW, chicken production reduces by 12 percent compared with
the baseline, even though the household in this area is confronted with the same price
changes as households in other regions. This may be due to the different method of
chicken production, manifested by the high proportion of raw feed in total chicken feed
(85 percent). Hence, with a price change in raw feed under the simulation of about 10
percent and industrial feed of 4.7 percent, obviously the household in NE+NW faces
higher prices for chicken feed, the main input for production.
126
Table 7-5: Ratio of Raw Feed and Price Change of Feed in Different Households
(percent)
RRD NE+NW Central South
Ratio of raw feed in total chicken feed 64 85 57.9 25.6
Price change of chicken feed* under
global trade scenario
6.79
8.05
6.39
5.47
Source: Household model simulation
* change in price of combination feed (based on ratio of raw and industrial feeds) in each region
With pig production, the household in the North mountainous area (NE+NW) uses a
similar technique to other households in other regions, in terms of raw feed in total
feed38. Therefore, with the same impacts of price change in all regions, households have
similar changes in production. However, since the household in the NE+NW area
restricts its rice and chicken production under the simulation, some resources are now
used to produce more pig, leading to higher total farm income. It should be noted that,
compared with other areas, the percentage increase in pig production of the NE+NW
household is quite large; however with modest output in the baseline, the absolute
change in pig output is only 46 kg. A similar situation occurs for chicken production,
with a 2.3 kg reduction in chicken production equating to approximately 12 percent of
total chicken output of the household.
On the consumption side, households in the RRD, NE+NW and the Central region have
similar reactions to the household in the South, with increased consumption in all broad
group of commodities and expenditures, even though prices of the “main food” group
and “industrial goods” are higher than the baseline. This is because of the strength of
income effects in the household. Under the simulation, the price of the “other food”
group decreases by nearly 13 percent, hence its own-price effect combined with the
income effect drives big changes in consumption of this group of commodities in all
regions.
38 The ratio of using raw feed in total feed for pig in the regions are around 92-95 percent, except the household in the South, which is more commercially oriented which, uses only 36.4 percent raw feed for pig production.
127
0%
10%
20%
30%
40%
RRD NE+NW the Central the South
regionch
ange
com
pare
with
bas
elin
e
Main foods Other foods Industrial goods & others
Figure 7-6: Changes in Consumption Quantities of Households in Different Regions
under Global Trade Scenario
Source: Household model simulation
In all regions, when more funds are available to allocate for “main food” consumption,
households increase consumption particularly pork, chicken, other meats, fish and
prawn. This sounds quite sensible since small household producers are often not rich
consumers.
-4%
0%
4%
8%
12%
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
Rice Pork Chicken Fish-shrimp Vegetable Other meat
Figure 7-7: Changes in Main Food Consumption of Households in Different Regions
under Global Trade Scenario
Source: Household model simulation
Inside the “main food” group, meat is still considered a luxury food compared with
other foods such as rice and vegetables. The expenditure elasticities of all meats range
128
from 1.1 to 1.5 in all regions, meanwhile the elasticities of rice and vegetables vary
between 0.6 to 0.8. Therefore, when income increases, expansion of rice and vegetable
consumption is quite limited compared with meat consumption. For households in
NE+NW, consumption of rice and vegetables do not increase as per the other regions
but decrease, as the income effect could not cover the negative effects of price
increases.
One factor contributing to increased welfare of households under the global trade
scenario is changes in time allocation of the household. Households in all regions chose
more leisure and less work than in the pre-simulation period.
Table 7-6: Changes in Labour Allocation of Households in Different Regions under
Global Trade Scenario Compared with Baseline (percentage)
RRD NE+NW Central South
Leisure 1.57 0.45 1.05 2.40
Labour supply -1.64 -0.35 -0.75 -1.69
Days working on-farm 3.03 84.42 3.87 3.55
Days working off-farm -2.16 -10.83 -1.25 -2.83
Source: Household model simulation
Relatively, the number of leisure days in households increase by about 1 to 3 percent.
These changes in leisure allow households to increase their utility by about 2 to 4.5
percent compared with the baseline. In order to have more leisure, households chose to
work less off-farm, while at the same time increasing work on-farm. The latter is driven
by the increased demand for farm labour, due to expansion of livestock production, as
mentioned above. For the household in the north mountainous area (NE+NW), as pig
production expanded by 27 percent, the number of days working on-farm increases by 2
man months in comparison with the baseline.
7.3.3. Impacts of Unilateral Trade Liberalisation on Households in Different
Regions
With the unilateral trade scenario, Vietnam reduces taxes for all imported commodities.
This allows Vietnam some benefit without negotiating with others, by reallocating
129
resources more efficiently. Unilaterally decreasing taxes has a marginal impact on many
prices of commodities in agricultural sectors, prices of outputs and production inputs:
about 1 percent compared with the baseline. Inside households, the reallocation of
resources for a more efficient production combination also happens. The household re-
structures agricultural production on-farm to gain more farm profit. Pig and chicken
production increases in almost all regions under this scenario. In RRD, rice production
reduces slightly and livestock production increase. Expansion of all pig, chicken and
rice production occurs in households in NE+NW and the South. In the Central region,
the household slightly reduces pig production, while increasing production of rice and
chicken.
-3%
-1%
1%
3%
5%
RRD NE+NW the Central the South
region
change c
om
pare
with
base
line
Rice Pig Chicken Farm profit
Figure 7-8: Changes in Agricultural Production of Households in Different Regions
under the Unilateral Trade Scenario
Source: Household model simulation
Restructuring farm production under the scenario helps households increase farm profit,
hence have more cash income for expenditure. Table 7-7 presents changes in
consumption of the households under the simulation. Price and income effects are
negligible in household consumption of “main food” and “industrial goods”.
Meanwhile, due to the impact of the trade scenario, the price of “other food” decreases
by 15 percent compared with the baseline: this price effect, along with the income effect
helps households in all regions to increase “other food” consumption by about 20
percent.
130
Table 7-7: Changes in Price and Consumption Quantity of Households in Different
Regions under the Unilateral Trade Scenario (percentage)
RRD NE+NW Central South
Change in “main food” price -0.32 0.02 0.63 1.26
Change in ”main food” quantity 1.71 1.60 1.38 1.95
Change in “other food” price -15.22 -15.22 -15.22 -15.22
Change in “other food” quantity 19.69 19.83 19.62 20.28
Change in price of “industrial and
other expenditures” group
-1.40
-1.40
-1.40
-1.40
Change in quantity of “industrial
goods and others”
1.74
3.13
3.07
3.46
Source: Household model simulation
In the main food group, rice and fish have higher prices, with all other individual
commodities having slightly lower prices, about 1 to 2 percent cheaper in comparison
with pre-simulation. Figure 7-9 shows that in all regions, households consume cheaper
commodities more and less rice and fish.
-3
-2
-1
0
1
2
3
4
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
(%)
Rice Pork Chicken Fish-shrimp Vegetable Other meats
Figure 7-9: Changes in Food Consumption of Households in Different Regions under
Unilateral Trade Scenario (percentage)
Source: Household model simulation
In general, by changing production and consumption behaviours, households in all
regions increase welfare under the unilateral trade scenario. Welfare of households
131
increases by 6 - 7 percent due to impacts of trade liberalisation, with the value of
welfare change in terms of money ranging from 600 000 to 900 000 VND per
household per year, depending on agro-ecological areas (see Appendix A7.2a).
Apart from changes in behaviour in production and consumption, time and labour
allocation inside the households also affects changes in welfare. The NE+NW and the
South are two regions which expand farm work, at about 1 percent compared with the
baseline. In comparison, households in RRD and the Central regions reduce the number
of working days on-farm, and work more off-farm. However, the deciding factor that
makes the household better off in terms of utility is the reduction in number of days of
labour supply and increased leisure days in all regions. It should be noted that behaviour
change of the household in commodity consumption and labour supply is decided by the
assumption made in the household model. With the application of the LES, whenever
households become richer and have more income available, they consume more of both
commodities and leisure.
Table 7-8: Changes in Time Allocation of Households in Different Regions under
Unilateral Trade Scenario Compare with Baseline (percentage)
RRD NE+NW Central South
Leisure 0.20 0.41 0.26 0.46
Labour supply -0.21 -0.31 -0.19 -0.32
Days working on-farm -3.08 1.25 -2.73 1.08
Days working off-farm 0.10 -0.51 0.09 -0.63
Source: Household model simulation
7.3.4 Impacts of Regional AFTA to Households in Different Regions
Under the simulation of free trade in the ASEAN region, the price of agricultural
outputs of pig, chicken and rice all increase by 3 to 5 percent. That leads to a household
response to increase farm production in most regions. Households in RRD, the Central,
and the South show similar reactions to price changes by expanding farm production of
rice, pig and chicken. As a result, farm profit in these regions increases by 8 to 14
percent in comparison with pre-simulation.
132
Table 7-9: Changes in Production Outputs and Total Farm Profit of Households in
Different Regions under AFTA Compare with Baseline (kg)
RRD NE+NW Central South
Rice output change 4.05 -107.94 16.39 35.32
Pig output change 4.99 41.02 4.24 13.33
Chicken output change 1.83 -3.95 1.82 4.47
Total farm profit change (%) 14.10 -3.48 11.66 8.25
Source: Household model simulation
Response of the household in NE+NW is similar with that under global trade
liberalisation, since the price of chicken feed becomes relatively more expensive than in
other regions, and the household reduces chicken production. The ecological conditions
for rice cultivation are not competitive in the mountainous area, which drives the
household to concentrate more resources on pig production.
-4%
2%
8%
14%
20%
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
Main foods Other foods Industrial goods & others
Figure 7-10: Changes in Quantity Consumption of Households in Different Regions
under AFTA in Comparison with Baseline (percentage)
Source: Household model simulation
On the consumption side, prices of the broad group of “other food” commodities, and
“industrial goods” reduce by 9.85 and 0.09 percent, respectively, under the simulation.
This leads to increased consumption of those groups in all regions. The increased level
of consumption of the “other food” group in the NE+NW is lower than other regions,
133
and at the same time consumption of industrial commodities reduces by 2 percent. This
is due to decreased farm profits in the household. Under the scenario, prices of the
“main food” group (weighted prices for individual foods (rice, pork, chicken, etc.) in
each region) increase by 1.1 to 2.8 percent. However, since households in the RRD, the
Central and the South have higher profits from farm production, they spend more,
irrespective of the slight increase in prices.
Inside the “main food” group, price increases of rice, fish, vegetable and other meats
lead to a decrease in consumption quantities of those individual commodities in all
regions. However, allocation of more funds for the “main food” group expenditure of
the household in the South compared with those in other regions cause negligible
changes in consumption of rice, fish and other meats in the household. As a result of
trade liberalisation, consumption prices of both pork and chicken are reduced by 0.41
percent. These price effects, combined with increased expenditure, increase pork and
chicken consumption in the South in the simulation.
-5%
-3%
-1%
1%
3%
5%
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
Rice Pork Chicken Fish-shrimp Vegetable Other meats
Figure 7-11: Changes in Food Consumption of Households in Different Regions under
AFTA in Comparison with Baseline (percentage)
Source: Household model simulation
Under the AFTA simulation, the household in the South has the biggest welfare
increase of more than 7 percent compared with the baseline. The households in RRD
and the Central increase welfare by more than 4 percent, while the household in the
northern upland (NE+NW) seems unaffected by trade liberalisation in terms of total
welfare (see Appendix A.7.2b). Changes in time allocation of households’ members are
134
shown in Table 7-10, with number of days of leisure increasing in all regions except
NE+NW. More labour is allocated to on-farm work, and less to off-farm work to meet
requirements from increasing agricultural production of households under the scenario.
Table 7-10: Changes in Time Allocation of Households in Different Regions under
AFTA Trade Scenario Compare with Baseline (days)
RRD NE+NW Central South
Leisure 2.45 -1.73 2.63 6.68
Days working on-farm 3.87 71.32 3.41 6.88
Days working off-farm -6.32 -69.59 -6.04 -13.56
Source: Household model simulation
7.3.5 Impacts of Expansion of the Regional AFTA (ASEAN plus China, Korea and
Japan) to Households in Different Regions
In the last chapter, results of the GTAP model showed that expansion of the regional
free trade area, ASEAN, to include China, Korea and Japan creates the second largest
increase in welfare for Vietnam in general. In the household model, the simulation also
shows that households in all regions of the country under the scenario are better off in
terms of welfare, with the biggest change in the South, where welfare of the household
is improved by 20 percent compared with the baseline. In the other three regions,
welfare increases by around 10 percent (see Appendix A7.2c).
Increases in output prices in the market of 10 to 13 percent are good incentives for
households to expand farm production. In all regions, livestock production and rice
cultivation increase, improving total farm profits of households by 20 to nearly 40
percent compared with the baseline.
With improvements in farm profits, households have more cash income available for
consumption. Consumption of all goods and commodities increases in terms of quantity
over all regions. Inside the “main food” group, due to more funds being allocated,
consumption of almost all individual foods increase under the simulation, even though
the increase in prices lead to negative effects on consumption (see Appendix A7.3a and
b for more detail).
135
-10%
0%
10%
20%
30%
40%
50%
RRD NE+NW the Central the South
regionch
ange
com
pare
with
bas
elin
e
Rice Pig Chicken Farm profit
Figure 7-12: Changes in Production of Households in Different Regions under
AFTA+3
Source: Household model simulation
Expansion of farm production requires more labour on-farm. As a consequence all
households increase labour on-farm by approximately 10 percent, and decrease off-farm
days by 2 to 4 percent compared with the base, depending on region (see summary
Table 7-15).
7.3.6 Impacts of Bilateral Trade Liberalisation with the United State on
Households
Bilateral trade liberalisation between Vietnam and the United States has almost no
impact on consumption prices or output and input prices in Vietnam. Table 7-11
presents price changes under the simulation used in the optimisation of the household
model.
Facing negligible changes in prices, households in all regions except the RRD, do not
respond with changes in production and consumption. Welfare of the household in RRD
changes by only 0.3 percent (see Appendix A7.5a for more detail). As a result, further
details are not discussed here.
136
Table 7-11: Changes in Prices of Consumption Commodities and Production
Inputs and Outputs in Vietnam under Bilateral Trade Liberalisation with USA
(percentage)
Production side Price
change
Consumption side Price
change
Rice and Paddy seedling 0.49 Rice 0.49
Live Pig 0.52 Pork meat 0.10
Live Chicken 0.70 Chicken meat 0.10
Raw feed for pig 0.38 Fish and prawn 0.87
Industrial feed for pig 0.40 Vegetable and fruit -0.01
Raw feed for chicken 0.38 Other meats -0.42
Industrial feed for chicken 0.35 Industrial goods and others 0.28
Chemical fertiliser and pesticides 0.43 Other food group -0.07
Unskilled labour 0.24
Source: GTAP simulation
7.3.7 Impacts of Bilateral Trade with EU and Multilateral Trade Liberalisation on
Households in Different Regions
The simulation results of these two trade liberalisations in the GTAP model show
similar trends in changes in prices of consumption commodities as well as prices of
outputs and inputs of agricultural production. Therefore, reactions of households to the
changes are similar. With higher prices of agricultural outputs under the simulations,
households have incentives to increase farm production (Figures 7-13 and 7-14).
In both optimisations, livestock production is improved in terms of quantity in all
regions. Rice production is not expanded in the household in the South under the
bilateral trade simulation, and decreases slightly in the RRD household under the
multilateral one.
137
0%
5%
10%
15%
20%
RRD NE+NW the Central the Southregion
chan
ge c
ompa
re w
ith b
asel
ine
Rice Pig Chicken Farm profit
-5%
0%
5%
10%
15%
20%
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
Rice Pig Chicken Farm profit
Figure 7-13: Changes in Production of
Households in Different Regions under
Scenario of Bilateral Trade Liberalisation
with EU
Source: Household model simulation
Figure 7-14: Changes in Production of
Households in Different Regions under
Multilateral Trade Scenario
The same trends in consumption of households under the scenarios are also presented.
The “other food” group has the biggest increase in terms of consumption quantity. The
increase under the bilateral trade scenario is due to the positive effect of income, which
dominates over the negative own price effect39, and consumption increase under the
multilateral simulation due to a combination of income and own-price effects (Figures
7-15 and 7-16)40.
-1%
1%
3%
5%
7%
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
Main foods Other foods Industrial goods and others
Figure 7-15: Changes in Consumption of
Households in Different Regions under
Scenario of Bilateral Trade Liberalisation
with EU
Source: Household model simulation
0%
4%
8%
12%
16%
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
Main foods Other foods Industrial goods and others
Figure 7-16: Changes in Consumption of
Households in Different Regions under
Multilateral Trade Scenario
39 Under the bilateral scenario, the price of the “other food” group increase by 0.97 percent in comparison with baseline. 40 Price of the “other food” group decreases by 5.95 percent in the multilateral trade liberalisation.
138
Funds allocated for the “main food” group in both scenarios increase by about 1-2
percent compared with the baseline. That is why consumption of each individual
commodity in the “main food” group increases by a maximum of 3 percent. In both
cases of liberalisation, the consumption quantity of each commodity in the “main food”
of the household in RRD reduces, since the price of each food group expenditures
increase insignificantly. This is also the case for the household in the Central region
under the multilateral scenario (Figures 7-17 and 7-18).
-2%
-1%
0%
1%
2%
3%
4%
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
Rice Pork Chicken Fish-shrimp Vegetable Other meats
-2%
-1%
0%
1%
2%
3%
4%
RRD NE+NW the Central the South
region
chan
ge c
ompa
re w
ith b
asel
ine
Rice Pork Chicken Fish-shrimp Vegetable Other meats
Figure 7-17: Changes in Main Food
Consumption of Households in Different
Regions under Scenario of Bilateral Trade
Liberalisation with EU
Source: Household model simulation
Figure 7-18: Changes in Main Food
Consumption of Households in Different
Regions under Multilateral Trade Scenario
The combination of reactions in production and consumption of households under the
two simulations make them better off in terms of welfare in all cases, apart from RRD
household under bilateral trade liberalisation with the European Union where there is a
small decline (see Appendices A7.2d and e)
In both scenarios, households have similar tendencies in using their time and in
allocation to work. Table 7-12 shows slight increases in leisure time. Due to the
requirements for an expansion in farm production, on-farm work of household members
increases in all regions, especially in RRD under the bilateral trade agreement, at the
same time a reduction in days worked off-farm is reported.
139
Table 7-12: Changes in Time Allocation of Households in Different Regions under
Alternative Liberalisations (percentage)
RRD NE+NW Central South Changes in
compare with
baseline VNM-
EU
Multi-
lateral
VNM-
EU
Multi-
lateral
VNM-
EU
Multi-
lateral
VNM-
EU
Multi-
lateral
Leisure 0.09 0.37 1.20 1.45 0.61 0.58 1.03 1.42
Labour supply -0.09 -0.39 -0.92 -1.11 -0.43 -0.41 -0.72 -1.00
On-farm work 52.87 5.98 5.11 7.87 4.02 6.46 3.87 5.99
Off-farm work -5.96 -1.10 -1.67 -2.22 -0.92 -1.16 -1.72 -2.52
Source: Household model simulation
7.4. Impacts of Trade Scenarios, Opportunities and Threats in Each
Region
The section above analyses in detail reactions of households to price and wage changes
due to trade liberalisations. In this section, the benefits and losses indicated in each of
the scenarios simulated are presented. Opportunities and threats to households from
changes in production and consumption are mentioned. Alternative specifications of
assumptions relating to the labour markets in the global general equilibrium model,
which lead to different simulation results for welfares levels and reactions of
households under the trade liberalisations, are also discussed.
In general, the results from alternative liberalisation scenarios in all regions show that
Vietnam’s small households in the livestock sector would benefit from trade
liberalisation. Table 7-13 presents the summary of welfare changes in households under
alternative scenarios.
The largest benefit that households have is when full trade liberalisation occurs over the
world and the second largest is when ASEAN expands its regional free trade agreement
to include Japan, China and Korea. The results are slightly different in the NE+NW,
where the biggest change in welfare occurs with the AFTA+3 scenario followed by
global trade liberalisation.
140
Table 7-13: Welfare Changes of Households in Different Regions under
Alternative Liberalisations Compared with Baseline (percentage)
Welfare change
of the household
Uni-
lateral
(1)
AFTA
(2)
AFTA
+3
(3)
VNM-
EU
(5)
Multi-
lateral
(6)
Global
(7)
RRD 6.41 4.63 13.82 -0.12 3.26 19.13
NE+NW 5.90 -0.03 12.18 4.76 5.62 11.56
Central 6.11 4.22 11.86 3.36 3.50 15.80
South 6.85 7.29 19.56 5.06 6.78 22.31
Source: Household model simulation
Welfare changes under VNM-USA (4) simulation were omitted since the scenario only made a negligible
change in RDD (0.3%) and had no impacts on the other regions.
The households in almost all regions do not show any response to bilateral trade
liberalisation with the United States, as it leads to very small price changes. Bilateral
trade liberalisation with the European Union also had an insignificant impact on
households in the RRD. Meanwhile liberalisation increases welfare of the other three
regions by 4.76, 3.36, and 5.06 percent in NE+NW, the Central, and the South,
respectively.
Unilateral trade liberalisation of Vietnam increases welfare in households in all regions
equally, by around 6 percent compared with the baseline. All other trade scenarios help
households in Vietnam.
On the production side, trade liberalisation scenarios increase prices of agricultural
outputs, with the biggest increases under AFTA+3 and global liberalisations. As a
result, households in all regions have a tendency to increase livestock production.
Chicken production increases in almost all scenarios with rates ranging from 5 to 10
percent over the regions. Unilateral trade liberalisation in Vietnam seems to have little
impact on changes in behaviour of households in raising chickens. Meanwhile the
expanded AFTA increases chicken output over the country. Pig production also expands
under alternative scenarios, with the rate up to 10 percent higher than the baseline. In
the NE+NW area, pig output in particular increases in the scenarios of AFTA and global
141
trade liberalisation, meanwhile production of chickens and rice are narrowed (see Table
7-14).
Table 7-14: Production Changes of the Households in Different Regions under
Alternative Liberalisations Compared with Baseline (percentage)
Production Changes in
the Households
Uni-
lateral
(1)
AFTA
(2)
AFTA
+3
(3)
VNM-
EU
(5)
Multi-
lateral
(6)
Global
(7)
Rice -2.25 0.15 -0.89 2.57 -1.23 -1.71
Pig 2.11 2.96 4.27 5.44 5.42 7.19
RRD
Chicken 1.34 9.79 42.83 11.77 16.93 5.39
Rice 0.34 -3.82 1.93 1.47 2.00 -4.28
Pig 0.67 24.31 7.73 2.80 3.21 26.95
NE+NW
Chicken 1.41 -15.85 11.06 4.51 6.99 -12.80
Rice 1.93 0.76 1.19 0.72 0.54 0.87
Pig -2.26 2.73 7.00 2.71 3.17 5.51
Central
Chicken 1.84 5.92 14.95 5.39 7.81 10.11
Rice 0.18 0.59 0.73 0.03 0.38 0.43
Pig 0.64 3.84 9.57 3.96 4.36 7.00
South
Chicken 1.91 7.64 16.36 4.39 9.45 10.57
Source: Household model simulation
Benefits from expanding farm production as well as an increase in the value of working
time of households gives more cash income to purchase more goods and commodities
under all trade scenarios (see Appendix A7.6). “Other food” consumption increases
significantly every time households have more available income. Price decreases in this
group of commodities under the scenarios, especially the 10 percent reduction in
AFTA+3 and the global trade scenario, also cause increased consumption. Consumption
of “main foods” and “industrial goods” increase under trade liberalisations, albeit at
modest levels.
142
For the “main foods” group, households consume more under AFTA+3 and global trade
scenarios in all regions. Pork, chicken and other meats are consumed more as soon as
households have available income. The household in the South consume more
individual foods in all scenarios, since their income effects are large enough to dominate
the opposite own-price effects. On the other hand, the RRD household reduces food
consumption in most simulations, except under scenarios of expanded AFTA and global
liberalisation (see Appendix A7.7).
Table 7-15: Changes of Time Allocation of Households in Different Regions under
Alternative Liberalisations Compared with Baseline (percentage)
Changes in Time
Allocation of
Households
Uni-
lateral
(1)
AFTA
(2)
AFTA
+3
(3)
VNM-
EU
(5)
Multi-
lateral
(6)
Global
(7)
leisure 0.11 0.09 0.75 0.05 0.20 0.82
days on-farm -3.08 0.14 10.05 52.87 5.98 3.03
RRD
days off-farm 0.10 -0.24 -2.77 -5.96 -1.10 -2.16
leisure 0.17 -0.06 0.79 0.50 0.60 0.19
days on-farm 1.25 2.52 10.12 5.11 7.87 84.42
NE+NW
days off-farm -0.51 -2.46 -2.89 -1.67 -2.22 -10.83
leisure 0.10 0.12 0.47 0.23 0.22 0.39
days on-farm -2.73 0.16 9.89 4.02 6.46 3.87
Central
days off-farm 0.09 -0.28 -2.06 -0.92 -1.16 -1.25
leisure 0.20 0.62 1.31 0.45 0.62 1.05
days on-farm 1.08 5.73 8.77 3.87 5.99 3.55
South
days off-farm -0.63 -2.47 -4.48 -1.72 -2.52 -2.83
Source: Household model simulation
One of the most important impacts of trade liberalisations on households is the way they
reallocate time between leisure, working on-farm and off-farm. With the application of
Closure C in all regions, households have a tendency to increase working days on-farm,
diminish off-farm days under trade scenarios, except the unilateral liberalisation. Labour
143
allocation between on-farm and off-farm is driven by demand by of the household for
increased farm production, when market prices of agricultural outputs become higher
under the impact of trade liberalisations.
In bilateral trade between Vietnam and the European Union for the RRD, and global
trade liberalisation for NE+NW, the expansion of farm production increases days
worked on-farm by more than 50 and 80 percent, respectively, compared with the
baseline, which limits the time that household labour could be devoted to off-farm jobs.
In all trade scenarios, households chose to increase leisure time. This is attributed to
increased household income. Since the household model applies LES in defining the
household’s utility, the household would choose less work and more leisure every time
they have more available income. The change in time allocation under trade
liberalisation helps households to improve welfare, with the increase in leisure
contributing about 20 to 25 percent of total welfare gained under the scenarios (see
Appendix A7.5).
In the previous chapter, at the national level under scenario simulations, the number of
jobs for unskilled labour increased (Table 6-12). In Appendix A7.11, the change in
unskilled labour jobs in sectors at the national level is reported, showing that labour in
live pig and live poultry sectors increase under simulations compared with baseline.
However, labour in the aggregate OMT sector, which produces pork and poultry meat
decreases in all trade liberalisation scenarios. It is therefore difficult to conclude at the
national level, whether jobs created in the livestock sector increase or not. Meanwhile,
the household model reports a decrease in labour supply from small household livestock
producers, therefore the increase in labour at national level can be attributed to increases
from labour in other sectors such as rice production, textile, electronic or service sectors
(see Appendix A7.11). The increase in labour in the livestock sector at national level (if
there is any) may come from commercial households and bigger scale producers or by
joining of landless labourers.
In the GTAP model simulations, the restructure of the economy due to effects of trade
liberalisation lead to changes in supply, demand and wages in the labour market.
However, the assumption about possible labour market expansion, or in other words, the
binding labour constraint, makes the change in wages more significant in some
144
scenarios. This leads to different impacts on household’s welfare under the context of
trade liberalisation. This is illustrated by the following comparison of household welfare
and behaviour in labour allocation, when households are faced with alternative labour
markets.
Table 7-16 shows the changes in wages of unskilled labour under trade scenarios with
two alternative assumptions of the labour market. In Closure A, the supply of labour is
fixed as exogenous, therefore labour wages increase when the demand for labour
increase. The assumption in Closure C is that Vietnam’s unskilled labour market has
potential to supply labour up to 112 percent compared to the baseline under trade
scenarios, without causing increased labour wages, then allowing labour wages to vary
when demand increases over that limit.
Table 7-16: Change Unskilled Labour Wage under Alternative GTAP Scenarios
with Different Assumptions of Labour Market (percentage)
Changes of
Unskilled Labour
Wages
Uni-
lateral
(1)
AFTA
(2)
AFTA
+3
(3)
VNM-
USA
(4)
VNM-
EU
(5)
Multi-
lateral
(6)
Global
(7)
Closure A 9.15 3.21 13.56 1.3 6.52 7.17 17.66
Closure C (R=12%) 0.31 -1.03 4.44 0.24 2.00 -0.23 8.23
Source: GTAP simulations
Clearly Closure A leads to higher wage levels. The differences in the changes of wages
as well as other input and output prices give different welfare changes under the trade
simulations.
Figure 7-19 presents changes in welfare of households in the South if alternative
assumptions on the labour markets are applied. Similar presentations for other regions
are in Appendix A7.10.
145
0
1000
2000
3000
4000
Unilateral AFTA AFTA+3 VNM-EU Multilateral Global
scenarioch
ange
com
pare
with
bas
elin
e (0
00V
ND
)Welfare change w ith Closure C Welfare change w ith Closure A
Figure 7-19: Welfare Changes of the South Household under Alternative Scenarios
with Different Assumptions of Labour Market
Source: Household model simulation
Fixing labour supply in the economy (Closure A) lead to increased wage in all scenarios
(Table 7-16) which is partly attributed to increased welfare of the household under
Closure A compared with Closure C. An obvious source of the differences is changes in
the household’s labour allocation (Figure 7-20)
Figure 7-20: Changes in Labour Allocation of Household in the South under
Alternative Scenarios with Different Assumptions on Labour Market
Source: Household model simulations
With Closure A, household labour can earn higher wages compared with Closure C.
Since agricultural output prices are not increasing as fast, the high wage pulls labour
146
from their farm. The household chooses to reduce agricultural production, and work off-
farm more for more cash income. Similar results in other regions are presented in
Appendix A7.9.
7.5. Summary
By linking the simulation results of the GTAP model with a household model, this
chapter examines how small livestock households react to changes in economic
policies, especially in the context of trade liberalisation. Analytical results allow one to
see how household behaviour changes when they are both consumers and producers,
taking into account how income changes (from profit, via production) influence
consumption to give a more accurate assessment.
Since four household models are constructed to represent four different agro-ecology
regions over the country, question remains whether or not there are differences in price
transmission between regional markets. A simple method of bivariate correlation
coefficients to measure price integration is applied, concluding that there are
approximately equal changes of prices and wages in all regions.
Regarding impacts of trade liberalisation on the household, the results from different
liberalisation scenarios run with Closure C, show that Vietnam’s small households in
the livestock sector would benefit from most trade liberalisations. The largest benefit for
households is if full trade liberalisation occurs over the world. Households do not seem
to have significant gains from bilateral trade liberalisation with the USA. Meanwhile
unilateral trade can help households increase their welfare.
On the production side, households tend to increase livestock production under trade
liberalisation scenarios, while rice production does not change much in output. Since
land expansion for cultivation is limited, productivity in rice is limited to increases in
variable inputs (given the assumption within the model that rice is the only extensive
land use).
On the consumption side, since more cash income is available under the trade scenarios,
households purchase more goods and commodities, and in particular increase
consumption of the “other food” group.
147
Welfare of the household is dominated by the effect of the household’s labour decision
(working or taking leisure), rather than increases in production profit and consumption
only. Under all trade scenarios, households increase their leisure, at the same time
working more on-farm to meet requirements of labour increase due to the expansion of
agricultural production, despite increases in off-farm wages.
The comparative illustration at the end of the chapter shows the importance of the
assumption made with respect to the national labour market in the GTAP model.
Closure C in the GTAP model is chosen in this study based on the unemployment and
under employment situation in many developing countries, and Vietnam. Even though
Closure A is often used in trade liberalisation simulations, applying that closure in this
study may lead to imprecise results, since the higher wage increase under trade
liberalisations would pull them out of on-farm work, instead of expanding the
production.
148
CHAPTER 8 : CONCLUSIONS
This study analysed the impacts of trade liberalisation on small livestock households in
Vietnam, using a multi-country general equilibrium model (GTAP) at the macro level,
linked with a household model at the micro level. The modelling process was
implemented in three steps:
1. Build a representative household model for each of four regions: RRD, North
mountainous area (NE+NW), Central, and South.
2. Simulate the effects of trade liberalisation scenarios on commodity and input prices,
using the GTAP model, after implementing SplitCom to disaggregate live pig and live
poultry from the general livestock group.
3. Take the simulated prices from the trade model and use them in the household model,
and optimise the household model to investigate household changes in welfare,
production and consumption, as well as in time allocation behaviours under the price
changes that the trade liberalisation scenarios imply.
8.1. Summary, Conclusions, and Policy Implications
The following summary, conclusions, and policy implications are drawn from the study:
Households involved in livestock production in Vietnam are mainly characterised by
small-scale production, not specialised in livestock activities, but raising livestock is
often considered a secondary activity combined with others such as cultivation. Small-
scale production, based on the traditional livestock husbandry method, is very popular
throughout the country. Though constrained by poor access to markets, very low scale
operations, poor access to improved genetics and high-quality forage and concentrated
feeds, and poor animal husbandry and animal nutrition, the smallholder sector supplies
the majority of meat in the market, and raising livestock is an important source of
household cash income.
149
In building the model at the micro level, a recursive household model with
characteristics of a semi-commercial family farm was constructed. In undertaking this,
some assumptions were made: the household is a price-taker in all markets and all
markets exist; commodities are assumed homogeneous, including labour; the total stock
of land and labour are treated as given; issues relating to intertemporal allocation and
risk are omitted. In the model, the role of the labour market is of particular importance.
The labour market was assumed to be an active market, where family and hired labour
are perfect substitutes and face the same wage rates regardless of whether they work on-
or off-farm.
The household was assumed to produce only three agricultural commodities: rice, pigs
and chicken. They consume rice, pork, chicken in the main food group, “other foods” as
well as “industrial goods” bought through the market. For each output produced by the
household, a CD production function was estimated. Consumption decisions were
specified using the LES for demand estimation of broad groups (including leisure), and
LA-AIDS for individual commodity demand in the “main food” group. Two aspects of
production and consumption were then linked together through farm profit.
Household behaviour in consumption and production are driven by elasticities, which
are taken from estimated econometric models. The results showed livestock raising
activities are very responsive to changes in output prices. The labour demand for
livestock production was also very responsive to changes in labour wages. The results
are broadly in agreement with those from previous studies of other countries in similar
conditions. On the consumption side, elasticities from the LES showed that the “other
foods” group are classified as a luxury.
Building the household model for each representative small livestock household was the
main contribution of the study. A recursive model with both consumption and
production linked together allows a more accurate assessment of the determinants of
household consumption, since the household’s response to a change in an exogenous
variable such as a commodity price consists of both restructuring of consumption
patterns attributed to expenditure and consumption substitution effects, and making
production decisions which will influence income. Choosing the LES functional form,
which is easily derived from a direct utility function, is very helpful since it allows one
150
to assess consumption demand of broad groups of commodities, including leisure, and
also help solving the complexity of problems in households allocating time between
work and leisure.
At the macro level, several scenarios for trade liberalisation in Vietnam including
unilateral, bilateral, regional, multilateral and fully global liberalisations were explored
in the GTAP model. All simulation scenarios were constructed so as to describe as
closely as possible the rules and regulations required by trade agreements for all trade
partners, including Vietnam.
One important step in applying the GTAP model was the disaggregation of livestock
into new sectors for live pig, poultry and other animals using SplitCom software. By
disaggregating the group into its different components, price signals of the separate
commodities were better captured, particularly since changes in prices of live pig and
poultry were different in all trade liberalisation scenarios, and the study was interested
in impacts of trade liberalisation on households who raise pigs and poultry.
In implementing the GTAP model, the labour market issue was also considered since it
has a very significant influence on simulated changes of social welfare and household
behaviour. In the GTAP, three different closures relating to the labour market were
explored: (1) a standard closure (Closure A) with the assumption of full employment
utilisation and full factor mobility in all factor markets; (2) in Closure B, where
unemployment exists, real wages were held constant and the supply of unskilled labour
adjusted following the policy change; and (3) in Closure C, unemployment also exists,
but the supply of unskilled labour is not unlimited. With trade liberalisations, Vietnam
can supply unskilled labour up to a maximum 112 percent compared with the baseline,
without causing any increase in wages, but if demand for labour increases beyond that
point, wages must increase to clear the market. In this study, Closure C was chosen as
the base situation, as it is believed to be a better representation of the Vietnamese
situation in resource allocation and usage of endowments. There were some sensitivity
results reported for Closures A and B, since implementing different closures in the
GTAP model showed significantly different effects in defining the scope and changes of
commodity/factor prices.
151
At the national level, the results of the trade liberalisation simulations in the GTAP
model show that Vietnam can get the largest benefit if full liberalisation occurs. The
expansion of the ASEAN free trade area to include China, Korea and Japan brings the
second largest welfare increase to the society. The voluntary trade liberalisation of
Vietnam in a unilateral liberalisation would generate some benefit without the need for
negotiating with others. However, the market access benefits are limited because other
countries do not open their markets. A bilateral trade agreement with USA does not
bring much benefit to the country as a whole, but has a positive impact on the textiles
sector41
All trade scenario simulations increased production in rice and livestock sectors. Some
export-oriented sectors, such as textiles and electronics also increased production under
most of the alternative scenarios. And since most of the output of these sectors is
exported, the sectors also increased exports in all scenarios. Trade liberalisation also led
to significant increases in imports for all commodities.
By linking simulated prices of the GTAP model with the household model, the results
of the simulations showed that Vietnam’s small households in the livestock sector
would benefit from almost all trade liberalisations. The largest benefit was generated by
full liberalisation - by comparison there was almost no gain from bilateral trade
liberalisation with the USA. Meanwhile, in line with the macro-economic level results
from GTAP, unilateral trade liberalisation led to increases in household welfare.
For the small households of the livestock sector, output prices of rice, pig and chicken
increased under trade scenarios - an incentive for households to produce more on-farm
thereby increasing farm profit and cash flow. At the same time, by removing import
taxes, inputs for production and consumption goods were relatively cheaper compared
with the baseline. This effect was particularly noticeable for the “'other food”
consumption, with the increase in consumption ranging from 10 to 30 percent.
Producing more agricultural products also required households to devote more labour
41 However, it should be noticed that since the base year of the simulation in GTAP is 2001, these conclusions may be affected by trade situation in that period of time. For example, the possibility of trade between Vietnam with the United States in agricultural goods was limited because of diplomatic relationships between two countries were only just normalised after a long period of time. If database of the model is upgraded, says to GTAP version 7 (with base year data of 2004), the conclusion may be different. This is one of the study limitations, which will be returned to later in the chapter.
152
time for on-farm work in almost all trade scenarios. There are also some benefits to
households due to trade liberalisation, which are not captured by the model but which
should be recognised. Households who are producers can exploit opportunities of
productivity-enhancement by easier access to improved production technologies.
One of the important issues contributing to increases in household welfare is the
decision of the household on time allocation. In almost all scenarios, the household
decided to take more leisure. This contributed about one quarter of the household’s total
welfare increases in the trade liberalisations.
The decision of the household in time allocation was driven by the type of utility
function chosen. By applying the LES, the household chose to work less and take more
leisure every time they had more income available. Since full income is defined by the
sum of non-farm income, net other income, a farm’s profits, and the value of the
household’s stock of time (wT), increased wage rates under trade scenarios made
households feel more wealthy, hence they spent more time for leisure than work.
Choosing an appropriate assumption about possible labour market expansion in the
general equilibrium trade model is very important. Assumptions about whether wages
or labour supply changes to clear markets have important implications for wages
simulated by the model. The study compared decisions of the household on labour
allocation under alternative trade scenarios with Closure A and Closure C. Assuming
full utilisation of labour at the national level, labour wages under Closure A were much
higher than that under Closure C, where the economy supplied up to 12 percent more
unemployment to the labour force. Higher wages under Closure A made the household
wealthier, in terms of full income, resulting in the household choosing more leisure time
by working less compared with that under Closure C. At the same time, since
agricultural output prices are not high enough, farm production does not promise
increased benefits, so higher wages pull out more household labour from their farm
under Closure A. The household chooses to reduce agricultural production, and work
more off-farm. This illustrates that choosing Closure C is appropriate, since it represents
Vietnam’s economy more precisely with wage increases due to trade liberalisations at a
more suitable level, that wouldn’t draw household labour away from on-farm work to
an excessive degree.
153
In summary, the research shows that Vietnam’s small households in the livestock sector
would benefit from almost all trade liberalisation simulation scenarios. What
implications do these results have for policy? We can envision that in the long run
Vietnam will keep moving towards a more open economy. The household seem
continuously get benefits, which are derived from an increase in livestock output prices
relative to production costs. The availability of improved production technologies for
production could contribute to increase competitiveness and income of the household in
a globalised economy. As experienced in other countries, which follow liberalisation,
increased specialization in production and trade with income growth and increasing
urbanization is happened. Higher labour costs, including for unskilled labour and
opportunities in the manufacture sector will act as pull factors attracting workers from
agricultural household, which promises a possible increase of non-agricultural income
of the household. Certainly, as the past experience showed, this transition is not
necessarily smooth.
In that context, developing enabling policies to help small households to better exploit
the opportunities that globalisation presents is not an easy exercise for the government.
Given Vietnam’s current policy is oriented to increasing market orientation in livestock
production, with a focus toward the development of industrial livestock production
system, how can the government facilitate and support the small household? Since
liberalisation leads to a tightening of the link between domestic and world markets, it is
desirable for government to have a policy to strength other sectors such as industry or
service, or the export sector, to generate jobs and absorb labour from the small
household/ agricultural sector, before exposing small households to a fiercer
competition with large-scale producers in the domestic economy and others in the
world. The alternative option of facilitating more productive smallholder production
systems through adoption of productivity-enhancing technologies may be more
beneficial, equitable, and sustainable in the long run.
8.2. Limitations of the Study and Potentials for Further Study
Constructing a household model was a major contribution to this study. With the
principle of maximising household utility in response to changes in external factors,
reactions of the household to production, consumption and time allocation are easily
154
captured. This capability is the power of the household model in assessing impacts of a
number of policy issues on the household, such as government intervention to input and
output prices, restrictions on exports or imports, decisions about subsidisation, etc. In
the context of urbanisation and industrialisation, migration or capacity of the industrial
zones in absorbing labour may impact households in terms of welfare which can be
assessed using the model.
The household model was constructed using a recursive form, in which the household
decision making process is divided into two separate stages. The benefit of the model is
through its treatment of the profit effect on consumption: the two sides of production
and consumption are linked, and then the reaction of the household to an external factor
change is captured more precisely. However, the model has some disadvantages itself.
Some assumptions made in the model, such as the household is a price-taker in all
markets and all markets exist and are perfect and commodities are homogeneous,
including labour, are not always suitable. In some cases, these assumptions may be
violated, for example in the situation of under- or unemployment in the agricultural
sector in the slack season. Therefore an extension to the model allowing for incomplete
markets may be useful. In particular, in the labour market, where there may be different
wages for on- and off-farm work, or limits to obtaining work may be good options for
future research.
The household model in the study is often used for assessment in the short run, since
some factors are fixed under the model. It was assumed that cultivation land area was
fixed, technology of agricultural production did not change, savings and exogenous
income of the household were also unchanged, but in the long term the above factors
may vary and be endogenous. Therefore relaxing these assumptions may be good
applications for future modelling. In this study, the impact of risk was not included.
However, the issue of risk, which may vary under alternative liberalisation scenarios,
may also influence production decisions, so extending the model to include risk issues
is another potential area of study.
In the study, LES and LA-AIDS models were used for the demand side, and Cobb
Douglas models were applied for production functions. These choices were based on
suitability for inclusion in the household model, and their ability to be solved in that
155
context. However, choice of functional forms for the econometric models may change
policy conclusions. For example, an assumption of the LES model implies the
household will consume more leisure as soon as they become wealthy, which is partly
defined by the value of total stock of time. This may not be strictly true, especially for
very poor households. Therefore a study that allows different values for labour wages
and 'cost' for leisure would be helpful. The choice of functional forms applied in the
model may be a useful area to explore further in development and application of the
household model.
To assess possible impacts of trade liberalisation on the household, a trade model at the
macro level was linked with the micro household model. Matching perfectly an
aggregate model such as GTAP with a household model is always difficult. In this
study, in order to match sectors, commodities as well as inputs and factors between the
models, some simplifying assumptions were made. Disaggregating the live pig and
poultry sectors as separate activities within the GTAP database was implemented using
SplitCom. Splitting commodities requires massive supporting data sets. If possible,
further work on the appropriate level of aggregation of commodities in GTAP for the
purpose of this research would be useful. The “SplitCom technology” is also a potential
application for trade liberalisation research using the GTAP model.
Choosing the closure rule in GTAP is required, and has significant implications for
simulated changes in the economy. In this study, several closures were used to compare
the sensitivity of results. Identifying a closure which is most suitable for the Vietnamese
situation is not straightforward. Assuming a maximum level of labour increase in
Vietnam of 12 percent was supported by some macro-economic data, however other
closures may also need to be explored in future research. Currently, the only variable
that was changed was labour, but there may be other variables, such as the treatment of
trade balances that could also be explored. Since the GTAP model applied in this
research is a static model, it cannot capture ongoing changes of the policy. The static
model only allows assessment at a point in time, , which does not allow a consideration
of the time path between equilibria to be considered.. Research with a dynamic GTAP,
which is available now, would be more ideal for future studies of this nature.
156
In conclusion, applying the GTAP model at the macro level and linking it with
household models at the micro level, the study met the objective of the study: to assess
impacts of globalisation and trade liberalisation on small households in the Vietnam
livestock sector. The study showed the welfare of the households is improved under the
trade scenarios. However, the results depend significantly upon assumptions made in
the models. The assumptions made about the operation of the labour market (both at the
aggregate level within GTAP, and on household decisions about work/leisure) are most
important. Acknowledgement of advantages as well as understanding of the limitations
of the models are necessary for further and deeper studies in the future.
157
APPENDICES
A3.1: Recursive Model Matrix
The annex is verbatim drawn from Singh, Squire and Strauss, 1986 to give more detail on how
the household model is separated between revenue and expenditure.
A household utility function is assumed to exist:
),.....,( 1 LXXUU i =1, …., (A3.1-1)
Where Xi is vector of commodity i consumption for different members of the household, XL
denotes leisure time.
Its utility is maximised subject to a budget constraint:
L
iii XpY
1
(A3.1-2)
Where Y is the household’s full income and pi’s are commodity prices (pL is wage rate):
RLpVqQqTpY L
N
iii
M
jjjL
11
(A3.1-3)
Where:
T= time endowment
Qj = output, j= 1,…,M
Vi = non-labour variable inputs, i=1,….,N
L = labour demand
qj = price of Qj
pi = price of Vi
R = exogenous income
It is assumed that L is labour demanded by household both family and hired, and they are
in which A: land cultivation for rice production D: labour requirement V: variable inputs G: feed for pig or chicken 2. Utility functions – LES (of each individual people in the household):
wi: budget share of a given food commodity pi: the price of commodity i i = rice, pork, chicken, fish and prawn, vegetable, and other meats M: household expenditure P: price index
186
A6.1: Initial Values and Percentage Changes in Vietnamese Exports from Alternative Scenarios with Closure A