FOOD DEMAND IN URBAN CHINA: AN

Transcription

FOOD DEMAND IN URBAN CHINA: AN
FOOD DEMAND IN URBAN CHINA: AN EMPIRICAL ANALYSIS USING MICRO
HOUSEHOLD DATA
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy
in the Graduate School of The Ohio State University
By
Kang Ernest Liu, M.S.
*****
The Ohio State University
2003
Dissertation Committee:
Approved by
Dr. Wen S. Chern, Adviser
Dr. Barry K. Goodwin
Dr. Brian E. Roe
Adviser
Department of Agricultural, Environmental,
and Development Economics
Copyright by
Kang Ernest Liu
2003
ABSTRACT
China, since its economic reform in 1978, has changed significantly as it makes
its transition from a centrally-planned to a consumer-oriented economy. This dramatic
shift in the economic structure has gradually increased household income and changed
consumption patterns in urban China. This study attempts to provide a better
understanding of heterogeneous consumer patterns in urban China by developing a multistage censored demand system using household data.
Specifically, this study attempts to develop an economic model considering
heterogeneous consumption patterns across households and commodity groupings and to
estimate econometric models of a Quadratic Almost Ideal Demand System (QAIDS)
using household data. Three methodologies are integrated in this study including
constructing a multi-stage demand system, incorporating demographic variables using the
‘ordinary budget share scaling and translation’ (OBSSAT), and employing a two-step
estimator to deal with zero consumption problems.
This study covers three provinces, Shandong, Jiangsu, and Guangdong in China
and used household data from 1993 to 1998, with 2,000 observations each year, provided
by the National Bureau of Statistics in China. Based on the Chinese food guide pyramid,
a three-level utility tree is constructed to divide 18 food items into five subgroups.
ii
An empirical analysis is conducted by estimating econometric models in a
sequence of six steps to examine the impact of the potential factors, e.g., income and
demographic variables on food demand. The results show the uniqueness of this study in
three dimensions. First, using the OBSSAT helps us to answer the problem of “how to
break down the heterogeneous consumption patterns in urban China?” In addition, our
findings also show that China should be treated as several markets instead of one. Second,
the QAIDS has not been applied to the study of food demand in urban China. Our results
show that the QAIDS is superior to the AIDS; however, the degree of importance for the
quadratic term decreases as other effects such as demographic and censoring effects are
considered in a demand system. Finally, 18 food items are broken down into five food
subgroups and are estimated by a multi-stage censored QAIDS. Including this large food
bundle in a demand system provides us detailed information of the relationship among
food items. For example, the demand for such detailed foods as yogurt, nuts, and bean
products included in this study, has never been investigated previously.
iii
Dedicated to my parents, to my God mom,
and to my friend, Jerry
iv
ACKNOWLEDGMENTS
I thank my adviser, Dr. Wen S. Chern, for his intellectual support, encouragement,
and enthusiasm which made this dissertation possible, and for his patience in correcting
both my stylistic and scientific errors. To Dr. Barry K. Goodwin and Dr. Brian E. Roe,
my committee members, whose generous suggestions and comments were invaluable to
this dissertation and to my future career. To Dr. Steven T. Yen, whose suggestions were
productive and fruitful to this dissertation. To Dr. Steve Wu, Dr. Raylene Kos, and Dr.
Patricia E. Enciso, my candidacy and oral examination faculty representatives, for their
wisdom and inquires which broadened this dissertation and research.
I am grateful to the department of Agricultural, Environmental, and Development
Economics for offering me a combination of graduate research and teaching associates
for my Ph. D. study, including a teaching mentoring program and to Dr. Bernard L. Erven
and Dr. Eugene Jones, my teaching mentors, for sharing their incredible teaching
experiences with me.
My greatest thanks also go to Mr. Gerald E. Baker, who has spent much of his
valuable time and effort helping me improve my English, to my family members: my
grandmother, parents, siblings, new-born cousin, and relatives who all share the
happiness in the completion of my Ph. D. degree, and to my friends at The Ohio State
University, with whom I shared the life of studying in the United States.
v
VITA
1992 ......................................................... B.S. Agricultural Economics, National Taiwan
University
1994 ......................................................... M.S. Agricultural Economics, National
Taiwan University
1996 – 1997 ............................................. Research Assistant,
Institute of Economics, Academia Sinica,
Taipei, Taiwan
1998 ......................................................... M.A. Economics, the Ohio State University
2002 ......................................................... Master of Applied Statistics, the Ohio State
University
1997 – Present ......................................... Graduate Teaching and Research Associates,
the Ohio State University
PUBLICATIONS
Research Publication
1. Liu, K.E. and W.S. Chern, “Estimation of Food Demand Systems: Evidence from
Micro Household Data in Urban China,” Agricultural and Resource Economics Review,
Selected Abstracts of the NAREA Meeting, 30 (Oct. 2001): 216-217.
2. Liu, K.E., J. Geaun, and L.F. Lei, “Optimal Hedging Decisions for Taiwanese Corn
Traders on the Way of Liberalization,” Agricultural Economics 25 (Sep. 2001):303-309.
3. Liu, E. (Research Assistant), “1999 Ohio Farm Income,” Ohio Agricultural Statistical
Service, United States Department of Agriculture, October 2000.
4. Liu, K.E. and W.S. Chern, “Nutrition Intakes in Urban and Rural China: Implications
for Consumers and Agricultural Producers,” Poster at annual meeting of the AAEA,
American Journal of Agricultural Economics, 81 (Dec. 1999): 1325.
vi
5. Liu, K.E. and W.S. Chern, “Food Consumption in Urban and Rural China: Quantity
vs. Quality,” Referred Poster at 45th Annual Conference of the ACCI, Consumer Interests
Annual, 45, 1999: 132.
6. Liu, E. (Research Assistant), “1998 Ohio Farm Income,” Ohio Agricultural Statistical
Service, United States Department of Agriculture, October 1999.
7. Liu, K. and Li-Fen Lei, “Hedging Taiwan’s Corn Imports on U.S. Futures Markets
under Exchange Rate Risk,” Selected Paper at annual meeting of the AAEA, American
Journal of Agricultural Economics, 77 (Dec. 1995): 1359.
FIELD OF STUDY
Major Field: Agricultural, Environmental, and Development Economics in consumption
and demand analysis and quantitative methods
vii
TABLE OF CONTENTS
Page
ABSTRACT .................................................................................................................... ii
DEDICATION ................................................................................................................ iv
ACKNOWLEDGMENTS .............................................................................................. v
VITA ............................................................................................................................... vi
LIST OF TABLES .......................................................................................................... x
LIST OF FIGURES ........................................................................................................ xiii
CHAPTERS:
1. INTRODUCTION ....................................................................................................... 1
1.1 Motivation and Objectives .................................................................................. 1
1.2 Background Information on China ..................................................................... 2
1.3 Food Demand Studies in Urban China ............................................................... 5
1.4 Summary ............................................................................................................. 7
2. A REVIEW OF METHODOLOGY ........................................................................... 12
2.1 Review of Consumer Theory .............................................................................. 12
2.1.1 Commodity Groupings and Weak Separability ...................................... 12
2.1.2 Incorporation of Demographic Effects ................................................... 15
2.2 Review of Econometric Models: Censored Demand Systems ........................... 18
2.3 Summary ............................................................................................................. 19
3. METHODOLOGY ...................................................................................................... 21
3.1 Economic Model ................................................................................................. 21
3.2 Specification of Econometric Model .................................................................. 30
3.2.1 Functional Form Specification: The Quadratic Almost Ideal Demand
System ..................................................................................................... 31
3.2.2 Incorporation of Demographic Variables ............................................... 32
3.2.3 Two-step Estimator ................................................................................. 36
3.3 Summary ............................................................................................................. 37
4. DATA SOURCES AND DESCRIPTION .................................................................. 38
4.1 Data Sources ....................................................................................................... 38
4.2 Data Description ................................................................................................. 41
4.3 Summary ............................................................................................................. 52
5. EMPIRICAL RESULTS ............................................................................................. 85
5.1 Engel Curve Analysis: Income and Food Consumption ..................................... 85
5.2 The Quadratic Almost Ideal Demand System .................................................... 87
5.3 Incorporation of Demographic Variables ........................................................... 89
5.4 Censored Demand System .................................................................................. 95
5.5 Second Stage Demand System Estimation ......................................................... 98
5.6 Summary .............................................................................................................103
viii
6. SUMMARY AND CONCLUSIONS ..........................................................................139
6.1 Objectives and Methodology ..............................................................................139
6.2 Empirical Results and Implications ....................................................................140
6.3 Limitations and Future Research ........................................................................141
BIBLIOGRAPHY ...........................................................................................................143
ix
LIST OF TABLES
Table
1.1
1.2
1.3
1.4
1.5
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
4.13
4.14
4 15
4.16
4.17
Page
Per Capita Annual Income and Engel Coefficients of Urban and Rural
Households, 1978-2000 ................................................................................
Per Capita Consumption Expenditures and Shares by Commodity Groups
in Urban China, 1985-2000 ...........................................................................
Per Capita Total Living Expenditures and Their Expenditure Shares by
Income Level in Urban China, 2000 .............................................................
Per Capita Food Expenditures of Urban Households by Income Level,
2000................................................................................................................
Demographic Profiles from National Population Censuses in 1964, 1982,
1990 and 2000 ...............................................................................................
Identification of Cities or Counties and Sample Sizes (Each Year) in
Guangdong, Jiangsu, and Shandong .............................................................
Total Living Expenditure (TLE) and Expenditure Shares among Eight
Broad Groups in Urban China, 1993-1998 ...................................................
Total Food Expenditure (TLE) and Food Expenditure Shares for Six Food
Subgroups in Urban China, 1993-1998 ........................................................
Food Consumption for 18 Basic Food Items in Shandong, 1993-1998 ........
Food Consumption for 18 Basic Food Items in Jiangsu, 1993-1998 ............
Food Consumption for 18 Basic Food Items in Guangdong, 1993-1998 .....
Prices for 18 Basic Food Items in Shandong, 1993-1998 .............................
Prices for 18 Basic Food Items in Jiangsu, 1993-1998 .................................
Prices for 18 Basic Food Items in Guangdong, 1993-1998 ..........................
Proportion (%) of Households with Zero Values in Three-level Utility
Tree in Shandong, 1993-1998 .......................................................................
Proportion (%) of Households with Zero Values in Three-level Utility
Tree in Jiangsu, 1993-1998 ...........................................................................
Proportion (%) of Households with Zero Values in Three-level Utility
Tree in Guangdong, 1993-1998 ....................................................................
Descriptive Statistics of Selected Demographic Variables, 1993-1998 .......
Comparison of Food Consumption for 18 Food Items in Urban China,
1998 ...............................................................................................................
Percentage of Households by Household size, 1998 ....................................
Comparison of Food Consumption for 18 Basic Food Items by Household
Size, 1998 ......................................................................................................
Percentage of Households by Number of Children in a Household, 1998 ...
x
8
9
10
10
11
53
54
56
57
58
59
60
61
62
63
64
65
66
68
68
69
70
4.18
4.19
4.20
4.21
4.22
4.23
4.24
4.25
4.26
4.27
4.28
4.29
4.30
4.31
4.32
4.33
4.34
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
Comparison of Food Consumption for 18 Basic Food Items by Number of
Children under Age 17 in a Household, 1998 ............................................... 70
Comparison of Percentage of Households by Income Level, 1998 .............. 71
Comparison of Food Consumption for 18 Basic Food Items by Income
Groups in Shandong, 1998 ............................................................................ 71
Comparison of Food Consumption for 18 Basic Food Items by Income
Groups in Jiangsu, 1998 ................................................................................ 72
Comparison of Food Consumption for 18 Basic Food Items by Income
Groups in Shandong, 1998 ............................................................................ 73
Comparison of Percentage of Households by Age Groups, 1998 ................ 73
Comparison of Food Consumption for 18 Basic Food Items by Age Group
in Shandong, 1998 ........................................................................................ 74
Comparison of Food Consumption for 18 Basic Food Items by Age Group
in Jiangsu, 1998 ............................................................................................ 75
Comparison of Food Consumption for 18 Basic Food Items by Age Group
in Guangdong, 1998 ...................................................................................... 76
Comparison of Percentage of Households by Gender of Household Head,
1998 ............................................................................................................... 76
Comparison of Food Consumption for 18 Basic Food Items by Gender,
1998 ............................................................................................................... 77
Distribution of Households in Percentage by Education Level of
Household Head, 1998 .................................................................................. 78
Comparison of Food Consumption for 18 Basic Food Items by Education
Level in Shandong, 1998 .............................................................................. 78
Comparison of Food Consumption for 18 Basic Food Items by Education
Level in Jiangsu, 1998 .................................................................................. 79
Comparison of Food Consumption for 18 Basic Food Items by Education
Level in Guangdong, 1998 ............................................................................ 80
Comparison of Households by Ownership of a Refrigerator, 1998 .............. 80
Comparison of Food Consumption for 18 Basic Food Items by Ownership
of a Refrigerator, 1998 .................................................................................. 81
Definitions and Descriptive Statistics of Variables Used in Engel Curve
Analysis .........................................................................................................104
Regression Results for Engel Curve Analysis, 1998 ....................................105
Definitions and Descriptive Statistics of Variables in the QAIDS Model,
1998 ...............................................................................................................108
The Wald Test Results of the QAIDS vs. AIDS, 1998 .................................110
Within-Group Expenditure and Price Elasticities, 1998 ...............................111
Definitions and Descriptive Statistics of Variables in the QAIDS Model
with Demographic Variables, 1998 ..............................................................113
Parameter Estimates for Expenditure and Prices in the QAIDS Model with
Demographic Variables, 1998 ......................................................................116
Parameter Estimates for Demographic Variables in the QAIDS Model
with Demographic Variables, 1998 ..............................................................117
xi
5.9
5.10
5.11
5.12
5.13
5.14
5.15
5.16
5.17
5.18
5.19
5.20
5.21
5.22
5.23
The Wald Test Results of the QAIDS Models with Demographic
Variables, 1998 .............................................................................................120
Within-Group Expenditure and Price Elasticities, 1998 ...............................121
Parameter Estimates in Probit Models, 1998 ................................................123
Parameter Estimates for the Censored QAIDS, 1998 ...................................124
Parameter Estimates for the Demographic Variables, Censored QAIDS,
1998 ...............................................................................................................125
The Wald Test Results of the Censored QAIDS with Demographic
Variables, 1998 .............................................................................................128
Within-Group Expenditure and Price Elasticities, Censored QAIDS, 1998 129
Variable Definitions and Descriptive Statistics, Second Stage Demand
System, 1998 .................................................................................................131
Wald Test Results of the Second-Stage QAIDS, 1998 .................................132
Between-Group Expenditure and Price Elasticities, 1998 ............................132
Parameter Estimates for the Second-Stage QAIDS with Demographic
Variables, 1998 .............................................................................................133
Wald Test Results of the Second-Stage QAIDS with Demographic
Variables, 1998 .............................................................................................134
Between-Group Elasticities, the QAIDS with Demographic Variables,
1998 ...............................................................................................................134
Unconditional Elasticities for 18 Food Items, the QAIDS, 1998 .................135
Unconditional Elasticities for 18 Food Items, the Censored QAIDS with
Demographic Variables, 1998 ......................................................................137
xii
LIST OF FIGURES
Figure
4.1
4.2
4.3
Page
A Map of China ............................................................................................ 82
An Example of Utility Tree for Urban Households in China ....................... 83
A Modified Utility Tree for Urban Households in China ............................. 84
xiii
CHAPTER 1
INTRODUCTION
China, economically one of the fastest growing countries in the world during the
last two decades, has attracted considerable attention from researchers and policy makers
debating on its ability to feed 1.2 billion people. The purpose of this study is to provide a
thorough understanding of food consumption in urban China. This chapter is divided into
three sections: the motivation and objectives of this study, an overview of the general
situations surrounding the economic and socio-demographic changes that have occurred
in China, and a brief review of literature on food demand in urban China.
1.1 Motivation and Objectives
This study is inspired by previous findings in estimating food demand in China
with extremely different expenditure elasticities. In the literature of food demand in urban
China, several studies have already shown high expenditure elasticities for grain in China
(e.g., Chen, 1996; Han and Wahl, 1998; Chern, 1997, 2000). These high expenditure
elasticities often translate into high income elasticities for grain, which many forecasters
would find unacceptable for predicting the long-term demand for grain in China (World
Bank, 1997). This study attempts to investigate and construct a suitable and reliable
model that offers a new understanding of food demand in urban China with a thorough
consideration of the demand system specification.
1
Since beginning its economic reform in 1978, China has changed significantly as
it makes its transition from a centrally-planned to a market-based, consumer-oriented
economy. With market liberalization, there have been significant changes in demographic
profiles, household income, and rural-urban income disparity. These changes, among
others, make policy analyses related to food consumption more complex for researchers
and policy makers. Given the great heterogeneity within the Chinese population, it is
important to develop a demand methodology that can rigorously incorporate micro
household data to provide more precise information about food consumption patterns.
Specifically, the purposes of this study are twofold. First, a demand system which
considers heterogeneous consumption patterns across households as well as commodity
groupings will be developed. This includes the incorporation of demographic attributes
and the two- or multi-step decision making processes. And second, more efficient
estimates are provided using the available household data from three provinces in China
(i.e., Guangdong, Jiangsu, and Shandong). An empirical analysis which deals with zero
consumption problems is conducted using the most recent dataset from 1998.
1.2 Background Information on China
China is the most populous country in the world and the question of how to feed
such a large population makes food and agriculture an extremely important issue. Since
the beginning of its economic reforms in 1978, China has had dynamic changes in several
aspects of its society such as freedom of consumption choices, the marketing system,
production technologies, a continuously increasing income, volatile prices, a
heterogeneous population and family structure, and a growing income disparity between
urban and rural areas. In 2002, China’s accession into the World Trade Organization
2
(WTO) will likely move China towards a more market-oriented and consumer-oriented
economy. All of these dramatic changes make the answers to “Who will feed China?”
(Brown, 1994) more complex, causing researchers, both inside and outside of China, to
study China’s food and agriculture with great interest.
China’s rapidly growing economy has caused household income levels to increase.
Table 1.1 presents household income and Engel coefficients for both urban and rural
areas since 1978. The steadily increasing household income with slowly decreasing Engel
coefficients indicates that food consumption is still very important in China (accounting
for 40% and 50% of household income in urban and rural areas, respectively). However,
the decreasing trends of Engel coefficients both in urban and rural China reflect that nonfood consumption has been gaining an increasing share.
The changing relationship between household income and food and non-food
consumption patterns is vividly illustrated in Tables 1.2 and 1.3 using both longitudinal
and cross-sectional trends. Table 1.2 presents the changing patterns of consumption
among eight broad groups between 1985 and 2000, whereas Table 1.3 shows different
expenditure shares among the eight broad groups by income level in 2000. Even though
Table 1.2 shows a uniformly increasing trend on all eight food and non-food expenditures,
both tables indicate a decreasing trend of expenditure share on food over time and across
income levels whereas most of the expenditure shares of non-food items increased over
time or income levels as indicated in Tables 1.2 and 1.3, respectively. For example, the
expenditure shares of transportation, postal, and communication services as well as
recreation, education, and cultural services have increased both over time and income
levels. This spending pattern implicitly indicates that, as incomes increase, Chinese urban
3
households are better able to satisfy their need for food, and thus are trying to improve
their living standard by increasing expenditures on non-food commodity and service
groups.
Looking closer at food expenditure in urban China, Table 1.4 presents the per
capita food expenditure and food expenditures on major food sub-groups such as grain,
aquatic products, and meat and poultry. Importantly, the range of the per capita food
expenditure among income levels is from 1,256 to 2,847 Yuan with an average of 1,958
Yuan, which means significant differences on food expenditure with respect to income
levels. More specifically, consumption patterns are different among the food subgroups.
For example, the differences in expenditure on grain with respect to income groups are
not as significant as those on animal protein food subgroups, such as meat and poultry,
aquatic products, and milk and dairy products. Obviously, households at the lowest
income level spent only 171 Yuan on grain and 283 Yuan on meat and poultry annually
whereas households at the highest 10% income level spent 212 Yuan on grain but more
than 530 Yuan on meat and poultry. This could indicate that Chinese urban households
might change their food consumption behavior from grain-based meals to more animal
products as income increases (Liu and Chern, 1999).
Demographic characteristics are important as they also affect food consumption
patterns. Table 1.5 exhibits some general demographic profiles from the national
population census, 1964-2000. China’s total population increased from 695 million
people in 1964 to 1.265 billion in 2000 with 36% of the total population living in urban
areas in 2000 compared to only 21% in the early 1980s. This huge migration from rural
into urban areas makes it more difficult to predict food demand in China. In addition, the
4
population structure by age group has changed dramatically due to the one-child policy
and improvements in hygiene and public health. For example, the population of people
aged 0-14 decreased from 41% in 1964 to 23% in 2000, while those aged 15-64 increased
from 56% to 70% and those aged 65 and over increased from 3.56% to 6.96%. Finally,
there have been notable improvements in education with remarkable reductions in
illiteracy (from 33.58% in 1964 to 6.72% in 2000) and an overall increase in the
distribution of the population at all levels of education. These combined changes in
demographic profiles have been shown to affect food consumption in the literature (e.g.,
Bhandari and Smith, 2000).
1.3 Food Demand Studies in Urban China1
To date, several studies have previously estimated food demand systems in urban
China employing either pooled aggregate data (e.g., Lewis and Andrews, 1989; Wang
and Chern, 1992; Chern and Wang, 1994; Gao, Wu, Li, and Samuel, 1995; Wales, and
Cramer, 1996; Fan and Chern, 1997; Chern, 2000; Liu and Chern, 2001a) or household
data (e.g., Zhao and Wahl, 1999; Gould and Sabates, 2001; Hasan, Rahs, Mittelhammer,
and Wahl, 2001; Liu and Chern, 2001b, 2001c). The analyses using pooled aggregate
data are different from those using household data. An assessment of these previous
studies is provided here.
First, the assumption of weakly separable preferences is necessary to make
estimation of food demand practicable. However, most of these studies either presumed
the separability condition held or only estimated the second stage of the two-stage
budgeting process. Only Wu et al. (1995) estimated the first stage demand system with
1
In this study, we focus on urban China instead of either rural China or both urban and rural China due to
data availability and urbanization since the economic reform.
5
two broad groups (i.e., commodities under study as the first group and all others in the
second group). Obviously, a reasonable specification of the first-stage budgeting process
is not seen in any of these studies of urban China.2
Next, demographic effects have been given limited attention in the studies of food
demand in urban China (only in Zhao and Wahl, 1999; Gould and Sabates, 2001; Hasan
et al., 2001; and Liu and Chern, 2001b, 2001c). Gould and Sabates used equivalence
scales to incorporate demographic attributes, whereas the others used demographic
shifters. Pollak and Wales (1981) and Lewbel (1985) proposed general methods to
incorporate demographic effects into theoretically plausible demand systems. Some of
their techniques allow demographic variables to affect income and prices directly but
have not been applied to any of the studies for both urban and rural China.
Finally, it is still uncertain which model specification is most preferable in
analyzing the food demand system in urban China. Chern and Wang (1994) compared the
linear expenditure system (LES) with the quadratic expenditure system (QES) and found
the estimated elasticities to be similar despite the nested test rejecting the LES. Chern
(1997) showed notable differences of estimated results between the LES and linear
approximate almost ideal demand system (LA/AIDS). Chern (2000) compared the
performance of the AIDS and LA/AIDS. Liu and Chern (2001a) compared these four
models and concluded that the simpler models such as the LES or LA/AIDS are preferred
in terms of prediction ability. Their conclusion is at odds with the well-known
shortcomings of the LA/AIDS (e.g., Buse, 1994; Hahn, 1994; Moschini, 1995) and the
2
In the studies of rural China, however, Fan, Wailes, and Cramer (1995) and Gao, Wailes, and Cramer
(1996) estimated two-stage demand systems with five commodity groups in the first stage by using
pooled provincial data and household data, respectively.
6
restrictive form of the LES (e.g., Barten, 1993). There have been other attempts to
estimate more flexible functional forms such as the quadratic AIDS (QAIDS), and the
Translog. It would be interesting to know how different the performance would be among
selected model specifications.
1.4 Summary
Motivated by significant differences in expenditure elasticities estimated from
previous studies in urban China, this study attempts to develop a more thorough model
that considers all potential impacts of demand system specifications on food demand
analyses. These improvements, some of which are new to the study of food demand in
urban China, will provide more insightful implications for capturing changes in food
consumption patterns and thus a better understanding will emerge as to whether China
should be treated as one or as several markets in the future. The results of this study
should fill some gaps in the current literature.
The rest of this study is organized as follows. Chapter 2 reviews literature
methodologically and provides direction for improvements to be conducted in this study.
In Chapter 3, an economic model is built and followed by an econometric model
specification. Chapter 4 discusses the data employed in this study, and in Chapter 5, the
empirical results are discussed. A conclusion is provided in the last chapter.
7
Income a
Year
(Yuan)
1978
343.4
1979
387.0
1980
477.6
1981
491.9
1982
526.6
1983
564.0
1984
651.2
1985
739.1
1986
899.6
1987
1,002.2
1988
1,181.4
1989
1,375.7
1990
1,510.2
1991
1,700.6
1992
2,026.6
1993
2,577.4
1994
3,496.2
1995
4,283.0
1996
4,838.9
1997
5,160.3
1998
5,425.1
1999
5,854.0
2000
6,280.0
a
Urban annual disposable income.
b
Rural annual net income.
Urban
Engel Coefficient
(%)
57.5
57.2
56.9
56.7
58.7
59.2
58.0
53.3
52.4
53.5
51.4
54.4
54.2
53.8
52.9
50.1
49.9
49.9
48.6
46.4
44.5
41.9
39.2
Income b
(Yuan)
133.6
160.2
191.3
223.4
270.1
309.8
355.3
397.6
423.8
462.6
544.9
601.5
686.3
708.6
784.0
921.6
1,221.0
1,577.7
1,926.1
2,090.1
2,162.0
2,210.3
2,253.4
Rural
Engel Coefficient
(%)
67.7
64.0
61.8
59.9
60.7
59.4
59.2
57.8
56.4
55.8
54.0
54.8
58.8
57.6
57.6
58.1
58.9
58.6
56.3
55.1
53.4
52.6
49.1
Source: China Statistical Yearbook, 2001, National Bureau of Statistics, China Statistics Press, p.304.
Table 1.1: Per Capita Annual Income and Engel Coefficients of Urban and Rural
Households, 1978-2000.
8
Total Living Expenditures
Food
Clothing
Household Facilities, Articles, and
Service
Medicine and Medical Service
Transport, Postal, and Communication
Services
Recreation, Education, and Cultural
Services
Housing and Utilities
Miscellaneous Commodities and Services
Expenditure Shares
Food
Clothing
Household Facilities, Articles, and
Service
Medicine and Medical Service
Transport, Postal, and Communication
Services
Recreation, Education, and Cultural
Services
Housing and Utilities
Miscellaneous Commodities and Services
Unit
Yuan
Yuan
Yuan
1985
673.20
351.72
98.04
1990
1,278.89
693.77
170.90
Year
1995
3,537.57
1,766.02
479.20
1999
4,615.91
1,932.10
482.37
2000
4,998.00
1,958.31
500.46
Yuan
57.87
108.45
296.94
395.48
439.29
Yuan
16.71
25.67
110.11
245.59
318.07
Yuan
14.39
40.51
171.01
310.55
395.01
Yuan
55.01
112.26
312.71
567.05
627.82
Yuan
Yuan
32.23
47.23
60.86
66.57
250.18
151.39
453.99
228.79
500.49
258.54
%
%
52.25
14.56
54.25
13.36
49.92
13.55
41.86
10.45
39.18
10.01
%
8.60
8.48
8.39
8.57
8.79
%
2.48
2.01
3.11
5.32
6.36
%
2.14
3.17
4.83
6.73
7.90
%
8.17
8.78
8.84
12.28
12.56
%
%
4.79
7.02
4.76
5.21
7.07
4.28
9.84
4.96
10.01
5.17
Source: China Statistical Yearbook, 2001, National Bureau of Statistics, China Statistics Press, p.305.
Table 1.2: Per Capita Consumption Expenditures and Shares by Commodity Groups in
Urban China, 1985- 2000.
9
Income Level a
III
IV
V
3,948 4,795 5,895
--- % --44.30 40.90 37.58
9.50 10.53 10.65
Mean
I
II
VI
VII
Total Living Expenditure (Yuan)
4,998 2,540 3,275
7,102 9,251
Expenditure Shares
Food
39.18 49.47 46.55
34.62 30.78
Clothing
10.01
7.48
8.52
10.71 10.09
Household Facilities, Articles and
8.79
4.66
5.63
6.51
7.85
8.93 10.99 13.19
Services
Medicine and Medical Services
6.36
6.41
6.07
6.28
6.26
6.34
6.23
6.90
Transport, Postal and Communication
7.90
5.62
6.49
7.13
7.37
8.26
8.84
9.48
Services
Recreation, Education and Cultural
12.56 11.29 11.92 11.88 12.46 12.87 13.03 13.23
Services
Housing and Utilities
10.01 11.87 11.24 10.33
9.82
9.82
9.60
9.35
Miscellaneous Commodities and
5.17
3.20
3.58
4.08
4.81
5.54
5.98
6.99
Services
a
Income level is divided into seven blocks with level I the lowest and VII the highest, where I, II, VI, and
VII contain 10% each and III to V contain 20% each.
Source: China Statistical Yearbook, 2001, National Bureau of Statistics, China Statistics Press, pp.306-7.
Table 1.3: Per Capita Total Living Expenditures and Their Expenditure Shares by Income
Level in Urban China, 2000.
Income Level a, b
Mean
I
II
III
IV
V
VI
VII
Food
1,958
1,257
1,524
1,749
1,961
2,216
2,459
2,847
Grain
189
172
179
187
188
192
202
211
Meat and Poultry
411
284
341
389
422
453
486
534
Eggs
57
45
51
54
58
61
63
67
Aquatic Products
144
89
107
128
144
168
179
205
Milk and Dairy Products
69
33
43
57
69
82
96
117
a
Income level is divided into seven blocks with level I the lowest and VII the highest, where I, II, VI, and
VII contain 10% each and III to V contain 20% each.
b
Unit: Yuan
b
Source: China Statistical Yearbook, 2001, National Bureau of Statistics, China Statistics Press, pp.306-7.
Table 1.4: Per Capita Food Expenditures of Urban Households by Income Level, 2000.
10
Year
1964
694.58
18.30
81.70
4.43
1982
1,008.18
20.91
79.09
4.41
1990
2000
Total Population (million persons) a, b
1,133.68
1,265.83
Urban (% of Total)
26.44
36.22
Rural (% of Total)
73.56
63.78
Average Family Size (persons/household)
3.96
3.44
Population by Age Group (%)
Age 0-14
40.69
33.59
27.69
22.89
Age 15-64
55.75
61.50
66.74
70.15
Age 65 and Over
3.56
4.91
5.57
6.96
Population Distribution by Education Level (%)
Junior College and Above
0.42
0.62
1.42
3.61
Senior Secondary/Secondary Technical School
1.32
6.78
8.04
11.15
Junior Secondary School
4.68
17.89
23.34
33.96
Primary School
28.33
35.24
37.06
35.70
Illiterate Population and Rate c
Illiterate Population (million persons)
233.27
229.96
180.03
85.07
Illiterate Rate (%)
33.58
22.81
15.88
6.72
a
Data in this table exclude the population of Hong Kong and Macao.
b
Total population from censuses includes military personnel. Military personnel are listed in the urban
population.
c
Illiterate population of the 1964 census refers to people 13 years old and over who could not read.
Illiterate population in 1982, 1990, and 2000 refers to people 15 years old and over who could not read or
could read very little.
Source: China Statistical Yearbook, 2001, National Bureau of Statistics, China Statistics Press, p.93.
Table 1.5: Demographic Profiles from National Population Census in 1964, 1982, 1990
and 2000.
11
CHAPTER 2
A REVIEW OF METHODOLOGY
This literature review consists of two parts. First, a review of consumer theory
related to commodity groupings and an incorporation of demographic effects is
conducted, then a review of econometric methods dealing with zero observation problems
is summarized.
2.1 Review of Consumer Theory
2.1.1 Commodity Groupings and Weak Separability
Restrictions on preferences are important and necessary to simplify consumers’
decision making process, either by aggregation (commodity groupings) or by separation
(e.g., weak separability) (Deaton and Muellbauer, 1980a). Placing restrictions on
preferences allows economists, conceptually and empirically, to focus on or to deal with
smaller, but more manageable units. This “aggregation over commodities” problem is
first described by Hicks (1936) and Leontief (1936) in a well-known “composite
commodity theorem,” which asserts that if a group of prices moves in parallel, then the
corresponding group of commodities can be treated as a single good (Deaton and
Muellbauer, 1980a, p.121). In practice, the composite commodity theorem is too
restrictive to be satisfied in any empirical study. Lewbel (1996) proposed a generalized
composite commodity theorem to aggregate commodities without separability. His
approach imposes weaker and more empirically plausible restrictions on price
12
movements than the Hicks-Leontief composite commodity theorem. More recently,
Blundell and Robin (2000) have developed a new concept called latent separability,
which groups goods without weak separability. Their latent separability concept is a
generalization of weak separability in which commodities are allowed to enter more than
one group, i.e., overlapping groups. For example, electricity is used together with meat
and other foods in one group while it is used together with gas (mainly domestic heating)
in another group– for both cooking and heating (Blundell and Robin, 2000, p. 79).
However, both Lewbel’s, and Blundell and Robin’s approaches focus on the
improvement of using time-series data. It would be difficult in this study to apply their
approaches to cross-sectional or pooled data with a shorter time period.
There exists a gap of 60 years between Hicks-Leontief (1936) and Lewbel (1996).
During those 60 years, economists paid more attention to separability, an assumption
used to divide vast numbers of commodities into fewer workable groups. Strotz (1957,
1959) first proposed a concept, a utility tree, which allows consumers to break a
simultaneous decision making process into two steps. 3 Commonly, the first step in
budgeting is to allocate expenditures among broad groups of commodities (Gorman,
1959). The second step is to exhaust the quantity of commodities within each group given
the expenditure available from the first step. Hence, with separable utility, two-stage
budgeting is feasible. Weak separability of the direct utility functions is both necessary
and sufficient for the second stage of two-stage budgeting (Deaton and Muellbauer,
1980a, p. 124). However, the first stage of the two-stage budgeting process is more
problematic. Gorman (1959) showed that in order to present the first stage of the two3
Utility tree can be extended to consist of more than two levels which allows a sequence of decision
making processes.
13
stage budgeting process using traditional maximization frameworks with a single price
index and a single quantity index for each of the broad groups, either (1) the utility
function is strongly separable with sub-utility functions of the generalized Gorman polar
form (GGPF) or (2) each sub-utility function is homothetic. However, both strong
separability and homotheticity impose severe restrictions, which are neither desirable nor
realistic. Several solutions are proposed in the literature. For example, Deaton and
Muellbauer (1980a) suggested an approximate solution, i.e., the price index is
approximated using Paashe or Laspeyres indices. Segerson and Mount (1985) used two
price indices instead of one.
Under weak separability, two levels of demand systems are estimated– broad
groups of commodities and within groups of commodities. It is important to investigate
(1) the relationships between the two levels (between groups and within groups) and (2)
commodities within the same group as well as among different groups. Deaton and
Muellbauer (1980a) showed the intergroup substitution which was proved by Gorman
(1971). In addition, some researchers utilized those relationships to propose tests of weak
separability in demand systems (Moschini, 1992; Moschini, Moro, and Green, 1994;
Nayga and Capps, 1994). Edgerton (1997) and Carpentier and Guyomard (2001), on the
other hand, emphasized the derivations of unconditional elasticities.
Most of the empirical studies utilized time-series data to investigate the two-level
demand system. Only Yen and Roe (1989) estimated a two-level demand system with
limited dependent variables. They followed Gorman’s suggestion to specify demand
functions as a generalized Gorman polar form in addition to a restrictive assumption of
strong separability. To date, there have been no additional studies showing further
14
improvement of commodity groupings and weak separability using household data. In
this study, a two-level demand system with a less restrictive assumption is conducted
employing Chinese urban household data.
2.1.2 Incorporation of Demographic Effects
Household data provide rich information of demographic characteristics, e.g.,
household size and composition, education, and residential location. There are two
reasons to incorporate demographic effects into a demand function (Muellbuaer, 1977).
One is to estimate the equivalence scales (ES) and the other is to improve estimates.
Estimation of the ES is usually used to provide information for income maintenance
programs or for studies of poverty and income distribution which is beyond the scope of
this study. This study, with an emphasis on how to achieve better estimates of elasticities,
pays more attention to incorporating demographic effects to enhance a demand analysis.
Estimation using pooled data without incorporation of demographic variables implicitly
assumes identical tastes (or preferences) among the observations. This is not consistent
with the observed facts in household data; hence, in order to get better estimates, several
techniques for incorporating demographic effects have been developed and further
discussed in the literature (see Pollak and Wales (1992) for more details).
Pollak and Wales (1978) suggested two ways to deal with demographic effects.
One is to use the unpooled data, i.e., to separate the entire data set into sub-samples with
identical demographic profiles and then estimate each sub-sample demand system
separately. Obviously, this approach presumes that “… all the parameters in the demand
system depend on the demographic variables, and does not need to specify the form of
the relationship between these parameters and the demographic variables” (Pollak and
15
Wales, 1978, p. 354). However, the major drawback of this approach is that it is not
possible to “draw inferences about households with one demographic profile from
observations on the behavior of households with different profiles” (Pollak and Wales,
1978, p. 354). Therefore, utilization of pooled data would provide advantages to
overcome this shortcoming.
The second approach is to use pooled data. The procedure of using pooled data to
improve estimates can be divided into three separate but interrelated steps (Pollak and
Wales, 1980): (1) to specify a demand system (e.g., the AIDS); (2) to specify which
parameters depend on the demographic variables and which do not; and (3) to specify a
functional form for each parameter which depends on the demographic variables. Pollak
and Wales (1980) proposed a general procedure called demographic translating (DT) and
compared it with demographic scaling (DS) proposed by Barten (1964). 4 Moreover,
Pollak and Wales (1981) described, estimated, and compared five general procedures to
incorporate demographic variables. In addition to the DT and the DS, the general
procedures included the Gorman specification (1976), the reverse Gorman specification,
and the modified Prais-Houthakker procedure.
Lewbel (1985) showed a unified approach– the modifying function procedure,
which included the five procedures suggested by Pollak and Wales (1981) as special
cases. Lewbel’s procedure, as described in Pollak and Wales (1992, p. 73), “modifies the
expenditure function by first replacing each price by a function that depends on all prices
and demographic variables and then subjecting the resulting expenditure function to a
4
Barten’s method (1964) is the first to attempt to systematically discuss how to incorporate the
demographic effects into a complete demand system and is called ‘demographic scaling’ in Pollak and
Wales (1980).
16
further transformation that depends on all prices and demographic variables.”
Unfortunately, Lewbel’s procedure is too general to apply empirically, and thus, it is rare
to find any empirical studies which apply this unified functional approach of
incorporating demographic variables into demand systems. However, recently, Bollino,
Perali, and Rossi (2000) have applied Lewbel’s procedure to extend the Gorman
specification (1976). They proposed four specifications–Extended Scaling, Extended
Gorman (EG), Extended Reverse Gorman (ERG), and Extended Quadratic Gorman
(EQG). Chung (2001) utilized Lewbel’s modifying function procedure to show a more
general procedure than Lewbel’s. In addition, his procedure allows interpretations of
household production and subsistence expenditure, which is beyond the scope of this
study.
From an empirical viewpoint, it should be interesting to investigate which
procedure is the best for incorporating demographic variables. Unfortunately, the answer
is not unambiguous. For example, Pollak and Wales (1980) compared the DT with the
DS and concluded that the DS always provides a higher likelihood value than the DT,
even though the two procedures could not be tested since they are not nested. 5 Pollak and
Wales (1981) concluded that the modified Prais-Houthakker procedure was the best
among the five procedures, as mentioned earlier, on the basis of the likelihood value. The
data set used in both papers was taken from the British household budget data for the
years 1968-1972 (81 observations). On the other hand, Bollino et al. (2000) utilized the
Italian data to test the procedures and found that the DS was rejected against more
general specifications, such as the Gorman specification and the reverse Gorman
5
Demographic translating and demographic scaling can be “artificially” nested in the Gorman specification
(1976) by introducing an additional parameter of linear combination.
17
specification. In addition, the EG and the ERG are rejected against the EQG– the most
general procedure in their comparison. However, due to the limitation of the data, they
only incorporated one demographic attribute– household size, and concluded that “even
the most general procedure (the extended quadratic Gorman specification) is unable to
fully capture the behavioral heterogeneity shown by the data” (Bollino et al., p. 275).
From their conclusion, it seems obvious that one demographic variable is unable to fully
explain the whole heterogeneity. In this study, we will use the ‘ordinary budget share
scaling and translation’ procedure proposed by Lewbel (1985), to incorporate
demographic variables into a demand system. This procedure will allow us to explain the
heterogeneous nature in household consumption patterns.
2.2 Review of Econometric Models: Censored Demand Systems
Censored regression techniques were developed following the pioneering work of
Tobin (1958) who dealt with zero observation problems in univariate cases; however, a
widely used system of equations in econometrics is more complicated in multivariate
cases of demand systems. The techniques used to estimate a censored demand system can
be classified into a one-step and a two-step approach. 6 As to the one-step approach,
Wales and Woodland (1983) provided a procedure based on the Kuhn-Tucker conditions
of constrained utility maximization, while Lee and Pitt (1986) suggested a dual approach
based on an indirect utility function. Due to high dimensional integrals, the estimation in
empirical applications is extremely complicated. Simulation methods have been applied
to simplify the calculation of the integrals (Kao and Lee, 1996; Dong and Gould, 1999;
Hasan et al., 2001), e.g. using the GHK simulator (Geweke, 1991; Börsch-Supan and
6
Due to similarities in estimation and computation methods, Amemiya (1985, pp. 360-411) classified the
Tobit models into five key types according to the form of the likelihood function.
18
Hajivassiliou, 1993; Keane, 1994). Meanwhile, the estimation method based on the
Simulated Generalized Method of Moments is developed (e.g., Fahs et al., 2001). This
method is still very time-consuming even when using a more powerful, modern computer.
Alternative two-step approaches are developed to reduce the burden of calculation
of the multiple integrals. Heckman (1976) proposed a two-step estimator in a twoequation generalization of the Tobit model. Heien and Wessells (1990) extended
Heckman’s sample selection model to evaluate the inverse Mills ratio (IMR) for each
commodity and observation in the first step and used the IMR as an instrument to
estimate the full demand system in the second step. Their approach has been applied to
many empirical studies (see examples in Yen and Kan, 2000). However, Shonkwiler and
Yen (1999) pointed out the inconsistency of the Heien and Wessells’ estimator and
suggested an alternative two-step procedure. Similarly, there are alternative two-step
procedures proposed in the literature, for example, Perali and Chavas (2000) utilize a
minimum Chi-square method to overcome the computational burden.
In this study, a two-step estimation approach proposed by Shonkwiler and Yen
(1999) is adapted to simplify the computational burden and to satisfy the consistency
property in econometrics. However, one problem with Shonkwiler and Yen’s approach is
that the adding-up restriction can not be imposed. More discussion on their method will
be provided later.
2.3 Summary
This chapter reviews the three most important demand specifications which have
a profound influence on deriving more accurate demand elasticities both theoretically and
econometrically. Some of them are quite advanced and parts of them have already given
19
direction for improvement. On the basis of these previous studies, in the next chapter, an
economic model will be introduced followed by an econometric model specification.
20
CHAPTER 3
METHODOLOGY
In this chapter, an economic model is developed to incorporate demographic
attributes into weakly separable preferences using Barten-like group equivalence scales,
and then an econometric model is provided using the AIDS model as an example.
3.1 Economic Model
In order to pay specific attention to constructing a demand model for nondurable
goods and services (commodities in brief) using household data, it is assumed that, for
each household, preferences are weakly intertemporally separable, leisure is weakly
separable from commodities, and durable commodities are weakly separable from
nondurable commodities. For simplicity, a two-level demand structure is assumed as well,
i.e., between groups and within groups. Motivated by Deaton and Muellbauer (1980a) as
well as Lewbel (1989), the variables and notations used in this section are defined as
follows:
N= the number of groups;
G= specific group with G=1, 2, …, N;
nG= the number of goods in group G;
pGi and qGi= the price and quantity of the ith good in group G, respectively;
pG and qG= the vectors of prices and quantities in group G, respectively;
21
p and q= the vectors of all prices and quantities, respectively;
PG and QG= the price and quantity indices for group G, respectively;
P and Q= the vectors of price indices (PG) and quantity indices (QG), respectively;
XG=
∑p
i∈G
X=
∑X
Gi
G
qGi = total expenditure in group G;
= total expenditure;
wGi= pGiqGi/XG= the budget share of the ith good in group G, relative to total
expenditure in group G;
WG= XG/X= the budget share of group G;
A= a vector of demographical attributes that affect a household’s preferences; and
AR= some constant value of A for a reference household.
Under this structure, the weakly separable utility function for any household is
given by:
(3.1) U [u1 (q1 | A),K, u G (qG | A),K, u N (q N | A)] = U [u1 ,K, u G ,K, u N ] ,
where U[.] is called the “broad-group” or the “between-group” utility function for all
groups and uG(.) is the “within-group” sub-utility function with the corresponding utility
value, uG. Since the physical quantities for broad groups do not exist, we assume their
quantity and price indices (QG and PG for G=1,2,…,N) by defining QG = u G (qG | A) and
PG = X G / QG , respectively (similar to Lewbel, 1989). Hence, the between-group
allocation problem can be rewritten as:
(3.2) Max U [Q1 ,K, QG ,K, QN ] s.t.
{QG }
∑
G
PG × QG = ∑G X G = X .
22
This is (conceptually) a standard utility maximization problem. The solution to this
problem can be expressed as a Marshallian demand system:
(3.3) QG∗ = QG (P1 ,K, PG ,K, PN , X | A) = QG (P, X | A) .
By duality, the corresponding indirect utility function, expenditure function, and Hicksian
demand system are Ψ[P, X | A] , C [P,U | A] , and H G (P,U | A) , respectively.
Given the optimal quantity index for group G, i.e., utility level ( uG∗ = QG∗ )
determined in the between-group allocation problem (the first step), the within-group
commodity allocation problem can be expressed as:
(3.4) Min X G = ∑ pGi qGi
{ qGi }
i∈G
s.t. uG (qGi | A) ≥ u G∗ .
This cost minimization problem can be solved to determine the conditional Hicksian
compensated demand systems. 7 By duality, the Marshallian demand systems can be
expressed as:
∗
= qGi ( pG , X G | A) ∀i ∈ G .
(3.5) qGi
Again, by duality, the corresponding “within-group” indirect utility function, expenditure
function, and conditional Hicksian demand system are ψ G [ pG , X G | A] , C G [ pG , u G∗ | A] ,
and hGi ( pG , u G∗ | A) , respectively.
This construction of the two-step utility maximization program under the
assumption of weak separability places no restriction on the between-group utility
function U[.], no restriction on each of the sub-utility functions uG(.), and no restriction
on the incorporation of demographic attributes into both U[.] and uG(.). Any of these
7
The ‘conditional’ demand system means the demand system is determined given that the utility level in
the first step is known.
23
specifications can be complicated, but once each of these is specified, both within-group
and between-group demand systems can be estimated empirically (Lewbel, 1989).
However, weak separability imposes strong restrictions on Slutsky substitution terms
(Goldman and Uzawa, 1964). These restrictions are crucial to investigate the relationship
between commodities in different groups via conditional and unconditional elasticities.
Following Deaton and Muellbauer (1980a) and Carpentier and Guyomard (2001),
the unconditional elasticities can be derived by calculating the unconditional Slutsky
substitution terms. The unconditional Slutsky substitution terms sij, i ∈ G , j ∈ H and
G, H = 1,K, N , can be expressed as:
(3.6) sij = scij + λGH ×
∂qGi ( pG , X G | A) ∂q Hj ( p H , X H | A)
⋅
,
∂X G
∂X H
G≠H
 0,
where scij = 
, which is the conditional Slutsky
∂qGi ∂pGj + qGj ⋅ ∂qGi ∂X G , G = H
substitution term.
We need to determine λGH in equation (3.6). Following Deaton and Muellbauer (1980a)
and Carpentier and Guyomard (2001), λGH with G ≠ H can be expressed as follows:
(3.7) λGH = ∑ p Hj ⋅
j∈H
∂X G
∂p Hj
where
∂ ( PG ⋅ QG )
∂p Hj
u = const
where
∂PH
∂p Hj
=
since
∂X H
∂p Hj
u =const
u = const
u = const
= PG ⋅
= ∑ p Hj ⋅
j∈H
∂QG
∂p Hj
∂ ( X H QH )
∂p Hj
∂ ( PG ⋅ QG )
∂p Hj
u = const
u =const
=
= PG ⋅
1
QH2
u = const
∂H G ( P,U | A) ∂PH
⋅
∂PH
∂p Hj

∂X
⋅ QH ⋅ H
∂p Hj

u = const
− XH ⋅
∂QH
∂p Hj
u =const
=
= q Hj by Shephard’s lemma and
24
∂u H
∂p Hj
u = const
∂QH
∂p Hj
,
u = const
u = const
 q Hj
,
=
 QH
=0.
Hence, λGH =
λGH = PG ⋅
∑p
j∈H
Hj
⋅PG ⋅
∂H G ( P,U | A) q Hj
P ∂H G ( P,U | A)
⋅
= G ⋅
⋅ ∑ p Hj q Hj , i.e.,
∂PH
QH QH
∂PH
j∈H
∂H G ( P, U | A)
∂H G ( P, U | A) X H
⋅
= PG ⋅ PH ⋅
∂PH
QH
∂PH
,
and
thus
the
Slutsky
substitution term between two commodities in different groups (i.e., G ≠ H ) can be
rewritten as:
(3.8) sij = PG ⋅ PH ⋅
∂H G ( P,U | A) ∂qGi ( pG , X G | A) ∂q Hj ( p H , X H | A)
⋅
⋅
.
∂PH
∂X G
∂X H
Multiplying both sides by p Hj qGi , equation (3.8) becomes:
p Hj sij
qGi
=
PG QG p Hj q Hj
XG
XH
 PH ∂H G ( P, U | A)   X G ∂qGi ( pG , X G | A) 

 ⋅ 

∂X G
∂PH

  qGi
 QG
 X ∂q Hj ( p H , X H | A) 

⋅ H
q

∂
X
Hj
H


~
At the left hand side, p Hj sij qGi = Σ ij is the unconditional compensated elasticities for
commodity i ∈ G with respect to the price of commodity j ∈ H , where G ≠ H . At the
other side of the equation, PG QG = X G is described earlier in equation (3.2);
wHj = p Hj q Hj X H , the second element of the right hand side, is the conditional budget
share of commodity j in H; and the last three parts, according to elasticity formulas, are
~
Σ GH , η Gi , and η Hj , which mean the between-group compensated elasticities for the
quantity index of group G with respect to the price index of group H, the conditional
expenditure elasticities of commodities i ∈ G and j ∈ H , respectively. To simplify the
~
notation, the unconditional (Hicksian) compensated elasticities, Σ ij , for i ∈ G with
respect to the price of commodity j ∈ H , and G ≠ H can be expressed as:
25
~
~
(3.10) Σ ij = wHj ⋅ Σ GH ⋅ η Gi ⋅ η Hj .
If the two commodities (i and j) are in the same group G, then equation (3.6)
becomes:
(3.10) sij = scij + λGG ×
∂qGi ( pG , X G | A) ∂qGj ( pG , X G | A)
⋅
.
∂X G
∂X G
Using the property of
∑
G
λGH = ∑H λ HG = 0 (Deaton and Muellbauer, 1980a, p. 136;
Carpentier and Guyomard, 2001, p. 224), λGG is determined by:

∂H G ( P,U | A) 
 .
(3.11) λGG = − ∑ λGH = − PG ⋅  ∑ PH ⋅
P
∂
G≠H
G
≠
H
H


By using the homogeneity of compensated demand functions for between-group
commodities:
(3.12)
∑P
H
⋅
H
∂H G ( P,U | A)
∂H ( P,U | A)
∂H ( P,U | A)
= 0 ⇒ ∑ PH ⋅ G
= − PG ⋅ G
,
∂PG
∂PH
∂PH
G≠H
equation (3.11) becomes:

∂H ( P,U | A)
∂H ( P,U | A) 
 = PG ⋅ PG ⋅ G
(3.13) λGG = − ∑ λGH = − PG ⋅  ∑ PH ⋅ G
.
∂PG
∂PH
G≠H

 G≠ H
Therefore, the Slutsky substitution term between two commodities in the same group can
be rewritten as:
(3.14) sij = scij + PG ⋅ PG ⋅
∂H G ( P,U | A) ∂qGi ( pG , X G | A) ∂qGj ( pG , X G | A)
⋅
⋅
.
∂X G
∂X G
∂PG
~
Hence, the unconditional (Hicksian) compensated elasticities, Σ ij , for i, j ∈ G can be
expressed as:
~
~
(3.15) Σ ij = ε~ij + wGj ⋅ηGi ⋅η Gj ⋅ Σ GG ,
26
~
where η Gi and ηGj are the conditional expenditure elasticity of commodities i and j. Σ GG
is the between-group compensated elasticities for the quantity index of group G with
respect to its own price index. ε~ij is the conditional Hicksian price elasticity of
commodity i with respect to the price of commodity j. Combining equations (3.9) and
~
(3.15), the unconditional (Hicksian) compensated elasticities, Σ ij , for i ∈ G with respect
to the price of commodity j ∈ H , and G, H = 1,2,K, N can be expressed as:
~
~
(3.16) Σ ij = ε~ij + wHj ⋅η Gi ⋅η Hj ⋅ Σ GH , where ε~ij = 0 if G ≠ H .
This is identical to equation (20b) in Carpentier and Guyomard (2001, p. 226).
Using the Slutsky equation, the unconditional Mashallian elasticities, Σ ij ,
corresponding to equation (3.16) can be derived as follows:
By duality, the Hicksian demand is equal to the Marshallian demand, viz.
(
)
(
(
) )
(3.17) hGi pG , uG∗ | A = qGi pG , CG pG , uG∗ | A | A .
Suppose again that i ∈ G , j ∈ H , and G ≠ H . Taking the derivative on both sides of
equation (3.17) with respect to p Hj and multiply it by p Hj qGi , we have:
(3.18)
(
)
(
)
p Hj ∂hGi pG , uG∗ | A
p Hj ∂qGi ( pG , X G | A) p Hj ∂qGi ( pG , X G | A) ∂CG pG , uG∗ | A
=
+
,
qGi
∂p Hj
qGi
∂p Hj
qGi
∂X G
∂p Hj
 X ∂q ( p , X | A)  ∂CG ( pG , u G∗ | A) p Hj
~
 ⋅
i.e., Σ ij = Σ ij +  G ⋅ Gi G G
.
⋅
∂X G
∂p Hj
XG
 qGi

From equation (3.7), we know that:
(
)
∂CG pG , uG∗ | A
∂H ( P,U | A) q Hj
= PG ⋅ G
⋅
(3.19)
.
∂p Hj
∂PH
QH
Substituting (3.19) into (3.18), we get:
27
∂H ( P,U | A) q Hj p Hj
~
(3.20) Σ ij = Σ ij + η Gi ⋅ PG ⋅ G
⋅
⋅
∂PH
QH X G
 P ∂H G ( P,U | A)  p Hj q Hj PG QG
~
 ⋅
.
⇒ Σ ij = Σ ij + η Gi ⋅  H
⋅
∂PH
 QG
 PH QH X G
Since PGQG=XG and PHQH=XH, we can simplify equation (3.20) again as:
~
~
~
~
(3.21) Σ ij = Σ ij + η Gi ⋅ Σ GH ⋅ wHj ⇒ Σ ij = Σ ij − wHj ⋅η Gi ⋅ Σ GH .
~
Substituting equation (3.9) and Σ GH = Σ GH + WH ⋅η G , the Slutsky equation for broad
groups G and H, into equation (3.21), we have:
~
~
(3.22) Σ ij = wHj ⋅ηGi ⋅η Hj ⋅ Σ GH − wHj ⋅ηGi ⋅ Σ GH
⇒ Σ ij = wHj ⋅ηGi ⋅ (Σ GH + WH ⋅ηG ) ⋅ (η Hj − 1) .
On the other hand, for i, j ∈ G , equation (3.18) becomes:
 X ∂q ( p , X | A)  ∂CG ( pG , uG∗ | A) pGj
~
 ⋅
(3.23) Σ ij = Σ ij +  G ⋅ Gi G G
.
⋅
∂X G
∂pGj
XG
 qGi

∂CG ( pG , uG∗ | A)
= qGj , equation (3.23) can be rewritten as:
Using Shephard’s lemma,
∂pGj
pGj qGj
~
= Σ ij + η Gi ⋅ wGj .
(3.24) Σ ij = Σ ij + η Gi ⋅
XG
Therefore, the unconditional Marshallian price elasticities become
~
~
~
(3.25) Σ ij = Σ ij − wGj ⋅ηGi = ε~ij + wGj ⋅η Gi ⋅ηGj ⋅ Σ GG − wGj ⋅ηGi = ε ij + wGj ⋅ηGi ⋅η Gj ⋅ Σ GG ,
since ε~ij = ε ij + wGj ⋅ηGi .
Again, using the Slutsky equation in elasticity form for “between-group” commodities,
~
Σ GG = Σ GG + WG ⋅η G
28
(3.26) Σ ij = ε ij + wGj ⋅ηGi ⋅ηGj ⋅ (Σ GG + WG ⋅ηG ) .
Combining equations (3.22) and (3.26), the unconditional Marshallian elasticities, Σ ij ,
for i ∈ G with respect to the price of the commodity j ∈ H , and G, H = 1,2,K, N can be
expressed as:
(3.27) Σ ij = ε ij + wHj ⋅η Gi ⋅ (Σ GH + WH ⋅η G ) ⋅ (η Hj − 1 + δ GH ) , where ε ij = 0 if G ≠ H , and
δ GH = 1 , if G = H and δ GH = 0 , if G ≠ H .
Finally, the unconditional expenditure elasticity can be expressed as:
(3.28) η i =
X ∂qGi  X G ∂qGi
⋅
=
⋅
qGi ∂X  qGi ∂X G
  X ∂X G 
 X ∂QG 
 × 
 = η Gi × 
 = η Gi × η G ,
⋅
⋅
  X G ∂X 
 QG ∂X 
where η G is the expenditure elasticity of “between-group” commodity G.
To sum up, equations (3.16), (3.27), and (3.28) represent the unconditional
Hicksian compensated price elasticities, the unconditional Marshallian uncompensated
price elasticities, and the unconditional expenditure (income) elasticities, respectively.
Below, the economic model is summarized in a series of steps in order to build up
empirical procedures in this study. First, since there is no restriction on the sub-utility
function, any theoretically plausible demand functions, say the QAIDS, are viable
candidates. Second, several techniques can be used to incorporate demographic attributes
into any theoretically plausible demand system: (1) demographically modified parameters,
(2) any of the five general procedures described by Pollak and Wales (1992), or (3)
general procedures introduced by Lewbel (1985). We will use Lewbel’s general
procedure to incorporate demographic variables in this study. Third, the estimation of
subgroup demand systems can be used to recover uG(.) for all groups G, for G=1,…,N,
29
and thus can be used to construct PG for all G. Hence, the between-group utility function
U[.] can be specified and thus the between-group budget share functions can be estimated.
Last, the unconditional elasticities can be calculated by utilizing the within- and betweengroup elasticities from the two stages. The rest of this chapter will specify an econometric
model following the above four steps.8
3.2 Specification of Econometric Model
Following Pollak and Wales (1992), in this study, an econometric model is
specified for estimating food demand in urban China in four related steps: (1) select a
functional form; (2) determine a method for incorporating demographic variables; (3)
select estimators dealing with zero consumption problems (if serious); and (4) specify the
error structures. Since the QAIDS has properties of both a flexible functional form
(Fisher et al., 2001) and a nonlinear Engel function, which is more appropriate to
household data (Banks et al., 1997), the QAIDS is chosen in this study.
On the basis of the economic model, a three-stage utility maximization is assumed
to simplify the construction of the decision-making process for Chinese urban households.
The first stage is to make choices among the eight broad groups. The second stage is to
determine the utilities obtained from the seven food sub-groups. And in the last stage,
with the predetermined utility level for each sub-group, each household decides the
optimal consumption levels of commodities within each sub-group to minimize the cost
(more details on commodity groupings will be discussed in the next chapter). In addition,
the indirect utility functions are assumed to be identical for all stages of the utility
maximization framework.
8
This two-stage economic model of utility maximization can be extended to construct a multi-stage (three
or more) demand system.
30
3.2.1 Functional Form Specification: The Quadratic Almost Ideal Demand
System (QAIDS)
The QAIDS (Banks et al., 1997) has an indirect utility function in logarithm as:
(3.29) ln V =
1
,
[κ ( p, X )]−1 + λ ( p)
where κ ( p, X ) =
ln X − ln a( p )
, V is the indirect utility, p is a price vector and X
b( p )
represents the expenditure. In addition, a(p), b(p), and λ(p) are price aggregators to be
specified later. Its corresponding expenditure function in logarithm is given by:
(3.30) ln X = ln a( p) +
b( p )
,
[ln V ]−1 − λ ( p)
where
(3.31a) ln a( p) = α 0 + ∑k α k ln pk +
1
∑ ∑ γ jk ln pk ln p j ,
2 j k
(3.31b) b( p ) = ∏ p kβ k , and
k
(3.31c) λ ( p ) = ∑ λk ln pk .
k
Applying Roy’s identity to equation (3.29) or Shephard’s lemma to equation
(3.30), the QAIDS in share form can be expressed as:
2
 X  λi   X  
 +
 ,
⋅ ln
(3.32) wi = α i + ∑k γ ik ln p k + β i ln
 a ( p )  b( p )   a ( p )  
where α, β, γ, and λ are parameters to be estimated. When all the λ’s are zero, the QAIDS
reduces to the AIDS. Thus, the AIDS is nested in the QAIDS, which can be tested based
31
on the statistical significance of λ or other statistical tests such as the likelihood ratio
(LR) test, Wald test, etc.
The expenditure elasticity is provided by:
(3.33a) Ei = 1 +
where
1 ∂wi
⋅
,
wi ∂ ln X
 X 
λ
∂wi
.
= β i + 2 ⋅ i ⋅ ln
∂ ln X
b( p)  a( p ) 
The Marshallian price elasticities ( EijM ) can be computed by:
(3.33b) EijM = −δ ij +
∂wi
1
⋅
,
wi ∂ ln p j
0, if
where δ ij = 
1, if
i≠ j
, and
i= j
2
∂wi
λ   X 
 ∂wi 
 .
= γ ij − α j + ∑k γ jk ln p k ⋅ 
 − β j ⋅ i ⋅ ln
∂ ln p j
b( p )   a ( p ) 
 ∂ ln X 
(
)
The Hicksian compensated elasticities ( E ijH ) can be calculated by using the
Slutsky equation, i.e.,
(3.33c) E ijH = EijM + w j Ei .
In this study, the QAIDS will be applied to five food subgroups and one food
group, which will be discussed in depth later. Therefore, six demand systems will be
estimated independently.
3.2.2 Incorporation of Demographic Variables
As mentioned in the previous chapter, Lewbel (1985) proposed unified
approaches to incorporating demographic or other effects into demand systems.
32
Following Lewbel (1985, pp. 9-11), in this study, the modifying functions are specified,
called ‘ordinary budget share scaling and translation’ (OBSSAT, in theorem 8), for our
empirical analysis. The modifying functions are expressed as:
~
(3.34a) X = f [ X * , p, r ] = β (r ) ⋅ ( X * ) [ s ( r ) / α ( r )] P ( p, r ) ,
(3.34b) pi* = hi ( p, r ) = γ (r ) ⋅ piα ( r ) ,
n
~
where P ( p, r ) = ∏i =1 piri , pi* is ith modified price, X* is a modified expenditure, r is a
function of demographic characters A, and α(r), β(r), γ(r), and s(r) are some specific
functions of r, which will be specified later. By theorem 4 (Lewbel, 1985, pp. 4-5), the
demand system in share form can be derived as:
∂f [.] X *
(3.35) wi ( X , p, r ) =
∂X * X
where
∑
∂h j (.) p i * * * ∂f [.] pi
,
wj (X , p ) +
j =1
∂pi p *j
∂pi X
n*
s(r )
s (r ) X
∂f [.]
~
= β (r ) ⋅
⋅ ( X * ) [ s ( r ) / α ( r )]−1 ⋅ P ( p, r ) =
⋅
,
*
α (r )
α (r ) X *
∂X
r ~
∂f [.]
= β (r ) ⋅ ( X * ) [ s ( r ) / α ( r )] ⋅ i P ( p, r ) , and
∂pi
pi
∂h j (.)
∂pi
= γ (r ) ⋅
α (r )
pi
⋅ piα ( r ) =
α (r )
pi
⋅ hi ( p, r ) ∀i = j
.
= 0 ∀i ≠ j
Hence equation (3.35) becomes:
wi ( X , p, r ) =
p
s (r ) X X * α (r )
hi ( p, r ) *i wi* ( X * , p * )
*
α (r ) X X pi
pi
+ β (r ) ⋅ ( X )
* [ s ( r ) / α ( r )]
r ~
p
⋅ i P ( p, r ) i
pi
X
(3.36) wi ( X , p, r ) = s (r ) ⋅ wi* ( X * , p * ) + ri ,
33
, and thus,
which is the equation 20 in Lewbel (1985, p.10), where s (r ) = 1 − ∑ j =1 r j , ri ≥ 0 ∀i ,
n
0 < s(r ) ≤ 1 .9,10 This budget share is a function which is independent from w*j ∀j ≠ i and
the modifying function pi* = hi (.) is independent from pj for all j ≠ i.11
If we specify α (r ) = s (r ) = a and γ (r ) = β (r ) = 1 , as indicated in Lewbel (1985,
p.10), then equation (3.36) can be simplified as:
(3.37) wi ( X , p, r ) = a ⋅ wi* ( X * , p * ) + ri ,
n
n
K
~ ~
where a = 1 − ∑ j =1 r j , X * = X / P , P ( p, r ) = ∏i =1 piri , pi* = pia , and ri = ∑k =1 δ ik Ak .
Therefore, for an empirical analysis, a demand system with incorporation of
demographic variables (i.e., equation 3.37) can be estimated as long as wi* is specified. In
this study, the QAIDS model with demographic variables can be expressed as:
(3.38) wi ( X , p, r ) = a ⋅ wi* ( X * , p * ) + ri ,
2
 X∗ 
λi   X ∗  
 ,

where w ( X , p ) = α i + ∑k γ ik ln p + β i ln
+
⋅ ln
∗ 
∗
∗ 
 a ( p )  b( p )   a ( p )  
*
i
*
∗
k
*
ln a( p ∗ ) = α 0 + ∑i =1α i ln pi* + 1 / 2∑i =1 ∑ j =1γ ij ln pi* ln p *j ,
n
n
n
b( p ∗ ) = ∏ ( pk∗ ) β k , and
k
n
~
~
ln X * = ln( X / P ) = ln X − ln P = ln X − ∑ j =1 r j ln p j .
The expenditure elasticity is given by:
9
See more restrictions in Lewbel (1985, p.10).
A violation of r j ≥ 0 may happen in empirical studies.
10
11
In the Extended Gorman (Bollino et al., 2000) wi is a function of wj* (i.e., not independent from wj*.)
34
(3.39a) Ei = 1 +
where
1 ∂wi
⋅
,
wi ∂ ln X
 X∗ 
∂wi∗
λi
∂wi
∂wi∗

] .
= a⋅
and
=
⋅
+
⋅
⋅
a
[
2
ln
β
i
∂ ln X
∂ ln X ∗
∂ ln X ∗
b( p ∗ )  a ( p ∗ ) 
The Marshallian price elasticities ( EijM ) can be computed by:
(3.39b) EijM = −δ ij +
0, if
where δ ij = 
1, if
∂wi
1
,
⋅
wi ∂ ln p j
i ≠ j ∂wi
∂wi∗
∂wi∗
,
] , and
= a ⋅ [a ⋅
−
r
⋅
j
i = j ∂ ln p j
∂ ln p ∗j
∂ ln X ∗
(
∂wi∗
= γ ij − α j + ∑k γ jk ln pk∗
∂ ln p ∗j
2
 ∂wi∗ 
λi   X ∗  

 .
⋅ 
−
⋅
⋅ ln
β
j
∗ 
b( p ∗ )   a( p ∗ ) 
 ∂ ln X 
)
The Hicksian price elasticities ( EijH ) can be calculated by using the Slutsky
equation, i.e.,
(3.39c) E ijH = EijM + w j Ei .
The elasticities with respect to demographic variable Ak can be expressed as:
(3.39d) E Ai k =
Ak ∂wi
⋅
,
wi ∂Ak
∂wi
∂wi*
∂r
∂a
= a⋅
+ wi*
+ i ,
where
∂Ak
∂Ak
∂Ak ∂Ak
∂r j
∂ri
∂a
n
n
= δ ik ,
= −∑ j =1
= −∑ j =1 δ jk ,
∂Ak
∂Ak
∂Ak
∗
∂wi* ∂ ln p j
∂wi*
∂wi* ∂ ln X ∗
n
,
+ ∑ j =1
⋅
=
⋅
∂Ak ∂ ln X ∗ ∂Ak
∂ ln p ∗j ∂Ak
35
∂ ln p ∗j
∂ ln X ∗
n
n
= ∑ j =1δ jk ln p j ,
= − ln p j ⋅ (∑l =1δ lk ) , and
∂Ak
∂Ak
∂wi∗
∂wi∗
and
as discussed earlier.
∂ ln X ∗
∂ ln p ∗j
3.2.3 Two-step Estimator
If any of the demand systems in the two levels (between and within groups) have
fewer households with zero consumption values, a standard maximum likelihood (ML)
estimator or iterated seemingly unrelated regressor (SUR) can be used to estimate the
parameters. However, if the zero observation problems are severe, techniques dealing
with a censored demand system should be considered. Following Shonkwiler and Yen
(1999) and Yen and Kan (2000), a two-step estimation procedure can be expressed as
follows:
Consider a structure in which censoring of each commodity i is governed by a
separate stochastic process z it′ τ i + υ it such that
w ( p, X | θ ) + ε it , if z it′ τ i + υ it > 0
(3.40) wit =  it
,
otherwise
 0,
where wit denotes the observed expenditure share, θ represents all parameters in a certain
demand system, zit is a vector of exogenous variables, τi is a conformable parameter
vector, and εit and υit are random errors.
The system of demand equations in share form can be written as:
(3.41) wit = E (wit ) + ξ it = Φ (z it′ τ i )wit ( p, X | θ ) + δ iφ (z it′ τ i ) + ξ it ,
where ξ it = wit − E (wit ) , with E (ξ it ) = 0 and ξit is heteroscedastic with variance
(Shonkwiler and Yen, 1999)
36
(3.42)
{
}.
var(ξ it ) = σ i2 Φ( z it′ τ i ) + [1 − Φ( z it′ τ i )] wit2 ( p, X | θ ) ⋅ Φ( z it′ τ i ) + 2 wit ( p, X | θ )δ iφ ( z it′ τ i )
{
}
− δ z it′ τ iφ ( z it′ τ i ) + φ ( z it′ τ i )
2
i
2
Therefore, the system (3.42) can be estimated with a two-step procedure: (1) obtain ML
probit estimates τˆi using the binary outcomes wit=0 and wit>0; (2) calculate Φ ( z it′ τˆi ) and
φ ( z it′ τˆi ) and then estimate θ and δ1, δ2, …, δn in the system:
(3.43) wit = Φ ( z it′ τˆi )wit ( p, X | θ ) + δ iφ ( z it′ τˆi ) + η it ,
by the ML or SUR procedure. Since the right-hand side of the system (3.43) does not add
up to one, therefore, the second-step estimation of the system should be based on the full
n-vector. This also indicates that the adding-up condition is not satisfied.
Elasticities can be calculated by taking derivation of equation (3.43). 12 Namely,
(3.44)
∂E ( wi )
∂wi
,
= Φ ( z it′ τˆi ) ⋅
∂ ln p j
∂ ln p j
(3.45)
∂wi
∂E ( wi )
= Φ ( z it′ τˆi ) ⋅
.
∂ ln X
∂ ln X
3.3 Summary
This chapter attempts to propose an approach to analysis of food demand using
recent Chinese urban household data. Indirect utility functions for the two levels of
demand structure and an approach to incorporating demographic variables are specified.
A limited dependent variable approach is then applied to capture the large proportion of
zero observations.
12
The elasticity formulas (3.45-3.46) indicate that the explanatory variables Z’s in the first step estimation
do not include the variables, p and X, in the second step estimation, which is applied in this study.
37
CHAPTER 4
DATA SOURCES AND DESCRIPTION
In this chapter, we explain in depth the data base employed in this study. Section
4.1 briefly describes data sources and section 4.2 discusses and compares the descriptive
statistics of the related variables, including quantities, prices, and expenditures of the
selected food categories and some important demographic variables.
4.1 Data Sources
The data base is taken from the Urban Socio-Economic Survey Organization,
National Bureau of Statistics (NBS) in China. This is a national uniform survey
conducted in every province in China. The main content of the survey includes detailed
household demographic information, cash flow, quantities and expenditures of major
commodities purchased in the market, the employment of each household member, yearend housing condition and the ownership of durable goods. In total, there are 1,548
variables describing each household.
The survey selected households using a two-stage stratified systematic random
sampling scheme. In the first stage, the cities and counties were stratified by population
size and then randomly selected from each province. In the second stage, given the
selected cities and counties, sub-districts (e.g., a street or road) were randomly selected
and then sample households were selected from these sub-districts. In total, 25,000
households in 226 cities and counties in urban China were selected by the NBS.
38
Since the selected households needed to keep accounts for a successive three year
period and be interviewed by local NBS staff, a rotation sampling scheme (one third of
the old sample households were replaced by new sample households each year) was
conducted to ease the burden of the selected households. The local NBS staff collected
the records each month and summarized them into annual data for national reports and
other analyses. Hence, the data base contains annual information which is different from
other typical household surveys (e.g., the Nationwide Food Consumption Survey
conducted by the U.S. Department of Agriculture) covering a shorter period of time, e.g.,
one to two weeks. In addition, because of the rotation sampling scheme, it is possible to
construct panel data for dynamic analyses.
This study employs the NBS data base from three provinces for six years, 19931998. The three selected provinces are Shandong (near Beijing), Jiangsu (adjacent to
Shanghai) and Guangdong (adjacent to Hong Kong), which represent diverse patterns of
food consumption in China. Figure 4.1 shows a map of China and a geographical
relationship of the selected provinces with other provinces in China. These three
provinces are in the costal area and among the most prosperous provinces in China. The
implications derived from the analysis using these three provinces would provide a
prospective blueprint for the rest of the provinces in China as well as for China as a
whole.
From this urban household survey, the commodities (138 food items and 190 nonfood items) are classified into eight broad groups: (1) Food, (2) Clothing, (3) Household
Facilities, Articles and Services, (4) Medicine and Medical Services, (5) Transport, Postal
and Communication Services, (6) Recreation, Education and Cultural Services, (7)
39
Housing and Utilities, and (8) Miscellaneous Commodities and Services. A utility tree is
used to present the relationships among these eight broad groups and their related
commodities. These eight broad groups are shown at the top layer of a utility tree in
Figure 4.2. Under each broad group, two or three sub-layers are used to aggregate the
data. Taking the ‘food’ broad group as an example, nineteen sub-groups, including grains,
vegetables, meats and poultry, and food away from home (FAFH), are used to represent a
second layer in the utility tree. This utility tree can be further divided into several layers.
For example, the ‘meats and poultry’ sub-group consists of one layer with nine basic
items (such as pork, beef, live chicken, etc.), whereas the ‘vegetables’ sub-group contains
two more layers with three sub-subgroups (fresh vegetables, dried vegetables, and
processed vegetables) in the third layer. In addition, the ‘fresh vegetables’ sub-subgroup
consists of 26 basic items (e.g., cabbage, carrots, tomatoes, etc.) in the fourth (the lowest)
layer, as shown in Figure 4.2.
However, the utility tree in Figure 4.2 does not match the purpose of our study.
To focus on food consumption in urban China, a weak separability assumption is made
on the basis of the Chinese food guide pyramid and dietary guidelines (Chen and Ge,
2000). A modified utility tree is shown in Figure 4.3. The broad group of food in Figure
4.3 contains six sub-groups, viz., (1) grains, (2) vegetables and fruits, (3) animal foods, (4)
dairy and bean products, (5) fats, oils and sweets, and (6) others, which include those
consumed-at-home food items, not being considered in the five sub-groups of interest and
FAFH. In this modified utility tree, each food subgroup consists of two to five food items,
respectively. The descriptive analysis in the next section will focus on food consumption
in the first five food subgroups.
40
4.2 Data Description
In this section, the descriptive statistics of the commodities in different layers of
the utility tree and some important demographic variables are compared among the three
provinces over a six-year period with an emphasis on the most recent dataset in 1998.
Shandong, Jiangsu, and Guangdong represent the northern, middle, and southern
regions, respectively, as shown in Figure 4.1. Table 4.1 shows the number of sample
observations drawn from each selected city and county in each of the three provinces for
each year. For example, three cities and two counties were chosen in Guangdong; six
cities and four counties in Jiangsu and five cities and three counties in Shandong. Within
each province, sample sizes are different for cities and counties according to its
population. Each year, 800 household data in Jiangsu, 650 in Shandong, and 600 in
Guangdong were collected respectively. The total sample size employed in this study is
over 12,000.
Table 4.2 shows the per capita total living expenditure in current Yuan (TLE) and
the expenditure shares among eight broad groups in Shandong, Jiangsu, and Guangdong,
respectively. Generally, the TLE’s in these provinces rose steadily from 1993 to 1998. In
addition, the average of the TLE in Guangdong is greater than that in Jiangsu and
Shandong. For example, in 1998, the TLE in Guangdong was more than ten thousand
Yuan, whereas the averages of the TLE in Jiangsu and Shandong were only 5,467 and
4,259 Yuan, respectively. However, the allocation of the TLE is similar in the three
provinces according to the expenditure shares for these eight broad groups. Overall, the
expenditure share for food took the largest portion, which accounted for over 40% of the
TLE, with over 50% in Jiangsu. The changing pattern of the expenditure share for food
41
reclaims Engel’s Law that expenditure share on food decreases when income increases.
However, the expenditure share of food in Jiangsu increased in the early years from 53%
in 1993 to 57% in 1995, the only evidence violating Engel’s Law. On the contrary, the
expenditure shares of broad groups 4-8 present an overall increasing trend. Even though
expenditure share of food is declining, it still accounts for the largest part of the TLE,
which provides an important reason to study food consumption in urban China.
Table 4.3 presents the total food expenditure (TFE) and its expenditure shares for
six food subgroups in the three provinces. The TFE has a pattern similar to the TLE, i.e.,
the TFE in Guangdong was still larger than those in Jiangsu and Shandong; however the
TFE did not increase steadily but fluctuated after 1996. The TFEs for Shandong and
Guangdong reached their peak in 1996 and 1997 respectively, meaning that households
in urban China started experiencing a transition of quantity-based to quality-oriented food
consumption patterns. As to expenditure shares for the six food subgroups, the patterns
are mixed for the six year period, which is similar in the three provinces. The first five
food subgroups of interest in this study account for over 50% of the TFE; except for
Guangdong from 1996 to 1998, with slightly below 50%. This indicates that other foods,
including food away from home, have increased their proportion of food expenditure,
which would be an important topic to study. However, due to data limitations, this study
focuses on food items which are consumed at home. Among the five food subgroups,
most of the households spent 6% to 12% on grains, 14% to 17% on vegetables and fruits,
and 20% to 26% on animal foods (such as pork, poultry, and aquatic products), which
represent the major three subgroups in the TFE. Food subgroup 4 accounted for only 1%
to 3% of the TFE. In addition, households in Jiangsu spent more on animal foods
42
percentage-wise as compared to the other two provinces. Households in Shandong, on the
other hand, spent more than the other two provinces on food subgroup 5 (fats, oils, and
sweets) with over 8% for the period. This may indicate different preferences on food
consumption among the three provinces in urban China.
Tables 4.4-4.6 present quantity consumed for eighteen food items among the three
provinces from 1993 to 1998. In general, food consumption patterns in Shandong are
different from those in Jiangsu and Guangdong even though Shandong is adjacent to
Jiangsu geographically. Among the eighteen food items, fresh vegetables are the most
consumed basic food item with more than one hundred kilograms per capita each year. In
terms of quantity consumed, fresh fruits are the second most consumed food item in
Shandong but the third in both Jiangsu and Guangdong. Flour is the second most
consumed food item in Shandong whereas rice is the second in both Jiangsu and
Guangdong. In addition, potatoes are eaten more than ten kilograms per capita in
Shandong but about nine kilograms in Jiangsu and under six in Guangdong. As for
animal food items, pork was the most common choice in all three provinces with an
average of thirteen kilograms in Shandong and twenty kilograms in both Jiangsu and
Guangdong, whereas beef and mutton are the least with an average lower than three
kilograms. However, in Shandong, the quantity consumed for eggs is more than pork
with the volume around seventeen kilograms per capita. More than ten kilograms of
poultry is consumed in Guangdong but less than three kilograms in Shandong. The
quantity consumed of aquatic products is similar in both Jiangsu and Guangdong with
more than seven kilograms but less than five in Shandong. Dairy products are not
commonly consumed by Chinese; however, people in urban China are more aware of the
43
nutrition values from dairy products and thus from 1993 to 1998 they consumed more
and more dairy products regardless of where they lived. Bean products are rich in protein
and are usually a substitute for dairy products, but their consumption levels remained
constant between 1993 and 1998. Province-wise, Jiangsu households consumed the most
in bean products with a per capita consumption of nearly ten kilograms each year. As for
fats and oils, sugar, nuts, and cakes, the patterns did not change much during the six years
of the survey period. To sum up, some common patterns are found among the three
provinces during the six year period, such as similar consumption levels of fresh
vegetables, pork, and dairy products. However, the overall consumption patterns of
households in Jiangsu and Guangdong are quite similar but different from Shandong.
The mean prices for the 18 basic food items in Shandong, Jiangsu, and
Guangdong are shown in Tables 4.7-4.9, respectively. For each household, these prices
(unit values) are calculated by dividing expenditures (in Yuan) by their quantity
consumed (in Kilograms). Prices of the food items which are not consumed by a
household are replaced with the mean prices of the city or county to which a household
belongs. Among these three provinces there exists a similar trend in prices for the
selected food items, i.e., the prices increase year by year with people in Guangdong
paying higher prices than their counterparts in the other two provinces. In addition, prices
for animal food items, e.g., pork, beef and mutton, and aquatic products, are higher than
the other food items. However, prices for cakes were also expensive especially in
Guangdong from 1996 to1998 with average prices over twenty Yuan/kg. There are
apparently regional differences in food prices in urban China.
44
Households with zero consumption present one of the most critical problems
facing econometricians when using household data. Under the utility tree proposed in this
study, it is interesting to investigate the seriousness of the zero consumption problems in
the data base. Tables 4.10-4.12 show the proportions of households with zero
consumption in the three levels of the utility tree in the three provinces, 1993-1998. For
the eight broad groups, no households had zero food consumption, whereas medicine and
medical service as well as miscellaneous commodities and services in Jiangsu
encountered a large portion of observations with zero consumption. In the early years,
Shandong had a more serious zero consumption problem in transport, post and
communication services, about 14% in 1993-1995. As to Guangdong, zero consumption
problems in the first layer are not as serious as the other two provinces. As for the second
layer, the six food sub-groups encountered very few households with zero consumption.
The reason is because these sub-groups are aggregated from the basic food items. This
aggregation operation dilutes the zero consumption. However, this problem can be easily
found within the 18 basic food items. Milk and dairy products have the largest number of
households with zero consumption showing 95% for yogurt in Shandong, 1995. However,
the proportion of households with zero consumption decreased dramatically over the
1993-1998 period in all three provinces. Besides milk in 1998, Shandong had more
households with zero consumption in sugar and beef and mutton with over 30%, whereas
Jiangsu had over 40% of households with zero consumption in flour and beef and mutton.
As to Guangdong, the zero consumption problems for flour were serious (72.7% in 1998).
Demographic variables are important factors to explain heterogeneous
consumption patterns. In this study, key demographic variables are selected and
45
prioritized on the basis of previous studies and our own judgment. The selected
demographic variables employed in the analysis are shown in Table 4.13, which includes
disposable income, household size, number of children under age 17, gender, age, and
education level of the householder, and ownership of a refrigerator.13 As discussed earlier,
disposable income in Guandgond is the highest with an average of over twelve thousand
Yuan in 1998; meanwhile, disposable incomes in both Shandong and Jiangsu are similar
with an average between five to six thousand Yuan. Household components, including
household sizes, the number of wage earners, and the number of children under age 17
are similar in the three provinces. In addition, more households in Guangdong own a
refrigerator and have larger housing space than Jiangsu and Shandong. The average age
of the householder in the three provinces is over 40 years old, and in Jiangsu, the average
is close to 50 years old.
In order to make a comparison on how the selected demographic variables affect
consumption patterns, the database from 1998 is utilized to discuss the differences among
the three provinces. Table 4.14 shows the provincial differences in the quantities of 18
foods. Per capita consumptions of grains and fruits in Shandong are higher than the other
two provinces except for rice. Per capita rice consumption in Shandong is less than one
fourth of that in either Jiangsu or Guangdong; however, flour consumption in Shandong
is at least three times of that in Jiangsu and over ten times of that in Guangdong. As to
animal food products, different patterns are shown in Table 4.14. Among the three
provinces, households in Shandong consumed the most eggs (17.33 Kg); households in
13
Disposable income should be an economic variable instead of a demographic variable; however, in this
study, disposable income is utilized to classify households into three groups for comparison. Hence, to
simplify the analysis, disposable income is treated as a demographic variable.
46
Jiangsu consumed the most pork (20.51 Kg) and aquatic products (near 10 Kg); and
households in Guangdong consumed the most beef and mutton (2.60 Kg) and poultry
(over 10 Kg). In addition, people in Jiangsu consumed more dairy and bean products than
the other two provinces. People in Jiangsu consumed more fats and oils at 9.13 Kg per
capita whereas in Guangdong people consumed only 2.81 Kg. People in Shandong
consumed more nuts and cakes than the other two provinces with an average of 4.41 Kg
and 5.67 Kg, respectively. These descriptive statistics clearly show regional differences
of food consumption in urban China.
Table 4.15 shows that approximately two thirds of the households consist of three
members, accounting for 23% in both Shandong and Jiangsu and 17% in Guangdong,
followed by two or four persons in an average household. Single-person households and
those with five or more members are not commonly seen. In this study, more attention is
paid to household sizes of two to four members which will be discussed in greater detail
in the next chapter.
The per capita food consumptions with different household sizes are compared in
Table 4.16. Most of the food items showed a decreasing trend of per capita consumption
with respect to household size. Some of the food items reached a peak where there were
one or two persons in a household; however, the difference in food consumption patterns
between households of one or two persons is very marginal. This is understandable since
most of these household members are adults. In addition, we need to note that there were
only 0.24% of households with a single person. As to a household with more than three
persons, usually the additional member is either a child or an elder family member;
dramatically reduced per capita food consumption is expected since these additional
47
family members usually do not need as much food as the two adult members.
Table 4.17 shows the percentage of households with a varying number of children
in the household, which is an important factor affecting food consumption especially on
some food items such as dairy products. Children have different needs and food choices
as compared with adults. It should be noted that urban China is affected by the one-child
policy, as a result, the survey data show that there were less than 4% of households
having more than one child in 1998 and approximately one third of households had no
children, especially in Jiangsu. This phenomenon may become a serious problem to
China in terms of demographic and social issues, which is beyond the scope of this study.
The per capita food consumptions with respect to the number of children are
presented and compared in Table 4.18. The average per capita food consumptions of
households with no children are higher than those with children among most of the food
items, as expected. The effect of children, from this database, is significant since children
usually do not require as much energy intake as adults. Households with three children
consumed the most beef, sugar, and cakes. This finding should be viewed cautiously
since the sample size is extremely low, accounting for only 0.24% of the sample.
Considering the income factor, the entire sample was divided into three income
groups: low, middle, and high, with roughly 20%, 60%, and 20% of the sample,
respectively. The households with income ranging between 4,320 and 10,140 Yuan per
capita in 1998 are classified as being in the middle income group. From Table 4.19,
Shandong has the largest percentage of households in the low income group whereas
Guangdong has 14.59% households in the high income group. Hence, this table shows
that among the three provinces, households in Guangdong were richer than those in
48
Jiangsu and Shandong, which is the same as was discussed earlier.
Tables 4.20-4.22 summarize the consumption of 18 food items by income groups
in the three provinces. Generally, the consumption volumes of most of the 18 food items
increase when income climbs upward, especially for fresh vegetables, fresh fruits, and
fresh milk; however, there are differences among the three provinces. As income
increases, per capita rice consumption increases in Shandong but decreases in Guangdong.
Flour consumption decreases sharply in both Shandong and Jiangsu but increases
marginally in Guangdong as income increases. As to animal protein food items,
consumption of all five items in Shandong increases when income changes from low to
middle but decreases from middle to high income levels. Most of the animal food items
in Jiangsu show a rising trend except for beef and mutton. But only eggs and aquatic
products increase as income increases in Guangdong. As to dairy and bean products, milk
increases sharply as it moves from a low to a higher income group. For example, the
average per capita fresh milk consumption in Jiangsu is 3.41 Kg for the low income
group but 16.17 Kg for the high income group. Yogurt is similar to fresh milk.
Consumption of fats and oils in Jiangsu and Guangdong increases as income increases;
however, Shandong shows a decreasing trend. All three provinces show an increasing
trend in nuts and cakes when income groups move from low to high. Sugar, however, has
a mixed relationship among the three provinces. To sum up, income, as indicated by
economic theory, has a critical influence on food consumption. Clearly, as income
increases, urban Chinese households appear to consume more fresh vegetables, fresh
fruits, fresh milk, nuts, and cakes, but the relationships are mixed for animal products.
Householders play an important role in food consumption. Several characteristics
49
about householder have been studied in the literature, e.g., Bhandari and Smith (2000).
The most important factors considered in this study are age, gender, and education level
of the householder.
According to dietary guidelines published in China, households are divided into
three groups according to the age of the householder. Group 1 consists of households
with the householder below 45 years of age; group 2 is between 45 and 60 and the last
group includes those over 60 years of age. Table 4.23 shows the distribution of
households among age groups by provinces. Most householders in Shandong are aged
below 45. Jiangsu has the most householders over 60 years old. In Guangdong,
households are distributed evenly between low and middle age groups, presenting a
similar situation to Jiangsu but having a smaller percentage of householders over 60.
Even though the age distribution is different among the three provinces, according
to Tables 4.24-4.26, the food consumption patterns are quite similar, i.e., most of the food
items increase their volume when the householder gets older, especially in Shandong. For
example, fifteen out of eighteen food items in Shandong present an increasing trend as
age goes up. Jiangsu and Guangdong show a similar pattern. The increasing trend may be
explained by a testable assumption that when householders get older, they prefer to eat at
home instead of eating in restaurants. However, dairy products, including fresh milk and
yogurt, exhibit a decreasing trend in Guangdong when age group gets older.
As shown in Table 4.27, 65% of householders are male. Jiansgu, unlike the other
two provinces, has fewer female householders in the sample. However, the consumption
differences in volume between male and female householders are not significant (Table
4.28). Specifically, only households with female householders in Shandong consume
50
more cakes than male householders. In both Jiangsu and Guangdong, households having
female householders consume more than male householders for seven out of 18 food
items, including fresh fruits, yogurt, and cakes. Again, the difference is marginal.
The distribution among low, middle, and high educational levels of household
heads (Table 4.29) are similar between Jiangsu and Guangdong, whereas Shandong has a
lower rate of householders with a low education level and have marginally more
householders with high education levels. Generally speaking, most householders have a
middle education level, which corresponds to a high school level or equivalent. Education
is usually an indicator of knowledge, i.e., householders with a high education level
possess more knowledge about nutrition and the impact of food consumption on health.
In addition, high education usually implies high income. It would be interesting to
investigate whether these two factors have an interaction in explaining food consumption
patterns in urban China.
Tables 4.30-4.32 present and compare the trends of 18 food items with respect to
education levels. Generally, the higher the education level, the lower the consumption per
capita. There are eleven, fourteen, and eight out of eighteen food items presenting a
decreasing pattern in Shandong, Jiangsu, and Guangdong, respectively. As for grains,
fresh vegetables, and fresh fruits, Shandong and Jiangsu show similar trends with all
having decreasing food items except for fresh fruits; however, in Guangdong, the trend is
different. As the education level increases, only rice and fresh vegetables decrease, but
the other four food items, such as flour, coarse grains, potatoes, and fresh fruits, increase.
As for animal protein food items, the trends among provinces are mixed. Jiangsu has all
food items with a decreasing trend when the education level increases. In Shandong, pork
51
and eggs decrease, whereas in Guangdong, pork, poultry, and aquatic products decrease
but eggs increase. The consumption of dairy products in high education level households
is the largest among the three provinces. As for the group of fats and oils, sugar, nuts, and
cakes, the general trend is decreasing as the education level increases.
In 1998, most of the households had refrigerators. From Table 4.33, only 12.59%
of households did not own a refrigerator. Ownership of refrigerators is an indicator of
modernization. A refrigerator is crucial to the consumption of dairy products and key to
some nondurable food items, such as fresh vegetables and fruits. As a result, from Table
4.34, the consumption of fresh milk and fresh fruits increases when a household owns a
refrigerator among the three provinces. In Guangdong, there are only five food items
which households having a refrigerator consume lower than those without a refrigerator.
This pattern is even clearer for grains in Jiangsu.
4.3 Summary
To sum up, this chapter describes the data sources and compares the descriptive
statistics. The demographic variables are important factors in explaining food
consumption in China, including household size, number of children in a household,
income groups, ownership of refrigerators, and some characteristics of householders,
such as age, gender, and education level. Each factor has its unique impact on food
consumption. In the next chapter, we will use econometric models to incorporate these
demographic variables into a demand system to reflect their impact on food demand
analysis.
52
ID
1
2
3
4
5
6
7
8
9
Total
a
Guangdong a
City/County
Sample Size
Guangzhou
300
Zhanjiang
100
Shenzhen
100
Puning
50
Shunde
50
600
Jiangsu b
City/County
Sample Size
Nanjing
200
Wuxi
100
Xuzhou
100
Nantong
100
Yangzhou
100
Yixing
50
Suqian
50
Taixing
50
Dafeng
50
800
Shandong c
City/County
Sample Size
Jinan
100
Qingdao
100
Jining
100
Dezhou
100
Qingzhou
100
Wendeng
50
Zhucheng
50
Weishan
50
650
In Guangdong, IDs 1-3 are major cities and IDs 4 and 5 are counties inside a major city.
b
In Jiangsu, IDs 1-5 are major cities while IDs 6-9 are counties inside a major city.
c
In Shandong, IDs 1-5 are major cities and IDs 6-8 are counties inside a major city.
Table 4.1: Identification of Cities or Counties and Sample Sizes (Each Year) in
Guangdong, Jiangsu, and Shandong.
53
Item a, b
Shandong
TLE
(in Yuan)
W1
W2
W3
W4
W5
W6
W7
W8
Year
1993
1994
1995
1996
1997
1998
2,019
(897)
0.502
(0.150)
0.178
(0.093)
0.088
(0.102)
0.021
(0.041)
0.024
(0.048)
0.087
(0.102)
0.058
(0.045)
0.021
(0.041)
2,742
(1,438)
0.493
(0.151)
0.182
(0.101)
0.085
(0.096)
0.028
(0.047)
0.031
(0.072)
0.073
(0.078)
0.065
(0.055)
0.028
(0.047)
3,520
(1,717)
0.485
(0.151)
0.172
(0.095)
0.096
(0.108)
0.030
(0.047)
0.037
(0.077)
0.077
(0.076)
0.063
(0.054)
0.030
(0.047)
4,028
(1,818)
0.474
(0.143)
0.171
(0.092)
0.079
(0.092)
0.037
(0.060)
0.043
(0.070)
0.090
(0.091)
0.069
(0.063)
0.037
(0.060)
4,096
(1,813)
0.453
(0.134)
0.164
(0.086)
0.079
(0.096)
0.039
(0.054)
0.050
(0.075)
0.098
(0.090)
0.080
(0.062)
0.039
(0.054)
4,259
(2,288)
0.440
(0.136)
0.143
(0.081)
0.076
(0.092)
0.040
(0.051)
0.060
(0.078)
0.117
(0.104)
0.082
(0.056)
0.040
(0.051)
Jiangsu
TLE
(in Yuan)
2,715
3,618
4,313
4,761
5,082
5,467
(1,456)
(1,924)
(2,125)
(2,326)
(2,424)
(3,591)
0.535
0.549
0.566
0.555
0.516
0.507
W1
(0.162)
(0.165)
(0.159)
(0.154)
(0.151)
(0.153)
0.136
0.130
0.120
0.110
0.109
0.098
W2
(0.082)
(0.080)
(0.077)
(0.075)
(0.072)
(0.068)
0.086
0.084
0.077
0.072
0.070
0.074
W3
(0.099)
(0.110)
(0.099)
(0.089)
(0.084)
(0.096)
0.015
0.016
0.019
0.023
0.028
0.032
W4
(0.030)
(0.027)
(0.035)
(0.041)
(0.046)
(0.059)
0.041
0.046
0.039
0.047
0.054
0.051
W5
(0.077)
(0.078)
(0.064)
(0.069)
(0.068)
(0.053)
0.075
0.068
0.075
0.076
0.092
0.100
W6
(0.086)
(0.072)
(0.084)
(0.082)
(0.090)
(0.095)
0.069
0.066
0.067
0.080
0.085
0.094
W7
(0.052)
(0.053)
(0.053)
(0.055)
(0.058)
(0.066)
0.015
0.016
0.019
0.023
0.028
0.032
W8
(0.030)
(0.027)
(0.035)
(0.041)
(0.046)
(0.059)
a
The eight broad groups are: (1) Food, (2) Clothing, (3) Household Facilities, Articles and Services, (4)
Medicine and Medical Services, (5) Transport, Postal and Communication Services, (6) Recreation,
Education and Cultural Services, (7) Housing and Utilities, and (8) Miscellaneous Commodities and
Services.
b
Standard deviations are in parentheses.
Table 4.2: Total Living Expenditure (TLE) and Expenditure Shares among Eight Broad
Groups in Urban China, 1993-1998.
(Continued)
54
Table 4.2: Continued
Item a, b
Guangdong
TLE
(in Yuan)
Year
1993
1994
1995
1996
1997
1998
4,793
6,949
8,139
9,040
9,356
10,084
(2,890)
(5,568)
(5,151)
(5,540)
(5,317)
(6,736)
0.554
0.534
0.534
0.517
0.495
0.476
W1
(0.162)
(0.171)
(0.159)
(0.156)
(0.145)
(0.152)
0.074
0.070
0.068
0.071
0.063
0.059
W2
(0.055)
(0.051)
(0.052)
(0.051)
(0.043)
(0.055)
0.086
0.095
0.079
0.075
0.068
0.070
W3
(0.089)
(0.098)
(0.082)
(0.071)
(0.071)
(0.076)
0.025
0.027
0.032
0.033
0.039
0.039
W4
(0.039)
(0.037)
(0.044)
(0.044)
(0.053)
(0.047)
0.056
0.051
0.051
0.049
0.056
0.070
W5
(0.082)
(0.077)
(0.068)
(0.057)
(0.057)
(0.070)
0.080
0.089
0.090
0.102
0.110
0.110
W6
(0.076)
(0.091)
(0.084)
(0.093)
(0.089)
(0.082)
0.084
0.090
0.094
0.102
0.111
0.124
W7
(0.071)
(0.093)
(0.078)
(0.076)
(0.079)
(0.084)
0.025
0.027
0.032
0.033
0.039
0.039
W8
(0.039)
(0.037)
(0.044)
(0.044)
(0.053)
(0.047)
a
The eight broad groups are: (1) Food, (2) Clothing, (3) Household Facilities, Articles and Services, (4)
Medicine and Medical Services, (5) Transport, Postal and Communication Services, (6) Recreation,
Education and Cultural Services, (7) Housing and Utilities, and (8) Miscellaneous Commodities and
Services.
b
Standard deviations are in parentheses.
55
Item a, b
Shandong
TFE (in Yuan)
WF1
WF2
WF3
WF4
WF5
WF6
1993
1994
1995
1996
1997
1998
964.49
(425.09)
0.077
(0.053)
0.163
(0.054)
0.232
(0.072)
0.017
(0.018)
0.088
(0.034)
0.423
(0.107)
1,271.20
(580.54)
0.098
(0.063)
0.163
(0.050)
0.229
(0.071)
0.015
(0.015)
0.097
(0.039)
0.399
(0.110)
1,613.65
(759.82)
0.098
(0.060)
0.164
(0.050)
0.231
(0.072)
0.016
(0.018)
0.090
(0.037)
0.400
(0.115)
1,810.41
(808.88)
0.094
(0.060)
0.162
(0.050)
0.224
(0.072)
0.019
(0.020)
0.083
(0.034)
0.418
(0.113)
1,756.74
(745.28)
0.074
(0.049)
0.165
(0.050)
0.228
(0.070)
0.021
(0.022)
0.084
(0.035)
0.427
(0.111)
1,737.02
(769.70)
0.066
(0.044)
0.151
(0.049)
0.218
(0.068)
0.026
(0.024)
0.087
(0.037)
0.451
(0.111)
1,349.50
(629.87)
0.094
(0.048)
0.145
(0.042)
0.261
(0.076)
0.018
(0.020)
0.074
(0.029)
0.408
(0.122)
1,808.29
(767.50)
0.109
(0.058)
0.149
(0.044)
0.262
(0.075)
0.018
(0.020)
0.076
(0.030)
0.386
(0.121)
2,278.91
(947.04)
0.117
(0.060)
0.149
(0.046)
0.261
(0.074)
0.018
(0.022)
0.072
(0.029)
0.383
(0.123)
2,473.69
(1,040.71)
0.117
(0.059)
0.149
(0.040)
0.257
(0.072)
0.021
(0.024)
0.068
(0.028)
0.388
(0.116)
2,440.01
(977.59)
0.079
(0.045)
0.145
(0.043)
0.259
(0.077)
0.024
(0.027)
0.069
(0.029)
0.424
(0.115)
2,506.57
(1,150.16)
0.083
(0.051)
0.140
(0.045)
0.238
(0.077)
0.028
(0.029)
0.073
(0.031)
0.438
(0.128)
Jiangsu
TFE (in Yuan)
WF1
WF2
WF3
WF4
WF5
WF6
Guangdong
TFE (in Yuan)
2,361.95
3,160.93
3,837.79
4,161.78
4,195.25
4,185.14
(883.50)
(1,263.44)
(1,530.16)
(1,597.27)
(1,703.47)
(1,710.61)
0.064
0.073
0.071
0.066
0.062
0.056
WF1
(0.033)
(0.039)
(0.037)
(0.036)
(0.032)
(0.032)
0.155
0.147
0.149
0.144
0.148
0.144
WF2
(0.052)
(0.046)
(0.041)
(0.042)
(0.047)
(0.047)
0.233
0.230
0.214
0.216
0.218
0.202
WF3
(0.085)
(0.083)
(0.073)
(0.078)
(0.087)
(0.083)
0.010
0.011
0.010
0.010
0.012
0.016
WF4
(0.016)
(0.015)
(0.014)
(0.011)
(0.015)
(0.018)
0.056
0.057
0.057
0.055
0.057
0.057
WF5
(0.026)
(0.027)
(0.026)
(0.025)
(0.025)
(0.024)
0.482
0.482
0.499
0.510
0.502
0.526
WF6
(0.129)
(0.129)
(0.117)
(0.120)
(0.129)
(0.126)
a
The six food subgroups: (1) grains, (2) vegetables and fruits, (3) animal foods, (4) dairy and bean products,
(5) fats, oils and sweets, and (6) others.
b
Standard deviations are in parentheses.
Table 4.3: Total Food Expenditure (TFE) and Food Expenditure Shares for Six Food
Subgroups in Urban China, 1993-1998.
56
Food
Item a, b, c
Year
1993
1994
1995
1996
1997
1998
14.83
14.87
13.30
15.57
14.78
13.71
Q1
(16.39)
(16.69)
(14.56)
(16.13)
(16.73)
(14.95)
32.07
38.88
38.13
35.99
27.56
24.18
Q2
(33.30)
(34.85)
(33.43)
(30.90)
(26.53)
(24.54)
3.029
2.826
2.834
2.528
2.395
2.440
Q3
(7.393)
(5.719)
(7.017)
(4.271)
(4.402)
(4.706)
11.78
11.80
12.61
13.77
13.04
12.96
Q4
(9.74)
(10.11)
(9.62)
(10.32)
(8.96)
(9.21)
106.95
110.49
107.82
112.57
107.00
103.96
Q5
(56.39)
(63.23)
(58.21)
(59.85)
(48.39)
(46.47)
59.73
60.40
59.55
62.94
74.26
74.77
Q6
(39.37)
(67.59)
(39.25)
(38.45)
(40.98)
(40.26)
14.14
12.38
13.88
13.46
11.46
12.77
Q7
(9.24)
(7.96)
(8.62)
(8.59)
(7.33)
(7.95)
1.95
1.72
1.87
2.95
2.68
2.10
Q8
(3.10)
(2.87)
(3.01)
(3.78)
(3.76)
(2.70)
1.56
1.97
2.07
2.65
2.76
2.47
Q9
(2.43)
(2.73)
(2.54)
(3.47)
(2.55)
(2.49)
17.37
19.02
18.36
17.72
18.90
17.33
Q10
(11.60)
(13.19)
(12.43)
(12.01)
(11.27)
(10.19)
3.77
4.07
4.47
4.51
4.04
4.35
Q11
(5.09)
(4.93)
(4.84)
(4.69)
(4.19)
(4.16)
5.43
4.85
5.91
6.28
6.65
8.78
Q12
(14.36)
(12.07)
(14.18)
(14.93)
(14.60)
(14.59)
0.11
0.07
0.12
0.29
0.46
0.80
Q13
(0.54)
(0.37)
(0.82)
(1.41)
(1.94)
(3.26)
6.32
6.69
6.97
7.00
6.31
6.30
Q14
(5.93)
(6.61)
(6.50)
(5.86)
(5.34)
(5.23)
5.85
6.19
5.82
5.18
5.54
5.40
Q15
(5.69)
(5.72)
(6.19)
(5.31)
(5.42)
(4.98)
1.17
1.01
1.15
1.22
1.14
1.12
Q16
(1.60)
(1.38)
(1.48)
(1.55)
(1.32)
(1.74)
3.62
3.60
3.68
3.82
3.78
4.41
Q17
(3.97)
(3.60)
(3.67)
(3.34)
(3.25)
(3.59)
5.99
6.25
6.23
6.76
5.76
5.67
Q18
(4.17)
(4.78)
(4.46)
(4.70)
(4.35)
(4.34)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.4: Food Consumption for 18 Basic Food Items in Shandong, 1993-1998.
57
Food
Item a, b, c
Year
1993
1994
1995
1996
1997
1998
74.84
73.66
74.58
76.76
64.45
65.38
Q1
(53.66)
(50.83)
(48.46)
(53.35)
(44.36)
(43.25)
8.77
9.26
9.60
8.10
6.81
6.79
Q2
(21.84)
(22.21)
(25.29)
(17.07)
(17.13)
(15.77)
1.659
2.013
2.060
2.255
2.135
1.971
Q3
(4.188)
(7.692)
(6.999)
(7.560)
(4.636)
(4.527)
9.53
9.20
9.31
10.80
9.92
9.89
Q4
(7.62)
(8.23)
(8.31)
(8.38)
(7.51)
(7.51)
117.49
117.51
122.50
120.99
116.53
115.52
Q5
(68.93)
(70.21)
(68.56)
(65.97)
(58.95)
(60.55)
50.92
56.04
58.98
49.27
58.66
63.99
Q6
(35.91)
(33.16)
(34.40)
(31.26)
(38.69)
(40.75)
22.90
20.87
22.33
22.04
19.33
20.51
Q7
(14.38)
(12.55)
(13.32)
(13.03)
(12.28)
(12.28)
1.71
1.39
1.59
2.39
1.97
1.61
Q8
(3.40)
(2.38)
(2.59)
(3.55)
(3.02)
(2.63)
5.55
7.08
6.99
6.55
7.89
7.19
Q9
(5.27)
(5.73)
(5.83)
(5.42)
(6.61)
(6.16)
12.32
13.35
13.29
13.38
14.67
13.20
Q10
(8.79)
(9.84)
(9.73)
(9.49)
(8.89)
(8.37)
7.91
7.04
7.74
8.03
8.11
9.45
Q11
(7.13)
(5.63)
(6.35)
(7.00)
(7.05)
(7.59)
6.18
5.50
6.45
7.69
8.12
9.57
Q12
(16.00)
(12.89)
(13.70)
(15.79)
(14.22)
(15.35)
0.53
0.53
0.61
0.80
1.19
1.74
Q13
(1.82)
(1.76)
(2.53)
(2.82)
(3.62)
(5.45)
9.18
8.97
9.12
9.24
8.16
8.08
Q14
(7.55)
(6.94)
(7.14)
(7.42)
(6.30)
(6.21)
8.96
8.54
8.42
8.61
8.48
9.13
Q15
(6.32)
(6.31)
(6.28)
(7.02)
(6.57)
(6.78)
2.66
2.39
2.43
2.48
2.30
2.42
Q16
(2.60)
(2.33)
(2.28)
(2.36)
(2.38)
(2.20)
2.69
2.73
3.36
3.22
2.93
3.64
Q17
(2.75)
(2.63)
(3.06)
(3.34)
(2.73)
(3.39)
3.36
3.19
3.23
3.26
3.12
3.16
Q18
(3.33)
(2.89)
(2.96)
(2.76)
(2.80)
(2.89)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.5: Food Consumption for 18 Basic Food Items in Jiangsu, 1993-1998.
58
Food
Item a, b, c
Year
1993
1994
1995
1996
1997
1998
59.50
59.09
56.37
55.31
53.74
52.01
Q1
(32.85)
(36.16)
(30.05)
(33.21)
(29.25)
(30.21)
1.79
1.56
1.13
1.11
1.43
1.47
Q2
(7.14)
(7.39)
(4.34)
(4.75)
(7.04)
(6.57)
0.895
1.219
2.223
1.820
2.242
1.823
Q3
(1.725)
(1.599)
(3.727)
(2.228)
(2.336)
(2.227)
4.95
5.38
5.49
6.04
5.97
5.98
Q4
(5.28)
(5.69)
(5.11)
(6.04)
(5.81)
(6.01)
106.27
107.53
109.08
114.44
109.21
112.16
Q5
(45.40)
(46.68)
(43.10)
(45.89)
(40.92)
(41.63)
31.03
34.72
35.89
39.89
42.62
48.22
Q6
(23.38)
(25.06)
(23.79)
(29.11)
(29.60)
(32.30)
19.57
19.63
18.86
19.01
19.22
20.05
Q7
(10.81)
(11.52)
(10.48)
(10.69)
(12.24)
(20.07)
2.86
2.65
2.32
2.47
2.49
2.66
Q8
(3.04)
(3.03)
(2.63)
(3.25)
(3.14)
(3.29)
10.53
11.48
11.45
11.25
11.65
10.80
Q9
(7.83)
(8.14)
(7.12)
(7.16)
(7.38)
(8.21)
8.27
8.77
8.52
8.25
9.37
8.25
Q10
(5.13)
(5.32)
(4.70)
(4.73)
(5.69)
(4.96)
6.92
7.69
7.53
7.35
8.00
9.11
Q11
(5.92)
(6.98)
(6.97)
(6.13)
(6.78)
(10.94)
3.89
3.84
3.14
2.80
3.95
5.54
Q12
(11.37)
(9.14)
(7.52)
(6.57)
(9.69)
(9.71)
0.37
0.35
0.55
0.46
0.68
1.03
Q13
(1.69)
(1.87)
(3.04)
(1.41)
(2.13)
(2.52)
4.13
4.60
4.75
4.62
4.78
4.84
Q14
(4.51)
(4.53)
(4.34)
(4.76)
(4.96)
(4.59)
6.59
6.64
6.94
6.88
6.77
7.04
Q15
(4.89)
(5.07)
(5.24)
(5.44)
(4.65)
(7.78)
2.81
2.78
2.74
2.82
2.88
2.85
Q16
(2.79)
(2.91)
(2.46)
(2.63)
(2.94)
(2.62)
1.75
1.98
1.88
2.35
2.25
2.45
Q17
(1.76)
(2.14)
(1.73)
(2.20)
(2.04)
(2.11)
4.20
4.47
4.90
4.36
4.36
4.84
Q18
(3.76)
(4.05)
(4.32)
(3.87)
(3.84)
(6.62)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.6: Food Consumption for 18 Basic Food Items in Guangdong, 1993-1998.
59
Food
Item a, b, c
Year
1993
1994
1995
1996
1997
1998
1.32
2.28
3.21
3.10
2.39
2.38
P1
(0.20)
(0.46)
(0.47)
(0.51)
(0.37)
(0.31)
1.22
1.72
2.10
2.25
2.28
2.06
P2
(0.22)
(0.24)
(0.23)
(0.30)
(0.41)
(0.49)
1.626
2.103
2.854
3.101
2.853
2.577
P3
(0.775)
(1.081)
(1.078)
(1.369)
(1.068)
(0.960)
1.00
1.38
1.77
1.84
1.75
1.71
P4
(0.49)
(0.60)
(0.63)
(0.73)
(0.79)
(0.59)
0.86
1.13
1.43
1.55
1.46
1.36
P5
(0.25)
(0.35)
(0.37)
(0.42)
(0.38)
(0.35)
1.19
1.51
1.90
1.94
1.82
1.62
P6
(0.43)
(0.55)
(0.65)
(0.79)
(0.74)
(0.68)
6.43
9.99
11.93
12.22
14.62
12.05
P7
(0.88)
(1.45)
(1.67)
(1.72)
(1.77)
(1.45)
8.75
11.92
14.13
13.12
13.68
14.26
P8
(2.15)
(2.90)
(4.07)
(3.88)
(3.86)
(4.04)
7.06
8.99
10.25
11.29
15.82
15.75
P9
(2.09)
(2.66)
(3.27)
(3.41)
(6.24)
(6.64)
4.69
5.15
5.93
6.86
5.31
5.50
P10
(0.49)
(0.65)
(0.52)
(0.72)
(0.51)
(0.49)
8.81
10.51
11.76
12.65
13.92
13.25
P11
(5.46)
(6.55)
(6.34)
(7.82)
(8.34)
(8.20)
1.60
1.97
2.41
2.95
3.14
3.17
P12
(0.80)
(1.49)
(0.78)
(1.48)
(1.11)
(1.40)
3.97
4.67
4.08
5.14
5.40
4.65
P13
(2.14)
(3.58)
(2.55)
(3.05)
(3.10)
(2.28)
1.53
1.80
2.01
2.37
2.70
2.60
P14
(0.58)
(0.67)
(0.75)
(0.87)
(0.94)
(0.79)
6.54
9.64
10.24
9.77
10.13
10.20
P15
(1.85)
(2.85)
(2.99)
(3.12)
(2.85)
(2.42)
3.39
4.66
5.39
5.27
4.72
4.70
P16
(0.76)
(0.95)
(1.77)
(2.09)
(1.43)
(1.59)
3.81
5.10
6.18
7.13
7.05
6.82
P17
(1.47)
(2.46)
(2.29)
(3.79)
(2.70)
(3.10)
5.65
7.33
9.10
10.08
10.83
11.14
P18
(1.62)
(2.28)
(3.05)
(3.32)
(4.52)
(6.18)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Yuan/Kilogram
c
Standard deviations are in parentheses.
Table 4.7: Prices for 18 Basic Food Items in Shandong, 1993-1998.
60
Food
Item a, b, c
Year
1993
1994
1995
1996
1997
1998
1.28
2.11
2.78
2.92
2.15
2.30
P1
(0.15)
(0.25)
(0.22)
(0.23)
(0.17)
(0.19)
1.25
1.86
2.22
2.54
2.44
2.29
P2
(0.25)
(0.35)
(0.55)
(0.65)
(0.69)
(0.63)
2.064
2.657
2.884
3.902
2.768
2.882
P3
(2.023)
(2.413)
(1.676)
(3.826)
(1.907)
(2.015)
1.64
1.94
2.56
2.52
2.35
2.43
P4
(0.74)
(0.71)
(1.03)
(0.85)
(0.85)
(0.83)
1.08
1.43
1.77
2.06
1.94
1.95
P5
(0.30)
(0.39)
(0.51)
(0.51)
(0.53)
(0.53)
1.51
1.91
2.15
2.55
2.33
1.98
P6
(0.80)
(0.72)
(0.90)
(1.05)
(1.06)
(0.83)
7.17
11.06
13.35
13.60
15.39
13.01
P7
(0.99)
(1.32)
(1.47)
(1.41)
(1.58)
(1.42)
9.92
13.08
15.24
14.23
14.33
14.67
P8
(2.50)
(3.61)
(4.34)
(3.97)
(3.86)
(4.42)
8.37
10.52
12.55
13.90
12.82
12.55
P9
(1.62)
(1.71)
(2.71)
(3.15)
(2.91)
(3.04)
5.20
5.87
6.97
8.00
6.14
6.36
P10
(0.39)
(0.62)
(0.73)
(0.84)
(0.80)
(0.71)
8.41
10.71
12.81
14.29
15.54
14.18
P11
(3.28)
(4.12)
(5.58)
(7.24)
(7.46)
(7.22)
2.29
2.99
3.50
4.04
4.48
4.95
P12
(0.82)
(1.45)
(0.99)
(1.09)
(1.46)
(1.60)
4.67
5.54
6.29
5.95
6.74
6.54
P13
(2.53)
(2.78)
(3.65)
(3.21)
(4.95)
(3.34)
1.16
1.57
1.69
1.98
2.06
2.09
P14
(0.40)
(0.57)
(0.59)
(0.77)
(0.66)
(0.67)
5.58
8.67
9.46
9.00
9.08
9.43
P15
(1.40)
(1.28)
(1.69)
(1.95)
(1.79)
(1.67)
3.42
4.63
5.60
5.09
4.97
4.67
P16
(0.55)
(0.57)
(1.17)
(0.97)
(1.14)
(1.16)
5.62
7.65
8.22
10.26
9.66
8.98
P17
(2.92)
(3.73)
(4.13)
(5.79)
(4.59)
(4.71)
8.02
10.42
13.10
15.24
16.63
16.61
P18
(2.91)
(4.35)
(5.34)
(6.87)
(7.99)
(8.15)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Yuan/Kilogram
c
Standard deviations are in parentheses.
Table 4.8: Prices for 18 Basic Food Items in Jiangsu, 1993-1998.
61
Food
Item a, b, c
Year
1993
1994
1995
1996
1997
1998
2.16
3.37
3.93
4.06
3.86
3.54
P1
(0.59)
(0.86)
(0.70)
(0.79)
(1.03)
(0.97)
2.36
3.14
3.34
3.91
3.52
3.72
P2
(0.77)
(1.02)
(1.09)
(1.50)
(1.05)
(1.27)
3.856
4.131
4.972
6.681
6.270
4.783
P3
(2.231)
(2.114)
(2.605)
(5.073)
(3.948)
(2.293)
2.65
3.17
3.72
3.80
3.82
3.68
P4
(1.51)
(1.72)
(1.87)
(1.79)
(2.00)
(1.89)
2.18
2.55
3.02
2.95
3.12
2.88
P5
(0.60)
(0.60)
(0.67)
(0.71)
(0.77)
(0.77)
4.43
5.60
6.65
6.66
6.51
5.71
P6
(1.77)
(2.53)
(2.21)
(2.20)
(2.23)
(2.21)
11.39
14.93
18.33
19.13
21.07
18.93
P7
(1.38)
(1.65)
(2.10)
(2.19)
(2.50)
(2.40)
11.22
15.35
20.29
22.11
20.61
20.05
P8
(2.38)
(3.84)
(5.39)
(5.57)
(5.33)
(5.85)
13.14
15.98
17.69
19.47
17.48
16.67
P9
(1.94)
(2.46)
(2.87)
(3.30)
(2.85)
(2.91)
6.70
7.49
8.31
9.44
7.75
7.63
P10
(0.75)
(0.89)
(1.01)
(1.07)
(1.14)
(1.14)
12.46
15.83
19.09
20.78
19.20
16.64
P11
(6.16)
(9.80)
(12.01)
(11.89)
(12.47)
(10.42)
4.47
6.31
6.77
7.58
8.33
8.61
P12
(3.90)
(4.76)
(2.99)
(3.24)
(3.05)
(3.04)
5.68
6.59
6.41
8.03
8.27
8.37
P13
(2.42)
(2.94)
(2.73)
(3.30)
(3.59)
(3.48)
2.52
2.81
2.97
3.42
3.50
3.48
P14
(1.05)
(1.09)
(1.31)
(1.52)
(1.58)
(1.52)
8.65
11.85
13.03
13.19
13.33
12.81
P15
(1.88)
(2.09)
(2.48)
(2.83)
(2.95)
(2.24)
3.52
4.72
5.55
5.53
5.50
4.76
P16
(0.85)
(0.95)
(1.24)
(1.38)
(1.48)
(1.40)
8.22
10.21
12.16
12.97
14.20
12.65
P17
(4.48)
(6.31)
(6.16)
(6.59)
(7.08)
(6.98)
13.28
16.58
19.30
23.33
24.24
22.41
P18
(7.25)
(9.48)
(10.17)
(14.13)
(13.69)
(11.33)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Yuan/Kilogram
c
Standard deviations are in parentheses.
Table 4.9: Prices for 18 Basic Food Items in Guangdong, 1993-1998.
62
Year
Item
1993
1994
1995
1996
1997
1998
Broad Group a
Z1
0.0
0.0
0.0
0.0
0.0
0.0
Z2
0.0
0.0
0.2
0.2
0.0
0.0
Z3
0.2
0.0
0.6
0.3
0.3
0.0
Z4
15.1
9.7
6.8
6.8
6.8
6.2
Z5
15.4
14.8
14.0
6.6
3.4
1.2
Z6
1.5
2.2
1.2
0.5
0.5
0.6
Z7
0.5
0.9
1.1
1.5
0.3
0.3
Z8
15.1
9.7
6.8
6.8
6.8
6.2
Food Group b
ZF1
0.2
0.0
0.6
0.3
0.3
0.5
ZF2
0.0
0.0
0.0
0.0
0.0
0.0
ZF3
0.2
0.2
0.0
0.0
0.0
0.0
ZF4
2.2
1.2
1.7
0.6
0.9
0.5
ZF5
0.0
0.2
0.2
0.2
0.2
0.0
ZF6
0.0
0.0
0.0
0.0
0.0
0.0
Food Item c
ZB1
15.7
16.2
16.2
16.6
20.8
16.2
ZB2
10.9
5.7
6.3
6.8
11.1
10.6
ZB3
24.8
30.6
39.7
37.8
36.3
35.1
ZB4
0.5
0.9
2.3
1.7
1.1
1.7
ZB5
0.2
0.0
0.0
0.0
0.0
0.0
ZB6
0.0
0.0
0.0
0.0
0.0
0.0
ZB7
1.8
2.0
1.5
3.1
2.0
2.3
ZB8
15.4
21.2
38.5
24.0
28.6
30.9
ZB9
36.2
29.5
33.4
30.5
15.2
18.0
ZB10
1.8
0.9
1.8
2.5
0.9
1.5
ZB11
7.7
8.5
18.0
15.8
16.0
13.7
ZB12
59.4
57.8
53.5
52.3
47.5
34.6
ZB13
83.8
89.8
95.2
89.5
84.8
76.2
ZB14
2.6
1.5
5.1
4.5
3.1
3.8
ZB15
9.7
11.1
17.4
19.2
19.8
20.9
ZB16
20.6
21.5
39.8
38.3
36.6
36.5
ZB17
4.5
7.2
11.7
8.6
8.6
6.0
0.8
1.1
4.0
1.5
4.2
3.1
ZB18
a
The eight broad groups are: (1) Food, (2) Clothing, (3) Household Facilities, Articles and Services, (4)
Medicine and Medical Services, (5) Transport, Postal and Communication Services, (6) Recreation,
Education and Cultural Services, (7) Housing and Utilities, and (8) Miscellaneous Commodities and
Services.
b
The six food subgroups: (1) grains, (2) vegetables and fruits, (3) animal foods, (4) dairy and bean products,
(5) fats, oils and sweets, and (6) others.
c
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 4.10: Proportion (%) of Households with Zero Values in Three-level Utility Tree in
Shandong, 1993-1998 (Sample Size: 650).
63
Year
Item
1993
1994
1995
1996
1997
1998
Broad Group a
Z1
0.0
0.0
0.0
0.0
0.0
0.0
Z2
0.9
0.8
1.5
0.9
1.3
1.4
Z3
0.1
0.6
0.4
0.5
0.1
0.5
Z4
17.9
15.8
17.0
15.9
13.0
12.3
Z5
7.9
8.4
7.4
6.6
3.3
3.1
Z6
2.1
2.4
3.1
2.5
2.1
2.8
Z7
0.3
0.0
0.3
0.3
0.3
0.0
Z8
17.9
15.8
17.0
15.9
13.0
12.3
Food Group b
ZF1
0.4
0.3
0.0
0.0
0.0
0.1
ZF2
0.0
0.0
0.0
0.0
0.0
0.0
ZF3
0.0
0.0
0.0
0.0
0.0
0.0
ZF4
0.6
0.3
1.8
0.3
0.3
0.6
ZF5
0.0
0.0
0.0
0.0
0.1
0.0
ZF6
0.0
0.0
0.0
0.0
0.0
0.0
Food Item c
ZB1
2.4
2.0
2.0
1.5
3.0
3.4
ZB2
34.9
32.1
37.3
36.3
44.6
49.0
ZB3
32.9
34.0
52.3
47.4
40.1
45.1
ZB4
0.8
0.4
2.9
1.3
0.9
1.6
ZB5
0.1
0.1
0.0
0.1
0.0
0.0
ZB6
0.1
0.0
0.1
0.3
0.0
0.1
ZB7
0.5
0.4
0.1
0.5
0.3
0.4
ZB8
36.4
36.4
47.9
42.6
43.0
46.8
ZB9
4.3
1.8
4.6
3.3
2.4
2.6
ZB10
0.4
0.5
0.8
0.6
0.3
0.5
ZB11
0.8
1.5
5.8
5.4
2.6
2.9
ZB12
64.5
62.4
60.0
60.4
53.0
43.6
ZB13
70.6
70.8
85.4
81.9
78.1
72.0
ZB14
1.0
0.4
2.6
0.9
1.3
2.1
ZB15
1.8
2.0
5.4
6.5
6.4
5.0
ZB16
8.3
8.3
17.8
18.4
16.4
13.8
ZB17
3.1
4.1
9.8
11.1
10.1
8.0
3.9
1.8
9.9
9.4
8.9
8.9
ZB18
a
The eight broad groups are: (1) Food, (2) Clothing, (3) Household Facilities, Articles and Services, (4)
Medicine and Medical Services, (5) Transport, Postal and Communication Services, (6) Recreation,
Education and Cultural Services, (7) Housing and Utilities, and (8) Miscellaneous Commodities and
Services.
b
The six food subgroups: (1) grains, (2) vegetables and fruits, (3) animal foods, (4) dairy and bean products,
(5) fats, oils and sweets, and (6) others.
c
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 4.11: Proportion (%) of Households with Zero Values in Three-level Utility Tree in
Jiangsu, 1993-1998 (Sample Size: 800).
64
Year
Item
1993
1994
1995
1996
1997
1998
Broad Group a
Z1
0.0
0.0
0.0
0.0
0.0
0.2
Z2
2.0
2.8
4.0
2.0
1.3
1.0
Z3
0.5
0.5
1.2
0.5
0.7
0.3
Z4
4.2
3.2
2.8
4.5
3.7
1.7
Z5
8.7
6.0
5.3
3.2
2.2
1.3
Z6
2.0
2.2
3.5
2.5
1.0
0.8
Z7
0.0
0.0
0.3
0.0
0.0
0.2
Z8
4.2
3.2
2.8
4.5
3.7
1.7
Food Group b
ZF1
0.3
0.2
0.0
0.0
0.0
0.0
ZF2
0.0
0.0
0.0
0.0
0.0
0.0
ZF3
0.2
0.0
0.0
0.0
0.0
0.0
ZF4
1.0
0.7
0.5
1.2
2.2
0.5
ZF5
0.0
0.2
0.2
0.2
0.2
0.2
ZF6
0.0
0.0
0.0
0.0
0.0
0.0
Food Item c
ZB1
0.8
0.7
0.8
0.7
0.5
1.0
ZB2
56.2
57.5
69.0
72.8
72.5
72.7
ZB3
32.5
24.3
37.8
32.2
25.8
28.7
ZB4
2.5
1.3
6.7
4.2
5.8
6.7
ZB5
0.3
0.2
0.2
0.0
0.0
0.0
ZB6
0.3
0.5
0.5
0.2
0.3
0.3
ZB7
0.5
0.3
0.3
0.0
0.0
0.0
ZB8
6.2
6.0
21.0
22.2
23.5
20.3
ZB9
1.0
0.3
1.7
1.7
1.5
1.3
ZB10
0.3
0.0
0.3
1.2
0.8
0.3
ZB11
1.3
0.5
4.3
3.3
5.2
2.5
ZB12
39.2
36.8
53.5
51.2
48.7
35.7
ZB13
74.0
76.0
85.0
81.0
80.7
72.7
ZB14
1.8
1.0
4.8
5.2
9.2
5.7
ZB15
4.7
5.3
5.7
6.8
7.5
8.2
ZB16
6.8
8.5
12.2
14.2
14.5
13.0
ZB17
6.0
3.8
19.8
11.8
14.2
11.2
4.0
5.8
9.7
11.0
7.3
7.0
ZB18
a
The eight broad groups are: (1) Food, (2) Clothing, (3) Household Facilities, Articles and Services, (4)
Medicine and Medical Services, (5) Transport, Postal and Communication Services, (6) Recreation,
Education and Cultural Services, (7) Housing and Utilities, and (8) Miscellaneous Commodities and
Services.
b
The six food subgroups: (1) grains, (2) vegetables and fruits, (3) animal foods, (4) dairy and bean products,
(5) fats, oils and sweets, and (6) others.
c
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 4.12: Proportion (%) of Households with Zero Values in Three-level Utility Tree in
Guangdong, 1993-1998 (Sample Size: 600).
65
Item a
Shandong
Unit
Disposable Income
Yuan
Household Size
Persons
Number of Wage Earners
Persons
Number of Children under Age 17
Persons
Number of Refrigerators
Units
Housing Area
Square
Meters
Age of Householder
Years
Education Level of Householder
Year
1995
1996
1993
1994
2,625
(953)
3.23
(0.78)
1.95
(0.78)
0.79
(0.58)
0.64
(0.50)
35.33
(15.27)
43.10
(10.91)
3.95
(1.38)
3,583
(1474)
3.22
(0.74)
1.97
(0.76)
0.79
(0.56)
0.69
(0.55)
38.80
(19.03)
42.86
(10.92)
3.87
(1.42)
4,563
(1882)
3.18
(0.68)
1.93
(0.73)
0.79
(0.55)
0.72
(0.50)
40.09
(19.57)
43.43
(10.43)
3.73
(1.39)
3,185
(1436)
3.10
(0.88)
1.84
(0.87)
0.70
(0.58)
0.78
(0.62)
29.33
(12.27)
46.26
(12.20)
3.97
(1.47)
4,388
(2021)
3.19
(0.89)
1.88
(0.87)
0.67
(0.54)
0.83
(0.56)
30.27
(13.78)
47.22
(12.02)
4.05
(1.48)
5,182
(2127)
3.13
(0.88)
1.79
(0.90)
0.66
(0.55)
0.78
(0.44)
30.51
(13.69)
48.19
(11.83)
4.06
(1.40)
1997
1998
5,192
(2333)
3.17
(0.67)
1.95
(0.65)
0.80
(0.53)
0.79
(0.45)
36.61
(14.55)
42.21
(9.74)
3.65
(1.29)
5,164
(2229)
3.21
(0.66)
1.99
(0.56)
0.83
(0.52)
0.83
(0.43)
38.71
(17.21)
41.06
(9.17)
3.65
(1.26)
5,434
(2758)
3.17
(0.64)
1.98
(0.53)
0.79
(0.53)
0.88
(0.38)
39.75
(17.93)
41.13
(8.81)
3.52
(1.28)
5,789
(2481)
3.12
(0.91)
1.77
(0.95)
0.59
(0.54)
0.86
(0.48)
30.52
(13.39)
49.41
(12.12)
4.07
(1.41)
6,325
(2831)
3.18
(0.89)
1.85
(0.88)
0.62
(0.54)
0.88
(0.38)
31.35
(13.61)
48.16
(12.18)
3.97
(1.43)
6,665
(2988)
3.11
(0.86)
1.80
(0.92)
0.59
(0.53)
0.87
(0.39)
31.51
(14.32)
49.11
(11.85)
3.97
(1.39)
Jiangsu
Disposable Income
Yuan
Household Size
Persons
Number of Wage Earners
Persons
Number of Children under Age 17
Persons
Number of Refrigerators
Units
Housing Area
Square
Meters
Age of Householder
Years
Education Level of Householder
a
Standard deviations are in parentheses.
Table 4.13: Descriptive Statistics of Selected Demographic Variables, 1993-1998.
(Continued)
66
Table 4.13: Continued
Item a
Guangdong
Unit
Disposable Income
Yuan
Household Size
Persons
Number of Wage Earners
Persons
Number of Children under Age 17
Persons
Number of Refrigerators
Units
Housing Area
Square
Meters
Age of Householder
Years
Education Level of Householder
a
1993
1994
5,763
(3205)
3.49
(0.92)
2.02
(0.77)
0.84
(0.60)
0.82
(0.43)
41.15
(23.49)
45.69
(10.65)
4.07
(1.49)
8,232
(5133)
3.45
(0.92)
1.99
(0.78)
0.78
(0.58)
0.86
(0.41)
43.22
(24.86)
46.61
(10.79)
3.96
(1.57)
Standard deviations are in parentheses.
67
Year
1995
1996
9,576
(5256)
3.46
(0.89)
2.04
(0.75)
0.75
(0.55)
1.05
(0.52)
43.74
(23.37)
45.93
(10.20)
3.88
(1.53)
10,948
(6112)
3.47
(0.88)
2.08
(0.71)
0.76
(0.55)
0.99
(0.45)
47.09
(26.22)
45.84
(9.72)
3.75
(1.50)
1997
1998
11,499
(6851)
3.48
(0.83)
2.09
(0.72)
0.74
(0.55)
0.95
(0.34)
47.43
(27.02)
45.52
(9.30)
3.62
(1.50)
12,326
(7493)
3.43
(0.79)
2.04
(0.66)
0.74
(0.56)
0.95
(0.34)
47.33
(26.16)
44.38
(9.19)
3.65
(1.41)
Food
Shandong
Jiangsu
Guangdong
Pooled Data
Item a, b, c
mean (s.d.)
mean (s.d.)
mean (s.d.)
mean (s.d.)
Q1
13.71 (14.95)
65.38 (43.25)
51.61 (28.61)
44.96 (39.06)
Q2
24.18 (24.54)
6.79 (15.77)
1.47 (6.57)
10.75 (19.72)
2.44 (4.71)
1.97 (4.53)
1.80 (2.17)
2.07 (4.06)
Q3
12.96 (9.21)
9.89 (7.51)
5.96 (6.00)
9.72 (8.18)
Q4
103.96 (46.47)
115.52 (60.55)
111.78 (40.65)
110.76 (51.20)
Q5
Q6
74.77 (40.26)
63.99 (40.75)
47.76 (30.34)
62.67 (39.28)
12.77 (7.95)
20.51 (12.28)
19.39 (12.04)
17.73 (11.52)
Q7
Q8
2.10 (2.70)
1.61 (2.63)
2.60 (2.96)
2.06 (2.78)
2.47 (2.49)
7.19 (6.16)
10.63 (7.10)
6.70 (6.46)
Q9
17.33 (10.19)
13.20 (8.37)
8.18 (4.67)
13.04 (8.91)
Q10
4.35 (4.16)
9.45 (7.59)
8.77 (7.14)
7.63 (6.92)
Q11
8.78 (14.59)
9.57 (15.35)
5.55 (9.72)
8.14 (13.78)
Q12
Q13
0.80 (3.26)
1.74 (5.45)
1.03 (2.52)
1.23 (4.12)
Q14
6.30 (5.23)
8.08 (6.21)
4.82 (4.58)
6.56 (5.63)
5.40 (4.98)
9.13 (6.78)
6.79 (4.78)
7.26 (5.92)
Q15
Q16
1.12 (1.74)
2.42 (2.20)
2.81 (2.42)
2.12 (2.25)
Q17
4.41 (3.59)
3.64 (3.39)
2.42 (1.92)
3.53 (3.20)
5.67 (4.34)
3.16 (2.89)
4.62 (3.77)
4.38 (3.81)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.14: Comparison of Food Consumption for 18 Food Items in Urban China, 1998.
Household Size
Shandong
Jiangsu
Guangdong
Pooled
--- % --1
2
3
4
5
6 or more
TOTAL
0.05
2.15
23.38
4.93
0.98
0.24
31.72
0.20
7.27
23.13
5.47
2.44
0.54
39.04
Table 4.15: Percentage of Households by Household Size, 1998.
68
0.00
1.37
17.42
7.42
2.54
0.49
29.23
0.24
10.79
63.93
17.81
5.95
1.27
100.00
Food
Item a, b, c
Household Size
1
2
3
4
5
6 or more
39.23
45.53
49.47
50.62
74.20
73.40
Q1
(54.29)
(34.90)
(34.80)
(33.69)
(33.14)
(96.21)
30.20
15.51
10.00
10.95
7.88
15.19
Q2
(40.80)
(28.26)
(18.43)
(18.80)
(13.34)
(20.46)
7.80
3.61
1.79
1.97
1.68
5.50
Q3
(11.88)
(8.17)
(2.83)
(3.20)
(2.94)
(9.01)
22.80
15.90
9.12
8.56
7.53
11.04
Q4
(9.52)
(12.71)
(6.95)
(7.46)
(6.42)
(8.08)
170.18
104.05
104.10
97.06
97.81
131.60
Q5
(74.03)
(43.53)
(42.17)
(31.41)
(32.59)
(32.30)
81.67
63.97
52.98
46.15
40.23
109.20
Q6
(51.62)
(51.51)
(37.89)
(32.80)
(29.89)
(22.10)
40.20
27.14
16.38
16.60
17.84
16.69
Q7
(36.18)
(15.98)
(10.12)
(9.80)
(10.68)
(10.49)
2.20
2.59
2.05
1.86
1.96
1.15
Q8
(2.17)
(4.69)
(2.47)
(2.33)
(2.65)
(1.54)
11.60
10.17
6.11
6.73
6.56
5.92
Q9
(5.73)
(8.77)
(5.82)
(6.66)
(5.87)
(4.68)
28.60
19.02
12.77
11.47
9.76
10.42
Q10
(10.26)
(12.14)
(8.58)
(6.66)
(6.42)
(4.93)
12.60
13.04
6.92
7.10
6.76
8.27
Q11
(11.50)
(9.93)
(5.93)
(6.85)
(6.05)
(6.19)
5.60
11.51
8.73
5.38
5.30
2.69
Q12
(7.70)
(21.16)
(13.24)
(10.47)
(10.15)
(5.71)
0.00
1.98
1.28
0.75
1.10
0.27
Q13
(0.00)
(7.32)
(3.49)
(3.76)
(3.68)
(0.67)
16.60
12.46
5.78
6.36
4.75
5.46
Q14
(8.62)
(8.01)
(4.73)
(5.16)
(3.57)
(4.19)
19.80
12.02
6.57
7.09
6.16
6.73
Q15
(18.42)
(8.67)
(5.21)
(5.00)
(4.20)
(4.31)
2.00
3.42
1.91
2.05
2.23
2.23
Q16
(2.24)
(2.92)
(2.10)
(2.13)
(1.92)
(2.23)
5.62
3.30
3.36
2.58
3.19
7.00
Q17
(6.32)
(4.68)
(2.94)
(2.77)
(2.20)
(2.56)
11.20
5.14
4.53
3.85
3.17
2.46
Q18
(10.94)
(5.01)
(3.70)
(3.35)
(2.61)
(1.70)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.16: Comparison of Food Consumption for 18 Basic Food Items by Household
Size, 1998.
69
Number of Children
Shandong
Jiangsu
0
1
2
3
TOTAL
8.25
21.91
1.46
0.10
31.72
16.89
21.47
0.68
0.00
39.04
Guangdong
Pooled
--- % --9.22
18.55
1.32
0.15
29.23
34.36
61.93
3.47
0.24
100.00
Table 4.17: Percentage of Households by Number of Children in a Household, 1998.
Number of Children
Food
0
1
2
3
Item a, b, c
mean (s.d)
mean (s.d)
mean (s.d)
mean (s.d)
57.13 (45.47)
38.81 (33.59)
35.03 (33.44)
34.20 (24.44)
Q1
Q2
13.46 (24.14)
8.93 (16.31)
16.18 (22.45)
15.40 (26.00)
2.49 (5.39)
1.82 (3.04)
2.52 (4.43)
1.00 (1.00)
Q3
11.67 (9.97)
8.66 (6.83)
9.23 (7.20)
9.60 (7.20)
Q4
133.85 (60.36)
99.15 (41.20)
91.63 (33.03)
77.40 (10.64)
Q5
Q6
69.03 (43.76)
60.22 (36.45)
44.41 (30.27)
47.00 (28.20)
21.65 (14.03)
15.83 (9.28)
13.03 (9.35)
15.60 (16.53)
Q7
Q8
2.24 (3.26)
1.97 (2.48)
1.80 (2.71)
2.80 (1.92)
7.99 (7.66)
6.08 (5.67)
5.13 (4.77)
2.80 (2.59)
Q9
Q10
15.08 (10.45)
11.96 (7.82)
12.48 (7.29)
11.20 (3.77)
9.89 (8.63)
6.55 (5.49)
4.80 (5.12)
5.00 (3.81)
Q11
Q12
8.37 (15.36)
8.25 (12.95)
4.28 (11.39)
3.20 (3.11)
Q13
1.21 (5.25)
1.29 (3.46)
0.42 (1.05)
0.00 (0.00)
8.90 (7.14)
5.33 (4.11)
5.65 (4.89)
4.20 (3.49)
Q14
9.24 (7.30)
6.25 (4.78)
5.59 (3.79)
7.20 (3.77)
Q15
Q16
2.60 (2.55)
1.87 (2.04)
1.96 (1.83)
2.60 (2.88)
4.45 (3.74)
3.05 (2.79)
2.99 (2.30)
2.20 (1.92)
Q17
Q18
4.37 (4.09)
4.44 (3.68)
3.49 (3.04)
4.60 (0.89)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.18: Comparison of Food Consumption for 18 Basic Food Items by Number of
Children under Age 17 in a Household, 1998.
70
Income Level
Shandong
Jiangsu
Guangdong
Pooled
--- % --Low
Middle
High
TOTAL
11.42
18.98
1.32
31.72
7.66
27.28
4.10
39.04
0.93
13.71
14.59
29.23
20.01
59.98
20.01
100.00
Table 4.19: Comparison of Percentage of Households by Income Level, 1998.
Income Level
Food
Low
Middle
High
a, b, c
Item
mean (s.d)
mean (s.d)
mean (s.d)
11.71 (13.80)
14.68 (15.51)
17.00 (14.94)
Q1
Q2
28.11 (25.30)
22.88 (24.14)
8.70 (13.54)
1.96 (3.00)
2.67 (5.44)
3.30 (5.16)
Q3
Q4
12.70 (8.41)
13.23 (9.83)
11.15 (6.13)
93.98 (40.30)
107.48 (47.07)
139.70 (62.38)
Q5
Q6
61.15 (31.91)
80.80 (41.68)
105.89 (46.84)
10.53 (6.57)
14.10 (8.40)
12.96 (8.14)
Q7
Q8
2.08 (2.79)
2.18 (2.71)
1.22 (1.12)
2.35 (2.48)
2.56 (2.52)
2.15 (1.96)
Q9
Q10
15.85 (7.83)
18.22 (11.12)
17.41 (12.90)
3.91 (4.18)
4.61 (4.17)
4.41 (3.59)
Q11
4.53 (9.27)
10.43 (15.05)
21.85 (27.96)
Q12
Q13
0.29 (0.81)
1.08 (4.07)
1.15 (2.97)
5.50 (4.08)
6.66 (5.56)
8.19 (7.87)
Q14
5.55 (4.30)
5.42 (5.41)
3.74 (3.72)
Q15
Q16
1.26 (2.30)
1.08 (1.34)
0.56 (0.89)
Q17
3.57 (3.11)
4.84 (3.71)
5.37 (4.38)
4.87 (4.05)
6.12 (4.43)
6.22 (4.51)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.20: Comparison of Food Consumption for 18 Basic Food Items by Income
Groups in Shandong, 1998.
71
Income Level
Food
Low
Middle
High
Item a, b, c
mean (s.d)
mean (s.d)
mean (s.d)
59.97 (36.08)
67.59 (43.99)
60.81 (49.44)
Q1
Q2
9.13 (17.06)
6.58 (16.16)
3.85 (8.50)
3.54 (7.66)
1.55 (3.32)
1.85 (2.69)
Q3
Q4
9.34 (6.70)
9.99 (7.54)
10.26 (8.68)
100.31 (48.79)
117.56 (59.65)
130.39 (78.81)
Q5
Q6
47.12 (30.35)
65.48 (38.45)
85.64 (57.35)
17.13 (7.69)
21.10 (12.65)
22.90 (15.27)
Q7
Q8
0.89 (1.53)
1.79 (2.89)
1.74 (2.11)
4.80 (3.63)
7.71 (6.29)
8.13 (7.83)
Q9
10.65 (5.34)
13.63 (8.51)
15.13 (10.79)
Q10
6.50 (4.68)
9.94 (7.52)
11.71 (10.52)
Q11
3.41 (7.36)
10.31 (15.13)
16.17 (22.55)
Q12
Q13
0.64 (2.61)
1.81 (5.42)
3.32 (8.39)
Q14
6.99 (5.08)
8.34 (6.55)
8.37 (5.66)
8.54 (4.77)
9.20 (6.70)
9.76 (9.83)
Q15
Q16
2.10 (1.76)
2.48 (2.22)
2.69 (2.70)
3.04 (2.53)
3.68 (3.45)
4.56 (4.17)
Q17
2.48 (1.85)
3.19 (2.92)
4.27 (3.84)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.21: Comparison of Food Consumption for 18 Basic Food Items by Income
Groups in Jiangsu, 1998.
72
Income Level
Food
Low
Middle
High
Item a, b, c
mean (s.d)
mean (s.d)
mean (s.d)
74.21 (31.91)
56.57 (28.81)
45.51 (26.62)
Q1
Q2
0.16 (0.37)
1.25 (7.05)
1.77 (6.31)
0.84 (0.90)
1.58 (2.00)
2.08 (2.34)
Q3
Q4
4.68 (5.98)
6.01 (6.23)
6.00 (5.80)
95.32 (35.08)
105.98 (36.93)
118.29 (43.22)
Q5
Q6
19.79 (11.87)
39.31 (25.33)
57.47 (32.00)
16.42 (7.45)
20.11 (12.36)
18.91 (11.95)
Q7
Q8
2.32 (2.16)
2.79 (3.35)
2.44 (2.59)
11.47 (6.28)
10.42 (7.21)
10.78 (7.05)
Q9
6.42 (4.05)
7.73 (4.32)
8.72 (4.95)
Q10
5.74 (2.79)
8.37 (6.38)
9.34 (7.90)
Q11
0.79 (1.55)
3.71 (7.26)
7.57 (11.45)
Q12
Q13
0.37 (0.96)
0.86 (2.16)
1.23 (2.86)
Q14
5.05 (4.22)
4.24 (4.06)
5.36 (4.98)
6.21 (4.09)
6.72 (4.46)
6.89 (5.12)
Q15
Q16
2.63 (2.29)
3.29 (2.64)
2.37 (2.11)
1.32 (1.20)
2.10 (1.87)
2.78 (1.93)
Q17
2.05 (1.96)
4.06 (3.41)
5.31 (4.01)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.22: Comparison of Food Consumption for 18 Basic Food Items by Income
Groups in Guangdong, 1998.
Age Group
Shandong
Jiangsu
Below 45
Between 45 and 60
Above 60
TOTAL
21.72
8.88
1.12
31.72
15.91
14.69
8.44
39.04
Guangdong
Pooled
--- % --15.76
11.32
2.15
29.23
Table 4.23: Comparison of Percentage of Households by Age Groups, 1998.
73
53.39
34.90
11.71
100.00
Age Group
Food
Below 45
Between 45 and 60
Above 60
Item a, b, c
mean (s.d)
mean (s.d)
mean (s.d)
11.64 (12.89)
17.83 (17.31)
21.22 (22.22)
Q1
Q2
20.53 (21.92)
29.92 (25.52)
49.22 (38.85)
1.91 (2.77)
2.77 (3.95)
10.00 (17.44)
Q3
Q4
11.50 (7.27)
14.58 (10.13)
28.17 (16.92)
94.91 (40.28)
117.10 (45.03)
174.91 (77.04)
Q5
Q6
70.13 (39.46)
82.84 (38.97)
100.65 (47.31)
11.48 (6.68)
14.50 (8.63)
23.91 (12.86)
Q7
Q8
1.96 (2.59)
2.24 (2.55)
3.74 (4.71)
2.54 (2.47)
2.27 (2.37)
2.70 (3.51)
Q9
15.95 (9.18)
18.82 (10.02)
32.30 (15.67)
Q10
3.79 (3.53)
5.19 (4.76)
8.52 (6.55)
Q11
8.27 (12.91)
8.03 (12.93)
24.61 (35.62)
Q12
Q13
0.78 (2.04)
0.79 (4.99)
1.30 (4.99)
Q14
5.02 (3.82)
8.25 (5.74)
15.61 (9.41)
4.56 (4.08)
6.70 (5.52)
11.26 (9.08)
Q15
Q16
1.07 (1.82)
1.10 (1.26)
2.22 (2.88)
3.81 (3.23)
5.34 (3.62)
8.57 (5.30)
Q17
5.57 (3.98)
5.41 (4.42)
9.74 (7.44)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.24: Comparison of Food Consumption for 18 Basic Food Items by Age Group in
Shandong, 1998.
74
Age Group
Food
Below 45
Between 45 and 60
Above 60
Item a, b, c
mean (s.d)
mean (s.d)
mean (s.d)
53.19 (35.97)
67.87 (42.34)
84.03 (49.77)
Q1
Q2
3.09 (9.25)
7.52 (15.62)
12.49 (22.55)
1.38 (3.27)
1.90 (3.71)
3.21 (6.98)
Q3
Q4
7.63 (5.43)
9.59 (6.56)
14.66 (9.93)
88.98 (42.13)
117.14 (53.20)
162.73 (71.82)
Q5
Q6
58.02 (34.35)
65.84 (40.01)
72.03 (50.54)
17.04 (8.80)
20.59 (12.23)
26.92 (15.12)
Q7
Q8
1.40 (2.25)
1.47 (2.09)
2.25 (3.79)
5.91 (4.44)
6.85 (5.07)
10.17 (9.06)
Q9
10.86 (6.04)
13.34 (8.05)
17.37 (10.75)
Q10
7.72 (5.33)
9.20 (7.24)
13.14 (10.16)
Q11
9.15 (13.52)
9.14 (14.75)
11.12 (19.15)
Q12
Q13
1.84 (4.60)
1.82 (6.53)
1.42 (4.81)
Q14
5.88 (4.07)
8.45 (6.15)
11.57 (7.76)
7.05 (5.29)
9.18 (6.37)
12.97 (8.17)
Q15
Q16
1.95 (1.72)
2.36 (2.06)
3.44 (2.84)
2.48 (2.46)
3.91 (3.10)
5.38 (4.42)
Q17
2.90 (2.43)
3.32 (3.24)
3.39 (3.04)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.25: Comparison of Food Consumption for 18 Basic Food Items by Age Group in
Jiangsu, 1998.
75
Age Group
Food
Below 45
Between 45 and 60
Above 60
Item a, b, c
mean (s.d)
mean (s.d)
mean (s.d)
Q1
47.85 (26.95)
54.49 (28.98)
64.00 (33.59)
Q2
1.39 (4.98)
1.66 (8.68)
1.11 (3.21)
1.91 (2.30)
1.78 (2.09)
1.16 (1.36)
Q3
Q4
5.42 (5.01)
6.39 (6.87)
7.70 (7.32)
Q5
107.98 (40.00)
113.43 (38.61)
131.02 (49.85)
Q6
48.21 (30.61)
46.91 (30.97)
48.91 (25.02)
Q7
18.50 (10.55)
20.03 (13.11)
22.57 (15.48)
Q8
2.34 (2.32)
2.77 (2.71)
3.66 (6.33)
Q9
10.38 (6.73)
10.66 (6.81)
12.34 (10.39)
Q10
8.08 (4.54)
8.12 (4.59)
9.27 (5.82)
Q11
7.99 (6.13)
9.28 (7.93)
11.84 (8.56)
Q12
6.53 (10.99)
4.48 (7.78)
3.95 (8.20)
Q13
1.06 (2.38)
1.00 (2.53)
0.93 (3.40)
Q14
4.58 (4.17)
4.96 (4.50)
5.89 (7.09)
Q15
6.31 (4.26)
6.97 (4.90)
9.34 (6.74)
Q16
2.51 (2.19)
3.02 (2.65)
3.93 (2.34)
Q17
2.26 (1.89)
2.65 (2.01)
2.39 (1.50)
Q18
4.80 (4.02)
4.55 (3.60)
3.70 (2.37)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.26: Comparison of Food Consumption for 18 Basic Food Items by Age Group in
Guangdong, 1998.
Gender
Shandong
Jiangsu
Guangdong
Pooled
--- % --Male
Female
TOTAL
17.08
14.64
31.72
31.04
8.00
39.04
17.03
12.20
29.23
65.15
34.85
100.00
Table 4.27: Comparison of Percentage of Households by Gender of Household Head,
1998.
76
Food
Item a, b, c
Shandong
Jiangsu
Guangdong
Male
Female
Male
Female
Male
Female
13.30
64.61
45.68
14.06
68.38
55.86
Q1
(14.31)
(15.69)
(42.60)
(45.73)
(29.87)
(25.66)
26.73
7.25
1.99
21.20
5.02
0.76
Q2
(25.77)
(22.69)
(15.73)
(15.86)
(8.20)
(3.00)
2.67
2.00
1.84
2.17
1.84
1.78
Q3
(3.88)
(5.51)
(4.74)
(3.58)
(2.10)
(2.28)
14.53
10.06
6.49
11.11
9.26
5.23
Q4
(10.03)
(7.77)
(7.51)
(7.48)
(6.75)
(4.67)
108.64
115.62
112.91
98.49
115.15
110.22
Q5
(44.95)
(47.69)
(60.77)
(59.87)
(42.15)
(38.49)
76.57
66.18
50.44
72.67
63.43
45.84
Q6
(41.56)
(38.66)
(39.55)
(45.16)
(29.40)
(31.47)
13.15
20.52
19.93
12.33
20.51
18.64
Q7
(8.23)
(7.61)
(12.26)
(12.35)
(12.18)
(11.83)
2.32
1.63
2.65
1.85
1.52
2.54
Q8
(2.97)
(2.31)
(2.70)
(2.33)
(3.38)
(2.24)
2.54
7.24
11.08
2.38
7.17
10.01
Q9
(2.54)
(2.42)
(6.36)
(5.31)
(7.49)
(6.46)
18.31
13.35
8.26
16.20
12.63
8.13
Q10
(10.30)
(9.96)
(8.22)
(8.93)
(4.78)
(4.51)
4.79
9.89
9.03
3.84
9.33
8.42
Q11
(4.48)
(3.69)
(7.62)
(7.51)
(7.36)
(6.81)
8.93
9.64
6.60
8.59
9.30
4.79
Q12
(14.66)
(14.53)
(15.51)
(14.76)
(9.62)
(9.78)
0.83
2.17
1.26
0.77
1.63
0.86
Q13
(4.06)
(1.97)
(5.29)
(6.00)
(2.29)
(2.80)
7.04
8.08
5.19
5.44
8.07
4.32
Q14
(5.55)
(4.70)
(6.20)
(6.27)
(5.10)
(3.66)
6.00
9.13
7.32
4.70
9.10
6.05
Q15
(5.16)
(4.68)
(6.63)
(7.37)
(5.22)
(3.99)
1.18
2.65
3.01
1.05
2.36
2.53
Q16
(2.13)
(1.12)
(2.13)
(2.45)
(2.57)
(2.15)
4.75
3.69
2.59
4.00
3.46
2.30
Q17
(3.84)
(3.22)
(3.45)
(3.16)
(1.79)
(2.07)
5.49
5.88
3.29
5.04
3.13
4.32
Q18
(4.32)
(4.35)
(2.81)
(3.20)
(3.71)
(3.82)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.28: Comparison of Food Consumption for 18 Basic Food Items by Gender, 1998.
77
Education Level
Shandong
Jiangsu
Guangdong
Pooled
--- % --Low
Middle
High
TOTAL
0.59
22.99
8.15
31.72
3.56
28.01
7.47
39.04
2.54
19.33
7.37
29.23
6.69
70.33
22.99
100.00
Table 4.29: Distribution of Households in Percentage by Education Level of Household
Head, 1998.
Education Level
Food
Low
Middle
High
a, b, c
Item
mean (s.d)
mean (s.d)
mean (s.d)
23.08 (30.39)
13.81 (14.75)
12.75 (13.78)
Q1
Q2
57.42 (38.46)
25.92 (24.70)
16.89 (19.66)
3.83 (6.18)
2.56 (5.10)
2.01 (3.16)
Q3
Q4
16.83 (12.47)
13.48 (9.63)
11.19 (7.35)
131.67 (25.84)
103.84 (47.97)
102.29 (42.74)
Q5
Q6
77.25 (37.16)
74.32 (39.87)
75.86 (41.75)
17.58 (13.59)
13.09 (7.90)
11.50 (7.40)
Q7
Q8
5.58 (6.20)
2.02 (2.62)
2.09 (2.37)
0.75 (0.97)
2.40 (2.44)
2.77 (2.64)
Q9
Q10
24.17 (7.08)
17.40 (10.16)
16.66 (10.34)
5.25 (6.45)
4.32 (4.11)
4.35 (4.12)
Q11
9.17 (10.59)
7.59 (14.16)
12.09 (15.56)
Q12
Q13
0.17 (0.58)
0.70 (3.36)
1.13 (3.06)
10.83 (8.24)
6.42 (5.34)
5.65 (4.48)
Q14
10.25 (8.00)
5.63 (5.02)
4.39 (4.32)
Q15
Q16
1.50 (2.15)
1.16 (1.89)
0.99 (1.19)
Q17
7.33 (5.68)
4.47 (3.66)
4.01 (3.06)
5.75 (8.05)
5.56 (4.40)
5.99 (3.79)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.30: Comparison of Food Consumption for 18 Basic Food Items by Education
Level in Shandong, 1998.
78
Education Level
Food
Low
Middle
High
Item a, b, c
mean (s.d)
mean (s.d)
mean (s.d)
79.19 (54.93)
66.30 (42.30)
55.33 (38.28)
Q1
Q2
17.82 (28.81)
6.18 (14.18)
3.84 (9.35)
5.29 (9.53)
1.66 (3.69)
1.57 (2.77)
Q3
Q4
13.88 (9.75)
9.60 (7.08)
9.08 (7.32)
151.29 (76.88)
113.21 (58.07)
107.14 (55.28)
Q5
Q6
65.15 (36.37)
62.06 (40.33)
70.69 (43.71)
24.51 (16.50)
20.51 (11.98)
18.61 (10.53)
Q7
Q8
1.73 (3.06)
1.66 (2.72)
1.35 (1.94)
8.55 (8.80)
7.22 (6.01)
6.41 (5.01)
Q9
16.63 (10.90)
13.01 (8.26)
12.27 (6.93)
Q10
10.26 (9.29)
9.41 (7.39)
9.22 (7.46)
Q11
4.68 (7.82)
8.55 (14.18)
15.75 (19.95)
Q12
Q13
0.36 (1.16)
1.50 (4.76)
3.30 (8.13)
Q14
9.27 (7.65)
8.12 (6.36)
7.34 (4.62)
12.23 (8.16)
9.13 (6.74)
7.64 (5.67)
Q15
Q16
2.89 (2.80)
2.52 (2.25)
1.83 (1.50)
4.84 (3.93)
3.53 (3.24)
3.48 (3.59)
Q17
3.07 (2.43)
3.06 (2.86)
3.58 (3.19)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.31: Comparison of Food Consumption for 18 Basic Food Items by Education
Level in Jiangsu, 1998.
79
Education Level
Food
Low
Middle
High
Item a, b, c
mean (s.d)
mean (s.d)
mean (s.d)
66.40 (32.12)
51.20 (27.18)
47.60 (29.57)
Q1
Q2
0.33 (0.68)
1.58 (7.55)
1.60 (4.63)
1.71 (1.86)
1.76 (2.19)
1.95 (2.22)
Q3
Q4
5.46 (4.06)
5.88 (5.94)
6.36 (6.69)
128.00 (52.43)
111.59 (38.06)
106.71 (41.49)
Q5
Q6
35.73 (22.47)
47.92 (30.07)
51.48 (32.45)
20.60 (13.36)
19.31 (12.27)
19.21 (10.96)
Q7
Q8
2.65 (2.71)
2.67 (3.18)
2.40 (2.38)
11.87 (6.72)
10.74 (7.33)
9.92 (6.54)
Q9
8.00 (5.38)
8.13 (4.61)
8.40 (4.56)
Q10
12.52 (8.85)
8.79 (7.23)
7.44 (5.66)
Q11
3.23 (8.60)
4.78 (8.33)
8.36 (12.54)
Q12
Q13
0.54 (2.01)
0.91 (2.26)
1.49 (3.17)
Q14
5.00 (4.35)
4.66 (4.59)
5.20 (4.62)
8.13 (5.40)
7.00 (4.78)
5.77 (4.39)
Q15
Q16
3.75 (2.42)
2.89 (2.48)
2.28 (2.10)
2.69 (1.83)
2.39 (1.86)
2.39 (2.09)
Q17
4.35 (3.60)
4.70 (3.82)
4.52 (3.70)
Q18
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Table 4.32: Comparison of Food Consumption for 18 Basic Food Items by Education
Level in Guangdong, 1998.
Refrigerator
Shandong
Jiangsu
Guangdong
Pooled
--- % --No
Yes
TOTAL
4.39
27.33
31.72
5.86
33.19
39.04
2.34
26.89
29.23
Table 4.33: Comparison of Households by Ownership of a Refrigerator, 1998.
80
12.59
87.41
100.00
Ownership of a Refrigerator
Jiangsu
Guangdong
No
Yes
No
Yes
No
Yes
13.49
13.74
74.17
74.02
63.83
49.66
Q1
(18.11)
(14.40)
(46.00)
(42.60)
(31.19)
(27.56)
31.90
11.85
1.48
22.94
5.90
1.35
Q2
(31.17)
(23.09)
(19.65)
(14.82)
(3.01)
(6.80)
2.73
5.01
1.92
2.39
1.44
0.48
Q3
(4.46)
(4.75)
(9.28)
(2.67)
(0.71)
(2.22)
15.16
12.78
6.03
12.60
9.38
5.21
Q4
(11.92)
(8.66)
(7.53)
(7.40)
(4.66)
(6.10)
103.61
104.01
134.66
112.56
112.15
102.88
Q5
(42.47)
(47.12)
(67.15)
(58.72)
(32.29)
(41.23)
73.29
75.01
65.31
48.98
56.52
33.77
Q6
(39.81)
(40.37)
(36.16)
(41.39)
(28.59)
(30.21)
11.71
12.94
21.90
29.73
20.27
18.49
Q7
(8.08)
(7.92)
(12.34)
(12.26)
(17.51)
(11.01)
2.20
1.71
2.70
2.09
1.01
1.46
Q8
(3.17)
(2.62)
(1.94)
(2.72)
(2.01)
(3.01)
2.66
7.30
10.77
2.44
6.52
9.02
Q9
(2.88)
(2.42)
(6.60)
(6.08)
(8.12)
(6.99)
19.22
14.80
8.71
17.03
12.92
8.14
Q10
(10.36)
(10.14)
(8.92)
(8.24)
(5.29)
(4.61)
4.60
9.47
9.27
4.31
9.30
8.73
Q11
(4.48)
(4.11)
(8.03)
(7.52)
(7.53)
(7.11)
9.35
10.45
5.88
4.62
1.71
5.22
Q12
(15.24)
(11.07)
(15.83)
(3.29)
(10.02)
(8.91)
0.92
1.97
1.09
0.78
0.42
0.35
Q13
(2.80)
(3.33)
(2.33)
(5.79)
(1.44)
(2.58)
6.49
10.59
4.92
6.27
7.63
3.75
Q14
(4.92)
(5.29)
(7.89)
(5.76)
(3.53)
(4.65)
6.26
10.48
6.88
5.26
8.89
5.71
Q15
(5.46)
(4.89)
(7.11)
(6.70)
(4.13)
(4.83)
1.20
2.45
3.17
1.11
2.42
2.78
Q16
(1.38)
(1.79)
(2.03)
(2.23)
(2.82)
(2.38)
3.94
4.48
4.73
2.50
3.45
1.52
Q17
(3.14)
(3.65)
(4.26)
(3.18)
(1.57)
(1.93)
5.94
3.23
4.82
5.63
2.79
2.33
Q18
(5.07)
(4.21)
(3.13)
(2.85)
(2.43)
(3.80)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Unit: Kilogram
c
Standard deviations are in parentheses.
Food
Item a, b, c
Shandong
Table 4.34: Comparison of Food Consumption for 18 Basic Food Items by Ownership of
a Refrigerator, 1998.
81
82
Figure 4.1: A Map of China.
82
Living Expenditure for Consumption
Food
Clothing
Household
Facilities,
Articles and
Services
Grains
Starches and
Tubers
Vegetables
Medicine and
Medical
Services
Transport,
Postal and
Communication
Services
Recreation,
Education
and Cultural
Services
Meat and
Poultry
83
Fresh Vegetables
Dried Vegetables
Processed Vegetables
1. Pork
2. Beef
3. Mutton
4. Live
Chicken
(9 items)
1. Cabbage
2. Carrot
3. Tomato
…
(26 items)
Figure 4.2: An Example of Utility Tree for Urban Households in China.
83
Housing and
Utilities
Food away
from Home
Miscellaneous
Commodities
and Services
Food
Processing
Living Expenditure for Consumption
Food
Clothing
Household
Facilities,
Articles and
Services
Medicine and
Medical
Services
Vegetables
and Fruits
Animal
Foods
1. Rice
2. Flour
3. Coarse
grains
4. Potatoes
1. Fresh
Vegetables
2. Fresh
Fruits
1. Pork
2. Beef
3. Poultry
4. Eggs
5. Aquatic
products
84
Grains
Transport,
Postal and
Communication
Services
Milk and Bean
Products
1. Fresh Milk
2. Yogurt
3. Bean and its
products
Recreation,
Education
and Cultural
Services
Fats and Oils
and Sweets
1. Fats and
Oils
2. Sugar
3. Nuts
4. Cakes
Figure 4.3: A Modified Utility Tree for Urban Households in China.
84
Housing and
Utilities
Others
Miscellaneous
Commodities
and Services
CHAPTER 5
EMPIRICAL RESULTS
This study attempts to analyze food demand systems using recent Chinese urban
household data from three provinces (Shandong, Jiangsu, and Guangdong) over a sixyear span, 1993-1998. In Chapter 3, an economic model is constructed under the
assumption of weak separability and an econometric model using the QAIDS is
illustrated. In this chapter, the empirical results are presented in five parts: Section 5.1
presents the Engel curve analysis to investigate the relationship of income and food
consumption. The quadratic almost ideal demand system is estimated in Section 5.2. In
Section 5.3, the QAIDS with demographic variables is presented and Section 5.4 attempts
to estimate the censored QAIDS. Second-stage demand system estimates are shown in
Section 5.5 with a comparison of unconditional elasticities. In the last section, a summary
is provided to conclude the empirical analysis.
5.1 Engel Curve Analysis: Income and Food Consumption
In the literature of empirical demand analysis using micro household data, the
QAIDS has proven to be a more suitable model than the AIDS to explain household
consumption behavior, e.g., Banks et al. (1997). However, whether the QAIDS model
properly applies to food demand analysis in urban China needs to be validated. In order
to investigate whether the QAIDS is an appropriate model to explain Chinese household
consumption, its nested model, the AIDS, is used as an alternative specification. As
85
indicated in Chapter 3, the QAIDS model has an additional quadratic term in the
logarithm of expenditure (log income, in short); therefore, a nested test examining the
quadratic term in log income is executed before price effects are considered. 14 This
assessment of the consumption-expenditure (income) relationship is called an Engel
curve analysis. Among several single-equation specifications, the Working-Leser form is
chosen since it satisfies the adding-up property.
The Working-Leser form equation is augmented by incorporating demographic
variables and can be expressed as:
(5.1)
wi = α i + ∑k =1 δ ki Ak + β i ln X + γ i (ln X ) 2 ,
K
where wi = budget share of food i for i=1,…,18,
Ak = kth demographic variable,
ln X = logarithm of expenditure, and
α i ’s, β i ’s, γ i ’s, and δ ki ’s are parameters to be estimated.
The definitions of variables employed in this analysis and their descriptive
statistics are listed in Table 5.1. Following Deaton and Muellbauer (1980a, pp.19-24),
each equation (5.1) is estimated independently utilizing the ordinary least squares
estimator (OLS), which automatically satisfies the adding-up property. For simplicity, the
dataset for 1998 is used in this analysis with a sample size of 2,049 observations.
Table 5.2 presents the estimation results.15 Since the purpose of this exercise is to
determine whether or not the demand system is quadratic in log income, more attention is
14
The test of quadratic terms will be executed in the following sections in order to investigate the impact of
other effects on the importance of quadratic terms in the QAIDS.
15
In this section, the analysis of parameter estimates of demographic variables is not offered due to its lack
of importance. Analyses of demographic variables will be conducted starting in Section 5.3.
86
paid to the coefficients γi’s in this analysis. Specifically, we are interested in testing the
hypothesis of γi = 0 against γi ≠ 0, for i=1,…,18. Table 5.2 shows that thirteen out of
eighteen food items have an estimated γi not statistically different from zero at the 0.05
level. Only the ri coefficients for rice, fresh fruits, pork, bean products, and fats and oils
are significant at the 0.05 level. This result also shows that at least one of the food items
in each group has a non-linear Engel curve. As a result, the QAIDS is used in the
remaining analysis including incorporation of demographic variables and censored
demand systems. In addition, a hypothesis testing for the quadratic term will also be
implemented. The following four sections will present the empirical results by
considering the original QAIDS model itself (Section 5.2) augmented by incorporation of
demographic variables (Section 5.3) as well as censoring (Section 5.4).
5.2 The Quadratic Almost Ideal Demand System
This section presents the empirical results of the QAIDS model without
considering demographic or censoring effects. As discussed earlier, eighteen food items
are included in the demand analysis; however, managing 18 equations in a demand
system is burdensome even with today’s high-speed computers. Therefore, utilizing the
economic theory to separate these 18 food items into several subgroups will avoid
estimation problems. Following food dietary guidelines and the food guide pyramid in
China (as mentioned in Chapter 4), eighteen food items are assumed weakly separable
from other food or non-food items and are divided into five groups– group 1: rice, flour,
coarse grains, and potatoes; group 2: fresh vegetables and fresh fruits; group 3: pork, beef
and mutton, poultry, eggs, and aquatic products; group 4: fresh milk, yogurt, and bean
87
products; and group 5: fats and oils, sugar, nuts, and cakes.16 These eighteen food items
account for about half of the total food expenditure as discussed earlier. Some food items,
such as beverages, wine, processed food, and food away from home are not included.
Pooled data from three provinces in 1998 are employed in this study. In total,
there are 2,050 observations available to use in the analysis; however, in order to make
the log income term meaningful, households with no expenditure on grains (group 1) are
not considered since log income (ln X) is not defined. Excluding those households with
zero expenditure in each group, the sample sizes are slightly different in the five food
subgroups, which consist of 2,045, 2,049, 2,049, 2,038, and 2,048 observations,
respectively. Group 4 has fewer observations due to zero consumption of dairy and bean
products as discussed in the previous chapter. The QAIDS is used as a functional
specification to examine the significance of the quadratic terms of log income. In addition,
homogeneity and symmetry conditions are imposed to ensure regularity conditions hold.
Without incorporation of demographic variables, budget shares, prices, and total
expenditures by groups are used for econometric analysis and are listed in Table 5.3. The
parameter estimates for the five subgroup models are not presented here. Table 5.4 shows
the results of testing λi = 0 vs. λi ≠ 0 by using the Wald test. This is actually a nested
test of the AIDS versus the QAIDS. From the Wald test, the null hypothesis, λi = 0 , is
rejected for eleven out of eighteen food items. This indicates that the QAIDS model fits
the urban Chinese household data better than the AIDS model. Only pork, poultry, eggs,
yogurt, fats and oils, sugar, and cakes do not reject λi = 0 at the 0.05 significant levels.
This finding is not consistent with that from the Engel curve analysis (Section 5.1)
16
From model diagnosis, another approach to dividing these food items is by separability tests.
88
especially for pork and fats and oils when price effects are additionally considered in this
section. However, at this point, our findings reclaim the previous findings from Banks et
al. (1997) that the QAIDS is a more suitable model than the AIDS to explain household
consumption behavior. It should be noted here that the demographic effects are not yet
considered, and the conclusion may change in the later analysis.
The conditional elasticities, including expenditure elasticities and Marshallian and
Hicksian price elasticities within groups are calculated (equations 3.33a-c) and presented
in Table 5.5. All the expenditure and own-price elasticities have a correct sign. The
conditional expenditure elasticities range from 0.483 for potatoes to 1.585 for fresh milk.
Rice, aquatic products, yogurt, and fresh milk have relatively high conditional
expenditure elasticities, which indicate that people will increase (decrease) consumption
on these food items relative to food items within each group as expenditure increases
(decreases). Marshallian own-price elasticities vary dramatically ranging from -2.757 for
flour to -0.257 for coarse grains. Most of the Marshallian own-price elasticities are close
to or lower than unity (-1), except for flour (-2.757) and rice (-2.066), indicating rice and
flour are price elastic. Most of the Hicksian cross-price elasticities are positive, especially
in groups 2, 4, and 5, which indicate that food items in those groups are net substitutes, as
expected. The discussion and comparison of these elasticities will be provided in detail
later in this chapter.
5.3 Incorporation of Demographic Variables
Following Lewbel’s unified approach to incorporating demographic variables
(Chapter 3), ordinary budget share scaling and translation is utilized to fit the QAIDS
(equation 3.38) using the dataset from urban China. Sample sizes in the five groups are
89
adjusted as in the previous section, and the five food groups are estimated independently
using the full information maximum likelihood estimator (FIML) in SAS. The definitions
and descriptive statistics of variables are presented in Table 5.6. Parameter estimates are
summarized in Tables 5.7 and 5.8.
There are four food items in group 1, including rice, flour, coarse grains, and
potatoes. Rice accounts for 62% of the budget share. Potatoes and flour make up 17%
and 16%, respectively, whereas coarse grains account for only 5%. The mean prices are
similar among the four grain food items with potatoes being the cheapest (2.567
Yuan/Kg). The average expenditure in group 1 is 169 Yuan per capita per year.
Demographic variables were previously discussed in Chapter 4, and thus, they will not be
repeated here. Food group 2 includes only fresh vegetables and fresh fruits,
corresponding to 60% and 40% of the budget share, respectively. Total expenditure in
this group is 386 Yuan, more than twice the expenditure found in the grain group. Food
group 3 includes five animal food items. The budget share for pork is the largest,
accounting for more than 40%. Beef and mutton have the smallest share at only 6%. The
other three animal food items (poultry, eggs, and aquatic products) have similar budget
shares of slightly over 15%. Prices of these animal food items are alike (approximately
15 Yuan/Kg) except for eggs (6.5 Yuan/Kg). In addition, the average expenditure on
these five animal food items totaled 575 Yuan per capita in 1998, the largest among the
five groups. Group 4 contains dairy and bean products. Bean products account for 52% of
the budget share in this group, whereas fresh milk and yogurt contribute 39% and 9%,
respectively. The average prices of dairy products are around five to six Yuan/kg whereas
the price of bean products is 2.66 Yuan/kg. The average total expenditure for this food
90
group was about 63 Yuan per capita in 1998, which is much lower than the first three
food groups. A possible reason for this difference is that most of the households in this
study did not consume dairy products as indicated by the large proportion of zero
consumption. Fats and oils, the most important food in the last group, account for 40% of
the budget share followed by cakes (36%), nuts (17%), and sugar (6%). The prices of
these four food items vary between 4.7 Yuan/kg to 16.5 Yuan/kg. The mean expenditure
of this group is 182 Yuan per capita, which is almost the same as group 1.
Table 5.7 shows parameter estimates for expenditure and prices in the QAIDS
with demographic variables. In order to reduce the computational burden and to satisfy
demand properties, homogeneity and symmetry properties are imposed in estimation. By
doing so, the elasticities can be compared with those obtained in the previous section.
Less than half of the parameters are statistically different from zero at the 0.05 significant
level. However, the fitted models for groups 1 and 4 look promising with more than half
of the parameters statistically different from zero.
Parameter estimates corresponding to demographic variables of equation 3.37 are
summarized in Table 5.8. In general, these selected demographic variables help to
explain consumption patterns in urban China since all the adjusted R2 are improved
dramatically. For example, the adjusted R2 of rice and poultry improve from 0.32 to 0.63
and 0.04 to 0.25 respectively. When interpreting these parameter estimates, it is necessary
to be cautious since their impact on budget shares (the dependent variables) is
compounded with other factors, e.g., prices within the group. This is revealed from
equation 3.39d. However, the parameter estimates still reflect a direct impact of
demographic variables on the budget share (from the ∂ri ∂Ak term). Overall,
91
demographic variables such as household size, income groups, age of the householder,
and region are the most important factors affecting food consumption patterns.
Specifically, household size has a positive impact on food items such as rice, flour, and
fats and oils but has a negative impact on potatoes and bean products. Taking group 1 as
an example, the parameter estimates indicate that budget shares of rice and flour will
increase 4.8% and 1.7%, respectively, but decrease 1.3% in potatoes when household size
increases one unit. Since rice and flour are major staple food items in China, this finding
is reasonable. The ratio of number of children to household size has one significantly
negative effect on flour. In addition, the ratio of number of children to household size has
a positive but insignificant impact on fresh milk and yogurt, indicating children are not
the determining factors affecting milk budget share, which is not as expected. The
comparisons of income groups provide mixed findings. Both high and middle income
groups have significantly negative estimates in fats and oils which means the higher the
income level, the lower the expenditure share on fats and oils. As for nuts, the higher the
income level, the more the expenditure share. In addition, the high income group has a
significantly positive effect on rice, coarse grains, and milk but a negative effect on sugar.
These dummy variables of income groups show that low income households have higher
budget shares of animal food items, which is not as expected; however, these estimates
are not statistically significant. Age has a negative impact on fresh milk and yogurt; but a
positive effect on bean products. This finding implies that elderly householders will
spend 0.7% more budget on bean products by reducing approximately 0.8% and 0.3%
budget on fresh milk and yogurt. This is consistent with consumption patterns found in
urban China indicating that milk products are still not a significant part of Chinese eating
92
habits. Moreover, age has a positive effect on potatoes, eggs, and nuts. Gender as well as
dummy variables of education level and owning a refrigerator have a very limited impact
on the allocation of budget shares. It was anticipated that the ownership of a refrigerator
would have an impact on dairy products; however, the econometric model for group 4
does not show any significant influence. The most significant contribution of the
demographic variables is from the regional dummy variables. Seven out of eighteen
coefficients are statistically significant from zero, which indicates that the differences
among the three provinces are notable. Relative to Guangdong, people in Shandong
spend significantly more on flour, eggs, bean products, nuts, and cakes but less on rice
and sugar; people in Jiangsu spend significantly less expenditure shares on rice, coarse
grains, beef and mutton, poultry, and fresh milk but more on fats and oils and nuts. For
example, budget shares of rice and flour are different among the three provinces, as
expected. The budget share of rice in Guangdong is about 96.9% higher than that in
Shandong and 55.7% higher than that in Jiangsu whereas the budget share of flour in
Shandong is 29.5% higher than that in Guangdong; hence, regional differences are an
important factor affecting consumption choices.
With incorporation of demographic variables, the results of hypotheses testing of
λi = 0 vs. λi ≠ 0 can be obtained from Table 5.9. Compared with Table 5.4, Table 5.9
shows that, with incorporation of demographic variables, the explanatory power of the
QAIDS has been diluted since only seven out of eighteen λ’s present statistically different
from zero, including four food items in group 1 and fresh milk, bean products, and nuts.
We also find that the QAIDS can be reduced to the AIDS in four food items (fresh
vegetables, fresh fruits, beef and mutton, and aquatic products) since the λ’s are not
93
statistically important when incorporating demographic variables. This may be because
the effect of demographic variables on budget share dominates the effect of the quadratic
term in log income.
The conditional elasticities within each group, including expenditure as well as
Marshallian and Hicksian price elasticities, are calculated (equations 3.39a-c) and
presented in Table 5.10. These elasticities are calculated ignoring demographic effects in
order to make a comparison of elasticities without demographic variables (Table 5.5).
The expenditure elasticities within each group have an expected positive sign and range
from 0.484 (for sugar) to 1.386 (for fresh milk). Many of them are close to unity, such as
fresh vegetables, fresh fruits, pork, beef and mutton, poultry, and cakes. Rice and flour
have relatively high expenditure elasticities (1.12 and 1.18, respectively). In addition, fats
and oils also have a high expenditure elasticity (1.22). These high expenditure elasticities
indicate that consumption of these food items will increase more than other food items
within each food subgroup in urban China as expenditures within each subgroup increase.
All the Marshallian own-price elasticities have a correct sign and range from -1.53 (for
flour) to -0.35 (for sugar). Again, many of them are close to unity (-1.0), especially in
groups 2 and 3. From the Hicksian price elasticities, food items within each group present
substitutes except flour, coarse grains, and potatoes in group 1, beef and poultry in group
3, and sugar and nuts in group 5 indicating complements. A comparison of elasticities in
Tables 5.5 and 5.10 shows that, with incorporation of demographic variables, most of the
expenditure and own-price elasticities in Table 5.10 move towards unity (either 1.0 or 1.0 depending on the situation) compared with those in Table 5.5 whereas cross-price
elasticities move towards zero.
94
5.4 Censored Demand System
It is necessary to consider the zero-consumption problem since micro household
data are employed in this study. As indicated in Chapter 4, the most serious zero
consumption problems occurred in six of the 18 selected food items– flour (36.3%),
coarse grains (24.0%), beef and mutton (20.7%), fresh milk (31.3%), yogurt (62.7%), and
sugar (9.9%), with the percentages of households with zero consumption in parentheses.
Hence, it is important to improve estimation by considering a censored demand system.
As discussed in Chapter 3, the two-step estimation procedure proposed by Shonkwiler
and Yen (1999) is adopted in this study.
Shonkwiler and Yen’s procedure consists of two steps. In the first step, the Probit
models in single equation are estimated using LIMDEP 7.0 econometric software in order
to compute the probability and cumulative density values for the six commodities with
serious zero consumption problems. The results are summarized in Table 5.11. Most of
the variables are defined as earlier. However the total food expenditure (TFE) and its
square (TFE2) as well as the square of age (AGE2) are included in the Probit analysis.
Similar to the results presented in the previous section, the dummy variables for regions
are still the most important factor, especially among flour, fresh milk, and yogurt.
Meanwhile, total food expenditure is another significant factor for coarse grains, beef and
mutton, fresh milk, yogurt, and sugar. The dummy variable for ownership of a
refrigerator indicates that households having a refrigerator are likely to consume more
beef and mutton, fresh milk, and yogurt but less flour. In addition, with more children in
a household, the probability of consuming more beef and mutton, fresh milk, and yogurt
will be higher. The higher the income groups, the higher probability of consuming more
95
fresh milk and yogurt but less sugar. Dummy variables for education have an impact on
coarse grains, beef and mutton, and yogurt, especially the difference between high and
low education level households. Age has an impact on flour, beef and mutton, and yogurt.
However, gender is not an influential factor at all in this Probit analysis.
In the first step, the probability and cumulative density values are obtained from
this Probit model and in the second step, the equation 3.43 is estimated with the QAIDS
as the chosen functional form. Again, with incorporation of demographic variables, the
censored QAIDS models are estimated for food groups 1, 3, 4, and 5, since group 2 has
not encountered zero consumption problems in our analysis. Using the seemingly
unrelated regressor (SUR), the parameters are estimated independently in four food
groups and are summarized in two parts. The first part contains the estimates of price and
expenditure parameters in the QAIDS model (Table 5.12), and the second part is for the
demographic variables (Table 5.13).
From Table 5.12, the additional information from the probability values in the
first step improves the fit of the QAIDS model to the Chinese urban household data. All
the parameters δ’s in the four food groups are statistically different from zero at the 0.001
level, indicating an important improvement to the fitted models. From Table 5.13, the
parameter estimates corresponding to demographic variables are of significant interest to
this study. Generally, demographic variables lose some explanatory power compared to
those with no censoring (Table 5.8), especially household size. Household size becomes
less important with only one parameter (fats and oils) statistically different from zero
with a positive sign. This means that as household size increases by one unit, the budget
share of fats and oils will increase by approximately 5.4%. The ratio of number of
96
children to household size becomes statistically unimportant when considering censoring.
Dummy variables for income groups are still important, and the parameter estimates
indicate that the richer the households, the more their budget shares on nuts, but smaller
shares on flour and fats and oils. The parameters of income groups for animal food items
are negative, which are the same as in Table 5.8. This indicates that richer households
spent less on animal food items, not as expected based on the descriptive statistics in
Chapter 4; however, they are not statistically significant. Age is still an important factor
with six parameter estimates significantly different from zero, including coarse grains,
pork, beef and mutton, eggs, nuts, and cakes. Gender and educational levels are not
important in this model. The parameter estimates of the ownership of a refrigerator on
coarse grains (-0.024) and beef and mutton (0.014) are significant, but those on fresh
milk and yogurt are -0.252 and -0.438, respectively, reflecting a negative impact of
having a refrigerator on fresh milk and yogurt consumption, which is not as expected;
however, they are insignificant. Among the demographic variables, the most notable
impacts are regional differences. This conclusion is similar to the previous analysis. The
food consumption patterns between Guangdong and Shandong are statistically different
for rice, flour, poultry, sugar, nuts, and cakes. In addition, the parameter estimates of the
dummy variable for Jiangsu, which indicates the difference between Jiangsu and
Guangdong, are statistically different from zero on beef and mutton, poultry, fats and oils,
and nuts. As a result, region, age, and income are the most important demographic
characterstics that explain consumption behavior in urban China.
Table 5.14 presents the Wald tests of λi = 0 vs. λi ≠ 0 in the censored QAIDS
with incorporation of demographic variables. Again, the purpose is to investigate whether
97
the quadratic term of log income is important or not. Interestingly, parameter estimates of
rice, potatoes, and nuts are statistically different from zero at the 0.10 level using the
Wald test. This finding shows that most of the budget shares are linear in log income.
The elasticity estimates are presented in Table 5.15. These conditional elasticities
with censoring are slightly different from those without censoring. Since group 2 has no
censoring problem, it is not necessary to calculate elasticities for this group, which is
identical to those elasticities without censoring (Table 5.10). It should be noted again that
these elasticities are calculated without considering demographic effects, which are set to
zero. Expenditure elasticities range from 0.73 (nuts) to 1.17 (fats and oils), which is
narrower compared to those from Table 5.10. Most of the expenditure elasticities are
close to unity. Own-price elasticities, either Marshallian or Hicksian, have a correct sign.
In addition, most of the Marshallian own-price elasticities are close to unity, ranging
from -0.736 (cakes) to -1.075 (rice). Hicksian own-price elasticities are below unity,
ranging from -0.945 (beef and mutton) to -0.394 (cakes). Cross-price elasticities, as
expected, are smaller than own-price elasticities (in absolute values) and many of the
cross-price elasticities are close to zero, meaning very small effects with respect to
changes of other prices. In addition, the Hicksian cross-price elasticities indicate that food
items within groups are net substitutes. This is slightly different compared with Table
5.10. The comparison of the elasticities between Tables 5.10 and 5.15 shows that the
own-price and expenditure elasticities in the censored QAIDS move closer to unity.
5.5 Second Stage Demand System Estimation
In order to investigate the relationship of food items in different groups and to
calculate unconditional elasticities for the 18 food items, it is necessary to estimate a
98
second stage demand system with the aggregate foods. The five food groups are shown in
Figure 4.3 and are estimated as a demand system in the second stage. The definitions and
descriptive statistics of the variables for the broad food groups are summarized in Table
5.16. The budget shares for the five food groups are 0.126, 0.277, 0.412, 0.046, and 0.139,
respectively. The average expenditure of these five groups was 1,374 Yuan per capita in
1998. The corresponding aggregate prices are 2.64, 2.31, 12.46, 3.69, and 11.15 Yuan/kg,
respectively; their calculation will be discussed shortly. The demographic variables are
the same as the previous analyses. In addition, zero consumption problems are not serious
due to aggregation, and thus, a censored demand system approach is not applied to the
second stage demand system estimation.
The aggregate price for each group is calculated using the geometrically weighted
average of prices within groups (Moschini, 1995):
(5.2)
log PhI = ∑i =1 wi0 log pih ,
nI
where PhI = aggregate price for group I for household h,
wi0 = fixed mean budget share for food i within group I, and
pih = price of food i for household h.
Applying the same procedure used for within-group demand systems, testing for
quadratic term is conducted first. The Wald test results are presented in Table 5.17.
Statistically, the quadratic term is significantly different from zero for groups 1 and 2 but
not for groups 3 to 5. Therefore, there exists a non-linear relationship between log income
and budget shares of grains, vegetables, and fruits.
99
The expenditure and price elasticities for the broad food groups are calculated by
imposing homogeneity and symmetry properties and are presented in Table 5.18. The
expenditure elasticitis are close to unity, similar to those from within-group elasticities.
Expenditure elasticity for animal protein food is the highest (1.113), which means that
Chinese urban households will expend more budget shares on animal protein food as total
expenditure on food increases. All the Marshallian and Hicksian own-price elasticities
have a right sign, and most of them are smaller than unity (in absolute values) except for
food group 4, the dairy and bean products, which is -1.35. This indicates that dairy and
bean products are very price sensitive. In addition, the Hicksian cross-price elasticities
show that most of the food groups are net substitutes, except for group 2, with groups 1
and 5 having negative cross-price elasticities.
Adopting the same approach to incorporating demographic variables as discussed
earlier, the parameter estimates for the broad food groups are shown in Table 5.19. The
fitted model is acceptable with more than 20 parameter estimates significantly different
from zero. The fitted model reveals that most of the demographic variables, which were
important in the previous sections, have less influence on the aggregate food groups when
explaining the consumption patterns with respect to broad food groups in urban China.
Notably, household size and age and gender of the householder are statistically
insignificant. Dummy variables for income groups and for regional differences are still
important when explaining consumption patterns with many parameter estimates being
significant at the 0.05 level. An interesting finding is that seven out of eleven parameters
are significant for group 4 of dairy and bean products, including the ratio of number of
children to household size and dummy variables of income groups, education levels,
100
region, and ownership of a refrigerator. The parameter estimates indicate that households
with higher income, higher education level, more children, and having a refrigerator
consumed more dairy and bean products. In addition, households in Shandong and
Jiangsu, compared with Guangdong, also have larger budget shares allocated in group 4.
Another interesting finding is that households having a refrigerator, being richer and
having a higher education level tend to consume fewer grains with negative parameter
estimates which are statistically significant.
Table 5.20 presents the Wald tests of the five food groups in the second stage
estimation. The results again show that the quadratic terms are important in food groups 1
and 2 but not in the remaining three groups. The corresponding expenditure and price
elasticities are calculated and presented in Table 5.21. When considering demographic
effects, the expenditure elasticities are closer to unity compared with those in Table 5.18.
Again, groups 2, 4, and 5 remain necessities, but the expenditure elasticity for group 3
drops slightly from 1.113 to 1.099. The own-price elasticities, either Marshallian or
Hicksian, have a similar movement towards unity; however, Hicksian cross-price
elasticities in Table 5.21 show all food groups being substitutes, which are slightly
different from those in Table 5.18.
The unconditional Hicksian and Marshallian price elasticities as well as
unconditional expenditure elasticities are calculated by equations (3.16), (3.27), and
(3.28), respectively. These unconditional elasticities are presented in Tables 5.22 and
5.23. Table 5.22 shows the unconditional elasticities derived from the QAIDS with no
demographic effects whereas Table 5.23 exhibits those elasticities from the censored
QAIDS with demographic variables. Generally, the unconditional elasticities from Table
101
5.23 are closer to unity than those from Table 5.22. In addition, for expenditure and ownprice elasticities, all elasticities have a right sign. From Table 5.23, aquatic products have
the highest unconditional expenditure elasticity (1.240) and nuts have the lowest (0.640),
whereas from Table 5.22, the highest unconditional expenditure elasticity is fresh milk
(1.559) and the lowest is bean products (0.507). Rice is also income elastic with 1.138
(Table 5.23). As for own-price elasticities, Table 5.22 shows a wide range from -2.673
(flour) to -0.238 (coarse grains); however, in Table 5.23, the range is narrower from 1.066 (fresh milk) to -0.619 (cakes) and only two food items have an unconditional ownprice elasticity over unity in absolute value (-1.004 for beef and mutton). Hence, the
unconditional elasticities are not fluctuated when censoring and incorporation of
demographic variables are considered. The unconditional Hicksian price elasticities
present a similar trend shown in Table 5.23. All the Hicksian own-price elasticities have a
negative sign. In addition, from the unconditional Hicksian cross-price elasticities, most
of the food items are net substitutes and many of them are close to zero, indicating weak
cross-price effects. In this study, the expenditure elasticity for the food broad group is
estimated using the Working-Leser form (equation 5.1). The expenditure elasticity for the
food broad group is 0.657 and thus the unconditional expenditure elasticity with respect
to total living expenditure can be calculated by augmenting equation (3.28) to the third
level, and the unconditional expenditure elasticities with respect to total living
expenditure are shown in Tables 5.22 and 5.23. Since all unconditional elasticities are
multiplied by 0.657, most of the unconditional elasticities become smaller than unity
except fresh milk (1.025) in Table 5.22. This finding indicates that most of the 18 food
102
items are normal goods, as expected. However, as for rice and flour, their income
elasticities are still higher than what was expected.
5.6 Summary
An econometric analysis is conducted in a sequence of six steps and shown in
Sections 5.1-5.5. A Working-Leser form of a single equation approach is used to
determine if a demand system requires the quadratic term in log income. The results
show that the QAIDS model works properly by fitting the dataset from urban China. Next,
a demand analysis is conducted considering 18 food items which are consumed at home.
The QAIDS is utilized to test the significance of the necessity of a quadratic term in log
income. According to the Wald test results, the QAIDS is accepted as a preferable model.
In order to compare elasticities, homogeneity and symmetry conditions are
imposed. Eleven demographic variables are created from eight different demographic
characters, such as household size, age, gender, and education level of the householder,
and provinces. The empirical results show that these demographic variables have a
significant impact on all six QAIDS models (five food subgroups and one aggregate food
group) under the framework of the two-stage demand system. Specifically, among the
considered demographic variables, the regional difference is the most in determining
consumption patterns among the three provinces. Other important factors are household
size, income levels and age of household head. The expenditure and price elasticities
show that a censored demand approach and a two-stage demand system improve these
estimates, which is important to policy makers for forecasting future food consumption
trends in urban China.
103
Variable
Description a
Mean
S.D. b
Dependent
w1
Budget share for food 1
0.082
(0.061)
Budget share for food 2
0.022
(0.039)
w2
Budget share for food 3
0.005
(0.009)
w3
Budget share for food 4
0.017
(0.012)
w4
Budget share for food 5
0.161
(0.049)
w5
w6
Budget share for food 6
0.116
(0.066)
Budget share for food 7
0.185
(0.078)
w7
w8
Budget share for food 8
0.025
(0.031)
w9
Budget share for food 9
0.065
(0.048)
Budget share for food 10
0.068
(0.045)
w10
w11
Budget share for food 11
0.070
(0.051)
Budget share for food 12
0.028
(0.044)
w12
w13
Budget share for food 13
0.006
(0.016)
Budget share for food 14
0.012
(0.010)
w14
Budget share for food 15
0.056
(0.038)
w15
w16
Budget share for food 16
0.008
(0.008)
w17
Budget share for food 17
0.023
(0.018)
Budget share for food 18
0.053
(0.044)
w18
Explanatory
X
Total living expenditure (in Yuan)
1,374.16 (668.98)
HS
Household size (in persons)
3.225
(0.785)
NC
Ratio of number of children to household size
0.212
(0.162)
INH
1 if high income household; 0 otherwise
0.200
(0.400)
INM
1 if middle income household; 0 otherwise
0.600
(0.490)
AGE
Age of household head (in years)
45.200 (10.744)
MALE
1 if male household head; 0 otherwise
0.652
(0.477)
EDH
1 if high education level; 0 otherwise
0.230
(0.421)
EDM
1 if middle education level; 0 otherwise
0.703
(0.457)
PR1
1 if Shandong; 0 otherwise
0.317
(0.466)
PR2
1 if Jiangsu; 0 otherwise
0.390
(0.488)
FR
1 if the ownership of refrigerator; 0 otherwise
0.874
(0.332)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Standard deviations are in parentheses.
Note: sample size= 2049.
Table 5.1: Definitions and Descriptive Statistics of Variables Used in Engel Curve
Analysis, 1998.
104
Parameter
α
β
γ
δ1 (HS)
δ2 (NC)
δ3 (INH)
δ4 (INM)
δ5 (AGE)
δ6 (MALE)
δ7 (EDH)
δ8 (EDM)
δ9 (PR1)
δ10 (PR2)
δ11 (FR)
Adj. R2
Rice
Coefficient
S.E. a
-0.735*** (0.164)
0.238*** (0.047)
-0.017*** (0.003)
0.002 (0.002)
-0.005 (0.008)
-0.032*** (0.004)
-0.014*** (0.003)
0.0005*** (0.0001)
0.001 (0.002)
-0.013* (0.005)
-0.008 (0.005)
-0.073*** (0.004)
0.002 (0.003)
-0.008* (0.003)
0.3551
Potatoes
Parameter
Coefficient
S.E.
α
0.048 (0.035)
β
-0.007 (0.010)
γ
0.0002 (0.0007)
δ1 (HS)
-0.0004 (0.0003)
δ2 (NC)
0.002 (0.002)
-0.003** (0.0009)
δ3 (INH)
δ4 (INM)
-0.003*** (0.0007)
0.0001*** (0.00003)
δ5 (AGE)
δ6 (MALE)
0.002** (0.0005)
-0.0008 (0.001)
δ7 (EDH)
δ8 (EDM)
0.0001 (0.001)
0.010*** (0.0008)
δ9 (PR1)
0.004*** (0.0007)
δ10 (PR2)
-0.002*** (0.0007)
δ11 (FR)
0.2322
Adj. R2
*p<.05;**p<.01;***p<.001.
a
Asymptotic standard errors are in parentheses.
Note: sample size= 2049.
Flour
Coefficient
S.E.
0.119 (0.106)
-0.020 (0.030)
0.0009 (0.002)
0.002* (0.001)
-0.018*** (0.005)
-0.012*** (0.003)
-0.012*** (0.002)
0.0002* (0.00009)
0.004* (0.002)
-0.012*** (0.003)
-0.006 (0.003)
0.042*** (0.003)
0.001 (0.002)
-0.008*** (0.002)
0.3493
Coarse Grains
Coefficient
S.E.
0.031 (0.029)
-0.006 (0.008)
0.0004 (0.0006)
0.0007** (0.0003)
0.001 (0.001)
-0.0005 (0.0008)
-0.002*** (0.0006)
0.00003 (0.00002)
0.0003 (0.0004)
-0.003*** (0.0009)
-0.003*** (0.0008)
0.001* (0.0007)
-0.0001 (0.0006)
-0.003*** (0.0006)
0.0474
Fresh Vegetables
Coefficient
S.E.
0.503** (0.163)
-0.090 (0.046)
0.006 (0.003)
-0.001 (0.001)
0.005 (0.008)
0.009* (0.004)
0.0007 (0.003)
0.0003* (0.0001)
-0.002 (0.002)
-0.002 (0.005)
-0.003 (0.005)
-0.019*** (0.004)
-0.012*** (0.003)
0.002 (0.003)
0.0270
Fresh Fruits
Coefficient
S.E.
0.756*** (0.198)
-0.177** (0.056)
0.012** (0.004)
0.00002 (0.002)
0.017 (0.010)
0.056*** (0.005)
0.024*** (0.004)
-0.0007*** (0.0002)
-0.008** (0.003)
0.019** (0.006)
0.013* (0.006)
0.005 (0.005)
-0.023*** (0.004)
0.009* (0.004)
0.1876
Table 5.2: Regression Results for Engel Curve Analysis, 1998.
(Continued)
105
Table 5.2: Continued
Parameter
α
β
γ
δ1 (HS)
δ2 (NC)
δ3 (INH)
δ4 (INM)
δ5 (AGE)
δ6 (MALE)
δ7 (EDH)
δ8 (EDM)
δ9 (PR1)
δ10 (PR2)
δ11 (FR)
Adj. R2
Pork
Coefficient
S.E. a
0.653** (0.252)
-0.142* (0.072)
0.011* (0.005)
0.005* (0.002)
-0.021 (0.013)
-0.036*** (0.007)
-0.006 (0.005)
-0.0001 (0.0002)
0.0001 (0.004)
-0.003 (0.008)
0.004 (0.007)
-0.032*** (0.006)
-0.006 (0.005)
-0.017** (0.005)
0.0590
Eggs
Parameter
Coefficient
S.E.
α
-0.083 (0.117)
β
0.051 (0.033)
γ
-0.004 (0.002)
δ1 (HS)
-0.0009 (0.001)
-0.003 (0.006)
δ2 (NC)
δ3 (INH)
-0.008* (0.003)
-0.006** (0.002)
δ4 (INM)
δ5 (AGE)
0.00005 (0.0001)
0.003 (0.002)
δ6 (MALE)
δ7 (EDH)
-0.006 (0.004)
-0.005 (0.003)
δ8 (EDM)
δ9 (PR1)
0.060*** (0.003)
0.021*** (0.002)
δ10 (PR2)
δ11 (FR)
-0.011*** (0.002)
0.3999
Adj. R2
*p<.05;**p<.01;***p<.001.
a
Asymptotic standard errors are in parentheses.
Note: sample size= 2049.
Beef and Mutton
Coefficient
S.E.
-0.163 (0.099)
0.052 (0.028)
-0.004 (0.002)
-0.0003 (0.0009)
0.009 (0.005)
-0.007** (0.003)
-0.003 (0.002)
0.0001 (0.00008)
0.002 (0.001)
-0.002 (0.003)
-0.002 (0.003)
0.004 (0.002)
-0.013*** (0.002)
0.007*** (0.002)
0.0615
Poultry
Coefficient
S.E.
-0.009 (0.142)
0.029 (0.040)
-0.002 (0.003)
-0.003* (0.001)
-0.002 (0.007)
-0.012** (0.004)
0.001 (0.003)
-0.0003* (0.0001)
0.003 (0.002)
-0.002 (0.004)
-0.0008 (0.004)
-0.057*** (0.003)
-0.037*** (0.003)
0.009** (0.003)
0.2134
Aquatic Products
Coefficient
S.E.
-0.257 (0.158)
0.054 (0.045)
-0.002 (0.003)
0.003 (0.001)
-0.006 (0.008)
0.007 (0.004)
0.009** (0.003)
-0.0002 (0.0001)
-0.002 (0.002)
0.006 (0.005)
0.006 (0.004)
0.012** (0.004)
0.034*** (0.003)
0.005 (0.003)
0.1328
Fresh Milk
Coefficient
S.E.
0.071 (0.142)
-0.026 (0.040)
0.002 (0.003)
-0.001 (0.001)
0.021** (0.007)
0.026*** (0.004)
0.015*** (0.003)
-0.0002 (0.0001)
-0.001 (0.002)
0.017*** (0.004)
0.004 (0.004)
0.013*** (0.003)
0.019*** (0.003)
0.011*** (0.003)
0.0787
(Continued)
106
Table 5.2: Continued
Parameter
α
β
γ
δ1 (HS)
δ2 (NC)
δ3 (INH)
δ4 (INM)
δ5 (AGE)
δ6 (MALE)
δ7 (EDH)
δ8 (EDM)
δ9 (PR1)
δ10 (PR2)
δ11 (FR)
Adj. R2
Yogurt
Coefficient
S.E. a
0.084 (0.051)
-0.023 (0.015)
0.002 (0.001)
-0.0002 (0.0005)
0.006* (0.003)
0.004** (0.001)
0.004*** (0.001)
-0.0001* (0.00004)
-0.001 (0.0008)
0.005*** (0.002)
0.002 (0.001)
0.0002 (0.001)
0.005*** (0.001)
0.002 (0.001)
0.0451
Sugar
Parameter
Coefficient
S.E.
α
0.046 (0.028)
β
-0.008 (0.008)
γ
0.0005 (0.0006)
δ1 (HS)
-0.0005 (0.0003)
0.0004 (0.001)
δ2 (NC)
δ3 (INH)
-0.003*** (0.0008)
-0.001** (0.0005)
δ4 (INM)
δ5 (AGE)
0.00006** (0.00002)
0.0002 (0.0004)
δ6 (MALE)
δ7 (EDH)
-0.001 (0.0009)
-0.0002 (0.0008)
δ8 (EDM)
δ9 (PR1)
-0.003*** (0.0007)
-0.001* (0.0006)
δ10 (PR2)
δ11 (FR)
-0.0002 (0.0006)
0.0296
Adj. R2
*p<.05;**p<.01;***p<.001.
a
Asymptotic standard errors are in parentheses.
Note: sample size= 2049.
Bean and its Products
Coefficient
S.E.
-0.054 (0.030)
0.018* (0.008)
-0.001* (0.0006)
-0.0003 (0.0003)
-0.001 (0.002)
-0.002* (0.0008)
-0.002*** (0.0006)
0.0002*** (0.00003)
0.0009* (0.0004)
0.0007 (0.0009)
0.0007 (0.0008)
0.008*** (0.0007)
0.002** (0.0006)
-0.001* (0.0006)
0.1825
Fats and Oils
Coefficient
S.E.
-0.425*** (0.121)
0.131*** (0.034)
-0.009*** (0.002)
0.0003 (0.001)
-0.012 (0.006)
-0.024*** (0.003)
-0.019*** (0.002)
0.0002 (0.0001)
0.005** (0.002)
-0.013*** (0.004)
-0.006 (0.003)
0.009** (0.003)
0.010*** (0.002)
-0.004 (0.002)
0.0980
Nuts
Coefficient
S.E.
0.043 (0.056)
-0.008 (0.016)
0.0004 (0.001)
0.0005 (0.0005)
-0.002 (0.003)
0.010*** (0.002)
0.004*** (0.001)
0.0001** (0.00005)
-0.0003 (0.0008)
-0.001 (0.002)
-0.0001 (0.002)
0.017*** (0.001)
0.008*** (0.001)
0.001 (0.001)
0.1244
Cakes
Coefficient
S.E.
0.373** (0.135)
-0.064 (0.038)
0.003 (0.003)
-0.005*** (0.001)
0.009 (0.007)
0.026*** (0.004)
0.012*** (0.003)
-0.0003** (0.0001)
-0.006** (0.002)
0.010* (0.004)
0.004 (0.004)
0.003 (0.003)
-0.013*** (0.003)
0.009** (0.003)
0.1587
107
Variable a, b
Description
Sample Mean
Sample S.D. c
Group 1
w1
budget share of food 1
0.619
(0.314)
budget share of food 2
w2
0.160
(0.238)
budget share of food 3
w3
0.050
(0.079)
budget share of food 4
w4
0.171
(0.165)
price for food 1 (in Yuan/kg)
p1
2.688
(0.783)
p2
price for food 2 (in Yuan/kg)
2.648
(0.863)
price for food 3 (in Yuan/kg)
p3
3.405
(1.892)
price for food 4 (in Yuan/kg)
p4
2.567
(1.394)
X
expenditure for group 1 (in Yuan)
168.720
(105.233)
Group 2
w5
budget share of food 5
0.597
(0.153)
budget share of food 6
w6
0.403
(0.153)
price for food 5 (in Yuan/kg)
p5
2.037
(0.824)
p6
price for food 6 (in Yuan/kg)
2.960
(2.235)
X
expenditure for group 2 (in Yuan)
385.545
(241.189)
Group 3
w7
budget share of food 7
0.444
(0.146)
budget share of food 8
w8
0.061
(0.073)
budget share of food 9
w9
0.155
(0.102)
budget share of food 10
w10
0.171
(0.119)
budget share of food 11
w11
0.169
(0.112)
price for food 7 (in Yuan/kg)
p7
14.447
(3.398)
p8
price for food 8 (in Yuan/kg)
16.498
(4.937)
price for food 9 (in Yuan/kg)
p9
14.779
(4.617)
price for food 10 (in Yuan/kg)
p10
6.462
(1.153)
price for food 11 (in Yuan/kg)
p11
14.584
(8.466)
X
expenditure for group 3 (in Yuan)
574.887
(331.615)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b.
Sample sizes for group 1= 2,045; group 2= 2,049; group 3= 2,049; group 4= 2,038; and group 5=2,048.
c
Standard deviations are in parentheses.
Table 5.3: Definitions and Descriptive Statistics of Variables in the QAIDS Model, 1998.
(Continued)
108
Table 5.3: Continued
Variable a, b
Description
Sample Mean
Sample S.D. c
Group 4
w12
budget share of food 12
0.389
(0.352)
budget share of food 13
w13
0.095
(0.189)
budget share of food 14
w14
0.516
(0.373)
price for food 12 (in Yuan/kg)
p12
5.393
(2.745)
p13
price for food 13 (in Yuan/kg)
6.788
(3.064)
price for food 14 (in Yuan/kg)
p14
2.663
(1.155)
X
expenditure for group 4 (in Yuan)
63.486
(77.966)
Group 5
w15
budget share of food 15
0.405
(0.232)
budget share of food 16
w16
0.061
(0.064)
budget share of food 17
w17
0.173
(0.126)
budget share of food 18
w18
0.360
(0.221)
price for food 15 (in Yuan/kg)
p15
10.651
(2.428)
p16
price for food 16 (in Yuan/kg)
4.708
(1.221)
price for food 17 (in Yuan/kg)
p17
9.324
(5.358)
price for food 18 (in Yuan/kg)
p18
16.518
(9.502)
X
expenditure for group 5 (in Yuan)
182.285
(101.931)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b.
Sample sizes for group 1= 2,045; group 2= 2,049; group 3= 2,049; group 4= 2,038; and group 5=2,048.
c
Standard deviations are in parentheses.
109
Test
λi = 0
vs.
λi ≠ 0
Wald Test Statistic
P-value
a
Food item 1
14.054
0.0002
Food item 2
8.28
0.0040
Food item 3
9.82
0.0017
Food item 4
231.03
<.0001
Food item 5
48.23
<.0001
Food item 6
48.23
<.0001
Food item 7
0.55
0.4570
Food item 8
7.36
0.0067
Food item 9
0.41
0.5200
Food item 10
0.54
0.4628
Food item 11
10.52
0.0012
Food item 12
40.34
<.0001
Food item 13
0.96
0.3275
Food item 14
34.21
<.0001
Food item 15
2.71
0.0997
Food item 16
0.25
0.6183
Food item 17
34.63
<.0001
Food item 18
0.55
0.4572
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 5.4: The Wald Test Results of the QAIDS vs. AIDS, 1998.
110
Group 1
Marshallian
Food
Price
of a
Total
a
Item
1
2
3
4
Expenditure
1
-2.066
0.664
0.009
0.179
1.214
2
2.765
-0.183
-0.704
0.878
-2.757
3
0.562
-0.534
-0.257
-0.282
0.511
4
1.101
-0.600
-0.079
-0.906
0.483
Hicksian
Food
Price
of
Item
1
2
3
4
1
-1.315
0.859
0.070
0.386
2
3.309
-2.616
-0.139
-0.554
3
0.879
-0.452
-0.231
-0.195
4
1.400
-0.523
-0.055
-0.823
Group 2
Marshallian
Food
Price
of
Total
Item
5
6
Expenditure
5
-0.773
-0.176
0.949
6
-0.336
-0.740
1.076
Hicksian
Food
Price
of
Item
5
6
5
-0.207
0.207
6
0.306
-0.306
Group 3
Marshallian
Food
Price
of
Total
Item
7
8
9
10
11
Expenditure
7
-1.001
0.107
0.068
-0.144
-0.091
1.060
8
1.057
-0.999
-0.483
-0.335
0.039
0.722
9
-0.007
-0.177
-1.084
0.230
-0.142
1.180
10
0.039
-0.067
0.058
0.060
0.504
-0.595
11
-0.415
-0.052
0.014
-0.119
-0.706
1.279
Hicksian
Food
Price
of
Item
7
8
9
10
11
7
-0.530
0.172
0.233
0.037
0.088
8
1.378
-0.955
-0.371
-0.211
0.160
9
0.517
-0.105
-0.902
0.432
0.057
10
0.263
-0.036
0.137
-0.509
0.146
11
0.153
0.026
0.212
0.099
-0.490
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 5.5: Within-Group Expenditure and Price Elasticities, 1998.
(Continued)
111
Table 5.5: Continued
Group 4
Marshallian
Food
Price of a
Total
a
Item
12
13
14
Expenditure
12
-1.065
-0.070
-0.450
1.585
13
-0.141
-0.928
-0.170
1.239
14
0.075
0.040
-0.630
0.515
Hicksian
Food
Price of
Item
12
13
14
12
-0.449
0.081
0.368
13
0.340
-0.809
0.469
14
0.275
0.089
-0.364
Group 5
Marshallian
Food
Price
of
Total
Item
15
16
17
18
Expenditure
15
-0.951
-0.016
0.004
-0.233
1.195
16
-0.150
-0.646
0.015
0.162
0.619
17
0.008
0.0004
-0.713
-0.142
0.846
18
-0.034
-0.043
-0.145
-0.698
0.920
Hicksian
Food
Price
of
Item
15
16
17
18
15
-0.466
0.057
0.211
0.198
16
0.101
-0.608
0.122
0.385
17
0.351
0.052
-0.566
0.163
18
0.339
0.014
0.014
-0.366
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
112
Group 1 (Sample size: 2045)
Variable a Description
Sample Mean Sample S.D. b
w1
Budget share for food 1
0.619
(0.314)
w2
Budget share for food 2
0.160
(0.238)
Budget share for food 3
w3
0.050
(0.079)
Budget share for food 4
w4
0.171
(0.165)
Price of food 1 (in Yuan/kg)
p1
2.688
(0.783)
p2
Price of food 2 (in Yuan/kg)
2.648
(0.863)
Price of food 3 (in Yuan/kg)
p3
3.405
(1.892)
Price of food 4 (in Yuan/kg)
p4
2.567
(1.394)
X
Total expenditure of group 1 (in Yuan)
168.720
(105.233)
HS
Household size (in persons)
3.225
(0.785)
NC
Ratio of number of children to household size
0.212
(0.162)
INH
1 if high income household; 0 otherwise
0.200
(0.400)
INM
1 if middle income household; 0 otherwise
0.600
(0.490)
AGE
Age of household head (in years)
45.221
(10.737)
MALE
1 if male household head; 0 otherwise
0.652
(0.477)
EDH
1 if high education level; 0 otherwise
0.229
(0.421)
EDM
1 if middle education level; 0 otherwise
0.704
(0.457)
PR1
1 if Shandong; 0 otherwise
0.316
(0.465)
PR2
1 if Jiangsu; 0 otherwise
0.391
(0.488)
FR
1 if the ownership of refrigerator; 0 otherwise
0.874
(0.332)
Group 2 (Sample size: 2049)
Variable a Description
Sample Mean Sample S.D. b
w5
Budget share for food 5
0.597
(0.153)
w6
Budget share for food 6
0.403
(0.153)
Price of food 5 (in Yuan/kg)
p5
2.037
(0.824)
p6
Price of food 6 (in Yuan/kg)
2.960
(2.235)
X
Total expenditure of group 2 (in Yuan)
385.545
(241.189)
HS
Household size (in persons)
3.225
(0.785)
NC
Ratio of number of children to household size
0.212
(0.162)
INH
1 if high income household; 0 otherwise
0.200
(0.400)
INM
1 if middle income household; 0 otherwise
0.600
(0.490)
AGE
Age of household head (in years)
45.200
(10.744)
MALE
1 if male household head; 0 otherwise
0.652
(0.477)
EDH
1 if high education level; 0 otherwise
0.230
(0.421)
EDM
1 if middle education level; 0 otherwise
0.703
(0.457)
PR1
1 if Shandong; 0 otherwise
0.317
(0.466)
PR2
1 if Jiangsu; 0 otherwise
0.390
(0.488)
FR
1 if the ownership of refrigerator; 0 otherwise
0.874
(0.332)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Standard deviations are in parentheses.
Table 5.6: Definitions and Descriptive Statistics of Variables in the QAIDS Model with
Demographic Variables, 1998.
(Continued)
113
Table 5.6: Continued
Group 3 (Sample size: 2049)
Variable a Description
Sample Mean Sample S.D. b
w7
Budget share for food 7
0.444
(0.146)
w8
Budget share for food 8
0.061
(0.073)
Budget share for food 9
w9
0.155
(0.102)
Budget share for food 10
w10
0.171
(0.119)
Budget share for food 11
w11
0.169
(0.112)
p7
Price of food 7 (in Yuan/kg)
14.447
(3.398)
p8
Price of food 8 (in Yuan/kg)
16.498
(4.937)
Price of food 9 (in Yuan/kg)
14.779
(4.617)
p9
Price of food 10 (in Yuan/kg)
6.462
(1.153)
p10
Price of food 11 (in Yuan/kg)
14.584
(8.466)
p11
Total expenditure of group 3 (in Yuan)
X
574.887
(331.615)
Household size (in persons)
HS
3.225
(0.785)
Ratio of number of children to household size
NC
0.212
(0.162)
1 if high income household; 0 otherwise
INH
0.200
(0.400)
1 if middle income household; 0 otherwise
INM
0.600
(0.490)
Age
of
household
head
(in
years)
AGE
45.200
(10.744)
1 if male household head; 0 otherwise
MALE
0.652
(0.477)
1 if high education level; 0 otherwise
EDH
0.230
(0.421)
1 if middle education level; 0 otherwise
EDM
0.703
(0.457)
1 if Shandong; 0 otherwise
PR1
0.317
(0.466)
1 if Jiangsu; 0 otherwise
PR2
0.390
(0.488)
1
if
the
ownership
of
refrigerator;
0
otherwise
FR
0.874
(0.332)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Standard deviations are in parentheses.
(Continued)
114
Table 5.6: Continued
Group 4 (Sample size: 2038)
Variable a Description
Sample Mean Sample S.D. b
w12
Budget share for food 12
0.389
(0.352)
w13
Budget share for food 13
0.095
(0.189)
Budget share for food 14
0.516
(0.373)
w14
Price of food 12 (in Yuan/kg)
5.393
(2.745)
p12
p13
Price of food 13 (in Yuan/kg)
6.788
(3.064)
Price of food 14 (in Yuan/kg)
2.663
(1.155)
p14
X
Total expenditure of group 4 (in Yuan)
63.486
(77.966)
HS
Household size (in persons)
3.223
(0.784)
NC
Ratio of number of children to household size
0.213
(0.162)
INH
1 if high income household; 0 otherwise
0.201
(0.401)
INM
1 if middle income household; 0 otherwise
0.599
(0.490)
AGE
Age of household head (in years)
45.170
(10.733)
MALE
1 if male household head; 0 otherwise
0.651
(0.477)
EDH
1 if high education level; 0 otherwise
0.230
(0.421)
EDM
1 if middle education level; 0 otherwise
0.703
(0.457)
PR1
1 if Shandong; 0 otherwise
0.317
(0.466)
PR2
1 if Jiangsu; 0 otherwise
0.390
(0.488)
FR
1 if the ownership of refrigerator; 0 otherwise
0.874
(0.331)
Group 5 (Sample size: 2048)
Variable a Description
Sample Mean Sample S.D. b
w15
Budget share for food 15
0.405
(0.232)
w16
Budget share for food 16
0.061
(0.064)
Budget share for food 17
w17
0.173
(0.126)
Budget share for food 18
w18
0.360
(0.221)
Price of food 15 (in Yuan/kg)
p15
10.651
(2.428)
p16
Price of food 16 (in Yuan/kg)
4.708
(1.221)
Price of food 17 (in Yuan/kg)
p17
9.324
(5.358)
Price of food 18 (in Yuan/kg)
p18
16.518
(9.502)
X
Total expenditure of group 5 (in Yuan)
182.285
(101.931)
HS
Household size (in persons)
3.225
(0.785)
NC
Ratio of number of children to household size
0.212
(0.162)
INH
1 if high income household; 0 otherwise
0.200
(0.400)
INM
1 if middle income household; 0 otherwise
0.600
(0.490)
AGE
Age of household head (in years)
45.203
(10.746)
MALE
1 if male household head; 0 otherwise
0.651
(0.477)
EDH
1 if high education level; 0 otherwise
0.229
(0.421)
EDM
1 if middle education level; 0 otherwise
0.704
(0.457)
PR1
1 if Shandong; 0 otherwise
0.317
(0.466)
PR2
1 if Jiangsu; 0 otherwise
0.391
(0.488)
FR
1 if the ownership of refrigerator; 0 otherwise
0.875
(0.331)
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
Standard deviations are in parentheses.
115
Group 1 (Sample Size: 2,045)
Parameter
Coefficient
0.274**
a1
-0.197**
a2
0.150***
a3
0.188***
b1
0.097***
b2
-0.032***
b3
-0.014***
L1
-0.008*
L2
0.002
L3
-0.250*
r1r1
0.089*
r1r2
0.025*
r1r3
-0.104*
r2r2
-0.008
r2r3
0.016*
r3r3
Group 2 (Sample Size: 2,049)
Parameter
Coefficient
0.369
a5
0.132
b5
-0.014
L5
-0.031
r5r5
S.E. a
(0.106)
(0.071)
(0.028)
(0.043)
(0.027)
(0.008)
(0.004)
(0.003)
(0.001)
(0.103)
(0.038)
(0.011)
(0.043)
(0.006)
(0.007)
Group 4 (Sample Size: 2,038)
Parameter
Coefficient
0.301**
a12
0.100*
a13
0.055*
b12
0.026*
b13
0.018*
L12
-0.002
L13
-0.032
r12r12
-0.009
r12r13
-0.001
r13r13
Group 3 (Sample Size: 2,049)
Parameter
Coefficient
S.E. a
0.379***
(0.065)
a7
-0.030
(0.051)
a8
0.264***
(0.038)
a9
0.236***
(0.060)
a10
0.039
(0.033)
b7
0.032
(0.025)
b8
-0.008
(0.019)
b9
-0.025
(0.017)
b10
-0.005
(0.005)
L7
-0.004
(0.003)
L8
0.002
(0.003)
L9
-0.0003
(0.002)
L10
-0.020
(0.022)
r7r7
0.014
(0.015)
r7r8
0.003
(0.005)
r7r9
0.013
(0.014)
r7r10
-0.001
(0.004)
r8r8
-0.010
(0.010)
r8r9
-0.003
(0.004)
r8r10
0.008
(0.009)
r9r9
-0.002
(0.003)
r9r10
0.002
(0.005)
r10r10
*p<.05;**p<.01;***p<.001.
a
Asymptotic standard errors are in parentheses.
Group 5 (Sample Size: 2,048)
Parameter
Coefficient
0.160
a15
0.200***
a16
0.304***
a17
0.155*
b15
-0.014
b16
-0.153***
b17
-0.012
L15
-0.003
L16
0.019**
L17
0.048
r15r15
-0.005
r15r16
0.019
r15r17
0.046
r16r16
-0.016
r16r17
0.035
r17r17
S.E. a
(0.254)
(0.110)
(0.012)
(0.052)
S.E. a
(0.101)
(0.051)
(0.023)
(0.013)
(0.008)
(0.002)
(0.025)
(0.008)
(0.005)
S.E. a
(0.108)
(0.041)
(0.084)
(0.061)
(0.011)
(0.044)
(0.010)
(0.002)
(0.006)
(0.037)
(0.007)
(0.018)
(0.025)
(0.010)
(0.023)
Table 5.7: Parameter Estimates for Expenditure and Prices in the QAIDS Model with
Demographic Variables, 1998.
116
Food Item
a
Demographic Variables b
NC
INH
Coefficient
S.E.
Coefficient
-0.048
(0.114)
0.106*
-0.066*
(0.034)
-0.028
0.011
(0.021)
0.027***
0.013
(0.034)
0.020
0.040
(0.332)
0.054
0.115
(0.166)
0.148
-0.306
(0.269)
-0.311
-0.005
(0.032)
-0.035
-0.163
(0.157)
-0.187
-0.090
(0.086)
-0.099
-0.090
(0.087)
-0.048
0.007
(0.168)
0.179*
0.049
(0.053)
0.042
-0.200
(0.130)
-0.066
-0.048
(0.104)
-0.186**
-0.007
(0.019)
-0.028*
-0.018
(0.020)
0.061***
0.045
(0.105)
0.063
HS
INM
Coefficient
S.E. c
S.E.
Coefficient
S.E.
1
0.048*
(0.023)
(0.046)
-0.073
(0.041)
2
0.017***
(0.005)
(0.016) -0.044*** (0.012)
3
0.002
(0.003)
(0.008)
-0.007
(0.008)
4
-0.013*
(0.005)
(0.017)
-0.001
(0.014)
5
-0.248
(0.238)
(0.198)
-0.086
(0.186)
6
-0.110
(0.118)
(0.099)
0.017
(0.088)
7
0.025
(0.027)
(0.211)
-0.373
(0.229)
8
0.002
(0.003)
(0.024)
-0.039
(0.025)
9
0.003
(0.016)
(0.122)
-0.219
(0.132)
10
0.002
(0.009)
(0.068)
-0.125
(0.074)
11
0.006
(0.009)
(0.064)
-0.085
(0.068)
12
-0.034
(0.047)
(0.084)
0.090
(0.067)
13
-0.019
(0.015)
(0.025)
0.023
(0.019)
14
-0.084*
(0.037)
(0.070)
-0.063
(0.053)
15
0.037*
(0.017)
(0.057)
-0.157**
(0.052)
16
-0.0009
(0.003)
(0.011)
-0.010
(0.010)
17
-0.0002
(0.003)
(0.012)
0.032***
(0.009)
18
0.011
(0.017)
(0.058)
-0.021
(0.049)
*p<.05;**p<.01;***p<.001.
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
HS=household size; NC= Ratio of number of children to household size; INH= 1 if high income
household; 0 otherwise; and INM= 1 if middle income household; 0 otherwise.
c
Asymptotic standard errors are in parentheses.
Table 5.8: Parameter Estimates for Demographic Variables in the QAIDS Model with
Demographic Variables, 1998.
(Continued)
117
Table 5.8: Continued
Food Item
a
Demographic Variables b
MALE
EDH
Coefficient
S.E.
Coefficient
S.E.
-0.092**
(0.034)
0.122
(0.087)
0.009
(0.010)
-0.011
(0.021)
-0.012*
(0.005)
0.0008
(0.011)
0.008
(0.010)
0.005
(0.019)
-0.026
(0.111)
-0.625
(0.650)
-0.028
(0.056)
-0.248
(0.324)
0.038
(0.047)
0.105
(0.113)
0.006
(0.006)
0.010
(0.014)
0.031
(0.029)
0.059
(0.067)
0.020
(0.015)
0.014
(0.034)
0.011
(0.016)
0.040
(0.032)
-0.011
(0.045)
-0.021
(0.097)
-0.011
(0.013)
0.029
(0.033)
0.030
(0.037)
0.040
(0.078)
0.065*
(0.027)
-0.106
(0.067)
0.003
(0.005)
-0.010
(0.012)
0.001
(0.006)
-0.003
(0.015)
-0.004
(0.028)
0.024
(0.066)
AGE
EDM
Coefficient
S.E. c
Coefficient
S.E.
1
0.0007
(0.002)
0.124
(0.081)
2
-0.0002
(0.0005)
-0.002
(0.019)
3
0.0004
(0.0003)
-0.0007
(0.009)
4
0.002*** (0.0004)
0.007
(0.017)
5
-0.005
(0.011)
-0.681
(0.709)
6
-0.006
(0.006)
-0.294
(0.353)
7
0.004
(0.002)
-0.040
(0.136)
8
0.0005
(0.0003)
-0.010
(0.017)
9
0.002
(0.001)
-0.034
(0.081)
10
0.001*
(0.0007)
-0.030
(0.041)
11
0.001
(0.0007)
-0.008
(0.037)
12
-0.008**
(0.003)
-0.028
(0.089)
13
-0.003**
(0.0009)
0.020
(0.031)
14
0.007**
(0.002)
0.040
(0.071)
15
0.001
(0.002)
-0.085
(0.065)
16
0.0005
(0.0003)
-0.002
(0.011)
17
0.002*** (0.0003)
0.003
(0.014)
18
-0.003
(0.002)
-0.035
(0.063)
*p<.05;**p<.01;***p<.001.
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
AGE=age of household head; MALE= 1 if male household head; 0 otherwise; EDH= 1 if high education
level; 0 otherwise; and EDM= 1 if middle education level; 0 otherwise.
c
Asymptotic standard errors are in parentheses.
(Continued)
118
Table 5.8: Continued
Food
Item a
Demographic Variables b
PR2
Coefficient
S.E.
-0.557***
(0.120)
0.002
(0.035)
-0.036*
(0.018)
0.010
(0.033)
0.402
(0.394)
0.313
(0.193)
-0.367
(0.232)
-0.061*
(0.028)
-0.301*
(0.136)
-0.038
(0.075)
-0.068
(0.070)
-0.129*
(0.061)
0.0006
(0.019)
0.087
(0.056)
0.124**
(0.040)
-0.013
(0.008)
0.051***
(0.009)
0.045
(0.043)
PR1
FR
Coefficient
S.E.
Coefficient
S.E. c
1
-0.969***
(0.129)
0.046
(0.042)
2
0.295***
(0.039)
-0.010
(0.011)
3
-0.033
(0.019)
-0.013
(0.007)
4
0.067
(0.036)
-0.014
(0.012)
5
0.425
(0.431)
-0.354
(0.366)
6
0.387
(0.211)
-0.146
(0.180)
7
-0.203
(0.132)
0.044
(0.077)
8
0.009
(0.017)
0.026*
(0.009)
9
-0.226
(0.078)
0.057
(0.046)
10
0.127**
(0.043)
-0.016
(0.025)
11
-0.063
(0.041)
0.030
(0.023)
12
-0.077
(0.064)
0.066
(0.072)
13
-0.030
(0.020)
0.022
(0.021)
14
0.176**
(0.054)
-0.088
(0.056)
15
-0.064
(0.037)
-0.039
(0.038)
16
-0.051***
(0.007)
-0.011
(0.007)
17
0.100***
(0.009)
0.002
(0.008)
18
0.132***
(0.035)
0.044
(0.038)
*p<.05;**p<.01;***p<.001.
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
PR1= 1 if Shandong; 0 otherwise; PR2= 1 if Jiangsu; 0 otherwise; FR= 1 if the ownership of refrigerator;
0 otherwise.
c
Asymptotic standard errors are in parentheses.
119
Test
λi = 0
vs.
λi ≠ 0 a
Wald Test Statistic
P-value
Food item 1
11.04
0.0009
Food item 2
6.23
0.0125
Food item 3
3.48
0.0622
Food item 4
20.01
<.0001
Food item 5
1.43
0.2313
Food item 6
1.43
0.2313
Food item 7
1.10
0.2949
Food item 8
1.42
0.2337
Food item 9
0.43
0.5101
Food item 10
0.03
0.4628
Food item 11
2.59
0.1075
Food item 12
5.00
0.0253
Food item 13
1.19
0.2758
Food item 14
4.92
0.0265
Food item 15
1.47
0.2255
Food item 16
2.02
0.1553
Food item 17
8.80
0.0030
Food item 18
0.17
0.6830
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 5.9: The Wald Test Results of the QAIDS Models with Demographic Variables,
1998.
120
Group 1
Marshallian
Food
Price
of a
Total
a
Item
1
2
3
4
Expenditure
1
-1.365
0.204
0.010
0.031
1.121
2
0.672
-0.105
-0.216
1.176
-1.527
3
0.479
-0.287
-0.605
-0.244
0.657
4
0.553
-0.161
-0.051
-0.838
0.497
Hicksian
Food
Price
of
Item
1
2
3
4
1
-0.672
0.384
0.065
0.222
2
1.400
-1.338
-0.047
-0.015
3
0.886
-0.182
-0.573
-0.131
4
0.860
-0.081
-0.026
-0.753
Group 2
Marshallian
Food
Price
of
Total
Item
5
6
Expenditure
5
-0.964
-0.014
0.978
6
-0.054
-0.979
1.033
Hicksian
Food
Price
of
Item
5
6
5
-0.380
0.380
6
0.563
-0.563
Group 3
Marshallian
Food
Price
of
Total
Item
7
8
9
10
11
Expenditure
7
-1.042
0.037
0.005
0.025
-0.030
1.006
8
0.260
-0.986
-0.175
-0.080
-0.044
1.023
9
0.004
-0.069
-0.958
-0.018
0.003
1.038
10
0.127
-0.019
0.029
-0.953
-0.024
0.839
11
-0.117
-0.020
-0.017
-0.067
-0.884
1.105
Hicksian
Food
Price
of
Item
7
8
9
10
11
7
-0.595
0.098
0.161
0.196
0.139
8
0.715
-0.923
-0.016
0.095
0.129
9
0.465
-0.005
-0.798
0.159
0.179
10
0.500
0.033
0.159
-0.809
0.118
11
0.374
0.048
0.154
0.122
-0.697
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 5.10: Within-Group Expenditure and Price Elasticities, 1998.
(Continued)
121
Table 5:10: Continued
Group 4
Marshallian
Food
Price of a
Total
a
Item
12
13
14
Expenditure
12
-1.170
-0.059
-0.157
1.386
13
-0.109
-1.019
-0.025
1.154
14
0.148
0.048
-0.877
0.681
Hicksian
Food
Price of
Item
12
13
14
12
-0.631
0.073
0.558
13
0.339
-0.909
0.570
14
0.413
0.113
-0.526
Group 5
Marshallian
Food
Price
of
Total
Item
15
16
17
18
Expenditure
15
-0.800
0.028
-0.016
-0.427
1.215
16
-0.120
-0.258
0.251
-0.357
0.484
17
-0.079
-0.137
-0.631
0.128
0.719
18
-0.166
-0.075
-0.115
-0.625
0.982
Hicksian
Food
Price
of
Item
15
16
17
18
15
-0.308
0.103
0.194
0.011
16
0.076
-0.327
-0.174
0.425
17
0.212
-0.093
-0.506
0.388
18
0.231
-0.015
0.055
-0.271
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
122
Variable
Constant
TFE
TFE2
HS
NC
INH
INM
AGE
AGE2
EDH
EDM
MALE
PR1
PR2
FR
Flour (2)
Coefficient S.E. a
-1.862** (0.583)
-0.027 (0.070)
0.002 (0.006)
0.086* (0.043)
-0.043 (0.241)
-0.088 (0.132)
-0.056 (0.095)
0.052* (0.022)
-0.0003 (0.0002)
0.075 (0.141)
-0.016 (0.127)
0.075 (0.069)
1.604*** (0.121)
0.347*** (0.093)
-0.341*** (0.102)
Fresh Milk (12)
Variable
Coefficient S.E.
Constant
-1.328* (0.565)
TFE
0.463*** (0.071)
TFE2
-0.034*** (0.007)
HS
0.065 (0.042)
NC
0.972*** (0.235)
INH
0.774*** (0.132)
INM
0.535*** (0.085)
AGE
-0.015 (0.021)
AGE2
0.00006 (0.0002)
EDH
0.189 (0.139)
EDM
0.045 (0.124)
MALE
-0.123 (0.069)
PR1
0.677*** (0.116)
PR2
0.207* (0.097)
FR
0.343*** (0.092)
*p<.05;**p<.01;***p<.001.
a
Asymptotic standard errors are in parentheses.
Coarse Grains (3)
Coefficient S.E.
-0.982 (0.567)
0.288*** (0.074)
-0.018* (0.007)
0.040 (0.044)
0.277 (0.243)
0.219 (0.137)
0.103 (0.087)
0.041 (0.021)
-0.0003 (0.0002)
-0.480** (0.159)
-0.355* (0.147)
-0.090 (0.071)
0.235* (0.119)
-0.192 (0.100)
0.073 (0.094)
Beef and Mutton (8)
Coefficient S.E.
-1.453* (0.591)
0.524*** (0.077)
-0.037*** (0.007)
0.075 (0.046)
0.737** (0.259)
-0.233 (0.149)
-0.117 (0.093)
0.048* (0.022)
-0.0004 (0.0002)
-0.346* (0.165)
-0.263 (0.151)
0.018 (0.076)
0.172 (0.133)
-0.681*** (0.112)
0.326*** (0.097)
Yogurt (13)
Coefficient S.E.
-0.559 (0.542)
0.301*** (0.069)
-0.020** (0.006)
0.053 (0.041)
0.952*** (0.228)
0.341** (0.126)
0.179* (0.087)
-0.060** (0.021)
0.0005* (0.0002)
0.361* (0.141)
0.128 (0.130)
-0.004 (0.065)
0.261* (0.109)
0.435*** (0.093)
0.354*** (0.097)
Sugar (16)
Coefficient S.E.
0.943 (0.733)
0.258** (0.091)
-0.018* (0.008)
-0.015 (0.059)
-0.287 (0.316)
-0.580*** (0.173)
-0.192 (0.110)
0.026 (0.028)
-0.0002 (0.0003)
-0.313 (0.235)
-0.167 (0.224)
-0.047 (0.088)
-0.627*** (0.154)
-0.177 (0.143)
-0.092 (0.129)
Table 5.11: Parameter Estimates in Probit Models, 1998.
123
Group 1 (Sample Size: 2,048)
Parameter
Coefficient
S.E.
α1
0.563**
(0.205)
α2
-0.022
(0.047)
α3
0.023
(0.043)
β1
0.122
(0.072)
β2
0.030
(0.019)
β3
-0.024
(0.015)
λ1
-0.010
(0.006)
λ2
-0.003
(0.002)
λ3
0.002
(0.001)
φ2
0.053***
(0.006)
φ3
0.108***
(0.005)
γ1,1
-0.043
(0.051)
γ1,2
-0.002
(0.003)
γ1,3
0.008
(0.009)
γ2,2
-0.004
(0.005)
γ2,3
-0.001
(0.002)
γ3,3
0.003
(0.003)
Group 3 (Sample Size: 2,049)
Parameter
Coefficient
S.E.
α7
0.324***
(0.084)
α8
0.034***
(0.008)
α9
0.263***
(0.044)
α10
0.245***
(0.074)
β7
0.077
(0.049)
β8
-0.002
(0.003)
β9
-0.014
(0.023)
β10
-0.026
(0.026)
λ7
-0.010
(0.007)
λ8
0.0009
(0.0006)
λ9
0.003
(0.003)
λ10
-0.0009
(0.003)
φ8
0.038***
(0.001)
γ7,7
0.0002
(0.013)
γ7,8
0.002
(0.003)
γ7,9
-0.007
(0.009)
γ7,10
0.020
(0.022)
γ8,8
-0.0006
(0.0008)
γ8,9
-0.0002
(0.0006)
γ8,10
-0.002
(0.002)
γ9,9
0.014
(0.015)
γ9,10
-0.008
(0.008)
γ10,10
0.003
(0.007)
*p<.05;**p<.01;***p<.001.
a
Asymptotic Standard Error is in parentheses.
Group 4 (Sample Size: 2,038)
Parameter
Coefficient
α12
0.177
α13
0.256***
β12
0.015
β13
-0.006
λ12
0.003
λ13
-0.0007
φ12
0.553***
φ13
0.062***
γ12,12
-0.0009
γ12,13
0.0002
γ13,13
-0.00007
Group 5 (Sample Size: 2,048)
Parameter
Coefficient
α15
0.233*
α16
0.078***
α17
0.278**
β15
0.127*
β16
0.002
β17
-0.135**
λ15
-0.011
λ16
-0.0008
λ17
0.016*
φ16
0.033***
γ15,15
0.032
γ15,16
-0.001
γ15,17
0.012
γ16,16
0.006
γ16,17
-0.001
γ17,17
0.021
Table 5.12: Parameter Estimates for the Censored QAIDS, 1998.
124
S.E.
(0.127)
(0.071)
(0.027)
(0.010)
(0.005)
(0.001)
(0.030)
(0.015)
(0.003)
(0.0008)
(0.0003)
S.E. a
(0.103)
(0.009)
(0.105)
(0.052)
(0.003)
(0.050)
(0.007)
(0.0008)
(0.006)
(0.001)
(0.028)
(0.002)
(0.012)
(0.004)
(0.002)
(0.018)
Food Item
a
HS
Coefficient
0.004
0.007
0.003
-0.016
S.E. c
(0.085)
(0.005)
(0.005)
(0.008)
Demographic Variables b
NC
INH
Coefficient
S.E.
Coefficient
-0.712
(0.706)
0.186
-0.062
(0.045)
-0.034**
0.051
(0.040)
0.026*
-0.054
(0.075)
0.026
S.E.
(0.185)
(0.012)
(0.011)
(0.021)
INM
Coefficient
-0.324
-0.048**
0.022
-0.030
S.E.
(0.270)
1
(0.018)
2
(0.015)
3
(0.030)
4
5
6
0.018
(0.027)
-0.088
(0.177)
-0.152
(0.126)
-0.177
(0.122)
7
0.003
(0.003)
0.013
(0.018)
-0.015
(0.013)
-0.017
(0.012)
8
0.0007
(0.016)
-0.036
(0.109)
-0.096
(0.076)
-0.106
(0.074)
9
0.0005
(0.008)
-0.027
(0.054)
-0.049
(0.039)
-0.064
(0.038)
10
0.005
(0.008)
-0.027
(0.056)
-0.008
(0.038)
-0.033
(0.035)
11
-0.131
(0.333)
-0.444
(1.376)
-0.373
(1.123)
-0.488
(1.197)
12
-0.181
(0.341)
-0.665
(1.403)
-0.638
(1.151)
-0.694
(1.226)
13
-0.396
(0.763)
-1.714
(3.142)
-1.469
(2.572)
-1.556
(2.739)
14
0.054*
(0.022)
-0.101
(0.153)
-0.225**
(0.087)
-0.159*
(0.071)
15
0.004
(0.003)
-0.015
(0.022)
-0.024*
(0.012)
-0.013
(0.010)
16
-0.001
(0.003)
-0.017
(0.022)
0.061***
(0.012)
0.032***
(0.009)
17
0.026
(0.021)
-0.005
(0.149)
0.022
(0.085)
-0.021
(0.066)
18
*p<.05;**p<.01;***p<.001.
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
HS=household size; NC= Ratio of number of children to household size; INH= 1 if high income
household; 0 otherwise; and INM= 1 if middle income household; 0 otherwise.
c
Asymptotic standard errors are in parentheses.
Table 5.13: Parameter Estimates for the Demographic Variables, Censored QAIDS, 1998.
(Continued)
125
Table 5.13: Continued
Food Item
a
AGE
Coefficient
S.E. c
-0.010
(0.010)
0.0004
(0.0006)
0.002*** (0.0005)
0.0006
(0.001)
Demographic Variables b
MALE
EDH
Coefficient
S.E.
Coefficient
S.E.
-0.206
(0.152)
0.208
(0.246)
0.0008
(0.010)
0.031
(0.016)
0.005
(0.009)
-0.006
(0.014)
-0.003
(0.017)
0.007
(0.025)
EDM
Coefficient
S.E.
0.271
(0.223)
0.030*
(0.014)
-0.015
(0.013)
0.018
(0.022)
1
2
3
4
5
6
0.004*
(0.002)
0.076
(0.060)
0.052
(0.101)
0.055
(0.093)
7
0.0006** (0.0002)
0.011
(0.006)
-0.005
(0.010)
-0.003
(0.009)
8
0.002
(0.001)
0.054
(0.036)
0.024
(0.061)
0.023
(0.055)
9
0.002*
(0.0006)
0.031
(0.018)
-0.003
(0.030)
-0.003
(0.028)
10
0.001
(0.0006)
0.023
(0.018)
0.022
(0.029)
0.019
(0.026)
11
-0.010
(0.012)
0.048
(0.154)
0.013
(0.176)
-0.037
(0.169)
12
-0.004
(0.013)
0.077
(0.158)
0.089
(0.185)
0.033
(0.180)
13
-0.006
(0.028)
0.176
(0.354)
0.074
(0.413)
0.019
(0.400)
14
-0.002
(0.003)
0.074
(0.038)
-0.062
(0.076)
-0.034
(0.070)
15
-0.000005 (0.0004)
0.005
(0.005)
-0.005
(0.011)
-0.002
(0.010)
16
0.002*** (0.0004)
-0.0004
(0.006)
-0.004
(0.012)
0.003
(0.011)
17
-0.006*
(0.003)
0.005
(0.038)
0.067
(0.074)
0.016
(0.067)
18
*p<.05;**p<.01;***p<.001.
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
AGE=age of household head; MALE= 1 if male household head; 0 otherwise; EDH= 1 if high education
level; 0 otherwise; and EDM= 1 if middle education level; 0 otherwise.
c
Asymptotic standard errors are in parentheses.
(Continued)
126
Table 5.13: Continued
Food
Item a
PR1
Coefficient
-1.383*
0.250***
0.045
0.036
S.E. c
(0.639)
(0.042)
(0.036)
(0.069)
Demographic Variables b
PR2
Coefficient
S.E.
-1.220
(0.736)
-0.018
(0.047)
0.038
(0.042)
-0.056
(0.078)
FR
Coefficient
0.062
-0.019
-0.024**
-0.014
S.E.
(0.147)
(0.011)
(0.009)
(0.016)
1
2
3
4
5
6
-0.322
(0.208)
-0.528
(0.310)
0.007
(0.060)
7
-0.033
(0.021)
-0.079*
(0.031)
0.014*
(0.006)
8
-0.315*
(0.124)
-0.403*
(0.185)
0.036
(0.036)
9
0.087
(0.063)
-0.085
(0.093)
-0.025
(0.019)
10
-0.098
(0.060)
-0.113
(0.091)
0.019
(0.018)
11
-0.299
(0.516)
0.051
(0.274)
-0.252
(0.757)
12
-0.250
(0.527)
0.212
(0.282)
-0.438
(0.776)
13
-0.497
(1.180)
0.405
(0.631)
-1.007
(1.734)
14
-0.043
(0.055)
0.111*
(0.053)
-0.081
(0.063)
15
-0.018*
(0.007)
0.0005
(0.008)
-0.010
(0.009)
16
0.096***
(0.008)
0.050***
(0.008)
0.0009
(0.009)
17
0.138**
(0.051)
0.027
(0.054)
0.003
(0.058)
18
*p<.05;**p<.01;***p<.001.
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
PR1= 1 if Shandong; 0 otherwise; PR2= 1 if Jiangsu; 0 otherwise; FR= 1 if the ownership of refrigerator;
0 otherwise.
c
Asymptotic standard errors are in parentheses.
127
Test
λi = 0
vs.
λi ≠ 0 a
Wald Test Statistic
P-value
Food item 1
2.28
0.0930
Food item 2
2.46
0.1169
Food item 3
1.97
0.1599
Food item 4
2.85
0.0911
Food item 5
Food item 6
Food item 7
2.39
0.1224
Food item 8
1.83
0.1767
Food item 9
0.77
0.3791
Food item 10
0.77
0.7843
Food item 11
2.24
0.1341
Food item 12
0.29
0.5891
Food item 13
0.28
0.5999
Food item 14
0.29
0.5881
Food item 15
2.58
0.1084
Food item 16
1.18
0.2768
Food item 17
6.35
0.0118
Food item 18
0.66
0.4152
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 5.14: The Wald Test Results of the Censored QAIDS with Demographic Variables,
1998.
128
Group 1
Marshallian
Food
Price
of
Total
Item a
1
2
3
4
Expenditure
1
-1.075
0.006
0.005
-0.004
1.068
2
0.0003
-1.009
-0.010
-0.008
1.026
3
0.142
-0.033
-0.945
-0.022
0.858
4
0.217
0.004
-0.024
-0.965
0.768
Hicksian
Food
Price
of
Item
1
2
3
4
1
-0.414
0.178
0.058
0.178
2
0.636
-0.844
0.042
0.167
3
0.673
0.104
-0.902
0.124
4
0.692
0.127
0.015
-0.834
Group 2
Marshallian
Food
Price
of
Total
Item a
5
6
Expenditure
5
-0.964
-0.014
0.978
6
-0.054
-0.979
1.033
Hicksian
Food
Price
of
Item
5
6
5
-0.380
0.380
6
0.563
-0.563
Group 3
Marshallian
Food
Price
of
Total
Item a
7
8
9
10
11
Expenditure
7
-0.972
0.005
-0.019
0.038
-0.047
0.995
8
0.009
-1.010
-0.020
-0.037
-0.004
1.062
9
-0.075
-0.003
-0.920
-0.054
-0.001
1.054
10
0.156
-0.003
0.016
-0.944
-0.035
0.810
11
-0.166
-0.003
-0.031
-0.091
-0.838
1.128
Hicksian
Food
Price
of
Item
7
8
9
10
11
7
-0.530
0.066
0.135
0.208
0.121
8
0.481
-0.945
0.144
0.145
0.175
9
0.393
0.061
-0.757
0.126
0.177
10
0.516
0.046
0.142
-0.805
0.101
11
0.336
0.066
0.144
0.101
-0.647
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
Table 5.15: Within-Group Expenditure and Price Elasticities, Censored QAIDS, 1998.
(Continued)
129
Table 5.15: Continued
Group 4
Marshallian
Food
Price of
Total
Item a
12
13
14
Expenditure
12
-1.012
-0.014
-0.030
1.055
13
0.008
-0.991
0.020
0.963
14
0.009
0.010
-0.977
0.958
Hicksian
Food
Price of
Item
12
13
14
12
-0.602
0.087
0.514
13
0.382
-0.899
0.517
14
0.381
0.102
-0.483
Group 5
Marshallian
Food
Price
of
Total
Item a
15
16
17
18
Expenditure
15
-0.893
-0.008
-0.008
-0.261
1.170
16
-0.010
-0.916
-0.027
-0.008
0.961
17
-0.024
-0.0003
-0.766
0.064
0.727
18
-0.107
-0.007
-0.098
-0.736
0.947
Hicksian
Food
Price
of
Item
15
16
17
18
15
-0.419
0.064
0.194
0.161
16
0.379
-0.857
0.140
0.338
17
0.270
0.044
-0.640
0.326
18
0.277
0.051
0.066
-0.394
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6)
fresh fruits, (7) pork, (8) beef and mutton, (9) poultry, (10) eggs, (11) aquatic products, (12) fresh milk,
(13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
130
Variable a Description
Sample Mean Sample S.D. b
w1
budget share of food group 1
0.126
(0.065)
w2
budget share of food group 2
0.277
(0.080)
budget share of food group 3
0.412
(0.095)
w3
budget share of food group 4
0.046
(0.049)
w4
budget share of food group 5
0.139
(0.057)
w5
aggregate price for food group 1 (in Yuan/kg)
2.642
(0.760)
p1
p2
aggregate price for food group 2 (in Yuan/kg)
2.311
(1.168)
aggregate price for food group 3 (in Yuan/kg)
12.461
(2.510)
p3
aggregate price for food group 4 (in Yuan/kg)
3.691
(1.331)
p4
aggregate price for food group 5 (in Yuan/kg)
11.152
(3.319)
p5
X
Total expenditure for the 18 food items (in Yuan)
1,374.16
(668.98)
HS
Household size (in persons)
3.225
(0.785)
NC
Ratio of number of children to household size
0.212
(0.162)
INH
1 if high income household; 0 otherwise
0.200
(0.400)
INM
1 if middle income household; 0 otherwise
0.600
(0.490)
AGE
Age of household head (in years)
45.200
(10.744)
MALE
1 if male household head; 0 otherwise
0.652
(0.477)
EDH
1 if high education level; 0 otherwise
0.230
(0.421)
EDM
1 if middle education level; 0 otherwise
0.703
(0.457)
PR1
1 if Shandong; 0 otherwise
0.317
(0.466)
PR2
1 if Jiangsu; 0 otherwise
0.390
(0.488)
FR
1 if the ownership of refrigerator; 0 otherwise
0.874
(0.332)
a
Group 1: grains; group2: vegetables and fruits; group 3: animal protein food; group 4: dairy and bean
products; and group 5: fats and oils, sugar, nuts, and cakes.
b
Standard deviations are in parentheses.
Table 5.16: Variable Definitions and Descriptive Statistics, Second Stage Demand
System, 1998.
131
Test
λi = 0
vs.
λi ≠ 0 a
Wald Test Statistic
P-value
Food Group 1
73.19
<.0001
Food Group 2
44.09
<.0001
Food Group 3
0.77
0.3794
Food Group 4
0.85
0.3578
Food Group 5
0.06
0.8138
a
Group 1: grains; group2: vegetables and fruits; group 3: animal protein food; group 4: dairy and bean
products; and group 5: fats and oils, sugar, nuts, and cakes.
Table 5.17: Wald Test Results of the Second-Stage QAIDS, 1998.
Marshallian
Food
Price
of a
Total
a
Group
1
2
3
4
5
Expenditure
1
-0.430
-0.438
-0.257
0.074
-0.046
1.096
2
-0.172
-0.634
0.053
-0.004
-0.124
0.880
3
-0.080
-0.029
-0.883
-0.020
-0.101
1.113
4
0.216
-0.055
-0.123
-1.352
0.331
0.984
5
-0.007
-0.229
-0.181
0.117
-0.521
0.821
Hicksian
Food
Price
of
Group
1
2
3
4
5
1
-0.292
-0.134
0.196
0.124
0.106
2
-0.061
-0.390
0.417
0.036
-0.001
3
0.060
0.279
-0.424
0.032
0.053
4
0.339
0.217
0.282
-1.306
0.467
5
0.096
-0.002
0.158
0.155
-0.407
a
Group 1: grains; group2: vegetables and fruits; group 3: animal protein food; group 4: dairy and bean
products; and group 5: fats and oils, sugar, nuts, and cakes.
Table 5.18: Between-Group Expenditure and Price Elasticities, 1998.
132
QAIDS
a1
a2
a3
a4
b1
b2
b3
b4
λ1
λ2
λ3
λ4
r1r1
r1r2
r1r3
r1r4
r2r2
r2r3
r2r4
r3r3
r3r4
r4r4
Coefficient
S.E. b
-0.523*
0.980***
0.378**
0.019
0.242**
-0.226**
-0.024
-0.004
-0.021**
0.018**
0.006
0.0001
-0.116
0.121
0.005
0.005
-0.094
-0.023
-0.001
0.059
-0.010
-0.002
(0.210)
(0.243)
(0.141)
(0.053)
(0.076)
(0.072)
(0.043)
(0.019)
(0.007)
(0.006)
(0.004)
(0.002)
(0.078)
(0.077)
(0.030)
(0.014)
(0.069)
(0.028)
(0.013)
(0.035)
(0.007)
(0.003)
a
Demographic Variables
HS
s1s1
0.001
(0.005)
s1s2
-0.011
(0.008)
s1s3
-0.021
(0.022)
s1s4
-0.002
(0.002)
s1s5
-0.012
(0.007)
Coefficient
S.E. b
Demographic Variables
NC
s2s1
0.001
(0.025)
s2s2
0.034
(0.036)
s2s3
0.049
(0.106)
s2s4
0.024**
(0.008)
s2s5
0.019
(0.032)
INH
s3s1
-0.047***
(0.014)
s3s2
0.058**
(0.020)
s3s3
-0.051
(0.061)
s3s4
0.027***
(0.005)
s3s5
-0.0007
(0.018)
INM
s4s1
-0.027*
(0.011)
s4s2
0.014
(0.013)
s4s3
-0.009
(0.043)
s4s4
0.017***
(0.004)
s4s5
-0.008
(0.013)
AGE
s5s1
0.0006
(0.0004)
s5s2
0.0003
(0.0006)
s5s3
-0.00004
(0.002)
s5s4
-0.00001
(0.0001)
s5s5
0.0003
(0.0005)
MALE
s6s1
0.0005
(0.008)
s6s2
-0.013
(0.012)
s6s3
-0.014
(0.035)
s6s4
-0.001
(0.003)
s6s5
-0.005
(0.011)
Coefficient S.E. b
Demographic Variables
EDH
s7s1
-0.039*
(0.019)
s7s2
-0.014
(0.027)
s7s3
-0.083
(0.074)
s7s4
0.024**
(0.007)
s7s5
-0.028
(0.023)
EDM
s8s1
-0.027
(0.017)
s8s2
-0.016
(0.024)
s8s3
-0.066
(0.065)
s8s4
0.008
(0.007)
s8s5
-0.023
(0.020)
PR1
s9s1
-0.027*
(0.013)
s9s2
0.037
(0.020)
s9s3
-0.036
(0.057)
s9s4
0.020*** (0.004)
s9s5
0.038*
(0.017)
PR2
s10s1
-0.004
(0.013)
s10s2
-0.026
(0.023)
s10s3
-0.079
(0.062)
s10s4
0.025*** (0.005)
s10s5
-0.011
(0.018)
FR
s11s1
-0.022*
(0.010)
s11s2
0.003
(0.014)
s11s3
-0.019
(0.042)
s11s4
0.012**
(0.004)
s11s5
-0.002
(0.013)
*p<.05;**p<.01;***p<.001.
Group 1: grains; group2: vegetables and fruits; group 3: animal protein food; group 4: dairy and bean
products; and group 5: fats and oils, sugar, nuts, and cakes.
b
Asymptotic standard errors are in parentheses.
a
Table 5.19: Parameter Estimates for the Second-Stage QAIDS with Demographic
Variables, 1998.
133
Test
λi = 0
vs.
λi ≠ 0 a
Wald Test Statistic
P-value
Food Group 1
10.28
0.0013
Food Group 2
8.31
0.0039
Food Group 3
2.33
0.1267
Food Group 4
0.01
0.9371
Food Group 5
0.74
0.3899
a
Group 1: grains; group2: vegetables and fruits; group 3: animal protein food; group 4: dairy and bean
products; and group 5: fats and oils, sugar, nuts, and cakes.
Table 5.20: Wald Test Results of the Second-Stage QAIDS with Demographic Variables,
1998.
Marshallian
Food
Price
of
Total
Group a
1
2
3
4
5
Expenditure
1
-0.637
-0.265
-0.113
0.021
-0.072
1.065
2
-0.098
-0.793
0.012
0.004
-0.017
0.891
3
-0.039
-0.051
-0.889
-0.024
-0.096
1.099
4
0.065
0.041
-0.197
-1.051
0.194
0.948
5
-0.040
-0.035
-0.186
0.062
-0.680
0.880
Hicksian
Food
Price
of
Group
1
2
3
4
5
1
-0.503
0.030
0.326
0.070
0.076
2
0.015
-0.546
0.379
0.045
0.107
3
0.099
0.254
-0.435
0.026
0.056
4
0.184
0.304
0.194
-1.007
0.326
5
0.071
0.208
0.177
0.102
-0.558
a
Group 1: grains; group2: vegetables and fruits; group 3: animal protein food; group 4: dairy and bean
products; and group 5: fats and oils, sugar, nuts, and cakes.
Table 5.21: Between-Group Elasticities, the QAIDS with Demographic Variables, 1998.
134
Food a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Cb
UF b
UB b
1.214
0.878
0.511
0.483
0.949
1.076
1.060
0.722
1.180
0.504
1.279
1.585
1.239
0.515
1.195
0.619
0.846
0.920
1.330
0.963
0.560
0.529
0.835
0.947
1.180
0.803
1.313
0.561
1.424
1.559
1.218
0.507
0.981
0.508
0.694
0.755
0.874
0.633
0.368
0.348
0.549
0.622
0.776
0.528
0.863
0.369
0.936
1.025
0.801
0.333
0.645
0.334
0.456
0.496
Marshallian
1
2
3
4
5
6
7
8
Price
9
a
of
10
135
11
12
13
14
15
16
17
18
0.054
0.335
-0.306
-0.225
-0.134
-0.023
-0.043
-0.069
-0.042
0.056
0.010
0.023
-0.012
-0.005
-0.012
-0.027
1 -1.689 0.779
0.008
0.017
-0.008
-0.004
-0.008
-0.020
2 3.038 -2.673 -0.151 -0.591 -0.222 -0.163 -0.097 -0.017 -0.031 -0.050 -0.031 0.040
0.004
0.010
-0.005
-0.002
-0.005
-0.012
3 0.721 -0.485 -0.238 -0.217 -0.129 -0.095 -0.056 -0.010 -0.018 -0.029 -0.018 0.023
0.004
0.009
-0.005
-0.002
-0.005
-0.011
4 1.251 -0.554 -0.061 -0.844 -0.122 -0.090 -0.053 -0.009 -0.017 -0.028 -0.017 0.022
5 -0.112 -0.027 -0.006 -0.019 -0.551 -0.051 0.032 -0.003 0.018 -0.022 0.027 -0.002 -0.0006 -0.001 -0.047 -0.006 -0.019 -0.044
6 -0.127 -0.030 -0.007 -0.021 -0.084 -0.598 0.036 -0.004 0.020 -0.025 0.030 -0.003 -0.0007 -0.002 -0.053 -0.007 -0.022 -0.050
0.074
-0.082
-0.092
-0.012
-0.003
-0.006
-0.038
-0.007
-0.019
-0.043
7 -0.049 -0.017 -0.004 -0.014 -0.022 -0.009 -0.959 0.123
-0.033
-0.012
-0.003
-0.010
-0.015
-0.006
1.085
-0.989
-0.479
-0.292
0.038
-0.008
-0.002
-0.004
-0.026
-0.005
-0.013
-0.030
8
-0.003
-0.006
-0.043
-0.007
-0.021
-0.048
9 -0.054 -0.019 -0.005 -0.016 -0.024 -0.010 0.039 -0.160 -1.078 0.300 -0.143 -0.014
-0.001
-0.003
-0.018
-0.003
-0.009
-0.021
10 -0.023 -0.008 -0.002 -0.007 -0.010 -0.004 0.059 -0.059 0.061 -0.566 0.060 -0.006
-0.003
-0.007
-0.046
-0.008
-0.023
-0.052
11 -0.059 -0.021 -0.005 -0.017 -0.027 -0.011 -0.365 -0.034 0.020 -0.044 -0.708 -0.015
0.041
0.006
0.018
-0.055
-0.032
-0.077
-0.019
-0.019
-0.067
-0.013
-1.768
-0.173
-0.201
0.272
0.016
0.073
0.163
12 0.277
0.032
0.004
0.014
-0.043
-0.025
-0.060
-0.015
-0.015
-0.052
-0.010
-0.691
-1.007
0.024
0.212
0.013
0.057
0.128
13 0.216
0.013
0.002
0.006
-0.018
-0.010
-0.025
-0.006
-0.006
-0.022
-0.004
-0.153
0.006
-0.549
0.088
0.005
0.024
0.053
14 0.090
0.016
0.037
-0.756
0.032
0.118
-0.015
15 0.006 -0.006 -0.002 -0.007 -0.159 -0.115 -0.092 -0.016 -0.029 -0.050 -0.028 0.087
0.009
0.019
-0.049
-0.621
0.073
0.275
16 0.003 -0.003 -0.001 -0.004 -0.082 -0.060 -0.048 -0.008 -0.015 -0.026 -0.015 0.045
0.004
-0.004
-0.002
-0.005
-0.112
-0.081
-0.065
-0.012
-0.021
-0.035
-0.020
0.061
0.012
0.026
0.146
0.034
-0.633
0.013
17
0.013
0.028
0.115
-0.006
-0.058
-0.530
18 0.005 -0.004 -0.002 -0.005 -0.122 -0.088 -0.071 -0.013 -0.023 -0.038 -0.022 0.067
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6) fresh fruits, (7) pork, (8) beef and mutton, (9) poultry,
(10) eggs, (11) aquatic products, (12) fresh milk, (13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
C means conditional expenditure elasticities and UF and UB indicate unconditional expenditure elasticities with respect to food expenditure and total living
expenditure, respectively.
Table 5.22: Unconditional Elasticities for 18 Food Items, the QAIDS, 1998.
(Continued)
135
Table 5.22: Continued
Hicksian
Food
a
1
2
3
4
5
6
7
8
Price
9
of a
10
11
12
13
14
15
16
17
18
136
0.061
0.357
-0.092
-0.071
0.112
0.010
0.043
0.020
0.051
0.093
0.018
0.040
0.062
0.005
0.019
0.043
1 -1.581 0.809
-0.051
0.081
0.008
0.031
0.015
0.037
0.067
0.013
0.029
0.045
0.004
0.014
0.031
2 3.116 -2.652 -0.146 -0.575 -0.067
-0.030
0.047
0.004
0.018
0.009
0.022
0.039
0.008
0.017
0.026
0.002
0.008
0.018
3 0.767 -0.473 -0.235 -0.208 -0.039
-0.028
0.044
0.004
0.017
0.008
0.020
0.037
0.007
0.016
0.025
0.002
0.008
0.017
4 1.295 -0.542 -0.058 -0.835 -0.037
-0.043
-0.008
-0.001
-0.005
-0.416
0.046
0.186
0.017
0.072
0.034
0.085
0.021
0.004
0.009
-0.0007
-0.00005
-0.0002
-0.0005
5
0.068
-0.488
0.211
0.020
0.082
0.039
0.097
0.024
0.005
0.010
-0.0008
-0.00006
-0.0002
-0.0005
6 -0.049 -0.009 -0.002 -0.005
0.002
0.005
0.168
0.128
-0.741
0.152
0.150
-0.002
-0.009
0.021
0.004
0.009
0.027
0.002
0.008
0.019
7 0.048 0.009
0.001
0.004
0.114
0.087
1.234
-0.969
-0.427
-0.238
0.094
0.014
0.003
0.006
0.019
0.001
0.006
0.013
8 0.032 0.006
0.002
0.006
0.187
0.143
0.282
-0.127
-0.993
0.389
-0.051
0.023
0.004
0.010
0.030
0.002
0.009
0.021
9 0.053 0.010
0.080
0.061
0.162
-0.045
0.097
-0.528
0.099
0.010
0.002
0.004
0.013
0.001
0.004
0.009
10 0.023 0.004 0.0008 0.002
0.002
0.006
0.202
0.155
-0.102
0.002
0.113
0.052
-0.607
0.025
0.005
0.011
0.033
0.003
0.010
0.023
11 0.057 0.011
0.014
0.044
0.195
0.149
0.211
0.020
0.082
0.039
0.097
-1.725
-0.164
-0.182
0.358
0.028
0.108
0.245
12 0.404 0.076
0.011
0.035
0.153
0.117
0.165
0.015
0.064
0.030
0.076
-0.656
-1.001
0.039
0.280
0.022
0.085
0.192
13 0.316 0.059
0.131
0.025
0.004
0.014
0.063
0.049
0.068
0.006
0.027
0.013
0.031
-0.139
0.009
-0.543
0.116
0.009
0.035
0.080
14
0.003
0.009
-0.001
-0.0008
0.089
0.008
0.034
0.016
0.041
0.114
0.022
0.049
-0.702
0.039
0.140
0.037
15 0.086 0.016
0.002
0.005
-0.0006
-0.0004
0.046
0.004
0.018
0.008
0.021
0.059
0.011
0.025
-0.021
-0.617
0.085
0.302
16 0.045 0.008
0.002
0.007
-0.0008
-0.0006
0.063
0.006
0.024
0.011
0.029
0.081
0.015
0.035
0.184
0.039
-0.617
0.049
17 0.061 0.011
0.002
0.007
-0.0008
-0.0006
0.068
0.006
0.027
0.012
0.031
0.088
0.017
0.038
0.157
-0.0004
-0.041
-0.490
18 0.066 0.012
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6) fresh fruits, (7) pork, (8) beef and mutton, (9) poultry,
(10) eggs, (11) aquatic products, (12) fresh milk, (13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
C indicates conditional expenditure elasticities and U means unconditional expenditure elasticities.
136
Food a
Cb
UF b
UB b
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
1.068
1.026
0.858
0.768
0.978
1.033
0.995
1.062
1.054
0.810
1.128
1.055
0.963
0.958
1.170
0.961
0.727
0.947
1.138
1.093
0.914
0.818
0.872
0.921
1.093
1.167
1.158
0.890
1.240
1.000
0.912
0.908
1.030
0.845
0.640
0.834
0.748
0.719
0.601
0.538
0.573
0.605
0.719
0.767
0.762
0.585
0.815
0.657
0.600
0.597
0.677
0.556
0.421
0.548
Marshallian
1
2
3
4
5
6
7
8
Price
9
a
of
10
137
11
12
13
14
15
16
17
18
0.029
0.089
-0.164
-0.119
-0.056
-0.006
-0.017
-0.029
-0.014
-0.001
0.0006
0.023
-0.025
-0.004
-0.016
-0.032
1 -0.862 0.064
0.081
-0.158
-0.114
-0.054
-0.006
-0.016
-0.027
-0.013
-0.001
0.0005
0.022
-0.024
-0.004
-0.015
-0.031
2 0.205 -0.954 0.014
0.013
-0.925
0.053
-0.132
-0.095
-0.045
-0.005
-0.014
-0.023
-0.011 -0.0009
0.0004
0.018
-0.020
-0.003
-0.013
-0.026
3 0.313
0.046
-0.006
-0.898
-0.118
-0.085
-0.040
-0.004
-0.012
-0.021
-0.010 -0.0008
0.0004
0.016
-0.018
-0.003
-0.011
-0.023
4 0.370
0.003
0.002
0.004
-0.008
0.009
-0.006
-0.0008
0.011
0.001
-0.0004 -0.007
-0.010
5 -0.062 -0.017 -0.004 -0.013 -0.832 0.056
0.002
0.004
-0.008
0.010
-0.007
-0.0008
0.012
0.001
-0.0004 -0.007
-0.011
6 -0.065 -0.018 -0.004 -0.013 0.085 -0.905 0.003
-0.004
-0.0008 -0.034
-0.005
-0.018
-0.038
7 -0.024 -0.008 -0.001 -0.005 -0.028 -0.022 -0.924 0.011 -0.006 0.074 -0.038 -0.020
0.005
-0.021
-0.004
-0.0008 -0.036
-0.005
-0.019
-0.041
8 -0.026 -0.009 -0.002 -0.006 -0.030 -0.024 0.061 -1.004 -0.007 0.002
-0.021
-0.004
-0.0008 -0.036
-0.005
-0.019
-0.041
9 -0.026 -0.009 -0.002 -0.006 -0.030 -0.023 -0.024 0.002 -0.907 -0.016 0.008
0.001
0.027
-0.914
-0.028
-0.016
-0.003
-0.0006 -0.028
-0.004
-0.015
-0.031
10 -0.020 -0.007 -0.001 -0.004 -0.023 -0.018 0.195
-0.004
-0.0009 -0.039
-0.006
-0.020
-0.044
11 -0.027 -0.009 -0.002 -0.006 -0.032 -0.025 -0.111 0.003 -0.017 -0.050 -0.828 -0.023
0.010
0.003
0.009
0.026
0.017
-0.094
-0.012
-0.031
-0.039
-0.031
-0.016
-0.023
0.107
0.013
0.020
0.065
-1.066
12 0.047
0.009
0.003
0.008
0.024
0.016
-0.086
-0.011
-0.029
-0.036
-0.029
-0.041
-0.993
0.026
0.098
0.012
0.019
0.059
13 0.042
0.009
0.003
0.008
0.024
0.016
-0.085
-0.011
-0.029
-0.036
-0.028
-0.040
0.008
-0.971
0.097
0.012
0.018
0.059
14 0.042
0.020
0.005
0.046
-0.786
0.018
0.088
-0.116
15 -0.029 -0.009 -0.002 -0.006 -0.023 -0.018 -0.099 -0.012 -0.033 -0.041 -0.033
0.017
0.004
0.038
0.078
-0.895
0.053
0.111
16 -0.024 -0.008 -0.001 -0.005 -0.019 -0.015 -0.081 -0.010 -0.027 -0.034 -0.027
-0.018
-0.006
-0.001
-0.004
-0.014
-0.011
-0.061
-0.008
-0.021
-0.025
-0.021
0.013
0.003
0.029
0.042
0.016
-0.706
0.153
17
0.016
0.004
0.037
-0.020
0.014
-0.019
-0.619
18 -0.024 -0.007 -0.001 -0.005 -0.019 -0.015 -0.080 -0.010 -0.027 -0.033 -0.027
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6) fresh fruits, (7) pork, (8) beef and mutton, (9) poultry,
(10) eggs, (11) aquatic products, (12) fresh milk, (13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
C means conditional expenditure elasticities and UF and UB indicate unconditional expenditure elasticities with respect to food expenditure and total living
expenditure, respectively.
Table 5.23: Unconditional Elasticities for 18 Food Items, the Censored QAIDS with Demographic Variables, 1998.
(Continued)
137
Table 5.23: Continued
Hicksian
a
Price
9
of a
10
138
Food
1
2
3
4
5
6
7
8
11
12
13
14
15
16
17
18
0.035
0.108
0.019
0.014
0.154
0.023
0.057
0.048
0.066
0.031
0.007
0.037
0.038
0.005
0.010
0.028
1 -0.769 0.089
0.100
0.018
0.013
0.148
0.022
0.055
0.046
0.064
0.030
0.007
0.036
0.037
0.005
0.010
0.027
2 0.294 -0.929 0.020
0.015
0.011
0.124
0.018
0.046
0.039
0.053
0.025
0.006
0.030
0.031
0.004
0.008
0.022
3 0.388 0.033 -0.921 0.068
0.014
0.010
0.111
0.016
0.041
0.035
0.048
0.022
0.005
0.027
0.028
0.003
0.007
0.020
4 0.436 0.064 -0.002 -0.884
-0.692
0.157
0.164
0.024
0.061
0.051
0.071
0.018
0.004
0.022
0.050
0.006
0.013
0.036
5 0.009 0.002 0.0006 0.002
0.233
-0.798
0.173
0.025
0.064
0.054
0.075
0.019
0.004
0.023
0.052
0.007
0.014
0.038
6 0.010 0.002 0.0006 0.002
0.004
0.013
0.147
0.105
-0.722
0.038
0.065
0.148
0.039
0.011
0.002
0.013
0.027
0.003
0.007
0.019
7 0.065 0.016
0.004
0.014
0.157
0.112
0.277
-0.975
0.069
0.081
0.087
0.011
0.003
0.014
0.028
0.004
0.008
0.020
8 0.070 0.017
0.004
0.014
0.156
0.111
0.190
0.031
-0.832
0.063
0.089
0.011
0.003
0.014
0.028
0.004
0.007
0.020
9 0.069 0.017
0.053
0.013
0.003
0.011
0.120
0.086
0.360
0.024
0.084
-0.854
0.034
0.009
0.002
0.010
0.022
0.003
0.006
0.016
10
0.005
0.015
0.167
0.119
0.118
0.034
0.064
0.033
-0.741
0.012
0.003
0.015
0.030
0.004
0.008
0.022
11 0.074 0.018
0.008
0.025
0.187
0.133
0.091
0.013
0.033
0.028
0.039
-0.010
-0.011
0.163
0.020
0.043
0.117
-1.038
12 0.128 0.032
0.008
0.023
0.171
0.122
0.083
0.012
0.031
0.026
0.036
-0.016
-0.988
0.038
0.149
0.018
0.039
0.107
13 0.117 0.029
0.116
0.029
0.008
0.023
0.170
0.121
0.082
0.012
0.030
0.026
0.035
-0.014
0.013
-0.960
0.148
0.018
0.039
0.106
14
0.004
0.011
0.142
0.102
0.091
0.013
0.034
0.029
0.039
0.049
0.011
0.059
-0.729
0.025
0.112
-0.062
15 0.055 0.014
0.003
0.009
0.117
0.083
0.075
0.011
0.028
0.023
0.032
0.040
0.009
0.048
0.125
-0.889
0.072
0.155
16 0.045 0.011
0.002
0.007
0.088
0.063
0.057
0.008
0.021
0.018
0.024
0.030
0.007
0.037
0.078
0.020
-0.692
0.187
17 0.034 0.009
0.003
0.009
0.115
0.082
0.074
0.011
0.027
0.023
0.032
0.040
0.009
0.048
0.026
0.020
-0.0004
-0.575
18 0.045 0.011
a
The 18 basic food items are: (1) rice, (2) flour, (3) coarse grains, (4) potatoes, (5) fresh vegetables, (6) fresh fruits, (7) pork, (8) beef and mutton, (9) poultry,
(10) eggs, (11) aquatic products, (12) fresh milk, (13) yogurt, (14) bean and its products, (15) fats and oils, (16) sugar, (17) nuts, and (18) cakes.
b
C indicates conditional expenditure elasticities and U means unconditional expenditure elasticities.
138
CHAPTER 6
SUMMARY AND CONCLUSIONS
In this chapter, a summary is provided restating the objectives and methodologies
used in this study as well as the empirical results and their implications. Limitations and
future potential research topics are presented in the last section.
6.1 Objectives and Methodology
China, since its economic reform in 1978, has changed significantly as it makes
its transition from a centrally-planned to a consumer-oriented economy. The dramatic
shift in the economic structure has gradually increased household income and changed
consumption patterns in urban China. This study attempts to provide a better
understanding of heterogeneous consumer patterns in urban China by developing a multistage censored demand system using household data.
Specifically, this study consists of two major tasks: (1) to develop an economic
model considering heterogeneous consumption patterns across households and
commodity groupings by constructing a multi-stage demand system; and (2) to estimate
an econometric model of a Quadratic Almost Ideal Demand System (QAIDS) dealing
with zero consumption problems using household data. Three methodologies are
integrated in this study including constructing a multi-stage demand system,
incorporating demographic variables using the ‘ordinary budget share scaling and
translation’ (OBSSAT), and employing a two-step estimator to deal with zero
139
consumption problems. Therefore, with these advancements, this study makes a
contribution of improving elasticity estimations and is useful for policy making and
forecasting future food demand in urban China.
6.2 Empirical Results and Implications
This study analyzes data from three provinces, Shandong, Jiangsu, and
Guangdong in urban China and uses household data (2,000 observations each year) from
1993 to 1998 provided by the National Bureau of Statistics in China. On the basis of the
Chinese food guide pyramid, a three-level utility tree is constructed that divides 18 food
items into five food subgroups.
An empirical analysis is conducted by estimating econometric models in a
sequence of six steps: 1. Engel curve analysis, 2. the QAIDS, 3. the QAIDS with
demographic variables, 4. the censored QAIDS with demographic variables, 5. the
second-stage demand system, and 6. calculation of unconditional elasticities. This
empirical analysis allows us to examine the impact of the effects of each factor, such as
income, prices, demographic variables, and zero consumption on the demand system.
The results show the uniqueness of this study in three dimensions. First, using the
OBSSAT helps us to answer the problem of “how to break down the heterogeneous
consumption patterns in urban China?” In addition, our findings also show that China
should be treated as several markets instead of one. Second, the QAIDS has not been
applied to the study of food demand in urban China. Our results show that the QAIDS is
superior to the AIDS; however, the degree of importance for the quadratic term decreases
as other effects such as demographic and censoring effects are considered in a demand
system. Finally, 18 food items are broken down into five food subgroups and are
140
estimated using a multi-stage censored QAIDS. Including this large food bundle in a
demand system provides detailed information of the relationship among food items. For
example, the demand for such detailed foods as yogurt, nuts, and bean products, included
in this study, has never been investigated previously.
6.3 Limitations and Future Research
This analysis is based on the dataset from three provinces in urban China and
attempts to provide an indicator for foreseeing future consumption patterns in urban
China. However, not all of the urban areas in China, especially the inner provinces, are as
prosperous as the three provinces in this study. Whether or not the economy in these
inner urban areas will grow quickly enough to catch up with the food consumption
patterns as indicated in the three provinces in this study is another issue. How that growth
in economy will affect food consumption in China is still uncertain at the present time.
Certain methodological issues are not considered in this study. For example, the
multi-stage demand system is not fully estimated since there are no price indices
available to estimate a demand system for broad groups. A remedy of estimating a single
equation is used to calculate income elasticities, which are useful for policy makers.
In addition, the methodology for deriving total income effects needs to be
developed in order to enhance the value of this study. Price variation reflects both quality
and quantity. How to decompose price or how demographic factors impact prices faced
by consumers are not addressed in this study. Price is another important issue needing in
depth study in the near future.
As for model selection, a comparison of models would be important since
different models produce different elasticities which can be used to predict future demand
141
and is critical for guiding domestic production as well as international trade. Predicting
performance is another criterion to determine the goodness of fit of the selected models.
Comparison of the predicting performance criterion has been done by Liu and Chern
(2001a) using pooled aggregate provincial data. It would be interesting to compare the
performance of demand systems using household data with aggregate provincial data. In
addition, policy makers would find it interesting to compare the forecasting of food
demand with others such as Brown (1994).
There are other general procedures available in the literature to incorporate
demographic variables into a demand system for a comparison of methodologies. A
comparison of ‘ordinary budget share scaling and translation’ with other commonly-used
procedures, e.g. five procedures in Pollak and Wales (1981) or Bollino et al. (2000)
would provide some empirical evidence for selection of procedures to incorporate
demographic variables.
This study did not investigate the issue of structural changes; however, studying
whether structural changes occurred in food consumption patterns between 1993-1998
may provide valuable information to help explain the transition of the economy in China.
142
BIBLIOGRAPHY
Alessie, R. and A. Kapteyn, “Habit Formation, Interdependent Preferences and
Demographic Effects in the Almost Ideal Demand System,” The Economic Journal, Vol.
101, Issue 406, May, 1991: 404-419.
Amemiya, T., Advanced Econometrics, Harvard University Press, 1985.
Banks, J., R. Blundell, and A. Lewbel, “Quadratic Engel Curves and Consumer
Demand,” The Review of Economics and Statistics, Vol. 79, No. 4, November, 1997:
527-539.
Barten, A.P., “Family Composition, Prices and Expenditure Patterns,” in Econometric
Analysis for National Planning, Hart, P.E., G. Mills and J.K. Whittaker eds., Butterworth:
London, 1964.
Barten, A.P., “Consumer Allocation Models: Choice of Functional Form,” Empirical
Economics, Vol. 18, 1993: 129-158.
Bhandari, R. and F.J. Smith, “Education and Food Consumption Patterns in China:
Household Analysis and Policy Implications,” Journal of Nutrition Education, Vol. 32,
No. 4, July/August, 2000: 214-224.
Blundell R. and J.M. Robin, “Latent Separability: Grouping Goods without Weak
Separability,” Econometrica, Vol. 68, January, 2000: 53-84.
Bollino, C.A., F. Perali, and N. Rossi, “Linear Household Technologies,” Journal of
Applied Econometrics, Vol. 15, 2000: 275-287.
Börsch-Supan, A. and V. Hajivassiliou, “Smooth Unbiased Multivariate Probability
Simulators for Maximum Likelihood Estimation of Limited Dependent Variable
Models,” Journal of Econometrics, Vol. 58, 1993: 347-368.
Brown, L.R., “Who Will Feed China?” World Watch, September/October, 1994.
Buse, A., “Evaluating the Linearized Almost Ideal Demand System,” American Journal
of Agricultural Economics, Vol. 76, November, 1994: 781-793.
143
Carpentier, A. and H. Guyomard, “Unconditional Elasticities in Two-Stage Demand
Systems: An Approximate Solution,” American Journal of Agricultural Economics, Vol.
83, No. 1, February, 2001: 222-229.
Chen, J., “Food Consumption and Projection of Agricultural Demand/Supply Balance for
1996-2005 in China,” Unpublished Master Thesis, Department of Agricultural,
Environmental, and Development Economics, The Ohio State University, 1996.
Chen, P.Y. and M.M. Veeman, “ An Almost Ideal Demand System Analysis for Meats
with Habit Formation and Structural Change,” Canadian Journal of Agricultural
Economics, Vol. 39, 1991: 223-235.
Chern, W.S., Estimated Elasticities of Chinese Grain Demand: Review, Assessment and
New Evidence, a report submitted to the World Bank, January 1997.
Chern, W.S., “Assessment of Demand-Side Factors Affecting Global Food Security,” in
Food Security in Asia, Edited by W.S. Chern, C.A. Carter, and S.-Y. Shei. Chapter 6, pp.
83-118, 2000.
Chern, W.S. and G. Wang, “The Engel function and Complete Food Demand system for
Chinese Urban Households,” China Economic Review, Vol.4, spring 1994, 35-57.
China Statistical Yearbook, 2001, National Bureau of Statistics, China Statistics Press,
2001.
Chung, C.-F., “Modelling Demand Systems with Demographic Effects Based on the
Modifying Function Approach,” Economics Letters, Vol. 73, 2001: 269-274.
Coulibaly, N. and B.W. Brorsen, “Monte Carlo Sampling Approach to Testing Nonnested
Hypotheses: Monte Carlo Results,” Econometric Reviews, Vol. 18, No. 2, 1999: 195-209.
Cox, D.R., “Further Results on Tests of Separate Families of Hypotheses,” Journal of
Royal Statistical Society, B24, 1962: 406-424.
Dameus, A., B.W. Brorsen, K.P. Sukhdial, and F.G.-C. Richter, “AIDS versus Rotterdam:
A Cox Nonnested Test with Parametric Bootstrap,” AAEA Annual Meeting, 2001.
Darrough, M.N., R.A. Pollak and T.J. Wales “Dynamic and Stochastic Structure: An
Analysis of Three Time Sereis of Household Budget Studies” The Review of Economics
and Statistics, Vol. 65, Issue 2, May, 1983: 274-281.
Davidson, R. and J.G. MacKinnon, “Several Tests for Model Specification in the
Presence of Alternative Hypotheses,” Econometrica, Vol. 49, Issue 3, May, 1981: 781793.
144
Davidson, R. and J.G. MacKinnon, “Some Non-nested Hypothesis Tests and the
Relations among them,” The Review of Economic Studies, Vol. 49, Issue4, October,
1982: 551-565.
Deaton, A., “Specification and Testing in Applied Demand Analysis,” The Economic
Journal, Vol. 88, 1978: 524-536.
Deaton, A. and J. Muellbauer, Economics and Consumer Behavior, Cambridge:
Cambridge University Press, 1980a.
Deaton, A. and J. Muellbauer, “An Almost Ideal Demand System,” American Economic
Review, Vol. 70, 1980b: 312-326.
Dong, D. and B.W. Gould, “Estimation of Censored Demand Systems: An Overview,”
Working paper at Department of Agricultural and Applied Economics, University of
Wisconsin-Madison, May 1999.
Edgerton, D.L., “Weak Separability and the Estimation of Elsticities in Multistage
Demand System,” American Journal of Agricultural Economics, Vol. 79, February, 1997:
62-79.
Fahs, R., S. Cardell, and R. Mittelhammer, “Semiparametric Estimation and Inference in
a System of Censored Demand Equation,” AAEA annual meeting, 2001.
Fan, J.X. and W.S. Chern, “Analysis of Food Consumption Patterns in China:
Nonparametric and Parametric Approaches,” Journal of Family and Economic Issues,
Vol. 18, 1997: 113-125.
Fisher, D., A.R. Fleissig and A. Serletis, “An Empirical Comparison of Flexible Demand
System Functional Forms,” Journal of Applied Econometrics, Vol. 16, 2001: 59-80.
Gao, X.M., E.J. Wailes, and G.L. Cramer, “Partial Rationing and Chinese Urban
Households Food Demand analysis,” Journal of Comparative Economics, Vol.22, 1996:
43-62.
Geweke, J., “Efficient Simulation from the Multivariate Normal and Student-t
Distribution Subject to Linear Constraints,” in Computer Science and Statistics:
Proceedings of the Twenty-Third Symposium on the Inference, American Statistical
Association, Alexandria, 1991: 571-578.
Goldman, S.M. and H. Uzawa, “A Note on Separability in Demand Analysis,”
Econometrica, Vol. 32, Issue 3, July, 1964: 387-398.
Gorman, W.M., “Separability Utility and Aggregation,” Econometrica, Vol. 27, Issue 3,
July, 1959: 459-481.
145
Gorman, W.M., “The Structure of Utility Functions,” Review of Economic Studies, Vol.
35, Issue 4, October, 1968: 367-390.
Gorman, W.M., “Tricks with Utility Functions,” in Essays in Economic Analysis, M.
Artis and R. Nobay eds., Cambridge: Cambridge University Press, 1976.
Gould, B.W., and R. Sabates, “The Structure of Food Demand in Urban China: A
Demand System Approach,” AAEA annual meeting, 2001.
Hahn, W., “Elasticities in AIDS Models: Comments,” American Journal of Agricultural
Economics, Vol. 76, November 1994: 972-977.
Han, T. and T.I. Wahl, “China’s Rural Household Demand for Fruit and Vegetables,” in
Food Markets in China: New Looks and Deeper Understandings, Proceedings of WCC101, Honolulu, Hawaii, January 12-13, 1998.
Hasan, M., R. Fahs, R.C. Mittelhammer, and T. Wahl, “Estimation of Systems of
Censored Limited Dependent Variables Models Using Simulated ML and GMM
Estimation Methods,” AAEA annual meeting, 2001.
Heckman, J.J., “The Common Structure of Statistical Models of Truncation, Sample
Selection and Limited Dependent Variables and a Simple Estimator for Such Models,”
Annals of Economic and Social Measurement, 5, 1976: 475-492.
Heien, D. and C.R. Wessells, “Demand Systems Estimation with Microdata: A Censored
Regression Approach,” Journal of Business and Economic Statistics, Vol. 8, 1990: 365371.
Hicks, J.R., Value and Capital, Oxford University Press, 1936.
Holt, M.T. and B.K. Goodwin, “Generalized Habit Formation in an Inverse Almost Ideal
Demand System: An Application to Meat Expenditures in the U.S.,” Empirical
Economics, Vol. 22, 1997: 293-320.
Kao, C. and L.-F. Lee, “Maximum Simulated Likelihood Estimation of Consumer
Demand Systems with Binding Non-Negativity Constraints,” Working Paper, March
1996.
Keane, M.P., “A Computationally Practical Simulation Estimator for Panel Data,”
Econometrica, Vol. 62, 1994: 95-116.
Keller, W.J. and J. van Driel, “Differential Consumer Demand Systems,” European
Economic Review, Vol. 27, 1985: 375-390.
146
Lee, L.-F. and M.M. Pitt, “Microeconometric Demand Systems with Binding
Nonnegativity Constraints: the Dual Approach,” Econometrica, Vol. 54, 1986: 12371342.
Leontief, W., “Composite Commodities and the Problem of Index Numbers,”
Econometrica, Vol. 4, 1936: 39-59.
Lewbel, A., “A Unified Approach to Incorporating Demographic or Other Effects into
Demand Systems,” Review of Economic Studies, Vol. 52, 1985: 1-18.
Lewbel, A., “Aggregation without Separability: A Generalized Composite Commodity
Theorem,” The American Economic Review, Vol. 86, Issue 3, June, 1996: 524-543.
Lewbel, A., “Identification and Estimation of Equivalence Scales under Weak
Separability,” Review of Economic Studies, Vol. 56, 1989: 311-316.
Lewis, P. and N. Andrews, “Household Demand in China,” Applied Economics, Vol. 21,
1989: 793-807.
Liu, K.E. and W. S. Chern, “Food Consumption in Urban and Rural China: Quantity vs.
Quality,” Referred Poster at 45th Annual Conference of the ACCI, Consumer Interests
Annual, Vol. 45, 1999: 132.
Liu, K.E. and W.S. Chern, “Food Demand in Urban China and its Implications for
Agricultural Trade,” Paper for presentation at the 2001 meeting of WCC-101,
Westerbeke Ranch, Sonoma, CA, April 8-10, 2001a.
Liu, K.E. and W.S. Chern, “Effects of Model Specification and Demographic Variables
on Food Consumption: Microdata Evidence from Jiangsu, China,” Paper for presentation
at the 11th Annual World Food and Agribusiness Forum, World Food and Agribusiness
Symposium of the International Food and Agribusiness Management Association,
Sydney, Australia, June 27-28, 2001b.
Liu, K. E. and W. S. Chern, “Estimation of Food Demand Systems: Evidence from Micro
Household Data in Urban China,” Agricultural and Resource Economics Review,
Selected Abstracts of the NAREA Meeting, Vol. 30, October, 2001c: 216-217.
Magnus, J.R. and A.D. Woodland, “On the Maximum Likelihood Estimation of
Multivariate Regression Models Containing Serially Correlated Error Components,”
International Economic Review, Vol. 29, Issue 4, November, 1988: 707-725.
Mathews, B.P. and A. Diamantopoulos, “Towards a Taxonomy of Forecast Error
Measures,” Journal of Forecasting, Vol. 13, 1994: 409-416.
147
Moschini, G., “A Non-Nested Test of Separability for Flexible Functional Forms,” The
Review of Economics and Statistics, Vol. 74, 1992: 365-369.
Moschini, G., “Unit of Measurement and the Stone Index in Demand System
Estimation,” American Journal of Agricultural Economics, Vol. 77, February 1995: 6368.
Moschini, G., D. Moro, and R.D. Green, “Maintaining and Testing Separability in
Demand Systems,” American Journal of Agricultural Economics, Vol. 76, February,
1994: 61-73.
Muellbauer, J., “Testing the Barten Model of Household Composition Effects and the
Cost of Children,” Economic Journal, Vol. 87, Issue 347, September, 1977: 460-487.
Naik, N.Y. and M.J. Moore, “Habit Formation and Intertemporal Substitution in
Individual Food Consumption,” The Review of Economics and Statistics, Vol. 78, Issue 2,
May, 1996: 321-328.
Nayga, R.M., Jr. and O. Capps, Jr., “Tests of Weak separability in Disaggregated Meat
Products,” American Journal of Agricultural Economics, Vol. 76, November, 1994: 800808.
Parks, R.W., “Systems of Demand Equations: An Empirical Comparison of Alternative
Functional Forms,” Econometrica, Vol. 37, 1969: 629-650.
Perali, F. and J.-P. Chavas, “Estimation of Censored Demand Equations from Large
Cross-Section Data,” American Journal of Agricultural Economics, Vol. 82, November,
2000: 1022-1037.
Pindyck, R.S. and D.L. Rubinfeld, Econometric Models and Economic Forecasts, Irwin
McGraw-Hill, 4th ed. Chapters 8, 13, and 14, pp. 379-462, 1998.
Pollak, R.A., “Habit Formation and Dynamic Demand Functions,” The Journal of
Political Economy, Vol. 78, Issue 4, July/August, 1970: 745-763.
Pollak, R.A. and T.J. Wales, “Estimation of Complete Demand Systems from Household
Budget Data: The Linear and Quadratic Expenditure Systems,” American Economic
Review, Vol. 68, 1978: 348-359.
Pollak R.A. and T.J. Wales, “Comparison of the Quadratic Expenditure System and
Translog Demand Systems with Alternative Specifications of Demographic Effects,”
Econometrica, Vol. 48, Issue 3, April, 1980: 595-612.
Pollak R.A. and T.J. Wales, “Demographic Variables in Demand Analysis,”
Econometrica, Vol. 49, Issue 6, November, 1981: 1533-1551.
148
Pollak, R.A., and T.J. Wales, “The Likelihood Dominance Criterion– A New Approach
to Model Selection,” Journal of Econometrics, Vol. 47, 1991: 227-242.
Pollak, R.A. and T.J. Wales, Demand System Specification and Estimation, Oxford
University Press, 1992.
Ray, R “A Dynamic Generalisation of the Almost Ideal Demand System,” Economics
Letters, Vol. 14, 1984: 235-239.
Segerson, K. and T.D. Mount, “A Non-Homothetic Two-Stage Decisio Model Using
AIDS,” Review of Economics and Statistics, Vol. 67, November, 1985: 630-639.
Shonkwiler, J.S. and S.T. Yen, “Two-step Estimation of a Censored System of
Equations,” American Journal of Agricultural Economics, Vol. 81, November, 1999:
972-982.
Strotz, R.H., “The Empirical Implications of a Utility Tree,” Econometrica, Vol. 25, Issue
2, April, 1957: 269-280.
Strotz, R.H., “The Utility Tree– A Correction and Further Appraisal,” Econometrica, Vol.
27, Issue 3, July, 1959: 482-488.
Theil, H. and R.H. Mnookin, “The Information Value of Demand Equations and
Predictions,” The Journal of Political Economy, Vol. 74, Issue 1, February, 1966: 34-45.
Theil, H., Economics and Information Theory, Chicago: Rand McNally and Company,
Chapters 6 and 7, 1967.
Tobin, J. “Estimation of Relationships for Limited Dependent Variables,” Econometrica,
Vol. 26, January, 1958: 24-36.
Wales, T.J. and A.D. Woodland, “Estimation of Consumer Demand Systems with
Binding Non-Negativity Constraints,” Journal of Econometrics, Vol. 21, 1983: 263-285.
Wang, Z. and W.S. Chern, “Effects of Rationing on the Consumption Behavior of
Chinese Urban Households during 1981-1987,” Journal of Comparative Economics, Vol.
16, 1992: 1-26.
World Bank, At China’s Table: Food Security Options, Washington, DC, 1997.
Wu, Y., E. Li and S.N. Samuel, “Food Consumption in Urban China: An Empirical
Analysis,” Applied Economics, Vol. 27, 1995: 509-515.
149
Yen, S.T. and K. Kan, “Household Demand for Fats and Oils: Two-step Estimation of a
Censored Demand System,” Working Paper, August 2000.
Yen, S.T. and T.L. Roe, “Estimation of a Two-Level Demand System with Limited
Dependent Variables,” American Journal of Agricultural Economics, Vol. 71, 1989: 8598.
Zhao, F. and T. Wahl, “Changing Animal Protein Food Demand in Urban China,” AAEA
annual meeting, 1999.
150

Similar documents