Egyptian import demand for United States wheat

MISSING IMAGE

Material Information

Title:
Egyptian import demand for United States wheat
Physical Description:
v, 84 leaves : ill. ; 29 cm.
Language:
English
Creator:
Pana-Cryan, Regina, 1961-
Publication Date:

Subjects

Subjects / Keywords:
Wheat trade -- Egypt   ( lcsh )
Demand (Economic theory)   ( lcsh )
Imports -- Egypt   ( lcsh )
Food and Resource Economics thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Food and Resource Economics -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 80-83).
Statement of Responsibility:
by Regina Pana-Cryan.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 026435541
oclc - 36686927
System ID:
AA00017664:00001


This item is only available as the following downloads:


Full Text











EGYPTIAN IMPORT DEMAND FOR UNITED STATES WHEAT


By

REGINA PANA-CRYAN
























A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1996















TABLE OF CONTENTS

ABSTRACT .................................................iv

CHAPTERS

1 INTRODUCTION....... .................................. 1

Background Information on Egypt....................1
The Land and the People.......................... 1
Political and Economic History..................6
Agricultural Policy Objectives and Instruments..... 8
Food Subsidies.................................. 8
The Ration System...............................10
Previous Studies on Egypt.........................12

2 PROBLEM STATEMENT AND OBJECTIVES..................15

The Current Situation in the Wheat Import Market
and the US Role ..................................15
Wheat and Wheat Flour..........................15
Problem Statement ................................20
Objectives .........................................22

3 THEORETICAL AND EMPIRICAL CONSIDERATIONS...........24

Demand Analysis in Theory and Practice............24
Exogenous Prices..................................26
Aggregation and Separability.....................36

4 METHODOLOGY......................................39

Approaches to Demand Estimation and General
Restrictions....................................39
A System-wide Differential Approach to Import
Demand........................... ................ 45
Preference Structure and Conditional Demand.......47
Preference Independence........................47
Block Independence.............................48
Blockwise Dependence...........................49
Rotterdam Model .................. ............. 50
A Geographical Import Allocation Model.............53

5 DATA, ESTIMATION AND RESULTS......................59

Data......... ....... .............................. 59
Testing Restrictions........................... 61
ii









Conditional Marginal Shares....................... 63
Expenditure Elasticities.......................... 65
Price Parameters........................... ........... 65
Price Elasticities................................ 66
Summary of AP Results............................ 68

6 SUMMARY, CONCLUSIONS, LIMITATIONS AND SUGGESTIONS FOR
FURTHER RESEARCH............................... 72

Summary and Conclusions........................... 72
Limitations and Suggestions for further Research..74

APPENDIX ROTTERDAM AP RESULTS ..... ..........................76

REFERENCES................................................ 80

BIOGRAPHICAL SKETCH........................................84


iii















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


EGYPTIAN IMPORT DEMAND FOR UNITED STATES WHEAT

By

REGINA PANA-CRYAN

December 1996



Chairman: Dr. Robert D. Emerson
Major Department: Food and Resource Economics



Wheat is Egypt's leading food staple and the country

faces water and land constraints. Most of the domestically

produced wheat is home-consumed and does not reach the market.

Egypt has been the third largest wheat importer in the world

and wheat imports are increasing.

The Egyptian import wheat market is important for US

wheat producers and exporters. Potential exists for increased

wheat exports to Egypt that could benefit not only US wheat

exporters but also the whole US agricultural sector, given the

importance of wheat. In addition, many developing countries

face situations similar to Egypt's; therefore, a model

developed for Egypt could be adjusted and used for other

countries as well.








In this study a geographical allocation model is

developed for Egyptian imported wheat, in order to assess the

US versus its competitors' position in this market.

The study concludes that the US is expected to expand its

exports of wheat to Egypt if more expenditure is allocated for

this good. Wheat imports from different countries seem to be

uniform substitutes. The estimated model predicts what

percentage change in quantities imported to Egypt from

different sources will result from a percentage change in the

wheat offering price by an individual supplier, assuming that

all other suppliers keep their offering prices constant.















CHAPTER 1

INTRODUCTION

Egypt was the largest import market for US wheat in 1995

(United States Department of Agriculture (USDA), 1996). This

dissertation explores the allocation of import demands for

wheat by Egypt to determine the potential for a continuing

demand for US wheat in this market. Since the import demand

must be considered within the national setting, summary

information about Egyptian agriculture, recent political and

economic history, and agricultural policies focusing on food

subsidies and the ration system and how these apply to wheat

are discussed in this chapter before proceeding to the

analytical and empirical models. In the last section of the

chapter a brief literature review is presented, that discusses

studies on Egypt, which focus on economic development and

trade and how these are affected by existing domestic policies

and the policies of trading partners.

Background Information on Egypt

The Land and the People

Egypt covers an area of 386,000 square miles, of which

less than 4% is habitable and occupied by 98% of the total

population. Egypt is the only developing country that is

almost completely dependent on irrigation. The only water










source is the Nile, which makes expanding the arable land base

very difficult, especially considering the continuing loss of

agricultural lands to urbanization (Robinson and Gehlhar, pp.

1 and 3). In addition, the water supply from the High Aswan

Dam is governed by the existing sharing agreement with Sudan

(World Bank, 1993). It seems that attempts to reclaim desert

land were high-cost and resulted in low-productivity

operations, so that the land suitable for the production of

most crops has already been brought into production (Antle, p.

173). Hansen agrees on the failure of reclamation; he

maintains that in the 1960s about three-quarters of all

agricultural investment was spent to reclaim 0.9 million

feddan (1 feddan = 1.038 acre = 0.42 hectare) in the Western

Desert, resulting in less than two-thirds of this land ever

being cultivated at marginal returns (Hansen, p. 120). Since

total cropland has remained the same since 1960, increases in

agricultural production are due to more intensive cultivation

(Dethier, p. 3). Environmental problems arise from the

extensive use of chemicals, poor drainage, and intensive water

use. There are fewer than 3 harvested feddan per agricultural

laborer.

Excluding agriculture, there are few natural resources in

the country. In the early 1970s, oil and gas fields were

discovered, which now provide much of the foreign exchange

earnings for the country. Foreign exchange is earned through










worker remittances, energy, tourism, cotton and textiles, in

order of importance.

Population has more than doubled since the 1950s, and

now exceeds 50 million (Table 1). While population has been

increasing by more than 2% per year since 1960, Gross National

Product (GNP) per capital increased moderately until 1987, but

then declined until 1991 (Table 1).

Agriculture's share in total employment declined from 55%

in 1965 to 40% in 1992, while agriculture's share in Gross

Domestic Product (GDP) fell from 29% to 18% for the same

period (Table 1). Agricultural exports as a share of total

exports have declined from 71% in 1965 to 13% in 1992 (Table

2). The aggregate growth rate and productivity of the sector

have declined since the mid 1960s. The increase in production

from 1978 to 1983 was caused by increases in production of

vegetables, fruits, livestock and dairy (Dethier, pp. 16-17).

There have been substantial increases in wheat yields during

the 1980s following the adoption of high yielding varieties.

Major producing crops are maize, rice, wheat, sugar cane

and cotton. Although self-sufficient in the 1970s, Egypt was

importing more than one-half of its food requirements by the

mid-1980s. Major imports are wheat and wheat flour, corn,

vegetable oils, sugar, and meat. The total and agricultural

import and export values for recent years are presented in

Table 2.











Table 1. Population, Agriculture's Share in GDP, GDP per
Capita and Labor Force, Egypt, 1965-92.


Population/a
Total Rural
1000 % of
total


29389
33053
36289
40875
46511
52426
54842


Agricult.
Share
in GDP/b


29
29
29
18
20
18
18


GNP
per
capita/c
$US


172
202
320
500
670
610
660


Labor Force/d
Total Agricult.
1000 % of
total


8130
9172
10037
11240
12786
14514
15275


Note: c/ Calculated by the Atlas Method, by the World Bank.
Sources: a/ Food and Agricultural Organization of the United
Nations (FAO), AGROSTAT, b/ World Bank, World Tables,
c/ World Bank, STARS, d/ FAO, AGROSTAT.


Year



1965
1970
1975
1980
1985
1990
1992










Table 2.
1965-92.


Year


1965
1966
1967
1968
1969

1970
1971
1972
1973
1974

1975
1976
1977
1978
1979

1980
1981
1982
1983
1984

1985
1986
1987
1988
1989

1990
1991
1992


Total and Agricultural Imports and Exports, Egypt,


------ Imports
Total Agric.
$US 1000


9335
10705
7920
6660
6377

7702
9198
8988
9075
23518

39345
38082
48162
67279
38386

48615
88395
90806
102783
107690

99645
115055
119410
86573
74477

92020
78620
82930


2947
3243
3471
2519
2208

2140
3163
2897
3197
12060

14174
13765
15474
19971
16684

23503
36355
32167
33037
39415

37098
33427
26136
31960
31930

30878
25849
26244


Agric.
% of
total

32
30
44
38
35

28
34
32
35
51

36
36
32
30
43

48
41
35
32
37

37
29
22
37
43

34
33
32


------ Exports
Total Agric.
$US 1000


6012
5991
5578
6199
7447

7614
7893
8252
11163
15165

14022
15220
17083
17375
18403

30469
32322
31211
32157
31408

37153
29351
20372
21204
26478

25850
36590
30500


4282
4256
3998
4350
5107

5125
5575
5117
7162
9824

7821
7336
8228
6638
6058

6773
7405
6729
7264
7561

6616
6694
6725
5138
5325

4270
3910
4012


Source: FAO, AGROSTAT.


Agric.
% of
total

71
71
72
70
69

67
71
62
64
65

56
48
48
38
33

22
23
22
23
24

18
23
33
24
20

17
11
13










Political and Economic History'

In 1952 King Farouk was overthrown, and in 1954 the

military took over under Colonel Nasser. Some of the wide-

ranging reforms imposed were the redistribution of land, the

promotion of industrial development and the expansion of

social welfare services. In 1956 a new constitution was

proclaimed. During the 1950s, the government under President

Nasser nationalized private industry, closed the economy to

Western investment and created a huge public sector, with 70%

of the country's industrial output being provided by State

enterprises. Administered prices and an extensive food

subsidy system were established.

In 1970 President Nasser died and was succeeded by his

Vice-President Anwar Sadat, who strengthened relations with

Israel and introduced a more liberal political and economic

regime. President Sadat started liberalizing the economy,

establishing an "open door" policy in 1974 to encourage

foreign investment. Price controls were loosened, and private

farmers, as opposed to cooperatives, were favored. The pace

of price liberalization policies had to be slowed after riots

in 1977, caused mainly by the increase in price for subsidized

bread. In 1981 Sadat was assassinated by Islamic

fundamentalists and was succeeded by his Vice-President Hosni

Mubarak. While the liberalization of the economy continues,



1 The information for this section is from the Europa
World Yearbook.









7

foodstuffs and petroleum are still heavily subsidized and

considerably strain the country's finances.

During the planning period, between 1960 and 1965, the

economy grew at an average annual rate of about 6%. Between

1965 and 1973 it slowed to an average annual rate of 3%, due

to warfare with Israel. From 1973-80, oil prices soared, the

economy boomed and the demand for imported goods, especially

foodstuffs increased. The subsequent drop in oil prices

depressed foreign exchange earnings and remittances, and GDP

grew at an average 1.4% between 1986 and 1989. Food import

dependency increased to 50%, although agricultural production

almost kept pace with the population growth rate at 2.4%. By

the late 1980s the need for economic reform was apparent. In

1987 and 1988 Egypt's debt of $6.5 million to Western

developed countries was rescheduled. The existing multiple

exchange rate was partly unified, some public enterprises were

reformed, and producer prices were raised. In 1991 the

country reached an agreement with donors for $380 million aid

under comprehensive accelerated reforms. The commitment to

these reforms resulted in the Paris Club restructuring the

country's private commercial external debt, the US and Gulf

State creditors canceling $13 billion in debt, and a set of

Western donors establishing a special compensation fund to

cover the cost of structural adjustment and to absorb the

displaced workers from Iraq and Kuwait.










Agricultural Policy Objectives and Instruments

Egyptian agricultural policy has two stated broad

objectives: first, to provide an adequate supply of food

staples to all income groups, and second to achieve self-

sufficiency in strategic food crops. Lesser objectives aim at

increasing farm income, insulating producers from

international price fluctuations, and conserving foreign

exchange. The worsening balance of trade and the

liberalization of the economy seem to call for a redefinition

of these objectives, especially those of self-sufficiency and

producer protection.

Policy instruments include price policy, quota

deliveries, input subsidies, exchange rate management and

trade controls. Eight different government entities are

responsible for decision making or program implementation.

Tariff rates have been declining for most goods by 10% a

year (Foreign Agricultural Service of the United States (FAS),

p. 6) and the products requiring import authorization by the

government have been reduced.

Food Subsidies2

These subsidies alter prices of goods both
domestically and relative to traded goods; they figure
prominently in government expenditures; they are an
instrument in income redistribution; they affect the
volume of imports and exports of food and other goods;
they affect investment and economic growth; they affect
the country's balance of payments; and they leave their
mark on foreign policy. Their size and pervasiveness

2 The information for this section is from von Braun and
de Haen, unless otherwise indicated.








9

make them important to consider in any analysis of the
economy. At the same time, their ramifications are so
wide spread that it is difficult to disentangle their
effects. (Scobie 1983, p. 11).

Scobie describes the development of the system as

follows. Ration cards were introduced during World War II,

and slowly a system of rationing and subsidies on consumer

commodities developed. Expenditures on these subsidies were

estimated to be less than 2 Egyptian pounds (LE) per capital

(in constant 1975 prices deflated by the consumer price index)

until 1973, or about 1% of national income. They increased to

LE 6.2 in 1972-74, and by 1981 reached LE 20 (Scobie 1983, p.

12). During 1967-1973 the country's resources were devoted to

military expenditures. After the 1973 war, however, the open

door economic policy was implemented, which was supposed to

also result in less state involvement in domestic economic

activity. Instead, government expenditures on social programs

were increased. Wheat and flour subsidies were around 45-60%

of total food subsidies (Scobie 1983, pp. 12-13), which was a

significant portion of total consumer subsidies.

Policies related to food subsidies are area allotments,

procurements, direct and indirect price controls and input

pricing and allotment.

The producer prices for cereals, industrial crops, oil

seeds and fodder are fixed, and have been historically below

world prices. Inputs (fertilizers, pesticides, access to

mechanization, etc.) have been subsidized and water is free,

as was already mentioned in the introductory section.








10

Procurement consists of compulsory delivery at fixed

prices. There has been variation of the required percentages

of production through time, and there have been data

inconsistencies for some years for some crops. The

requirement for wheat had been between 15-20%, but dropped to

around 7% after 1977. Wheat procurement prices were similar

to import prices, when converted at official exchange rates.

After 1973-74 world prices increased, with procurement prices

increasing more slowly. In 1975 import prices started

declining, while procurement prices kept increasing. This led

to the eventual reduction of wheat procurement and increasing

imports. By 1994 all procurements were voluntary, including

that of cotton, for the first time (FAS, p. 10).

The Ration System3.

Wheat flour and bread are available at subsidized prices

to all consumers with no restrictions, while rice, tea,

cooking oil and sugar comprise the "basic ration." They are

rationed into monthly quotas and are sold at subsidized prices

through ration cards. There is allowance for an additional

ration at subsidized, but higher prices. Contrary to the

usual situation in other developing countries with similar

systems, the rural areas in Egypt are reasonably well served

by government controlled subsidized food distribution. More

than 90% of the households possess ration cards. Excluded are


SThe information about the ration system is from the
household survey conducted by Alderman and von Braun in 1981-
82, unless otherwise indicated.










households with more than 10 feddan, and those whose head is

working abroad. Over-reporting is in general not a problem.

Subsidized flour is available in specialized flour shops

and, to a lesser extent, in the government cooperative shops

that distribute ration card goods. Frequently, it is only

available at fixed days of the month, with regulations that

change according to location and over time. In 1981-82, more

than 75% of urban and about 25% of rural households had access

to subsidized bread without quantity restrictions, and one-

fourth of rural and one-fifth of urban households indicated

limitations in bread supply.

Alderman and von Braun argue that bread is not used for

animal feed, because even the cheapest bread is more expensive

than the farm gate price for corn, because it is considered

immoral to feed it to animals, and finally because it is not

available in bulk. Calculating percentages of waste (bread

used for backyard or "urban" agriculture, and bread thrown

away) they conclude it amounts to less than 4% of the amount

spent on wheat subsidies.

Urban consumers purchase a greater proportion of their

food than rural ones through outlets with fixed prices

(cooperatives, government flour shops, licensed bakers).

Ration distributions are rarely a family's only source for the

commodities. Therefore, rations at subsidized prices can be

considered income transfers. So, many households purchase

much beyond ration levels, and these quantities often exceed








12

the rationed ones (Alderman and von Braun, p. 24). More bread

is consumed in urban, and more flour in rural areas. In grain

equivalents, rural households purchase more, mostly unmilled

wheat.

Previous Studies on Egypt

Many recent studies on Egypt focus on political economy

issues. A thorough work in this area was conducted by Hansen

(1991), who compares Egypt's and Turkey's policies,

institutions and politics, in this way presenting a

comprehensive study on the contemporary political economy of

the two countries. Among the issues he examines are the lack

of growth in total factor productivity in both countries, the

different degrees of success in educating the people of each

country, the role of domestic politics on development and the

difference in agricultural growth in the two countries. With

respect to this last issue, he claims that Turkey's growth can

be attributed to the expansion of the cultivable land until

1960. In contrast, Egypt has faced a land constraint since

the beginning of the century. He concludes that Turkey's

inability to expand cultivable land after 1960 is the reason

for the sluggish agricultural growth observed since then.

Another thorough study by Dethier (1989) examines Egypt's

agricultural pricing policy from a political economy

perspective. His main conclusion is that the stated

objectives of the Egyptian agricultural policy, to provide

adequate basic foods to the population and achieve food self-








13

sufficiency, conflict. This creates increased inefficiency

and rising government expenditures. He claims that price

stabilization was achieved, but it increasingly required

allocation of foreign exchange to food imports, as well as

deficit financing.

Most of the economic studies of Egypt are descriptive.

For example, Handoussa (1991) studied the effect of foreign

aid on Egypt's development, and the World Bank (1993)

constructed a framework in order to assist the Egyptian

government in developing the agricultural sector. The

International Food Policy Research Institute (IFPRI) published

a series of works in the early 1980s, that focused on Egypt.

These include the descriptive studies already mentioned above,

by Alderman and von Braun (1984) and von Braun and de Haen

(1983), that discuss the effects of food pricing policies and

rationing on Egyptian agriculture, income distribution and

consumption.

Quantitative economic studies on Egypt are mostly

conducted through the use of Computable General Equilibrium

Models (CGE). For example, Robinson and Gehlhar (1994) used

such a model to examine the effects of economic liberalization

on water and land, two scarce resources in Egypt. Two IFPRI

publications by Scobie (1981, 1983) utilize complete demand

systems to assess the impact of government policies on foreign

exchange and trade, and specifically food imports. The

earlier work assumes a foreign exchange constraint and, based









14

on that, allocates total import expenditures on total wheat

imports, other imports and foreign reserves. Then the effects

of consumer subsidies, foreign aid and domestic supply on the

wheat import market are assessed. Whereas this study is

comprehensive and interesting, and ideally would be updated

and expanded for the present study's purposes, data is limited

with respect to both accuracy and availability. Obtaining

consistent data series for consumer and producer subsidies

seems to present the most difficulties, but other data series

are problematic as well, for example, finding consistent and

comprehensive aid figures.

As will be discussed in the next chapter, this study will

assess the US position in the Egyptian wheat import market

with respect to its competitors. As far as the author knows,

there have been no previous studies examining this specific

problem in any framework.















CHAPTER 2

PROBLEM STATEMENT AND OBJECTIVES

The Current Situation in the Wheat Import Market
and the US Role

In 1992 and 1993, wheat, wheat flour and corn imports from the

US accounted for 77% of total US agricultural imports to Egypt.

Table 3 presents the US share of quantity of wheat and wheat flour,

and Table 4 presents the US share of value of total food, wheat and

wheat flour imports to Egypt. These shares include aid shipments.

In 1993, the value of unmilled wheat and wheat flour imports to

Egypt from the US was $156.6 and $57.9 million, respectively. The

following subsection examines the wheat market in detail.

Wheat and Wheat Flour'

The government has a monopoly in imports and has been

responsible for the distribution of both domestically procured and

imported wheat. In 1991, about two-thirds of total commercial

wheat was distributed by the government. The same year, nonbread

use accounted for 5% of total commercial wheat. Out of the portion

distributed by the government, about 66% was used for balady bread,

the cheapest kind, 14% for shami, which is more expensive, and the

rest for fino, a high priced, nonarabic bread, and flour.




1 The information for this section is from Parker and
Shapouri, unless otherwise indicated.

15











Table 3. Share of US to World
Imports to Egypt, 1965-93.


Year

1965
1966
1967
1968
1969

1970
1971
1972
1973
1974

1975
1976
1977
1978
1979

1980
1981
1982
1983
1984

1985
1986
1987
1988
1989

1990
1991
1992
1993


Wheat

61.86
80.89
8.41
0.00
0.00

0.00
0.00
0.00
9.55
29.31

38.66
49.26
50.82
37.90
52.20

46.94
56.75
67.08
48.85
25.93

29.54
38.00
49.49
43.17
67.88

28.49
42.18
71.56
52.32


Quantity of Wheat and Wheat Flour


Wheat
Flour

30.70
52.04
1.87
0.00
0.00

0.00
0.00
0.00
3.58
0.00

13.71
40.76
42.69
36.60
33.63

51.56
29.38
38.06
82.12
30.31

33.20
44.44
56.70
54.58
61.66

47.86
73.58
51.79
42.42


Source: Calculated from United Nations, Commodity Trade Statistics.










Table 4. Share of US to World Value of Total Food, Wheat and Wheat
Flour Imports to Egypt, 1965-93.

Wheat Total
Wheat Flour Food

1965 63.33 33.80 40.02
1966 80.79 55.31 43.84
1967 8.38 2.11 5.46
1968 0.00 0.00 0.50
1969 0.00 0.00 2.31

1970 0.00 0.00 2.56
1971 0.00 0.00 1.76
1972 0.00 0.00 2.25
1973 12.18 5.98 10.96
1974 28.82 0.00 27.67

1975 40.84 10.62 30.93
1976 46.18 42.95 35.45
1977 51.27 41.78 36.30
1978 40.85 40.83 28.52
1979 61.15 37.25 34.30

1980 50.01 52.30 36.34
1981 57.73 26.77 37.98
1982 67.17 39.98 39.37
1983 48.82 82.06 35.67
1984 25.97 31.53 22.13

1985 25.62 35.40 23.49
1986 34.43 47.36 28.10
1987 45.17 59.44 29.39
1988 46.19 55.24 25.23
1989 66.69 64.54 36.63

1990 28.33 41.97 23.46
1991 42.39 75.57 27.27
1992 67.85 55.05 34.25
1993 46.68 41.51 30.96

Source: Calculated from United Nations, Commodity Trade Statistics.








18

Egypt has been a recipient of US export assistance under

programs such as PL-480, the Export Credit Guarantee Program (GSM-

102), and the Export Enhancement Program (EEP). Australia and the

European Union (EU) have also been providing wheat and wheat flour

aid to Egypt. The quantities in metric tons (MT) in grain

equivalents (to convert flour to grain equivalent one needs to

divide flour quantities by 0.73) for wheat and wheat flour from

different donors are shown in Table 5.

The US is the leading supplier of wheat for Egyptian imports

and competes with Australia and the EU for that market. After

1991, private imports of wheat were allowed, but the bulk of wheat

imports are still handled by the government. The Australian wheat

is preferred for its high quality, and the US wheat follows in

order of preference, due to its higher moisture content, and

variability in its quality. The main quality consideration is live

insects and insect damage.

Wheat is the country's leading food staple, accounting for

half of the calories in the daily diet as bread and wheat products.

Most of the domestically produced wheat is consumed at its

production site, with less than 10% sold to the government.

Imported and domestically procured wheat is processed in government

run mills and sold at subsidized prices. Producer prices have

increased, yields have increased and government provides mechanical

drill planting. As land rents are deregulated and the input

distribution system is privatized, the cost of production is

increasing.











Table 5. Aid Shipments of Wheat and Wheat Flour in Grain
Equivalents, by Selected Donors, Egypt, 1974-93.


Year


US/a


EU Australia


----------- MT -----------


1973
1974

1975
1976
1977
1978
1979

1980
1981
1982
1983
1984

1985
1986
1987
1988
1989

1990
1991
1992
1993


1300 181350
5088 100980


441377
690490
1193610
983451
1081650

1035325
1060840
1168967
985056
974947

977907
984940
1024226
793038
602829

755537
1459000
456045
632611


104400
98320
96800
137113
202342

145802
154035
178010
205367
187535

195340
120000
301913
284747
135200

150573
305789
134000
130147


0
0

0
10000
0
20000
22300

20000
25000
40000
40000
80000

40000
40000
50000
50000
50000

0
60500
48416
0


Sources: a/Quantities for years 1974-90, are PL-480 shipments from
Parker, J. and S. Shapouri, Determinants of Wheat Import Demand:
Egypt, p. 8. The rest are from FAO, AGROSTAT.








20

Egypt is the third largest wheat importer in the world after

the former Soviet Union and China2. Egypt's milling capacity is

expanding, and imports of wheat flour are declining. Several

Egyptian companies are in the process of importing flour milling

equipment in order to run their own mills and avoid renting public

enterprise sector milling capacity (FAS, p. 27).

Problem Statement

Wheat is Egypt's leading food staple and the economy's

liberalization is resulting in increased costs of production.

Limited water availability is another constraint for domestic

production. Since all agriculture is irrigated in Egypt, and water

use for irrigation has been free, it seems that with the economy's

liberalization the price of water might become positive. Even if

that does not happen, it would be more profitable to grow crops

with higher returns per feddan than wheat. For example, Sarris

claims, that since past studies show that self-sufficiency in wheat

is technically infeasible, a shift of production toward high valued

tradable products, like horticultural ones, is advisable (Sarris,

p. 55). While wheat yields have been increasing, due mainly to the

adoption of high-yielding varieties, the country faces a land

constraint. As mentioned above, most of the domestically produced

wheat is home-consumed and does not reach the market. Egypt has

been the third largest wheat importer in the world, private wheat

imports are increasing, and private mills are starting to operate.



2 This statement is consistent with the opening statement of
the previous chapter; the latter refers to US exports.









21

The Egyptian import wheat market is important for US wheat

producers and exporters. US agricultural exports to Egypt account

for 41% of its total exports to North Africa, and amounted to $766

million in 1992, and $661 million in 1993 (USDA, 1994). The value

of unmilled wheat and wheat flour exported from the US to Egypt in

1993 was $214.5 million. In addition, due to reasons explained in

the previous paragraph, there is potential for increased wheat

exports to Egypt. The expansion of the Egyptian wheat import

market could be beneficial not only for wheat exporters in the US,

but for the whole US agricultural sector, since production and

exports of grains in general, and wheat specifically, have been and

will continue to be very important for US agriculture. Therefore,

the potential for increased access to export markets is of

considerable interest. A model of wheat import demand for Egypt

that incorporates the US and its major competitors for this market

will be used to assess the US role in this market. Also, some

insight on how wheat imports relate to considerations such as food

subsidies, foreign aid and domestic wheat production will help

assess how wheat imports would be affected by likely future changes

in those items.

Furthermore, situations similar to Egypt's are faced by other

developing countries. Constraints in basic food production due to

limited resources and heavily distorted domestic markets are common

in developing countries. Governments in developing countries are

starting to realize that agriculture has to be a productive sector

of the economy, and that heavy consumer subsidies and other









22

interventions are unsustainable in the long run. At the same time,

international development agencies and donors are insisting on

liberalization of the economy as a prerequisite for continuing

assistance to developing countries. The US is a major competitor

in the world grains market and an important source of grain imports

for many developing countries. The framework developed here to

examine the Egyptian wheat import market could be adapted for other

developing countries that import grains from the US and have

economic distortions that may be eliminated.

Objectives

Specifically, the proposed import demand model will address

the following:

1. Assess the US' versus its competitors' positions with

respect to the Egyptian import wheat market, by examining in detail

the substitutability between wheat from different sources of

origin.

2. Assess the effects of policy changes in Egypt and in the

supplier countries, with respect to their implications for current

and future Egyptian import demand for wheat.



In the following chapters, the connection between the problem

at hand and empirical demand analysis is first established. In

order to do that, theoretical and empirical issues concerning

demand analysis are examined and related to the Egyptian case. The

proposed methodology is then presented, with subsections examining

specific issues, such as incorporating theoretical restrictions









23

into the empirical model, describing the system-wide differential

approach, preference structures and how they can be used to derive

conditional demand systems, and the Rotterdam model and how it is

used to build a conditional geographic import demand system. Then

the data and estimation results are presented. Finally,

conclusions and suggestions for further research are discussed.














CHAPTER 3

THEORETICAL AND EMPIRICAL CONSIDERATIONS

Demand Analysis in Theory and Practice

Most literature reviews on demand analysis stress that,

while applied work has to build on theory, the tools that

theory provides at its present state generally do not deal

explicitly with the problems encountered in practice. Surveys

of developments on combining demand theory and applied demand

analysis are presented by Brown and Deaton (1972), Powell

(1974), Theil (1980) and Barten (1977).

The first of the above mentioned works, is a thorough

survey of models of demand analysis. The authors mention that

agricultural commodities were the first to be used for applied

analysis, because single-equation partial equilibrium analysis

is best suited to examine "a homogeneous commodity with a

single quantity dimension, stable consumer preferences, and

relatively large fluctuations or trends in supply which are

independent of the current market price; and these conditions

are most nearly met by many agricultural staples" (Brown and

Deaton, p. 1147). Referring to the empirical issues of demand

analysis encountered since the 1950s, the authors claim that

focus has shifted from the classical approach, which is

concerned with what values the price and income elasticities








25

of a commodity should be, to methodological questions that

deal with how these elasticities should be measured.

They define as "pragmatic" the ad hoc approach, where the

variables we are directly interested in are included in the

analysis, while others are ignored or summarized (Brown and

Deaton, p. 1151). Then they claim that it is better to

construct models for which the restrictions implied by theory

can be imposed and tested, for which the pragmatic approach

does not account. Following this line of argument, a model of

a system of demand equations not a single equation approach is

to be used for this study. This way, the theoretical

restrictions can be imposed and tested and the interactions

between wheat from different sources of origin can be

explicitly accounted for. In addition to conforming to demand

theory, the imposition of these restrictions also serves to

reduce the dimensionality of the estimation problem and to

deal with certain shortcomings of the available data, as

Barten claims in his survey of demand systems (Barten, 1977,

p. 23).

In the sections that follow, several issues are addressed

that have important implications in applied demand analysis,

and for which theory does not provide a clear direction. The

most important of these issues are identification and

aggregation among consumers and commodities.








26

Exogenous Prices

Brown and Deaton claim that strong assumptions must hold

regarding supply conditions, in order to assure that one

estimates demand, instead of a function that represents supply

or a mixture of supply and demand. Prices are usually assumed

fixed by producers or on world markets, with supply

forthcoming at that price. Then demand equations are written

with quantities dependent on prices and income.

For the Egyptian case, an argument can be made in favor

of the exogeneity of world prices, since for trade purposes,

Egypt is reasonably assumed to be a small country. Despite

the fact that Egypt is the third largest wheat importer, it

seems that it can not affect the prices its suppliers charge.

Furthermore, there have been no quantitative controls for

imported wheat to Egypt, only tariffs, which strengthens the

argument in favor of exogenous prices. In the proposed

framework, savings are ignored and income is equivalent to

total expenditure, so that the allocation of that quantity is

examined and not of income (Brown and Deaton, p. 1189).

To further examine the issue of world prices, one needs

to look into the conditions that prevail in the domestic

markets of the wheat suppliers to Egypt, namely the US, EU and

Australian domestic markets.

Australia can be considered a small country for wheat

exports due to the volume of their exports compared to the

volume of exports that dominate the world market and originate








27

in US, Canada and EU. Also, the only domestic distortion is

the existence of the Australian Wheat Board (AWB), which buys

all the wheat produced and sets the domestic price higher than

the one that prevails in the export market. This way the

domestically demanded quantity falls and more wheat is

allocated to the export market.

Figure 1 depicts the situation graphically. Domestic

demand and supply are represented by curves D and S,

respectively. Given the world price for wheat (Pw) total

demand is the section of D above Pw that represents domestic

demand, and Pw from that point on. Domestic quantity demanded

is Qd and domestic quantity supplied Qs. Then exports are Qs

minus Qd.

The AWB sets the domestic consumers' price at Pc, which

causes the domestic quantity demanded to decrease to Qd'.

Domestic producers receive the weighted average of the

domestic (Pc) and export price (Pw), Pp. Then domestic

quantity supplied increases to Qs'. Exports increase to Qs'

minus Qd'.

The distortions in the Australian wheat market are

minimal when compared to the distortions prevailing in the EU

and US markets. Calculated in percentages for 1992, wheat

producer and consumer subsidy equivalents were 7 and zero for

Australia, respectively, while for the EU they were 47 and

-38, and the US 37 and -15,respectively (Organization for










































Od' Od as sw a


Figure 1. Australia as a Small Exporter of Wheat.








29

Economic Cooperation and Development (OECD), 1995, pp. 202-

203, 214-215, 250-251).

The US policies that affect the export price for wheat

are deficiency payments and export subsidies. Because the US

is a large country for trade purposes, its domestic policies

affect world prices. Both deficiency payments and export

subsidies tend to lower world prices, even though that happens

for different reasons and in different magnitudes depending on

the policy.

Figure 2 depicts the US deficiency payments scenario.

The first graph shows the domestic US situation, the middle

graph shows the excess demand and supply (Trade), and the last

graph shows what happens to the rest of the world (ROW). With

a deficiency payment in the US, the domestic producer price

will increase to P'. Excess supply will shift from S to S',

so that the world price will decrease to P''. Domestic and

foreign consumers pay P' and domestic producers receive P'

and supply more. Foreign producers are worse off, because

they receive P' and produce less, and US taxpayers are also

worse off.

An export subsidy is presented in Figure 3. It would

raise the producers' price to P' in the US, and domestic

consumers will have to pay P' too. The effect of the export

subsidy on the ROW is to depress the world price to P'', which

is lower than the P" that results from a deficiency payment.


















































Figure 2. The US as a Large Exporter of Wheat and the Effects
of a Deficiency Payment.


















































Figure 3. The US as a Large Exporter of Wheat and the Effects
of an Export Subsidy.








32

The most important EU policy influencing the wheat export

market is export subsidies. The difference between the US and

the EU with respect to export subsidies is that without the

subsidies the EU would be an importer. So, the situation for

the EU is the same as the one for the US in Figure 3, except

for Pw in the first graph being below the point where domestic

supply and demand intersect.

On the Egyptian side, it seems that aid or political

leverage is not a determining factor for the import prices

that Egyptians face from the US, EU and Australia. When

compared with another of the top three world wheat importers,

China, it is observed that the trends in prices they face are

similar, despite the fact that the countries are very

different in size, geographical location and political

leverage. Tables 6 and 7 show volumes and prices (calculated

by dividing value by volume) of wheat imports for China and

Egypt. More details on how aid is treated in this study are

discussed in subsequent sections.

Because Egypt is a small country for trade purposes,

domestic wheat supply does not affect world prices, and

neither do domestic consumer subsidies. Figure 4 depicts the

Egyptian situation. Before any intervention, supply is the

part of domestic supply below Pw and Pw from that point on.

The Egyptian government sets the domestic consumer price at Pc

and the domestic producer price at Pp. Then the quantity

supplied decreases from Qs to Qs' and the quantity demanded










Table 6. Wheat Import Quantities and Prices from Selected
Suppliers, China, 1973-93.


$/MT


$/MT


Australia
MT $/MT


1973 2648648 115.33
1974 1905493 135.09


1975
1976
1977
1978
1979

1980
1981
1982
1983
1984

1985
1986
1987
1988
1989

1990
1991
1992
1993


0
0
0
1847165
1481818

5800684
7070873
6485523
2335435
3818957

707929
3192
1563507
5768226
8293176

3919284
4586557
3334444
2588981


0.00
0.00
0.00
148.98
158.94

197.09
197.41
177.53
177.89
165.71

150.74
130.61
96.37
122.79
171.45

162.57
107.71
134.18
117.60


6 183.33 768462
205007 157.05 1317839


0
0
0
0
90280

134164
632354
706408
813454
27000

320632
114374
567649
29996
1611679

2173848
1265591
1340261
56915


0.00
0.00
0.00
0.00
133.71

158.63
187.52
162.23
132.90
174.70

117.07
106.75
114.60
180.02
172.36

167.03
113.93
127.00
180.01


1327551
933509
2984537
2434736
2968131

1997812
1242804
2113110
416391
2429362

1254856
2785944
4431620
397187
1676585

1386170
1364645
219065
618722


Source: Calculated from United
Statistics.


Nations,


Commodity Trade


Year


75.80
159.12

162.63
146.40
102.58
99.81
136.13

177.45
213.04
171.41
157.84
166.86

147.47
119.59
97.48
114.66
182.52

176.49
119.18
163.79
145.93










Table 7. Wheat Import Quantities and Prices from Selected
Suppliers, Egypt, 1973-93.


$/MT


$/MT


Australia
MT $/MT


1973 142273 119.50 510747 105.25 552062
1974 659692 259.78 674872 263.42 720354


214.56
156.07
137.08
155.97
129.57

221.10
249.62
241.17
192.62
197.60

100.19
133.02
96.10
131.07
187.54

176.09
123.74
134.59
127.89


643252
434031
33862
352647
448524

869179
399233
66690
511265
263377

1002
0
188602
365911
289035

998034
191156
110000
318774


192.31
184.88
162.57
138.55
81.71

194.99
224.90
161.91
176.16
168.89

125.75
0.00
159.52
110.54
159.71

160.71
170.35
155.10
137.84


833567
717455
506568
385684
237155

243057
776029
709303
679186
1102481

1254326
1805813
1405498
1591094
658692

1805461
1912098
1242054
755395


67.34
273.26

190.32
169.01
156.58
120.51
116.24

195.50
246.35
246.69
208.27
216.59

127.08
156.01
102.62
115.23
208.25

184.93
117.77
161.48
169.33


Source: Calculated from
Statistics.


United Nations,


Commodity Trade


Year


1975
1976
1977
1978
1979

1980
1981
1982
1983
1984

1985
1986
1987
1988
1989

1990
1991
1992
1993


1036246
1161329
1229222
1110820
1175468

998040
1754962
1958584
1258829
705752

690228
1293910
1798248
1543623
2083433

1269639
1736255
3552272
1224543



















































Figure 4. Egypt as a Small Importer of Wheat.








36

increases from Qd to Qd'. The effect of aid will be similar

to that of increased income (Scobie, 1981; Handoussa, 1991).

The demand curve D will shift out to D'. Pw is the average

price of imports from different sources that the Egyptians

face, for the study's purpose. It is assumed to be very close

to the world price referred to in the analysis for Australia,

US and EU, so that the connection between Egypt and its wheat

suppliers is through Pw.

To summarize, for the purpose of this study, the wheat

world price that Egypt faces every year is assumed to be the

average price that results from dividing total value over

total quantity of wheat imported from all sources. As a small

country, Egypt does not influence this price. The similarity

of import prices faced by China, implies that there are no

political variables greatly affecting the levels of these

prices either.

Aggregation and Separability

With respect to the aggregation problem, a distinction

has to be made between aggregation among consumers and

aggregation among commodities. According to Barten, "actual

data usually refer to consumption for a limited number of

highly aggregated groups of goods, for a whole economy, and

for a relatively small number of years. The theory of

consumer demand, however, is about individual behavior with

respect to the choice of quantities of a potentially large

number of elementary goods" (Barten, 1977, p. 23).








37

To deal with aggregation over consumers, most of the work

done in applied demand analysis either ignores it altogether,

or explicitly assumes a representative consumer. Ignoring the

issue obviously does not solve it, while the shortcomings of

the representative consumer assumption are pointed out by many

authors. Exact aggregation over consumers implies very

restrictive assumptions, for example, all consumers' Engel

curves must be parallel straight lines. The methodology

advocated in this study, explicitly accounts for the issue.

As demonstrated by Barnett (1979; details on Barnett's

approach in the methodology section) the proposed model is an

aggregate rather than an individual consumer model, built with

no restrictive assumptions on individual consumers'

preferences.

The aggregation over commodities is usually treated by

assumptions on separability, which refers to the degree of

dependency among individual commodities and groups of

commodities separabilityy is discussed further in the

methodology section). The consumer is assumed to exhibit

independent preferences for distinctively different

commodities, so that for example his consumption of shoes

would be independent of his consumption of butter. On the

issue of separability among different groups of commodities,

researchers' opinions vary. Moschini, Moro and Green (1994)

tested for separability for food items and found that the

widespread practice of modeling food items in terms of








38

conditional demand systems seems to be justified (they also

present a literature review on the issue). Some of the works

that found evidence against the separability hypothesis

between imported and domestic goods are Winters (1984), and

Brenton (1989), but these works involve the Almost Ideal

Demand System (AIDS) model that does not nest separability.

Truett, Truett and Apostolakis (1994) tested for separability

of domestic inputs and imports using a translog cost function

for Mexico, and could not reject it at the 5% level of

significance but they could at 10%. The issue of separability

is important for the derivation of conditional demand systems,

as developed in the methodology section.

For the purposes of the study at hand, a decision had to

be made on how to treat the separability issue. There seems

to be evidence that the markets for Egyptian imports and

domestic goods are separable. According to ez Elarab (1982),

Egyptian imports were found to be weakly separable from

domestic goods. This implies that the aggregate utility

function is first maximized subject to GNP, and then two

subaggregate utility functions are maximized: one in order to

determine the demand for domestic and the other for imported

goods.

As discussed in detail later, in order to estimate the

allocation model, two further stages are assumed that utilize

the separability concept. Imports are broken down to food and

other and food is further broken down to wheat and other.
















CHAPTER 4

METHODOLOGY

In this section, the different approaches to estimate

demand systems are briefly discussed, and the general

restrictions that theory requires to be imposed on these

systems are presented. A discussion of the general

differential model and the development of conditional

empirical models follows. Then, the empirical models to be

estimated will be presented, namely the Absolute Price Version

(AP) and the Uniform Substitutes (UB) Version of the Rotterdam

model, as different versions of a geographical allocation

model.

Approaches to Demand Estimation and General Restrictions

As claimed by Clements, the estimation of systems of

demand equations to determine empirical regularities in

consumption patterns is one of the primary aims of applied

demand analysis (Clements, p. 1). There are several

approaches to derive systems of demand equations.

The older approach to demand analysis was to directly

specify the demand for a good as a function of income and

prices, without any reference to the utility function. A

popular example is the double log system, which is attractive

due to its simplicity, but does not satisfy the adding up

39








40

constraint for all values of income, meaning that if income

rises sufficiently, expenditure on one good will exceed

income.

Another approach of specifying demand systems is directly

deriving them from a specified utility function. The consumer

has a fixed amount of income to spend on n goods and services

at fixed prices. To determine the quantities bought,

assumptions about the consumer's preferences must be made,

implied through his utility function. Maximizing the

consumer's utility subject to his budget constraint, results

in a unique demand function for each good. These demands are

functions of income and prices.

According to Hicks, the utility function that is

maximized represents a preference ordering, which implies that

the demand functions derived are invariant under monotone

increasing transformations of the utility function (Hicks, p.

307). It also implies that there could be no declining

marginal utility for a good, because this property is not

invariant under monotone increasing transformations of the

utility function (Theil, p. 3).

Given that a consumer's commodity set is divisible, does

not include negative commodities, and is unbounded from above,

and his preferences satisfy the axioms of comparability,

transitivity, continuity, monotonicity, strict convexity and

differentiability, the consumer's "well-behaved" utility

function can be defined.








41

To answer whether the utility function really exists, if

the consumer wants to maximize it and if he does, whether he

is he able to do so, Phlips maintains the following arguments.

The utility function is a formal concept used by the

economist, not the consumer, in order to describe the

consumer's observed behavior in the market and to forecast his

future behavior. Reality, expressed in statistical data, is

intelligible when interpreted through a formal hypothesis

imposed on the observed data. The economist, not the

consumer, maximizes utility to find the optimal quantities the

consumer purchases. The conclusions derived from the

maximization of the utility function, refer to the demand

equations, taking the form of restrictions imposed on these

equations. Finally, he claims that "... applied consumption

analysis appears then as the art of constructing and

effectively utilizing interesting theoretical restrictions in

the (econometric) estimation of demand equations" (Phlips, p.

27).

When a demand system is directly derived from maximizing

a utility function of a specific functional form, it satisfies

the general restrictions of demand functions. These general

restrictions refer to homogeneity of degree zero, adding up,

negativity, and symmetry.

Homogeneity of degree zero means that if prices and

income are multiplied by a positive constant, the quantity

demanded does not change. It implies that the sum of all









42

direct and cross elasticities with respect to prices for any

commodity must equal minus its income elasticity.

Adding up means that the budget constraint must be

satisfied, which implies that the marginal propensities to

consume must sum to one.

Negativity and symmetry refer to the Slutsky equation,

that shows that the change of the quantity demanded of a good

to a change in its price, or any other good's price, can be

decomposed into an income effect and a substitution effect.

With respect to the substitution effect, one can conclude that

the own substitution effect is negative and that the matrix of

substitution effects is symmetric.

An example of a demand system directly derived from a

specified utility function is the linear expenditure system

(LES), which results from maximizing a Klein-Rubin utility

function (Clements, p. 9). The LES does not admit inferior

goods (it implies that as a consumer becomes richer, food

becomes more of a luxury). This implication is true for all

demand systems derived from a preference independent utility

function (more details on preference independence are

presented in the relevant section).

A utility expressed as a function of income and prices is

an indirect utility function (can be derived by substituting

demands into utility). Then the negative ratio of the

derivative of utility with respect to prices to the derivative

of utility with respect to income gives the quantities








43

demanded (Roy's theorem). The translog indirect utility

function is an example.

The consumer's cost function is the dual to its indirect

utility (can be derived by substituting cost as a function of

utility and prices for expenditure in indirect utility), and

shows the minimum expenditure to reach a specific level of

utility given the prices. Differentiating the cost function

with respect to prices gives the demands (Shephard's lemma).

An example is the AIDS model by Deaton and Muellbauer (1980).

While the direct derivation of demand equations from the

consumer's utility maximization results in these demand

equations automatically satisfying the theoretical

restrictions necessary, this method has shortfalls. First,

there is only a limited number of known utility functions that

one can use in order to solve the consumer's utility

maximization problem, which greatly narrows one's available

options. In addition, one cannot always derive demand

equations from these utility functions, or the demand

equations estimation might be problematic.

Another approach can be used in order to better suit the

particular problem at hand, i.e., increasing one's

flexibility, while satisfying the theoretical restrictions.

According to this approach, demand equations can be directly

specified, and the restrictions that make them theoretically

plausible can be imposed afterwards.








44

These restrictions that theory implies, are most

conveniently expressed in terms of derivatives of the demand

equations (i.e., quantities of goods expressed as functions of

prices and total income), or in terms of elasticities, which

are derivatives of the logarithmic version of the demand

equations (Barten, 1993).

The differential approach to estimating a demand system

is advocated in this study, because of its numerous

advantages. It connects consumer theory to empirical

analysis, it is flexible, it allows for a choice of functional

form (see general model below), and it allows for a

hierarchial allocation that links demand for goods from the

consumer to the import level demand and to sources of imported

goods (see conditional geographic import demand system below).

Theil claims that the differential approach is an

allocation or decomposition theory, that considers a total

budget and how it is allocated into amounts spent on specific

goods, like total imports of a country allocated over the

exporting countries. He compares the system-wide approach to

a view from a mountain top. While this has advantages, its

disadvantages include difficulty of incorporating ad hoc

explanations (for example, housewives hoarding nonperishable

goods fearing a war) (Theil, pp. 5-6).









45

A System-wide Differential Approach to Import Demand

Only the final stages of the mathematical derivations for

the following subsections are presented in the text; details

can be found in Theil (1980).

The general differential model can be expressed as:

wi d(log q,) = 68 d(log Q) + 4 E (6, 0,8,) d(log p,)
)-1

where, p, and q, are the price and quantity of commodity i

(i=l, ..., n), w, is the budget share for i equal to (p, q,)/M,

M is the total expenditure, 8, is the marginal share of the ith

good, that is, the additional amount spent on this good when

income increases by one dollar, with E 6, = 1, and d(log x) =


dx/x for x either p or q, and all logs are natural logarithms.

The term d(log Q) is the Divisia volume index and the

term d(log P) the Divisia price index, and they are weighted

means of logarithmic quantity or price changes, respectively,

with weights equal to the corresponding budget shares, so
n n
that E w, d(log qj) = d(log Q) and E w, d(log p1) = d(log P).
1-1 1-1

"Changes" refer to any kind of displacement (time or other)

and d(log Q), d(log P) are abbreviations. The Divisia volume

index is the true index of real income which compares the

utility level of (M+dM, p+dp) with that of (M,p) at the

prevailing prices p, while the Divisia price index is the true

cost of living index which compares p+dp with p, at the

prevailing utility level u(M,P).








46

The ratio of the marginal share to the budget share is

the income elasticity of the ith good (0B/w,) =

p,(aq,/aM)(M/(plq,)) = (alog q,)/(alog M). The weighted mean of

the income elasticities with the budget shares as weights is
n n
E w, (8,/wi) = E 8, = 1. A good is a luxury or a necessity,
i-1 1-1

if its income elasticity is larger or smaller than one,

respectively. A negative income elasticity, i.e., a negative

marginal share implies the good is inferior.

The term 1/0 is the income elasticity of the marginal

n n n n
utility of income, and E 61, = 8, and E E 8O, = E 81 = 1, where
j -1 l lJ 1 1

80,'s are the normalized price coefficients.

Following mathematical manipulations, the general

restrictions of a demand system can be derived. Among those,

the Slutsky equation derived, explicitly contains a specific

and a general substitution effect (in addition to the income

effect). The specific substitution effect refers to the

specific relation in terms of u1" (the (ij)t" element of the

inverse of the Hessian matrix U of the utility function u(q))

between goods i and j, while the general substitution effect

refers to the competition of all goods for an extra dollar of

the consumer's income. It can be shown that the substitution

matrix is symmetric and negative semidefinite (Phlips, pp. 52-

53).

Looking at the general differential model, the

relationship between the Hicksian substitutes and complements,








47

and the specific substitutes and complements can be

established. Since 0 < 0 if 8,6 < 0, an increase in the jth

relative price with real income constant, increases the demand

for the ith good. [1,,] is symmetric so that 1,, < 0 implies 86,

< 0, and an increase in the ith relative price, increases the

demand for the j"t good. Then i and j are specific

substitutes. If O,j, 6,, are both positive, an increase in the

relative price of either, decreases the demand for the other,

so that they are specific complements. Hick's definitions

refer to the total effects, so they refer to (8O 8O,3)

(Theil, p. 17).

Preference Structure and Conditional Demand

Preference Independence

Preference independence (PI) refers to the case of an

additive utility function, u(q) = u,(ql) + ... + u,(q,). Then

the marginal utility of each good is independent of the

consumption of all other goods, which means that U and U-1 are

diagonal and 1,, = 0 for i not equal to j, and On = O1. Then

no good is a specific substitute or complement of another, 8,,

is positive as a diagonal element of a positive definite

matrix, and 8n = ,8 shows that the marginal share of each good

is positive, that is, there are no inferior goods under

preference independence. Define d(log P') = E 80 d(log p,), as
1-1

the Frisch price index, that uses marginal shares as weights.

Luxuries have a greater weight in the Frisch index, and the








48

reverse is true for necessities. An inferior good results in

its weight being negative in the Frisch index, so that an

increase in the good's price lowers that index. It is the

true marginal price index which compares two price vectors p

and p+dp at the same utility level u(p,M) (Theil, p. 26).

Then the budget shares move in the direction of the

corresponding marginal shares as the consumer becomes richer

(Theil, p. 27).

Substituting the Frisch price index into the general

differential model one gets

w, d(log q,) = O~ d(log Q) + 0 6, d(log p,/P').

If this last equation is divided by the budget share,

Oj/w, is the income elasticity of demand for the ith good and

( (0,/w,) is the own price-Frisch-deflated elasticity. Under

PI the own price elasticity is proportional by 0 < 0 to income

elasticity (Theil, p. 18).

Block Independence

If the goods can be grouped into G groups G < n, S,, ...,

SG, so that each good belongs to one group, u(q) = u1(q,) + ...

+ uQ(q,). Then the marginal utility of each group depends only

on the quantities consumed of goods that belong to the same

group and U and U-' are block-diagonal. The general model

becomes

w, d(log q) = 6, d(log Q) + 0 E (80 08,8) d(log p,),
's with i and j in different

with 801, = 0,, i e Sg. All 0,'s with i and j in different
1 6 Sg









49

groups vanish, so that no good is a specific substitute or

complement of a good that belongs to another group (Theil, p.

19).

To derive a demand equation for all goods in a group,

define W, = E w,, g = E 8,, d(log Q,) = E (w,/W,) d(log q),
i 6 Sg 1 E 1s Sg
G
d(log P,) = E (6,/e,) d(log p,), d(log Q) = E Wg d(log Qg),
i e Sg g -
G
d(log P') = E ), d(log P,*), so that
g 1
g=1
W, d(log Q,) = E, d(log Q) + 4 Q, d(log (Pg/P')). (1)

Inferior goods cannot occur under block independence

(BI), because 0g > 0 (Theil, p. 99).

Blockwise Dependence

When the utility is an increasing function f, not the sum

of the group utility function, u(q) = f(u,(q,), ..., u,(qG)),

and (a2U)/(a(p,q,)a(pjq)) = a,, i e S,, j E S,, g not equal to

h.

Then a change in the marginal utility of a dollar spent

on the ith good caused by an extra dollar spent on the jth good

(which belongs to a different than i group), is independent of

i and j and the same for all pairs of goods in the two groups.

Then

G
W, d(log Q,) = 8, d(log Q) + 0 E g,, d(log (P,*/P')) (2)
h-1
h = 1
where

Gh = A OE j, (3)
1 F Sg j E Sh
G G G G
E 0g, = ,8, EE 8,, = E O, = 1 and h = 1, ..., G, g = 1, ..., G.
h g h g -









50

Because of (3), Sg, S, are specific substitutes or complements,

when O,, is negative or positive, respectively.

Because now Eg can be zero, 8,/eg is replaced with 8,*

(then if Og = 0, O = 0, for i E Sg) so that the Frisch price

index of S, is d(log P,*) = E 80, d(log pj). Then O is the
I E Sg

conditional marginal share of the ith good within its group.

It shows the proportion allocated to the ith good, when the

additional amount from an increase in income by one dollar

spent on Sg is Og.

After some mathematical manipulation, the conditional

demand equation for the it" good within its group is derived:

w, d(log q,) = 0,* Wg d(log Qg) + 0 E 8,, d(log (p,/P')). (4)


The consumer first uses (2) to allocate income among

groups of goods. For this, knowledge of real income and the

price indexes of the groups (not price changes of individual

goods) is needed. Then, the consumer allocates total

expenditure for each group, within the group. So, the change

in the consumption volume of the group (from the first step)

and price changes of all goods in the group are required.

Rotterdam Model

Following Theil (Theil, pp. 156-161), the Divisia and

Frisch indexes are converted to their finite change versions:
n n
w = 1/2 (w,.,,1 + w,,), DPt = E w it Dpit, DQt = Z w it Dqqt,
1 1 1 1 -
DP,* = S 09 Dpt, where the first term is the average budget
1 = 1









51

share between period t and period t-1, i=l, ..., n, the second

term is the finite version of the Divisia price index, the

third the finite version of the Divisia quantity index and the

fourth the finite version of the Frisch price index. Again,

the left hand side terms of these indexes should be considered

abbreviations for their right hand sides.

If the general differential model is converted into its

finite change version, rr, = 0 (8,, 8, 0,) the parameters IT,,

8, are assumed constant and a disturbance term is added, the

absolute price version of the Rotterdam model is
n
W I Dq1t = 8, DQt + E rri Dpt + et. (5)


The coefficients of the second right hand side term add

up to zero over j, and are called the Slutsky coefficients of

the Rotterdam model. Theil shows that the disturbances et,

..., nt add up to zero, and all the disturbance covariances

are proportional to the corresponding Slutsky coefficients.

The income elasticity is given by dividing the marginal

budget share by the budget share, so that the sign of the

income elasticity depends on the sign of the marginal budget

share, i.e., a commodity is inferior when they are both

negative and non-inferior when they are both greater than or

equal to zero. For a non-inferior good to be a necessity

(luxury), the marginal budget share has to be less than or

equal (greater than) to the budget share, and the income

elasticity less than or equal (greater than) to one.









52

Therefore, a good can change from luxury to necessity,

depending on the change in the budget share, but it cannot

change from a non-inferior good to an inferior good. While

this is a limitation of the Rotterdam model, at the level of

aggregation that the data are usually used, inferiority is

rarely observed.

A desirable feature of the Rotterdam, is that all the

general restrictions of the demand functions can be built in

it, while the specific ones showing special preference

structures are special cases of it.

The use of the Rotterdam parametrization allows for

nested testing for homogeneity, symmetry, uniform substitutes

and strong separability hypotheses, unlike alternative

differential approach parametrizations (for example, the

Almost Ideal Demand System by Deaton and Muellbauer, 1980).

The Rotterdam parametrization has been criticized because

using it to test the homogeneity hypothesis often results in

its rejection. However, Laitinen (1978) showed that it is a

bias in the F test, rather a limitation of the model that

causes that, and demonstrated that if a corrected version of

the test is used, homogeneity is not rejected.

Another common criticism is integrability, the absence of

which in the aggregate implies that the model's theoretical

properties are not known. It is dealt with in conjunction

with the aggregation over consumers issue discussed above, in

Barnett (1979). He used a random coefficient model to perform








53

an aggregation analysis for an individual consumer. The

coefficients used to identify each consumer are not assumed

constant. Barnett's result is that the aggregate model (which

has constant parameters) is a Taylor series approximation to

a new theoretical construct, existing under conditions far

weaker than those necessary for aggregate integrability.

A Geographic Import Allocation model

While the differential approach's different

parametrizations have been used in various estimations of

consumer's demand models, its application in estimating import

demand has not been widely used. This is particularly true

for imports of the same type products which originate from and

can be differentiated by location of production.

Specifically, the Rotterdam parametrization for import demand

estimation of similar goods by country of origin has been used

by Theil and Clements (1978), Clements and Theil (1987), Lee

et al. (1990) and Seale et al. (1992). However, none of these

imposed the separability condition of uniform substitutes.

The advantages of this form of separability are many.

Firstly, the method recognizes the close similarity of the

same type product, imported from different sources. Further,

it is extremely parsimonious in its use of parameters, much

like the Armington model. However, unlike the Armington

model, which suffers from conceptual and theoretical problems,

the differential approach is based on a solid, theoretical

foundation (Alston et al.; Davis and Kruse). The model









54

developed is also globally separable unlike that of the

Deaton-Muellbauer model. The uniform substitutes restriction

on a differential demand system was developed by Theil (1980)

for estimating the demand for similar goods, such as brand

names of the same type of good. The hypothesis has been

empirically utilized only by Brown (1994), but not for import

allocation or for a conditional demand system. In this study,

an import demand model is developed of a similar type good

from different sources of origin in a conditional framework.

Thus, the functional form we develop is different from that of

Brown.

The concept of multistage budgeting (Barten (1969); Seale

et al.) was utilized. For the first stage, the separability

structure exhibited by Egyptian domestic goods and Egyptian

imported goods is assumed to be blockwise dependence. Then

domestic goods and imported goods are functions of total

expenditure and the relative price between these goods. Next,

imports expenditure is allocated among imported goods, and

finally, expenditure of a good is allocated among that good

based on country of origin. Thus, a good is differentiated

geographically by place of production.

Following Seale et al., a conditional geographic import

demand system can be developed, according to which, total

expenditure for good i, E,, is allocated among the countries

of origin (j=l, ..., n), so that E,, is the expenditure on good

i from country j. Thus, demand for i from the j countries can









55

be estimated conditionally on E,, the expenditure spent on

imported i.

The estimation of the conditional demand among countries

shows the effects on the conditional trade shares when the

consumption of good i changes due to a change in relative

prices of i among sources j, or when total expenditure on i

changes.

Let q,, and p, be quantities and prices of good i from

the j countries of origin, and let w,, w, be the import

(budget) shares of good i in S, (containing countries j from

which i is imported), and good i from country j, respectively.

Also, E, = E Ej, so that w, = E w,. Then 8, is the marginal
J c Si j c Si
share of good i from country j, and 8, = Z E 0,,, is the
j Sil k z Si
marginal import share of good i. 8,, is the additional amount

spent on good i from country j when total expenditure on all

j e S, increases by one dollar. The ratio of that to the

budget share (08,/w,,) is the expenditure elasticity of the ith

good from the jth country.

The conditional differential import demand for good i

from country j is, assuming weak separability among the good

from different sources,

w," d(log q,) = 0,, d(log Qj) + E r7k* d(log pik) (6)
k c Si
where wj is the conditional import share of good i from

source j (w,* = w,1/w,), 9O* is the conditional marginal import

share for good i from country j (8 = 8O,/8), ,,,*'s are the

conditional Slutsky price parameters, d(log Qj) = E wj
j C S1









56

d(log q,,) is the Divisia quantity index for S, and

d(log P,) = E w,, d(log p,,) is the (conditional) Divisia price
j z Si
index, where d(log X,) = dX,/X,, with X representing either Q

or P. The adding up, homogeneity and symmetry conditions are

E 8i, = 1, E rrijk = 0, and rr,, = nk', respectively.
j C S1 k z Si

The Rotterdam model's conditional absolute price version

(AP) is derived by assuming the 8,'s and r,,''s constant and

adding an error term, et


W ijt* Dqjt = O8, DQit + Z TTijk Dpit + Eajt (7)
k c Si

where w t,' = 1/2 (w,,.t./ + w,,t') and DXjt, = log (X,) -

log (X,,t_-), with X equal to q, p or Q. To estimate the

system, one equation is arbitrarily omitted, and the parameter

estimates remain invariant to the equation omitted (Barten,

1969). Estimating (7) with homogeneity and symmetry imposed

involves estimating n-l marginal shares and n(n-l)/2 price

parameters.

Under blockwise independence', the structure of the

Slutsky price parameters is

Ijk* = 1(ak 01i/06,) (8)

where Ok* = Ok/ws, and 4, = 0O,/W, is a factor of

proportionality specific for the group S,.






1 For the derivation of the blockwise dependence case,
see Seale, 1996.








57

The ,,,''s can be restricted by various assumptions to

make the model much more parsimonious in terms of degrees of

freedom. One such restriction is based on the separability

conditions due to uniform substitutes. Imports of wheat, for

example, from different sources are similar but differentiated

products, much like brand-name goods. If wheats from

different sources are uniformly substitutable, then the

marginal utility of a dollar spent on wheat by each source

declines by the same amount when an additional dollar is spent

on any other wheat by source (Theil, 1980, p. 200). It can be

shown that equation (6) for uniform substitutes (by

restrictions (8)) becomes (Theil, 1980, pp. 166, 200)

w," d(log q.,) = 0," d(log Qj) + 0 80,' d(log p,,/P,') (9)

where 0,* = 0,/(l-RO,), R is a positive value and d(log P,') =

E O8ik d(log pk) is the Frisch conditional price index (with
k c Si
conditional marginal shares as weights).

The finite version of the Uniform Substitutes (UB)

model becomes

w jt Dqj = 86,O DQt + 0,* Oi D(pjt,/Pt') + Ei (10)

Models (7) and (10) are nested and the likelihood ratio

test can be used to test the UB hypothesis for goods i from

different sources. Finally, it should be noted that when R=0,

the model is restricted to one of preference independence

(strong separability). (See Seale et al. for estimation of a








58

separable allocation model under the assumption of preference

independence.)















CHAPTER 5

DATA, ESTIMATION AND RESULTS

Data

For the estimation of the model, import data from the United

Nations (UN) was used. The description of the wheat and wheat

flour commodities was based on the Standard International Trade

Classification. Revision I. Wheat flour was assumed to be a

different commodity than wheat, for reasons explained in the

introductory chapters (also see estimation of Egyptian import

demand for flour, Pana-Cryan and Seale, 1995).

As mentioned in earlier sections, in this study aid is treated

as an income transfer. In the estimated model, amounts of aid to

Egypt by donor were subtracted from the UN data that include aid,

in order to compare imports by source net of aid shipments (Pana-

Cryan and Seale include aid amounts in the estimation). The aid

shipments reported in AGROSTAT show totals of wheat and wheat flour

in grain equivalents. To derive wheat aid only amounts, the

percentage of wheat imports to total wheat and flour imports (in

grain equivalents) was estimated for each year and country of

origin. Then this percentage was applied to the gross aid figures

in order to derive wheat aid only by country of origin.

Table 8 presents the commercial quantities of wheat imported

(net of aid) and their values by selected suppliers. Since the










Table 8. Commercial Quantities and Values of Imported Wheat by
Selected Suppliers, Egypt, 1973-93.


EU Australia


Year


1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993


1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993


---------- $1
17001 53754
171376 177772
222337 123707
181245 80244
168498 5505
173250 48860
152304 36650
220667 169481
438077 89787
472342 10798
242474 90064
139455 44482
69157 126
172118 0
172804 30085
202316 40447
390737 46163
223572 160394
214847 32564
478108 17061
156608 43939


Other


MT ---------------
552062 213614
720354 145728
833567 124823
707455 19451
506568 512770
365684 1022740
214855 340601
223057 2340
751029 143458
669303 148803
639186 104056
1022481 627398
1214326 367191
1765813 270308
1356735 123930
1113223 319457
608692 30957
1805461 375670
1851598 238239
1193638 2646
726502 208481


000-----
37178
196845
158641
121255
79317
46478
27567
47517
191173
174976
141453
238791
159402
281716
144236
183348
137172
333877
225188
200562
127908


31597
48712
39674
9748
75337
155543
32535
3605
39822
45082
22644
114222
41247
46028
35403
11900
11793
71245
34201
8926
7066


Sources: United Nations, Commodity Trade Statistics, FAO, AGROSTAT.


141067 396473
654604 608575
632983 589935
583149 378024
305835 27520
424632 311598
327812 363565
179338 752468
916948 358490
1030197 53245
764327 375265
182996 233463
147007 817
658978 0
1140399 126556
857484 429821
1635289 235859
721453 899246
558643 18204
3118350 50419
559603 280506








61

main interest of the study regards the position of the US in the

Egyptian import market, the years selected start in 1973 because

there were no US exports to Egypt during the previous war years.

In 1986 there were no imports from the EU, so that year was

excluded as well (the model estimated does not accept zero values).

The AP and UB versions of the conditional Rotterdam were

estimated using the seemingly unrelated regressions procedure in

Times Series Processor (TSP).

Testing Restrictions

Model (7) was estimated with no restrictions, with homogeneity

imposed and with symmetry imposed for wheat imported to Egypt by

country of origin. The logs of the concentrated likelihood

functions for these estimations are presented in Table 9. The

likelihood ratio test (LR) is minus twice the log ratio of the

restricted to the unrestricted concentrated likelihood function and

is asymptotically distributed as a chi-squared with degrees of

freedom equal to the number of restrictions. Homogeneity could not

be rejected when tested against the unrestricted model (a=0.01).

Also, symmetry was tested against the model with homogeneity

imposed, and again could not be rejected (a=0.01). Finally, the

uniform substitutes restriction was imposed and the calculated

statistic was 12.58, which is less than the critical value of the

chi-squared at a=0.01 (the hypothesis could not be rejected at this

level of significance).

It should be noted that this is a nested testing framework, so

that each nested test is conditional on the failure to reject the









62

Table 9. Log of Likelihood Functions for Conditional Rotterdam
Model Under Different Restrictions.


Restrictions Log Lik.

Free Rotterdam 46.857 (15)
Homogeneity 46.403 (12)
Symmetry 43.482 (9)
Uniform Substitutes 37.192 (5)

Note: In parentheses are the numbers of free parameters for each
model.








63

previous hypothesis. At the level of uniform substitutes, the

overall significance is approximately the sum of the significance

levels of all three tests.

Whereas the coefficients of determination R2 for the

individual equations are meaningless, testing for goodness of fit

for the whole system is possible using modified measures of R2. An

example is the one by McElroy (1977), that can be calculated

according to the formula R.2 = l-1/(l+W/(T-k)(n-1)), where W is the

Wald test statistic, T is the number of observations, k is the

number of regressors in each equation, and n is the number of

equations in the system (Bewley, p. 188). Taking into account all

the shortcomings of a measure like that (Greene, p. 513) and also

that the statistic will vary depending on which equation of the

system is dropped, the statistic was calculated with the rest of

the world (ROW) equation dropped because it was the least

interesting for the purpose of this study. The values for the

statistic are .93 and .98 for the symmetry AP and the UB version,

respectively.

Conditional Marginal Shares

The conditional marginal import shares show how an additional

dollar spent on imported good i is allocated among sources for that

good. The parameter estimates are reported in Table 10 for the

Rotterdam UB case, and Table A-1 in the Appendix for the AP case.




1 In the sections that follow the UB results will be
presented. A summary of the AP results will be discussed at the
end of the present chapter.










Table 10. Parameter Estimates for Egyptian Wheat Imports by
Sources, Rotterdam UB, 1973-85, 1987-93.


Exporting
Country


Slutsky Price Coefficients


Conditional Factor
Marginal of
Shares Propor-
tionality


US AUS
-0.309 0.185
(0.050) (0.056)
-0.246
(0.068)


EU
0.070
(0.024)
0.034
(0.014)
-0.115
(0.036)


ROW
0.053
(0.014)
0.026
(0.009)
0.010
(0.004)
-0.089
(0.023)


0.550 -1.248
(0.060) (0.194)
0.270
(0.068)
0.102
(0.037)
0.078
(0.024)


Note: In parentheses are asymptotic standard errors.


US

AUS

EU

ROW








65

In the UB version all estimates are positive and significantly

different from zero. According to that version, for an additional

dollar of wheat imported into Egypt, 55 cents would be spent on US

wheat, 27 cents on Australian and 10 cents on European wheat.

The hypothesis of equality among the conditional marginal

shares for US, Australian and European wheat imported to Egypt is

rejected (a=0.05).

Expenditure Elasticities

The conditional expenditure elasticities are calculated by

dividing the conditional marginal import shares by the mean of the

average import shares. They show how a 1% change in expenditure

for a good would affect the percentage change in quantities of the

good imported from each country. The results are reported in

columns one in Tables 11 and A-2 for the UB and AP versions of the

model, respectively.

In the UB case, all results are significantly different from

zero. According to that version of the model, a 1% expansion in

the Egyptian imported wheat market will cause an increase in wheat

quantities imported from the US, while the import shares from other

sources will decline. However, the hypothesis of equality among

the expenditure elasticities for wheat imported to Egypt from

different sources could not be rejected.

Price Parameters

The conditional Slutsky price parameters for imported wheat by

country of origin are reported in columns one through four in

Tables 10 and A-1 for the UB and AP versions, respectively. The








66

factor of proportionality was used to calculate the UB version

coefficients. Along the diagonals are the own-price estimates and

the off-diagonals suggest substitution if positive and

complementarity if negative according to the Hicksian definitions.

In the UB case the own-price estimates are negative as expected.

They are also significant. The hypothesis of equality among own-

price parameters for wheat from different sources is rejected.

The cross price results are all positive and significantly

different from zero for the UB case. Therefore, the results imply

that wheat imports from different sources are substitutes.

Price Elasticities

Three kinds of price elasticities were calculated for the AP

and the UB versions of the model, namely, Slutsky, Cournot and

Frisch.

The conditional Slutsky price elasticities are calculated by

dividing the estimated Slutsky price parameters by the mean of the

average import shares and are reported in Tables 11 and A-2. They

show the percentage change in quantities demanded that result from

a 1% change in price, holding real expenditures on imported wheat

constant. They correspond to a Hicksian demand curve and are all

significant for the UB version. The hypothesis of equality among

the conditional Slutsky price elasticities for wheat from different

sources of origin could not be rejected.

The conditional Cournot price elasticities are calculated


using,











Table 11. Egyptian Expenditure and Slutsky Price Elasticities of
Import Demand for Wheat by Source, Estimated at Sample Means,
Rotterdam UB.


Exporting
Country


US

AUS

EU

ROW


Expenditure
Elasticities


1.201
(0.130)
0.880
(0.221)
0.750
(0.270)
0.787
(0.244)


Slutsky Price Elasticities


US
-0.674
(0.110)
0.604
(0.183)
0.515
(0.173)
0.540
(0.145)


AUS
0.404
(0.122)
-0.801
(0.220)
0.252
(0.100)
0.265
(0.094)


EU
0.153
(0.052)
0.112
(0.044)
-0.840
(0.268)
0.100
(0.046)


ROW
0.116
(0.031)
0.085
(0.030)
0.073
(0.033)
-0.905
(0.228)


Note: In parentheses are asymptotic standard errors.











C1j = (j 0* w )/ w



where n7, the conditional Slutsky coefficients, w and w a


average conditional budget shares and 8,* conditional marginal

share. These elasticities express substitution and income effects

resulting from price changes, holding nominal income constant, and

correspond to a Marshallian demand. Tables 12 and A-3 present the

results for the UB and AP versions, respectively.

According to the UB version, the Cournot own-price

elasticities are negative and significantly different from zero.

However, the hypothesis of equality among the own-price

elasticities for different suppliers could not be rejected.

Frisch price elasticities are calculated according to the

formula F,, = (rr + r 8O, 68)/ w i, holding the marginal utility


of income constant. The results are presented in Tables 13 and A-4

for the UB and AP versions of the model, respectively.

The cross-price elasticities vanish for the UB case. The own-

price elasticities are all negative and significant, but the

hypothesis of equality among them could not be rejected.

Summary of AP Results

The Rotterdam AP version conditional marginal shares estimates

are positive. They are also significantly different from zero,

except for the EU estimate. Compared to the UB case, they are











Table 12. Egyptian Cournot Price Elasticities of Import Demand for
Wheat by Source, Estimated at Sample Means, Rotterdam UB.


Exporting
Country


US

AUS

EU

ROW


Cournot Price Elasticities


US
-1.224
(0.102)
0.201
(0.112)
0.171
(0.084)
0.180
(0.077)


AUS
0.036
(0.143)
-1.071
(0.282)
0.022
(0.088)
0.024
(0.092)


EU
-0.012
(0.055)
-0.008
(0.040)
-0.942
(0.301)
0.007
(0.036)


ROW
-0.002
(0.032)
-0.002
(0.023)
-0.001
(0.020)
-0.983
(0.249)


Note: In parentheses are asymptotic standard errors.











Table 13. Egyptian Frisch Price Elasticities of Import Demand for
Wheat by Source, Estimated at Sample Means, Rotterdam UB.


Exporting
Country


US

AUS


Frisch Price Elasticities


US
-1.498
(0.245)


AUS


-1.098
(0.394)


ROW


EU




-0.936
(0.332)


Note: In parentheses are asymptotic standard errors.


ROW






-0.981
(0.269)








71

lower for all sources of origin except the US, for which the AP

estimate is higher than the UB one.

The expenditure elasticities estimates are significantly

different from zero, except for the EU estimate. With respect to

the magnitude of the estimates, the AP estimates are lower than the

UB ones, except for the US case, where the reverse is true.

The own-price estimates are all negative and significant. The

AP case includes a negative cross-price estimate (AUS-ROW), but it

is not significantly different from zero.

The own-price Slutsky elasticities are negative and

significant, while the cross-price estimates are positive except

for Australia and the ROW, for which they are negative but

insignificant.

The Cournot own-price elasticities are negative and

significantly different from zero. They are also similar in the

two models.

The Frisch own-price elasticities are higher for all suppliers

except the US, and the cross-price estimates are all negative

except the ones including the US.














CHAPTER 6

SUMMARY, CONCLUSIONS, LIMITATIONS AND SUGGESTIONS
FOR FURTHER RESEARCH

Summary and Conclusions

The geographic wheat import allocation model for Egypt was

utilized to assess the US position in this market compared to other

major Egyptian suppliers, namely the EU and Australia. Testing

among alternative hypotheses, the UB version of the model could not

be rejected. The UB hypothesis implies that when an additional

dollar is spent on wheat by one source, the marginal utility of a

dollar spent on wheat by each other source declines by the same

amount.

When an additional dollar of wheat imported into Egypt is

spent, the conditional marginal share estimates suggest that 55

cents of it would be spent on US wheat, 27 cents on European wheat

and 10 cents on Australian wheat. The negative cross price

parameters imply that Egyptian wheat imports from different sources

are substitutes.

Domestic and export policies in the wheat markets for the

Egyptian suppliers can potentially affect Egyptian wheat imports.

Australia is considered a small country and policy changes in its

wheat market do not affect the Egyptian imports. The EU and the US

policies affect the wheat world price however, which in turn causes

changes in the quantities of world wheat exports and ultimately,

72








73

Egyptian wheat imports. For example, following the analysis in

chapter 3, if the US eliminates deficiency payments to its wheat

producers, the world price for wheat will increase. The estimated

model however, assumes four wheat exporting areas and a single

import market which is only a portion of the wheat traded on the

world market. One interpretation of the resulting "world price" in

chapter 3 resulting from the removal of the US deficiency payment

program is simply the "offering price" by the US, assuming no

policy changes by other exporters'.

The estimated model predicts what would happen in the Egyptian

wheat import market when the offering wheat price from an

individual supplier changes, assuming the rest of the suppliers'

prices remain constant. According to Table 11, a 1% increase in

the US wheat offering price resulting from such a policy change as

above will reduce the Egyptian wheat quantity imported from the US

by 0.67%, while the EU wheat quantity imported to Egypt will

increase by 0.52% and the Australian wheat quantity imported to

Egypt will increase by 0.60%.

An alternative scenario would be the removal of export

subsidies by the EU, resulting in an increase in their offer price

to the world market. A 1% increase in the EU offering wheat price,

for example, will decrease wheat imported to Egypt from the EU by

0.84%, while the US wheat imports to Egypt will increase by 0.15%

and the Australian imports to Egypt by 0.11%.


1 This abstracts from the other mixture of programs affecting
the US grain market for the sake of illustrating the application of
the estimated model.








74

The assumption that a policy change by one supplier will not

cause a change in other suppliers' policies refers to the short

run. In the long run competitors policies will change as well, but

the model does not examine that scenario.

If the Egyptian government did not set the wheat prices for

consumers and producers, the domestic supply would also increase,

and the consumers would face higher wheat prices. While the

quantity of imports would decrease, the elimination of price

interventions in Egypt would not affect the wheat world price.

Limitations and Suggestions for Further Research

The limitations of the study that are mentioned refer to the

Rotterdam model's choice and the assumption of Egypt being a small

country. While these are considered the most important

limitations, there exist certainly more, as is the case in most

modelling efforts.

The advantages and disadvantages of the Rotterdam model were

explored in the theoretical and methodological sections. Its

desirable properties as a geographical allocation model were the

main reason it was chosen for this study. Some of the model's

limitations include its inability to incorporate ad hoc

explanations, the fact that a good can change from a luxury to a

necessity, but it cannot change from a non-inferior to an inferior

good, and the model's logarithmic formulation, which means it does

not accept zero data values.

Another limitation of the study refers to the assumption that

Egypt is a small country for trade purposes. While it was assumed








75

that Egypt can not affect the world price for wheat, an argument

can be made against this point. Not only is Egypt the third

largest wheat importer, political reasons might provide it with

additional leverage to negotiate and receive lower wheat imports

price than what the small country case would predict. More in

depth research on whether or not the small country assumption is

reasonable is needed.

The models estimated are conditional on expenditure allocated

for wheat imports. If relevant data were available one could

calculate unconditional expenditure elasticities, that would show

what happens to the wheat import market with changes in total

expenditure. For each stage (stages are described in the

development of the geographic allocation model) one could obtain

expenditure elasticities that if multiplied would result in the

unconditional ones. For example, the conditional expenditure

elasticity for US wheat imported to Egypt multiplied by the

expenditure elasticity for all imported wheat would be the

unconditional elasticity for US wheat imported to Egypt.















APPENDIX


ROTTERDAM AP RESULTS


Table A-i. Parameter Estimates for Egyptian Wheat Imports by
Sources, Rotterdam AP, 1973-85, 1987-93.


Conditional


Exporting
Country


US

AUS


Slutsky Price Coefficients


AUS
0.257
(0.055)
-0.257
(0.059)


EU
0.087
(0.040)
0.021
(0.024)
-0.117
(0.043)


US
-0.434
(0.066)


ROW
0.090
(0.023)
-0.021
(0.019)
0.008
(0.023)
-0.077
(0.022)


Marginal
Shares


0.344
(0.089)
0.353
(0.067)
0.134
(0.080)
0.169
(0.061)


Note: In parentheses are asymptotic standard errors.


ROW










Table A-2. Egyptian Expenditure and Slutsky Price
Elasticities of Import Demand for Wheat by Source, Estimated
at Sample Means, Rotterdam AP.


Exporting
Country


US

AUS

EU

ROW


Expenditure
Elasticities


0.751
(0.195)
1.152
(0.219)
0.979
(0.586)
1.713
(0.620)


Slutsky Price Elasticities


US
-0.947
(0.145)
0.560
(0.119)
0.191
(0.088)
0.196
(0.050)


AUS
0.837
(0.178)
-0.838
(0.192)
0.068
(0.080)
-0.068
(0.061)


EU
0.640
(0.295)
0.153
(0.179)
-0.854
(0.313)
0.061
(0.167)


ROW
0.910
(0.234)
-0.210
(0.190)
0.084
(0.231)
-0.784
(0.223)


Note: In parentheses are asymptotic standard errors.











Table A-3. Egyptian Cournot Price Elasticities of Import
Demand for Wheat by Source, Estimated at Sample Means,
Rotterdam AP.


Exporting
Country


US

AUS

EU

ROW


Cournot Price Elasticities


US
-1.291
(0.116)
0.309
(0.129)
0.192
(0.337)
0.126
(0.315)


AUS
0.330
(0.157)
-1.191
(0.240)
-0.147
(0.265)
-0.735
(0.266)


EU
0.088
(0.098)
-0.089
(0.086)
-0.988
(0.332)
-0.173
(0.188)


ROW
0.122
(0.057)
-0.181
(0.065)
-0.036
(0.187)
-0.953
(0.248)


Note: In parentheses are asymptotic standard errors.











Table A-4. Egyptian Frisch Price Elasticities of Import
Demand for Wheat by Source, Estimated at Sample Means,
Rotterdam AP.


Exporting
Country


US

AUS

EU

ROW


Frisch Price Elasticities


US
-1.269
(0.050)
0.343
(0.077)
0.220
(0.065)
0.175
(0.114)


AUS
0.229
(0.051)
-1.345
(0.079)
-0.278
(0.067)
-0.965
(0.117)


EU
0.065
(0.019)
-0.124
(0.030)
-1.017
(0.025)
-0.202
(0.044)


Note: In parentheses are asymptotic standard errors.


ROW
0.038
(0.025)
-0.311
(0.038)
-0.311
(0.038)
-1.146
(0.056)















REFERENCES


Alderman, H. and J. von Braun. The Effects of the Egyptian
Food Ration and Subsidy System on Income Distribution and
Consumption. Research Report 45, International Food Policy
Research Institute, Washington D.C., 1984.

Alston, J. M., C. A. Carter, R. Green, and D. Pick. "Whither
Armington Trade Models?". American Journal of Agricultural
Economics 72 (May 1990), pp. 455-67.

Antle, J. "Agriculture in the National Economy". In The
Agriculture of Egypt ed., G. M. Craig. Oxford University
Press, New York, 1993.

Barnett, W. A. "Theoretical Foundations for the Rotterdam
Model". Review of Economic Studies 46(1) (1979), pp. 109-30.

Barten, A. P. "Maximum Likelihood Estimation of a Complete
System of Demand Equations." European Economic Review 1
(1969), pp. 7-73.

Barten, A. P. "The Systems of Consumer Demand Functions
Approach: A Review." Econometrica 45 (1977), pp. 23-51.

Barten, A. P. "Consumer Allocation Models: Choice of
Functional Form." Empirical Economics 18 (1993), pp. 129-58.

Bewley, R. Allocation Models: Specification, Estimation and
Application. Ballinger Publishing Co., Cambridge,
Massachusetts, 1986.

Brenton, P. A. "The Allocation Approach to Trade Modelling:
Some Tests of Separability Between Imports and Domestic
Production and Between Different Imported Commodity Groups".
Weltwirtschaftliches Archiv 125(2) (1989), pp. 230-51.

Brown, A. and A. Deaton. "Surveys in Applied Economics:
Models of Consumer Behavior". Economic Journal 82 (1972),
pp. 1145-1236.

Brown, M. G. "Demand Systems for Competing Commodities: An
Application of the Uniform Substitute Hypothesis." Florida
Agricultural Experiment Station Journal Series No. R-03158,
1994.











Clements, K. W. "Alternative Approaches to Consumption
Theory." In Applied Demand Analysis by H. Theil, Ballinger
Publishing Company, Cambridge, Massachusetts, 1987.

Clements, K. W. and H. Theil. "A Simple Method of Estimating
Price Elasticities in International Trade." Economics Letters
1 (1987), pp. 133-137.

Davis, G. C. and N. C. Kruse. "Consistent Estimation of
Armington Demand Models." American Journal of Agricultural
Economics 75 (August 1993), pp. 719-723.

Deaton, A. and J. Muellbauer. "An Almost Ideal Demand
System." American Economic Review 70 (1980), pp. 312-26.

Dethier, J. J. Trade, Exchange Rate. and Agricultural Pricing
Policies in Egypt. World Bank, Washington D.C., 1989.

Europa World Yearbook 1989. Europa Publications Limited,
London, 1989.

ez Elarab, M. "An Estimation of a Complete Demand System for
Egyptian Imports Under a Separable Utility Function." Ph.D.
Thesis, North Carolina State University, 1982.

Foreign Agricultural Service of the United States.
Agricultural Statistical Annual Report: Egypt Report 1994,
Washington D.C., 1995.

Food and Agricultural Organization of the United Nations.
AGROSTAT (Computer Files) Rome, 1994.

Greene, W. H. Econometric Analysis. Macmillan Publishing
Co., New York, 1990.

Handoussa, H. "The Impact of Foreign Aid on Egypt's Economic
Development, 1952-1986." In Transitions in Development: The
Role of Aid and Commercial Flows ed., U. Lele and I. Nabi,
International Center for Economic Growth, San Francisco,
California, 1991.

Hansen, B. The Political Economy of Poverty, Equity. and
Growth: Egypt and Turkey. Oxford University Press for the
World Bank, Oxford, 1991.

Hicks, J. R. Value and Capital. Second ed., Oxford
University Press, Oxford, 1946.

Laitinen, K. "Why is Demand Homogeneity so Often Rejected?"
Economics Letters 1 (1978), pp. 187-191.










Lee, J.-Y., J. L. Seale, Jr., and P. A. Jierwiriyapant. "Do
Trade Agreements Help U.S. Exports? A Study of the Japanese
Citrus Industry." Agribusiness 6 (1990), pp. 505-14.

McElroy, M. B. "Goodness-of-Fit for Seemingly Unrelated
Regressions." Journal of Econometrics 6 (1977), pp. 381-87.

Moschini, G., D. Moro and R. D. Greene. "Maintaining and
Testing Separability in Demand Systems." American Journal of
Agricultural Economics 76 (February 1994), pp. 61-73.

Organization for Economic Cooperation and Development.
Agricultural Policies, Markets and Trade in OECD Countries:
Monitoring and Outlook 1995. Author, Paris, 1995.

Pana-Cryan, R. and J. L. Seale, Jr. "A Geographic Allocation
Model for Egyptian Import Demand for Wheat, Corn, and Flour."
Paper presented at the annual meetings of the American
Agricultural Economics Association, San Antonio, Texas, 1996.

Parker, J. and S. Shapouri. Determinants of Wheat Import
Demand: Egypt. United States Department of Agriculture,
Agriculture and Trade Analysis Division, Commodity Economics
Division, Washington, D.C., November 1993.

Phlips, L. Applied Consumption Analysis. North-Holland,
Amsterdam, 1983.

Powell, A. A. Empirical Analytics of Demand Systems. D. C.
Heath and Co., Lexington, Massachusetts, 1974.

Robinson, S. and C. Gehlhar. Land Water and Agriculture in
Egypt: The Economywide Impact of Policy Reform. Trade and
Macroeconomics Division, International Food Policy Research
Institute, Washington D.C., September 1994.

Sarris, A. "Structural Adjustment and Agricultural
Development in Egypt: Policies, Prospectives and Options." In
Agricultural Policy Analysis in Egypt:Selected Papers ed., F.
K. Bishay, S. Nassar and D. Abdou. Food and Agricultural
Organization of the United Nations Economic and Social
Development Paper 129, Food and Agricultural Organization of
the United Nations, Rome, 1994.

Scobie, G. M. "Government Policy and Food Imports." Research
Report 29, International Food Policy Research Institute,
Washington D.C., 1981.

Scobie, G. M. "Food Subsidies in Egypt: Their Impact on
Foreign Exchange and Trade." Research Report 40,
International Food Policy Research Institute, Washington D.C.,
1983.










Seale, J. L. Jr. "Uniform Substitutes and Import Demand for
Products Differentiated by Source." Paper presented at the
University of California, Davis, January, 1996.

Seale, J. L. Jr., A. L. Sparks and B. M. Buxton. "A Rotterdam
Application to International Trade in Fresh Apples: A
Differential Approach." Journal of Agricultural and Resource
Economics, 17(1) (1992), pp. 138-149.

Theil, H. The System-Wide Approach to Microeconomics. The
University of Chicago Press, Chicago, 1980.

Theil, H. and K. W. Clements. "A Differential Approach to US
Import Demand." Economics Letters 1 (1978), pp. 249-252.

Truett, L. J., D. B. Truett and B. E. Apostolakis. "The
Translog Cost Function and Import Demand: The Case of Mexico".
Southern Economic Journal 60(3) January 1994, pp. 685-700.

TSP International. Time Series Processor (Computer Software),
1993.

United Nations. Commodity Trade Statistics (Computer Files),
1994.

United Nations. Standard International Trade Classification.
Revision I. Author, New York, 1963.

United States Department of Agriculture. Foreign Agricultural
Trade of the United States. January/February 1994. Author,
Washington D.C., 1994.

United States Department of Agriculture. Foreign Agricultural
Trade of the United States. January/February 1996. Author,
Washington D.C., 1996.

von Braun, J. and H. de Haen. "The Effects of Food Price and
Subsidy Policies on Egyptian Agriculture." Research Report
42, International Food Policy Research Institute, Washington
D.C., 1983.

Winters, L. A. "Separability and the Specification of Foreign
Trade Functions". Journal of International Economics 17
(1984), pp. 239-263.

World Bank. Arab Republic of Egypt: An Agricultural Strategy
for the 1990s. Country Study, Author, Washington D.C., 1993.

World Bank. STARS (Computer files), 1994.

World Bank. World Tables, 1993-94 ed. Johns Hopkins
University Press, Baltimore, 1994.















BIOGRAPHICAL SKETCH

Regina Pana-Cryan was born in 1961 in Athens, Greece.

After graduating from high school, she attended the American

College of Greece in Athens, were she studied business

administration, and received a B.S. in accounting and finance

in December 1987. While enrolled in the American College of

Greece, she attended the University of Thessaloniki in

Thessaloniki, Greece, where she studied agriculture and

received a BS in plant science in February 1989. In August of

the same year she enrolled at the University of Florida, Food

and Resource Economics Department. She graduated with an MS

in December 1991 and enrolled in the Ph.D. program in the same

department. She is currently a prevention effectiveness

fellow at the Centers for Disease Control and Prevention,

National Institute for Occupational Safety and Health, in

Atlanta, Georgia.








I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and luity, as
a dissertation for the degree octp- of Phi s .



R bert D. Emerson, Chair
Professor of Food and Resource
Economics




I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.



Jaies L. Seale, Jr., Cocnair
Professor of Food and Resource
Economics




I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.



Max R. Langham
Professor Emeritus of Food and
Resource Economics




I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.



Elias opoulos
ProfeSsor of Economics








This dissertation was submitted to the Graduate Faculty
of the College of Agriculture and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.



December 1996
Dean, College of agriculture


Dean, Graduate School



















LD
1780
1996








UNIVERSITY OF FLORIDA
B l ll I I l l tltl 7ll Ill l1111111
3 1262 08555 0647




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EX8SCVZ8B_W58D9C INGEST_TIME 2013-10-09T22:04:38Z PACKAGE AA00017664_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES