Previous Page Table of Contents Next Page


4. DEMAND CHARACTERISTICS AND DIETARY PATTERNS

4.1 The place of fish in the food basket

Fish is preferred over meat in almost every household in Bangladesh. 'A fish diet is an integrated part of all festive occasions for the members of the Hindu community. Amongst Muslims, sending of large sized fishes to the bride's residence along with sweets, dresses, etc. at the advent of wedding is an essential part of the wedding ceremony' (Ali, 1991, p. 19). Fish is nutritionally equivalent to meat in protein, high in essential minerals and low in saturated fats (Islam and Dewan, 1987). The place of fish in the food basket and desirable levels of intake are recorded in Table 12.

Table 12. Dietary pattern in Bangladesh compared to desirable level

Food

Current intake

Desirable

Intake (g)

Energy (Kcal)

% of total energy

Intake (g)

Energy (Kcal)

% of total energy

i) Cereals

465

1607

82.4

372

1281

55


- Rice

423

1464

75.1

312

1076

46

- Wheat

42

143

7.3

60

205

9

ii) Potato

42

37

1.9

130

115

5

iii) Pulses

14

49

2.5

66

231

10

iv) Animal Food

60

55

2.8

126

115

5


- Fish

24

24

1.2

50

50

2.2

- Meat

10

11

0.6

22

24

1.0

- Egg

3

5

0.3

7

11

0.5

- Milk

23

15

0.7

47

30

1.3

v) Fruits

34

34

0.7

57

57

2.5

vi) Vegetables

57

25

1.3

132

57

2.5

vii) Added oil

7

63

3.2

38

342

15

viii) Sweeteners:

20

80

4.1

28

112

5


- Sugar

6

24

1.2

8

32

1.4

- Gur

12

48

2.5

17

68

2.9

- Molasses

2

8

0.4

3

12

0.7

Total

699

1950

100

949

2310

100

SOURCE: Hyder and Husain, The Bangladesh Observer, 25 Sept. 1999, p.5
Table 12 shows that food intake on average is 74% (699 g) of the recommended quantity (949 g). In terms of energy, actual intake is 84% of what a person needs. Share of intake from cereals is much higher than the desirable level (465 g versus 372 g). Recommended intake from animal sources is 126 g (it should cover 5% of total energy requirement), while actual intake is 60 g (less than half of the required quantity). Fish consumption more than doubled from 24 g[15] to 50 g daily. Of total energy required (2310 kcal), desirable energy share from fish is 50 kcal (2.2%). Current fish consumption provides only 24 kcal of energy (meeting only 1.2% of the required energy). Thus, the consumption of fish has to increase. We see a large market potential for further growth in the supply and production of fish to meet domestic requirements for protein. The demand for fish is relatively stable throughout the year, but supply fluctuates due to the biological nature of the product and seasonal changes in production. Data from the Household Expenditure Survey (Table 13) reveal that during 1995-96, per caput daily food intake was 914 g, which was an increase from the 886 g in 1991-92. Fish consumption amounted to 44 g (1995/96), up from the 35 g in 1991/92. These figures indicate that fish has become significantly more available (27% increase in per caput consumption) within a period of five years. Consumption of fish in rural areas increased substantially by 130%, and in urban areas it expanded by 108%. Availability and access to fish rose in rural areas. Increased supply should also have reached urban areas. However, although higher income affects consumption of fish positively, consumers in urban areas may benefit less from higher incomes than rural consumers[16]. In urban areas, per caput daily intake of fish was 52 g in 1995-96 (more than the required amount), whereas intake of fish was 42 g in rural areas. As expected, as income increased, consumption of cereals decreased more in urban areas than in rural areas. Demand for meat, poultry and eggs increased much more in urban areas than in rural areas (Table 13) as people's income increased.

Protein intake in terms of different food items (g) is shown in Tables 4.3 and 4.4 (in percentage terms). Nationally, protein intake per caput in 1995/96 was 66 g, whereas protein intake from fish amounted to 14% of the total (9.15 g). Protein share from fish in 1991-92 was 11% of total (6.78 g). Between 1995/96 and 1991/92, increase of protein intake was 35% (for the same period, advances in the consumption of milk and milk products were also substantial). The share of fish protein consumed in urban areas in 1995/96 amounted to 15% of the total (10.68 g) and in rural areas to 14% (8.85 g out of a total of 65.38 g in rural areas).

Table 13. Average per caput daily intake (g) by food item

Food groups

National

% change

Rural

% change

Urban

% change

1991-92

1995-96

1991-92

1995-96

1991-92

1995-96

Cereals

516.2

508.7

-1.45

523.3

521.9

-0.27

470.7

442.1

-6.08

Potato

43.7

49.5

13.27

41.4

46.7

12.80

58.3

64.4

10.46

Vegetables

137.4

152.6

11.06

135.3

154.4

14.12

150.9

142.9

-5.30

Pulses

17.9

14.0

-21.79

17.3

12.9

-25.43

21.7

19.4

-10.60

Milk/milk products

19.1

32.3

69.11

18.5

30.3

63.78

23.2

42.1

81.46

Meat/poultry/eggs

12.9

15.1

17.05

11.8

12.0

1.69

20.3

30.0

47.78

Fish

34.5

43.8

26.96

32.5

42.2

29.85

47.8

51.7

8.16

Spices (onion, chillies others)

43.5

37.2

-14.48

42.4

35.5

-16.27

50.6

45.4

-10.28

Edible oils

10.1

9.9

-1.98

9.1

8.4

-7.96

16.4

17.0

3.66

Fruits

16.9

27.6

63.31

15.9

25.3

59.12

23.4

38.8

65.81

Sugar

8.8

9.3

5.68

8.6

9.1

5.81

10.8

10.1

-6.48

Miscellaneous*

25.2

13.8

-45.24

22.0

11.8

-46.36

44.3

26.6

-39.95

Total

886.2

913.8

3.11

878.1

910.5

3.69

938.4

930.8

-0.80

NOTE: * Includes tea, soft drinks, biscuits, betelnut and betel leaf etc.

SOURCE: Household Expenditure Survey 1995-96, BBS, 1998, p. 46

Of total daily protein intake after cereals (59%), fish provided 14% of protein. Fish was the second highest provider of protein in the country in 1995-96. In urban areas, fish provided 15% of all protein intake (second highest provider of protein) during the same period. Share of fish protein has been higher than that supplied by livestock and poultry products together, both in urban and rural areas. Thus, fish is the second highest protein provider in the country after cereals. (Table 15).

4.2 Consumption preference: freshwater vs. marine fish

Most people in the country prefer freshwater fish. The taste for freshwater fish can possibly be attributed to the abundance of freshwater fish in rivers, beels, flood plains, haors, etc., which are easily accessible. At one point in time, almost the entire fish production in the country consisted of freshwater fish. Rohu is the preferred of all carps among consumers with higher income. Among marine fishes, only hilsa is widely popular. Consumers prefer hilsa from rivers caught during spawning season when they migrate upstream. In fact, hilsa is perceived to be a freshwater fish. In inland markets, traders pass hilsa from marine waters as fish caught in the river Padma (Ali, 1991 p. 33).

Most of the 260 species of freshwater fish and 20 species of freshwater prawns are greatly favoured by consumers. Of the major carps, rohu, catla and mrigal, particularly from rivers or other water bodies, command the highest prices. Generally, consumers always prefer bigger fish from rivers or open water bodies. These fish fetch higher prices in the market than cultured fishes from ponds. Carps, whether they are from open water bodies or from ponds, can easily be identified by their bright colour and their texture.

Consumption of marine fishes is largely confined to coastal communities in areas like Chittagong, Noakhali and coastal islands. In recent years, due to improved transportation and promotional activities by the Bangladesh Fisheries Development Corporation (BFDC), marine fishes have become more acceptable to consumers in Dhaka, Khulna and Chittagong cities. Consumers in Chittagong city and in rural areas are more accustomed to marine fishes than people in other areas.

Table 14. Per caput daily protein intake (g) by food items during 1991-92 and 1995-96

Food group

National

% change

Rural

% change

Urban

% change

1991-92

1995-96

1991-92

1995-96

1991-92

1995-96

Cereals

39.78

38.63

-2.89

40.23

39.53

-1.74

36.90

34.08

-7.64

Potato

1.33

1.48

11.28

1.26

1.40

11.11

1.77

1.90

7.34

Vegetables

3.28

4.49

36.89

3.18

4.49

41.19

3.88

4.46

14.95

Pulses

4.74

3.56

-24.89

4.61

3.31

-28.20

5.56

4.84

-12.95

Meat/Poultry/Eggs

2.51

3.08

22.71

2.27

2.45

7.93

4.10

6.26

52.68

Fish

6.78

9.15

34.96

6.52

8.85

35.74

8.49

10.68

25.79

Milk/Milk Products

0.84

2.08

147.62

0.77

1.90

146.75

1.30

3.00

130.77

Spices

1.61

1.23

-23.60

1.59

1.21

-23.90

1.71

1.36

-20.47

Edible oils

1.03

1.13

9.71

1.09

1.24

13.76

0.64

0.61

-4.69

Fruits

0.43

0.65

51.16

0.40

0.57

42.50

0.63

1.04

65.08

Sugar/Brown Sugar

0.01

0.01

0.00

0.01

0.01

0.00

0.01

0.01

0.00

Miscellaneous*

0.38

0.52

36.84

0.36

0.43

19.44

0.50

0.96

92.00

Total

62.72

66.01

5.24

62.29

65.38

4.96

65.49

69.19

5.65

NOTE: * Includes tea, soft drinks, biscuits, betelnut and betel leaf, etc.

SOURCE: Computed from BBS 1998, p. 50

Table 15. Percentage distribution of protein intake by food items during 1991-92 and 1995-96

Food groups

National

Rural

Urban

1991-92

1995-96

1991-92

1995-96

1991-92

1995-96

Cereals

63.42

58.52

64.58

60.46

56.35

49.25

Potato

2.12

2.24

2.02

2.14

2.70

2.75

Vegetables

5.23

6.80

5.11

6.90

5.92

6.45

Pulses

7.56

5.40

7.40

5.06

8.49

6.99

Meat/Poultry/Egg

4.00

4.67

3.64

3.74

6.25

9.05

Fish

10.81

13.86

10.47

13.53

12.96

15.43

Spices

2.57

1.86

2.55

1.86

2.61

1.97

Fruits

0.68

0.98

0.64

0.87

0.63

1.50

Sugar

0.02

0.02

0.02

0.00

0.05

0.01

Milk/Milk Products

1.34

3.15

1.24

2.90

1.99

4.34

Edible oils

1.64

1.71

1.75

1.89

0.98

0.88

Miscellaneous*

0.79

0.61

0.58

0.65

0.76

1.39

Total

100.00

100.00

100.00

100.00

100.00

100.00

NOTE: * Includes tea, soft drinks, biscuits, betelnut and betel leaf, etc.

SOURCE: Computed from BBS 1998, p. 50

Consumption of marine fishes is confined to marine species like pomfret, Indian salmon, snappers and jew fish. All marine species, except pomfrets and hilsa, command relatively low prices.

Consumers of all income groups always prefer fresh fish to iced or frozen fish. Iced fish is preferred to frozen fish. BFDC once transported frozen sea hilsa to internal markets for sale during the lean season. This practice was not profitable, since consumers showed little acceptance of frozen fish or fish that was previously frozen (Coulter and Disney, 1987).

4.3 Consumption preference: pond fish vs. fish from open water bodies

A survey on three primary retail fish markets in Mymensingh town and its periphery (Mymensingh Bara Bazar, Machhua Bazar and Sutiakhali village market) revealed the following price differences per kg of fresh fish from open water bodies and pond fish available in the markets (Table 16).

Table 16. Average price levels of different species (Mymensingh markets during last week of September, 1999) by sources (cultured and captured)

Species

Pond (kg/Tk)

Open water bodies (river)

difference in price per kg

% difference

Rohu

110

130

20

18

Catla

100

115

15

15

Mrigal

90

100

10

11

Pangas

90

140

50

56

Silver carp

90

100

10

11

Grass carp

70

80

10

14

Small fish*

70

85

15

21

SOURCE: Field survey, 1999.

NOTE: For carps, price differences will be higher as size increases.

* Small fish include indigenous species like Mola, Ketchki, Tengra, Pabda, Bachha, Kajuli, etc.

Among the high-valued fish, the price difference between captured (from open water bodies) and cultured fishes is highest for pangas. Price difference for small, fresh fish (low-value fish), captured or cultured (pond fish), are also higher than for highly-value fish like rohu and catla. Small, fresh fish are more preferred by rural and urban low- and medium-income families than by richer families. Per kg price difference between small, fresh fish from cultured and those from captured sources exceeds 20%.

4.4 Fish preferences by species

Respondents' (116 from 43 villages) preferences for certain fish species under different economic conditions are depicted in Table 17. As their first choice, 55% of consumers reported a preference for small, fresh fish in the market, 19% chose live fish, followed by 11% of consumers who made rohu their fish of choice, and 7% liked mrigal best, if available.

Table 17. Fish preference in rural areas

Fish species and kind

Preference order

Pref. 1

Pref. 2

No.

%

No.

%

Small fresh fish

64

55.17

21

18.10

Live fish

22

18.97

31

26.72

Rohu

13

11.21

27

23.28

Mrigala

8

6.90

4

3.45

Catla

2

1.72

13

11.21

Silver carp

2

1.72

5

4.31

Aire

3

2.59

4

3.45

Snakehead

-

-

2

1.72

Chital

2

1.72

9

7.76

Total

116

100.00

116

100.00

SOURCE: Field Survey, 1999.
As their second choice, 27% of consumers picked live fish, 23% rohu and 11% catla. Consumers liked snakehead fish the least. Demand for fresh, small fish is high, both in rural and urban areas, particularly by medium- and lower-income groups. Small fish can be bought in small quantities (easily divisible), which lower-income groups can more easily afford. Greater supply of small, indigenous fishes will certainly increase poor people's access to fish. Live fish is preferred by all groups of buyers. Research priority should focus on identifying consumer preferences in terms of demands in the market.

4.5 Change in fish prices over the years

Even if demand increases, prices may not rise, as long as supply of a commodity keeps pace with higher demand. Time-series data since 1985/86 on prices per t for rohu (BBS reported no other fish prices for this period) at wholesale and retail markets in Dhaka show that real prices (deflated by non-food consumer price index of the middle class income group of Dhaka city) increased 1.88% per annum at the retail level and 3.71% at the wholesale level (Table 18). Retail prices rose less than wholesale prices (retail markets are more competitive than wholesale markets; retail markets respond more to price increases). Higher nominal prices have been associated with inflation. Real price increases at the retail level have been much lower than the increase in the demand for fish (annual demand for fish has risen at least around 4.15% a year - with a population increase of more than 2% per annum; for explanations see section 7). Thus, the increase in supply and availability of fish (carps) in Bangladesh has been adequate.

Table 18. Real and nominal price (annual average) changes in Rohu (1985/86 to 1995/96)

Prices

Growth rate
Per annum (percent)*

t-ratio

R2

Real

Wholesale

3.71

3.34 hs

0.55

Retail

1.88

3.19 hs

0.53

Nominal

Wholesale

9.82

15.06 hs

0.96

Retail

7.99

12.87 hs

0.94

SOURCE: Bangladesh Statistical Yearbook 1997.

NOTE: * Estimated through fitting semi-logarithmic equation (exponential growth equation).

We analysed growth-trends of deflated prices of (by non-food CPI of the middle class income-group of Dhaka city) representative fish species in the wholesale markets of Dhaka (retail prices not reported). The results are reported in Table 19. Supply of many species, including carps, is usually low in May and peaks in November. Thus, we considered fish prices of the second week in both May and November for our annual growth analysis for the period of 1989/90 to 1995/96.

We saw no significant growth of deflated prices for leading fish species in the Dhaka markets during the peak season of fish supply. During the lean period of May, only catla and shingi prices showed significant yearly growth trends[17]. Supply of shingi fish might not have increased, because this species is not yet being cultured (this is true for koi also). In general we can say that fish prices did not increase during that period in real terms. A plausible explanation could lie in the increased supply of fish in the country (this may not be true for indigenous fish species) and a greater supply of substitute foods like poultry, poultry products and vegetables.

Table 19. Trends in fish prices (deflated) during the period 1989/90 to 1995/96

Fish species

Coefficient of time

t ratio

R2

Estimator

Auto regressive process

Rohu

May

0.06

2.47

0.74

ML

2nd order

Nov.

0.02

0.30

0.29

ML

2nd order

Catla

May

0.07

6.96hs

0.91

ML

2nd order

Nov.

0.05

2.74

0.63

ML

2nd order

Hilsa

May

0.04

0.36

0.72

ML

2nd order

Nov.

0.06

1.13

0.65

ML

2nd order

Shrimp

May

0.04

1.64

0.55

ML

2nd order

Nov.

0.02

0.63

0.66

ML

2nd order

Koi

May

-0.54

-0.92

0.52

ML

2nd order

Nov.

0.64

2.33

0.61

ML

2nd order

Shingi

May

0.10

25.04hs

0.99

ML

2nd order

Nov.

0.03

1.45

0.49

ML

2nd order

NOTE: Estimated through fitting a semi-logarithmic equation (exponential growth equation on time). Equations fitted well with second order auto-regressive error model and were estimated by using Maximum Likelihood Estimator (OLS cannot be used, if the error is auto correlated).

SOURCE: Reports of weekly prices by the Department of Agricultural Marketing - Various Issues. Data given in Appendix 5.

4.6 Income elasticity of demand for fish

Income elasticity of demand for fish is a measure of effect of change in income on the demand for fish. Usually income elasticity of demand for most of the goods is positive which implies that an increase in income leads to an increase in the demand for the goods. Income elasticity of demand for a commodity is defined as:

Change in income affects a commodity consumed not so much quantitatively but rather qualitatively (for example, coarse-quality rice may be replaced by fine-quality rice with increase of income). Therefore, the following two elasticity measures are better suited to measure income elasticity:
i)

= Value expenditure elasticity of demand

ii)

= Value income elasticity of demand

Formula i) captures elasticity on total income actually spent over a particular time, and the magnitude of elasticity can be postulated as being higher than formula ii), where total available income may not be spent completely.

We estimated income elasticities of demand for fish and for all food (for comparison) for the year 1995-96, using the Household Expenditure Survey data on different land holding groups (7 groups; data are provided in Appendix 6). Results of the income elasticity of demand for fish and income elasticity of demand for all food (for comparison) are presented in Table 20.

Table 20. Income elasticity of demand for fish 1995-96

a) Dependent variable = Expenditure on fish

Independent variable

B Value

t ratio

R2

Estimator

Auto-regressive process

i) Total expenditure

0.79

17.22 hs

0.99

ML

2nd order

ii) Total Income

0.65

20.20 hs

0.99

ML

2nd order


b) Dependent variable = Expenditure on all food

Independent variable

B Value

t ratio

R2

Estimator

Auto-regressive process

i) Total expenditure

0.60

28.80 hs

0.99

OLS

No auto corr.

ii) Total income

0.50

31.11 hs

0.99

ML

2nd order

NOTE: Both dependent and independent variables were in natural logarithm in the regression equation.

SOURCE: Appendix 6.

Value expenditure elasticity of demand for fish and for all food in 1995-96 was 0.79 and 0.60 respectively. That is, a 10% increase in total expenditure will boost fish expenditure by 7.9%, while expenditure on all food items will increase by 6%. Increase in expenditure on fish will increase more than on all other food items consumed together. Value income elasticity of demand for fish has also been higher (0.65) than for all other food items (0.50). Fish appears to be a superior commodity above all other foods.


[15] According to the household expenditure survey, daily fish consumption per caput is 43.8 g, and average per caput daily food intake is 913.8 g, and calorie intake is 2254 (HES, 1995-96 PP. 46-47). Over 40% of the population (around 50 million people) regularly consume fewer than 2 122 kcal per day, and around half of these people may live in extreme poverty consuming fewer than 1 805 kcal per day (project memorandum, DFID 1997-Annex-2).
[16] Expenditure elasticity is unitary for fish consumption. That is, a 10% increase in per caput total expenditure due to increase in income, raises per caput expenditure on fish or freshwater fish by the same percentage amount, i.e. 10% (Rahman and Ray, 1990, p.10). See the present estimates in section 3.6.
[17] The OLS estimates do not provide any significant growth for any fish species both in May and November. However, OLS estimates have not been reported due to low R2 values of equations.

Previous Page Top of Page Next Page