Agricultural Diversity, Dietary Diversity and Nutritional ...

Agricultural Diversity, Dietary Diversity and Nutritional Intake: An Evidence on Inter-linkages from Village Level Studies in Eastern India

Anjani Kumar Sunil Saroj R K P Singh Shiv Jee 24th Annual Conference on Agriculture for Nutritional Security 15-17 December 2016 ICAR- Indian Veterinary Research Institute Izatnagar, Uttar Pradesh

Outline of the presentation Background and objectives  Data  Methodology  Findings  Conclusions and Recommendations 

Background Food intake from own production

Agriculture

High value crops

Nutrition

Livestock rearing

Reduced real food prices

Source: World Bank, 2007; Gillespie and Kadiyala, 2012; Viswanathan et. al., 2015; Slavchevska, 2015 

The evidence on the links between agriculture and nutrition has been vague

Contd.. 

India is facing a paradoxical situation. 



Despite the importance and potential of agriculture in improving nutrition of the farming households. 



Witnessing one of the highest economic growths with a much slower decline in undernutrition.

Existing understanding about linkages between agriculture and nutrition is extremely weak in India.

In the rural India (particularly in less-developed regions), food intake is closely tied to on-farm agricultural production. 

But the paucity of unit-level data that combine information on both nutrition and agriculture constraints at the national level.

Objectives  To 

food consumption patterns, agricultural production and dietary diversity

 To 

assess the relationship

between quality of food intake and agricultural production portfolio

 To 

examine

analyse

the impact of dietary diversity on nutritional intake

Data   

Used the panel data for 2010-11 to 2014-15 – part of village level studies (VLS) of the BMGF sponsored ICAR-ICRISAT collaborative project. Three states – Bihar, Jharkhand and Odisha Two districts from each state – 6 districts   



Two villages from each district – 12 villages   

 

Patna and Darbhanga - Bihar Ranchi and Dumka - Jharkhand Dhenkanal and Bolangir - Odisha Arap and Bhagakole – Patna; Susari and Inai - Darbhanga Dubaliya and Hesapiri – Ranchi; Dumariya and Durgapur - Dumka Sogar and Chandrasekhapur – Dhenkanal; Ainlatunga and Bilaikani - Bolangir

Forty households from each village – 480 Hhs Comprehensive data at household, individual and plot levels 

Household characteristics, expenditure, income, consumption, investment, employment and farming practices.

Methodology 

Simpson Index of Dietary Diversity (SIDD) 𝐧

𝐏𝐢𝟐

𝑺𝑰𝑫𝑫 = 𝟏 − 𝐢=𝟏





𝐏𝐢 is the proportion of the ith food item in total monthly consumption of all food commodities by the members of household.

Agriculture Production Diversity (AgPD) 

𝐏𝐢 is the share of cultivated area of the ith crop in the total cultivated area. 𝐧

𝐏𝐢𝟐

𝑨𝒈𝑷𝑫 = 𝟏 − 𝐢=𝟏

 

Index ranges between 0 and 1 Larger the value of index; larger the diversity in dietary diversity and crop production

Contd.. 

Multiple Panel Regression Model 

Hypotheses examined  Agricultural production diversity influences the household dietary diversity  𝑺𝑰𝑫𝑫𝒊𝒕 = 𝛂 + 𝛃𝑨𝒈𝑷𝑫𝒊𝒕 + 𝛄𝑬𝒊𝒕 + 𝒖𝒊  Household dietary diversity influences the nutritional (calorie or protein) intake  Calorie Intake: 𝑪𝑰𝒊𝒕 = 𝜶 + 𝜷𝑺𝑰𝑫𝑫𝒊𝒕 + 𝜸𝑬𝒊𝒕 + 𝒖𝒊  Protein Intake: 𝑷𝑰𝒊𝒕 = 𝜶 + 𝜷𝑺𝑰𝑫𝑫𝒊𝒕 + 𝜸𝑬𝒊𝒕 + 𝒖𝒊 Where;

   

𝑨𝒈𝑷𝑫𝒊𝒕 - agricultural production index (ranging from 0 to 1) 𝑺𝑰𝑫𝑫𝒊𝒕 - dietary diversity index (ranging from 0 to 1) 𝑬𝒊 - socio-economic and demographic characteristics of the households 𝒖𝒊 - error term and assumed to be normally distributed

Contd.. 

Binary Panel Logistic Regression 

Hypothesis examined 

Variables affecting the probability of being deficient in calorie and protein intake  𝐘 = 𝐥𝐧

𝐩 𝟏−𝐩

= 𝛃𝟎 +

𝛃𝐢 𝐗 𝐢

Where;  p represents the probability of a household being deficient in calorie and protein intake  βs are the regression coefficients estimated by the maximum likelihood method  Xs represent the explanatory variables and include several socio-economic and demographic characteristics of the farm households

FINDINGS

Average monthly expenditure and share of food in consumption expenditure across farm-size groups (Rs/capita) 2000

Bihar

1626

1500 1000 500

Bihar

Jharkhand

640

642

799

0

International Food Policy Research Institute

894

717

60 50 40 30 20 10 0

53

Jharkhand 54

55

53

53 39

Consumption of various food commodities (kg/capita/month) 15 Bihar

Jharkhand

Odisha

Eastern India

12 9

6 3 0 Cereals

Pulses Edible oils Fruits

Veg.

Milk

MFE

Sugar

Spices Dry fruits

Share of home-produced food commodities (%) Bihar

80

Jharkhand

Odisha

Eastern India

70

65

60 50

54

40 30

26

23

20

20 11

10

7

0 Cereals

Pulses

Edible oils

Fruits

Vegetables

Milk

MFE

Share of cereals purchased from PDS (%) Bihar

40

Jharkhand

Odisha

Eastern India

35

30 25 20

28 22

19

22 18

15 11

10 5 0 Labour Households Marginal

Small

Medium

Large

All

Simpson index of dietary diversity Labour Households

Marginal

Small

Medium

Large

0.9

0.8

0.78

0.73

0.72 0.65

0.7

0.6 0.5 0.4 0.3 0.2 0.1 0 Bihar

Jharkhand

Odisha

Eastern India

All

Simpson index of crop production diversity 0.6

Labour Households 0.6

Marginal

Small

Medium

Large

0.5 0.4

0.3 0.3

0.3 0.2 0.2 0.1 0.0 Bihar

Jharkhand

Odisha

Eastern India

All

Per capita calorie (kcal) and protein (g) intake in a day Calorie

Protein

3200

90

2400 kcal for rural area

2800 2400

80 70

2425 2276

2399

2368

2000

60.7 g for rural area

68

60

61

61

Odisha

Eastern India

54

50 1600 40 1200 30 800

20

400

10

0

0 Bihar

Jharkhand

Odisha

Bihar

Eastern India Labour households

Medium

Marginal

Large

Small

All

Jharkhand

Incidence of calorie and protein deficiency (%) Calorie

Protein

90

90

80

80 71

70 60

63

50

40

40

30

30

20

20

10

10

0

0 Jharkhand

Odisha

71

60

50

Bihar

78

70 61

59

84

56

Bihar

Eastern India Labour households

Medium

Marginal

Large

Small

All

Jharkhand

Odisha

Eastern India

Determinants of household dietary diversity Significant variables Agricultural production index log (age of the household-head) (in years) log (household size) (No.) Milk-producing household (Yes = 1, otherwise = 0)

Coefficient 0.059*** -0.039** 0.038*** 0.029***

SE (0.017) (0.017) (0.011) (0.009)

log (annual per-capita total expenditure) (in ₹)

0.041***

(0.007)

log (market distance from home) (km) Constant Observations (No.) Number of groups Wald chi2(20)

0.026*** 1.106*** 2033 488 122.51

(0.006) (0.129)

Determinants of household nutritional intake Calorie

Protein

Significant variables Dietary diversity index

Coef. 0.069***

SE (0.022)

Coef. 0.144***

SE (0.024)

Gender of the household-head (male = 1, female=0)

-0.062***

(0.021)

-0.057**

(0.022)

Caste affiliation (SC/ST = 1, others =0) log (household size) (No.) Vegetarian dummy (did not consume any non-vegetarian food = 1, otherwise = 0) Milk producing household (Yes = 1, otherwise = 0) log (annual per-capita total expenditure) (in ₹) log (share of food expenditure to total expenditure) log (household assets value) (in ₹) Access to PDS (Yes = 1, otherwise = 0) log (market distance from home) log (dependency ratio) Constant Observations (No.) Number of groups Wald chi2(20)

-0.027** -0.086***

(0.011) (0.013)

-0.063***

(0.014)

0.054***

(0.014)

0.086***

(0.015)

-0.035*** 0.241*** 0.226*** 0.011*** -0.018* -0.037*** -0.030* 4.964*** 2,033 488 1484.8

(0.009) (0.008) (0.010) (0.002) (0.010) (0.007) (0.015) (0.149)

0.278*** 0.221*** 0.007*** -0.054*** -0.038***

(0.008) (0.011) (0.002) (0.011) (0.007)

0.862*** 2,033 488 1624.6

(0.161)

Determinants of calorie and protein deficiency Calorie Significant variables Dietary diversity index log (age of the household-head) (in years) Gender of the household head (male = 1, female=0) Caste affiliation (SC/ST = 1, others =0) log (household size) (No.) Vegetarian dummy (did not consume any non-vegetarian food = 1, otherwise = 0) Marginal farm-size (Yes = 1, otherwise = 0) Milk producing household (Yes = 1, otherwise = 0) log (annual per-capita total expenditure) (in ₹ ) log (share of food expenditure to total expenditure) log (household assets value) (in ₹ ) Access to PDS (Yes = 1, otherwise = 0) log (share of home produced food) log (market distance from home) Constant lnsig2u Observations (No.) Number of groups

Protein

dy/dx -0.109** -0.077* 0.091** 0.086*** 0.187***

SE (0.050) (0.043) (0.044) (0.026) (0.026)

dy/dx -0.234*** -0.072* 0.078*

SE (0.049) (0.042) (0.044)

0.080***

(0.025)

-0.166***

(0.032)

-0.217***

(0.028)

-0.102*

(0.061) -0.042** -0.386*** -0.343*** -0.0091* 0.110*** 0.031*** 0.106*** 53.43*** 0.998*** 2,033 488

(0.020) (0.017) (0.026) (0.004) (0.022) (0.009) (0.017) (5.370) (0.230)

-0.372*** -0.371*** -0.020*** 0.085***

(0.015) (0.024) (0.004) (0.023)

0.118*** 45.89*** 1.035*** 2,033 488

(0.016) (4.300) (0.203)

Conclusion and Recommendation     

Agricultural production diversity is a major determinant of dietary diversity Dietary diversity – Strong effect on nutritional intake Agricultural research and policy efforts need to be broadened More diverse agricultural systems are also good for dietary intake and it will have positive impact on nutritional outcome. Policy recommendations   

Promote agricultural diversification to enhance food and nutritional security Promote diversity in agricultural production rather than focusing on increase in production of selected staple crops. In-depth research is needed to explicitly decipher the relationship between agricultural activities and nutritional outcome

THANK YOU