Remoteness, welfare and nutrition

17
Remoteness, Welfare and Nutrition: Evidence from Rural Ethiopia David Stifel Lafayette College 15 June 2015 Addis Ababa, Ethiopia

Transcript of Remoteness, welfare and nutrition

Remoteness, Welfare and Nutrition:Evidence from Rural Ethiopia

David StifelLafayette College

15 June 2015Addis Ababa, Ethiopia

How does remoteness (accessibility to markets) affect…

• Household well-being?• Nutrition?

• Outcomes – mother & child anthropometrics• Dietary diversity• Food security

Data – Ethiopia Rural Transport Survey 2011

• Survey of 851 households in northwestern Amhara• Sample area selected purposefully

• Without a passable road• Land quality similar throughout• Households use the same major market

• Remoteness defined relative to this market town• Households’ circumstances differ because of differences in transport costs… …not because of land characteristics.• Technical purpose: Quasi-experimental setting addresses issue of endogeneity

and facilitates a causal interpretation of results

Ethiopia Rural Transport Survey 2011

HOUSEHOLD WELFARE & REMOTENESS

Household Per Capita Consumption & Transport Costs

050

010

0015

0020

0025

00B

irr /

pers

on

0 20 40 60 80Transport Cost (Birr/kg)

Per Capita Household Consumption FoodNon-Food

Household Food Consumption & Transport Costs

050

010

0015

0020

0025

00B

irr /

pers

on

0 20 40 60 80Transport Cost (Birr/kg)

Per Capita Food Consumption PurchasedOwn-Consumption

NUTRITION & REMOTENESS

Outcomes: Mothers’ BMI

• Mean BMI = 19.2• Percent underweight = 34.2• Statistically not different by

remoteness

1015

2025

30B

MI (

we

ight

/sq

uare

d-h

eigh

t)

0 20 40 60 80Transport Cost (Birr/Quintal

bandwidth = .8

Note: Underweight is BMI < 18.5

Outcomes: Child Stunting

• 36 percent stunted• Statistically not different by

remoteness

-6-2

02

6H

eig

ht-f

or-

Age

Z-s

core

0 20 40 60 80Transport Costs (Birr/Quintal)

bandwidth = .8

Stunting: HAZ < -2

Outcomes: Child Wasting

• 17 percent wasted• Statistically not different by

remoteness

-6-2

02

6W

eig

ht-f

or-

He

ight

Z-s

core

0 20 40 60 80Transport Costs (Birr/Quintal)

bandwidth = .8

Wasting: WHZ < -2

Dietary Diversity

020

4060

8010

0P

erce

nt

Least Remote Quintile 2 Quintile 3 Quintile 4 Most Remote

Percent of Households Consuming Particular Food Items

Oil Milk Meat Fruit

01

23

45

6A

vera

ge N

umb

er o

f Fo

od G

rou

ps

0 20 40 60 80Transport Cost (Birr/kg)

All Household Members Chldren under age 5

Average Number of Food Groups in Household Diet

Food Security

0.1

.2.3

Per

cent

of H

ous

eho

lds

Least Remote Quintile 2 Quintile 3 Quintile 4 Most Remote

Worry Not Enough Food in Past 30 Days

1-2 Times 3-10 Times10+ Times

Food Security

0.1

.2.3

Per

cent

of H

ous

eho

lds

Least Remote Quintile 2 Quintile 3 Quintile 4 Most Remote

Ate Smaller Meals in Past 30 Days

1-2 Times 3-10 Times10+ Times

0.1

.2.3

Per

cent

of H

ous

eho

lds

Least Remote Quintile 2 Quintile 3 Quintile 4 Most Remote

Ate Fewer Meals in Past 30 Days

1-2 Times 3-10 Times10+ Times

IN SUMMARY

Remoteness in our survey area…

• Negatively affects household consumption• Lower production – less own consumption• Lower marketed surplus – less purchased

• Has no discernable effect on nutrition outcomes• Mothers’ BMI• Child stunting and wasting

• Negatively affects…• Dietary diversity• Food security

Tension in the results…

• Remoteness• Negatively affects

• Food consumption• Dietary diversity• Food securty

• No discernable effect on nutrition outcomes

• Why no effect on nutrition outcomes?• Low access to sanitation & healthcare throughout the survey area?• Threshold effects – Food consumption & dietary diversity levels too low?• Your thoughts?