Secondary Cities and - World Bankpubdocs.worldbank.org/pubdocs/publicdoc/2016/5/... · Cities &...

Secondary Cities and Agricultural Transformation: Evidence from Ethiopia

World Bank – May 19, 2016

Joachim Vandercasteelen, Seneshaw Tamru, Bart Minten, and Johan Swinnen

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Urbanization

2

• Urbanization is quickly increasing in SSA o Growth and concentration in medium-sized citieso Economic development and poverty

• Composition of urbanizationo Agglomeration in mega citieso More dispersed pattern leads to more inclusive growth

• Conference topico Indirect mechanismo Broader impacts

Secondary Towns

3

• Migrants’ outcomeso More accessible non-farm sectoro Migration less costly

• Hinterland effectso Remittances, wage effects, etc.o Consumption linkage

• Higher urban demand• Agricultural production in rural hinterland

• Rural development o Small scale and rural farming

Cities & Agriculture in Ethiopia

4

• Ethiopia is urbanizing fasto Urban share will triple in next decade o People are concentrating in cities

• Addis Ababa • Secondary towns

• Teff is major staple cropo Produced by 6 million farmerso Consumed daily by 2/3 Ethiopianso Economically superior good

Cities are growing fast in Ethiopia

5

0

100

200

300

400

500

600

Mek

elle

0

1000

2000

3000

4000

5000

6000

Add

is A

baba

1950 1960 1970 1980 1990 2000 2010 2020 2030Year

Population in thousands

Market Access has improved

6

Conceptual framework

P (teff price)

µf (Transportation costto City 1)0

Agriculture and 1 City

Agriculture and 2 Cities

9

• Now: Urban proximity to 2 citieso Presence of city 2 (Secondary Town) at μo Same price and per unit transport costo , and , .

Teff price

Transportation cost to City 1

μ1 μfnμ2

0

Impact of City 1 Impact of City 2

μf1

Pf,1(μ) Pf,2(μ)

Prices and 2 cities

Teff price


μ1 μfnμ2

0


μ1'μf1

Pf,1(μ) Pf,2(μ)

Prices and 2 cities

Teff price

μ2

0


Pf,1(μ) Pf,2(μ)

μ3

Impact of City 3

μfnμ11μf1 μ1

Pf,3(μ)

Prices and many cities


Teff price

μ2

0


Pf,1(μ) Pf,2(μ)

Prices and many cities

μ3

Impact of City 3


Survey Data

16

• Teff value chain survey in 2012 (cross sectional)o 36 randomly selected villages in major teff producing areas o 720 randomly selected farmers

• Urban proximityo Transport cost o Cost of transporting 1 quintal of teff (ETB/quintal)

• Cities in survey areao Capital: Addis Ababao Secondary towns: Nazareth and Bahir Dar o Select city that maximizes net prices : ,

Teff Transaction Trip

17

Donkey Trip Truck Trip

Self-reported cost GIS Road Network

max , )?

μcf

Total Transport Cost T

mf, pf

Farmers and cities

18

ResultsIs there an effect of Secondary Towns?

Simulation of model

20

• Compare model simulation with empirical observation

• Teff price farmers receive in trader marketso Predicted: ,

o Observed: ,

• Truck cost from trader town to Addis Ababa

21

Predicted vs. Observed teff prices

400

600

800

1000

1200

1400

-100 -50 0 50 100Truck Cost to Addis Ababa

Predicted teff price net of transport cost

950

1000

1050

1100

1150

1200

1250

1300


Observed teff price

22

Other outcomes

0

50

100

150

200

250


Use of DAP (kg/ha)

0

10

20

30

40

50

60


Use of Improved Seeds (kg/ha)

800

1000

1200

1400

1600

1800

2000

2200


Teff Land Productivity (kg/ha)

5000

10000

15000

20000


Teff Profits (birr/ha)

ResultsSize of effect

Urban proximity & Agricultural Outcomes

24

• Effect of (total) transport cost To Y α β ∗ T X P Z ε ,o Y : improved seed usage, productivity and profitso X : controlso P : output and input prices o Z : zone fixed effects

• Effect β dependent on type of city?o Y α β ∗ T γ ∗ S π ∗ S ∗ T X P Z ε ,o S dummy for Secondary Townso β is effect for farmers shipping to Addis Ababao β π is effect for farmers shipping to Secondary Town

Endogeneity issues

25

• Heterogeneity between farmers in So Household and teff productiono Double Robust method

• Estimation of Propensity Score• Weighted regression

• Endogenous location of road networko Better infrastructure in areas with higher economic potentialo Instrumental Variable (IV) Approach

• ‘Natural path’ walking distance

Preliminary results

26

Secondary TownTransportation Cost to Addis

Ababa

Transportation Cost to

Secondary Town

Difference in effect

Teff price (ETB) –*** –*** –*** 0

Wages (ETB) 0 –*** –*** 0

Fertilizer (kg/ha) 0 –** –*** 0

Improved Seed (kg/ha) 0 0 –*** 0

Labor (days/ha) 0 –* –*** <0**

Yield (kg/ha) –*** –*** –*** >0**

Labor Productivity (kg/day) –** –*** 0 >0**

Teff Profits (ETB/ha) –*** –*** 0 >0**

Conclusions

Cities & Agricultural transformation in Ethiopia

28

• Impact of urban proximity on teff production decisions

• Difference between type of citieso Effect on prices is similaro Strong effect on input use in Secondary Townso Strong effect on labor productivity in Capital

• Way forwardo Improve empirical resultso Welfare implicationso Other mechanisms

Thank you!

Propensity Score

30

Secondary town (Yes=1)0.00

(0.00)0.03

(0.16)0.14

(0.08)-0.67***

(0.09)0.02

(0.04)0.05

(0.04)-0.10(0.11)

-0.55**(0.24)

-0.53***(0.11)

0.16***(0.05)

-0.29***(0.10)0.06

(0.07)-0.09***

(0.03)0.82

(0.70)

Age of head (years)

Educated head (years=1)

Head is from Oromia (yes=1)

Size of the household (number)

Children in the household (number)

Gender of head (male=1)

Constant

Household owns radio (yes=1)

Household owns tv (yes=1)

Household owns mobile phone (yes=1)

Wealth index (PCA of household assets)

Farm assets (ln of ETB)

Land owned (ha)

Farming Ability (.)

Balancedness of matching

31

.2 .4 .6 .8 1Propensity Score

Untreated: Off support Untreated: On supportTreated

Natural Path

32

First Stage result

33

First stageTransportation Cost

(ETB/quintal)Transportation Cost *

Secondary Town1.38*** 0.81***(0.26) (0.15)0.36 1.27***(0.26) (0.24)‐2.15 25.12***(5.79) (5.36)

ControlsPricesR² 0.590 0.890Shea's Adjusted Partial R² 0.244 0.330

Natural Path (hours)

Natural Path * Secondary Town

Secondary Town

yesyes

Preliminary results

34

• Teff outcomes are lower in Secondary Towns

• Effect of distance o Teff price and fertilizer use: similar

o Fertilizer and labor use: larger for farmers shipping to a Secondary Town

o Improved seed use, teff yield, labor productivity and profits: larger for farmers shipping to Addis

Prices log of teff prices (ETB/quintal) log of wage (ETB/day)

OLS IV OLS IV Urban proximity

Truck Cost (ETB/quintal) ‐4.52*** ‐3.35** ‐3.77 ‐5.85 (0.92) (1.33) (4.76) (5.68)

Wu‐Hausman F(3,715) 10.49*** 11.75*** Urban proximity interacted Truck Cost to Addis Ababa (ETB/quintal)

‐6.61*** ‐6.34*** ‐14.50* ‐16.24** (1.30) (1.35) (7.29) (7.56)

Truck Cost to Secondary Town (ETB/quintal)

‐5.75*** ‐5.54*** ‐6.53 ‐11.25** (0.90) (0.91) (4.95) (5.62)

Difference in effect 0.86 0.80 7.97 4.98 (1.41) (1.41) (5.80) (5.74)

Secondary Town (yes=1) ‐158.29*** ‐154.84*** ‐618.27*** ‐541.90***(47.90) (48.69) (211.83) (203.47)

Wu‐Hausman F(3,715) 0.36 45.93*** Urban proximity interacted + controls Truck Cost to Addis Ababa (ETB/quintal)

‐6.38*** ‐6.50*** ‐12.96 ‐14.89* (1.30) (1.30) (8.47) (8.24)

Truck Cost to Secondary Town (ETB/quintal)

‐5.05*** ‐5.12*** ‐6.57 ‐14.29** (0.94) (0.98) (4.96) (5.57)

Difference in effect 1.33 1.38 6.39 0.60 (1.51) (1.58) (7.10) (6.38)

Secondary Town (yes=1) ‐171.93** ‐173.79*** ‐523.40 ‐318.21 (63.24) (66.57) (327.23) (281.25)

Wu‐Hausman F(3,715) 0.97 58.94***

Urban proximity and prices

35

InputsOLS IV OLS IV OLS IV

‐1.80*** ‐2.30*** ‐0.36** ‐0.32* ‐1.07*** ‐0.64(0.35) (0.64) (0.13) (0.18) (0.28) (0.50)

Wu‐Hausman F(3,715) 4.30** 2.84** 3.55**

‐3.34*** ‐5.67*** ‐0.56*** ‐0.85*** ‐0.18 ‐0.79(1.10) (1.60) (0.14) (0.24) (0.62) (0.76)

‐1.64*** ‐2.67*** ‐0.35** ‐0.50** ‐1.20*** ‐1.00***(0.32) (0.58) (0.15) (0.21) (0.26) (0.29)1.70 3.00** 0.21 0.35 ‐1.03 ‐0.22(1.13) (1.51) (0.20) (0.27) (0.63) (0.69)

‐140.01** ‐209.08*** ‐24.43** ‐32.23** 33.49 ‐4.96(57.23) (76.51) (11.36) (14.71) (25.58) (26.78)

Wu‐Hausman F(3,715) 17.73*** 6.22*** 2.90**

‐1.77* ‐4.67** ‐0.04 ‐0.42 ‐0.41 ‐1.21*(0.96) (1.88) (0.19) (0.38) (0.45) (0.71)

‐1.91*** ‐3.59*** ‐0.47*** ‐0.87*** ‐1.53*** ‐1.59***(0.34) (0.63) (0.13) (0.19) (0.31) (0.42)‐0.14 1.08 ‐0.43* ‐0.45 ‐1.12** ‐0.38(0.98) (1.71) (0.23) (0.31) (0.47) (0.64)‐17.58 ‐90.99 9.55 8.06 29.90 ‐9.00(50.00) (83.93) (12.99) (17.69) (26.21) (33.31)

Wu‐Hausman F(3,715) 12.92*** 7.43*** 1.14

Secondary Town (yes=1)


Transportation Cost (ETB/quintal)


Labor (days/ha)

Urban proximity

Urban proximity interacted

Urban proximity interacted + controls

Transportation Cost to Addis Ababa (ETB/quintal)

Transportation Cost to Secondary Town (ETB/quintal)



DAP (kg/ha) Improved Seed (kg/ha)


Urban proximity and inputs

36

OutcomesOLS IV OLS IV OLS IV

‐14.63*** ‐10.62** ‐0.06*** ‐0.05 ‐89.28*** ‐22.24(1.89) (4.73) (0.01) (0.04) (21.46) (51.35)

Wu‐Hausman F(3,715) 3.80* 0.39 10.80***

‐26.47*** ‐33.12*** ‐0.17*** ‐0.20*** ‐230.05*** ‐273.62***(4.96) (6.27) (0.05) (0.05) (55.12) (71.80)

‐13.43*** ‐15.38*** ‐0.05*** ‐0.10*** ‐76.16*** ‐63.92*(2.07) (3.18) (0.01) (0.03) (24.80) (36.83)13.03** 17.74*** 0.11** 0.10* 153.89*** 209.70***(4.96) (5.58) (0.05) (0.06) (51.91) (59.37)

‐1,048.02*** ‐1,287.77*** ‐8.30*** ‐8.03*** ‐11,732.28*** ‐14,388.72***(211.71) (252.91) (2.41) (2.85) (2,285.03) (2,691.07)

Wu‐Hausman F(3,715) 1.62 1.81 2.25*

‐19.62*** ‐28.60*** ‐0.13** ‐0.17*** ‐150.60* ‐189.81**(5.56) (7.17) (0.05) (0.06) (78.75) (94.89)

‐11.26*** ‐12.91*** ‐0.01 ‐0.03 ‐56.41** ‐20.76(2.17) (4.34) (0.02) (0.04) (22.60) (51.17)8.36 15.70** 0.12** 0.14** 94.19 169.04**(5.61) (6.14) (0.05) (0.07) (70.54) (72.92)

‐686.70** ‐1,077.26*** ‐7.24** ‐8.52** ‐7,627.23** ‐11,293.37***(297.87) (324.88) (2.98) (3.62) (3,617.66) (3,800.74)

Wu‐Hausman F(3,715) 1.62 0.27 2.56*



Yield (kg/ha) Labor Productivity (kg/day) Teff Profits (ETB/ha)

Urban proximity










Urban proximity and productivity

37

Urban proximity and welfare

38

WelfareOLS IV OLS IV

‐0.00 ‐0.01 ‐0.01 ‐0.01(0.00) (0.01) (0.01) (0.01)

Wu‐Hausman F(3,1177) 0.10 0.41

‐0.01 ‐0.02 ‐0.03*** ‐0.04***(0.01) (0.02) (0.01) (0.01)‐0.00 ‐0.01 0.01 ‐0.00(0.01) (0.01) (0.01) (0.01)

Difference in effect 0.00 0.00 0.04*** 0.04***(0.01) (0.02) (0.01) (0.01)

Wu‐Hausman F(3,1177) 0.84 0.33

‐0.00 0.02 ‐0.04*** ‐0.05**(0.01) (0.02) (0.01) (0.02)‐0.01 ‐0.02 ‐0.00 ‐0.03**(0.01) (0.01) (0.01) (0.01)‐0.00 ‐0.04 0.03** 0.02(0.01) (0.03) (0.01) (0.02)

Wu‐Hausman F(3,1177) 1.85 1.50







log of Wage income (ETB) log of Commerce income (ETB)

Urban proximity



Secondary Cities and - World Bankpubdocs.worldbank.org/pubdocs/publicdoc/2016/5/... · Cities &...

Documents

Transcript of Secondary Cities and - World Bankpubdocs.worldbank.org/pubdocs/publicdoc/2016/5/... · Cities &...