ETHIOPIAN DEVELOPMENT RESEARCH INSTITUTE

Implications of

Agricultural Productivity Growth for

Structural Change and Employment in Ethiopia

Paul Dorosh, Sherman Robinson and James ThurlowIFPRI

EDRI-ESSP SeminarAddis AbabaApril 9, 2015

Overview

1. Recent growth trends

2. Measuring structural change

3. Moving to a time-based view of employment

4. Modeling future growth and structural change

5. Summary and conclusions

2

1. Recent Growth Trends

3

Rapid Economic Growth Since 2002

• Total GDP growth averaged 10% per year (2001/02-13/14)

– About 7.5% annual growth in per capita termsReal GDP per capita, 2002/03-13/14

(constant 2012/13 prices)

239

541584

1,322

0

200

400

600

800

1,000

1,200

1,400

20

01

/02

02

/03

03

/04

04

/05

05

/06

06

/07

07

/08

08

/09

09

/10

10

/11

11

/12

12

/13

13

/14

GD

P p

er c

apit

a (U

S$)

US$

US$ PPP

4

Sectoral Patterns of Growth

• Growth occurred throughout the economy

– Particularly strong growth in industry and services

Initial GDPshare (%)

Final GDP share (%)

GDP growth rate (%)

Contribution to increase in

GDP (%)

Total 100.0 100.0 10.0% 100.0

Agriculture 45.6 36.2 7.2% 28.6

Industry 13.9 17.5 12.7% 20.1

Services 40.6 46.3 11.8% 51.3

Real GDP, 2002/03-13/14

5

Key Growth Sectors

Initial GDPshare (%)

Final GDP share (%)

GDP growth rate (%)

Contribution to increase in

GDP (%)

Total 100.0 100.0 10.0 100.0

Agriculture 45.6 36.2 7.2 28.6

Mining 0.5 0.6 11.3 0.7

Manufacturing 5.4 5.2 9.8 5.3

Utilities 2.4 2.0 8.8 1.8

Construction 5.6 9.7 16.4 12.4

Trade 12.4 13.7 11.2 15.0

Hotels, catering 2.4 5.3 18.3 7.4

Transport, communication 5.8 6.4 11.8 6.6

Finance 0.8 1.9 17.0 2.5

Business, real estate 8.6 9.0 11.3 9.9

Public administration 4.7 4.1 8.8 4.0

Education 2.8 3.0 11.7 3.2

Health 0.9 1.1 11.8 1.3

Other services 1.9 1.5 7.2 1.3

***

*

= 72.2%

*

Real GDP, 2002/03-13/14

6

Agricultural Growth

• Agriculture is still a major driver of economic growth

– Grew at about 7% per year during 2001/02-13/14

• Four-fifths of agricultural GDP growth was from increased crop production

Initial agricultural GDP share

(%)

Final agricultural GDP share

(%)

Contribution to increase in agricultural

GDP (%)

Agriculture 100.0 100.0 100.0

Crops 59.3 68.4 81.9

Livestock 29.8 24.8 17.3

Forestry, Fishing 10.8 6.8 0.9

Real Agricultural GDP, 2002/03-13/14

Source: Diao, Thurlow and Verduzco Gallo (2015)7

Rising Crop Yields• Cereals generated more than half of crop GDP growth

during 2001/02-13/14

– Driven by land expansion and rising yields

– Cereals alone were responsible for almost 13% of total GDP growth in Ethiopia

• Overall, rising crop yields (land productivity) accounted for more than half of total crop GDP growth

– i.e., equal to 13% of total GDP growth in Ethiopia

All crops Cereals Non-Cereals

Crop GDP share in 2001/02 100% 57.1% 42.9%

Contribution to total crop GDP growth 100% 54.1% 45.9%

From cultivated land expansion 36.9% 21.1% 15.8%

From rising crop yields 55.4% 36.1% 19.3%

From reallocating land to higher value crops 7.7% -3.1% 10.8%Source: Diao, Thurlow and Verduzco Gallo (2015)

Decomposition of real crop GDP growth, 2001/02-13/14

8

Summary

• Ethiopia is one of the fastest growing economies in the world

• Economic growth has occurred throughout the economy

• There has been tremendous growth in industry and services– As a result agriculture’s share of GDP has fallen over the last

decade

• Nevertheless, agriculture remains a major driver of growth– Most of which is due to improvements in land productivity

9

2. Measuring Structural Change

10

Evidence of Structural Change• Positive structural change is defined as a movement of

workers from low to high productivity sectors (e.g., from agriculture to non-agriculture)

• Ethiopia’s two recent Labor Force Surveys reported a large decline in agriculture’s share of total employment between 2004/05 and 2012/13

– From 80% to 73% over 8 years

– Is this evidence of rapid, positive structural change in Ethiopia?

Employment in 2004/05

(1000s)

Employment in 2012/13

(1000s)

Employment share in

2004/05 (%)

Employment share in

2012/13 (%)

Total employment 31,435 42,404 100.0 100.0

Agriculture 25,208 30,817 80.2 72.7

Industry 2,090 3,134 6.7 7.4

Services 4,127 8,453 13.1 19.9

Employment, 2004/05 and 2012/13

Source: Ethiopia’s Labor Force Surveys11

Private Household Jobs (1)

• Sharp rise in the employment share for the sector called “Private Households”– +2.8mil. workers over 8

years (= 26% of all new jobs)

• Sector includes work within the household, e.g., wood/water collection; cleaning; child care.

• Increase probably due to a change in the definition of employment between surveys:– Had to work 4+ hours per

week in 2004/05– Only had to work 1+ hours

per week in 2012/13

2004/05 2012/13

Total 100.0 100.0

Agriculture 80.2 72.7

Mining 0.3 0.4

Manufacturing 4.9 4.5

Utilities 0.1 0.5

Construction 1.4 2.0

Trade 5.2 5.4

Hotels, catering 2.5 1.1

Transport, communication 0.5 1.0

Finance 0.1 0.3

Business, real estate 0.2 0.7

Public administration 1.2 0.7

Education 0.9 1.6

Health 0.3 0.6

Other services 1.6 1.2

Private households 0.8 7.3

Employment share (%)

12

Private Household Jobs (2)

• Rural women accounted for most (80%) of the increase in Private Household employment between 2004/05 and 2012/13

• 2012/13 survey probably captured more “household chores” than the 2004/05 survey (possibly done by young women in rural households)

Total Male Female

National total 2,843 411 2,432

(+1,143%) (+1,780%) (+1,078%)

Urban areas 244 59 185

(+121%) (+290%) (+101%)

Rural areas 2,598 351 2,247

(+5,649%) (+1,3617%) (+5,176%)

Increase in Private Household jobs, 1000s (% changes in parentheses)

13

Adjusted Employment Shares

• Impose Private Household shares from 2004/05 onto 2012/13 – At the detailed male/female,

rural/urban level

• Private Household jobs in 2012/13 fall from 3.1 to 0.4 mil.– Up from 0.25mil in 2004/05

• Total employment falls from 42.4 to 39.7 mil.– Job growth is 3% p.a. (not 4%)

• Agriculture’s employment share is now 77.6% rather than 72.7%– Implies a slower rate of decline

in agricultural employment

Original Adjusted

Total 100.0 100.0

Agriculture 72.7 77.6

Mining 0.4 0.5

Manufacturing 4.5 4.8

Utilities 0.5 0.6

Construction 2.0 2.1

Trade 5.4 5.8

Hotels, catering 1.1 1.2

Transport, communication 1.0 1.0

Finance 0.3 0.3

Business, real estate 0.7 0.7

Public administration 0.7 0.7

Education 1.6 1.7

Health 0.6 0.6

Other services 1.2 1.3

Private households 7.3 1.0

Employment share (%)

14

Worker Productivity by Sector,

2012/13

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

0 10 20 30 40 50 60 70 80 90 100

Ave

rage

GD

P p

er w

ork

er (

US$

)

Share of employment (%)

PRHMAN

OSV

TRD

AGR

AGR: Agriculture

MIN: Mining

MAN: Manufacturing

CON: Construction

TRD: Trade services

HOT: Hotels, catering

TRC: Transport, communication

FIN: Finance

PAD: Public administration

EDU: Education

HSW: Health, social work

OSV: Other services

PRH: Private households

CON

HOT

EDU/HSW

FIN/PAD/TRC

15

Measuring Structural Change

• GDP growth can be expressed as changes in GDP per worker, which is a measure of labor productivity

• We can decompose growth in GDP per worker into two components:

– Within-sector: Increases in average GDP per worker assuming workers do not move between sectors

– Between-sector: Productivity increases caused by workers moving from lower to higher productivity sectors (e.g., agric. to manufacturing)

• A positive “between-sector” component implies a positive contribution of structural change to GDP growth

16

Decomposing Productivity Growth• Positive structural change accounted for 25% of

the increase in worker productivity between 2004/05 and 2012/13

– Mainly people exiting agriculture for non-farm jobs

– Agriculture still generated 26% of worker productivity gains (from higher yields)

Within-sectorproductivity

growth

Structural change (labor reallocation)

TotalShare of total

change

All sectors 353 116 469 100.0%

Agriculture 136 -14 123 26.2%

Industry 41 47 88 18.8%

Services 175 83 258 55.1%

Share of total change 75.3% 24.7% 100.0%

Increase in average value-added per worker (US$) between 2004/05 and 2012/13 caused by…

17

Positive Structural Change (04/05-

12/13)

agrmin

man

egw

con

trdhot

trcfin

rbs

pad

edu

hsw

osv

prh

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0

Val

ue-

add

ed p

er w

ork

er (

log

dev

iati

on

fro

m a

vera

ge)

Change in employment share (%-point)

Slope: 0.346

t-stat: 2.21 AGR: Agriculture

MIN: Mining

MAN: Manufacturing

EGW: Energy, gas, water

CON: Construction

TRD: Trade services

HOT: Hotels, catering

TRC: Transport, communication

FIN: Finance

RBS: Real estate, business

PAD: Public administration

EDU: Education

HSW: Health, social work

OSV: Other services

PRH: Private households

18

Summary

• The decline in agriculture’s employment share over the last decade is not as rapid as recent surveys suggest– Need to maintain consistent employment definitions in order to track

progress

• Nevertheless, rapid economic growth in Ethiopia was associated with positive structural change– Mainly from a declining importance of agricultural jobs– But also from a relative shift out of manufacturing into construction and

higher value-added services

– Agriculture’s employment share is falling by 0.33 percentage points per year

• This is lower than China’s rate of structural change (0.9 percentage points per year) when it transitioned from $500 to $1000 per capita

• At current trends, agriculture will continue to employ a majority of workers until 2097 (i.e., 77.6% - 84 years x 0.33 = 50%)

19

3. Moving to a Time-Based View

of Employment

20

Workers vs. Time Spent Working• Most surveys only capture people’s primary occupation

or job

– This includes 2012/13 Labor Force Survey

– But many farmers split their year between family farms and non-farm enterprises (and sometimes they work as wage labor in the nonagricultural sector)

• 2011/12 Ethiopia Rural Socioeconomic Survey (ERSS) provides information on farmers’ secondary non-farm jobs

%

Total 100.0

Family farming only 73.3

Farming + wage/non-farm work 13.5

Non-farm work only 13.2

Share of rural employed workforce holding farm and non-farm jobs, 2011/12

Source: 2011/12 Ethiopia Rural Socioeconomic Survey

= 86.8% in agriculture(≈ 89.6% in 2012/13 LFS)

21

Estimating Labor Full-Time Equivalents• Rough estimate of workers’ time allocations using simple

assumptions:

– Farming activities = 8 months of work

– Farmers’ non-farm jobs = 4 months of work (i.e., off-season)

– Solely non-agricultural jobs = 12 months of work (i.e., year round)

• Agriculture’s share of rural employment falls from 87% (measured in people’s primary jobs) to 77% (measured in time spent on the job)

People share (%)

People (1000s)

Time (years)

Time share (%)

Total employment 100.0 33,428 25,261 100.0

Family farming only 73.3 24,502 ×⅔ 16,335 64.7

Farming and wage/non-farm work 13.5 4,510 ⇒ 4,510 17.9

Farming component 13.5 4,510 ×⅔ 3,007 11.9

Non-farm work component 0 0 ×⅓ 1,503 6.0

Non-farm work only 13.2 4,416 ⇒ 4,416 17.5

Agriculture employment share 86.8% 76.6%

Measuring rural employment in full-time equivalents

22

Summary

• Secondary nonfarm jobs are often overlooked (esp. in rural economy)– These could easily account for a significant share of the time spent working– More detailed surveys and analysis are needed

• Three implications:

– Agriculture’s employment share may be lower than primary job statistics suggest

– Average agricultural value-added per “worker year” is higher than average valued-added per worker• This partially narrows gap in worker productivity across sectors (and reduces the

benefits from structural change)

– Rural non-farm jobs could be a potential source of structural change • One that does not require migration to small towns and major cities• And one that is difficult to capture in existing labor force surveys

23

4. Modeling Future Growth and

Structural Change

24

Conceptual Framework

Non-farm employment

Urban Sectors

Farm employment

Non-farm Underemployment

Rural Sectors

Migration

Poor

Non-poor

Rural households

Poor

Non-poor

Urban householdsNational markets

Rural “home” goods

Foreign markets

La

nd

, la

bo

r a

nd

ca

pit

al i

nco

mes

La

bo

r an

d ca

pita

l in

com

es

Government + Investment

25

Producers and Factor Markets

• Multi-sector, multi-region production structure– 2010/11 Social Accounting Matrix (i.e., adapted EPAU/EDRI SAM)– 20 sectors (6 in agriculture) in 3 regions (rural areas, small towns, major

cities)

• Detailed labor markets (time-based)– Rural: Farm work (fully-employed during season); Non-farm work (surplus)– Urban: Non-farm professional, skilled and unskilled (full-employed year-

round)

• Rural-to-urban migration:– Based on relative real wages (i.e., rural farm to urban unskilled)

Non-farm employment

Urban Sectors

Farm employment

Non-farm Underemployment

Rural Sectors

Migration

26

Households and Product Markets

• Rural/urban, poor/non-poor households (i.e., bottom two quintiles)– Earn land, labor and capital incomes according to their factor

endowments

• Purchased goods in national markets (potentially supplied from abroad)– Rural households first source goods from local rural markets

Poor

Non-poor

Rural households

Poor

Non-poor

Urban householdsNational markets

Rural “home” goods

Incomes Incomes

27

Recursive Dynamics• Model run over the period 2010/11 to 2024/25

• Endogenous capital accumulation and allocation

– Past investment influences current availability of capital (after depreciation)

• Exogenous updating:

– National population growth (regional population depends on migration flows)

– Labor, land and capital productivity growth (factor-specific, not TFP)

Baseline scenario

Alternative scenario

Outcome variable

(e.g., GDP)

2010/11 2014/15 2024/2528

Four Simulations

Baseline + Faster urban

growth+ Faster

Migration+ Agricultural productivity

Population growth 2% 2% 2% 2%

Exogenous labor supply growth

Rural 2% 2% 2% 2%

Towns 3% 3% 3% 3%

Cities 4% 4% 4% 4%

TFP growth

Rural 2% 2% 2% 2%

Towns 2.5% 2.5% 2.5% 2.5%

Cities 3% 3% 3% 3%

Base rural-urban migration rate 0.5% 0.5% +1%-point +1%-point

Capital productivity growth

Towns +5%-points +5%-points +5%-points

Cities +6%-points +6%-points +6%-points

Agricultural labor productivity +4%-points

Annual growth rates, 2014/15 to 2024/25

29

Results: Economic Growth

Baseline + Faster urban

growth+ Faster

Migration+ Agricultural productivity

Total GDP per capita (%) 4.71 8.30 8.32 8.96

Rural 3.56 3.47 3.46 4.96

Urban 5.93 12.56 12.16 12.16

Agric. GDP per capita (%) 1.89 1.84 1.71 3.99

Population (%) 2.00 2.00 2.00 2.00

Rural 1.95 1.95 1.88 1.85

Urban 2.29 2.28 2.71 2.83

Real food price index 0.38 0.87 0.89 0.74

Average annual growth rate (%) or deviation in price index (%-point)

30

Results: Wages and Migration

Baseline + Faster

urban growth+ Faster

Migration+ Agricultural productivity

Person-based labor supply (%) 2.34 2.34 2.35 2.36

Rural 1.88 1.88 1.69 1.64

Urban 3.84 3.82 4.42 4.57

Time-based labor supply (%) 3.19 3.60 3.59 3.80

Rural 3.03 3.55 3.39 3.61

Urban 3.84 3.82 4.42 4.57

Rural-urban wage ratio, 2024/25

Person-based 1.56 1.72 2.18 1.46

Time-based 1.80 2.16 2.75 1.74

Average annual growth rate (%) or deviation in price index (%-point)

31

Results: Employment Shares

Baseline + Faster urban

growth+ Faster

Migration+ Agricultural productivity

Person-based (%)

2013/14 78.1 78.1 78.1 78.1

2024/25 74.3 74.3 72.7 72.3

Annual change (%-point) -0.35 -0.34 -0.49 -0.53

Time-based (%)

2013/14 67.4 67.4 68.1 68.1

2024/25 58.5 56.1 55.0 53.4

Annual change (%-point) -0.80 -1.03 -1.19 -1.33

Agriculture’s share of total employment (%)

• Exceeds China’s 0.9%-points per year• Large increase over person-based rate of structural change

32

Results: Household Welfare

Baseline + Faster urban

growth+ Faster

Migration+ Agricultural productivity

National 3.7 6.3 6.2 7.1

Poor 4.5 7.2 7.2 6.8

Non-poor 3.4 5.9 5.8 7.3

Rural 3.8 5.8 5.9 6.8

Poor 4.5 7.1 7.3 6.7

Non-poor 3.4 5.1 5.1 6.9

Urban 3.1 6.9 6.5 7.3

Poor 3.7 8.3 7.0 7.9

Non-poor 3.2 6.9 6.7 7.6

Real household per capita consumption growth rate (%)

33

5. Summary and Conclusions

34

Summary and Conclusions1. Ethiopia is one of the fastest growing countries in the world

– A quarter of growth since 2004/05 was due to positive structural change

– This pace of change is slower than other countries during their rapid growth

2. Faster urban growth and rural-urban migration has limited effect on the pace of structural change, without investments in agricultural productivity

3. Labor-saving agricultural investments accelerate structural change by…– Releasing labor to urban areas– Preventing food prices from eroding real wage gains for new urban

migrants– Generating demand for rural non-farm products and creating jobs for

surplus rural labor during the off-season

35