Urbanization and Public Investment: Implications for Growth and Poverty Reduction in Africa

1

Urbanization and Public Investment: Implications for Growth and Poverty Reduction in Africa

Paul A. Dorosh*Director, Development Strategy and Governance

International Food Policy Research Institute

Analytical Exchange on “Strategic Consequences of Urbanization in Sub-Saharan Africa to 2025”

Arlington, VirginiaApril 3, 2014

The simulation model results presented here derive from Dorosh and Thurlow (2014), “Can Cities or Towns Drive African Development? Economywide Analysis for Ethiopia and Uganda”. World Development.

2

Presentation Outline

1. Urbanization and Connectivity in sub-Saharan Africa– Primary and secondary cities– Agglomeration and travel time

2. Implications of Urbanization and the Allocation of Public Investments: Model Simulations

3. Concluding Observations

3

Cities with Population Greater than 100K Sub-Saharan Africa 2000

* Average annual growth 1960-2000 for cities with data for both years.Source: Calculated from Henderson and Wang World Cities Database.

Population # Cities

Total Pop

(mns)

Average Pop

(mns)

Share of Urban

SSA

Average Annual

Growth*

> 10 mn 1 13.40 13.40 9.3% 7.4%

5 - 9.99 mn 1 5.06 5.06 3.5% 6.2%

2 - 4.99 mn 11 29.50 2.68 20.5% 4.8%

1 - 1.99 mn 25 34.81 1.39 24.2% 5.1%

0.5 - 0.99 mn 26 17.47 0.67 12.2% 4.3%

0.25 - 0.499 43 15.13 0.35 10.5% 4.1%

0.1 - 0.249 178 28.25 0.16 19.7% 2.9%

Total 285 143.63 0.50 100.0% 4.5%

4

Ethiopia: Population of Cities with over 100,000 Inhabitants, by City Size

Source: Calculated from Henderson and Wang World Cities Database and Ethiopia Population Census.

1970 1980 1990 2000 2008 -

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

2.0mn+1.0 - 1.99m500-999K300-499K200-299K100 - 199K

5

Ethiopia Cities with Population Greater than 100K (Population and Number of Cities)

Source: Calculated from Henderson and Wang World Cities Database and Ethiopia Population Census.

Population100 - 199K 200-299K 300-499K 500-999K 1.0 - 1.99m 2.0mn+ Total

1970 - - - 730 - - 730 1980 230 - - - 1,182 - 1,412 1990 611 252 - - 1,793 - 2,656 2000 700 - 310 - - 2,639 3,649 2008 1,188 429 342 - - 2,738 4,697

Number of Cities100 - 199K 200-299K 300-499K 500-999K 1.0 - 1.99m 2.0mn+ Total

1970 0 0 0 1 0 0 11980 2 0 0 0 1 0 31990 5 1 0 0 1 0 72000 5 0 1 0 0 1 72008 9 2 1 0 0 1 13

6

Road Infrastructure and UrbanizationTravel Time 1984

Source: Schmidt and Kedir (2009).

7

Road Infrastructure and Urbanization Travel Time 1994

Source: Schmidt and Kedir (2009).

8

Road Infrastructure and Urbanization Travel Time 2007

Source: Schmidt and Kedir (2009).

9

Ethiopia: UrbanizationOfficial Agglomeration

Estimate Index(percent) (percent)

1984 11.4% 3.7%1994 13.7% 7.1%2007 15.9% 14.2%

(mns people) (mns people)1984 4.55 1.481994 7.33 3.802007 11.72 10.50

(growth rate) (growth rate)1984-1994 4.9% 9.9%1994-2007 3.7% 8.1%1984-2007 4.2% 8.9%

Ethiopia is urbanizing faster than people think!!!

10

Ethiopia: Alternative Urbanization Estimates

0

2

4

6

8

10

12

14

1984 1994 2007

(mill

ions

)

Agglomeration Index Official CSA

11

Policy Questions

• To what extent will increased migration to urban centers accelerate economic growth via agglomeration effects?(i.e., spillovers from densely concentrated populations)

• Will urbanization lead to higher national poverty?(i.e., rural “brain-drain”; increased competition in non-farm product markets; lower agric. production and higher food prices)

• What are the trade-offs between investing in either rural areas, small towns, or major cities?(i.e., in terms of national economic growth or poverty reduction)

12

Drivers of Migration and Productivity Growth

• Rural-urban migration depends in part on expected relative incomes in rural and urban areas, and more broadly, on expected welfare (including social, health and economic benefits and costs)

• Positive agglomeration effects in cities may contribute to higher total factor productivity.

• Public investment in infrastructure enhances productivity of private investment.

13

Economic Structure: Ethiopia and Uganda

• Models built on consistent economy-wide data bases (social accounting matrices) with substantial disaggregation by economic sectors and household groups– 85% of the population resides in rural areas in each country– 20% of GDP generated in large cities (population > 50,000)– Poverty rates are significantly higher in rural areas

Population (%) GDP (%) Poverty rate (%)Ethiopia Uganda Ethiopia Uganda Ethiopia Uganda

National 100.0 100.0 100.0 100.0 40.0 40.0Rural 84.5 84.7 53.3 62.4 41.8 44.3Towns 10.2 9.3 26.4 18.0 35.5 23.7Cities 5.2 6.1 20.4 19.6 19.2 4.8

14

Characteristics of cities, towns and rural areas in Ethiopia and Uganda, 2005

Notes: Figures are calculated using sub-national SAMs. Population from household surveys. Poor population based on national poverty lines set at the 40th population percentile per capita expenditure level. Workers are separated into occupation groups.Source: Dorosh and Thurlow (2014).

Ethiopia Uganda Rural Towns Cities All Rural Towns Cities All Population (mil.) 60.3 7.1 3.6 71.0 23.0 2.5 1.7 27.2 Share (%) 84.9 10.0 5.1 100.0 84.7 9.3 6.1 100.0 Poor population (mil.) 25.2 2.5 0.7 28.4 10.2 0.6 0.1 10.9 Share (%) 88.7 8.9 2.5 100.0 93.8 5.5 0.7 100.0 National GDP shares (%) 53.3 26.4 20.4 100.0 62.4 18.0 19.6 100.0 Regional GDP shares (%) 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Agriculture 81.4 17.9 0.0 48.1 37.2 5.3 0.1 24.2 Industry 3.2 15.7 27.6 11.5 18.0 30.5 36.2 23.8 Services 15.3 66.4 72.4 40.4 44.8 64.2 63.7 52.0 Workers (mil.) 22.3 2.6 1.3 26.2 9.5 1.1 0.7 11.3 Skilled 0.1 0.4 0.2 0.7 0.3 0.1 0.1 0.5 Semi-skilled 0.9 1.2 0.8 2.9 1.5 0.4 0.4 2.4 Unskilled 21.2 1.0 0.4 22.6 7.7 0.5 0.2 8.4

15

Highlights of CGE Model Structure

• Sectoral and rural economy linkages– National product markets (some non-tradable local services)– Open economy model (i.e., imperfect substitution)– Fixed technical coefficients (i.e., IO table)

• Surplus labor in rural areas– Higher wages in urban industrial sectors

• Internal migration– Limited treatment of remittances

• Agglomeration and technical change

16

Accelerated Rural-Urban MigrationSimulation Results

Increasing urbanization by 10 percentage points (to 31 percent in Ethiopia; 29 percent in Uganda) results in…– Faster annual GDP growth in urban industry (+2.5 to

3.0 percentage points)– Faster annual GDP growth in agriculture (+0.7 to 1.3

percentage points) due to increased market demand for agricultural products

– Increase in overall per capita consumption for poor and non-poor by 0.3 to 0.4 percentage points

– Lower urban average per capita welfare as migrants to cities increase urban labor supplies

17

Public Investment Scenarios

• Raise share of public capital allocated to either cities, towns or rural areas by 10%

• Reallocation of existing resources– Increase budget allocation for target sector– Proportionately reduce allocation to other sectors

• Urban investment increases public capital (KPC) but reduces investment in agriculture

• Agriculture investment has a TFP-spending elasticity of 0.15 (causes agricultural output to increase/decrease for agric./urban sims)

Source: Dorosh and Thurlow (2014), “Can Cities or Towns Drive African Development? Economywide Analysis for Ethiopia and Uganda”.

18

Ethiopia: Development Budget 1999/00 and 2007/08

Source: Ethiopia Ministry of Finance and Economic Development data.

1999/00 2007/080

5

10

15

20

25

30

35

RuralUrbanOther

(bn

2007

/08

birr

)

19

Re-allocation of Public InvestmentSimulation Results

Increasing the share of public investment going to urban infrastructure results in…– Faster overall annual GDP growth (by about 0.2%

points) – More rapid growth in industry and services– Slower growth in agriculture– Higher levels of national welfare (price-adjusted

real consumption)– Higher levels of national poverty (since most of

the poor reside in rural areas and their incomes are linked to the agricultural sector)

20

Simulation Results: Household Welfare in Ethiopia and Uganda, 2005

Note: Welfare measured by equivalent variation, which is a consumption-based measure that controls for price changes.Source: Dorosh and Thurlow (2014).

Baseline growth

(%)

Deviation from baseline (%-point) Urban-

ization Investment scenarios

Cities Towns Rural Ethiopia Annual welfare change 1.73 0.41 0.08 0.10 0.10 Poor 1.17 0.25 -0.03 0.04 0.30 Non-poor 1.86 0.44 0.10 0.12 0.05 Rural areas 1.38 0.90 -0.02 0.04 0.30 Towns 0.72 -1.22 0.14 0.19 -0.14 Cities 1.69 -1.95 0.29 0.15 -0.22 Uganda Annual welfare change 2.45 1.31 0.31 0.20 0.01 Poor 1.50 1.05 0.16 0.08 0.29 Non-poor 2.59 1.35 0.33 0.21 -0.03 Rural areas 1.81 1.68 0.23 0.12 0.19 Towns 2.21 -0.71 0.30 0.34 -0.25 Cities 3.64 -0.25 0.59 0.31 -0.36

Residents of the areas of increased investment enjoy the largest welfare gains.

The biggest poverty reduction effects result from rural investments.

21

Concluding Observations• Migration/agglomeration are potential sources

of economic growth

• Model simulations indicate that:– Urbanization reduces the gap between average rural

and urban incomes – But without supporting investments, there is an

“urbanization of poverty” and rising urban inequality

• There is a growth-equity trade-off to increasing the share of public investment in urban areas (i.e., faster growth but lower welfare for the average poor household)

22

Sources

Chamberlin, Jordan and Emily Schmidt. “Ethiopian Agriculture: A Dynamic Geographic Perspective” in Dorosh, Paul and Shahidur Rashid (eds.), Food and Agriculture in Ethiopia: Progress and Policy Challenges. Philadelphia, PA: University of Pennsylvania Press.

Dorosh, Paul and James Thurlow. 2014. “Can Cities or Towns Drive African Development? Economywide Analysis for Ethiopia and Uganda”. World Development. http://dx.doi.org/10.1016/j.worlddev.2013.10.014 Dorosh, Paul and James Thurlow. 2012. “Agglomeration, Growth and Regional Equity: An Analysis of Agriculture- versus Urban-led Development in Uganda”, Journal of African Economies, 21(1): 94-123.