Download Study

General

Study Overview

Title:

Route to Development: Impacts of Road Network Improvements on Agricultural Intensification in Mozambique

Study is 3ie funded:

No

Study ID:

RIDIE-STUDY-ID-5f0d3682b0446

Initial Registration Date:

07/13/2020

Last Update Date:

12/20/2022

Study Status:

Ongoing

Location(s):

Mozambique

Abstract:

Investments in roads in underdeveloped rural areas are often assumed to stimulate agricultural production, ultimately allowing farmers to earn higher incomes. However, robust causal evidence on this is scarce. In part, this is because of a lack of data: few countries collect sufficiently comprehensive administrative data to permit detection of effects over large areas, and extensive direct data collection in remote rural areas is prohibitively expensive. This impact evaluation will measure the effect of a program of road improvements on agricultural intensification in rural Mozambique, combining freely available remote sensing data with high-frequency measurements of the state and usage of infrastructure on the ground. We will focus on the construction and maintenance of rural roads of various sizes in 12 priority districts in the provinces of Nampula and Zambézia. We will distinguish the causal effects of the intervention from other background trends that may also influence agricultural intensification by exploiting sharp changes in travel times generated by repairs and upgrades to roads and river crossings. We will leverage remote sensing data as well as advances in machine learning to measure changes in land use that would indicate agricultural intensification, such as increased conversion of land to fields, shorter fallowing periods, higher dry season vegetation indices, or changed land clearing practices. The approaches we use to measure agricultural intensification have the potential to be extended to other contexts where administrative data is lacking and data collection is costly. 

Registration Citation:

Categories:

Agriculture and Rural Development
Transportation

Additional Keywords:

Remote Sensing

Secondary ID Number(s):

Principal Investigator(s)

Name of First PI:

Paul Christian

Affiliation:

World Bank

Name of Second PI:

Anna Tompsett

Affiliation:

Stockholm University

Study Sponsor

Name:

DFID

Study Sponsor Location:

United Kingdom

Research Partner

Name of Partner Institution:

Administração Nacional de Estradas (ANE) – National Road Authority

Type of Organization:

Government agency (eg., statistics office, Ministry of Health)

Location:

Mozambique

Intervention

Intervention Overview

Intervention:

The program under evaluation focuses on the construction and maintenance of rural ‘feeder’ roads in 12 priority districts in the provinces of Nampula and Zambézia in Mozambique. The “Integrated Rural Feeder Road Development Program” is financed by the World Bank and implemented by the National Road Authority. Feeder roads connect communities to larger ‘access’ roads running between local administrative posts and district capitals, and at baseline were often in bad condition and impassable for larger vehicles. The upgrades are expected to provide farmers with better access to markets for agricultural inputs and outputs, either directly or through intermediaries. Within districts, feeder roads were selected for upgrading based on a prioritization score including factors such as agricultural potential, population or facilities along the road, as well as estimated upgrading costs. In each district, approximately 15 roads out of 30 candidates have from February 2021 onwards received upgrades.  A large percentage of the investments are construction and repair of bridges and culverts that improve accessibility, particularly during periods of heavy rain or flooding. During the duration of the project, we also expect periodic rehabilitations of the access roads.

Change History for Intervention

Changed On

Previous Value

12/20/2022

The program under evaluation focuses on the construction and maintenance of rural ‘feeder’ roads in 12 priority districts in the provinces of Nampula and Zambézia in Mozambique. The “Integrated Rural Feeder Road Development Program” is financed by the World Bank and implemented by the National Road Authority. Feeder roads connect communities to larger ‘access’ roads running between local administrative posts and district capitals, and at baseline are often in bad condition and impassable for larger vehicles. The upgrades are expected to provide farmers with better access to markets for agricultural inputs and outputs, either directly or through intermediaries. Within districts, feeder roads were selected for upgrading based on a prioritization score including factors such as agricultural potential, population or facilities along the road, as well as estimated upgrading costs. In each district, approximately 15 roads out of 30 candidates will receive upgrades.  A large percentage of the investments are construction and repair of bridges and culverts that improve accessibility, particularly during periods of heavy rain or flooding. During the duration of the project, we also expect periodic rehabilitations of the access roads.

Theory of Change:

Many believe that agricultural intensification is a precondition for poverty reduction in smallholder-based agricultural economies. The motivation for the intervention in our study context is the idea that local inaccessibility and high transport costs are a central barrier to agricultural intensification in the study area, which is identified as an area with high agricultural potential but very low current agricultural intensity. 

The underlying theory of change can be summarized as follows:

• Improving road surfaces and making river crossings passable will reduce inaccessibility and transport costs;
• Reduced transport costs and reduced periods of inaccessibility will improve access to modern inputs, such as improved seeds, pesticides, fertilizers, or irrigation equipment, either by i) directly reducing transport costs, improving availability and decreasing input prices, or ii) stimulating the entry of intermediaries into these markets, thus pushing down markups.
• Similarly, reduced transport costs and reduced periods of inaccessibility may allow farmers to sell products at higher prices or in greater quantities. 
• Improved access to modern inputs or output markets will make it possible for farmers to intensify agriculture. 

Multiple Treatment Arms Evaluated?

No

Implementing Agency

Name of Organization:

Administração Nacional de Estradas

Type of Organization:

Public Sector, e.g. Government Agency or Ministry

Program Funder

Name of Organization:

World Bank-IDA with co-financing from the Government of Mozambique

Type of Organization:

Foreign or Multilateral Aid Agency

Intervention Timing

Intervention or Program Started at time of Registration?

No

Start Date:

06/30/2020

End Date:

11/30/2022

Change History for End Date

Changed On	Previous Value
12/20/2022	11/30/2020

Evaluation Method

Evaluation Method Overview

Primary (or First) Evaluation Method:

Difference in difference/fixed effects

Other (not Listed) Method:

Additional Evaluation Method (If Any):

Other (not Listed) Method:

Method Details

Details of Evaluation Approach:

The causal relationship of interest is between transport costs and agricultural intensification. The general challenge in measuring this causal relationship is that transport costs may be related to agricultural intensification via many other channels than the causal relationship. To isolate variation in transport costs that is independent of other factors affecting intensification, our identification strategy exploits the fact that improvements to the road network create sharp changes in transport time for some locations, but not for others.

Our approach to analysis will compare changes in market access and land use in areas that experience upgrades to changes in market access and land use in comparable areas that do not.  We will evaluate a series of progressively more conservative approaches to estimation, narrowing the comparison progressively from all treated roads to all non-treated roads in program districts, to roads just above and just below program cutoffs, to areas which lie just on either side of river crossings which are upgraded during the program.  Narrowing the comparison group reduces potential sources of bias at the cost of reducing the sample size.  We will evaluate the plausibility of the comparison groups by comparing trends in treated and control groups before the intervention begins, exploiting the availability of remote sensing data back to 2016.  We will pre-specify our approaches to estimation and the specification tests we will use to evaluate their validity in a pre-analysis plan. 

Outcomes (Endpoints):

Our primary outcomes of interest are remotely-sensed indicators of agricultural intensification. Firstly, we calculate the share of cultivated area in total land available as a measure of adaptation at the extensive margin. Secondly, we use vegetation indices on cultivated land during the growing season as a measure of plant productivity reflecting input use.

For additional information, we use dry season vegetation indices to measure cultivation outside of the primary growing season and sowing season vegetation indices as indicators of weeding practices. Furthermore, we can use the cultivated area maps to construct indicators of the length of fallowing.

We will collect these measures continuously throughout the intervention period.

Our secondary outcomes are intermediate measures of the effects of the intervention. We will throughout the upgrading period collect primary data on the state of road infrastructure as measured through travel times, roughness indicators and number of inaccessible links. These measures will also form the base of remotely-sensed measures of road quality before, during and after the upgrading period. Furthermore, we will monitor market activity in terms of the number of participants in a set of weekly markets along the project roads, again combining ground-collected with remotely-sensed data.

These measures let us quantify the changes in market access, both directly through improved roads and indirectly through the number of intermediators trafficking these roads and markets.

Change History for Outcomes (Endpoints)

Changed On

Previous Value

12/20/2022

Our primary outcomes of interest are remotely-sensed indicators of agricultural intensification. Firstly, we calculate the share of cultivated area in total land available as a measure of adaptation at the extensive margin. Secondly, we use vegetation indices on cultivated land during the growing season as a measure of plant productivity reflecting input use. For additional information, we use dry season vegetation indices to measure cultivation outside of the primary growing season and sowing season vegetation indices as indicators of weeding practices. Furthermore, we can use the cultivated area maps to construct indicators of the length of fallowing. We will collect these measures continuously throughout the intervention period. Our secondary outcomes are intermediate measures of the effects of the intervention. We will throughout the upgrading period collect primary data on the state of road infrastructure as measured through travel times, roughness indicators and number of inaccessible links. Furthermore, we will monitor market activity in terms of prices, goods traded and participants in a set of markets along the project roads. These measures let us quantify the changes in market access, both directly through improved roads and indirectly through the number of intermediators trafficking these roads.

Unit of Analysis:

Pixels will be aggregated into polygons along road segments (the unit of analysis), around river crossings or within populated areas. The size of these areas of identification of spatial trends.

Hypotheses:

Primary hypotheses

• Improvements to rural roads lead to agricultural intensification.

Secondary hypotheses

• The intervention reduces transport times
• The intervention reduces inaccessibility (eg, impassible roads due to wash-outs in the rainy season)
• The intervention increases the number of both sellers and buyers in rural markets

Heterogeneity analyses

• Impacts of improved roads are stronger when roads improvements are associated with relatively bigger changes in market access measures.

Change History for Hypotheses

Changed On

Previous Value

12/20/2022

Primary hypotheses Improvements to rural roads lead to agricultural intensification. Secondary hypotheses The intervention reduces transport times The intervention reduces inaccessibility (eg, impassible roads due to wash-outs in the rainy season) The intervention increases the number of both sellers and buyers in rural markets The intervention increases input availability and reduces the prices of inputs The intervention increases output prices Heterogeneity analyses Impacts of improved roads are stronger when roads improvements are associated with relatively bigger changes in market access measures.

Unit of Intervention or Assignment:

The unit of intervention is a road segment, with an average length of just over 10km.

Number of Clusters in Sample:

The program is implemented at the road segment level.

Number of Individuals in Sample:

For road level analyses: approx. 300 road segments, stretching for 3100km.

Size of Treatment, Control, or Comparison Subsamples:

We expect approximately 150 of the 300 road segments to be treated.

Supplementary Files

Analysis Plan:

Other Documents:

Data

Outcomes Data

Description:

We will use two main datasets for our primary analysis. The first consists of repeated measurements of quality and usage indicators of road segments along the project roads, collected during the duration of the upgrading works and supplemented by remotely-sensed measures of road quality. The second uses Sentinel-2 satellite imagery to derive measures of the outcomes of interest for areas along the project roads. In secondary analyses, we will use remotely-sensed market activity measures.

Change History for Description

Changed On	Previous Value
12/20/2022	We will use two main datasets in our analysis. The first consists of repeated measurements of quality and usage indicators of road segments along the project roads, collected during the duration of the upgrading works. The second uses Sentinel-2 satellite imagery to derive measures of the outcomes of interest for areas along the project roads.

Data Already Collected?

No

Data Previously Used?

Data Access:

Data Obtained by the Study Researchers?

Data Approval Process:

Approval Status:

Treatment Assignment Data

Participation or Assignment Information:

No

Description:

We have collected this information from the documentation provided by the National Road Authority to the funder of the intervention.

Change History for Description

Changed On	Previous Value
12/20/2022	We will collect this information from the documentation provided by the National Road Authority to the funder of the intervention.

Data Obtained by the Study Researchers?

Yes

Change History for Data Status

Changed On	Previous Value
12/20/2022	No

Data Previously Used?

No

Data Access:

Not restricted - access with no requirements or minimal requirements (e.g. web registration)

Data Obtained by the Study Researchers?

Yes

Data Approval Process:

Approval Status:

Data Analysis

Data Analysis Status:

Study Materials

Upload Study Materials:

Registration Category

Registration Category:

Prospective, Category 1: Data for measuring impacts have not been collected

Completion

Completion Overview

Intervention Completion Date:

Data Collection Completion Date:

Unit of Analysis:

Clusters in Final Sample:

Total Observations in Final Sample:

Size of Treatment, Control, or Comparison Subsamples:

Findings

Preliminary Report:

Preliminary Report URL:

Summary of Findings:

Paper:

Paper Summary:

Paper Citation:

Data Availability

Data Availability (Primary Data):

Date of Data Availability:

Data URL or Contact:

Access procedure:

Other Materials

Survey:

Survey Instrument Links or Contact:

Program Files:

Program Files Links or Contact:

External Link:

External Link Description:

Description of Changes:

Study Stopped

Date:

Reason: