1. Background and purpose
Forecast-based Financing (FbF) is a funding modality and way of working whereby resources are released based on forecast information before a climate-related disaster happens, to implement planned activities which reduce risks, enhance preparedness and response, and make disaster risk management overall more effective.
The Bangladesh Red Crescent Society (BDRCS) setup the FbF mechanism and developed Early Action Protocols (EAP) for floods and cyclones to help vulnerable households reduce the anticipated hazard impacts, based on forecast information. The EAP for cyclone was activated on 18 May 2020 for cyclone Amphan and early actions were implemented by BDRCS and CPP, reaching around 18,300 beneficiaries in 64 shelters under 10 districts with Early Actions. The EAs included distribution of dry food, safe drinking water, ORS, mask, hand sanitizer and soap. The EAP for floods was activated on 25 June 2020. BDRCS reached 3789 households in 10 Unions under 3 districts with cash support for early actions. BDRCS also supported 70 households to evacuate their assets.
The purpose of this contract is to engage a consultant experienced in quantitative survey data analysis following a quasi-experimental study design. The consultant will analyse the survey data to assess whether and to what extent forecast-based assistance with cash grants, food and non-food items, and evacuations helped the beneficiaries to avoid or reduce negative impacts of the cyclone and floods compared to similarly vulnerable and affected households who did not receive forecast-based assistance.
2. Scope of work and research questions
The consultant is asked to perform a state-of-the-art analysis of the available quantitative data sets for each intervention that have been collected following a quasi-experimental study design. The beneficiary group and a comparison group per each intervention are represented in the sample as shown below:
Overview of total sample size and distribution across groups for floods
- Group for floods: BDRCS / number of beneficiaries: 3,800 households / sample size: 200 households
- Comparison group / N/A / 220 households
Overview of total sample size and distribution across groups for cyclone Amphan
- Group for cyclone: BDRCS / number of beneficiaries: 18,300 households / sample size: 240 households / from # of shelters: 24
- Comparison group / N/A / 220 beneficiaries households / 22 shelters
The consultant will analyze the main research questions pertaining to the effectiveness and efficiency of the interventions vis-a-vis the comparison group. Therefore, advanced statistical methods suitable for publication in a peer-reviewed journal are required to test results for statistical significance, and to account for potential sources of bias in the dataset. The two interventions (floods and cyclone) have to be assessed separately.
Research questions:
- To what extent are the BDRCS intervention group and the comparison group comparable, for each intervention? What are the main differences between the groups? How can they be statistically accounted for?
- Have the interventions been effective in helping vulnerable households to protect their lives and livelihoods from the cyclone or flood, as measured by animal morbidity and mortality?
- Have the interventions been effective in helping the beneficiaries to avoid negative coping strategies, including: a) Emergency sales of livestock (destocking) or crops at deflated prices; b) Selling other valuable assets (destitution sales); c) Taking on - potentially unsustainably high or high-interest - new debt?
3. Specific tasks and deliverables
A suitable statistical software package should be used by the contractor to analyse the survey data (Stata, SPSS, R, etc). The syntax files (.do / .SPS / .R) documenting all operations performed on the data set should be stored and submitted together with the data analysis report.
Data cleaning: cleaning procedures should be documented in syntax files, including identification of outliers or missing data points.
Balancing and weighting the sample data: Since it can be expected that there will be differences between the intervention groups and the comparison groups, steps must be taken to account for these differences. In past analyses of FbF projects, propensity score matching (PSM) and the use of bias-corrected matching estimators have been found to be suitable solutions, although the survey contractor may suggest a different approach. All estimates from the survey data must be weighed by applying base weights that are adjusted for non-response error and unequal selection probabilities (a detailed description of the process by which provinces, districts and households were selected will be provided to the survey contractor).
Impact data analysis: For each questionnaire section, the survey contractor will produce summary tables and bar charts for all main variables, showing the difference between the intervention groups and the comparison households side-by-side (in one table and bar chart, respectively). Tables must clearly indicate the statistical significance of the differences between (intervention and comparison) means, or lack thereof, by including p-values or following a consistent highlighting process, for example, by including asterisks (*). Means comparisons should be adjusted to control for multiple comparisons; Holm’s correction method has proven to be a suitable solution in past analyses.
Final products: The results of the analyses should be summarized in an analytical summary report, plus a PowerPoint presentation, and should be complemented with all tables and bar charts provided in an annex.
All work must be original and follow the highest research and ethics standards.
Deliverables
- Complete set of analytical tables and charts for all questionnaire variables & summary PowerPoint presentation; 4 days
- Summary write-up of key findings on the impact of the interventions on the beneficiaries’ socio-economic well-being and livelihood (25 pages max., excluding annexes); 18 days
- Summary presentation of findings to BDRCS/RCCC/IFRC; 4 days
- Submission of complete, cleaned electronic dataset, statistical syntax files and all supplementary documentation; 2 days
4. Reporting and work arrangements
The contractor will formally report to the International Federation of Red Cross Red Crescent Societies (IFRC), Asia Pacific Regional Office (APRO), Kuala Lumpur, Malaysia. The work will be substantively supervised by the BDRCS, Red Cross Red Crescent Climate Centre (RCCC) and IFRC focal points for this study. Upon signing the contract, and once the contract is activated, regular progress reports against the consultant’s work plan must be submitted to the supervisor in a form and frequency to be determined by the supervisor.
Payment is subject to satisfactory performance and completion of all deliverables.
All submissions will be made electronically (email, Skype, phone, etc.), unless requested otherwise by the BDRCS/IFRC team.
5. Duration of the consultancy
The contract is entered between the International Federation of Red Cross Red Crescent Societies (IFRC), Asia Pacific Regional Office (APRO), Kuala Lumpur, Malaysia, and the contractor. The estimated level of effort is 28 working days. The approximate start date is 10 February 2021. All tasks are expected to be implemented and completed, with final products delivered by 22 March 2021.
6. Required qualifications and experience of the consultant
● Graduate or advanced (PhD) degree in relevant disciplines such as statistics, demography, economics, sociology or related disciplines
● Proven experience in quantitative research for impact assessment
● Proven experience in analyzing probabilistic sample survey data, including sample data weighting and propensity score matching (submit work samples together with application)
● Experience managing large datasets
● Knowledge of R, Stata, SPSS (or similar) and quantitative data analysis methods and documenting analytical operations in syntax files
How to apply:
The deadline for applications is 31 January 2021 at 23:59 hrs Kuala Lumpur time.
Interested applicants should submit the following documents electronically to raymond.zingg@ifrc.org:
● Short technical proposal summarizing:
○ Qualifications and relevant experience with similar assignments
○ Reflections on the research design
○ Draft work plan showing the suggested timing of carrying out the specific tasks
● List of publications and completed quantitative survey analyses, in tabular format
● 2 samples of previous work that show the expertise and experience of the consultant; samples must be the original work of the contractor (e.g. analytical reports or peer-reviewed journal articles).
● Financial proposal including detailed budgets for professional fees and reimbursable expenses (if any) for all tasks related to this assignment. No payments will be made outside the contractually agreed budget.
Incomplete applications or applications that are received late will not be considered. Only shortlisted candidates will be contacted.