Annex: SPRING Pathways to Better Nutrition Budget Methods - Nepal

SPRING has adapted guidance from several sources to develop a methodology for extracting nutrition-specific and nutrition-sensitive funding data from donor and government budget documents. This document is an annex to the Pathways to Better Nutrition (PBN) Case Study Series that outlines SPRING’s methodology and includes limitations of the data presented, as well as references for further information.

Background

The objective of SPRING’s “Pathways to Better Nutrition” (PBN) analysis of Nepal’s nutrition budgets is to provide stakeholders with:

1. An estimate of funding budgeted for Nepal’s Multi-Sector Nutrition Plan (MSNP) activities in FY 2013/2014, FY 2014/2015, and FY 2015/2016. This will be useful for comparison to the estimated costs to implement the MSNP plan, and for understanding gaps in nutrition funding. The data can also be used to plan government and donor nutrition funding, and to advocate for greater and more consistent nutrition funding.
2. Information on which activities are prioritized financially each year within the MSNP. This includes information on funding sources for each nutrition-related activity, whether funding has been shifted from other activities, and the balance of government and donor funding for the nutrition activities.
3. Budgeting tools and guidance to help nutrition stakeholders in Uganda more explicitly track and advocate for nutrition funding. This can help with reporting not only within Uganda but also for groups such as the “Scaling Up Nutrition” (SUN) Movement, which prioritizes financial tracking in its monitoring and evaluation of countries.

Defining Budget Analysis

Political will for nutrition must be reflected through financial support at the national and subnational level (USAID 2014). There are several steps involved in tracking financing support. Costing a national nutrition plan provides estimates for what amount of funding is necessary to implement nutrition activities; analysis of current budgets (government and donor) provides estimates for what funding is actually allocated to implement nutrition activities; analysis of expenditures to estimate what percent of allocated funds were spent; and expenditure tracking to find why funds did not reach their intended destination.

The government of Nepal and their partners have supported the first step of this process—the costing of the MSNP in Nepal. SPRING is primarily focused on the second step: estimating what funding is allocated to implement the nutrition activities in the MSNP, and to the extent that there are available data, how much of that funding was spent. This is what SPRING generally means by ‘budget analysis’ for purposes of this brief.

Budget analysis can be defined as applied analysis of government and donor budgets with the explicit intention of impacting a policy debate or furthering policy goals (The International Budget Project 2001). This work can include efforts to improve budget literacy of policymakers, program planners, and other key stakeholders. In the case of Nepal, SPRING’s budget analysis is meant to better inform the stakeholders advocating for the MSNP of their available resources. This can lead to more effective advocacy for greater nutrition funding, more transparency in how those funds will be spent, and clearer negotiation for donor funding.

To the extent possible, SPRING is also addressing what percent of funds were spent for nutrition activities. This will depend on the data available in Nepal and the strength of the government expenditure tracking systems. SPRING will not address the final step of financial analysis, which is to identify the reasons behind bottelnecks in spending, as this type of work is best done through other methods, such as the World Bank’s Public Expenditure Tracking Survey (PETS) or Public Expenditure Reviews.¹

Defining Nutrition Activities

The scope of nutrition is quite difficult to define, yet a clear definition is needed for budget analysis and financial tracking. The MSNP is used as the definition of the boundaries of activities that can be included for this analysis. There are several advantages to this, as well as a few drawbacks.

The MSNP contains an explicit implementation matrix (Annex I of the MSNP) that defines the interventions in support of the MSNP, expected outputs, the government agency responsible for leading each activity, and other participants. There is also an approximate cost assigned to each activity is included in the consolidated cost action plan within Annex I of the MSNP. The advantages of using this scheme are that the activities are set for the five-year period of the MSNP, allowing SPRING to follow the same set of activities over that time.² It also means that estimated financial allocation and expenditures can be compared to the costing for the plan. Finally, by having both the qualitative and budget research teams work from the same document, it aligns the budget analysis with the qualitative assessment of prioritization.

One drawback is that some activities that SUN includes on its “nutrition-sensitive” list for the Development Assistance Committee (DAC) guidance will not be counted in our analysis. Each country has latitude to include or exclude any of these activities, and in the MSNP certain activities received less emphasis. Qualitative enquiry can probe the reasons for the differences between the SUN definition and what appears as sensitive in the MSNP (see Appendix 1 for the SUN list), but for the budget analysis, excluded activities will not count toward the total estimated nutrition allocation or expenditure.

Where SPRING had any confusion of meanings or terms, we brought these questions up in our interviews with the lead agency assigned to the MSNP activity. In a few cases, new activities not in the MSNP were identified by key informants in the interviews as nutrition activities. When this occurs SPRING allows the inclusion of any budget items related to that activity that still fall within the SUN definitions of nutrition-sensitive. All final budget lists are validated by the ministry or donor responsible, as a last check on the validity of the budget analysis.

SPRING did an abbreviated examination of funding at the district level as well. With the institution of the MSNP line item in the 2014/15 fiscal year for the 6 priority districts, this allows anyone to follow the designated nutrition funding flows for the MSNP. This code does not cover implementation of nutrition activities, but of coordination, training, and planning activities to roll out the MSNP. SPRING also collected the same qualitative information as at national level on nutrition activities at the district level in three districts, Parsa, Achham, and Kapilvastu. Donor partners were tracked and where district funding data are available, they will be reported in the final analysis of the PBN case study.

Methods

The PBN case study is a prospective mixed-methods study. Budget analysis is an integral part of the study design, to compare with results of the qualitative data on activity prioritization and feed further inquiry into planning for nutrition. There are no standard documented methods for extracting budget data, especially for a subsector such as nutrition. For its methodology for extracting nutrition-specific and nutrition-sensitive funding data from donor and government budget documents, SPRING adapted guidance from several sources:

• SUN donor network guidance for tracking global investments in the Development Assistance Committee database (DAC) (SUN Donor Network 2013).
• Examination of the MSNP implementation matrix (GoN and NPC 2012).
• Cross-country learning with SPRING’s Uganda PBN case study.
• Consultation with the Nepali government ministries and key donors.

SPRING will collect and analyze budget data for three budget cycles: 2013/2014; 2014/2015; and 2015/2016. Data will be collected at the national level for government, donor, and UN groups, and an abbreviated look at secondary budget documents will occur in three districts for government and selected donor funds. Figure 1 gives an overview of the process of data collection, validation, and analysis.

Figure 1: Summary of SPRING’s Budget Methodology

The process for data extraction and analysis described below was used to address Objectives 1 and 2 of the budget analysis. SPRING will document this process and develop tools to help others replicate this analysis by the end of the study to meet Objective 3.

Data Collection

National Level

National-level data were gathered during baseline data collection in July 2014 and will be repeated for the next two budget cycles. The team conducted qualitative and budget interviews with stakeholders from the six key groups named by SUN for scaling up nutrition activities:

• Government (ministries as well as the nutrition coordinating body, National Planning Commission
• Donor agencies
• CSOs
• Business/private sector
• UN groups
• Academic/research institutions

SPRING requested “red book” budgets, supplemental documents, work plans, and any other documents needed to identify nutrition funding for each of the groups bolded from the above list. For the other groups, SPRING inquired about approximate funding for their nutrition work and source of funding but did not pursue the full budgeting exercise.

There are overlapping funding lines in these groups, particularly for donor and UN agencies. Many bilateral donors provide funding to UN agencies and to the Government of Uganda. When funding UN agencies, bilaterals rarely identify the funding as nutrition, which means the UN agency decides how to allocate those funds within the larger category of giving. SPRING chose to follow donor and UN funds at the project level, rather than starting from the top, i.e., global allocation level. Off-budget donor and UN activities can be identified through the MOF’s Aid Management Portal. This captures all external financing (on- and off-budget as well as a few CSOs). All on-budget financing of MSNP activities was identified within the Red Book, the official budget document across government ministries. On-budge donor funds appear in both of these sources, and were triangulated in the case of disagreement and verified via interviews with the donor. All data from government sources and all major donors were validated in follow up interviews after initial analysis to ensure that the correct activities, amounts, percentages, and sector assignations were used.

District Level

Due to resource constraints, the district level analysis in Nepal was not as extensive as in the Ugandan PBN study. In Nepal, SPRING collected government proposed, confirmed and spent budgets for the MSNP line item funds starting in the 2014/15 fiscal year. This line item was discussed in all relevant district level interviews with:

• Government (the seven sectors funded through this line item)
• UNICEF representative (who provided MSNP funding for the health sector)

In addition, SPRING was able to collect qualitative information on the budgets and budgeting process from:

• Primary nutrition donor agencies operating in each district
• CSOs that carry out donor activities in each district

Where full district government budgets were also available, these will be used to augment the final analysis, however the primary stream of budget data being followed at the district level in Nepal is the MSNP line item and the primary donor agency funding.

Data Processing and Analysis

National and District Level

Nutrition-Specific Activities

Within the sources above and the activities in the MSNP, SPRING largely follows the USAID Nutrition Strategy definition of nutrition-specific activities:

• Management of severe acute malnutrition
• Preventive zinc supplementation
• Promotion of breastfeeding
• Appropriate complementary feeding
• Management of moderate acute malnutrition
• Periconceptual folic acid supplementation or fortification
• Maternal balanced energy protein supplementation
• Maternal multiple micronutrient supplementation
• Vitamin A supplementation
• Maternal calcium supplementation 

This matches the list provided in the executive summary of the 2013 Lancet Series (Lancet 2013). SUN guidance for the identification of nutrition-specific activities was also based on the Lancet Series’ (2008 and 2013) set of interventions.

The SUN guidance for tracking global investment in nutrition (Mucha 2012; SUN Donor Network 2013) does not provide a definition past use of the “basic nutrition” DAC purpose code. In the DAC, the definition of this code is:

Direct feeding programs (maternal feeding, breastfeeding and weaning foods, child feeding, school feeding); determination of micro-nutrient deficiencies; provision of vitamin A, iodine, iron etc.; monitoring of nutritional status; nutrition and food hygiene education; household food security.
(OECD website, “Purpose Codes: sector classification” and “2012 CRS purpose codes_excel EN”).

According the guidance given by SUN, 100 percent of the funds assigned to a “nutrition-specific” activity will be counted toward the total (no weighting applied).

Nutrition-Sensitive Activities

The SUN financial tracking guidance outlines its approach for identifying and weighting nutrition-sensitive activities from the DAC. SPRING modified this guidance to align with the MSNP and to be relevant for both government and donor funding. These modifications have added one additional step (2). The overall approach and SPRING’s modifications to SUN’s approach can be summarized by Figure 2:

FIGURE 2: SPRING’s Modified Validation Approach

Broken out, this can be explained as follows:

1. Identify the pool of potentially nutrition-sensitive projects and budget line items: SUN suggests using a combination of DAC codes and a key word search on the CRS database for donor activities. The lists of DAC codes and key words are presented in Appendices 2 and 3. SPRING MODIFICATION: SPRING’s roster of potentially nutrition-sensitive activities is derived from the defined activities in the UNAP implementation matrix. While many areas overlap with the DAC descriptions, there is some divergence, and the level of detail is greater in the UNAP than in the DAC (see “defining nutrition activities” section above).
2. Integrated or Stand Alone Activity: very often in government budgets, and sometimes in donor activities, nutrition-relevant activities are “integrated” into larger non-relevant activities. Therefore SPRING had to undertake this additional step to allow for counting of partial components of the larger line item. In validation interviews, or via reading of project documents, SPRING endeavored to find out what percentage of a line item is nutrition-relevant.
3. Nutrition-Sensitive vs Specific: SUN suggests reviewing the projects selected in Step 1 by assessing individually each project document. The objectives, expected results, and indicators are examined to determine whether the project is nutrition-sensitive. SUN requires the activity to pass three criteria: 1) project must intend to improve nutrition for women, adolescent girls, or children; 2) project has a significant nutrition objective OR nutrition indicator(s) (see Appendix 3); and 3) project must contribute to explicit nutrition-sensitive outcomes (through activities, indicators, and results; see Appendix 1). SPRING MODIFICATION: SPRING modifies the list of nutrition-sensitive outcomes to match the UNAP activity outputs. If there is an activity on SUN’s list that is not in the UNAP, that activity would not be counted in SPRING’s budgeting, unless it is given as a nutrition activity in our interviews.² For instance, in SUN’s list, improving access to education/school to adolescent girls is not a UNAP activity, so a project with that as its only nutrition-sensitive outcome would not be counted in SPRING’s budgeting. A school feeding program would be counted however, as that is a UNAP activity
4. For Nutrition-Sensitive, is it “Dominant” or “Partial”: Through the same review of project documents, classify the “intensity” of nutrition-sensitivity into two sub-categories: nutritionsensitive dominant or nutrition-sensitive partial. SPRING MODIFICATION: If no other information for a project is available, SPRING will use SUN’s weighting scheme (100 percent of funding is counted if a project’s main objective, results, outcomes, and indicators are nutritionsensitive; 25 percent if secondary objective, results, outcomes, and indicators are nutritionsensitive). However, SPRING has access to work plans or donor budgets and if there is insufficient information in these document to determine the approximate percent, SPRING will ask stakeholders to define breakdown for accounting. If SPRING still cannot define percent after these consultations, the SUN weighting scheme be applied. Documentation of our decisions will be made for each activity.

Data Validation Process

SPRING is taking a two-pronged approach to ensure high-data quality. First, within our team, the following steps are taken in order to ensure inter-rater reliability:

1. Regular group extraction meetings
2. Feedback on ambiguous terms to OPM and NPA for guidance
3. Notation and documentation in extraction sheets
4. Cross-referencing figures from multiple sources, where available

Once extraction is completed, SPRING confirms the validity of the extracted ministry and donor budget data through meetings with the key informants for that ministry or donor. Every effort is made to cross-validate data with the sector focal point seconded to the MOF. Any projects or activities that cannot be validated by the country or global team (donors) or line ministry and NPC (government) will be dropped from the analysis. Any unlisted projects named by the key informants will require supplemental documentation in order to be added to the analysis.

Exchange Rates

MOF reports off-budget donor funding in current-year USD. However, all ministry budget data is reported in current-year Nepali Rupees (NRs). SPRING is reporting final estimates in both USD and NRs. Inter-bank exchange rates from the Nepal Rastra Bank will be used for the conversions, averaged over the first month of the fiscal year.

Deflation/Inflation Rates and Base Year

National level analysis will begin at 2013/2014. For yearly reporting, no modifications are made to the reported figures in USD but for aggregated reporting of more than one year or reporting trends, SPRING uses 2013/2014 as the base year and succeeding years are adjusted to base-year dollars. Inflation rates are averaged over the fiscal year using the World Bank GDP-Deflator/Nepal Central Bureau of Statistics Producer Price Index.

Limitations

Missing Data and Non-Response

In the Aid Management Portal, while the list of donor projects was comprehensive and total project commitments were reported for all entries, actual yearly commitments and disbursements were missing for many activities. Where we could not complete this data from our interviews, SPRING imputed the missing data from the total project commitment figure divided by the number of project years (an s-curve will also be assessed as an alternative). Since the majority of donor projects are on-budget, and those figures are recorded in the red book, this applies primarily to off-budget donor figures.

Data Quality

Setting aside from the yearly actual commitments and disbursements data in the Aid Management Portal, the data recorded in the government records appeared to be of good quality. In validation interviews with government officials, only one correction was made to a disbursement figure for 2013/14 and 2014/15.

Changes over Time

SPRING is comparing data over several budget cycles, so it is important to use the same standards each round for comparability. However, as ministries become more aware of nutrition and “nutrition-sensitive” activities via the roll out of the MSNP, their accounting for activities may change and a greater number of activities may be identified as nutrition-sensitive, even if they existed in previous budgets. SPRING is making every effort to return to previous years’ data after each new round to check that “new” activities are indeed new and not just re-categorized.

It is also important to note that with the April 2015 earthquake, there will be a structural break in the trend SPRING is examining for this study between the second and third study year. Thus, while this information is still very useful for the objectives set out by the study, one would not expect to see a steady upward march for nutrition funding over this time period.

Subjectivity of “Sensitive”

Defining "nutrition-sensitive" can be complicated. Within the data analysis team, SPRING ensures inter-rater reliability through regular group extraction meetings to discuss ambiguous activities listed in MSNP and cross-verifies final lists with the source ministry or donor organization.

Evolution of Nutrition Designation

Changes in the designation of nutrition-sensitive categories and how to weight them in this type of analysis, at the global and national levels are likely. The MSNP is not expected to change until 2017, but modifications could be made. If SUN guidance continues to evolve, SPRING will work with NPC to evaluate whether the improvement in accuracy is worth the loss of comparability over time.