Cash or in-kind? Why not both? Response Analysis Lessons from Multimodal Programming


Humanitarians maintain “do no harm” as a fundamental operating principle. However, CaLP North America’s 2016 needs assessment found that the level of rigor and documentation applied in identifying and evaluating evidence about the effectiveness, appropriateness, efficiency, feasibility, and risk of different response options to decide on the “best fit” for a given humanitarian problem and context varies greatly between organizations. This process, known as “response [options] analysis,” justifies response decisions and demonstrates due diligence efforts to do no harm. As such, it is increasingly required for proposals to donors, including the recently released USAID Office of Food for Peace’s Emergency Food Security Program 2017 Annual Program Statement.

This research reviews lessons learned about response analysis from multimodal responses, that is, responses in which practitioners determined that more than one response modality between cash, vouchers, and in-kind, was a “best fit” or in which the conclusions about “best fit” changed over the course of the project.1 The research hypothesizes that comparing the reasons for choosing different types of response within the same project and among the same beneficiaries should provide concrete examples of the relative importance of different criteria in response and, by extension, the conditions under which cash or vouchers or in-kind assistance may be most appropriate.

This work does not aim to inform technical or operational considerations of how to do cash transfers or multimodal programming. Rather, this work is intended to help analysts, advisors, and decision-makers develop and articulate the nexus between emergency context and response through concrete examples of response analysis, both at project design and throughout a response.

The cases analyzed in this study span the globe and include low-, middle-, and high-income countries. The crises include sudden-onset and slow-onset natural disasters, as well as sudden-onset and protracted conflict. While some lessons may be reflected in multiple case studies, the main messages from each case are:

Beneficiary-centered response is a common theme in current political discourse; humanitarians tend to design responses to fit the average needs of a typical beneficiary. The first case study with the Canadian Red Cross response to the Alberta wildfires in 2016 provides a concrete example of what one approach to beneficiarycentered response looks like.

The South Sudan case study with World Vision illustrates the importance of beneficiary preference and buy-in to any assistance project, particularly in conflict contexts in which protection is a major concern. This case also demonstrates why response analysis is not a one-off analysis at project design, but a continuous analysis throughout the project cycle.

Similarly, CRS in Guatemala found that strong response analysis allows assumptions, such as about the significance of real and perceived risks to beneficiary safety or of restrictions vs. messaging on dietary quality, to be identified, documented, monitored, tested, and revised based on evidence.

In the DRC, both CRS and Samaritan’s Purse have very similar USAID-funded emergency programs with clearly defined and well-practiced assessment and response analysis processes that explicitly allow for the most appropriate response, switching modalities over the course of implementation as appropriate.

In Nepal, the experience of the Red Cross Red Crescent Movement illustrates the significance of organizational capacity and preparedness in response analysis. Also in Nepal, CRS demonstrated an alternative approach to beneficiary-centered response allowing for variation in market appropriateness in different areas.

Oxfam in Myanmar demonstrated the value of evidence-based response analysis to building buy-in for response changes among beneficiaries, authorities, and donors, particularly in sensitive and protectionchallenged environments.

General conclusions drawn from multiple cases include:

An analysis of multimodal programming expects that there may often be multiple “right” responses in a given context. It is because the various goods and services needed may involve different supply chains with different regulatory environments (imported commodities, locally produced commodities, telecoms, banksetc.), shock affectedness, and cultural contexts (including willingness to pay) that different modalities may be appropriate.

Despite these differences, the general information categories needed for response analysis are quite similar across sectors suggesting that a harmonized approach to response analysis would be appropriate.

Interestingly, the case studies did not demonstrate financial cost savings as a major decision-making factor for choosing between cash, vouchers, or in-kind assistance. Rather than cost, the most influential arguments in deciding a given modality mix included: time to delivery, organizational capacity (human resource availability or experience), market supply response capacity (trader capacity), security, and beneficiary preference.

Respondents felt the human-resource cost of preparing or maintaining multiple pipelines for more or less simultaneous delivery was high, suggesting that it may not be reasonable to assume efficiencies through simultaneous distribution of these multiple types of goods and services, particularly at the beginning of a project or in a true emergency context (Nepal, Guatemala). On the other hand, switching modalities once multiple pipelines and products are established is relatively straightforward (Canada, South Sudan, Myanmar, DRC).

Careful assessment and validation; slow, progressive change; and well-communicated contingency planning contribute to trust and change acceptance. Engaging stakeholders in rigorous situation and response analysis contributes to acceptance and trust and is particularly important in volatile and sensitive contexts.


In addition to illustrating lessons about how organizations decide on a modality or project design from a strategic response analysis perspective, the case studies also generated some lessons about enablers and barriers in implementing multimodal programming.


The case studies demonstrate that a favorable “mindset” or level of “buy-in” was a significant factor in being able to implement simultaneous modalities or to switch. The case study research provides indicative evidence about buy-in among the following stakeholders: senior technical advisors and management, beneficiaries (South Sudan), local authorities (Myanmar), organization’s staff (Guatemala), and donors (Myanmar).

In general, systematic, evidence-based response options analysis, both at project design and as a part of project monitoring, may be more inclined to result in multimodal response decisions. Similarly, different mindsets or levels of support for multimodal programming may be rooted in differences in prioritization of the criteria being evaluated in response options analysis.

Support from senior management and technical leadership was pervasive, though it may not be explicitly mentioned in the case studies. These stakeholders are most comfortable with optimizing across multiple response analysis criteria in diverse contexts. They demonstrate a commitment to evidence-based response options analysis with balanced, context-appropriate priorities by using clear, reliable, timely information from assessments as inputs to situation and response analysis for project design and monitoring (adaptation) committing to identify and test assumptions and concerns to build evidence and experience to inform response communicating response analysis and project adaptations clearly to all stakeholders to ensure support.

However, senior management support for multimodal programming of multiple types of pipelines may not be effective without the commitment of additional human resources as compared to programming a single commodity type. It takes more effort, particularly preparedness effort, to deliver from multiple pipelines with different actors and regulatory environments (e.g., imported rice vs. mobile money) than it does from one pipeline or from multiple pipelines within the same regulatory environment (e.g., imported rice and imported wheat). Leadership may demonstrate support by committing funds to piloting or to full-scale implementation as well.

Myanmar and South Sudan demonstrate that when support from local authorities or beneficiaries is low, it will take more time than expected to build confidence and trust in change. In these cases, ethnic tension was at least a partial driver of the emergency context. As such, differences in ethnicity between actors predominant at different points in the supply chain prompted a higher prioritization of security and protection concerns among beneficiaries and authorities. Reducing change expectations by taking incremental steps (i.e., piloting in smaller areas and not switching from 100% in-kind rations to 100% unconditional cash all at once) and investing in contingency pre-positioning of preferred methods are likely to be appropriate in this context.

If an organization’s staff has reservations about the different modalities applied or the switch, their ability to creatively identify the best solutions during implementation may be compromised. Experience is a strong argument for this audience. CRS’s approach to building internal support for cash and voucher programming throughout the organization is widespread piloting, demonstrated here in the Guatemala and DRC cases using electronic vouchers. CRS uses strong response analysis to identify staff assumptions and acknowledge their concerns. Through a small pilot, staff experience the advantages and disadvantages of change and work together through additional mitigation strategies that might be appropriate. This process may help to inform and align response options analysis priorities in a given context. The proof is in the pudding.

In addition to funding explicitly multimodal programs, donors may enable or inhibit modality switching, in particular through the degree of budget management authority or flexibility they delegate to implementing organizations. For example, it would be difficult for an implementer to change modalities in response to changing context if the donor requires approval for changes of more than 10% to predefined budget lines because of the differences in the structure of the pipelines for different types of commodities.

Another factor influencing donor support is the degree of prioritization of cost or cost efficiency in proposal analysis. Preparing and delivering multiple pipelines and maintaining them as a contingency investment may not be the cheapest response option in the short run. As such, the degree to which a donor prioritizes (or allows implementers to prioritize) response analysis criteria other than cost (e.g., appropriateness, effectiveness, timeliness, beneficiary preferences, economic spillover or multiplier effects, sustainability or development, etc.) may influence its support for multimodal programming. Finally, the authority of field-based donor representatives to fit response analysis priorities or financial systems poorly adapted to multimodal programming, may be limited.


Several operational and preparedness-specific factors support the enabling of multimodal programming.

To consider different cash, voucher, and in-kind response options equitably in each context requires appropriate preparation of multiple pipelines. That is, if commodities are pre-positioned and similar pre-agreements are not in place for cash or vouchers, then cash and vouchers cannot be given equitable consideration in response options analysis even if they may otherwise be similarly or more appropriate, effective, cost-efficient, or cost-effective.

Several case studies mentioned the relative administrative ease of switching modalities once appropriate framework agreements were in place. Maintaining these agreements and good relationships with these providers allows implementers to pause and restart pipelines as needed for modality switching. Strong relationships with vendors may help identify the most appropriate division of procurement responsibility between the implementing partner and service providers. In the Myanmar case, the decision to switch a portion of transfers to cash reduced the procurement demands on the implementing partner and its suppliers, resulting in an unexpected improvement of the timeliness of suppliers’ rice distributions.

Operational capacity was an important deciding factor in response analysis in several cases, including Guatemala, Myanmar, and the American Red Cross in Nepal. Technical and operational teams should have the skills, standard operating procedures, and confidence to implement either cash, vouchers or in-kind (or a combination thereof) as appropriate. The Red Cross Movement implemented this “one team” approach in Nepal after finding in the Philippines that maintaining separate teams of cash/voucher and in-kind experts exacerbates internal coordination obstacles.

Organizations with access to resources for piloting, whether internal or external, will be able to test response analysis assumptions and develop and maintain internal capacity and confidence. Such resources may also serve as contingency funds allowing for a modality switch in the event of a change of context for a project without sufficient budget flexibility.

Barriers to multimodal programming

The most significant obstacle to programming multiple modalities simultaneously or to switching modalities would be an institutional commitment to specialize in only one modality. Specialization in one modality of assistance, without strong response options, analysis investment, and clear rules for both engagement and nonengagement risks, results in an organization built as a hammer that interprets both screws and bolts as types of nails. Such an organization would not be flexible enough to adequately respond to changing circumstances, as was noted in South Sudan. This kind of organizational commitment to a single modality, while relatively uncommon, can present an institutional barrier to multimodal programming.

As noted in several case studies, particularly those in which multiple modalities were intended to be delivered simultaneously but were not (Guatemala, Nepal), the effort to prepare multiple pipelines is greater than for one single pipeline. In addition, while simultaneous delivery is more efficient, it may not be possible, particularly in the earliest stages of an intervention. While these factors do not prevent multimodal programming, the time and human resource costs should be properly considered during response analysis.In conclusion, further study relating experience from these case studies and CaLP’s Organizational Cash Readiness Tool (OCRT) may be useful as a future area of research.