Input Subsidy Programs and Climate Smart Agriculture: Current Realities and Future Potential

The evidence base remains thin but the weight of the available evidence suggests that ISPs have had either no effect on or have reduced African smallholders’ use of CSA practices. Empirical evidence across many case studies shows mixed results for many CSA practices considered. In addition, studies show the difficulties posed by delivery mechanisms that provide inputs too late for effective and efficient use by farmers. Finally, the absence of robust agricultural extension services in many African countries makes the diffusion and implementation of CSA practices even more challenging.

More specifically, evidence suggests that ISPs did not affect Ghanaian farmers’ investment in soil and water conservation, broadly defined (Vondolia et al. 2012), nor did they affect Malawian or Zambian smallholders’ use of manure (Holden and Lunduka 2010, 2012; Levine 2015). And while Malawi’s ISP had no statistically significant effect on intercropping (Holden and Lunduka 2010), Zambia’s ISP has reduced intercropping in general, but not intercropping involving legumes (Levine 2015). Moreover, Zambia’s ISP has negatively affected crop rotation and fallowing (ibid; Mason et al. 2013). The program has contributed to continuous cultivation of mono-cropped maize over time and within seasons, which leave smallholders more vulnerable to climate shocks – the antithesis of CSA. ISPs may increase maize yields in the short run except during extreme weather conditions (see Holden and Lunduka 2010; Mason et al. 2013; Chibwana et al. 2014; Mason et al. 2015; among many others). However, if results similar to Zambia are obtained elsewhere, these yield gains could be coming at the cost of lower soil organic matter and higher soil acidity, both of which will result in lower yields and fertilizer use efficiency in the medium to long run (Marenya and Barrett 2009; Burke 2012).

Empirical evidence on the effects of ISPs on crop diversification is mixed. For example, while Chibwana et al. (2012) and Mason et al. (2013) find that ISPs in Malawi and Zambia, respectively, incentivize households to devote a greater share of their cropped area to maize, other studies from Malawi suggest the opposite (Holden and Lunduka 2010; Karamba 2013) or that ISPs have no statistically significant effect on crop diversification (Karamba 2013). Most likely, the effects of ISPs depend on the range of inputs provided. ISPs that focus less on a specific crop and support a broader range of alternative crops, in particular legumes that add biomass and moisture retention to soil, may generate better outcomes with respect to crop diversification and soil fertility, responsiveness of crops to inorganic fertilizer and other benefits (Snapp et al. 2010).

While ISPs may contribute to sustainable productivity growth by maximizing fertilizer to crop output efficiency, their track record has been disappointing. Jayne et al. (forthcoming) conclude that most African governments to date have focused more on increasing African farmers’ use of fertilizer than on providing support for its efficient use.

Another feature of many ISPs that is decidedly not climate smart is perennial late delivery of subsidized fertilizer and seeds to beneficiary farmers (Xu et al. 2009; Lunduka et al. 2013; Mason et al. 2013; Namonje-Kapembwa et al. 2015). Late delivery is particularly common when ISP inputs are disseminated through dedicated ISP distribution systems that largely sideline existing input distribution networks. This is how fertilizer for Malawi’s ISP and both fertilizer and seed for Zambia’s ISP were distributed until 2014/15 and 2015/16, respectively, when each country started piloting agrodealer-based voucher redemption systems (Logistics Unit 2015; ZMAL 2015a; b). Late delivery of ISP inputs results in late planting and/or late fertilizer application, reducing yields and leaving beneficiary households more vulnerable to climate shocks (Xu et al. 2009; Namonje-Kapembwa et al. 2015; Arslan et al. 2015).

Most public agricultural extension systems are seriously under-provisioned to perform their multiple mandates of providing new management advice to farmers, learning from their efforts and difficulties of implementation and liaising with adaptive research systems to generate and disseminate new productive and sustainable practices, including SI practices. Some African public extensions are virtually defunct. Therefore, it should not be surprising that despite heavy spending on ISPs, their impacts on crop yields have been smaller than anticipated (ibid). In Zambia and Malawi, for example, a one-kilogram increase in subsidized fertilizer raises smallholder households’ maize output by an average of only 1.88 kg and 1.65 kg, respectively (Mason et al. 2013; Ricker-Gilbert et al. 2011). This low crop yield response to fertilizer is a major reason for the relatively low benefit-cost ratios of the ISPs in Malawi (1.08) and Zambia (0.92) (Jayne et al. 2017).

In response to some of these limitations, many ISPs are currently transforming to more flexible, private-sector, inclusive systems. This creates possibilities for ISPs to be restructured in ways that incentivize farmers to adopt particular SI practices and also bring about system-wide changes that promote resilience. The remainder of this section examines this potential of ISPs, however the discussion is largely conjectural given the limited evidence that ISPs as implemented to date have achieved such benefits.

A handful of ex ante analyses have explored how ISPs might compare to other programs to promote farmers’ use of practices that may be climate smart. For example, Marenya et al. (2012) use 30-year crop simulation models for maize, rice, and sorghum calibrated for several districts in Kenya, Malawi, and Uganda to compare changes in the net present value (NPV) of adopting various soil fertility management (SFM) strategies under two sets of policy regimes: a 50% fertilizer subsidy and carbon credits priced at $4, $8, or $12 per metric ton of carbon sequestered in the soil. The SFM strategies considered include various combinations of inorganic (N) fertilizer, animal manure, and crop residue retention – practices that may be ‘climate smart’ in some contexts. Their results suggest that carbon credits, especially when priced at $8 or $12/mt, produce larger NPV increases than the 50% fertilizer subsidy. While carbon markets are virtually non-existent in Africa, this analysis suggests monetary incentives play an important role in stimulating adoption of climate smart practices. This leaves room for ISPs to deliver monetary incentives to such ends. Yet, this in turn requires that extension systems are capable of delivering appropriate management information and that adoption is effectively monitored, which seems very challenging.

In later work, Marenya et al. (2014) use choice experiments to measure Malawian smallholder farmers’ preferences for various hypothetical policy incentives to adopt soil conservation practices, namely minimum tillage with legume intercropping: cash payments, two different types of index-based crop insurance contracts, and fertilizer subsidies.² Results suggest that most farmers preferred fertilizer subsidies to cash payments or crop insurance. In addition, farmers generally preferred cash payments to crop insurance, even when the expected payout from the crop insurance was higher than the cash payment. We must be careful, however, in generalizing these results, as they are specific to the choice sets used in the experiments. For example, the expressed preference of fertilizer subsidy over cash payments is likely driven by the fact that cash payment options (ranging from MK 800 to MK 2000) were lower compared to fertilizer subsidy (MK 2000) because of the expected yield gains with fertilizer. Even still, both cases suggest that under the right conditions some combination of conditional subsidy or conditional cash payment can incentivize adoption of farm management practices. Whether or not this leads to a permanent behavioral change, or whether public entities are capable of monitoring adherence to the conditions, remains an open question.

Finally, there is the question about whether raising crop productivity through inorganic fertilizer use might reduce the rate at which forests are converted into farmland and therefore reduce the agricultural sector’s contribution to GHG emissions. Recent evidence has begun to question the logic that agricultural productivity growth can arrest rapid farm area expansion and thus conserve the world’s forests and grasslands (Hertel 2011; Robertson and Swinton 2005; Byerlee et al. 2014). Instead, a generally positive area response to improved profit incentives is likely to create new pressures for further area expansion and conversion of forest and grasslands to farmland. Policy incentives could play a potential role here. In theory, ISPs could be structured in such a way as to oblige beneficiaries to reduce or maintain the amount of area under cultivation. However, it is not clear whether such rules would impose unreasonable demands on food insecure rural households or whether they could be adequately monitored or enforced.

In summary, while ISPs can be theoretically structured in ways that promote farm-level management changes, the oversight, enforcement, and extension costs needed to make this work are high, and may increase the already substantial opportunity costs of large public expenditures on ISPs.

As the development community understandably pushes hard to make progress in helping African farmers, there are major risks of overgeneralization about what kinds of farming practices really contribute to ex ante risk management and ex post coping strategies. Africa is heterogeneous with respect to its climate conditions, soil types, market access conditions, and factor price ratios. Some parts of Africa are still land abundant; labor and capital may be binding constraints in such areas. Other agricultural areas of Africa are densely populated, facing land pressures and rising land prices. In some of these areas, labor is relatively abundant and hence labor-intensive CSA practices may hold some potential to be scaled-up and incentivized through ISPs. However, in areas with good market access conditions and proximity to urban areas, economic transformation processes are bidding up labor wages and making it difficult for farmers to adopt labor-intensive CSA practices unless they also provide high returns to labor. The heterogeneous conditions of farming systems in Africa warrant great caution against overgeneralization in promoting technologies through ISPs or on their own based on blanket recommendations across wide domains.

As an example, minimizing soil disturbance through no or minimum tillage (MT)³ strategies are frequently promoted in Africa as a means to mitigate soil erosion, increase soil water retention capacity, and to slow the rate of soil organic carbon (SOC) decomposition, and thus achieve yield growth and stability (Branca et al. 2011; Chivenge et al. 2007). However, yield and soil quality effects of MT practices vary substantially depending on soil type and association of MT with other land management practices, namely crop residue retention and incorporation. Several studies have shown that MT practices lead to an accumulation of SOC in the surface layers of soil (0–10 cm), rather than in the root zone (Sisti et al. 2004; Chivenge et al. 2007; Carter and Rennie 1982; Hernanz et al. 2002; Doran 1980). Carter and Rennie (1982) find that microbial biomass and potential mineralizable carbon and nitrogen are high in surface soils where MT is practiced. Conversely, these soil properties are higher in lower soil depths when conventional tillage (CT) is applied. The magnitude and location of the SOC pool are important for yield growth and stabilization. As Lal (2006) shows, every 1 mt/ha increase in the SOC pool in the root zone is associated with a 30–300 kg/ha increase in maize yields and a 10–50 kg/ha increase in rice yields. Improving SOC pool in the root zone can simultaneously enhance soil’s water retention capacity (Mbagwu 1991; Fernández-Ugalde et al. 2009), increase its cation exchange capacity, and thus nutrient retention (Carter et al. 1992), and improve soil aggregation and susceptibility to erosion (Lal 2006; Paul et al. 2013). Thus, further development of MT technologies may be needed to achieve its potential benefits.

Another potential limitation of MT is that without associated investments in crop residue retention and/or crop rotation, fields tilled using MT frequently experience no yield improvement (Hernanz et al. 2002) or in some cases a dramatic drop in yield relative to CT (Rusinamhodzi et al. 2011; Raimbault and Vyn 1991; Paul et al. 2013). When MT practices are applied in conjunction with crop residue retention, legume rotation, and/or nitrogen fertilizer application, the yield effects of MT tend to be higher than those achieved through CT, but again this is highly dependent on prevailing agro-ecological conditions (Raimbault and Vyn 1991; Govaerts et al. 2005; Dalal et al. 1991; Triplett et al. 1968).

As discussed in Section 3, ISPs in the region are not designed to cope with the high level of regional and farm level heterogeneity in input needs and management requirements. Significant region-specific modifications in the composition of ISP inputsm coupled with region-specific farm management promotion strategies will be required for ISPs to contribute meaningfully to CSA goals, which in turn implies significant modification in the logistical design, implementation and cost of ISPs.

A more obvious way in which ISPs can influence overall productivity is through the injection of greater levels of nitrogen (N) into African soils, where nitrogen is often the limiting nutrient factor (Snapp et al. 2010). Rusinamhodzi et al. (2011) in their summary of evidence on conservation agriculture shows that in 73% of the field studies, high levels of nitrogen fertilizer were required to achieve improved yields under these practices. However, recent advances in soil science and agronomy research show that massive nitrogen (N) injections may not be economically feasible for farmers or be social welfare raising without farmer adoption of complementary soil management practices that allow N to be efficiently utilized by plants (Snyder et al. 2009). Thus, the challenge for large-scale programs, such as ISPs, is promoting carbon management practices together with nitrogen to achieve high nitrogen efficiency (Tittonell and Giller 2013). Paul et al. (2013) demonstrate that without sufficient biomass production (often stimulated by inorganic fertilizer application) SI practices of MT and residue retention do not have an effect on yield stability or SOC. Thus, an ongoing challenge is maintaining a large enough N pool in soils containing little organic carbon, which increases N leaching and gaseous loss pathways, adversely affecting CSA goals (Drinkwater and Snapp 2007). Unfortunately, large-scale efforts to promote SI practices that build up soil organic carbon are largely absent from government programs, are largely untested over the wide range of soil types and agro-ecologies found in the region, and are sometimes discounted by some as not being viable from the standpoint of low-resource farmers.

These several examples underscore the lack of consensus within the crop science community about what viable CSA and SI packages appropriate for heterogeneous smallholder agricultural systems should look like. In addition, there is a great deal of uncertainty over how climate will change in the region over the coming decades (Powlson et al. 2016). For these reasons, we conclude that African governments and the development community need an improved empirical evidence base that establishes the practices that actually promote CSA and SI objectives under the wide range of diverse and uncertain farming conditions found in the region. A precondition for making progress on this front is much greater public expenditure on agricultural R&D and adaptive research across the various economic/biophysical micro-climates. While necessary, increased public funding to agricultural R&D is not sufficient. But without a better evidence base on how practices perform under various conditions, the risk is that ISPs may be misguided in choosing which practices to promote.

First, as mentioned earlier, by expanding and stabilizing the demand for specified input types and quantities, ISPs can potentially help to overcome some of the persistent risks to commercial legume seed multiplication in the region. Ensuring adequate supplies of these seeds on the market is critical to achieving crop diversification, organic nitrogen fixation, and rotations. However, this potential benefit is mitigated by the trend, among donors and governments, to move toward more open voucher systems. Thus, in many ways there are important trade-offs to consider when promoting particular ISP distribution modalities. While open vouchers are desirable from a farmer choice perspective, restricted-choice vouchers for particular inputs, such as legume seeds, may be necessary to support system-wide improvements in legume seed supply chains. Restricted-choice vouchers may be justified in some instances where there are major beneficial externalities associated with promoting certain inputs and where the social benefits of doing so may greatly outweigh the short-term financial benefits from the perspective of individual farmers. The two approaches may be combined; for example, farmers could be provided an open voucher in addition to a restricted-choice voucher for legume seed. Similar system-wide benefits may accrue by using ISPs to create farmer demand for specific drought-tolerant seed varieties or soil amendments such as lime or inoculants, which are currently not widely used by farmers.

A second way in which ISPs may promote system-wide CSA resilience is through promoting “market-smart” private investments, which could increase private investments in input supply chains and extension services. By encouraging private sector input supply chain development, market-friendly ISPs can foster improved input access conditions for farmers, thus over time making them less dependent on public input supply systems. Private input systems are potentially less prone than public systems to delivery challenges associated with logistical and financial constraints (Jayne and Rashid 2013). There is clear potential for ISPs to promote system-wide investments that are both climate-smart and market-smart and synergistic in their promotion of community resilience to climate variability.

Finally, the move toward digital platforms for delivering ISPs, such as electronic vouchers (‘e-vouchers’), create opportunities to use ISPs as delivery mechanisms for other sorts of products, such as weather indexed insurance. This requires that ISP farmer registries collect a wide range of information on beneficiaries, including geographic location and bank information. With this sort of information, ISPs can defray the screening costs of identifying farmers and managing insurance pay-outs when necessary.

Unfortunately, some aspects of ISPs may work against climate change mitigation even as they promote resilience objectives. ISPs increase the quantities of fertilizer manufactured and used in the agricultural production process (holding all other factors constant) and therefore ISP proposals that include increased fertilizer use must account for the additional GHG emissions. Inorganic fertilizer use contributes to GHG emissions both through the soil chemical and biological processes and through the production of synthetic fertilizer. According to a recent estimate, 56% of global non-carbon dioxide GHG emissions occur from agricultural production, and roughly 12% of agricultural GHG emissions occur from fertilizer use (IPCC 2014). The additional contribution to GHG emissions caused by the manufacturing of synthetic fertilizer is also significant (see Appendix 1). Thus, the net impact of ISPs on GHG emissions will depend on the effectiveness with which ISPs can be used to promote adoption of CSA practices that raise soil organic carbon, sequester carbon and depress the rate of forest conversion to farmland and offset the adverse effects of increased fertilizer use on GHG emissions. The empirical evidence on these issues is weak and more detailed research is needed. Appendix 1 provides some empirical estimates of the increased GHG emissions caused from additional use of synthetic fertilizers.

Moreover, there is the issue of opportunity costs. Nationwide ISPs tend to be expensive, and they can bid away scarce public funds that could otherwise be used to buffer communities from the effects of climate variability (e.g., irrigation, agricultural research and extension systems, weather insurance, etc.) or to support ex post coping responses (e.g., disaster relief programs). In Africa, where irrigation only accounts for 4% of arable land (You et al. 2012) and where there is huge unmet potential for irrigation expansion, ISPs would seemingly compete against public investment in water control and other ex ante risk management strategies. Future research is again needed to determine whether smart ISPs may be structured in ways that leverage private sector investments in CSA inputs and services and produce benefits that outweight those generated from other proven types of public investments in agriculture.

Key knowledge gaps include understanding why farmers are not adopting CSA practices or are subsequently dis-adopting them (which could then point to potential interventions to overcome these constraints); determining which practices are profitable for whom and under what conditions; understanding the interactions between CSA practices and ISP inputs (e.g., do selected CSA practices increase fertilizer use efficiency?); identifying cost-effective, enforceable, and scalable ways to implement a potential CSA precondition requirement for ISPs; and comparing the cost-effectiveness of such a requirement to that of other approaches to promote CSA. Given the very mixed results of ISPs, the rampant elite capture and diversion of inputs intended for the programs, and the high price tag and opportunity cost of ISPs in general and in relation to other programs and investments to develop and stimulate uptake of CSA technologies (see Jayne and Rashid 2013; Lunduka et al. 2013; Mason et al. 2013; among many others), linking CSA promotion to ISPs may be a risky proposition.

There are three overarching challenges to be addressed for ISPs to effectively contribute to CSA objectives. First is the limited understanding of workable approaches for internalizing the externalities associated with GHG-emitting land management decisions of millions of resource-poor farmers in developing countries. This is a problem for social scientists to resolve by developing ways for carbon markets to be linked to smallholders in Africa and that can provide farmers monetary incentives for the adoption of particular GHG mitigating practices, may be a viable strategy for achieving widespread farm management change, but much remains to be worked out before viable programs could be implemented in most of sub-Saharan Africa.

The second challenge is the currently limited on-shelf technologies and management know-how to improve smallholder yield stability and growth in the face of increasing climate variability. Most on-shelf technologies and practices being promoted as being “climate smart” appear to help at the margin, but cannot be relied upon to meaningfully stabilize harvests in the face of major droughts or floods or to arrest the degree of distress migration often associated with it. More effective water and soil fertility management techniques appropriate for the situation of low-resource farmers are needed, and this will requires significantly increased investment in localized, adaptive research for the wide range of smallholder farming systems in sub-Saharan Africa. This is a challenge both for the scientific research community and for policy makers to make the necessary long-term funding commitments to adaptive agricultural research and development programs.

The third challenge is the near absence of effective bi-directional learning and extension systems to help farmers profitably adopt and adapt proven farm management practices. This again presents challenges for policy makers to make the necessary long-term funding commitments and to social scientists to design extension systems that effectively link scientists and farmers disaggregated by particular agro-ecologies and degrees of resource constraints.

Addressing these three challenges is a tall order. For this reason, we believe that much greater progress is needed in each of these three areas before it could be practical or effective to try to use ISPs as a vehicle to make agriculture more climate-smart. This conclusion is not meant to stifle progress where progress can be made, but is rather to point out the scope of the challenges before us. It will take time for the proposals made here to generate meaningful impacts. This is why there is no time to waste in getting started.