Returns of sustainable land management interventions: Evidence from Southern Ethiopia

ABSTRACT

Ethiopia has been implementing sustainable land management (SLM) measures throughout the country to reverse land degradation. Despite the implementation of various SLM measures in Ethiopia, very little is known about whether these measures are effective from adopting farmers’ perspectives. This study examines the costs, benefits, and returns of SLM practices implemented in the Barcha-Adado watershed in Southern Ethiopia. Data for the analysis were obtained primarily from a survey of 231 barley-growing farmers. Using a cost-benefit analytical framework, the returns of SLM practices were evaluated using the net present value (NPV) and benefit-cost ratio (BCR) over the period 2019 to 2046. The major SLM measures implemented by farmers are soil bunds, fanyajuu bunds, and fanyajuu bunds stabilized with grass strips. The average establishment cost of SLM measures was $171 per hectare. On a per-hectare basis, farmers who implemented SLM practices experienced a 28% higher crop yield and an NPV of $1,491.6 compared to non-adopting farmers. The BCR of SLM practices is 5.16, implying that the present value of the benefits is more than five times higher than the present value of the costs of investment in SLM measures. Returns are the highest for fanyajuu bunds and the least for fanyajuu bunds stabilized with grass strips. Study findings suggest that SLM interventions significantly enhance crop productivity and generate considerable financial returns for adopting farmers. The findings would serve as evidence for the local decision-makers and may provide incentives to scale up the benefits of SLM practices to degraded farmlands in the country.

KEYWORDS:

• land degradation
• SLM
• costs
• benefits
• returns
• Ethiopia

1. Introduction

The Economics of Land Degradation (ELD) Initiative (2013) defines land degradation as a ‘reduction in the economic value of ecosystem services and goods derived from the land’. Recently, the United Nations Convention to Combat Desertification (UNCCD) (Vogl et al., 2017) also defines land degradation as ‘any reduction or loss in the biological or economic productive capacity of the land resource base’. It is caused by anthropogenic (human) activities, exacerbated by natural processes, and often magnified by and closely intertwined with climate change and biodiversity loss (Mansourian & Berrahmouni, 2021; United Nations Convention to Combat Desertification, 2014). The problem of land degradation is particularly severe in Africa. According to Mansourian and Berrahmouni (2021), as much as 65 percent of productive land in Africa is degraded, while desertification affects 45 percent of the region’s land area.

Land degradation has been one of the challenges that Ethiopia has faced for centuries. Ethiopia is known for its historic agriculture, but also the associated, widespread, and ongoing land degradation. The Ethiopian highlands account for 45% of the land area and provide the livelihood of about 80% of Ethiopia’s population (Gebreselassie, 2014). In particular, the older agricultural areas of the northeast have long been particularly affected, but the highest soil erosion rates are currently being observed in the western parts of the highlands (Hurni, et al., 2015). Evidence suggests that more than 85 % of the land is degraded to various degrees (Gebreselassie et al., 2016). More recently, using data from the FAO TERRASTAT, Kirrui et al. (2021) indicated that Ethiopia is the most seriously affected (25% of its territory degraded) country in the SSA.

Land degradation has been vastly detrimental to agricultural ecosystems and crop production (ELD Initiative & UNEP, 2015). In Ethiopia, it is a major cause of low and declining agricultural productivity, rural poverty, and food insecurity in Ethiopia (World Bank, 2020). In particular, rural smallholder farmers and households suffer the most from land degradation as their activities directly depend on healthy soils, tree cover, and clean water (Mansourian & Berrahmouni, 2021).

Every year, nearly three million hectares of Africa’s forests are lost, leading to a 3 percent loss of GDP associated with soil and nutrient depletion. Owing to land degradation and the associated loss of productivity, Africa spends more than $ 35 billion on food imports annually (Mansourian & Berrahmouni, 2021). In Ethiopia, evidence also suggests that the productivity of the Ethiopian economy is being seriously eroded by unsustainable land management practices in crop and animal feed production systems (Berry, 2003). Without accounting for downstream and off-site effects such as flooding, and damage to infrastructure resulting from erosion, the minimum estimated annual costs of land degradation in Ethiopia range from 2 to 3 percent of agricultural GDP (United Nations Convention to Combat Desertification, 2014).

Based on the total economic value (TEV) framework, the total cost of land degradation due to land use and cover changes (LUCC) during the period 2001–2009 was $ 35 billion. In monetary terms, the annual cost of land degradation associated with land use and cover change in Ethiopia is estimated to be about $4.3 billion. This loss of ecosystem values is equivalent to 22.5% of the country’s 2007 GDP (Gebreselassie et al., 2016). The decline in the total terrestrial ecosystem value and the annual cost of land degradation is a significant loss for a country that depends heavily on agriculture (World Bank, 2020). This suggests that land degradation in the productive agricultural Ethiopian highlands has become a major peril to the development of the nation’s smallholder-dominated agriculture. It appears to be a major impediment to achieving food security and improving livelihoods (ELD Initiative & UNEP, 2015).

Since 2008, the Ethiopian government in collaboration with development partners and donors has been undertaking huge investments in the Sustainable Land Management Program (SLMP) aiming to reduce land degradation, restore degraded landscapes, and increase agricultural productivity (GIZ, 2018). Through its two phases (SLMP-I and SLMP-II), the Sustainable Land Management Program has been implemented in close to 135 critical watersheds in six regional states of the country (World Bank, 2020). The primary reason to implement soil and water conservation (SWC) technologies in mountainous regions is to reduce the movement of soils, water flow velocity, and the broader effects of erosion, such as the siltation of rivers, lakes, and dams. SWC techniques also reduce soil loss from farmers’ plots, preserving critical nutrients and increasing crop yields, and this is the chief selling point for farmers (Kassie et al., 2008). The SLM practices implemented thus far are designed to decrease erosion and increase agricultural yields in the highlands of Ethiopia, thereby improving rural household welfare (Schmidt et al., 2014). Because SWC technologies serve not only the social good but also increase on-farm yields, they are considered ‘win-win’ (Kassie et al., 2008).

The SLMP has been implemented in 31 watersheds in the southern nations, nationalities, and people regional (SNNPR) state. The geographic focus of this study is the Barcha-Adado watershed, one of the SLM intervention watersheds found in the Gedeo administrative zone of Southern Ethiopia. The Bank-financed SLM program had been implemented in the watershed from the 2011/12 to 2016/17 period. Actions against land degradation involve both significant costs and economic benefits. Despite the implementation of SLM interventions in several watersheds at the country level, the costs, benefits, and returns of SLM interventions in Ethiopia in general and the Southern Highlands of Ethiopia, in particular, are rarely documented. For informed decision-making in the agricultural sector and scaling-up of Ethiopia’s experience in sustainable land management practices, it is important to understand the costs, benefits, and financial returns of SLM practices from economic perspectives at a watershed level. Thus, this article assesses the costs, benefits, and financial returns of SLM interventions in the Barcha-Adado watershed located in the Southern Highlands of Ethiopia using the cost-benefit analysis method.

This study is broadly linked to the literature that documents the effectiveness and returns of action against land degradation. However, there are several research gaps in the literature that need to be addressed.

First, most of the available studies tend to focus on the short-term benefits and costs of SLM practices. In particular, they focus dominantly on quantifying the impacts of soil and water conservation (SWC) measures on crop yield and other short-term benefits ((Adimassu et al., 2014; Kassie et al., 2007, 2008). Some studies also document the effect of physical soil and water conservation practices in reducing run-off, soil erosion, and nutrient depletion (Adimassu et al., 2017; Ayele et al., 2016; Gebreselassie & Belay, 2013). A recent study, for instance, by Mironova et al. (2022) found that soil fertility improves following COVID-19-related movement restrictions and quarantine measures. They reported that soil nutrients (NPK) in the post-quarantine period increased quantitatively; nitrogen increased by 25%, phosphorus by 30%, and potassium almost doubled. This is seen as a positive impact of urban ecosystems’ exclosure from human interaction.

Study results on the crop yield and revenue impacts of SLM practices are largely inconclusive to have any relevance for policy-making at the local and national levels. For instance, Kassie et al. (2007) found that plots with stone bunds are more productive than those without such technologies in semi-arid areas but not in higher rainfall areas of northern Ethiopian highlands. Tesfaye et al. (2016), using farm household data from the Gedeb watershed located in northern Ethiopia, also reported that farmers adopting various SWC measures have benefited from significant increases in crop productivity. Jabal et al. (2022) examined the impact of climate change on crop productivity and suggested that sustainable agronomic measures and land conservation are vital to enhance crop production in the face of climate change. On the other hand, using cross-sectional multiple plot observations, study results by (Kassie et al., 2008) found that the value of crop production for plots with fanyajuu bunds was lower than for plots without bunds in the northwest highlands of Ethiopia. Likewise, Adimassu et al. (2014) also noted that soil bunds reduced crop yield by 7 percent relative to plots without soil bunds in the Central Highlands of Ethiopia. Such observation has led some scholars, for instance, Adimassu et al. (2017) to conclude that the impact of SWC measures in the highland areas of Ethiopia has been inconsistent and scattered.

On the inconsistency of previous studies’ results, Adimassu et al. (2017) noted that the crop productivity impact of SLM practices depends on the type of SWC measures implemented. They found that while the impact of physical soil and water conservation practices on crop yield was negative, which is mainly attributed to the reduction of the effective cultivable area by soil/stone bunds, SWC measures combined with agronomic measures, on the other hand, were found to have a positive impact on crop yield. In general, the inconclusive findings indicate that the impacts of SWC measures on crop yield and returns vary by the type of SLM technology adopted, agroecology, and location of watersheds. This suggests the need to revisit the issue. Thus, there is a need for a study that examines the benefits, costs, and financial returns of SLM practices over a long period.

Secondly and importantly, the costs, benefits, and returns of SLM practices in the Southern Highlands of Ethiopia are rarely documented. This is because the extant studies draw evidence primarily from the northern and north western highlands of Ethiopia. For instance, using cost-benefit analysis, Tesfaye et al. (2016) evaluated the financial viability of SWC measures (soil bunds, fanyajuu bunds, and stone bunds) in the Gedeb watershed located in northern Ethiopia. They reported that SWC measures generate positive returns for farm households in the watershed. Using cost-benefit analysis, Mekuria et al. (2011) reported positive financial returns from the exclosure establishment in the Tigray regional state of Ethiopia. At a country level, study reports by Hurni et al. (2015) and Tilahun (2020) demonstrated that investment in SLM practices generates considerable additional financial returns in Ethiopia.

As the available studies are either highly aggregated at a country level and/or location-specific, the generalizability of the findings is limited and may have little relevance to understanding the returns of SLM practices in the context of the study area. In addition, the few available studies on SLM practices in the southern highlands of Ethiopia assess either the impact of SWC measures on crop yield or determinants of SWC adoption decisions (Ayalew, 2011; Tadesse & Kassa, 2004), overlooking the financial returns of SWC measures. For this reason, more region-specific studies are crucial to better understand the costs, benefits, and financial profitability of SLM measures in the Southern Highlands of Ethiopia.

Thirdly, existing literature focuses notably on a single type of SLM practice in isolation (Kassie et al., 2007, 2008), Adimassu et al. (2014); Matere et al. (2016). In some cases, studies use the generic term SLM measures. However, evidence suggests that as profitability varies with the type of SLM technology implemented, returns and economic viability of SWC measures are also viewed to be inconsistent (Adimassu et al., 2017). Also, the profitability of SWC measures depends very much on the location, agroecology, type of crops grown, also on farm-specific attributes, the prices of inputs used, and outputs produced (Lutz et al., 1994b). Case studies by Lutz et al. (1994b) and Lutz et al. (1994a) also show mixed evidence that conservation is profitable in some case study sites but not in others in Central America and the Caribbean. Financial returns are also noted to depend on the types of costs and benefits considered, and the period over which returns are projected (Tilahun, 2020). Thus, it is noteworthy to compare various types of SLM practices in terms of costs, benefits, and returns using evidence from a new case study area.

Fourthly, studies that employ an experimental design to compare plots with SLM with a control group of plots without SLM measure and integrate ecological valuation of benefits with a cost-benefit analytical framework are scant in Ethiopia.

The foregoing discussion suggests some research gaps and hence the need to revisit the issue using evidence from one of the SLM intervention watersheds in the less studied Southern Highlands of Ethiopia and SLM intervention watersheds of the country. Thus, using the cost-benefit analysis technique, this study assesses the costs, benefits, and financial returns of adopting SLM practices relative to a business-as-usual scenario in the Barcha-Adado watershed in southern Ethiopia.

This study differs from the existing studies on returns of actions against land degradation in Ethiopia in several dimensions and its contribution to the literature is manifold.

First, unlike previous studies that focus on short-term benefits and costs, this study projected the long-term financial returns of SLM practices over a longer project period. Understanding the profitability of SLM interventions is important from adopting farmers’ perspective since new agricultural technologies need to be profitable to the farmer if they are to be adopted and sustained (Kassie et al., 2011).

Second and important, it documents new evidence on the costs, benefits, and returns of SLM practices in the Southern Highlands of Ethiopia, which is usually overlooked in the literature. That is, in terms of geographic scope, unlike the extant studies that focus on the northern or north-western highlands of Ethiopia, this article draws evidence from the southern highlands of Ethiopia. In this regard, this study would particularly bridge the knowledge gap on the returns of SLM practices in the southern highlands of Ethiopia and contribute to a better understanding of the returns of SLM practices in Ethiopia.

Third, in contrast to those studies that estimate the returns of a specific SLM practice in isolation, this study compares the costs, benefits, and returns of three types of SLM practices (soil bunds, fanyajuu bunds, and fanyajuu bunds with grass strips). Such comprehensive evidence on the returns of various types of SLM practices may help in understanding the variability in returns across SLM types. It is also instrumental for farmers to identify and adopt the most beneficial SLM practices that could increase crop yield and farm income, and enhance food security and welfare of farmers.

Fourth, from a policy relevance point of view, it is argued that disentangling productivity differences between adopters and non-adopters of SWC measures is crucial for understanding household-level responses to land degradation and for designing appropriate policy interventions (Nyangena & Köhlin, 2008). Thus, the evidence drawn using a quasi-experimental set up is expected to support policymakers as well as farmers at the local and national levels to make informed decisions regarding the contribution and long-term profitability of investment in SLM practices.

Finally, methodologically, this study combined economic valuation of ecological benefits and cost-benefit analysis techniques to evaluate the returns of SLM practices for adopting farmers relative to a baseline group of non-adopters (business-as-usual scenario). Given the paucity of empirical studies that combine the assessment of changes in ecological benefits with a cost-benefit analysis, this study is expected to contribute to the pool of literature on the methods of assessing the returns of actions against land degradation. In terms of the analytical/conceptual framework used to estimate the returns of SLM practices, this study employed a cost-benefit analytical framework. The approach is linked to the literature that uses a cost-benefit analytical framework to evaluate the financial/economic returns of adopting SLM practices. The basic principle of the method is to compare the flows of costs and benefits with and without SLM practices (Lutz et al., 1994b; Mcharo & Maghenda, 2021; Pagiola, 1994). Using the stream of benefits and costs, the framework allows estimating the returns of SLM practices by comparing the net benefits of adopting SLM practices to the net benefits of a business-as-usual scenario (Hurni et al., 2015; Tilahun, 2020).

The remainder of the article proceeds as follows. Section two presents the data and methods for the cost-benefit analysis (CBA) of SLM practices. Section three provides the results and discussion of the CBA of SLM measures. The last section draws conclusions and policy implications.

2. Materials and methods

2.1. Description of the study site

Out of the 31 SLM intervention watersheds in the Southern Nations, Nationalities, and people’s Region (SNNPR), this study uses evidence from the Barcha-Adado watershed. The Barcha-Adado watershed is located in the Bule district of the Gedeo administrative zone in the SNNPR state. The Barcha-Adado watershed covers an area of 10,169 ha. The watershed is located in the Gedeo highlands. The micro-watersheds considered in this study, Wachilacha and Foldawe, are located in the Kochore and Suqo Kebeles of the district,¹ respectively (Figure 1).

Figure 1. Location map of the study sites in the Barcha-Adado watershed, Bule Woreda, Gedeo Zone.

The mean annual rainfall of the study area ranges from 1,401 mm to 1,800 mm and the mean annual temperature is between 12°C and 20°C. The climate of Bule Woreda is 65% highland (Dega) agroecology and 35% mid-highland (Woina Dega). The area has two rainy seasons namely, the short rainy season (from March to May) and the long rainy season (from July to October) (Bule Woreda Office of Agriculture and Natural Resource, 2019). A considerable part of the watershed (about 75%) lies in highland agroecology. The selected sub-watersheds have elevations ranging from 2800 to 3030 meters above sea level.

2.2. Data sources

The Bank-financed SLM program had been implemented in the Barcha-Adado watershed from the 2011/12 to 2016/17 period. While the first two years were the preparation period, the remaining three years were used for the construction of SWC structures on degraded land. The profitability of SWC measures is evaluated using data collected on the different attributes of benefits and costs of SLM interventions. To this end, the study makes use of both farm household surveys and secondary data.

2.3. Sampling strategy

Evaluating the returns of actions against land degradation involves selecting the SLM-treated/conserved and control/non-treated sub-watersheds, and obtaining data from farm households located in each group. Firstly, a reconnaissance survey is undertaken to identify the study site containing both the conserved and non-conserved micro watersheds. Then, the study area is generally divided into two sites: farmland treated with SWC measures and farmland without SWC measures. There are 17 micro-watersheds with SLM interventions in the watershed. The treated micro watersheds are similar in agroecology (highland), land use, and period of SLM implementation. In addition, a large group of control watersheds is available in proximity to the SLM watershed, providing a watershed-scale quasi-experimental design.

Secondly, four micro watersheds are selected from Suko and Kochore kebeles of the watershed: two with SLM and two without an SLM intervention. From the intervention site, two representative micro watersheds, namely Wachilacha and Foldawe, are randomly selected for assessment. Then, for benchmarking—business as usual scenario, two non-conserved micro watersheds, namely Bake semania and Bule laga, are selected from the adjacent cultivated land. The selected adjacent cultivated land was comparable to the treated ones in the pre-intervention period. The location map of the study sites in the Barcha-Adado watershed is given in Figure 1.

Finally, the survey covered 231 randomly selected farm household heads, of which 154 are from the treated site and the remaining 77 are from the non-treated micro watersheds. The sample of farm households taken from each representative micro watershed is determined using probability proportional to size.

2.4. Data collection

Primary data on relevant variables are collected from a survey of farm household heads located in the treated and control micro watersheds. Data on crop yield, farm inputs, land size, size of cultivated land, type of SWC measures, and other key attributes were collected from farmers using a structured questionnaire in 2019. Survey data is obtained both from SLM-adopting and non-adopting farm households located in the conserved and adjacent non-conserved croplands in the watershed, respectively. The survey has been undertaken between August and September 2019 using a structured questionnaire. Farm household data are collected using a pre-tested questionnaire. Secondary data on the size and costs of SWC measures are obtained from SLM implementing agencies and the district’s office of agriculture and natural resources conservation. The data are then used to quantify the benefits and costs of SLM interventions.

2.5. Analytical framework: Cost-benefit analysis

Sustainable land management (SLM) refers to the adoption of land-use systems that enhance the ecological support functions of land with appropriate management practices, and thus enable land users to derive economic and social benefits from the land while maintaining those of future generations (ELD Initiative, 2013). SLM has both costs and benefits. A key issue that is instrumental to decision-making is whether the benefits justify the costs of intervention. To this end, this study employs the cost-benefit analysis (CBA) technique to assess the financial effectiveness of SLM practices or soil and water conservation measures. The SLM practices examined in the CBA are soil bunds, fanyajuu bunds, and fanyajuu bunds stabilized with grass strips.

Previous studies on the economic returns of SLM in Ethiopia focus primarily on quantifying the impacts of SWC measures on crop yield and other short-term benefits, reducing run off, gain in soil nutrients, etc (Adimassu et al., 2014; Kassie et al., 2007, 2008). The disadvantages of these studies, among others, are that (a) they mostly consider short-term benefits of SLM, notably crop yield and revenue, and in other cases, the value of soil nutrients, (b) they fail to account the costs of implementing SWC/SLM measures, which is vital to determine financial returns, (c) a related point is that they failed to project the long term financial returns and the variation in the returns of various types of SLM practices. Moreover, from a decision-making point of view, such assessments that account only for short-term benefits are only partial as they offer little evidence regarding the financial returns or profitability of adopting SLM technologies. The advantage of employing the cost-benefit analytical approach is that it considers both the costs and benefits of SLM interventions and also enables to evaluate if such interventions are profitable at the watershed level in the long run in the study area. For this reason, the cost-benefit analysis method is employed in this study.

Cost-benefit analysis provides a coherent framework for integrating information on the biophysical and economic environments faced by farmers. The basic principle of the methodology is to compare the flows of costs and benefits with and without soil conservation (Lutz et al., 1994b; Mcharo & Maghenda, 2021; Pagiola, 1994). Estimating the costs and benefits of SLM under both scenarios is at the heart of the cost-benefit analysis technique.

CBA of soil conservation can be carried out either from the private/farmers’ perspective (financial CBA) or society as a whole (economic and social CBA). Under the social CBA, all the costs and benefits of a given activity must be considered (Lutz et al., 1994a, 1994b). On the other hand, in the private CBA, returns are evaluated from individual decision-making agents’ perspectives. In this approach, only the costs and benefits that accrue to the agent making the decision (i.e. adopting farmers themselves) about resource use are considered. Costs and benefits are valued at the prices the farmers face (Lutz et al., 1994a, 1994b).

This study assesses the financial viability of investment in SWC measures from the adopting farmers’ perspective (or farm level). CBA of SWC measures is conducted from the private/farmer’s point of view. The private returns as seen by individual farmers take into account the on-site effects of SLM interventions on major crop production. This approach is preferred for two major reasons. First, decisions about land use are ultimately made by the farmers themselves and not by social planners or government agencies. Farmers decide how to use their land in light of their objectives, production possibilities, and constraints, not based on any theory of the social good. Understanding the incentives (and disincentives) individual farmers face is necessary, therefore, to understand patterns of resource use and to formulate appropriate responses to problems. Second, land use problems generally depend heavily on site-specific biophysical characteristics, which can vary significantly even within small areas (Lutz et al., 1994a, 1994b). Moreover, CBA at the farm level is the most appropriate approach because the SLM program’s major objective was to increase the agricultural productivity and income of smallholder farmers.

2.5.1. Steps employed in the cost-benefit analysis

The CBA evaluates the monetary incentives to farmers to adopt SLM technologies by estimating the economic value of changes in barley yield due to the use of one of the SWC measures on their cropland as compared to the business-as-usual scenario. Analytically, a CBA of SLM has the following components: identification and definition of effects (costs and benefits), quantification in physical terms of the effects, valuation of effects, determination of time horizon, weighing of the costs and benefits in time (discounting), determination of evaluation criteria, and sensitivity analysis (de Graaff, 1996).

The analytical framework used to guide the CBA has five components. These analytical steps are summarized as follows. Firstly, the benefits and costs of the action are defined. The major benefit considered is crop production (i.e. barley). While crop yield is used to estimate the benefits, the establishment and maintenance expenses of SWC measures are used to capture the costs of SLM practices. Secondly, the monetary value of benefits and costs of SLM interventions in the treated sub-watersheds and the control group was estimated using their market prices. Thirdly, the incremental costs and benefits associated with the SLM interventions are projected for farmers with SWC measures relative to farmers without SWC. Farmers without SWC refer to farmers in the business-as-usual scenario, where no SWC is adopted to reverse land degradation. Fourthly, to determine the added value of the investment in SWC measures in the treated sites, the incremental net benefits generated from the SLM interventions are discounted. Based on the value of the discounted incremental net benefit, the SWC measure with the greatest return is identified and a comparison is made regarding the financial viability/returns of each SLM intervention from the farmer’s perspective. Finally, the sensitivity of the cost-benefit analysis results to changes in various parameters is evaluated to test the robustness of the result to uncertainties and shocks.

2.5.2 Valuation of benefits and costs

The first step of a cost-benefit analysis is to define what constitutes costs and benefits (Castillo et al., 2020). Costs and benefits are valued using the market prices method. Since SWC measures have been primarily implemented to increase crop productivity in smallholder-owned croplands, benefits were captured using barley yield. Crop production is one of the major provisioning ecosystem services. Barley is used since it is produced by all sampled farm households both in the non-conserved and conserved study sites. Accordingly, benefits here refer to the market value of barley production.

The monetized revenues were obtained as the product of barley yield per hectare and the local price of barley. Using the market price method, the monetary value of benefits from barley production is given as follows,

${\mathrm{B}}_{\mathrm{t}}={\mathrm{P}}_{\mathrm{t}}\ast {\mathrm{Q}}_{\mathrm{t}}$

where Q is crop yield (kg) measured as kilograms of barley produced per hectare of cultivated land annually, t refers period over which benefits are projected (2019–2046), and P is the average local farmgate price of barley. The local price of barley was obtained from the farmers and the local agricultural and natural resource office.

Costs of SLM refer to costs of establishing and maintaining SWC measures to prevent and/or reverse land degradation in the area. The costs of SLM interventions include costs of labor (man-days), and material inputs used for the construction and maintenance of each of the SWC measures. Thus, the market value of the costs of implementing each SLM measure (Cit) is given by

$C_{it}=E{C}_i+M{C}_{it}$

where EC_i denotes the initial establishment/construction cost incurred for each SLM measure. This is the cost incurred in the year of implementation per hectare per year. Maintenance costs refer to the recurrent costs of labor and materials required to maintain the functionality of the SWC measures. Following standard practices in the literature such as Hurni et al. (2015) and consultation with SLM experts of the district, maintenance cost was estimated at 10% of the establishment cost. The cost of maintenance is assumed to remain constant over the planning horizon of the project.

Accordingly, for SLM-adopting farmers, costs refer to labor costs involved in the initial construction and maintenance of SWC measures and costs of barley production. Except for money spent on the purchase of grass seedlings, all other costs imply labor costs.² On the other hand, for farmers without SWC measures, costs incurred for the production of barley, i.e. only fertilizer costs are considered. In this sense, the additional financial outlays involved in SWC measures are the incremental costs of SLM intervention in the treated area.

2.5.3. Cost-benefit appraisal criteria

Estimating the financial viability of SLM involves quantifying the changes in benefits and costs arising from the intervention. That is, net benefits from crop production without SWC measure and with SWC practices were compared to choose the most financially viable option(s). That is, CBA was conducted using incremental net benefits from SLM interventions (ELD Initiative, 2019), which is obtained by deducting net benefits in the baseline scenario from the net benefits with the SLM scenario. For this purpose, the incremental benefits and costs of SLM are estimated. The incremental benefits and costs due to SLM practices are obtained as

$\begin{array}{rl}& \mathrm{I}\mathrm{n}\mathrm{c}\mathrm{r}\mathrm{e}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{b}\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{f}\mathrm{i}\mathrm{t}\mathrm{s}\mathrm{o}\mathrm{f}\mathrm{S}\mathrm{L}\mathrm{M}=WithSLMbenefits\text{break}-WithoutSLMbenefits\\ \text{break}& \mathrm{I}\mathrm{n}\mathrm{c}\mathrm{r}\mathrm{e}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{s}\mathrm{o}\mathrm{f}\mathrm{S}\mathrm{L}\mathrm{M}=WithSLMcosts\text{break}-WithoutSLMcosts\end{array}$

Incremental benefits and costs are estimated annually as the difference between without-SLM and with-SLM farmers’ benefits and costs, respectively. In this study, the incremental net benefit from SLM interventions is derived as follows,

$\begin{array}{rl}& IncrementalnetbenefitofSLMpractices\text{break}=IncrementalbenefitsofSLM\text{break}-IncrementalcostsofSLM\end{array}$

That is, the incremental net benefit of SLM practices can be obtained by deducting the incremental cost of SLM from the incremental benefits of SLM. The decision to adopt then depends on whether positive incremental returns are generated from a given SLM practice.

Alternatively, the incremental net benefit of SLM can be calculated as follows,

$IncrementalnetbenefitofSLMpractices\text{break}=WithSLMnetbenefits\text{break}-WithoutSLMnetbenefits$

The net benefit derived from SLM is computed as follows,

$WithSLMnetbenefits=WithSLMbenefits\text{break}-WithSLMcosts$

Similarly, for the business-as-usual scenario, the net benefit can be computed as follows:

$WithoutSLMnetbenefits=WithoutSLMbenefits\text{break}-WithoutSLMcosts$

Some key assumptions are made to estimate the incremental net benefit of SLM practices. To calculate the incremental net benefit of SWC measures, the current observed incremental benefits from barley production and incremental costs are assumed to remain constant over the lifetime of the project. Nonetheless, since the maximum lifetime of soil bunds is estimated to be 12 years, farmers adopting soil bunds on their croplands are assumed to face investment costs twice. The financial returns are estimated based on conservative assumptions about future barley yield and benefits over the lifetime of the SWC measures;

1. Given crop production fundamentals, the current yield gap between the treated sub-watershed and the control group will remain the same for the economic lifespan of the SLM project.

2. Based on the prior assumption, the monetized value of incremental barley yield due to SLM is also assumed to remain constant throughout the lifetime of the SWC measure.

It should be noted, however, that the future differences in yield between the two groups are expected to be substantially larger than the current one. This is because the long-term effect of SWC measures on crop productivity is often higher than the immediate short-term impacts. Notwithstanding this, the approach allows a conservative estimate of SLM returns. Indeed, some of the assumptions made about future crop yield and SLM benefits appear to yield a lower estimate of returns. Such assumptions are common in the cost-benefit analysis literature (Tesfaye et al., 2016).

The financial viability of SLM interventions is assessed using the least possible level of benefits obtained from barley production alone. To evaluate the returns of SLM interventions, the benefit-cost ratio (B/C ratio) and the net present value (NPV) are employed. The benefit-cost ratio (BCR) measures the ratio of the present value of incremental benefits to incremental costs. The net present value(NPV) of adopting the SLM measure is given by

$NPV=\sum _{t=0}^T\frac{\left(B_t-C_t\right)}{{\left(1+r\right)}^t}$

Where T refers to the period over which costs and benefits are estimated, r is the real discount rate. B and C refer to incremental benefits and costs for farmers implementing SLM measures compared to a similar group of non-adopting farmers.

It would be in the farm household’s financial interest to adopt the new system if the net present value of the incremental return from switching were positive (NPV >0). A positive NPV estimate for a given conservation measure can be interpreted as showing that adoption of that measure would profit the farmer (Lutz et al., 1994a). Likewise, if the BCR is greater than one, investment in that specific conservation technology is profitable for farmers. However, if BCR is less than one or NPV is negative, the incremental costs outweigh the benefits of SLM, meaning that investments in SLM measures were not economic for adopting farmers.

2.5.4. Discount rate and project duration

Once cost-benefit appraisal indicators are chosen, the next steps in the CBA are choosing the discount rate and project lifetime. Regarding the period over which cost and benefits are projected, the maximum lifetime of SWC measures is used. The time horizon over which the cost-benefit analysis was carried out was 27 years, which is the maximum lifetime of fanyajuu bunds. Following Tesfaye et al. (2016), the comparison of NPV and BCR of SLM practices is based on the longest lifetime of 27 years. However, soil bunds are assumed to have only 12 years of life. Thus, to make the three SLM measures comparable over this period, farmers are assumed to face investment costs for soil bunds twice. Thus, using 2019 as a reference year, benefits, costs, and returns were projected for the period 2019–2046.

To discount costs and benefits occurring at different points of the project lifetime, a real interest rate of 7.3%, which was the discount rate in Ethiopia, is applied. This was calculated from the lending interest rate and inflation rate π following (Fisher, 1930) as$r=\hspace{0.17em}\frac{i-\pi }{1+\pi }$. Following Mekuria et al. (2011), the real discount rate r was determined based on the inflation rate and lending interest rate information for the recent consecutive years (2011–2017).³

The (geometric) mean of the inflation rate between 2011 and 2017, which is estimated based on data taken from the World Bank’s World Development Indicator database, is 11.9 %. Then, this rate is used to calculate the real discount rate. Accordingly, the real discount rate is found to be 7.3 %. The real discount rate is comparable with the discount rate used by the National Bank of Ethiopia (NBE). The discount rate, also called the bank rate, used by the NBE was 7 % in 2018/19 (Tradingeconomics.com, 2020).

3. Results

3.1. Estimated costs of SLM practices

This section presents the type of SWC technology adopted by farmers and their estimated costs. The initial establishment and maintenance costs of each SWC measure for the selected sub-watersheds are analyzed. Table 1 reports the average per-hectare costs incurred for constructing and maintaining each physical structure in the watershed.

Table 1. Estimated costs of SLM interventions

Major SWC measures	No. of households implemented(Percentage)	Treated area size(km or ha)	Mean EstablishmentCost ($)	Mean MaintenanceCost ($)
Soil bund	27(17.64)	14.25km	175.1	17.5
Fanyajuu bund	26(16.99)	13.3km	233.05	23.3
Grass strips on bunds		9.58km	28.77	2.87
			17.19^a
			11.66^b
Fanyajuu bund with grass strips	100(65.35)	9.58km	261.8	26.1
Mean cost of SLM per hectare			171.1	17.1

Note: ^a and ^b refer to labor cost and seedling cost of grass planting. The average cost of fanyajuu bund stabilized with Desho grass planting per hectare is obtained as the sum of the per-hectare labor cost of fanyajuu bund construction and grass planting along the bund.

Table 1 shows that the most widely practiced SLM measure was fanyajuu bunds stabilized with grass, where fodder grass is planted along the bunds to stabilize the bund and also help farmers benefit from grass production. This was practiced by 65.3% of the households and implemented on 9.58 ha of land during the intervention period. The average cost of this structure was $261.8 per hectare. Fanyajuu bund stabilized with grass is the most expensive structure among the SWC measures implemented by farmers. This is because its construction is arduous and labor-intensive. Fanyajuu bunds (without grass planting) and soil bunds were implemented by about 17% of the households. Investment in the fanyajuu bunds structure was found to cost $233 per hectare. Fanyajuu bund stabilized with grass is the second most costly technology. Soil bunds technology was implemented on 14.25 hectares of land. The mean cost of the establishment was found to be $175 per hectare. As compared to the other SWC structures, this is the most affordable technology.

In terms of the costs of establishment, other studies in Ethiopia have reported significantly lower investment costs of SWC measures at a watershed level. Tesfaye, Brouwer, van der Zaag, & Negat (2016) estimated the initial investment cost of soil bunds and fanyajuu at about $29 and 87 per hectare, respectively, compared to the per-hectare costs of $175 and $233 in this study. This might arise from differences in topography and location of the study sites, the time gap involved, and the associated increasing costs of labor over time.

Table 1 also reports the average cost of implementing SWC measures in the case study area. For the treated sites, the mean initial investment cost of implementing SWC measures on farmlands in the study site was $171. The average per hectare investment costs of SWC measures in the study area were significantly lower than those reported by Hurni et al. (2015) for Ethiopia, who estimated national average construction costs as high as $315 per 1 km (or hectare) of physical structures. This might be attributed to the fact that this study considered only labor costs. Nevertheless, if the (opportunity) costs of own equipment and tools used in the construction of SWC measures were incorporated in the estimation of costs, the costs would have been substantially higher than the one reported here. Notwithstanding the variation in the reported costs of SWC measures in Ethiopia, the initial construction cost is still considerable, particularly from resource-poor smallholder farmers’ perspectives. This remains one of the major explanations why poor smallholder farmers in Ethiopia are not investing to a large extent in the conservation of their croplands.

3.2. Benefits of SLM

Valuation of benefits from a given project is vital to determine whether the benefits of a project justify its cost, or whether a project is a cost-effective way of meeting a given objective. Many of a project’s costs and benefits are likely to be environmental, and valuation allows these costs and benefits to be included in the overall analysis (Silva & Pagiola, 2003). The incremental or marginal benefits deriving from SLM measures are quantified using the monetary value of major crop production (barley) for households taking on SWC measures. Table 2 reports the mean yield in the SLM program and non-SLM areas and estimates the mean increment in crop yield and revenue for adopting farmers. The mean incremental barley yield is the average change in crop yield in the SWC sites relative to the business-as-usual scenario.

Table 2. Benefits of SLM interventions: barley production

Scenario	Mean yield (Kg) per ha	Mean incremental barley yield (Kg) per ha	Mean incremental monetary value of the benefit (USD) per ha*
Without SLM	2,182
With SLM	2,733	551	150.75
Soil bunds	3,495	1,313	359.23
Fanyajuu bunds	3,417	1,235	337.89
Fanyajuu bunds &grass	2,253	71	19.24

*The average local farm gate price of barley per 100kg was ETB 800 (or USD 27.35).

Table 2 shows that the mean barley yield per hectare was 2,733 and 2,182 kg in the intervention site and comparison group, respectively. The data revealed that farmers who adopted SLM interventions between 2012 and 2017 experienced a yield increment of 551 kg per hectare in 2019 compared to similar non-adopting farmers in the area. On average, this shows that barley yield increased by more than 28% for the group adopting SLM technologies. The observed differences in crop yield between the treated and control groups are assumed primarily to reflect the differences in productivity arising from the adoption of SWC measures. This is in accord with the notion that differences in the market value of output are also viewed to reflect differences in crop yield between the treated and control groups.

The mean yield and benefit increment obtained in this study are in line with the empirical literature on the yield impact of SWC measures. Ayalew (2011) found that SLM practices increased crop yield by 22% on some farms and 15-fold on other farms within one year of intervention in the Wolaita area of southern Ethiopia. Similarly, study results by Asrat and Simane (2017) found nearly similar results from the Dabus sub-basin in the Blue Nile basin in Ethiopia. They reported that households that implemented SLM practices within the period (2004–2009) experienced a 24.1% higher value of crop production over non-users in 2016. Similarly, from a synthesis of empirical studies on the impact of SWC practices on crop yield, Adimassu et al. (2017) noted that stone bunds increased crop yield by 322 kg per hectare.

The results also revealed that crop productivity varies across the type of SLM technology adopted by farmers. Soil bunds have the highest impact on crop productivity (60%), followed by fanyajuu bunds (56%) and fanyajuu bunds stabilized with grass strips (3%). In terms of the quantity of output produced (kg) on a per-hectare basis, the incremental yield was as high as 1,313 for barley plots with soil bunds. The lowest increase in barley yield (71 kg per ha) was observed for plots with fanyajuu bunds stabilized with grass.

Ayalew (2011) also reported a yield increment of over 50% on fanyajuu and soil bunds after 3 years of construction in southern Ethiopia. In the Gedeb watershed of Ethiopia, Tesfaye et al. (2016) reported a gain in crop productivity of 24 and 17% from fanyajuu and soil bunds in the Gedeb watershed, respectively. However, the impacts on crop yield found in this study contradicted Adimassu et al. (2014) who reported that soil bunds do not increase crop yield in the Central highlands of Ethiopia. Soil bunds rather reduced crop yield by about 7 percent per hectare as compared with control plots. The attributed the reduction in yield in SLM-treated plots to the reduction of the cultivable area by 8.6 percent due to the construction of soil bunds.

The yield increment for plots that implemented fanyajuu bunds with grass strips, which is only 3%, is significantly lower than the average yield increment for the treated area. The below-average yield increment from fanyajuu bunds stabilized with grass appears to show two possible factors. The first and most important reason is related to the fact that fanyajuu bund construction and planting grass along the bund is soil-intensive and results in loss of cultivable land. That is, fanyajuu bunds reduce the area under cultivation by a significant percentage. It is argued in the literature that the effective use of unutilized land (the area occupied by bunds) by planting multipurpose grasses and trees on the bunds may offset the yield lost due to a reduction in planting area (Adimassu et al., 2017). If farmers are to benefit from installing bunds, productivity must not only increase but must increase by more than is lost by the reductions in cultivation area (Kassie et al., 2008). The lower productivity may arise from the substantial loss of cultivable land for the construction of bunds and planting grass which is uncompensated by a significant rise in productivity.

The second complementary factor is that the regular function of the structure tends to be affected by livestock tethering/grazing on the grass planted along the bund, particularly after crop harvest. The grass planted on the bund was meant primarily to stabilize the structure and benefit the farmers as a source of fodder for their livestock, albeit mainly through cutting and carrying the grass. Nevertheless, rather than the cut-and-carry system, some farmers are observed practicing tethering their cattle on/near the bund and on crop residue after crop harvest, which may reduce the crop yield impact of the structure. In addition, the economic value of feeding cattle on grass strips was not accounted for in this study as none of the farmers reported additional benefits from the sale of grass. Therefore, the lesser increment in barley productivity for farmers implementing this bund might be associated with the use of grass for livestock grazing and substantial loss of cultivable land to construct bund and fodder grass planting.

3.3. Cost-benefit analysis results

The financial effectiveness of SLM practices for adopting farmers in the watershed is analyzed using cost-benefit analysis techniques over a relatively longer project period. For this purpose, the stream of incremental net benefits of the selected SLM intervention was estimated relative to the baseline scenario of no conservation. This involves comparing costs and benefits in a business-as-usual scenario, which is non-conserved smallholder farmland without SWC measures, with an enhanced scenario of conserved cropland with SWC measures. Using costs and benefits data in both scenarios, the incremental/added financial net benefit of SWC measures is determined for barley-cultivated conserved croplands relative to the baseline scenario. The CBA approach allows a detailed assessment of the benefit-cost ratio and net present value of SLM interventions. The approach allows examining whether SWC leads to positive returns.

Table 3 shows that the total NPV deriving from SLM practices in the intervention sites is projected to be positive for the discounting period 2019 to 2046. The aggregate NPV of investments in SLM practices is $71,210. That is, investment in SWC measures has led to aggregate incremental returns of $71,210 for SLM-adopting farmers relative to the business-as-usual scenario. This indicates that the present value (PV) of the benefits of SLM outweighs the present value (PV) of the costs of SLM interventions over the project lifetime.

Table 3. Returns of SLM practices in the Barcha-Adado watershed (2019–2046): NPV and benefit-cost ratio of SLM

Appraisal indicators	SLM, average	Soil bund	Fanyajuu bund	Fanyajuu with grass strips along the bund
Total return from barley production (USD)
Total Incremental Benefits^a	199,586.8	86,952	69,474.5	15,167.4
Total Incremental Costs^a	30,263.8	11,693	3,515.4	10,527.7
NPV	71,210	31,932	28,696	623.8
B-C ratio	5.16	5.88	15.02	1.1
Mean NPV of SLM(USD/ha): barley production
NPV($/ha)	1,491.6	3,433.5	3,632.5	20.8

Note: “^a” refers to undiscounted incremental benefits and costs over the project period.

Returns, however, were observed to vary with the type of SLM measure adopted by farmers. The additional total net benefit (discounted) from implementing soil bunds, fanyajuu bunds, and fanyajuu bunds stabilized with grass strips was higher by $31,932, 28696, and $623.8 compared to the control group, respectively. The aggregate NPV was highest for investments in soil bunds, followed by fanyajuu bunds relative to the baseline of no SLM intervention.

Evaluated against the benefit-cost ratio criterion, Table 3 suggests that the BCR of investments in one of the SLM practices is 5.16, on average. This implies that the present value (PV) of the aggregate benefits of investment in SLM measures is more than five times higher than the present value (PV) of the total costs of intervention. Table 3 also indicates that there is a substantial variation in returns among the SLM technologies in terms of the BCR. The variation in returns ranges from a marginal benefit-cost ratio of 1.1 for fanyajuu bunds stabilized with grass strips to as high as 15 for farmers adopting fanyajuu bunds. Notwithstanding the variation in returns, the findings show that all SLM practices have BCRs greater than one, indicating that the PV of incremental benefits of SLM measures outweighs the PV of incremental costs for adopting farmers. This proves that investments in SLM can generate positive financial returns for adopting farmers in the long term. The benefit-cost ratio is highest for fanyajuu bunds, followed by soil bunds. The least financial return is found to be generated from fanyajuu bunds stabilized with grass strips.

The BCRs of soil bunds and fanyajuu bunds projected in this study are considerably higher than those reported in Tesfaye et al. (2016) in the Gedeb watershed in the Upper Blue Nile basin in Ethiopia. Using increased crop revenue as a major benefit, they reported benefit-cost ratios of 1.24 and 1.03 for soil bunds and fanyajuu bunds, respectively. Tesfaye et al. (2016) results, however, are comparable to the BCR obtained from fanyajuu bunds stabilized with grass strips in this study. Moreover, the average BCR projected for the study area is higher than those reported in Tilahun (2020) for Ethiopia. At a country level, Tilahun (2020) reported that the BCR of SLM interventions for preventing soil Nitrogen, Phosphorus, and Potassium (NPK) depletion and NPK losses in Ethiopia are 4.05 and 4.60 for the discounting periods 2020–2030 and 2020–2040, respectively. For the Gedeo Zone, the area where this study site is located, Tilahun (2020) found a BCR of 7.02. This is considerably higher than the average BCR obtained from implementing any of the SLM measures in this study, but fanyajuu bunds.

On a per-hectare basis, Table 3 shows that the NPV of SLM interventions is projected to be $1,491.6 per hectare. This means that investments in SLM practices can generate an additional NPV of $1,491 per hectare compared to the return in the business-as-usual scenario. It is noted that the returns to SLM investments were found to vary by the type of conservation measure implemented. The NPV is highest for fanyajuu bunds ($3,632.5 ha^_1), followed by soil bunds ($3,433.5 ha^_1) and fanyajuu bunds stabilized with grass strips ($20.8 ha^_1). The NPV of investments in SLM practices is $56 per hectare per year (ha⁻¹yr⁻¹). The highest net return is projected to be derived from fanyajuu bunds ($134.5 ha⁻¹yr⁻¹), followed by the soil bunds ($127 ha⁻¹yr⁻¹); whereas the least return was generated from fanyajuu bunds stabilized with grass strips (0.77 ha⁻¹yr⁻¹).

Despite some differences in the magnitude of returns, the results of this study are in accord with the empirical literature. Nkonya et al. (2014) found that the NPV derived from SLM interventions on maize cropland was $ 119 per hectare per year in Bhutan. In a detailed review of empirical studies in Ethiopia, Adimassu et al. (2017) synthesis of results showed that the NPV of most physical SWC practices was positive and varied between 1,675 and 5,283 ETB ha⁻¹ yr⁻¹. Though returns are not reported on a per-hectare basis, Tesfaye et al. (2016) reported positive returns from implementing soil bunds and fanyajuu bunds in the Gedeb watershed of Ethiopia. Similarly, Matere et al. (2016) reported that adopting fanyajuu bunds on maize-pigeon pea-producing farmlands in semi-arid areas of Kenya was profitable. On the contrary, Adimassu et al. (2017) showed that returns from graded soil bunds and fanyajuu bunds were negative in Ethiopia. They noted that returns from fanyajuu bunds became economically profitable when grasses were planted on bunds and the price for these grasses was considered in the cost-benefit analysis.

To visualize the returns of scaling up of SLM measures in the whole watershed, the aggregate NPV of SLM practices is projected over the next 27 years across 6,043.8 hectares of cropland that needs intervention. Using the mean per-hectare NPV of SLM, the results are extrapolated to the total cropland area of the Barcha-Adado watershed. Using the average return for the case study area, investments in the conservation of croplands in the entire watershed would generate additional returns with a net present value of $9,014,932 over 27 years compared to the business-as-usual scenario. The watershed level NPV was found to vary with the type of SLM measures adopted on barley-cultivated plots. The establishment of soil bunds and fanyajuu bunds on croplands in the watershed would generate additional NPV of $21,954,103.5 and 20,751,387, respectively.

3.4. Sensitivity analysis

The reported results so far were based on the assumption of a constant discount rate, prices, and yield over the project lifetime. Since the CBA result may be sensitive to changes in the parameters, sensitivity analysis is performed by relaxing the assumptions made regarding discount rate and output. According to Tilahun et al. (2016), the real discount rate may change due to changes in either inflation or the nominal interest rates or both while crop output levels may change due to farm-related or other factors. Thus, the sensitivity of the NPV to changes in the discount rate and output was carried out.

To investigate the sensitivity of NPV to changes in the Ethiopian real discount rate and yield, a 25% change in the discount rate and yield is simulated. The simulation is made using a moderate change in the parameters following the literature on SLM studies in Ethiopia (see Mekuria et al., 2011). Using the simulation exercise, the effects of changes in yield and Ethiopian real discount rate on net revenues from crop production were evaluated.

The effects of changes in yield and real discount rate on the NPV of SLM interventions were simulated. Table 4 shows that a 25% increase in crop yield would substantially increase the NPV of soil bunds, fanyajuu bunds, and fanyajuu bunds integrated with grass strips to $ 6,290.5, 6,324, and 1,795, respectively. On the other hand, a 25% decrease in crop yield would result in a considerable decrease in the returns of soil bunds, fanyajuu bunds, and fanyajuu bunds integrated with grass strips to $ 680, 940, and −1,754 per hectare, respectively.

Table 4. Sensitivity of NPV ($/ha) to changes in the discount rate and barley yield

Factors		SLM measures, crop production
	Changes	Soil bund	Fanyajuu bund	Fanyaju with Grass	SLM, average
	Baseline	3,433.5	3,632.5	20.8	1,491.6
Barley Yield	−25 %	680.3	940.6	−1,754	−661.3
	25 %	6,290.5	6324.3	1795.6	3,664.8
Real discount rate	−25 %	4,276.5	4,227	−49	1,829
	25 %	2,991.3	3,007.7	−54.6	1272.6

If the real discount rate in Ethiopia increased by 25%, Table 4 also reveals that the net return of SLM interventions would decline substantially over the project lifetime. Nonetheless, SLM interventions would still be economically attractive except for fanyajuu bunds stabilized with grass strips when the Ethiopian real discount rate increases.

The results in Table 4 demonstrate that the average NPV of SLM intervention was highly sensitive to a 25% decline in crop yield, resulting in a negative net financial return. In addition, the simulation exercise also demonstrated that fanyajuu bunds stabilized with grass strips were highly sensitive to the decline in crop yield of 25% and an increase in the discount rate of 25% where its NPV would turn from a little over 20 $/ha to negative 1,754 and 54 $/ha, respectively. This suggests that SLM intervention, on average, would generate higher net revenue and thus be financially feasible for higher crop yield per hectare and a modest real discount rate.

3.5. Discussion

The adoption of SLM practices involves both significant costs as well as benefits. The financial returns are determined, among others, by the establishment and maintenance costs of SLM measures, the benefits derived from implementation, and the period over which the stream of benefits and costs are projected.

As the extant studies focused solely on the northern and north-western highlands of Ethiopia, the effectiveness of SLM practices in the Southern Highlands of Ethiopia is rarely documented. This study contributes to the literature by providing new evidence on the costs, benefits, and returns of SLM practices in the Southern Highlands of Ethiopia.

This study found that the implementation of SLM measures involves considerable establishment and maintenance costs for smallholder farmers. In the selected study sites, only 33.3 percent of the farmers are non-adopters. This suggests that the adoption rate of SLM practices is low. According to the World Bank (2008), the adoption rate of SLM practices is low due to the significant costs and labor required for the initial setup of SWC structures, as well as a lack of awareness and limited access to credit. Amsalu and de Graaff (2007) highlighted that the costs of adopting SLM are substantial, with the upfront expenditure necessary for terracing being prohibitive for farmers. Similarly, Teshome et al. (2016) also revealed that the labor input required for SLM measures was high, which could impede the adoption and continuing use of such techniques. The high costs of implementation are one of the major reasons that discourage resource-poor farmers from adopting SLM practices. Furthermore, understanding and executing these new measures necessitate financing for training, awareness-raising, and monitoring programs.

This study found that implementing SLM practices increases crop yield, on average, by 28 percent relative to a control group of non-adopters. SLM practices generated benefits in terms of increased crop yield and revenue. The positive changes in benefits are in accord with the literature. Schmidt et al. (2014) found that households that invested in SLM infrastructure on their agricultural plots between 1992 and 2002 in northern Ethiopia and subsequently maintained those structures had a 24 percent higher value of production in 2010 than farming households that did not make such SLM investments. Similarly, Vogl et al. (2017) noted that SWC measures benefit farmers through higher production and increased revenues for Kenya.

The benefits are found to vary by the type of SLM technologies implemented by the farmers. Implementing soil bunds increases crop yield by 60%, followed by fanyajuu bunds (56%). Fanyajuu bunds stabilized with grass strips are found to be the least effective structure in enhancing crop yield (3%). The findings of this study differ from previous studies that examined the benefits of a single type of SLM practice in isolation (Kassie et al., 2007, 2008), Adimassu et al. (2014) (Matere et al., 2016)., One of the key gaps of these studies is their failure to compare and analyze the variation in ecological benefits of different types of SLM technologies implemented in Ethiopia. In this regard, the findings of this study contribute to filling in this gap and thereby allow a greater understanding of the variation in the benefits of different types of SLM practices and also could help farmers in identifying the most effective technology.

The results showed that farmers with SWC have derived a substantial amount of additional benefits from crop production. The incremental benefit from SLM interventions is viewed to represent a difference in crop productivity between the intervention area and the control group. With such meaningful positive returns in yield and value of crop production, it should be in the best interest of farmers to implement SWC structures. The increment in yield⁴ due to SLM is derived importantly from the contribution of SWC measures to the reduction of soil erosion by water and thereby reduced loss of soil fertility. It is also widely noted in the literature that SWC measures through reduced erosion and losses of soil and associated nutrients are found to improve soil fertility and water retention (Vogl et al., 2017). The results revealed that there is positive productivity gain from SLM interventions in the intervention area. It is widely noted that the main effect of SLM is to increase production by decreasing runoff and erosion, as well as increasing water capture in the soil (Schmidt et al., 2014). The intervention enabled farmers to save the loss of farm inputs, particularly seed, fertilizer/compost, which had been washed away by runoff in the pre-intervention period. Such productivity improvements are considered an incentive for farmers to adopt the technology to conserve their land. Perhaps, if farmers are to benefit from implementing SWC measures on their cropland, crop productivity must significantly increase as compared to the baseline scenario.

In summary, despite the high costs farmers face in implementing SLM practices, this study indicates that when these management practices are implemented, they result in increased crop yields due to savings from reduced soil erosion and hence enhanced soil fertility. Widening the adoption of such practices is thus critical for reducing land degradation and enhancing environmental sustainability, greater agricultural productivity, and improved livelihoods and food security of farmers in an agrarian economy that is heavily reliant on the productivity of smallholder farmers.

The BCR and NPV results reveal that the present value of benefits outweigh the present value of costs of implementing SLM measures in the Southern Highlands of Ethiopia in the long term. This indicates that actions against land degradation were financially attractive in the long term. The financial returns are high given the fact that only a single ecological benefit, crop production, was considered in the estimation of benefits. Here it should be noted that the increment in NPV derived from implementing SLM practices could be higher if multiple crops and other ecological benefits were considered in the analysis. The results corroborate the findings of Hurni et al. (2015) and (Tilahun, 2020) that investments in SLM intervention generate positive and high returns in Ethiopia.

Financial returns are observed to vary with the type of SLM practice implemented by farmers. Using both the NPV and BCR appraisal criteria, the most financially attractive investment on a per-hectare basis is the fanyajuu bunds investment. The least return is found in plots that adopted fanyajuu bunds stabilized with grass strips. The results show that profitability depends on the type of SLM measure adopted by farmers. Tesfaye et al. (2016) also reported variations in financial returns of three types of soil and water conservation measures in Ethiopia. However, in their study, soil bunds were found to be the most profitable one, followed by fanyajuu bunds.

Fanyajuu bunds stabilized with grass strips appear to be less profitable in comparative terms because, apart from the relatively lower incremental crop revenue, the construction of fanyajuu bunds with vegetative measures was the most costly labor-intensive technology compared to the other two structures. Moreover, the value of grass was not considered in the valuation of benefits. Perhaps, none of the farmers in the area reported additional income earned from grass sales planted to stabilize fanyajuu bunds. For this reason, additional benefits that could be generated from the sale of grass strips planted on fanyajuu bunds are not considered in the valuation of benefits. Though it seems a strict assumption, as noted by Tesfaye et al. (2016), it still gives a conservative estimate of the benefits.

This study shows that the private benefits of SLM practices, measured solely based on benefits in terms of the incremental value of crop revenue, clearly outweigh their initial establishment and maintenance costs. The financial return of SLM evaluated even using barley production alone suggests that the present value of benefits of SLM is significantly higher than the present value of costs of SLM interventions. That is, though the CBA considered a conservative estimate of benefits that could be generated from investments in SLM measures, SLM practices could generate considerable profit for adopting farmers. This suggests the importance of projecting the stream of benefits over a longer time horizon. Perhaps, in contrast to those studies that considered only short-term benefits (Adimassu et al., 2014; Kassie et al., 2007, 2008) and those that used a relatively shorter project period of 10 or 20 years, for instance, Tilahun (2020), this study projected the stream of benefits, costs, and returns over a relatively longer period of 27 years. This is important since the returns of SLM accrue significantly in the long term. This notion agrees with Tesfaye et al. (2016) analysis that a longer time perspective is needed to realize positive private financial returns.

The sensitivity analysis performed also shows the robustness of the profitability of SLM practices to changes in crop yield and discount rate. In particular, the returns (NPV) of SLM practices are fairly robust to changes in crop yield and discount rate, except for fanyajuu bunds combined with grass strip technology.

As other factors that influence adoption become less relevant in the absence of substantial profitability (Hurni et al., 2015), the positive financial returns can be taken as key evidence to convince non-adopters to implement SLM practices. According to the World Bank (2008), investments must be low cost, low risk, and enhance earnings to be appealing to smallholders. This evidence is also vital to scale up the benefits of SLM practices in the study area and other similar watersheds in the southern highlands of Ethiopia.

Moreover, at the watershed level, the considerable return obtained from scaling up SLM measures to the whole cropland in the Barcha-Adado watershed demonstrates that investments in SWC measures at larger scales could be economically worthwhile in the long term. It should be noted that the returns of SLM practices could have been much higher if the economic values of other ecological benefits such as soil nutrients, carbon sequestration, and other societal benefits were considered in the cost-benefit analysis.

The results demonstrate that while the initial expenditures of implementing SLM in Ethiopia can be prohibitively expensive, the long-term benefits make SLM measures economically feasible and beneficial. Strategic policy and financial support for SLM practices are vital for improving land productivity, food security, and enhanced welfare gain in the future.

4. Conclusion

The benefits, costs, and financial returns of SLM practices were rarely documented in the Southern Highlands of Ethiopia. This study examines the returns of SLM interventions using cost-benefit analysis in the Barcha-Adado watershed in Southern Ethiopia. The major SLM measures practiced in the selected case study area were soil bunds, fanyajuu bunds, and fanyajuu bunds stabilized with grass trips. Such actions against land degradation involved considerable costs and benefits. Methodologically, the study employed a cost-benefit analytical framework. CBA is conducted using incremental net benefits from SLM interventions, which are obtained by deducting net benefits in the business-as-usual scenario from the net benefits in the SLM scenario. To compare costs and benefits in the SLM intervention area with the business-as-usual scenario, the study applied a quasi-experimental design. Four representative micro watersheds are selected from the watershed: two with SLM (treated group) and two without an SLM intervention (business-as-usual scenario).

The approach followed in this study involves estimating the economic values of benefits and costs as well as performing a cost-benefit analysis (CBA) of investments in sustainable land management (SLM) practices in southern Ethiopia. The stream of benefits, costs, and returns of SLM practices were projected for the period 2019 to 2046. To discount the stream of costs and benefits occurring at different points of the project period, a real interest rate of 7.3% was applied. The returns of SLM interventions were evaluated using the benefit-cost ratio (BCR) and the net present value (NPV). The SLM program is worth undertaking if the present value of incremental net benefits is positive, i.e. if the net benefits are greater for the with-SLM scenario than for the without-SLM scenario.

The study’s results are important from several dimensions. First, it provides new evidence on the costs, benefits, and returns of SLM practices over a relatively longer time (2019–2046) in the Southern Highlands of Ethiopia. The findings can contribute to filling in the knowledge gap on the costs, benefits, and returns of SLM measures in the Southern Highlands of Ethiopia. Second, this study documented not only the yield impact but also the costs and returns of three distinct types of SLM technologies implemented on smallholder cropland in the Southern Highlands of Ethiopia. In this regard, the results allow an understanding of the variation in costs, benefits, and returns of different types of SLM technologies implemented in Ethiopia. This in turn can support policymakers in designing more effective SLM technology and can facilitate the implementation of the most viable SLM technology by farmers. Third, the returns of SLM practices are assessed over a relatively longer time perspective and the total return of implementing SLM practices at the watershed level is estimated. The estimated returns of investments in SLM measures in the watershed are instrumental in showing the economic significance of improving the productivity of croplands for local decision-makers.

Study results show that SLM interventions are found to involve considerable costs for the establishment and maintenance of SLM technologies in the form of labor and material costs. This study found that the average establishment cost of SLM practices was $171 per hectare, albeit the fact that the initial investment cost of each conservation structure varies from an average of 175 USD for soil bunds to about 262 USD per hectare for plots that practiced plantation of fodder grass on fanyajuu bunds.

Following SLM, a significant increment in crop yield and revenue was observed for farmers implementing SLM measures. On average, SLM-adopting farmers experienced a 28% higher value of crop yield and revenue compared to non-adopting farmers. In terms of incremental crop yield, soil bunds are found to be the most effective SLM technology, followed by fanyajuu bunds. The results demonstrate that the construction of SLM measures on degraded cropland can generate substantial benefits for farmers in the form of increased crop yield, which would be vital to increasing crop productivity, reducing food security and poverty, and enhancing the welfare of subsistence farmers.

The CBA shows that the mean incremental NPV and BCR of investments in SLM practices are projected to be 1,491 USD per hectare and 5.16, respectively. On a per-hectare basis, the highest returns are projected to be found in plots that implemented fanyajuu bunds, followed by soil bunds. The CBA results reveal that investments in SLM measures can generate a positive financial return for adopting farmers compared to the business-as-usual scenario in the Southern Highlands of Ethiopia. Investments in SLM measures across the whole cropland in the watershed are projected to generate an additional total net present value of $9,014,932 relative to the business-as-usual scenario. This suggests that further investments in SLM across degraded croplands would be profitable at a larger scale.

To validate the robustness of study results, a novel quasi-experimental approach and sensitivity analysis are used. First, a novel quasi-experimental design was used to choose the micro watersheds. To disentangle the benefits and returns resulting from SLM interventions, the quasi-experimental approach is used to assure comparability between the SLM intervention site and the business-as-usual scenario (micro watersheds without SLM measure). To ascertain this, control micro watersheds in the same agroecology that had a history of similar land degradation and similar farm characteristics before the implementation of SLM practices were utilized as a ‘business as usual’ scenario. Secondly and importantly, the sensitivity analysis shows the robustness of the returns of SLM practices to changes in crop yield and discount rate in the discounting period. On this basis, meaningful conclusions can be drawn from the study.

SLM practices were found to lead to a significant increment in crop productivity and revenue for farmers taking SWC measures compared to non-adopting farmers This implies that SLM has the potential to improve food security and livelihood of smallholder farmers and can contribute to efforts of poverty reduction. In this sense, scaling up land conservation practices can contribute to the overall growth of smallholder agriculture and the improved welfare of the local community.

The evidence shows that there is considerable variability in the costs, benefits, and returns of various types of SLM practices. Using both the NPV and BCR appraisal criteria, the most financially attractive investment on a per-hectare basis was fanyajuu bunds investment, followed by soil bunds. Notwithstanding the variation in returns among SLM measures, the present value of benefits ensuing from the adoption of SLM practices significantly outweighs the present value of costs of establishing and maintaining the SWC structures in the long run. Thus, actions against land degradation are found to be financially viable for smallholder farmers in the Southern Highlands of Ethiopia in the long term.

The positive effects of SLM on crop productivity and substantial financial returns reveal that it is worth conserving agricultural land from degradation. Therefore, investments in SLM practices can be taken as ecologically desirable as such interventions enhance land productivity and are economically profitable as they generate considerable financial returns through conserving agricultural land and natural resources from degradation. The positive short-term gains and long-term returns can serve as vital evidence to persuade non-adopting farmers in the area to implement SLM practices on their cropland. Moreover, since investments in SLM are financially viable from farmers’ perspectives, this evidence serves as a further stimulus for adopting farmers to invest more and/or maintain SWC measures. Perhaps, there is a demonstration effect from adopting farmers to non-adopting farmers as the latter can learn from the former.

In conclusion, even though adopting and implementing sustainable land management practices in the watershed involves significant costs, the benefits in terms of increased crop yield and farm income, improved livelihoods, environmental sustainability, and financial returns highlight its viability and desirability. However, to effectively address the substantial costs facing smallholder farmers and ensure the continued adoption and success of the SLM practices, strategic planning, and supportive policies are essential. Moreover, regular maintenance of the existing SWC structures would be instrumental in sustaining the positive returns from SLM practices in the long term.

The findings have important ramifications for policymakers to make evidence-based decisions to scale up SLM interventions, enhance farm income and food security of farmers, and achieve land degradation neutrality in the southern highlands of Ethiopia and elsewhere. The major policy implications of the study are highlighted as follows.

First, since SLM interventions improve land productivity and generate considerable financial returns for farmers, the findings would serve as evidence for the local decision-makers and may provide incentives to convince farmers to adopt SLM measures in the future.

Second, the benefits of SLM practices need to be scaled up to enhance the welfare of farmers. This is because such investments in SLM can contribute to enhancing crop productivity and income and achieving food security and poverty reduction in the study area. Thus, given its implications for achieving food security and reduction of rural poverty, enhancing the productivity of croplands through SLM practices needs to be a top priority for farmers and the local government.

Third, study findings have important ramifications for scaling up land conservation practices in the face of ongoing land degradation, where a considerable portion of agricultural land in the area needs conservation. However, because implementing SLM measures is costly for resource-poor farmers, poor smallholder farmers may not be able to afford to make significant investments to conserve agricultural land from land degradation. Thus, there is a need for technical and resource support to help farmers increase their investment capacity and know-how, hence facilitating adoption. Moreover, in the face of resource constraints, a participatory approach to SLM at the peasant association and watershed level needs to be practiced.

Fourth, mainstreaming SLM in the study area can contribute to building a land degradation-neutral economy at the local, regional, and national levels, which is in line with the United Nation’s land degradation neutrality (LDN) goal envisaged to be achieved by 2030.

This study has some limitations. One of the limitations is that it considered only one aspect of the benefits of SLM measures, i.e. crop production. Moreover, returns are also estimated based on a conservative assumption about the incremental benefits of SLM over time. That is, the monetized value of incremental benefits due to SLM is assumed to remain constant over the discounting period 2019–2046. However, since the long-term effect of SLM measures on crop productivity is usually higher than the immediate short-term impacts, the incremental benefits are expected to grow substantially over the project period. The dimension of benefits considered and the assumption about incremental benefits may underestimate the returns of investment in soil and water conservation measures.

Other studies in the future may account for other ecological benefits such as the value of soil nutrients, and forest products, and social benefits such as carbon sequestration in the estimation of benefits. Moreover, since the experimental design employed in this study is quasi-experimental, differences in yield may not exactly capture the net effect of SLM interventions. In this regard, future research may need to employ other impact evaluation techniques such as propensity score matching that are capable of controlling other potential sources of differences in crop yield and benefits.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Kebele is the smallest administrative unit in Ethiopia.

2. A shovel/spade is the major tool used for the construction and maintenance of soil conservation measures. However, because the shovel serves various other purposes in farming too and is not only used for soil conservation measures (Tesfaye et al., 2016), farmers were assumed to spend no additional money on the purchase of shovels and other tools in the study area. For this reason, the costs of such equipment were not considered in the estimation of costs of SLM.

3. Since the inflation rate in Ethiopia is generally higher as compared to the lending interest rate, the average real interest turns out to be either negative or very low. To take this factor into account and use an appropriate discount rate, the maximum lending rate of 20% observed in 2017/18 is used (National Bank Ethiopia, 2017/18). Moreover, the maximum lending rate was applied since microfinance institutions provide loans to rural people at rates higher than the national average interest rate (see.

4. This result shall be interpreted with caution. It is difficult to assess empirically productivity effects of SLM technology adoption based on non-experimental analysis. Farmers are likely to select land management practices on their plots based on their endowments, level of awareness and education and farm-specific attributes of their plots (often unobservable). In addition, farmers might be systematically selected by policy makers and development agents to adopt the technology based on their propensity to participate in the adoption of technologies. Given that adoption is likely endogenous, simple comparisons of mean differences in productivity on plots with and without use of particular land management practices has no causal interpretation.