4.1. The cash flow patterns of agroforestry practices
Table 4 and Table A1-A4 present the cash flows of the different investment alternatives in the study area. The mean total undiscounted net cash flow andmean annual net cash flow of agroforestry practices are higher than monoculture cereals. Millet has the lowest mean total net cash flow compared to other practices. In agreement with this finding Chemuliti et al. (2019) in Kenya reported that trees have higher mean total net cash flows and annual net cash flows than cereal crops. The mean total net cash flow and annual net cash flow for coffee-shade practice are higher than other options. In all three agroforestry practices, the highest expense occurred in the first cash flow.
. Except for the coffee-shade practices at the beginning of cash flow, farmers ensure positive net cash annually (Fig. 2a-c). In line with this study, Martinelli et al. (2019) reported a positive cash flow for biodiverse agroforestry systems in Brazil except in the first year of the project. Martinelli et al. (2019) concluded that biodiverse agroforestry arrangements help farmers secure positive cash flow, reduce their cash shortage during establishment periods, and encourage them to spend on environmentally sound investment alternatives. Henceforth, the positive cash flow in our study area might be due to the diversity of species and outputs in practices. However, in the work of Nigussie et al. (2020), in Ethiopia's Upper Blue Nile Basin, cash flows were negative at the beginning and end of year three, and positive at year one and year two forsingle tree species dominated Acacia decurrens-based agroforestry system. Cash flows were also fluctuating for uniform multipurpose Eucalyptus woodlots in central Ethiopia (Duguma, 2013). This is linked to the establishment of a stand from a single tree species. Negative cash flows can potentially limit cash availability for farmers during establishment, which forces poor farmers to rely on annual cash crops (Nigussie et al., 2020). Hence, biodiverse agroforestry practices, which have no negative cash flows, are more motivating for both better-off and poor growers.
However, the cash flow patterns varied with the agroforestry practices. In homegarden agroforestry practices, the cash flows show three trends (Figure 2a). The first is a decrease from cash flow zero to cash flow three. This is due to the diversity of components in this practice, increasing competition, and subsequent management costs. Duguma (2013) noted that homegarden requires repetitive and continuous care such as weeding, cultivating (hoeing), and watering to reduce competition. The second trend was a sharp increase in revenue and net cash flow from the fourth to eighth years. This is due to the subsequent branching and coppicing ability of the dominant income sources, i.e., fruits and dogwoodleaves grow and accumulate high biomass. The leaves and shoots harvested annually or biannually for sale and household use compensate for the costs. The last trend, which started after the 8th year, was a gradual increase in revenue and net cash flows. It is a stability stage where the branching capacity of woody components reaches the climax to produce additional biomass production.
Dogwoodpracticeshave an overall uniform increasing cash flow pattern for revenue and net income (Fig. 2b). Farmers leave some branches and leave untouched for the sustainability of standing trees and for activating shoots for next year's production. This gives farmers a continuous supply of marketable dogwood leaves. Duguma (2013) also reported most women intentionally leave Catha edulis and R. prinoides stemsuntouched during the harvesting period to systematically organize the continuity of income. The cash flow pattern for the expenditure decreases in the first six cash flow periods, which reveals most of the expenditures in this system were incurred during the establishment stage.
The net cash flows of coffee-shade practices showed positive except for the first cash flow (Fig. 2c). From year four through year seven, the net cash flows climbed dramatically. The subsequent years had positive net cash flows, with considerable fluctuation in the quantity of revenue. This is because coffee yields fluctuate year to year in the coffee-shade practice. Farmers noted that the yield of coffee is higher one year and lower in the next year due to the physiological nature of the species and environmental conditions. Hence, the coffee practice has disadvantages of income fluctuation compared to other agroforestry practices. In general, the cash flow pattern of agroforestry practices was affected by the diversity of components, environmental conditions, and management characteristics (Guleria et al., 2017; Martinelli et al., 2019).
Table 4 here
Figure 2 here
4.2. Financial returns from different investment alternatives
4.2.1. Scenario I: Baseline profitability
The NPV and EAI ha-1 of homegardens and coffee-practices were positive across all households at r = 10%, and 20%. But NPV and EAI ha-1 of homegardens were not positive for 1% of households at r=30% and 50%. The NPV and EAI ha-1 of households at r = 10% had a skewness of 1.01 and a kurtosis of 3.30. This indicates the distribution is slightly positively skewed and has light positive tails and fewer outliers. In financial analysis, kurtosis measures the financial risk (extremely large and extremely small returns). A small positive kurtosis in our study signals a low level of risk because the possibilities of extreme positive returns are relatively low. Dogwood had positive NPV and EAI ha-1 across all households and discount rates. At r=10%, it has a normal distribution (skeweness = 0.05, and kurtosis = 2.05). Education level and wealth status significantly and positively affect AF returns (Table 5), which may most likely be due to better resource endowments (i.e., more land, more knowledge) to invest (Ajayi et al., 2009; Nigussie et al., 2020). The NPV ha-1 of cereal crops at r=10% were positive over households but slightly skewed to the left (skewness = - 0.07 to -0.52) and light kurtic (kurtosis = 2.34 to 3.09). The summary statistics of NPV and EAI ha-1 (r=10%) of agroforestry and cereal mono-cropping practices are shown in Table A5.
The mean NPV, EAI ha-1, and BCR of all three agroforestry practices and cereal crops at r = 10%, 20%, 30%, and 50% are presented in Table 6. The ANOVA result of NPV and EAI ha-1 at r=10% shows there was a significant difference (P < 0.05) among AF practices (Table 7). The coffee-shade practice has the highest NPV and EAI ha-1, 34.65%, and 26.45% higher than Dogwood and homegarden, respectively. BCR of the coffee shade practice is two-fold the value of dogwood and 1.5-fold of homegarden.
The results indicate all three agroforestry practices are humbly profitable at the existing principal interest rates from smallholder farmers’ perspective. The greater than unity BCR values reveal benefits of agroforestry practices offset the costs. The high NPV, EAI, and BCR in coffee practice were realized because of lower initial investment cost, high-value coffee products, and no trade-off between the shade tree and coffee yield. Agroforestry practices, which do not compete between components, achieve better production and benefits (Phimmavong et al., 2019). The lowest NPV and BCR in dogwoodpractice could be associated with relatively highest (establishment and management) costs (41 times higher than the cost of coffee at r = 10%). NPV is positive for agroforestry practices in Brazil (Martinelli et al., 2019), and in Acacia decurrens-based agroforestry in the Fagita Lekoma district, Northwestern Ethiopia (Mekonnen et al. (2021). However, Dalemans et al. (2019) concluded that cultivating Pongamia (Millettia pinnata) agroforestry in southern India is profitable for only small‐scale production. Hence, adopting agroforestry practices is useful for enhancing the financial opportunities of smallholder farmers. In addition, by improving household subsistence use, agroforestry can improve food security in rural areas (Phimmavong et al., 2019).
The estimated NPV value for agroforestry practices in our study was higher than the short-term (5-year rotation) Acacia cum teff agroforestry system (79,285 ETB ha-1) in Upper Blue Nile, Ethiopia. NPV was similar to a 30-year investment smallholders woodlot project (286, 200 ETB ha-1) in the Amhara region, Ethiopia (Matthies and Karimov, 2014); and tree and shrub growing homestead agroforestry (183,970 ETB ha-1)in Central highlands of Ethiopia (Duguma, 2013) at 10% discount rate. However, Duguma (2013) found comparably a higher NPV (1,215,585 ETB ha-1) for 15 years’ investment in smallholder woodlots which could be due to better land productivity, high stem density ha−1, and good market access. Planting on eroded land is significantly less profitable than planting on uneroded land (Matthies and Karimov, 2014). Comparing the returns from other countries abroad, the NPV of a 7-year rotation Eucalyptus-rice agroforestry (US$799 ) in Lao PDR at a 12% discount rate (Phimmavong et al., 2019) is lower than our study (the US[1] $6641 for dogwood, $7474 for homegarden and $10,162 for coffee-shade). The lower return could primarily be due to the short rotation time and relatively higher discount rate in Eucalyptus-rice agroforestry. Matthies and Karimov (2014) found a 22% decrease for every 1-year decrease in rotation interval time in Ethiopia's Amhara region. Increasing rotation interval time to the optimum level would contribute to smallholders attaining higher earnings.
Due to market imperfection, tree management knowledge gap, and poverty status, farmers prefer to depend on productions that require a lower initial investment, but a higher NPV: IEC ratio (Molua, 2005). The establishment financial requirement of coffee-shade agroforestry practice is still preferred, as it requires 41% and 18% less cost than dogwood and homegarden, respectively (Table 5, and Table A1-A3). Coffee-shade had the highest NPV: IEC ratio, two-fold to the homegarden and two and half-fold to dogwood. This implies coffee-shade practice has a great potential for enhancing income with the existing market constraint. However, homegardens of our study have various additional advantages, such as using family labor to manage the field, household wastes for fertilizer, and accessibility for monitoring, which were also reported by Duguma (2013) in the central highlands of Ethiopia.
Financially cereal mono-cropping practices are profitable to farmers (positive NPV, EAI, and greater than unity BCR) (Table 6). Like this study, Duguma et al. (2010) found greater than unity BCR value for cereal farming practices in the central highlands of Ethiopia. Nevertheless, their BCR finding is relatively lower (1.21) than the value in our study (1.59-2.12) at a 10% discount rate, which may most likely be due to variation in the opportunity cost of labor. NPV ha-1 value of maize is 2.8, and 1.5 times higher than the value of millet and teff, respectively in the Lay Armachiho district. It has three-fold and two-third-fold EAI compared to teff and millet, respectively. The greater NPV and EAI ha-1 achieved for maize was due to higher yield production (36.00 ha-1). Nevertheless, teff has the highest national and local market demand than millet and maize, maintaining the cash requirement of households. According to the focus group discussion, teff and millet have the additional benefit of lower loss during storage. Group discussants noted millet is consumed by the growers’ families and therefore realizes their subsistence benefit. It is also resistant to the variability of weather conditions.
The results indicate that NPV, EAI, and BCR of agroforestry practices were higher than cereal crops (Table 6 and Table 7). The coffee shade practice has two, three, and six-fold EAI compared to maize, teff, and millet, respectively, a result consistent with Duguma (2013). Mekonnen et al. (2021) in the upper Blue Bile basin also figured out that NPV ha-1 of A. decurrens cum teff agroforestry practice is superior to monoculture cereals. Bado et al. (2021) in West Africa Sahel reported that the income of farmers increased from 5% to 14% due to agroforestry. The higher financial reward for agroforestry practices could be linked to lower fertilizer and labor inputs, diversified benefits, and improved fertility and productivity of fields (Table A1-A4). Agroforestry have lower fertilizer and labor inputs, diversified benefits, and improved fertility and productivity of fields than cereal crop practices in many areas of the tropics (Ajayi et al., 2009; Nigussie et al., 2020; Bado et al., 2021; Roshinus et al., 2021). Small-scale agroforestry offers an ideal blend of low-cost inputs (Mercer et al., 2014). However, in the lower Mississippi region of the USA, Mercer et al. (2014) found that the returns were lower from agroforestry than agriculture on most marginal lands, demonstrating the severity of land degradation limits the profitability. Generally, the overall NPV and BCR trend of the six investment alternatives shows coffee-shade > homegarden > Dogwood> maize mono-cropping >teffmono-cropping > millet. A better positive return in agroforestry over cereal monoculture practices is preferable to land compensation payments to farmers (MoFED, 1998) and formal rural microfinance institutions.
Table 5 here
Table 6 here
Table 7 here
4.2.2. Scenario II: Land lease system
Table 1 shows the financial profitability of the six investment alternatives invested on leased land considering 15 years time span. Producing cereal crops on lease land is a common practice in Ethiopia and the study area. Land leasing is a system in which the lessee pays money to the leaser (landowner) at the beginning of the agreement to invest an investment on land for a specified period (usually for > 1 year). The lessee produces crops or trees and collects revenues covering all other production. Both agroforestry and cereal monocropping practices have a positive NPV, EAI, and greater than unity BCR at a 10% discount rate (Table 8). However, the financial return of agroforestry practices is better than cereal monoculture practices. The results confirm landless (land-constrained) farmers can engage in long-term investments in leased lands and be profitable. In terms of benefit distribution, the NPV and EAI are higher for lessees than lessors. Agroforestry practices have a better financial incentive than cereal monocultures. Contrary to this finding, Nigussie et al. (2020) in Northwestern Ethiopia reported that, in an equal benefit and fertilizer cost-sharing system, a higher share is distributed for lessors than lessees in the acacia cum teff sharecropping system.
Table 8 here
4.3. Sensitivity analysis
To assess the sensitivity of the financial performance parameters, we used discount rate and major input and output variables. From the input side, labor cost covers 70%, 62%, and 75% of the cost for homegarden, dogwood, and coffee-shade practices. Labor costs are the major decisive factor in adopting agroforestry practices in many studies (Ajayi et al., 2009; Duguma, 2013; Matthies and Karimov, 2014). From the output side, leaves and fruits/seeds are the major outputs covering 72% of revenue for homegarden, 83% for dogwood, and 87% for coffee-shade practices. As a result, we performed a sensitivity analysis by varying the discount rate, labor cost, and revenue from perennial components. We assessed the effects of these changes on the NPV of all three agroforestry and three cereal monoculture practices.
4.3.1. Sensitivity to variation of discount rates
The financial attractiveness of agroforestry practices declined as the discount rates increased (Table 6). At a 20% and 30% discount rate, the NPV of agroforestry practices decreases by 52-56%, and 72-79%, respectively. Although the profitability of agroforestry practices decreased meaningfully, their financial performance is still better than cereal practices up to a 30% discount rate. This result is consistent with the findings of Nigussie et al. (2020) and Rasul and Thapa (2006). However, at a 50% discount rate, the financial profitability of monoculture (maize and teff ha-1) outperforms agroforestry practices. The greater weight given to benefits received in the early years of the growing cycle increases practices' sensitivity to discount rates (Ajayi et al., 2009). In the prevailing poverty and socio-political situations, farmers may attach high value to present income and discount the future income at a much higher rate. Hence, it is important to understand the discount rates used in selecting and deciding investments.
4.3.2. Sensitivity to change in labor cost
In all the investment alternatives, the change in labor cost affects the NPV of farmers’ investments (Table 9). Modest change (10%) in labor cost affects the NPV of homegarden more than dogwoodand coffee-shade practices. For instance, increasing the labor cost of a homegarden by 10% decreases the NPV by 2.35 above the baseline. The finding of Duguma (2013) is consistent with our result that homegarden is labor intensive and changes in the labor cost affect its NPV more than other alternatives. In such labor-intensive practices, households lacking available labor, such as elderly farmers, might not be benefited. However, agricultural crop production, particularly millet, is more sensitive (3.6%-8.3%) to labor cost variation than agroforestry practices (0.6-2.35%). This finding supports the reports of Duguma (2010) that cereal farming in the central highlands of Ethiopia was sensitive to a higher labor cost variation. Rich farmers and labor-endowed farmers gain more profit than the poor. Coffee-shade practice is the lowest sensitive practice to labor change compared to all investment alternatives. This suggests that environmentally sound and financially profitable practices should be promoted while minimizing the costs of production by increasing labor productivity, especially for the poor and less resource-endowed households.
4.3.3. Sensitivity to change in revenues
Financial return (NPV) can also be affected by changing prices, the output generated, or both (Martinelli et al., 2019). Revenue can represent these variables. Thus, the impact of changing revenues on financial returns was conducted at a 10% discount rate.
All three agroforestry practices and cereal mono-crops are sensitive to variation in revenue (Fig 3). At 10% interest, changing the revenue of perennial components by 10% affects the NPV of coffee practice by 10.17% from the baseline, which is greater than the impact on dogwood (9.15%)and homegarden practices (8.2%). However, the effect on the NPV of dogwood washigher than the effect on homegarden and coffee practices if the revenue of perennial components changed by ≥ 20%. For example, at a 10% interest rate, changing the revenue of perennial components by -20% affects the NPV of dogwood by 21.17% and the NPV of coffee-shade by 17.17%. The impact of reasonable change in revenue on profitability confirms the importance of improved farm management and germplasm. Field management, silvicultural practices, and quality germplasm improve the productivity of farm components and total outputs (Matthies and Karimov, 2014). Mercer et al (2014) stated efficient combinations of the annual and woody perennial crops produce a production possibility frontier (PPF) at a given level of input. The complementary relationship of components reduces costs for weeding, and fertilizer while increasing tree and crop production in agroforestry. Net benefits from agroforestry are also enhanced by increasing the value of the outputs, usually improving the market price. Efficient local markets are usually used to construct optimal profitability frontier (Mercer et al.,2014). Hence, the government should provide supports that improve productivity and financial returns of agroforestry practices to farmers such as providing farm management training, quality germplasm, and establishing an efficient market.
Compared to agroforestry practices, cereal crops are more sensitive to changes in revenue.Millet is the mostsensitive investment alternative, almost three-fold to the percentage change of homegarden and dogwoodpractices NPV at -10% revenue variation. The results indicatethat cereal crop practices are not profitable at a very high revenue decline (by 50%) (Fig 3). These indicate significant short-term changes induced by droughts and other events significantly impacting cereal crops' financial returns.The impact of input and output value fluctuations on the net return of cereal farming was also considered in the central highlands of Ethiopia (Duguma et al., 2010). However, Duguma et al. (2010) noted a proportional rise in input and output values increases the net return.
Figure 3 here
4.5. Constraints of agroforestry practices’ Establishment and Profitability
Figure 4 illustrates factors that determined the establishment and profitability of agroforestry practices in the Lay Armachiho district of Northwestern Ethiopia. The fluctuation of prices for products was the most cited constraint in our study. The fluctuation of prices for agroforestry products was marked by ethnic-based conflict between Amhara and Qemant within the district and neighboring districts. The conflict influences the traders' transport access and security by closing roads. Besides, additions of unwanted materials into products meaningfully deteriorate the value of products.
Diseases, the second most cited constraint, were major problems for fruits such as M. indica, C. sinensis, and C. aurantiifolia and improved coffee varieties. White mango scale insect (Aulacaspis tubercularis), and leaf and fruit spot of citrus (Pseudocercospora angolensis) were identified as diseases for M. indica, and citruses, respectively. The third constraint, i.e., lack of improved varieties, identified by farmers was connected with the problems of diseases. The impacts of diseases on products of agroforestry forced farmers to demand improved disease-resistant varieties. Farmers among the constraints for establishing agroforestry practices also mentioned the shortage of land and materials such as water pumps, hoses, seedlings, and seeds. Shortage of land for agroforestry is associated with land use priority for cereal crops.
Our reported constraints, consistent with Matthies and Karimov (2014), revealed that smallholders faced bargaining over price and access to mark constraints in the Amhara region of Ethiopia. Similarly, Molua (2005) in Cameroon identified market prices as the main determinant of agroforestry production. Martinelli et al. (2019) in Brazil also reported unexpected natural risks such as diseases were the factors for the economic performance of agroforestry systems. Nigussie et al. (2016) determined income, soil fertility management, and soil and water conservation were the deciding factors for planting A. decurrens in the Fagita Lekoma district of North-Western Ethiopia. Duguma (2013) in the central highlands of Ethiopia, on the other hand, listed lack of proper planning and knowledge and land shortage as the main constraints for adopting agroforestry land uses. Miller et al. (2017) also stated related constraints for on-farm tree adoption and management, including gender, land, and labour endowments.
Figure 4 here