Development evaluation of Lesotho agricultural input subsidy policy based on rural households ’ food security and access to inputs : Evidence from Mohale ' s Hoek District Thapelo

The prevalence of poverty in Lesotho is a pressing issues raising concern in terms of rural households’ food security and welfare. In that regard, this study attempts to establish or identify the determinants of farmers’ access to inputs via the government. The second part incorporates the use of subsidized inputs and measure the impact of the government input subsidy program on food security. The motive behind this study is to fill a gap in the economics literature in this area, which often does not assess the impacts of agricultural input subsidy programs (despite their widespread prevalence) and its impact on food security. The Combination of econometric approaches was applied to a selected sample of rural households actively engaged in maize production, and the findings revealed the determinants of farmers’ access to inputs as being cropland allocation, type of fertilizer and ecological zones. The vast majority of rural households doesn’t receive or get subsidized inputs and out of that small portion of the recipients, half of them don’t use their inputs (chemical fertilizers). This phenomenon is due to the fact that, rural households may not be in a position to use their inputs because of late delivery, or to use optimally because they do not perceive the benefits.


INTRODUCTION
Smallholder farmers and resource-constrained households (HHs) in Lesotho have been consistently obtaining low yields from their staple crops, and that adversely affected food production as well as considerably reducing incomes.In an attempt to mitigate the situation by improving the welfare of rural households and spurring crop production, the government of Lesotho (GoL) endorsed universal input subsidy policy (UISP) on *Corresponding author.Email: molatolithapelo@gmail.com(0086) 18771134747.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License agricultural inputs from 2001/02 after food-security emergency was declared due to poor agricultural production.Based on the literature, agricultural input subsidy policy (AISP) is not yet an ultimate solution; prevailing government policies, which involve the subsidized provision of tractor services, seed, and fertilizer, seem to increase unit costs for smallholder farmers due to low market price for grains (Schwab and Porter, 2009).
In addition, Mohlatsane et al. (2009) also, discovered that, even though the government provides subsidies, subsistence farmers still do not afford the input costs as commercial banks find it risky to give them credit.There are also views that, targeting input subsidies to the poor and smallholder farmers is potentially more efficient than universal as targeted subsidies are directed to different needs of farmers instead of universal subsidies which might not be a response to or address needs of farmers (Mohlatsane et al., 2009;Houssou and Manfred, 2011).Dorward (2009) also states that whether targeted or universal; subsidies are only effective where there are limited secondary markets in which recipients sell subsidized inputs to non-recipients.Dorward and Chirwa (2013) further reiterate that, targeted farm input subsidy irrespective of considerable improvements, there are still ongoing challenges regarding implementation, improving targeted outcomes and impacts.Huang et al. (2011) findings revealed that, about two-thirds of households who were subsidy recipients by the year 2008 believed that there is no link to the amount of their produce or cultivated land.While in Iran, food subsidy programs emerged as major developmental issues since consumers benefit from lower food prices (Karami et al., 2012).AISP contributes to farms efficiency and profitability; hence, there is a need for broader evaluations of the role of public support for various components of farm's performance (Bojnec and Latruffe, 2012).
There are different ways to evaluate input subsidy policy; taking access to input subsidies as the evaluation point, the findings reveal vulnerable and resourceconstrained farmers being marginalized compared to well-off farmers in possession of bigger land size and have market accessibility or more income (Chirwa et al., 2011;Jayne et al., 2011).Only findings from Nigeria appear completely different from other studies in Africa due to the fact that Nigeria has successfully linked rural farmers with input suppliers (Liverpool, 2012).Political elites benefit more than less politically-connected households (Banful, 2011;Mason and Ricker, 2012) and the extension agents are biased toward better-producing households (Pan and Christiaensen, 2012).
GOL has been and continues to offer subsidized inputs to farmers but, this initiative and efforts by the government fail to achieve significant increases in national food output.Since 2012, 18 500 vulnerable farmers in Lesotho have received agricultural inputs through the Ministry of Agriculture and Food Security / Food and Agricultural Organization (MAFS/FAO) Emergency and Resilience Programmed.These inputs included a 5-kg bag of maize seeds and a 5-kg bag of bean seeds, Crop estimates were indicating that the domestic production will contribute less than 10% of the annual national cereal requirements for 2012/13 and that the Lesotho rural populations are in dire need of food (Lesotho flash Appeal, 2012;Lesotho Food Insecurity Situation Report, 2012).Prior to that, Lesotho Demographic and Health Survey (2009) showed chronic malnutrition (stunting) amounting to 39, 2% of the total population.Later on in two years' time, the percentage of the population living on less than $1.25 per day was 47.59% (Lesotho flash Appeal, 2012).
Owing to the prevailing situation in Lesotho aforementioned, the study attempts to establish the determinants of farmers' access to inputs via the government and other sources.A number of studies about an evaluation of agricultural input subsidy policy have been conducted in various countries across Africa; Ghana, Malawi, Nigeria, Tanzania and Zambia.However, Chirwa et al. (2011) and Jayne et al. (2011) were using access to input subsidies as the evaluation point.Therefore, in order to fill a gap in the economics literature in this area, which often does not assess the impacts of agricultural input subsidy programs (despite their widespread prevalence) and its impact on food security, the study seeks to measure or analyze the impact of the government input subsidy program as compared with other input subsidy programs on food security.That is, who actually benefit from input subsidy?And how subsidies can benefit rural households without benefiting farm inputs retailers?

CONCEPTUAL FRAMEWORK
Input subsidies from government stores either go directly to farmers or through seed retailers.Rural households (RHH) could be divided or separated into two groups; needy and less-needy based on their income, food security, welfare and whether they hire labor in or out, crop purchases or sales.Besides government input subsidy program, there are other input subsidy programs offered by non-governmental organizations like WFP, UNICEF and World Vision.RHH receiving social grants are regarded as needy or vulnerable and in most cases perceived to be eligible for being recipients of subsidized inputs.However, it may depend on different scenarios and perspectives as Dorward et al., (2010) regard vulnerable households as those that are female-headed, elderly headed and child-headed households.Input subsidy beneficiaries may use subsidies accordingly and increase their production while others might resale them or use them alternatively, which is, displacement use as depicted in Figure 1.
Targeting, implementation and geographic distribution  Amounts of inputs provided depend on certain factors, for instance, in 2006/2007 the government subsidized chemical fertilizer, these fertilizers were supplied directly to farmers who bought a minimum of 200 x 50 kg bags as an incentive to encourage efficiency in terms of economies of scale.Those who bought less than 200 x 50kg were not offered transport, while others bought indirectly through traders who sold them at a price determined by MAFS.Alternatively, for better implementation of input subsidy, during the same period 2006/2007 growing season, block farming scheme which is a system that encourages planting one crop on a large scale in one locality was launched.This was done as a response to the Government policy of commercialization of agriculture, to assist block farmers in accessing credit from one of the local banks facilitated by the Agricultural Development Fund (government-guaranteed loans) aimed at providing capital to farmers for increasing the country's food security.The year 2007/2008 farmers were to obtain the input supplies and farming operations at 70% of the selling price and 40% on irrigation equipment.Recently, the government of Lesotho (GoL) has engaged in sharecropping program where the government provided all farming inputs and adopted a mechanized approach that necessitated the consolidation of blocks of farmers' fields and yields are divided between government and farmers, where GoL takes 70% of the yield and farmers take 30%.

Study site
The study was undertaken in Mohale's Hoek, which is one of the 10th districts of Lesotho with the land size of 3,530 km 2 and total Taking the analysis of cereal production, consumption and surplus / deficit over a period of four years as portrayed in Figure 2, it's clearly notable that Mohale's Hoek is heading towards severe food shortage once experienced in 2011 / 2012.Considering the state of food security in Lesotho as a whole, statistics reveal the increasing number of food insecure population (Figure 3).

Data collection
Data collection was conducted in Mohale's Hoek district from July up to October 2014 randomly picking one constituency from each Agro-ecological Zones.The survey design adopted stratified sampling procedure structured implicitly into recipients / nonrecipients dichotomy and four constituencies representing four Agro-ecological Zones.The focus was based on maize production as the staple food of Basotho people, hence regarded more important in terms of rural households' food security and welfare.
Farm visits and open-ended questionnaires employed were specifically targeted to individual input subsidy beneficiaries as well as key informants.Owing to the terrain of Lesotho and time constraint, it was not possible to include all constituencies during the survey; therefore, in this case, the researcher opted to pick one constituency from each zone in order to have a representative of all agro-ecological zones in the data set.Finally, four constituencies namely; Mpharane, Mekaling, Mohale's Hoek and Taung respectively were covered during the survey.

METHODOLOGY
Rural households' welfare is analyzed based on their production performance, but in this case emphasis is on maize production as the staple food for Basotho people.Therefore, welfare was determined by changes in farmers' maize production also regarded as farmer' income, which can be written as  (Matlon, 1987;Wong et al., 1991).Given the low levels of fertilizer use in Lesotho and the demonstrated contribution of fertilizers to increasing crop yields and land productivity, the increased use of fertilizers has great potential for boosting food production (Wilfred: 1996), and overcome a problem of food insecurity prevalent in rural households of Lesotho.The effectiveness of ISP is measured by the impact of the fertilizers, hybrid seeds and machinery subsidy (D) on rural households (RHH) maize produce based on the quantity of hybrid seeds planted and fertilizers applied per hectare (Q); compared to the RHH produce, and total income.All these can be summarized as: Lesotho mostly use universal input subsidies but assuming trade-off by some of the recipients, commercial farmers would turn to benefit more and they are expected to increase their production (food supply) which ultimately could result in a drop of imports from South Africa, and consequently a decline in local market price giving rise to poverty alleviation and of course an improvement in food security and self sufficiency.Continuous dependency on imports from South Africa means failure of Lesotho farmers to alleviate poverty irrespective of input subsidies and this phenomenon could imply number of things like poor distribution channels leading to leakage of subsidized inputs to South African farms, or political manipulations meant to tarnish government reputation or put government into disrepute in order to score political points or gain political mileage (Figure 4).

Econometric model
The model encompasses two ways of accessing subsidized inputs in the quest for answers to the questions raised; that is, the household is a beneficiary of input subsidies whether received vouchers from FAO or inputs from licensed seed retailers.
Where; for household i, ASI is access to subsidized inputs, HS is a vector of household situation including living conditions and durable assets, FP is a vector of farm properties like land size, cultivated land, and type of farmer, while PV is a vector of poverty and vulnerability variables, X is a vector of other control variables including participation in the labor market (employment), business enterprises, remittances and εi includes the effect of unobserved factors on the access of individual i as well as the impact of measurement errors (error term).Access to subsidized inputs is a dichotomous variable being the beneficiary of input subsidies; equation (3) was estimated using the Logit model (linear probability model).

Descriptive statistics
In an overall sampling of rural households actively     fertilizer because of late delivery, or to use optimally, because they do not perceive the benefits (Druilhe and Barreiro-Hurlé, 2012).
The mean amounts of maize production and mean values of outcome variables are presented in Table 3 by subsidy recipient and maize production.Input subsidy beneficiaries seem to have high maize production on average than non-beneficiaries, for instance, subsidy recipients cultivated more land compared to nonrecipients with the difference of 2.17 hectares in the year 2014 and 2.57 hectares in 2013.Similarly in terms of surplus or deficit and self-sufficiency, non-recipients produced an average of 1,457 metric ton of maize with a shortfall of 1,935.6 metric tons compared to the recipients and also on the side of self-sufficiency results revealed that non-recipients production can only sustain them for a period of 3 years, which is 2.2 years less compared to the recipients.
Lesotho government effort to nurture farmers' production through input subsidy program is heavily criticized and viewed as myopic since farmers were not initially engaged in probing their needs and challenges which they are faced with.Late delivery of subsidized inputs and lack of transparency in allocation of subsidized inputs contribute a lot towards food insecurity, reducing performance both for poverty alleviation and improved production objectives.In some other constituencies, there were clear instances of fraud and political manipulations hence in general, only 44.7% of farmers recognized ISP as a right policy while the rest criticize it for lack of transparency viewing it as myopic and also expressed concern over late delivery of input subsidies.Figure 5 shows more clarification concerning farmers' views and perceptions about input subsidy policy (ISP).
Based on the results (Figure 5), rural households may not be in a position to use their inputs due to late delivery, or to use optimally because they do not perceive the benefits.Input subsidy programs are rendered ineffective by the government universal approach which gives rise to unfair treatment as some regions get more advantage over other regions in terms of infrastructure and market structure.In addition, differences in geographical region play a major part.Farmers in the mountain areas and foothills are less likely to access input subsidies than those in the lowlands and Senqu river valley due to differences in geographical zones of Lesotho and poor infrastructure.This phenomenon is one of the contributing factors to poor maize production because plowing season varies across the regions, maize plantation in Mountainous Region is from September to October, and Foothills is from September to November, Lowlands and Senqu River Valley is October to November.

Results for ISP participation
Moreover, Senqu River Valley experiences early frost.The results (Table 5) can also be interpreted in terms of odds ratios, where farmers allocating more land for maize production are 1.2 times more likely to access input subsidies.This coincides with the fact that, most of the fields (land) belonging to the poor rural households are left fallowing for a long period of time since they don't afford inputs cost even at a subsidized price unless they can be in share cropping.It is noticed that farmers using chemical fertilizers are 3 times more likely to access input subsidies, which is not always the case in rural areas where most households don't even know the state of their soil fertility or the appropriate chemical fertilizers for their fields because no soil analysis test have ever been conducted.In the nutshell, resource-constrained households are marginalized as also noted by Chirwa and Dorward (2013) and Mvula et al. (2011) that, the vast majority of vulnerable and poor rural households are neglected.Farmers in remote areas are discriminated (extension agents are not willing to stay in those areas) mainly due to poor roads infrastructures (distance travelled and difficulties in transportation), therefore collecting subsidized inputs cost rural households relatively same amounts of money they would spend if buying inputs from private seed retailers located in their vicinity (Chirwa et al., 2011 andJayne et al., 2011).

Regression results for factors influencing rural households maize production
In order to get rid of outliers from the data set, outlier labeling rule was utilized with 2.2 multiplier effect instead of 1.5 suggested by Turkey (1977) as Hoanglin et al. (1986) discovered that 1.5 often identifies outliers that are not actually outliers and, therefore, proposed or advised people to use the value of 2.2 provided sample sizes were not huge and distribution was normal.
Based on multiple regression results in Table 6, land size and source of income show a strong statistical significance.On an average, a unit increase in the size of land leads to an increase in maize production by 0.18 metric tons.These findings suggest that the size of the landholding affects maize yield as a function of the scale of maize area (economies of scale).The source of income as a means of capital for farmers appears to have a greater impact with the magnitude of 0.32 which is far beyond the magnitude of landholding, this affirms Johnston (1996) findings viewing landlessness as not a perfect proxy for poverty, but rather, access to nonagricultural income being an important indicator of the economic situation of the household because such an income crucially influences a household's ability to make use of the agricultural assets that it holds.
Off-farm income improves households' welfare according to the following three outcomes; source of income, increase maize production and hence household's total income also increases.The inverse relationship between household's economic status and their production performance reveals relative deprivation of farmers compared to their counterparts or reference group, a unit change in farmers' economic status result in 0.17 metric tons decline in their produce confirming that less wealthy households depending on farm produce as their income were most disadvantaged.The negative magnitude for a type of farmer showed that subsistence farmers were losing 0.31metric tons on average compared to commercial farmers and that increased severity of poverty and relative deprivation among subsistence farmers.

CONCLUSION
The government of Lesotho (GoL) has regularly resorted to input subsidies as a way of increasing food production, notably through the block farming program and its predecessors, however, only a few minority access such inputs while displacement use is also more prevalent (Pearson chi -square results presented in Table 2).Proper use of inputs significantly increase farmers' production (t -test results), therefore, GoL has to put in place the follow-up strategies in ensuring use of inputs accordingly.Additionally, access to non-agricultural income is an important indicator of the economic situation of the household as such an income crucially influences a household's ability to make use of the agricultural assets that it holds (depicted from multi-linear regression results), therefore, GoL should put more focus on job creation.The determinants of farmers' access to inputs include cropland allocation, type of fertilizer and ecological zones.The vast majority of rural households doesn't receive or get subsidized inputs and out of that small portion of the recipients, half of them don't use their inputs (chemical fertilizers).This phenomenon is due to the fact that, rural households may not be in a position to use their inputs because of late delivery, or to use optimally because they do not perceive the benefits.GoL universal approach gives rise to unfair treatment as some regions get more advantaged over other regions in terms of agro -ecological economic conditions and rural livelihoods conditions, infrastructure and market structure.Lesotho is not exceptional from other South African countries like Ghana, Malawi, Tanzania and Zambia where similar studies have been conducted.Vulnerable and resource-constrained households are still being marginalized compared to well-off households in possession of bigger land size and have market accessibility or more income.

RECOMMENDATION
GoL has to make use of all geographical regions, improve infrastructure and open market for farmers' produce.
Farmers should be involved in the probe regarding their needs and challenges which they are faced with; top down approach doesn't address farmers' problems.Lesotho ISP has potential to change and improve rural households' food security and welfare, but there are a number of issues which need to be addressed; decentralizing targeting to local authorities does not improve targeting efficiency, as compliance with program guidelines is not assured and local elites capture most of the benefits, reducing targeting performance both for poverty alleviation and improved production objectives.Therefore, information transfer between rural households and extension agents should be revised, and lastly but not least, political disparities should be downplayed in order to bring unity among local people.

Figure 1 .
Figure 1.Conceptual framework for investigating rural households' food security, welfare and their access to input subsidies.(Source: Author)

Figure 2 .
Figure 2. Cereal production and consumption pattern over a period of four years (Source: Statistical Report No: 26 of 2011 / 2012 & No.31 of 2013 / 2014 NOTE: The deficit/surplus is only for farming households not for the whole population).
population 174,924 (Bureau of Statistics, 2012) located in the southern part.The district encompasses four agro-ecological zones of Lesotho namely, lowlands, foothills, mountains and the Senqu River valley.In terms of the level of food insecurity at the height of the lean season October to December, Mohale's Hoek compared to other districts came first with 62% share of population in dire need of food assistance, indicating Southern Lowlands and the Senqu River Valley as the most food insecure areas (FAO/WFP : 2007).
[ = (  −   )] and real income [  ∑   = (  −  )  ∑   ].Maize output (Y) is a function of a number of fertilizers, hybrid seed planted, machinery and other factors, x.While a number of fertilizers applied, hybrid seeds planted and machinery utilized is a function of the subsidy and other factors, z.In Lesotho, farmers are extremely land -scarce hence it's vital to shift from low -input systems characterized by relatively low land productivity.High-input systems would restore fertility via fertilizer application Zhong et al. (2013) estimated a similar model, Chirwa et al. (2011) also focused on access to fertilizer coupons irrespective of whether recipients utilized them or not to purchase subsidized fertilizers, while SOAS (2008) model focuses on those households that have utilized their coupon to purchase fertilizers.The second part incorporates the use of subsidized inputs and access in terms of location of resource centers.Access to input subsidy is measured based on the Chirwa et al. (2011) index:

Figure 4 .
Figure 4. Research model illustrating Lesotho agricultural input subsidies distribution (Source: Author)

Figure
Figure 4: Research

Table 1 .
Households' age group and gender.
It's also noted there were more female farmers (59.1%) than males as shown in Table1.The percentages of rural households receiving subsidized inputs are summarized in (Table2).Based on the findings, 62.1% of households do not receive subsidized inputs; showing that decentralizing targeting to local authorities does not improve targeting efficiency.Compliance with program guidelines is not assured and local elites capture most of the benefits, reducing targeting performance both for poverty alleviation and improved production objectives.Only a small privileged group of individuals amounting to 37.9% receive subsidized inputs, but surprisingly half of the subsidy beneficiaries still use organic fertilizers instead of chemical fertilizers.This phenomenon is due to the fact that, households may not be in a position to use

Table 2 .
Percent distribution of households using chemical fertilizers by subsidy recipient

Table 3 .
Comparing means of outcome variables, recipients & non-recipients of subsidy.

Table 4 .
Variables used in regression analysis.

Table 5 .
Binary logistic model coefficients and odds ratios.
Xu et al. (2009)(2012)tive statistics for variables used in the regression models.Based on the research findings presented on Table5, cropland allocation, type of fertilizer and ecological zones are the determinants of farmers' access to inputs.Meaning that, vulnerable rural households are marginalized though they were regarded as more eligible for being input subsidy beneficiaries.For instance, households which afford cultivation costs (cultivate more land) and utilize chemical fertilizers are more likely to access input subsidies than the disadvantaged households.This is coherent withChibwana et al. (2012)regarding cropland allocation effects in Malawi, Ricker-Gilbert et al. (2011) andXu et al. (2009)concerning input subsidies in Malawi and Zambia.