Households in the hanging in livelihood pathway earn a living from traditional pastoralism and small agro-pastoralism. The highest percentage of their earnings comes from livestock, making other sources negligible. A striking trend of this group was their continued decrease in numbers over time. In 2009, 27.81% of households were categorized as hanging in. By 2015, that figure had fallen to 11.53%. According to Lind et al. (2020), this group of pastoralists is common in areas where there is good access to rangeland and water sources, but market access is limited. People in such areas are hanging in customary pastoralism, and their incomes drive demand for other non-pastoral products or services such as crop produce, natural products, constructions services, and others (Headey et al. 2012).
The dropping out pathway represents households with the lowest herd size and cash incomes. It was the largest livelihood group that cumulatively represented 28.96% of the panel respondents. In the base year, 28.35% of households were categorized into this group. The proportion of households grew over the years, implying that relatively more households became dropouts over time. This group’s striking feature was its drastic rise from 22.40% in 2010 to 39.07% in 2011, a trend attributable to the devastating 2010/2011 drought that began in late 2010 and peaked in 2011. Drought not only deteriorates livestock health or results in high mortality, but it also pushes pastoralists out of their production system, forcing them to seek alternative sources of livelihood. Therefore, the increasing number of households dropping out depicts the increasing frequency and the severity of the drought. For instance, the 2010/2011 drought was thought to be the worst in the last 60 years, but the subsequent ones were even more severe, with 2017 having been the worst of all. Such a trend threatens pastoralism as a livelihood. As evidenced by the results of this study, the increasing number of households dropping out and decreasing number of households hanging in on pastoralism is an indication of a general exit from pastoralism over time.
Households in the moving up group represent the wealthiest in terms of both herd size and cash income. Also, from the income profiles of these households, much of their income comes from the sale of livestock and livestock products. Over time, the number of households categorized as being in this pathway decreased to 24.37% of households from the initial proportion of 27.81%. However, the rate of decrease over time is slower with patterns of increase, followed by a decrease, for instance, from 36.69% in 2010 to 23.59% in 2011 and then up to 28.35% in 2012. In northern Kenya, moving up activities are evident in Moyale where pastoralists are engaging in cross-border trade and other high-value livestock marketing activities with their counterparts from Ethiopia’s eastern lowlands (Mahmoud 2013).
Moving out is the second largest group among the four livelihood pathways. Furthermore, it is also the second group (after dropping out) that depicts a trend of growth in numbers over time. At the starting period (2009), 16.02% of the sampled households were categorized into this group. Although there were instances of a slight decrease in some periods, the group generally registered an upward trend reaching 31.66% in the final year. Also, households in this group had high incomes with most of it coming from sources not directly linked to pastoralism such as salaried employment, sale of crops, business, and petty trading. Improved connections with large centres, small-town expansion, and acceptance of non-traditional livelihoods by the younger generation are some of the factors that promote household engagement in moving-out activities (Lind et al. 2020). The general trends observed from the proportions of these four livelihood pathways are illustrated in Fig. 3.
Socio-economic characteristics of pastoralist households by livelihood pathways
A number of socio-economic characteristics were associated with the four pastoralist livelihood pathways. Table 3 provides the descriptive summary statistics for continuous household characteristics by livelihood pathways. The F-statistic values included in the table are the analysis of variance (ANOVA) tests for significant differences in the means of selected variables between the four livelihood pathways. As a precaution against extreme values including many zero-valued observations in some variables such as household head’s education years, household education stock, amount of savings, remittances, and size of land irrigated, the study used inverse hyperbolic sine transformation (IHS) as used by Bellemare and Wichman (2020). This procedure, to some extent, is the same as taking the log of the variable, but it allows the retention of observations with zero values.
As provided in the summary statistics in Table 3, the mean age of household heads varied significantly across the four livelihood pathways. The dropping out group had the highest mean age while moving up had the lowest, reflecting participation in market-oriented keeping and moving of large herds among young pastoralists. Dropping out had the highest mean of household size, while hanging in had the smallest. These results indicate that a relatively older head leads an average drop-out household with many household members who increase pressure on an already small income and fewer animals due to severe droughts and other factors.
Schooling years of household head and household education stock (sum of schooling years over all household members) showed a significant difference between the four livelihood pathways. Moving up had the highest mean of schooling years of the household head while dropping out had the lowest average years of schooling. On the other hand, moving out had the highest average household education stock, while hanging in had the lowest average. This highlights a vulnerability among drop-outs concerning access to or their search for education. According to Teshome and Bayissa (2014), dropping out of pastoralism is mainly due to unexpected events such as drought, conflict, and chronic poverty, and households that drop out often end up in old peri-urban impoverished areas where access to education and other services is a challenge. Such households’ activities are generally inferior survival strategies that yield low income that might not be sufficient for them to seek a better social status and education (Little et al. 2010). On the other hand, higher education years in the moving up and out pathways imply more cash income opportunities and greater access to education among those groups. Generally, moving up and out livelihood activities are common in urban areas where better living conditions, job opportunities, and access to education are among the pull factors that attract households to such areas (Lind et al. 2016).
The amount of savings significantly differed across the four livelihood pathways. The moving up group had the highest mean of household savings, while the dropping out group had the lowest mean. This finding highlights the vulnerability faced by households in the dropping out group. Should they face an income shortfall, these households have fewer resources to fall back on. Few savings are likely a result, in part, of their low income and could also be a signal of less access to savings institutions.
Moving out had the highest mean of irrigated land while hanging in had the lowest mean. Moving out involves participation in value-added activities such as micro-dairying, small-holder crop farming, and other activities; hence, irrigation projects play a very vital role for households in this category, especially when it comes to the growing of fodder, watering crops, and other high income-generating activities that may not do well in some of the world’s harshest lands that hardly receive adequate rainfall throughout the year.
To further show the difference in socio-economic characteristics by livelihood pathways, Table 4 provides the summary statistics for discrete and dummy variables.
The hanging in group had the highest percentage of households headed by females while moving out had the lowest. The highest percentage of female-headed households among traditional mobile pastoralists reflects a common practice of sending male heads to satellite camps while females take care of the basecamp (Jensen et al. 2017).
Settlement status varied significantly across the four livelihood pathways. Households practising traditional pastoralism (hanging in) had the highest mobility, while those in the moving out group had the least. These results affirm that the traditional adaptability and risk management strategy of mobility still plays a significant role among households that depend on livestock as their primary livelihood. This finding is consistent with the observation of McPeak et al. (2012) regarding livestock mobility among pastoralists that it must be supported and not hindered by policy or other changes.
Access to credit provides financial capital to venture into both livestock-based livelihood activities and those that are non-livestock-based. The majority of households that accessed credit were in the moving up group, with those dropping out having the lowest access rate to credit. Arguably, credit provides financial capital that increases the likelihood of household engagement in highly remunerative strategies such as moving up livelihood activities. Indeed, as Umeta and Temesgen (2013) note, helping pastoralists get credit enables them to participate in lucrative livestock and even non-livestock livelihood ventures.
Index-based livestock insurance is an intervention that aims to help pastoralists manage drought-related shocks. Moving out and up livelihood pathways had the highest insurance uptake, while the hanging in and dropping out groups had the lowest. This illustrates the idea that buying insurance enables pastoralists to build herds for trade, and any losses likely to be incurred in the event of drought will be catered for by indemnity received once the index reaches strike points set for payments to be released to all insured clients. Matsuda et al. (2019) also observed that insurance could increase herd size.
Land ownership varied significantly across the livelihood pathways. The findings of this study indicate that the moving out group had the highest percentage of land ownership while the hanging in group had the lowest. Ideally, traditional mobile pastoralism (hanging in) requires extensive and free access to commonly managed pastures. Individual land ownership is thus not expected in such groups. On the other hand, having higher productive physical assets holdings is necessary for households transitioning into or engaging in more remunerative market-oriented livelihood activities such as moving out.
Pastoralist livelihood pathway transitions
In the context of change over time, it is possible to estimate the probability that a household categorized under a given livelihood pathway in the previous period (t−1) is likely to remain in the same group or transition to another category in the current period (t). Table 5 presents the computed proportionate transitions between the four livelihood pathways. Rows show the initial values at t−1, while the columns show the final values at time t.
The proportionate transitions between pastoralist livelihood pathways results showed that the transitioning of households from one pathway to another was very dynamic. This is interesting because targeting households in one pathway does not make sense since they would transition anyway. Furthermore, apart from the clear exit trend from pastoralism (hanging in) over time, no clear pathways or sets of transitions seem much more common than others.
Three main trends were observed from these livelihood transitions. First, dropping out, moving up, and moving out are all very resilient in that once a household falls into one, it is unlikely to leave. For moving up and moving out, this is great, but for dropping out, the stability indicates something like a poverty trap, which once a household is in it, it is difficult to exit. Secondly, it is tough/uncommon to transition into the hanging in group, as shown by the very low values in the 2–4 rows of the first column. This could reflect that it requires a great deal of capital or maybe because people do not want to, but still interesting that this is the least likely livelihood category to transition into. Thirdly, it seems there was a lot of back and forth between dropping out and moving out and that moving out is the only way out of dropping out (poverty). Perhaps, this suggests that livelihood interventions aimed at alleviating poverty among pastoral households should support activities that are important to helping people move from dropping out to moving out and infrastructure that allows those moving out to avoid falling into the poverty trap by dropping out.
Effects of drought on pastoralist livelihood pathways
Exposure to covariate drought shock has an impact on livelihood activities among pastoralists. The vegetation index best captures the severity of drought. Table 6 provides the effect of drought on pastoralist livelihood pathways unconditionally, controlling for previous livelihood pathways and controlling for other variables in Tables 3 and 4. The interest in the analysis was to determine the predicted change in the likelihood of each outcome (livelihood pathways) as CZNDVI changed from bad (drought) to good (rainy) season. Estimates of each outcome were computed with CZNDVI at 10% and 90% levels, holding other covariates at their means, the idea being that, keeping everything else at the mean, the difference between 10 and 90% is the estimated impact of CZNDVI alone.
As CZNDVI increased from a drought season where CZNDVI was at a 10% level to a good season where CZNDVI was at 90% level, the probability of households dropping out decreased from a predicted probability of 37.9% to a probability of 28.7% and the likelihood of moving up and moving out increased from 22.2 to 25.0% and 22.6 to 34.3%, respectively. When controlled for the previous pathway, an increase in CZNDVI decreased the probability of dropping out from 47.0% to a probability of 25.6% and increased the likelihood of moving up and moving out from 11.7 to 25.4% and 30.7 to 35.5%, respectively. Controlling for other covariates in Tables 3 and 4, an increase in CZNDVI reduced the probability of dropping out from 47.2 to 24.2%, increased the likelihood of moving up from 11.3 to 27.2%, and increased the likelihood of moving out from 31.5 to 34.9% relative to the likelihood of hanging in (the reference group) holding other factors constant.
Overall, the results revealed that forage scarcity or availability, as indicated by the changes in CZNDVI, is a key driver of pastoralist livelihood pathways. These results imply that frequent droughts leading to a devastating loss of herds drive households into being stockless; hence, they drop out and cannot participate in livestock-based lucrative ventures such as the sale of animals and even value-added livelihood diversification activities. The impact of drought is significant across all pastoralist livelihood pathways; hence, all households are vulnerable to drought over time, including those that have moved up and out of traditional pastoralism. Drought still pushes them into a destitute outcome (dropping out). This suggests that while helping households transition into market-oriented and relatively successful groups such as moving out reduces the impact of drought to some extent, there might be a need to support households in those groups as well to protect them from falling back to poverty by dropping out. Different measures that help build resilience and reduce vulnerability to drought or other shocks might be helpful to non-herders too.