Mongolia: Anticipation of harsh winter, 2018 -2019 Impact Assessment

Originally published


Mongolia is unique. 30% of the population earn their livelihoods from herding livestock, leading a traditional nomadic life to enable their animals to access pasture throughout the year. Weather conditions can be extremely harsh, with long, stormy winters where temperatures can drop as low as -40C or -50C. Herders, and therefore the wider Mongolia economy, are vulnerable to the impacts of extreme winter weather. Dry summers can mean livestock do not gain the weight needed to survive winter, combined with harsh conditions this can cause mass livestock mortality known as 'dzud'.

Harsh winters pose a serious risk to livelihoods and herder wellbeing. They can also be predicted, making it possible to act early to reduce harm and loss. In December 2018, using a dzud risk map created by the Mongolian National Agency of Meteorology and Environmental Monitoring, the Start Fund allocated £200,000 in anticipation of harsh winter. World Vision and Save the Children intervened in seven districts in the West of Mongolia. They provided vulnerable herders with unconditional cash, hay and fodder for livestock and a booklet on protecting children and family members.

Research questions were developed to explore each of these links in turn, to understand attribution and the different types of impact the project had.

  1. Can Start members access dzud forecasting information of sufficient quality? Was the prediction correct in terms of geography, groups effected, intensity?
  2. Was the anticipation alert earlier than other harsh winter related interventions? Was the intervention considered timely?
  3. Did beneficiary communities prepare for and cope with the impacts of dzud differently to non-beneficiary communities? If so, how? Did beneficiary communities experience different types or severity of impact from the winter conditions? If so, how?
  4. Can we demonstrate that the cost of the cash transfer and fodder is less than the loss experienced by non-beneficiaries, through increased interest on loans, lost livestock & lost newborns, and diminished livestock condition?

Sampling and data sources

Different data sources were identified for each question. A household survey was carried out reaching 95 beneficiary households and 93 control households, across seven of the fifteen soums (districts) in the four intervention aimags. Households were intended to be sampled according to the same criteria as the initial distribution, herding their own livestock and owning fewer than 200 sheep head unit (SHU each). During implementation, the lack of availability of herders resulted in 27 families with herds larger than 200 SHU being sampled, of whom 21 were in the control group. Their data has been removed for questions where statistical tests have been conducted, including all of those quantifying livestock.

In addition to owning less than 200 sheep head unit, we ensured that families with infants, disabled members, children in higher education, pregnant and lactating women and senior citizens were represented across the sample. Local Government officials supported the identification of relevant respondents. The research would otherwise not be possible, where families within one soum can be hours apart and do not have fixed addresses.