Weather data pilots struggle to survive

Most weather-index insurance pilot projects in Africa fail within three years because of inadequate budgets that don’t allow for scale, weather-station maintenance or training, according to Bernard Pacher, CEO of ADCON Telemetry.

Too many pilot projects focus already limited budgets on the purchase of equipment and its operation for two to three years, instead of taking a longer-term view, he told the Fin4Ag conference in Nairobi.

Budget constraints also mean they are unable to reach a scale that is big enough to meet insurers’ requirements or to achieve economies of scale, he says. They also face difficulties commercialising their data.

A typical project may, for example, dot five weather stations in the climatically most risky areas of a country. An insurer will be unwilling in this case to extend insurance based on the data they collect, because the sample is too small and may trigger unjustified payouts in lower-risk areas. Also, regular site visits are unsustainably expensive if stations are positioned 1,000 kilometres apart, Pacher notes.

To succeed, pilot projects need to name a project champion, who bears the risk of the pilot and therefore has motivated to make it work. They also need a robust IT environment and to be able to build a large network of standardised weather stations that produce reliable, comparable data. It is equally important to invest in regular staff training and ensure that extension officers have the transport and fuel to visit stations and “solve the problem of the last mile,” he says.

There are success stories, although most still face challenges. Syngenta Foundation has 110 weather stations in Kenya and aims to add another 200-plus. The foundation had the budget and foresight to scale up immediately, and to locate its stations only around 20 kilometres apart so that data is dense and a number of station visits can be made in any one day, Pacher says.

One obstacle that Syngenta faces in terms of expanding its coverage in Kenya, both in terms of weather data and insurance, is that it is not allowed to ‘compete’ with the Kenya Meteorological Department (KMD) as a data provider. Proposals by Syngenta under which it would add its 110 stations to KMD’s 18 if, in return, KMD paid for their operation have also failed to meet with KMD’s approval, despite this being a “win-win solution,” he says. And other options, under which Syngenta could team up with entities such as the Rockefeller Foundation, could also put it in conflict with KMD, he says.

Zimbabwe’s EcoFARMER meanwhile makes intelligent use of mobile phone technology to generate revenue that can be used to invest in and maintain weather stations. Under the scheme, farmers can subscribe to a SIM card that costs only slightly more than a standard plan but gives them access to a stream of weather data.

A mobile phone company that was able to copy the EcoFARMER model could, for example, be able to generate $150,000 a month to finance and maintain weather stations by charging 300,000 farmers just 50 cents per month, Pacher notes.

Another interesting project is a joint venture of Mexicos’ Instituto Nacional de Investigaciones Forestales Agricoles 7 Pecurarias (INIFAP)/Secretariat of Agriculture, Livestock, Rural Development, Fisheries and Food (SAGARPA) and Cordinadora Nacional de las Fundaciones Produce (COFUPRO), which has around 1,000 stations throughout the country.

Although its maintenance and calibration budget is very small – around $100,000-$150,000 for 1,000 stations, compared with around $1 million for 130 stations under a comparable project in Oklahoma – its decision to make the data it produces free and available to anyone on the internet has been very useful to small producers and the move may eventually pay off.

“No smallholder farmer has the funds to pay for weather data,” he notes. Only when a data provider has “established a culture of credibility” and been able to show a farmer that data is benefiting him over three to five years, will that farmer pay for it.