What’s in a number? Divergent outreach reporting

Microinsurance is a new sector and understanding its extent of reach and overall performance through the lens of data and numbers is crucial for its development. For the sector to develop, stakeholders need data to understand reach, evaluate performance and design appropriate products which are valuable to clients. The numbers tell the story of the impact that microinsurance is making or is expected to make. However, there remains a degree of inconsistency in the number quoted as total microinsurance coverage; the overarching number within this sector. The often cited figures are 500 million, a number first published in Protecting the Poor, A microinsurance compendium (vol.II) in 2012, and 260.34 million, as cumulated from the first round of landscape studies in Latin America (2011), Africa (2012) and Asia (2013). It is important to rationalise this gap between these two figures to have a more consistent data-based narrative of the sector.

The two sources approach the number with different methodologies. The compendium uses back-of-the-envelope calculations to estimate 500 million, based on data from 2006 and 2009, as well as various region specific 2011 data. These are used to extrapolate sector growth in different regions, and thus estimate total coverage ranges for each region.

Over the past three years, the landscape studies have collected data on 1,264 microinsurance products, from 838 providers, in 102 countries throughout the developing world. They offer a more concrete, but significantly lower statistic with 260.34 million total coverage. All three landscape studies followed the same methodology:

  • Primary research with all microinsurance providers in each country through telephonic/online survey
  • Secondary research on profile of insurance/financial inclusion and microinsurance scenario of the countries along with the analysis of their regulation affecting microinsurance sector
  • Interview with sector experts in all countries on status and prospect of microinsurance in the country

These different methodologies are categorized as ‘back-of-the-envelope estimate’ for the 500 million and ‘data-driven figure’ for 260.34 million, and will be referred to as such throughout this article. These different methods prompt three main sources of divergence:

  • Definition of microinsurance and how to account for state-assisted microinsurance;
  • Quantifying total coverage in terms of risk or individuals covered, and;
  • Limitations in data collection.

The definition conundrum
Divergence in definition is the crux of the statistical gap.  The two methods initially approach microinsurance definition in similar ways, by (1) taking a mixed approach to operational definition, and, (2) identifying the importance of products having been developed intentionally for low-income populations; designed for their needs, preferences and characteristics.

Back-of-the-envelope estimation

Landscape studies Latin America, 2011 and Africa, 2012

Landscape study Asia, 2013

Microinsurance is the protection of low-income people against specific perils in exchange for regular premium payments proportionate to the likelihood and cost of the risk involved. The operational features of the definition includes:

Target group

Product definition

Provider definition

Distribution channels

Microinsurance is insurance that is modest in both coverage and premiums. Premium levels must be based on the types and amounts of risks insured. Additionally, in order to be considered microinsurance, products must meet the following characteristics:

Target population: the product is developed intentionally to serve low-income people

Non-government risk carrier: the government is not the risk carrier

Goal of sustainability: the objective of product results is ultimately profitability or commercial sustainability.

Minimal subsidies: the product must require no, or at most minimal direct subsidies

Microinsurance is insurance developed for and targeted for the low income segments of the populations that have been excluded by the mainstream insurance markets. The risk is underwritten either by a community based, semi-formal or formal insurer, while the premiums are paid by individuals fully or partials i.e. they receive a degree of subsidies (not 100%) either by the government or other entities.

Risks vs. individuals
The next important factor is the difference in total coverage quantification. The estimate counts coverage in terms of total risk covered. The data-driven number adjusts property coverage (like house or theft insurance, livestock insurance and even agriculture insurance) information to the number of policy holders. This can make a large difference in coverage counted, as demonstrated by the 2013 Asia landscape’s experience of collecting data in Mongolia. In Mongolia, 13,500 farmers have coverage on their 2.5 million livestock. When counted in terms of risks, Mongolia has a total coverage of 2.5 million, but when adjusted to individual policy holder that total coverage is only 13,500.

Which way of these is more appropriate in portraying total coverage? If coverage is portrayed as simply the number of people with policies, it does not give a full picture of the robustness or completeness of microinsurance coverage. If Mongolia’s 13,500 individual-adjusted total coverage were compared to another country’s 20,000 policy holders who only have life coverage and no protection for their property, the latter would be portrayed as having higher total coverage than Mongolia, which again may not be fair.

On the other hand, portraying coverage in terms of total risks can potentially distort data. The total coverage ratio measures the percentage of the population in a given country that is covered by microinsurance, and is calculated by dividing total coverage by the country’s population. In dividing by the total number of people, this calculation makes the assumption that total coverage is counted in terms of people. If total risk covered is divided by population, it will portray a disproportionately higher percentage of people covered by microinsurance. If this approach were taken in Mongolia, which has a population of only 2.8 million, 89% of the population would be portrayed as covered by microinsurance when in reality only 0.68% are covered. The argument for an individual-centric approach to counting coverage is that, if the primary concern of microinsurance is focused on people, the overarching statistic should be representative of people. How can this data be portrayed in a way that is representative of people while retaining the important level of detail that is included in total risk? This is a question to consider as we move ahead.

Accessing reliable data
The final significant differentiating factor is limitations within data collection; a common challenge in any numerical study. Barriers to data collection are a long-term challenge in microinsurance because of an inherent aversion to data sharing among insurers. The problem stems from the competitive nature of the sector, and a perceived low return in benefit from sharing data. Not being able to collect complete data is a significant problem for the data-driven figure, resulting in an unquantifiable gap between real total coverage and collected total coverage data.

Additionally, the time over which data was collected limits accuracy by not accounting for changes over the 1-2 year data stagnation between studies. By using annual compound growth statistics from landscape study reports, the growth adjusted figure for total microinsurance coverage in 2013 is 289.97 million.

Neither the estimate nor the data-driven figures can fully represent total coverage on their own. The back-of-the-envelop 500 million is not a direct product of data collected from the field, which limits the extent to which it can be used in the precise calculations needed to analyse performance and improve products. On the other hand, lack of access to complete data and limited collection of state-assisted microinsurance data means that 260.34 million does not capture the full scope of microinsurance coverage. When used together, however, these divergent numbers can play an important role in the advancement of the sector. Their stark difference illuminates the challenges of collecting data that represents the reach of coverage. The gap in statistics reveals the issue of whether to quantify coverage as total risk or total people covered, or how to combine them to portray both depth and reach of coverage. The issue of whether state-assisted microinsurance should be included in the count exposes a deeper ‘identity crisis’ of sorts, wherein the sector struggles to determine the role of government in microinsurance.

Should state-assisted microinsurance be counted as part of total microinsurance coverage?
With this conflict in definition as a main source of discrepancy in total coverage statistic, this is an important question to consider. The advocates of market-led microinsurance initiatives have the central belief that microinsurance can only be a long-term solution when the business case for such an initiative is a sustainable or profitable one. The drive to build a sustainable business case through scale and efficiency has been the key force behind innovations such as mobile insurance, which have extended microinsurance coverage like never before. The concern is that a government-led model is not sustainable and will be damaging to the growth of the sector in the long run.

Those who include state-assisted microinsurance data make the case that, while the market-based approach is appropriate for some of the target group, it is not accessible to the poorest segment of the population. Certain business lines do not currently have the capacity to be sustainable in a commercial setting and so government involvement is needed in order to offer such products. It is important for the two types of microinsurance to coexist and form partnerships to ultimately extend access to uninsured, low-income populations.

The way forward
The upcoming three studies, as part of the  World Map of Microinsurance (WMM) Programme , have already identified the issue of microinsurance definition, and are approaching it with a system of classification based on subsidy level, rather than excluding state-assisted microinsurance altogether. The new data will break down microinsurance in three categories of (1) fully commercialized (no or very low subsidy), (2) significantly subsidized but with risk taken by insurers, and (3) social protection schemes which are fully operated by the government. This adjustment will give a more complete picture of microinsurance and get us closer to the real number of total coverage without minimizing the important distinction between state-assisted and market-led schemes.

The problematic gap between the two prevalent numbers can be used to loosely quantify the holes in sector data. The real number of total coverage is likely to be closer to the compendium’s 500 million, so when 260.34 million is counted it effectively demonstrates the limitations of current data. If anything, these divergent statistics serve as an indicator of need for more complete and unified data. It is a justification for investment into a large-scale data-centred programme like the Network’s World Map of Microinsurance Programme. This programme will address the three issues raised here with determining total coverage. Firstly, it will serve as a focal point for discourse on sector-wide challenges, bringing stakeholders together to address such complicated issues as these. The accumulation of data in one place will facilitate better informed, and more collective decision-making. The second function that this programme fulfils is addressing the aversion to data sharing by making the results valuable to insurers. With any luck, documented improvements from access to data will diminish perceived disadvantages of data sharing, and enable future studies to expand on the data they collect.  Through the World Map of Microinsurance programme, this statistical divide can bring together microinsurance stakeholders across the board in a way that perpetuates the development of microinsurance throughout the world.

Written by Manoj Pandey, Knowledge Coordinator (March 2013 - September 2015) and Sarah McMinimy, Knowledge Intern at the Microinsurance Network (June - September 2014), July 2014.