A Study on life microinsurance in Brazil

This study examines the feasibility of life microinsurance in Brazil. Microinsurance is aimed at people with per capita income of up to two minimum wages (MW), according to the definition made by the Brazilian government’s agency for superintending private insurance (SUSEP). The main source of data was the Brazilian National Household Survey (PNAD) from year 2008. The target population consists of 41 million family heads. For microinsurance coverages for death and funeral expenses, monthly premiums would range from US$0.36 to US$4.13 for those with per capita income of less than 0.25 MW. For the income range between 1 and 2 MW per capita, premiums vary between US$1.32 and US$15.21. The potential yearly market is approximately US$1.8 billion, with a potential profit of approximately US$274 million. These results provide significative evidence that life microinsurance can offer protection for low income people and also can be feasible in Brazil.


INTRODUCTION
he Brazilian insurance market has grown considerably in recent years; in fact, there was an annual real increase in premium volume of 9.5% in 2008.
However, despite this growth, the insurance sector still represents only 3.3% of the country's gross domestic product (GDP) (Fundación Mapfre, 2009).Most insurance business is concentrated in two groups: (i) car insurance policies, which represent 29.6% of the total; and (ii) VGBL (a type of private pension plan) products, which correspond to 34.5% of the total.
According to article 757 of the Brazilian Civil Law, insurance is defined as a contract through which "the insurer assumes the obligation, through the payment of a premium, to guarantee the insured's legitimate interest, related to the person or thing, against predetermined risks".The insurance premium is calculated on the basis of the risk involved in the contract.This risk depends on a range of factors -such as the insured's age, housing, and marital status, among others.
Most risks can affect economic agents in different ways.In particular, low-income people tend to be more vulnerable to some risks; moreover, their financial situation often means that they are unable to insure themselves against significant risks.From a social perspective, this situation can be problematic because it means that a particularly vulnerable subgroup of the population lacks basic financial protection against death or invalidity from certain catastrophic events, such as droughts, floods, collapses, and so on (Dercon et al., 2008).A similar view has been suggested, presenting the notion of microinsurance as one of the risk management tool, in the pioneering work by by Siegel, Alwang, Canagarajah (2001).
A similar view had previously been expressed by Townsend (1995), albeit without mentioning the precise term microinsurance.
Although there is no single agreed definition of 'microinsurance', its objective is to reach the low-income public by offering insurance policies with lower premiums and more rapid payouts than are available under traditional policies.It is normally offered to people who do not have access to conventional insurance policies (such as informal workers), which usually live in greater vulnerability conditions.The verification of this vulnerability and the inefficiency of more primitive risk management techniques (such as neighbours' assistance in the poorest regions) was one of the factors motivating the development of microinsurance in many countries.
However, many people in the target population of microinsurance are not in such extreme poverty that they fit into social assistance and income-transfer programs (Churchill, 2006).
In contrast to the rest of the Brazilian insurance sector, microinsurance has not yet been fully regulated, although legislation to this effect was under consideration at the time of writing.Despite the lack of formal regulation, microinsurance is acknowledged as part of the Brazilian government's policy on 'microfinance', which aims to provide loans and other financial services for low-income people., aimed at facilitating and expanding credit among formal and informal micro-entrepreneurs and access to financial services for the low-income population, as well as to increase the number and participation of credit cooperatives in the financial system and to decrease informality and interest rates in financing.
Discussions on this policy in 2003 gave rise to the term 'popular insurance' to designate mass products with low insured amounts and premiums.In 2006, SUSEP (the Brazilian government's agency for regulating private insurance) became a member of the International Association of Insurance Supervisors Consultative Group to Assist the Poorest (IAIS-CGAP), which addresses various initiatives, including microinsurance.This was seen as an important step for the development of microinsurance in Brazil.
In parallel with these regulatory developments, the Brazilian economy went through a combination of factors has occurred in recent years that may be unparalleled in the country's recent history, e. g. a prolonged period of low inflation, GDP growth of around 5% (except in 2009, due to the global financial crisis), and the expansion of income-transfer programs like Bolsa Família.These macroeconomic developments have been associated with an increase in the real income of low-income families, which has raised the demand for consumption goods and insurance products.Despite these positive developments, there are doubts about the status of the market for microinsurance; indeed, it is uncertain whether the product is actually feasible.
In view of this background, the objective of this study is to ascertain the feasibility of life microinsurance in Brazil.In accordance with Vaughan and Vaughan (2007), feasibility is assessed in this study in terms of: (i) the demand side (whether the premium to be paid is appropriate in terms of the consumers' income); and (ii) the supply side (whether the product is profitable for the companies offering it).In other words, the study calculates the life microinsurance premiums for some coverages offered.On the basis of these premiums and the number of potential consumers, the potential market size is obtained.
The remainder of the paper is organised as follows.The next section presents the theoretical background for the study, including the definitions of microinsurance and an overview of microinsurance around the world and in Brazil.In the third section, the research methodology is described.In the fourth section, the results of the analyses are presented, including an assessment of the feasibility of the proposed microinsurance products.In the fifth section, the results are subjected to a sensitivity analysis in which two other life mortality tables are used.Finally, in the last section, the conclusions are presented.

DEFINITIONS
According to ILO (2009), the term microinsurance appeared in the literature for the first time in the 1990s, derived from the microfinance, with which it shares a similar target public.
Because microinsurance is a relatively recent concept, there is, as yet, no consensus on a single definition of the term.Nevertheless, several descriptions and definitions of the concept have been offered in the literature.According to Chandani (2008, p. 5) IAIS-CGAP (2007) also defined the target public of microinsurance as: (i) low-income families; (ii) people working in the informal sector of the economy; and (iii) low-income workers from the formal sector.The nature of this definition reveals some degree of overlapping among these three groups.
All definitions presented emphasize microinsurance as a tool that can be used to increase social protection, targeting the low-income population.To fulfill these purposes, it is apparent that microinsurance premiums should be much lower than traditional insurances.In addition, there is a need for close proximity between the consumer and sales channels because a significant proportion of the business conducted by low-income people is conducted in person.This need for close proximity requires the use of alternative distribution channels that are convenient for the target populationsuch as supermarkets, community leaderships, betting shops, and so on.One example is a microinsurance sold by the South African insurance company Hollard.This company sells a funeral microinsurance, with additional coverages for death and accidents, at supermarket counters in poor regions.For coverages around US$ 1,600 for each of these risks, the total premium is less than US$ 8 per month.

MICROINSURANCE AROUND THE WORLD
Although it is a relatively recent concept, microinsurance is now well disseminated around the world.In many countries, microinsurance models are already an integral part of the implementation process of a comprehensive social protection system (Jacquier et al., 2006).Indeed, according to Roth, Mccord;Liber (2007), only 23 of the 100 poorest countries in the world do not have microinsurance programs.Despite this growth, among the 77 countries that do have such programs, only two have more than 21 providers (India and Mali), two have between 11 and 21 providers, and 73 have fewer than 11 providers (Roth Mccord;Liber, 2007).
In Central and South America, 7.8 million people (representing about 10% of the entire population) are covered by microinsurance; the majority of these, 6.7 million people, are from Colombia and Peru (Roth, Mccord;Liber, 2007).In Peru, most of the microinsurance is conducted through credit unions and relates to life insurance (more than 3.3 million); in Colombia, most of the life insurance policies (2.5 million) are linked to an entity that also offers funeral insurance.In Africa, microinsurance is quite limited, corresponding to a mere 4% (3.5 million) of all lives insured of the 100 poorest countries in the world.Moreover, approximately 1.6 million of those insured are living on less than US$2 per day.(Roth, Mccord;Liber, 2007).
Most of the lives insured by microinsurance in the 100 poorest countries come from Asia, which includes more than 67 million lives insured, approximately 58 million of whom live on less than US$2 per day (Roth, Mccord;Liber, 2007).The majority (more than 58 million) is concentrated in India (more than 30 million) and China (more than 28 million).
Nevertheless, more than 97% of poor people in Asia do not have any kind of microinsurance coverage.
Table 1 provides a summary of these data.Of the products traded in the three regions (Americas, Africa and Asia), a majority relates to life insurances (Roth, Mccord;Liber, 2007).Low-income population, for the purpose of microinsurance in Brazil, is the population segment with a monthly per capita income of up to two minimum wages, whose occupational position can be classified in the formal as well as in the informal economic sector.
The second partial report (October 2008) identified some barriers to microinsurance in Brazil, as well as noting the laws and standards of the current regulatory structure that might impinge on microinsurance.The Work Group concluded that the existing barriers were not derived from insurance legislation; rather, they derived, from such factors as taxation legislation, from insurance companies' high policy costs, and occupational legislation (SUSEP, 2008).Two other reports were also presented: one on microinsurance stakeholders and their respective roles, and the other on probable microinsurance products and their minimal parameters.
In f) Any combination of the criteria above.

LIFE MICROINSURANCE
The first Brazilian standard to address mass insurance policies directed at low-income people referred to life insurance.Circular letter SUSEP 267/2004 established functioning rules and criteria for operating popular life insurance.This standard defined a maximum limit of US$5,466 for amount insured in the basic guarantee.In addition, elements of the product's actuarial technical note were defined, such as secondary coverages, specification of rates or pure premiums, statistics used for their definition, rate reassessment criteria, loads, technical provisions, and so on.The standard product characteristics proposed by Circular Letter SUSEP 267/2004 were: • Death by any cause as basic coverage; • Funeral aid and basic food aid as additional coverages; • Maximum insured amount of US$5,466; • Issuing of individual certificates for collective insurance policies; • Encouragement of alternative commercial channels, such as electricity, gas or telephone bills.
It's worth noting that Circular 440, cited above, also defined the ceiling of the insured amount for life insurance.The value for death by any cause is about US$ 12,000, in 2012.
This figure seems to be overestimated, given the per capita income of low-income population in Brazil, as reported in the next section.

SAMPLE AND DATA COLLECTION
The The first step was to ascertain how many people qualified as potential microinsurance consumers.As described in section 2.3, the new Susep regulation (Circular 440) left to the insurance companies the definition of the target audience, based on a broad set of possible criteria.This new regulatory direction highlights a possible trade-off.On the one hand, the insurers will be free to define many taylor-made products for the needs of specific groups of consumers.On the other hand, makes less clear the definition of the potential consumers, based on their income.Given this new outlook, we made a choice to follow the criteria previously defined by Susep, i.e., the target audience for the life microinsurance policies analysed in this paper are the earners of per capita income of two minimum wages or less.
The minimum wage at the time of data collection for PNAD 2008 (September 2008) was BRL415 (equivalent to about US$227).Four income ranges (Y) were defined in terms of MW multiples.This choice is due to the importance of the minimum wage as an indexer for the lowest income ranges and the so-called lighthouse effect (as the definition given by Souza and Baltar, 1982-3) of the MW, even for informal workers.As shown in Table 2, approximately 150 million Brazilians were found to fit into the income range for microinsurance defined by SUSEP.Given that the large majority of traditional insurance consumers earn more than two MW, it was important to ascertain whether significant differences (apart from income) existed between those who were eligible for microinsurance and those who were not.some of the characteristics of these two groups.It is apparent that there were important differences between people eligible for microinsurance and those not eligible, with regard to observable characteristics, such as literacy and race.For example, only 2.65% of those eligible for microinsurance have studied more than 15 years or more of education, in opposition to 25.84% of those not eligible.Likewise, a difference can be observed between both categories, with a greater predominance of white people (69.30%) among those not eligible for microinsurance, against 43.18% in the eligible group.Another considerable difference refers to the percentage of social insurance, and private pension plan contributors.
There was clear evidence that people eligible for microinsurance were less likely to be covered by social protection mechanisms, and thus more subject to risks, than those who were ineligible.These statistics refer only to people who were in the economically active population (EAP)that is, employed at the time of data collection for PNAD 2008.It is apparent that the group eligible for microinsurance included a greater proportion of informal workers and a smaller proportion of employers.In addition, the proportion of people with zero income (basically workers for self consumption and non-paid workers) was greater in the target public for microinsurance.Finally, Table 5 shows the size of families in both groups.It is apparent that, on average, the families of people eligible for microinsurance are larger than those of ineligible people.This is in accordance with the common observation that poorer people have a larger number of children.As these families have less access to credit and formal risk protection mechanisms, like the usual products the insurance market offers, this is an evidence of this groups' needs for access to alternative risk protection mechanisms even more, such as microinsurance.As explained above, because this is an insurance for the low-income population, it can be reasonably assumed that neither policy emission costs nor additional financing of the premium will be charged (which are common in mass products such as car insurance policies).For the pricing model of the proposed product, the adopted premises are based on the common practices in the insurance market and current legislation.

MARKET SIZE IN BRAZIL
The target public of microinsurance in Brazil was defined for the purposes of this study as the number of persons (aged 18 years to 80 years) responsible for a family (or considered as such by other members.of the family) who have a maximum per capita income of two minimum wages or less (according to SUSEP rules).Families with zero income were not included in the calculations.It was assumed that none of the families would purchase more than one life insurance.
Table 6 shows the potential market according to the per capita income ranges (Y), given in multiples of minimum wage (MW).The data show that the size of the market for life microinsurance in Brazil was approximately 41.8 million people.

PROPOSED PRODUCT PREMIUM -BASELINE SCENARIO
The Gross Premiums for both were based on what was designated as a Baseline Scenario.In this scenario, in which certain premises were assumed in accordance with the Brazilian insurance market and the characteristics of the target population of low-income families.
In choosing an appropriate mortality table, it was assumed that the mortality of the lowincome population would be compromised by poorer health, fewer years of education, lower income, inadequate housing and so on (Waldron, 2007) The Commercial Load (θ) consists of the administrative expenses (DA), commercial expenses (DC) and profit margin (ML).The average values for the Brazilian insurance market are 19%, 28% and 20%.These figures were calculated based on data collected in SES -Susep Statistical System.This database contains the balance sheets of all insurers in Brazil.Since life microinsurance is a new insurance product, it is expected that operational costs are higher and profit margin is lower than the market as a whole.We decided to be conservative in our choices.The values adopted for these variables were 20%, 30% and 15% respectively.The parameter q, which is given by the sum of DC, DA and ML, was thus 65%.
For the Statistical Load (α), 10% was adopted i .In principle, this parameter should reflect variations in mortality or loss ratios, which could affect the insurance company's result.In addition, it should be reminded that the IBGE mortality table, although the most appropriate available table, might underestimate the mortality of people eligible for microinsurance because this table is calculated on the basis of the entire population's mortality experience (rather than only the low-income population, as in the present study).
Table 7 shows the premises of this so-called Baseline Scenario:  To avoid an annual increase in premiums with increasing age (given an annual rise in the probability of death of the insured person), five age ranges were defined within the overall span of 18 years to 59 years (or more).In accordance with local Brazilian Statute of the Elderly, no divisions were considered after the age of 60 years.Table 9 shows the number of people per age range and income range.As observed, the number of people in the first age range is smaller than in the other groups.This is the case because it was defined that only heads of family would be considered as the target public, and it is more common for these to be concentrated in the higher age ranges.For each range, the premium to be paid corresponds to the weighted average of the premium calculated in terms of the number of people in each age range.Premiums payable for each of the four per capita income levels.The load percentages and profit margin are assumed to be the same in each age range.
In addition to the monthly premiums payable for each coverage, Tables 10-13 also indicate the number of people eligible for microinsurance and the monthly market potential.
For example, in    The greater the life expectancy, the lower the premiumbecause the insured person will pay for a longer period.In the baseline scenario described above, the IBGE's male Mortality table was used, which reflects mortality conditions for Brazil in 2008.However, insurance companies typically use the mortality tables elaborated by the Society of Actuaries (SOA), which are based on the demographic pattern of the USA and the accumulated experience of companies in the insurance market.Because the Brazilian experience differs from the American experience, and because this difference increases the possibility of adverse risk rate oscillations, it is therefore important to impute a statistical load percentage in the calculations when using an SOA mortality table.For these reasons, the present study subjected the obtained results to sensitivity analysis by pricing the proposed microinsurance products using two other mortality tables.
In Brazil, where the use of mortality tables in the life insurance business (and also in private pension plans) is regulated by SUSEP, most life insurance premium calculations use the AT-83 Annuity Table, which reflects mortality conditions in the USA in 1983.This table was used for the sensitivity analysis.However, the parameters of this table might not be appropriate for the microinsurance proposed herebecause the probabilities of death in the table are always less than those recorded by IBGE in Brazil.Given the socio-demographic characteristics of the eligible population for microinsurance, it was therefore decided that pricing would also be calculated with the AT-49 table, which has mortality rates that are closer to the target public of the product under analysis.Indeed, this latter table was used for various Brazilian insurance products until some years ago.
Table 14 summarises the three scenarios using the three tables noted above (IBGE-2008 male; AT-49; and AT-83).To define the statistical load percentages, allowances were made for potential differences between the mortality rates in each of the tables and the real mortality rates of the target population.For example, the life expectancy in AT-83 is greater because this is a more recent tablethus reflecting advances in housing, nutrition, and health conditions).As a consequence, there is likely to be a greater disparity between this table and the mortality rates in the proposed Brazilian target public.A greater statistical load rate therefore needs to be used to reflect the greater uncertainty associated with the use of this  In making the calculations, the same methodology was adopted as described above for the 'baseline scenario'.The results for scenarios 2 and 3 for each income band are shown in Tables 15-18.It is apparent that there is little difference in premiums between the baseline Scenario and AT-49, but that there is a much larger differential between the Baseline Scenario and AT-83.This reflects the differences in the tables' mortality rates, particularly in older age groups.
With regard to the potential market size, the figure for scenario 2 was US$1.64 billion per year and that for scenario 3 was US$1.04 billion per year.These figures represent annual profits of approximately US$246 million and US$156 million, respectively.Despite the considerable variations in the various mortality tables, it would seem that these markets are attractive.Moreover, the monthly premiums remains quite low.In other words, both supply and demand conditions were maintained, even when other mortality tables were used.
June 2012, Susep released a set of five long-awaited legal norms (Susep 2012a, 2012b, 2012c, 2012d e 2012e), which described and clarified many features about the microinsurance operations in Brazil.Most doubtless, the more important is Circular 440.This norm defined a somehow surprising change, compared to previous legal regulations, Susep no longer defined low-income population according to their per capita income, as it had been done before.Susep left the definition of the potential consumers to the insurance companies, that are allowed to choose of the following criteria: a) Demographic: age, income, gender, family size, schooling, occupation, among others; BBR, Braz.Bus.Rev. (Engl.ed., Online), Vitória, v. 10, n. 3, Art.6, p. 130 -155, jul.-sep.2013 www.bbronline.com.brb) Geographic; c) Social aspects (e.g.eligibility for social transfers programs); d) Economic aspects (specific occcupations); e) Behavioural aspects (life style, consumer decisions etc);

TABLE 1 : NUMBER OF INSURANCE POLICIES PER PRODUCT AND REGION Region Life Health Accidents and Disability Property
Manuamorn (2007)ickery (2008)creative examples of rural microinsurance in Asia.These include:Cai et al. (2009), who studied microinsurance in China;Giné, Townsend;Vickery (2008)andManuamorn (2007), who studied rural insurance in India, where farmers are indemnified against rains being above or below certain levels; Goldberg and Ramanathan (2008), who published a review on Latin America; and Matul et al. (2010) who conducted a similar review of microinsurance in Africa.More recently Churchill and Matul (2012) presented an entirely new version of their classical work, written in 2006, with a comprehensive summary of the new developments in microinsurance.BBR, Braz.Bus.Rev.(Engl.ed., Online),Vitória, v. 10, n. 3, Art.6, p. 130 -155, jul.-sep.2013  www.bbronline.com.br Its goal is to investigate various socio-economic characteristics of the Brazilian populationincluding education, work, income, housing, health, migrations and so on.PNAD 2008 had collected data from approximately 100,000 Brazilian homes and 400,000 people.
main information source for the present empirical study comprises microdata from the 2008 National Household Sample Survey (PNAD).This rich research is conducted annually by the Brazilian Institute of Geography and Statistics (IBGE) (except in years when the Demographic Census is held), using a representative sample of Brazilian homes and BBR, Braz.Bus.Rev.(Engl.ed., Online),Vitória, v. 10, n. 3, Art.6, p. 130 -155, jul.-sep.2013www.bbronline.com.brinhabitants.

TABLE 3
Table 4 compares the eligible and ineligible groups in terms of occupational status.

TABLE 6 : TARGET PUBLIC ACCORDING TO PER CAPITA INCOME (MULTIPLES OF MINIMUM MONTHLY WAGE) Per capita income Y per MW range
. A shorter life expectancy than other groups in society was thus expected.The IBGE male table published in 2008 was considered the mortality table that best represents the biometrical risk profile of individuals eligible for microinsurance.The choice of this table was in accordance with the practice of the Social Security Ministry, which uses this table to calculate retirement benefits by contribution time.

TABLE 7 : BASELINE SCENARIO Item Value
This figure was based on the assumption that funeral expenses would be similar for all people in the microinsurance target public.In contrast, the insured amount for death coverage (CM) varied with income.This figure was based on the income of the head of a family over a period of six months from the date of death.This figure was calculated for each of the four income ranges (defined according to minimum wage multiples) as six times the highest income in each range.These numbers are shown in Table8.For example, the payout figure in the first line of the table showing an insured amount of US$ 340.28 for CM was obtained by multiplying 0.25 MW (US$56.71)by 6 months.

TABLE 9 : NUMBER OF PEOPLE IN MICROINSURANCE TARGET PUBLIC PER AGE RANGE AND MW RANGE Per capita income Y per Age range (years)
Table 10, the potential demand in the age range between 18 and 28 years was