Formal vs informal insurance in South Africa

1
Formal vs informal insurance in South Africa

• Erlend Berg
• EC501 Presentation
• 6 March 2006

2
Summary

• Research question What determines the choice
  between formal and informal insurance?
• Why does it matter Should we try to graduate
  people into the formal sector? Why/why not?
• Tentative conclusion Informal insurance may
  dominate formal insurance due to information
  advantages, but its supply may be restricted.

3
The take-up of life insurance
4
The take-up of life insurance
5
A more apt comparison?
6
South African burial societies Informal insurance

• Informal or semi-formal risk-sharing and/or
  savings devices
• Typically composed of people with something in
  common ethnic origin, profession, church,
  kinship, neighbourhood
• Primary function is to assist with funeral costs
• But may also provide food, services in this
  context
• Serve social functions, not just when deaths
  occur
• Covers the member and family members
• Premia and payouts tend to be fixed irrespective
  of risk??
• Most societies collect money at regular meetings
• But some collect money only when deaths occur
• Three types according to link with formal sector
• Traditional
• Commercial
• Hybrid
• Strict rules and codes of conduct

7
Burial societies are popular amongst the black

• Burial societies are predominantly black (96)
• Almost 1/3 of adult blacks are members
• Many more are covered?
• Variation in size of membership
• Most societies are small
• Average society has 88 members
• Some societies have more than 1000 members
• It is not unusual to be a member of more than one
  burial society
• Or to sign up for formal cover in addition

8
Formal cover is also available, in principle

• Funeral policies are formal versions of burial
  societies
• But no social aspect
• Life policies are typically larger-scale
• Higher premia, more cover
• Formal insurance may be less accessible than
  burial societies
• Fixed cost, minimum purchase?
• Limitations in physical presence?
• These categories are starting to blur
• ..but still Many black South Africans have a
  choice between formal and informal funeral
  insurance

9
What characterises burial society members?

• Compared to holders of formal funeral insurance,
  burial society members are more likely to be
• Female
• Less educated
• Poorer
• Unemployed
• Unbanked
• Saving informally
• Unmarried, no children, living with parents
• They are not very different in terms of
• Age
• Retirement status

Data is from 2003/4 and covers black South
Africans aged 16
10
OLS a first cut at explaining participation
Subsample holding formal or informal cover
Full sample
11
What characterises the insured?

• Burial societies appear to be democratic
• Primarily driven by gender, age and family
  situation
• Education, employment, income, rural/urban
  location seems to matter less
• High correlation with stokvel activity
• Take-up of formal cover predicted by
• Gender, education, employment, income, family
  situation, rural/urban location and other links
  to formal finance
• Age and participation in stokvels not predictive

12
How are formally insured different from the
informal?

• Given participation in some form of funeral
  insurance, what determines whether you go
  informal or formal?
• Predicts formal insurance
• Education
• Employment
• Rural (negative)
• Other links to formal finance
• Stokvel participation (negative)
• No predictive power
• Gender
• Age
• Family situation

13
How can this be understood in terms of theory?

• Burial societies bundle insurance and social
  goods
• But even from a pure insurance point of view,
  burial societies differ from formal insurance
• Like most informal financial arrangements, it is
  based on superior (local) information
• Therefore, it can be more flexible in terms of
  payments
• How can this be captured in a traditional
  economic theory framework?

14
How formal and informal insurance differ

• Introduce a framework of uninsurable income
  uncertainty
• Formal insurance is inflexible
• The fixed premium is spd
• Actuarially fair
• Rigidity linked to information constraints
• Informal insurance has a lot more information
• Attendance, commitment, monitoring
• The premium is a fixed proportion of income, ay

15
The classic insurance story A shock d occurs
with prob p
u(x)
EU w/o insurance
y
y-d
y-pd
x
(occurs with probability 1-p)
(occurs with probability p)
16
Actuarially fair insurance is welfare-improving
u(x)
It does not matter whether the premium is charged
as a fixed amount spd, or as a fraction of
income aypd. They are equivalent and
actuarially fair.
EU with insurance
EU w/o insurance
y
y-d
y-pd
x
(occurs with probability 1-p)
(occurs with probability p)
17
Introduce an uninsurable uncertainty in income
u(x)
Pre-shock income is ye with probability ½, and
y-e with probability ½.
x
ye
ye-d
ye-pd
(occurs with probability 1-p)
(occurs with probability p)
18
Insurance is still welfare-improving
u(x)
EU with insurance
EU w/o insurance
x
ye
ye-d
ye-pd
(occurs with probability 1-p)
(occurs with probability p)
19
But now the form of insurance matters!

• Under income uncertainty, it matters how the
  premium is charged
• As a fixed premium spd
• Or as a fraction of income a(ye), where aypd
• In expected terms, both are actuarially fair
• Esspd regardless of income
• E(a(ye))aypd
• But a risk-averse consumer will prefer the latter
  type

20
Zooming in
u(x)
EU with insurance
EU w/o insurance
x
ye
ye-d
ye-pd
(occurs with probability 1-p)
(occurs with probability p)
21
Why informal insurance is better
EU with informal insurance
EU formal insurance
(ye)(1-a)
ye-s
EU without insurance
y-pd
22
The advantage of informal insurance

• Informal insurance may allow you to pay when you
  can afford it
• Essentially an income-smoothing feature
• This feature relies on an information advantage
• Parallells other forms of informal arrangements,
  and group lending
• But informal insurance may be supply-restricted
• There are only so many societies you can join
• Through your church, your street, etc
• So as income increases, you may have to branch
  into the formal market
• Can this explain the observed pattern?

23
The take-up of life insurance