Communitybased risk management arrangements: Implications for Social Funds

1
Community-based risk management arrangements
Implications for Social Funds

• Ruchira Bhattamishra, World Bank
• Christopher B. Barrett, Cornell University
• November 10, 2008

2
Motivation I

• There is growing recognition in the development
  community that
• vulnerability to adverse shocks is a defining
  characteristic of poverty
• risk management is central to poverty reduction
  policy because uninsured risk exposure is both
  cause and consequence of poverty.

Uninsured risk
Poverty
3
Motivation II

• There is also growing recognition in the
  development community that
• community-based and community-driven development
  can be effective in filling key gaps between
  national- and household-level strategies.
• But theres a major gap in the literature on
  community-based risk management arrangements
  (CBRMAs) hence this survey.

4
Objectives

• Convey importance of social protection to
  economic growth and poverty reduction
• Provide catalog of existing community-based risk
  management arrangements (CBRMAs)
• Discuss strengths and weaknesses of CBRMAs
• Stimulate discussion of implications for SP
  projects, especially as implemented by Social
  Funds

5
Essential Role of Social Protection

• Insuring downside risk can stimulate significant
  technology uptake, investment and other behaviors
  that foster growth, especially among the poorest
  and most risk averse subpopulations.
• Effects of social protection are especially
  pronounced in places characterized by poverty
  traps with multiple welfare equilibria
• Ex post effectprevents the numbers of
  chronically poor from growing in the wake of a
  shock that destroys productive assets
• Ex ante effectimproves incentives, crowds in
  private asset accumulation and new technology
  uptake, making sustainable escape from poverty
  more feasible.
• Both effects reduce the number of households
  needing assistance, which lets true humanitarian
  assistance go further.
• Lets look a bit more at the economics that
  underlie these claims

6
The Potential of Social Protection

• Barrett, Carter and Ikegami (2008)
• offer simulation-based evidence on the impacts of
  social protection.
• (Solid blue is with social protection, green is
  autarky, red is targeted transfers.)

7
Typology of risks I

• Covariate risk vs. Idiosyncratic risk
• Covariate risk affects households in the same
  locate at the same time (e.g., weather,
  disasters, war, prices, financial crises, etc.).
• Idiosyncratic shocks are (the component) specific
  to one household (e.g., illness, crop yield
  shocks, property loss due to fire or theft,
  etc.).
• CBRMAs can help households cope with
  idiosyncratic shocks, less so with covariate
  shocks, unless risk is reduced or transferred
  outside the community.
• The empirical literature suggests that
  idiosyncratic risk is considerable, implying
  significant scope for risk pooling within
  communities.
• Even covariate risk can be managed within
  communities through risk reduction efforts (e.g.,
  through NRM) and external risk transfer (e.g.,
  through index-based insurance).

8
Typology of risks II

• Asset risk (loss of human capital, livestock,
  land, etc.) vs. income risk (loss of current
  period revenue)
• Long-term (structural ) vs. one-off (transitory)
• Stunting due to drought in Zimbabwe in 1980s
  caused 14 percent reduction in lifetime earnings
  (Alderman et al.2006).
• Implications for poverty persistence poverty
  traps or at least very slow recovery and thus
  high persistence of shock-induced poverty.
• CBRMAs can not only address one-off income risk
  but also, and perhaps more importantly,
  longer-term asset risk by developing community/
  individual capabilities.

9
Risk management

• Typology of strategies under SP Social Risk
  Management (SRM) framework (Holzmann and
  Jorgensen 1999)
• 1) Risk reduction ex ante reduce exogenous
  income variability and/or probability of asset
  loss (e.g., water control, EGS).
• 2) Risk mitigation ex ante reduce endogenous
  income variability and/or probability of asset
  loss through portfolio diversification,
  insurance, hedging, etc.
• 3) Risk coping ex post behavioral adjustment
  (investment and consumption adjustment, asset
  sales, borrowing) and/or risk transfer.
• Households employ a combination of strategies.
  But poor households typically have limited
  recourse to 1 and 2 and resort to 3, often
  through transfers (gifts, food aid, etc.).

10
Importance of risk sharing

• Incomplete financial markets leave uninsured
  risk.
• Sale of assets restricted to those that have
  assets, typically not the poorest, and may seek
  to asset smooth.
• In the event of common shock, assets and income
  may move together, limiting ability to
  consumption smooth.
• The poorest (such as disabled, female-headed
  households) often unable/unwilling to access
  public works programs.
• Informal risk sharing often the only avenue open
  to poor households, but social invisibility/exclus
  ion a problem for the poorest and most marginal
  populations (e.g., Santos and Barrett (2008) in
  Ethiopia, Vanderpuye-Orgle and Barrett
  (forthcoming) in Ghana).

11
Definitions

• Community based risk management arrangements
  (CBRMAs)
• Define community loosely in order to include
  agents whose relations have an informal and
  non-market character
• Include all coordinated strategies used and
  managed by social groupings of individuals for
  the purpose of protection against the adverse
  effects of various types of risk.
• Include both indigenously developed, informal
  and externally-initiated, semi-formal
  arrangements.

12
Definitions (cont.)

• Key similarities between indigenous and
  externally-driven CBRMAs Use of interpersonal
  relations in management contract enforcement.
• Key differences see below

13
Contextualizing CBRMAs in the SRM matrix
14
Contextualizing CBRMAs in the SRM matrix (cont.)
15
Role for Social Protection

• Limitations of indigenous CBRMAs
• Exclusion of poorest or other marginalized
  sub-populations
• Inability to manage covariate risk
• Role for SP intervention
• SFs can potentially build on existing
  institutional networks to support CBRMAs
• Can use large size of networks for risk pooling
  purposes
• Can build on experience with participatory
  approaches to develop innovative, demand-driven
  risk management products.

16
Range of possible approaches

• Promote inclusion by provision of subsidies
• Support start-up of viable MFIs
• Expand menu of projects
• Support provision of risk-reducing public goods
• Support risk coping after covariate shocks via
  intermediaries

17
Enabling inclusion

• Identify cleavages in existing CBRMAs.
• Can design safety nets schemes explicitly aimed
  at reducing costs of social interaction between
  different social groups.
• E.g., Macedonia Community Development Project.
• Can subsidize cost to poorest households to
  enable their inclusion.
• E.g., subsidize ex ante contributions for health
  insurance associations.

18
Provision of subsidies (community-level)

• Cover start-up costs of viable financial
  institutions.
• E.g., microfinance institutions.
• Some relevant insights from behavioral economics
• Cognitive difficulties in assessing risk choice
  bracketing representativeness etc.

19
Expanding menu

• Include innovative programs.
• Can go beyond thinking of SF as instrument for
  primarily developing brick-and-mortar outputs
  and providing basic services.
• Design safety nets schemes explicitly aimed at
  creating behavioral change and supporting risk
  management, both for one-off income risk as well
  as long-term asset risk.
• E.g., can develop PTAs in addition to building
  schools.

20
Reducing exposure to covariate shocks

• Provision of risk-reducing public goods and
  services through community arrangements
• Builds longer-term capacity of community.
• Addresses not only one-off income risk but also
  more long-term asset risk.

21
Supporting risk coping after covariate shocks via
intermediaries

• Build capacity of communities to tap into
  reinsurance markets
• Emphasize risk management rather than crisis
  management.
• Underwrite start-up costs associated with
  creating risk-transfer products
• E.g., underwrite cost of developing data series
  for pricing index-based insurance products. These
  are non-manipulable, suitable for risk-layering.
  In addition, they can support risk coping for
  both slow-onset (e.g., drought) as well as
  sudden-onset risk (e.g., earthquake).
• Use community information for effective two-tier
  allocation of disaster assistance (Alderman
  2001).

22
Potential problems

• Specific problems affecting Social Fund
  intervention for risk-management
• Administrative concerns
• Other problems (which affect community
  initiatives in general)
• Scalability, crowding-out, etc
• Manipulation by local elites
• Corruption
• Limits of community decision-making
• See also Mansuri and Rao (2004), Conning and
  Kevane (2002), Ensminger (2007).

23
Administrative concerns

• Differences in administering periodic investment/
  preparing proposal vs. overseeing regularly
  running program.
• Expanding menu to include innovative programs
  implies need for new training for program
  managers.

24
Scalability concerns

• Range of CBRMAs
• Differences in membership and leadership
  structure, the nature of activities, history,
  longevity, etc.
• Differences in political economy and
  socio-economic environment.
• Differences between informal and semi-formal
  arrangements in level of technical/financial and
  accounting assistance required.
• CBRMAs will have different abilities to
  effectively absorb external assistance, depending
  on the nature of activities, history, etc
• Lack of existing evidence on impact of scaling
  up

25
Crowding-out concerns

• Disruption of existing CBRMAs
• Rockefeller effect external assistance can
  change characteristics of a previously
  effectively functioning group.
• E.g., Gugerty and Kremer (2008) study in Kenya.

26
Vulnerability to manipulation

• Project benefits captured by local elites
• Corruption
• Decentralized, community-led approaches can
  result in rampant misappropriation of project
  benefits, given the absence of well-functioning
  checks and balances, remote governance
  structures, absence of media, and low levels of
  education.
• E.g., Ensminger (2007) study in Kenya.

27
Limitations of community management

• Limitations in technical decision-making and
  management
• Positive impact of community participation in
  non-technical decision-making (such as
  targeting/project choice) but not so for
  technical decision-making.
• E.g., Khwaja (2004) study in Pakistan.
• More complex accounting/financial knowledge
  needed for semi-formal versus informal
  CBRMAs.

28
Moving forward.

• Need for econometric or experimental evidence
  comparing community-based models with other
  models (e.g., social marketing, public-private
  partnerships, etc.) that do not use community for
  project design or project delivery.
• Compare impact, cost-effectiveness Social Funds
  can provide valuable crucible for such analyses.
• Address key questions
• What are some of the main constraints and
  opportunities for SF programs supporting risk
  management?
• How can these interventions achieve the right
  scale of implementation?