Research on

1
Research on macro issues in governance and
development Philip KeeferDevelopment Research
GroupThe World BankWorkshop on Alternative
institutional approaches to ensuring effective
governanceGlobal Development NetworkJune 2008
2
What do I mean by. . . ?

• Governance
• 1) Performance of public sector?
• 2) Extent to which government serves broad
  public interest?
• Me Second
• Macro level analysis
• 1) Focus on higher vs. lower level governments?
• 2) Focus on citizen/beneficiary-politician
  relationships vs. service provider/bureaucrat -
  citizens/beneficiaries links?
• Me Second

3
Governance impact or determinants?

• Broad choices in governance and development
  research
• Collect further evidence that bad governance
  hurts development
• Identify causes of bad governance.
• If you believe Hausman-Rodrik-Velasco, the first
  is unsettled and deserves more attention.
• My view first is relatively settled the latter
  more pressing (and more fitting for the GDN).

4
Causes Bad laws, low capacity

• Most donor work on governance focused on bad
  laws, low public sector capacity.
• Bad laws (e.g., contradictory, vague,
  inefficient).
• Not a promising area of policy-relevant research.
  
• De jure changes have limited impact when politics
  unsympathetic to the public interest (Yakovlev
  and Zhuravskaya).
• Bad bureaucracies (poor capacity) and governance.
• More scope for research.
• Limited effect of de jure changes suggests low
  probability of success.
• But low capacity can be a reason that politicians
  dont act in the public interest (Huber and
  McCarty).

5
Causes Political market imperfections

• Deeper (more macro) causes of bad
  governance.
• Political market imperfections, a la
  Keefer/Khemani poor political incentives to
  serve the public interest.
• Citizen information
• Credibility of political promises
• Social characteristics (polarization).
• Research on the deepest causes
  history/geography.
• Compelling
• Good for identification
• Research on political mkt imperfections also
  furthers the deepest agenda what are the
  causal links from history/geography to incentives
  of current politicians?

6
Political mkt imperfections Citizen information

• We know
• media exposure associated with higher transfers
  (Svensson/Reinikka, Besley/Burgess, Stromberg)
• Information can increase demand (e.g., for
  health)
• We dont know
• Does information (per se) improve public good
  provision?
• Are info effects endogenous to other political
  factors (e.g., political credibility, access to
  gov. data, political salience)?
• Mechanisms What information? Do people who are
  more exposed to information know more? Do media
  lower costs of coordination?

7
Political mkt imperfections Political credibility

• We know
• Clientelist policies/political promises
  predominate in poor/young democracies (Keefer,
  Keefer/Vlaicu, Wantchekon) incompatible with
  public good provision.
• Political parties are main vehicle through which
  politicians bind themselves to provide benefits
  to large groups of citizens.
• We dont know
• When political parties emerge that have an
  identity independent of leader and bind
  candidates to party.
• When, specifically, politicians/parties believe
  that public good failure will cost them votes.
• Whether NGOs can turn into parties (Proshika?).

8
Political mkt imperfections Polarization

• We know
• Social fragmentation undermines public good
  provision (from Banerjee-Pande to Alesina, et
  al.)
• We dont know
• Is polarization endogenous to politics lack of
  credibility or information?
• Is polarization a product of history/culture
  (Hoff, et al.)?
• Is it violence, not ethnic fragmentation per se,
  that matters? Effects of violence for political
  ends (hartals, lynchings, etc.) on government
  accountability/governance? When does violence
  emerge?
• Are ethnic-based parties better than no parties
  at all (Ghana vs. Benin)?

9
Some tradeoffs to consider

• Ideal research project
• actionable policy recommendations with large
  impact on growth and development
• empirical design that leaves no doubts about
  causality and external validity.
• Real world tradeoffs

10
Some tradeoffs to consider policy aims

• Actionable vs. worthy of action
• Fundamental (more political) sources of
  governance failure are less actionable.
• Policy recommendations from research strategy
  that ignore politics are less likely to be
  action-worthy.
• Possible resolution
• explicitly discuss political context in which
  actionable reforms are feasible
• explicitly discuss whether small but actionable
  reforms (e.g., incremental, micro, sub-national,
  legal or bureaucratic) put us on a path to more
  fundamental change.

11
Some tradeoffs to consider definitive vs.
compelling

• Definitive causality vs. compelling development
  impact.
• Do we want to make definitive statements, even if
  the interventions are narrow?
• Or do we want to address bigger governance
  issues, even with less definitive statements?
• Eternal struggle. GDN should require that
• narrow, definitive research discusses external
  validity seriously, and
• broader, less definitive research discusses
  alternative explanations seriously.

12
GDN and macro-governance low risk

• What program best fits the GDN (amenable to local
  scholarship, policy relevant, but still cutting
  edge)?
• Low-risk information and governance.
• Allows for within-country, data-rich studies.
• Many open questions and GDN could lead the way in
  emphasizing political pre-conditions for success
  (external validity).

13
GDN and macro-governance higher risk

• Higher-risk violence/law-breaking and
  governance.
• Important (from Benin to Bihar) and under the
  radar
• Feasible American South (Alston and Ferrie).
• Not only hartals absence of censuses and
  vote-rigging are likely to be a key political
  reason for poor governance.
• Sources of political credibility?
• Theory good, but still hard to take to the data.
• However -- NGOs and political mobilization for
  public goods an important issue.