Transcript Clientelism, credibility and the policy choices of young

Governance and Growth
Philip Keefer
Development Research Group
The World Bank
Disclaimer: The views expressed here are those of the author and not
those of the World Bank, its management or executive board.
My version of governance


Degree to which governments are responsive to
citizen needs/demands.
Encompasses routinely measured governance
outcomes:
o
o
o

Rule of law and the security of property rights
Bureaucratic quality
Corruption
But we could easily add
o
o
o
Regulation/Investment Climate
Education, health, infrastructure quality
Correlation btw. rule of law and Human Development
Index=.72
Distractions in the governance debate

Are measured governance outcomes important for
growth?
Theory and evidence are at least as strong as the
evidence for other growth determinants.

Are the measures of selected governance
outcomes accurate?
No alternative to current broad measures of corruption,
bureaucratic quality and rule of law/security of property
rights.
Two more pressing questions



What explains significant outliers with less-thanstellar measured governance (East Asia)?
When does good governance emerge?
Political calculations of leaders underlie answers to
both questions.
o
o
In autocracies: does the regime stability gained by
faster growth offset the lost rents from better
governance?
In democracies: do broadly targeted policies (e.g.,
growth-promoting good governance) offer greater
electoral advantage than narrowly targeted policies
benefiting a few. . . at similar or higher cost?
What do successful autocracies do
differently?


Autocrats can’t decree growth – decrees do not
provide guarantees to investors.
Autocrats must either
o
o

promise investors they will not be expropriated and
promise regime supporters rewards tomorrow if they
refrain from expropriation today (=good governance)
. . . and/or offer extraordinary rates of return.
o
E.g.: Even mSouth Korea/Taiwan/Singapore “state-led”
growth required credible promises to investors.
Syria/Algeria have trouble making these. Difference?
Credible commitment in autocracies

How do autocrats do this?
o
o
o
o


Personalized (family, etc.) – few benefit.
Third party (warships from investor home countries)
Intra-elite checks and balances (many armed but
cooperating coalitions)
Institutionalized ruling parties – most benefit.
Only some benefit from good governance in
autocracies (e.g., members of ruling party, foreign
investors).
Arrangements can be sufficient for growth, but
governance indicators mediocre b/c governance
good for a minority only.
What matters in party institutionalization?

What to look for in instit’ized ruling parties.
o
o
o

Intra-party checks and balances (China now).
Sunk investments in intra-party evaluation and promotion
(China, Indonesian military, Singapore civil service).
Coordination privileges for party members – e.g., to
replace non-performing leader (China – segmented info)
(Gehlbach and Keefer).
Evidence that institutionalized ruling parties expand
investor pool:
60 million Communist Party members = 5 percent of
Chinese population; 6,000 Saudi royals < 1/4000 of
population. Measured governance higher in SA, though.
Fast growth possible in non-institutionalized
autocracies – but not sustainable

Porfirio Díaz (19th-20th century): Mexico grew
reasonably fast, but poor “measured”
governance. Why?
Haber et al:
o
o
o
o


policies raised investor rates of return: monopoly rents
(e.g., in banking);
crony (personalized) relationships;
armies of investor home countries;
powerful (armed) governors could protect their own
rights.
No party institutions.
Growth was not sustainable and end result was
Mexican Revolution.
Egypt more typical

Slow growth, no institutionalization.
o
o
o
o
o


No clear party structure (like China, Vietnam).
No intra-party checks on government leaders.
No decentralized military hierarchy (like Indonesia): US
sergeant has as much discretion as Egyptian colonel
No meritocratic bureaucracy (like Singapore)
No militarily strong governors (like Mexico)
Why not? Rents (foreign aid/oil)?
Measurement problems:
o
KK rule of law index, 1996 = .23.
Index = -.47 in (much) faster growing China and
Vietnam (3/4 std. deviation worse).
Chinese success and institutionalization


Under Mao, ferocious intra-party conflict.
Since Mao, and increasingly:
o
o
o


regularized cadre advancement, better information
among cadres than before and
more regularized succession and intra-party checks.
Helps that revolutionary leaders dying off – charisma
counts for less, provincial leaders more powerful.
Leadership can credibly promise cadres that they
will be rewarded for growth.
These innovations occurred after top leadership
(Deng Xiaoping) decided growth was key to
longevity (Keefer).
Some systematic evidence


Property rights in non-democracies improve as
age of political party increases relative to tenure
of leader.
Party institutionalization (age of parties, “rubber
stamp” legislature) less likely when natural
resource rents are lower.
(Gehlbach and Keefer)
Why don’t all autocracies grow fast?

Governance hard to improve when:
o
o

Rents too high, no incentive to pursue growth/share
power (Middle East/North Africa).
If leader is charismatic, risk of overthrow is low, don’t
need growth (Mao, Kenyatta, Nyerere, Chávez ??).
Generally difficult to create institutionalized ruling
parties.
Lessons for governance reform in autocracies


Institut’nalized autocracies: rents to ruling party
members and information controls essential to
regime stability
Incompatible with public sector reform,
democracy promotion.
o
o

These may undermine governance.
NB: most autocracies don’t have such parties.
Focus on rents essential in poor-performing
autocracies.
o
o
o
Rents discourage institutionalization (and democracy).
Aid flows = rents?
Monitoring and publicizing rent flows helpful (oil
receipts)
Lessons for governance reform in autocracies

Donors and crony arrangements
o
o
o
Support for them possibly harmful in non-
institutionalized autocracies (rents increase);
Opposition possibly counter-productive in
institutionalized autocracies (undermines regime).
In any case, donor focus should be on spreading
opportunities for cronyism as widely as possible
(existing cronies have no problems).
What about democracies?



Democracy at the center of current debates on
governance/security of property rights.
Many governance indicators focus on
democracy/voice.
Much scholarly debate implies that democracies
fail because they redistribute too much
Przeworski, Limongi and Cheibub; Acemoglu and
Robinson.

Focus incomplete.
Insufficiency of democracy is clear
Poor non- Poor
Rich
democracies democracies democracies
2.7 (25)
2.9 (34)
4.1 (49)
Corruption 1997
(0 – 6, least corrupt = 6)
Bureaucratic quality 2.3 (28)
2000 (0 – 6, 6 = highest)
Rule of law 2000
3.7 (28)
(0 – 6, 6 = highest)
2.4 (30)
4.6 (51)
2.9 (30)
4.6 (51)
Objective indicators send the same message
Poor non- Poor
Rich
democracies democracies democracies
Days in customs, 2004
Days to enforce a
contract, 2004
6.8 (15)
410 (37)
8.15 (16)
416 (30)
5.23 (17)
331 (50)
Not obvious that excessive redistribution
associated with poor democratic performance
Poor non- Poor
Rich
democracies democracies democracies
Total gov’t.
.25 (21)
expend.’s/GDP, 1998
Gross sec. school
39.8 (34)
enrollment, 1998
School spending/GDP, .038 (16)
1998
.23 (21)
.32 (38)
45.7 (25)
95.8 (48)
.025 (24)
.027 (36)
Political market imperfections better explain
why poor democracies struggle

Citizens can’t hold government accountable when
they are uninformed.
Politicians can’t get credit for improving citizen welfare
when citizens cannot distinguish effects of political
decisions on their welfare from all other influences.

Citizens can’t hold government accountable when
promises of political competitors not credible.
When challengers can’t credibly promise to pursue
different policies than incumbent.

Governance agenda in democracies should focus on
these (Keefer and Khemani).
Evidence that information makes a difference


Uganda Expenditure Tracking Survey
Newspaper circulation
o
o


reduces corruption (Adsera, Boix, Payne)
increases government responsiveness to food shortages (Besley and
Burgess on India).
Radios increase access to welfare payments (Stromberg on
US Great Depression).
Current WB research: does info improve public good
performance; what kind?
Gov’ts discourage information in poor
democracies
Poor non- Poor
Rich
democracies democracies democracies
16 (40)
26 (29)
196 (58)
Average newspaper
circulation per 1000,
1995
Market share of state- 63% (27)
owned newspapers,
1998
40% (14)
4.2% (47)
Low political credibility in poor democracies

What do politicians do when they can’t make
credible promises?
o
o
o
o
Target resources to groups of voters who believe them;
Seek out “patrons” who can make credible promises to
“clients”
Leads to more targeted goods, fewer public goods,
more rent-seeking (Keefer and Vlaicu).
Explains clientelism.
Credibility evidence: Young democracies

Poor democracies disproportionately young
6 years vs. 26 years.


Politicians in young dems. less credible (on average)
In fact: young democracies exhibit “low credibility”
policies:
o
o
o
o
low public good provision,
high targeted goods,
high rent-seeking/corruption.
Only credibility explains this constellation of policy
outcomes
2000 elections and public spending in Ghana
45%
40%
35%
30%
25%
20%
15%
10%
5%
05
20
03
20
01
20
99
19
97
19
95
19
93
19
91
19
19
89
0%
gross secondary enrollment
public investment/GDP
central government wage bill/GDP
tax revenues/GDP
2000 elections and governance in Ghana
4.5
4
3.5
3
2.5
2
1.5
1
0.5
corruption
rule of law
20
05
20
03
20
01
19
99
19
97
19
95
19
93
19
91
19
89
0
bureaucratic quality
Political market imperfections explain modesty
of Ghana electoral effects on policy

No broadly credible (programmatic) parties
o
o
o
o

NDC/NPP supporters have same market preferences
(Afrobarometer).
Political campaigns ignore policy differences.
Party and party policy stances irrelevant to legislative
campaigns.
Personal handouts by candidates hugely important.
Voters uninformed
o
o
Afrobarometer 2005: 60% NEVER get news from
newspaper vs. 14% in South Africa
Afrobarometer 1999: Up to 10 years of education (vast
majority), 36 percent could identify Economic Recovery
Program (ERP); more than ten years: 66 percent.
Lessons for governance reform in democracies




Traditional public sector reform agenda (e.g.,
meritocracy) unsustainable when political market
imperfections severe (e.g., place a premium on
patronage).
More effective donor policy focuses on political
market imperfections directly
Information easiest to address, but credibility key.
Information – governance link is the focus of
much ongoing Bank research.
Lessons for governance reform in democracies

Efforts to increase info directly:
o
o
o
o

support collection/dissemination of information on policy
impacts to disaggregated groups of citizens.
include information components in projects.
monitor information environment (e.g., newspapers,
radios, access to gov. information)
address policy barriers to gov’t information
Give priority to reforms that are easy for citizens
to monitor – allow faster reputation-building.
o
o
Infrastructure before public sector financial mgt.
Link PSM reforms to visible, sectoral reforms.
Lessons for governance reform in democracies

Don’t let crises distract from long-run goal of
alleviating political market imperfections.
o
o

E.g., despite woeful state of voter info, donor focus in
Ghana is emphatically infrastr. crisis.
Problem: there are always crises.
Tailor “demand” side interventions to the specifics
of political market imperfections.
Not info for info’s sake; civil society for civil society’s
sake.
Implications for quick wins and other strategies



Much discussion of “quick wins” as a way to
improve governance.
In the context of growth, where political credibility
towards investors is the key governance issue,
what “quick wins” improve credibility?
Costly actions by politicians that have a payoff
only if investors enter (infrastructure, subsidies for
foreign investors = China, Singapore).
Conclusion



Donor growth-governance work still needs to be
concerned with credibility issues.
This is consistent with emphasis in widely-used
governance indicators.
But need sensitivity to political particulars.
o
o
In non-democracies, growth can occur when a “large
enough” minority enjoys good governance (East Asia).
Donor objective: support conditions to achieve “large
enough”.
In democracies, growth fails when political market
imperfections prevent too large a share of voters from
holding politicians accountable.
Donor objective: mitigate the imperfections.
Conclusion-2


These arguments concur with others in pointing to
the insufficiency of current governance
prescriptions/ diagnostics.
Reasoning is different, though.
o
o
o
Political market imperfections, not capacity and fiscal
restraint, drive outcomes.
Both capacity and fiscal resources are policy choices, not
constraints.
Implies distinct reform strategies.