Transcript Institutional and Governance Reviews and the Role

Political Economy and Growth
Philip Keefer
Development Research Group
PREM Growth Course
26 March 2009
Why not just focus on economics?
Partial reforms don’t deliver big growth bang for buck.
Economic explanation:
• Wrong reforms (e.g., not the “bottleneck” reforms).
• Exogenous constraints (capacity; land-locked;
neighbors; no knowledge)
• Synergies absent (Reform more than sum of parts).
• None distinguish fast- and slow-growers, though.
Political explanation:
• Governments have limited interested in growth
• Explains partial, poorly implemented, halting, wrong
reforms; inattention to exogenous constraints; lack
of synergies.
• Assertion: if one size does not fit all, it’s more likely
because of politics than economics.
What do politicians need to do to support
growth?
Growth requires: Policies that give a substantial fraction of
entrepreneurs access to
• finance
• skilled workforce
• land and transportation
• regulatory approvals and security from opportunistic
behavior by government officials
• (etc.)
When do politicians pursue these?
• When political success depends more on broad public
than on special interest support
• When they can align incentives of public officials who
implement these policies.
What dimensions of politics matter?
2 key political market imperfections (PMIs): lack of
information; lack of credibility (Keefer and Khemani, WBRO
2005)
Citizen information
• Uninformed citizens can’t hold governments
accountable for poor performance.
• Discourages pro-citizen (pro-growth) policies (no
political credit)
• Encourages pro-special interest policies (no blame).
• e.g., Grossman and Helpman: special interests
finance advertising to reach uninformed citizens.
Lack of Political Credibility
Low political penalty for reneging on promises. Why?
• Challengers can’t promise to do better.
• Incumbents bear no reputational loss.
• Citizens cannot hold government officials collectively
responsible.
Policy consequences
• Low credibility promotes narrow targeted (clientelist)
policies, discourages broad public policies (Keefer and
Vlaicu, JLEO 2008).
• Explains policies in young democracies: more
corruption, less rule of law, less secondary education,
larger public sector wage bill/GDP (Keefer, AJPS,
2007).
Political parties key
Institutionalized parties allow broadly credible promises:
• Instill collective responsibility;
• Generate reputational costs from reneging
Give governments leverage over officials:
• harder for officials to play divide and conquer;
• appointments less likely to be cronyistic;
• government-official “contracts” more credible.
Parties institutionalized if:
• Voters vote for party, not just candidates.
• Parties raise money.
• Party members can replace party leaders.
Per capita growth ~2.5 percentage points higher in
democracies with programmatic parties.
Same issues important in autocracies
Autocrats who want to pursue growth need to bring
government officials on board.
But how can they credibly promise to reward officials for
success?
Institutionalization (of parties, bureaucracies, military)
(Gehlbach and Keefer).
• Autocrat allows large sub-group of citizens to organize
◦ (60 million members of the Chinese Communist
Party)
• Increases their ability to overthrow autocrat if he
reneges.
• Increases their willingness to invest.
• Autocrat can write pro-growth contracts with party
members
Ruling party institut’n and investment
Dependent
variable:
Private investment/GDP
Base
sample
More
autocratic
Age of ruling party
0,088
(0.01)
0.109
(0.00)
Ruler years in
office
0.072
(0.43)
-0.050
(0.70)
Length of non0.012
democratic episode (0.92)
0.103
(0.53)
Political instability
0.677
(0.90)
0.880
(0.89)
N, countries
107, 86
87, 76
R-squared
0.28
0.27
• OLS, SEs corrected for clustering, p-values reported
• Observations: non-democratic episodes
• Source: Gehlbach and Keefer (2007)
Applications: Benin, Ghana, MENA
Benin, Ghana: CEMs concerned with growth. Key concern:
What are the incentives of political decision makers to
pursue growth?
MENA: Flagship on the private sector in the region. Key
concern: What are the incentives of political decision
makers to encourage private investment?
Benin – democratic “success”, growth
“failure”
0
.005
.01
.015
.02
Growth in Benin versus
African and all other democracies, 1992-2006
Benin
All democracies, 1992
African democracies,1992
Benin: Governance weak
-.8
-.6
-.4
-.2
0
.2
Governance in Benin versus all other democracies, 2006
All other democracies
Regulatory quality
Corruption
Benin
Rule of law
Gov. expenditures don’t increase human
capital
0
20
40
60
80
Education and public spending
Benin versus other democracies, 2005
All other democracies
Benin
Gross secondary school enrollment
Government expenditures/GDP, percent
Benin: low information, low credibility
African politics often characterized as “neopatrimonial” or
“clientelist”.
• More likely when citizen information low and no loss of
credibility from slow growth.
• Benin exhibits both of these traits.
Benin: Citizens uninformed
Afrobarometer question on whether fees are obligatory in
public primary schools, public health clinics.
• 65-75% answered correctly in 17 countries other than
Benin.
• 31-51% answered correctly in Benin.
Many media outlets, but . . .
• Under-capitalized.
• Little spent collecting news about government
policies.
• Governments are willing to apply pressure against
critical media.
Benin: parties not institutionalized
Parties not programmatic (2006, Database of Political
Institutions):
• 60 percent of biggest parties in all democracies are
programmatic.
• None of Benin’s parties.
Parties unstable (not institutionalized).
• ~100 parties disappeared in 2005, ~75 new parties in
2007.
• Youngest governing coalition (FCBE) won the most seats
in 2007; oldest (PRD) lost a seat.
Benin: Low citizen attachment to parties
Citizens indifferent between parties when parties not
programmatic/institutionalized.
66% of 1200 Benin respondents to the Afrobarometer
survey indicate they do not feel close to any party;
twice as many as in Ghana.
Politicians know parties don’t matter, so they switch often.
“I prefer to start my own party rather than accept a
subordinate role in the party of another.”
Consequence: 95% of Beninese respondents: politicians
rarely or never keep their promises (Ghana: 82%).
Benin: Presidents weakly accountable for
growth
Presidents are the key decision makers in Benin.
Absence of parties means campaigns based on individual,
not party characteristics.
Two consequences.
• Rely on clientelist payouts to voters and donations
from special interests: attenuates incentives to pursue
growth.
• Difficult to govern: Can only trust those vested in
president’s success since no institutionalized party.
Institutions exacerbate low accountability
Rules in Benin marginalize legislators.
• President has sole authority to propose budgets AND
(very unusual): de facto authority to declare budget
amendments by deputies to be invalid, b/c cost
implications insufficiently documented.
• President of the National Assembly almost entirely
controls agenda of Assembly and (very unusual) 2/3
vote necessary to remove him.
Deputies indicate little legislative oversight of budgets,
little interest/capacity in advancing policy agendas.
Citizens have no reason to hold deputies accountable for
growth failures if they’re essentially powerless to influence
policy.
Benin: policy implications
Institutional
• Increase vote threshold for parties
• Increase responsibility of National Assembly for budget
Information
• Increase collection and dissemination of information
about gov’t decisions.
◦ Start with budget – currently even the National
Assembly is poorly informed about budget
implementation.
• Make it easy for media to report on government
performance.
Improve the quality of education
• Current policy emphasizes quantity
Ghana: could do better
• Growth better than SSA average; signif. FDI; big reforms
w/r/t red tape, education, health.
• But:
• Growth far lower than East Asian countries at similar levels
of development.
• Governance no better or worse than average of all other
countries.
• Public spending not productive
• Secondary school enrollment 20 percentage points
lower than comparators.
• Public investment and government employment are
several percentage points of GDP higher than average;
• But public investment has not been growth-maximizing.
Puzzle: limited impact of competitive
elections (2000)
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
Puzzle: limited impact of competitive
elections (2000)
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
What explains limited policy responsiveness
to elections?
Again, political market imperfections:
• Uninformed citizens,
• Limited political credibility
• not fully (but more than in Benin) institutionalized
parties
Leading to:
• continuing preference for policies serving targeted
constituencies;
• slow improvement of public goods;
• continued significant rent-seeking
Ghana: Information
• No newspapers:
Afrobarometer 2005: 60% NEVER get news from
newspaper vs. 14% in South Africa
• Low education – and it matters:
◦ Up to 10 years of education (vast majority), 36
percent could identify Economic Recovery Program
(ERP);
◦ more than ten years of education: 66 percent.
(Afrobarometer 1999)
Ghana: Non-credible electoral promises
Few, stable parties in Ghana (in contrast to Benin)
• Less dependent on charismatic leaders; more party
discipline; members have more influence on party
leadership
But:
• No programmatic differences
◦ preference for the market similar between NDC and
NPP, their supporters.
◦ Broad public campaigns focus on competence/
corruption, not promises regarding future policies.
• Party and party policy stances irrelevant to unusually
large number of legislative campaigns.
• Enormous importance of personal handouts by
candidates, not parties.
Why is Ghana doing well?
• Despite strong political incentives to pursue clientelist
policies, Ghana is doing better than neighbors. Why?
◦ No dominant political party (not Senegal).
◦ Two main political parties are relatively
“institutionalized” (not Benin).
◦ Strong traditional restraints on arbitrary behavior by
key “patrons” – chiefs and elders.
◦ Continually falling risks of extra-institutional threats
to the regime.
Policy implications-info and education
Information
• Accelerate transparency reforms and the dissemination
of relevant information to citizens.
• More aggressive collection and dissemination of
government performance information
• Examples:
◦ test scores and class sizes by school
◦ mortality by hospital; etc.
◦ spending by community.
Quality and quantity of education
• Bias in the system towards construction and quantity
improvements.
• Need access expansion, but also dramatic quality
improvement.
MENA: What are the growth issues?
• By many measures, MENA and East Asia are similar.
• But private investment, exports, growth far lower in
MENA.
• Why?
• Overall political environment does not favor growth,
even if specific policy reforms are evident.
Governance not obviously the problem: WBI
Governance Index
-1
-.5
0
.5
1
WBI Governance Index
1996
1997
1998
1999
2000
2001
year
MENA Oil
No competitive elections
East Asian Countries
2002
2003
2004
2005
MENA Non-Oil
Richer Countries
ICRG Index – better in MENA, non-oil than in
East Asia
2.5
3
3.5
4
4.5
ICRG Rule of Law + Corruption
1996
1997
1998
1999
2000
2001
year
MENA Oil
No competitive elections
East Asian Countries
2002
2003
2004
2005
MENA N on-Oil
Richer Countries
Domestic credit to private sector high in non-oil
MENA (but smaller as fraction of total)
0
20
40
60
80
Domestic Credit to the Private Sector/GDP, 2005
MENA Oil
No competitive elections
East Asia
MEN A N on-oil
Middle Income
So why are MENA growth and exports low?
Incumbent firms favored
• credit access very concentrated.
• trade barriers much higher in MENA.
Incumbent workers favored
• labor protection much stiffer
Government can’t make credible commitments to new
entrants.
Trade barriers high in MENA
Overall Trade Restrictiveness, Manufacturing
(Tariff and NT Barriers)
Malaysia
Indonesia
China
Tunisia
Morocco
Jordan
Egypt
Algeria
0
0.1
0.2
0.3
Source: Kee, Nicita and Olarreaga 2006
0.4
0.5
0.6
0.7
0.8
Labor protection extends to public sector
• 1997 (Schiavo-Campo, et al) (old but best)
• Algeria, Bahrain, Egypt, Jordan, Lebanon, Morocco, Syria,
Tunisia, West-Bank Gaza and Yemen:
◦ government (central + non-central) employment = 3.3
percent of population.
◦ avg. govt. wage = 3.4 times GDP/capita.
• China
◦ government emp. = 1.7 percent of population.
◦ average wage = 1.3 times GDP/capita.
• Indonesia
◦ government empl. = 1.0 percent of population.
◦ average wage = 1.6 times GDP/capita.
Civil service is clue that growth is more
politically important in East Asia
• Merit matters more in East Asia.
• Singapore:
• Entrance to civil service rigorously meritocratic.
• 2006: 40% of average civil service and 50% of senior
civil service total compensation performance-based.
• Performance criteria are related to growth objectives.
• China:
• Meritocratic recruitment.
• Growth second only to “maintenance of order” as
promotion/bonus criterion (mayors/governors).
• Where does this occur in MENA? Perhaps Dubai.
Can’t write growth-promotion contracts
with civil servants in MENA
• Contracts with leaders not credible.
• MENA: few/no formal obstacles to arbitrary decisions
by rulers.
• East Asia: Ruling parties more institutionalized (e.g.,
leadership transitions are regularized – China, Malaysia)
and/or consequences of slow growth are high
(Singapore).
• Delegation to officials is key, but leaders unwilling.
• Delegation gives officials greater ability to overthrow
ruler.
• MENA: Centralized control is key.
• East Asia: Broad delegation.
Institutionalization in East Asia
China:
• Mao opposes CCP institutionalization.
• ruling party institutionalization in post-Mao period
• elimination of competing organization (Red Guard)
• creation of formal cadre evaluation system.
Indonesia:
• In 1969, Suharto reforms military:
• Joint command (reduces coordination costs)
• Increases delegation downwards (to very local
commanders).
MENA Contrast: military institutionalization
Sadaam Hussein replaced successful officers with
incompetent loyalists, established competing armed
forces,
• at substantial cost to military readiness and
• despite external threats. (Hashim 2003).
• Little delegation of authority to lower level
officers
• Little trust across units or between enlisted men
and officers.
Similar to observer claims about military organization
throughout the region (Egypt, Syria).
MENA parties less institutionalized
• Age of ruling party = institutionalization.
• Takes time to institutionalize; and
• Non-institutionalized parties disappear.
-20
0
20
40
60
Ruling Party Age - Leader Tenure, Non-democracies, 2004
MENA Oil
MENA Non-oil
Middle I ncome
Large East Asia
Large MENA Oil
N o compet itive elections
East Asia
Why the difference?
High rents
• oil
• foreign aid
Lack of intra-elite cohesion
• Competing clans rather than, as in China,
shared experience of “The Long March.”
• Abu Dhabi: of nine Amirs since 1818, five
murdered, two deposed.
• Sharjah: eight leaders since 1803, two
murdered, three deposed.
• Jordan: coup threats led King Hussein to “deinstitutionalize”.
How to maintain support with low
growth?
No institutionalization
Low growth, but discourages overthrow
Redistribute to key interests to forestall revolt.
• Spending on military: Large oil countries, 5.8%
GDP; 1.3% in large East Asia (2004).
• Gov. wage bill: Egypt 7.8%, Morocco 12%; East
Asia<6% (2005).
MENA: Reform implications-coordination
• Focus on inter-ministerial coordination.
• Ruler strategies discourage coordination.
• Philippe de Meneval discusses Bank project to
promote coordination in Morocco.
MENA: Reform implications-investment
promotion
• Pursue reforms that offer guarantees to investors but
don’t challenge leader aversion to institutionalization
• Reforms over limited geographic jurisdictions, such
as Export Processing Zones,
The civil service overseeing regulatory activities
can be transparently and meritocratically
recruited without increasing coup threat.
• Accelerated growth-oriented public investment.
◦ Value of inv. lost to leaders if they renege on
their commitments to investors;
◦ And infrastructure cannot “rebel.”