Do institutions matter?
Institutions defined (again)
• Douglas North: ”Institutions are the humanly
devised constraints that structure human
interaction. They are made up of formal
constraints1, informal constraints2 and their
enforcement characteristics. Together they define
the incentive structure of societies and
specificially economies”.
– 1) rules, laws, constitutions
– 2) norms, behavior, conventions
• … or simply ’the rules of the game’
Institutions defined (again) cont.
• Can also be useful to distinguish between:
– Economic institutions (individual property rights,
contracts that can be written and enforced, patent
laws etc.)
– Political institutions (democracy vs nondemocracy, electoral rules, extent of checks and
balances etc.)
• Large literature documenting positive correlations between
certain institutions and economic performance.
The identification problem
• But this hardly demonstrates a causal effect of institutions on economic
– The kind of question we are interested in answering is:
”… if the UK were to switch its electoral rule from majoritarian to proportional,
how would this affect the size of its welfare state or its budget deficitis?” (PT)
”If Argentina were to abondon its presidential regime in favor of a parliamentary
form of government, would this facilitate the adoption of sound policy towards
economic development?” (PT).
How would changing institutions in Nigeria to those of Chile affect economic
performance? (AJR)
How would incresing the ’social infrastructure’ of Zaire to that of Switzerland affect
productivity? (HJ)
How did the introduction of universal suffrage affect redistribution levels in western
Europe? (AR)
Does inequality cause underdevelopment? (Easterly)
• Hard questions to answer empirically
• Institutions are endogenous to economic performance.
– Causality runs both ways
– Omitted variable bias
• Fundamental problem: cannot observe the counterfactual
The identification problem cont.
Figure from Przeworski04 (’institutions matter?’)
The identification problem cont.
• What we can learn from correlations and OLS
regressions is limited.
• Acemoglu (2005) makes this argument
forcefully in a review of Persson and Tabellini
• Agents understand that different policies will map
into different outcomes
– Policy endogeneity
– ”If state policy making is purposeful action, responsive to
economic and political conditions within the state, then it
may be necessary to identify and control for the forces
that lead policies to change if one wishes to obtain
unbiased estimates of a policy’s incidence” (Besley and
Case 2000,EJ)
• Agents understand that different institutions will
map into different policies and outcomes.
– Endogeneity of institutions
Endogeneity cont.
• Large literature treat institutions as exogenous (e.g.
institutions are ’predetermined’ or ’given by history’)
• However, the political economy approach suggests
that the same factors that make policies unappealing
to treat as exogenous is relevant for institutions.
– E.g. The introduction of democracy was not ’random’
• Papers that try to deal with this problem includes
Mauro (1995), Hall and Jones (1999), Acemoglu,
Johnson and Robinson (2001) and Persson and
Tabellini (2003)
Proximate vs fundamental causes
• Vast differences in prosperity across countries
– Output per worker 32 times higher in top five countries relative to to
bottom five countries (HJ).
• What can explain this difference?
• 1) Physical capital differences
– Poor countries don’t save enough
• 2) Human capital differences
– Poor countries don’t invest enough in education
• 3) ’Technology’ differences
– Poor countries don’t invest enough in R&D and technology adaption, and fail to
organize production efficiently
• 1 – 3 are however proximate causes of differences in
• What are the fundamental causes?
• How do we distinguish between the effects of institution from
the conditions that give rise to them?
• Jared Diamond
– Proximate causes: guns, germs and steel
– Fundamental causes: geography
• Availability of crops and animals
• Axes of communication across continents
– Geographical differences have determined the
timing and nature of settled agriculture. This have
shaped societies ability to develop complex
Geography cont.
Hall and Jones argue:
”By social infrastrcture we mean the institutions and government policies that
provide the incentives for individuals and firms in an economy”.
– The incentives can encourage productive or predatory behavior…
Institutions are not exogenous but there are potential sources of exogenous
variation in history.
– European colonization as a ’natural experiment’.
– Use history to estimate the causal effect of institutions on growth.
HJ use Western European influence to solve the endogeneity problem
– Distance from equator (latitude), which is correlated with ’western influence’ (good)
– The extent of euriopean languages spoken today in these countries.
These instruments are not entirely convincing (see the disc. In Acemoglu 05JEL).
We focus on the work of Acemoglu, Johnson and Robinson.
AJR approach rests on three premises
1. Different types of colonization policies created different
sets of institutions.
’Extractive states’ e.g. Congo, Nigeria, Tunisia
No protection of private property
No checks and balances against government.
Confiscatory taxation
Neo-Europes, i.e. US, NZ, AUS.
Protection of private property
Law and order
2. Feasibility of settlements affected the probability of
different sets of institutions.
Do not argue that Eur. Influence in itself was positive or negative
3. Early institutions affect current institutions
AJR approach rests on three premises
Settler mortality
• In the tropics, 80 percent of European deaths were caused by
malaria and yellow fever.
– Europeans in Africa, India and Caribbean faced very high death rates.
– But death rates for the adult local population were much lower
(developed immunity).
– Yellow fever in W. Africa was ’a strangers disease’.
– … and had little effect on the health and economy of indigenous
• Many of the areas that were colonized in the tropical zone
were richer and more densely populated in 1500 than the
temperate areas later settled by Europeans.
Current institutions
• Protection from expropriation risks.
– Vary from 0 to 10, for each country for each year.
– AJR use average over 1985-1995
– Extractive state  low value on this index.
• Broad measure of property rights.
• Data from Political Risk Services
– Private company which assess the risk that foreign
investment will be expropriated in different
The equation of interest
Plain OLS estimates of eq (1) find a strong
assocation between R and Y. Also when controlling
for latitude and continent dummies.
Why doesn’t this necessarily show a causal effect?
1. Rich economies may prefer (or afford) better institutions
(the causality runs the other way)
2. Omitted variables
3. Measure of institutions created ex post
4. Attenuation bias (due to poor institution measure)
Does this suggest an upward or downward bias in
• Solution, treat R as an endogenous variable.
• The first stage:
• What is the identifying assumption?
IV cont.
• IV estimates suggest a large and statistically
significant effect of institution on economic
– Differences in institutions account for over 75% of the
variation in income per capita today
– Improving the institutions of Nigeria to the level of Chile
would lead to a 7-fold increase in Nigeria’s income.
• (in practice the difference is 11-fold)
– No impact of geography, when institutions are treated as
an endogenous variable
• ”These results suggest that Africa is poorer than the rest of the
world not because of pure geographic or cultural factors, but
because of worse institutions” (p. 1372 AJR).
IV cont.
• Results robust to:
– Different subsamples
– Controlling for continent dummies, geography.
– Controlling for current prevalence of malaria and
life expectancy
– Only using yellow fever as an instrument
(eradicated today)
Conditions in the colonies
• A key aspect of AJR is that it is not the identity of the
colonizer that matters, but conditions in the colonies.
– Clearly depart from HJ where European influence in itself
is deemed beneficial
• Related to the work of Engerman and Sokolof (1997).
– The Caribbean islands illustrate the adverse affects of
Europeans, which set up repressive regimes based on
slavery and forced labor.
– Factor endowments such as geography, climate and soil
conditions help explain the builings up of ’good
institutions’ in nortern am. And ’poor institutions’ in Lat.
From ES 2000 JEP
North am. Of relatively ”marginal econ.
Interest compared with the extraordinary
opportunities available in the Caribbean
and Lat. am” (ES2000)
Haiti probably richest country in the world
(on per cap basis) in 1790
Why did US/CAN experience sustained
econ. Growth in 18th and early 19th
(while others did not attain this goal until
late 19th or 20th century, if ever…)
Factor endowments
• Some econ. Historians have explained the divergence due to
the identity of the colonizer.
– But striking differences within the identity of the colonizer (e.g. US,
Can vs. Barbados, Jamaica, e.g. Argentina vs Peru) calls for other
• AJR explain the divergence as a result of institutions.
• ES: Factor endowments (incl. Climate, soil, density of native
– Predisposed North am. Colonies towards relatively equal distributions
and corresponding institutions favoring a broad range of the
population in commercial activity.
– Predisposed Lat. Am, Caribbean colonies to highly unequeal
distributions and institutions that protected the elite.
• The suitability for cultivating sugar and other highly valued commodities
 economics of scale  extensive use of slaves, and the densely populated native population.
• ”In those societies that began with extreme
inequality, elites were better able to establish
a legal framework that insured them
disproportionate shares of political power, and
to use that greater influence to establish rules,
laws, and other government policies that
advantaged members of the elite relative to
non-members – contributing to persistence
over time of the high degree of
Inequality and per capita income
• A recent paper by Easterly (JDE 2007) draws
on ES to try to answer the question: does
inequality cause underdevelopment?
Inequality and per capita income
• Measures of factor endowments as
instruments for ’structural inequality’.
Sugarcane  labor intensive  high historical ineq.
Wheat  capital intensive  low historical ineq.
• What is the identifying assumption?
• Do we believe in this?
Acemoglu's review essay of PT:
P&T's book:
Besley and Case00:
Jared Diamonds book 'Guns, germs and steel‘:
Engerman and Sokoloff 2000:
Links to Hall and Jones, Acemoglu, Johnson and Robinson, Mauro, Engerman and Sokoloff97 are on the reading list

Does institutions matter?