on “institutional clusters” by jellema & roland emmanuel raynaud inra sad & atom (u....
DESCRIPTION
Background Set the priorities for institutional design Unbundle institutions and check their relative importance for economic growth (Acemoglu & al, 2005) Pb: subsets of institutions likely to be correlated multicollinearity issue Your move / contribution Construction of institutional clusters based on objective measures Disentangling legal, political & cultural institutions Avoid multicollinearity pb but other drawbacksTRANSCRIPT
On “Institutional Clusters” On “Institutional Clusters” by Jellema & Rolandby Jellema & Roland
Emmanuel RaynaudEmmanuel RaynaudINRA SAD & ATOM (U. Paris I)INRA SAD & ATOM (U. Paris I)
Background Background
Institutions matter for economic growth Need “good” institutions
What are “good” institutions? THE trade-offIndividuals are expected to install rules only after they have created mechanisms to control them (Barzel, 2004)A government strong enough to protect property rights and enforce contract is also strong enough to confiscate the wealth of its citizens (Weingast, 1995) The development of market-based exchanges relies on “contract-enforcement” institutions and “coercion-constraining” institutions (Greif, 2005)
BackgroundBackground
Set the priorities for institutional designUnbundle institutions and check their relative importance for economic growth (Acemoglu & al, 2005)Pb: subsets of institutions likely to be correlated multicollinearity issue
Your move / contributionConstruction of institutional clusters based on objective measuresDisentangling legal, political & cultural institutionsAvoid multicollinearity pb but other drawbacks
CommentsComments Describe in more details how one Describe in more details how one
empirical strategy overcome the empirical strategy overcome the weaknesses of the other.weaknesses of the other.
What is best? To have good summary What is best? To have good summary measure at the cost of potential measure at the cost of potential covariation or to have results free of covariation or to have results free of multicollinearity pb at the cost of unatural multicollinearity pb at the cost of unatural interpretation of (some) principal interpretation of (some) principal components? components?
Furthermore what is first in the Furthermore what is first in the introduction is second in section 3.introduction is second in section 3.
Food for thoughtFood for thought
From institutional clusters to “institutional From institutional clusters to “institutional profiles” (or institutional profiles” (or institutional systems) ) a first a first step toward empirically assessing step toward empirically assessing institutional institutional complementaritiescomplementarities à laà la Aoki? Note that the Aoki? Note that the relevant question is no longer the importance relevant question is no longer the importance of specific institutions for econ growth but the of specific institutions for econ growth but the measure of a “system or fit effect”measure of a “system or fit effect”If yes, a different paper?If yes, a different paper?Relevant analogy on this issue: Relevant analogy on this issue: complementarities among practices within complementarities among practices within firms (Ichniowski & al, 1997, AER, firms (Ichniowski & al, 1997, AER, the effects the effects of human resource management practices on of human resource management practices on productivity: a study of steel finishing linesproductivity: a study of steel finishing lines).).
Food for thoughtFood for thought
Policy implications / recommendationsPolicy implications / recommendations Simple LOT implication (for developing Simple LOT implication (for developing
countries): Shift to common law b/c CL bettercountries): Shift to common law b/c CL better Fit theory implication: Shift to CL will not Fit theory implication: Shift to CL will not
have the expected benefits (if any) unless have the expected benefits (if any) unless other institutions also change other institutions also change
We should compare alternative institutional We should compare alternative institutional systemssystems
At the same time, different institutions have At the same time, different institutions have different ability to change (in the short run). different ability to change (in the short run).
Slow vs fast moving institutions (Roland, 2004) (not necessarily easy to implement for empirical
works)Slow-moving institutions
Fast-moving institutions
Cultural clusters
Legal & political clusters
Food for thoughtFood for thought
Cultural clusters probably more Cultural clusters probably more difficult to change than the other two difficult to change than the other two (in the short run).(in the short run).
(Modified) Fit theory implication: (Modified) Fit theory implication: adapt political, legal cultural to each adapt political, legal cultural to each other other adapt fast-moving adapt fast-moving institutions (legal, political) to slow-institutions (legal, political) to slow-moving ones (cultural).moving ones (cultural).
Food for thoughtFood for thought
Example 1: Corsica
Formal law: law of the French republic (or of the colonial state depending on your preferences)Law in practice: Clan-based society with code of honor What drives local people’s behavior? Let you guess Formal law (legal regime) is not necessarily the “law” that people rely on in practice
Food for thoughtFood for thought
Example 2: even in developed Example 2: even in developed countries, national business sectors countries, national business sectors are riddle with “relational” are riddle with “relational” contracting and other “private contracting and other “private institutions” that set rules & related institutions” that set rules & related enforcement devices (see Bernstein, enforcement devices (see Bernstein, 1992, 1998, 2001). 1992, 1998, 2001).
How to measure this in international How to measure this in international comparisons? Case studies?comparisons? Case studies?
At lastAt last
Civil law tradition is Civil law tradition is at least good for at least good for
somethingsomething Mark Weist, 2002, Mark Weist, 2002,
“Legal “Legal determinants of determinants of
world cup world cup success”, U. success”, U.
Michigan Law Michigan Law School WP.School WP.
Institutions matters for economic growthInstitutions matters for economic growth Through which channels?Through which channels? Weaknesses of previous analysis: aggregate measures of institutions based Weaknesses of previous analysis: aggregate measures of institutions based
on subjective evaluationon subjective evaluation Which institutions are the most relevant for economic growth? Disentangle Which institutions are the most relevant for economic growth? Disentangle
different types of institutions but institutional variables are likely to be different types of institutions but institutional variables are likely to be correlated (multicollinearity issue). correlated (multicollinearity issue).
Paper make progress on these issues by using factor analysis: using Paper make progress on these issues by using factor analysis: using existing (objective) measures of specific institutions / tease out relevant existing (objective) measures of specific institutions / tease out relevant covariations / deal with multicollinearity covariations / deal with multicollinearity construct small number of construct small number of institutional clustersinstitutional clusters
Identify robust clusters of institutions (what robust mean?). Trade-off Identify robust clusters of institutions (what robust mean?). Trade-off between coverage & comprehensiveness. between coverage & comprehensiveness.
Two approaches were used: (i) compute principal components for three Two approaches were used: (i) compute principal components for three sets of institutions: political, judicial and cultural. Good synthetic measures sets of institutions: political, judicial and cultural. Good synthetic measures for each categories but potential multicollinearity among variables; (ii) for each categories but potential multicollinearity among variables; (ii) construct institutional clusters (combining previous categories of construct institutional clusters (combining previous categories of institutions) with objective measures of institutions avoiding institutions) with objective measures of institutions avoiding multicollinearity issues (creation of “institutional profiles”?). multicollinearity issues (creation of “institutional profiles”?).