N - OECD

Comments on STRI Project
OECD Experts Meeting on STRI,
OECD, 2-3 July 2009
Joaquim Oliveira Martins
OECD
Motivation
• The construction of policy/reform/institutional indicators has
developed in economic literature motivated by empirical research on
the determinants of economic performance and concern to develop a
culture of evidence-based decision making.
• The OECD has contributed largely to this research effort. Some
OECD institutional indicators, such as the Product Market
Restrictions (PMR) or the Employment Protection Legislation (EPL),
have become a basic tool-kit for economists.
• The STRI project is in line of this work. Right focus on ex-ante
indicators. It developed a sound analytical approach. It has used the
network of OECD experts to incorporate policy-relevant information.
Main points
• Method
– The question of weighting & robustness checks
• Relating indicators to performance (Top-down
approach)
– Interactions of policies
– Market structure
• Possible uses
– Research & Policy analysis
– Peer pressure vs. Trade negotiations
1. Method
•
Composite indicators always contain a certain
degree of arbitrariness and subjectivity. STRI
achieves a good compromise
•
The problem of weighting:
–
•
Mixed approach (EW+EJ): this is reasonable as equal
weights are not neutral and their effect depend a lot
from the nesting structure; PCA is often not accepted
Random-weights' approach requires caution
–
–
–
Confidence intervals are biased & asymmetric (m=1/N)
Apparent "robustness" may reveal problems with
scoring
Weak discrimination across countries (but STRI may
be better than PMR)
Aggregate PMR scores, 2008
Source: Wölfl, Wanner, Kozluk and Nicoletti (2009)
2. Relating indicators to Performance
•
Policy complementarities:
–
–
–
When policies are mutually interdependent, changing
one without changing another may not materialize
expected benefits
Link with 2nd Best
ECO is integrating this dimension in the PMR
indicator
 But complementarity complicates policy
evaluation because performance of one
reform area may be closely related to
progress in other policies
Policy complementarities: restrictions
often reinforce each other
Figure 4. Product market regulation and employment protection legislation 1
Employment protection legislation
4.0
Correlation 0.73
t-statistic 4.72
Portugal
Greece
3.5
3
rd
cluster
Italy
Spain
France
3.0
Norway
Germany
Japan
2.5
Sweden
2
nd
Austria
Netherlands
cluster
Finland
2.0
1.5
Belgium
Denmark
Switzerland
Ireland
1.0
Australia
New Zealand
Canada
United Kingdom
0.5
st
1 cluster
United States
0.0
0.0
0.5
1.0
1.5
2.0
2.5
Product market regulation
Source:
(2005)
1. The
scale ofNicoletti
indicators&isScarpetta
0-6 from least
to most restrictive
Relating indicators to Performance
• Non-linearities: Laffer-curve effects &
Imperfect competition
Outcome
Performance
Examples:
-Tax policy
- Optimal tariffs
- Strategic Trade policy
- Competition vs. Innovation
(Aghion & Howitt model)
Policy
Relating indicators to Performance
•
Sectors are characterised by different types of
competition/market structure that could be
brought into the picture
 The effect of restrictions is likely to be different when
output growth (Q=N.X) takes place through the
extensive margin (more firms or products, N) or the
intensive margin (more output per firm or product, X)
(cf. Sutton’s taxonomy of fragmented vs. segmented
markets)
 Idem whether competition is mainly on costs/prices or
product differentiation/innovation
Relating indicators to Performance
•
Modes of trade in services and STRI
–
Use models of vertical trade to explain why
depending on threshold levels of STRI there could be
a shift across modes (e.g. 1 to 3)
3. Possible uses
•
Parallel with the PMR indicator
–
–
–
•
The PMR indicator has been extensively used for
academic research
Could stimulate production of new & better data
Peer-pressure: there has been a movement towards
decreasing product market restrictions in OECD
countries mapped by the PMR
Could be difficult using the STRI for trade
negotiations
–
–
Problem of assessing significance of ΔSTRI
Analysing interaction of regulations could be useful
to a more systemic view of policies
Thank you!