Title: Composite Indicators for the Measurement of Economic Performance
1Composite Indicators for the Measurement of
Economic Performance
P. Roberti F. Oropallo ISTAT Productivity,
Competitiveness and the New Information
Economy Business , Systemic and Measurement
Issues NESIS FP5 ISTAT Rome June 26, 2003
2We live with too many Indicators
- A host of indicators that have been proposed to
monitor this goal. - They can be grouped in five areas
- General economic background
- Employment
- Economic Reform
- Social Cohesion
- Environment
- LISBON OBJECTIVES
- to become the most competitive and dynamic
knowledge-based economy in the world, capable of
sustainable economic growth ...and social
cohesion
3Indicators can beone-dimensional
ormulti-dimensional
- Composite indicators have received increasing
attention in recent years. - Various methodologies have been developed to
handle aggregation and related problems - Aggregation systems
- Deciding on the phenomenon to be measured
- Selection of sub-indicators
- Assessing the quality of the data
- Assessing the relationships between the
sub-indicators - Testing for Robustness and Sensitivity
4State of the Art Methodological Issues
- A number of methodologies can be applied for the
development of composite indicators. They
include - Multiple linear regression models
- Principal components analysis and factor
analysis - Cronbach alpha
- Neutralization of correlation effect
- Efficiency frontier
- Distance to targets
- Experts opinion (budget allocation)
- Public opinion
- Analytic Hierarchy Process
- JRC EC (2002) - State-of-the-art Report on
Current Methodologies and Practices for Composite
Indicator Development Joint research Centre
European Commission - Institute for the
Protection and Security of the Citizen
Technological and Economic Risk Management
I-21020 Ispra (VA) Italy - Prepared by the
Applied Statistics Group June 2002
5Is there a clear Framework yet?
- How can a framework be developed?
- By defining an analytical framework with precise
properties!
6Property of Indicators
- Micro founded
- Scope fulfilling
- Purpose oriented
- Well- behaved
- Consistent
- Decomposable
- Multidimensional - Composite
7MICRO FOUNDED
- that is, based on a comprehensive database that
embraces all aspects of enterprise features (ad
hoc surveys can cover some aspects) - Integration of different data sources of micro
data - Quality test of the integration process
(matching procedures / estimation) - In the second part of this presentation examples
of the opportunities opened up by an Integrated
Database are shown (Diecofis Project
www.istat.it/diecofis Year of reference is
1998-2000) - Sources are (1) Structural Business Statistics
(2) Administrative data (Foreign Trade,
Commercial account, Fiscal and Social security
data)
8SCOPE FULFILLING
- Indicators can measure size, change and
dispersion - Changes in an Indicators value can have
different causes and lead to different
conclusions depending on underlying combinations - Socio-economic phenomena have different and
complex dimensions. - Different indicators can serve different
purposes, i.e. measure - - heterogeneity/dispersion
- ?-performance (moving toward the mean)
- - overall systemic performance
- ß-performance (generalised move
upward/downward) - overall comparative performance (e.g. catching
up/lagging behind) - stratification (to evaluate differences in
systemic structures and whether they represent a
stratum - as with Yitzhakis decomposition)
9 PURPOSE ORIENTED
- Socio-economic phenomena may have many facets and
change can result from a combination of different
patterns - Appropriate indicators may be needed in different
circumstances
10WELL-BEHAVED AND CONSISTENT
- Indicators inconsistency may arise for different
reasons - Reference to condition of
- - Lorenz dominance (focusing on relative
differences) - - Pareto superiority (focusing on levels)
- - stochastic dominance (focusing on both
dimensions)
11DECOMPOSABLE MULTIDIMENSIONAL
- To be able to
- study patterns
- take into consideration more than one
dimension/aspects
12Developments in poverty analysis are a good
example of the possible problems and avenues to
solve them
Number of people ? Income gaps ? Welfare
dimensions ? Multidimensional aspects Headcoun
ts ? Gaps ? FGT indices (squared gaps, etc.) ?
Multidimensional indices
13Developing a similar framework is important for
the analysis of systemic performance and the
benchmarking of economic textures
- Since many factors and forces are at work to
determine, condition and produce different
outcomes. - The quote that follows can serve to grasp the
problem and possible approach to address it
14The fact that many of the smaller EU economies
do either better or worse than the larger ones is
partly due to larger EU economies contributing
more to the overall EU mean than smaller
economies, which means that they are less able to
diverge from the mean. A second explanation is
due to structural conditions. The industrial
distribution of small economies is often
concentrated in a few sectors, while larger
economies are more diverse. This can shift the
scores towards the mean for many indicators in
large economies, while small economies can
exhibit either a high or low innovative capacity,
depending on the sectors that dominate the
economy. Of course, this shift towards high or
low technology sectors is not accidental, but
reflects both public and private institutions
seeking out areas of comparative advantage and
high profitability. (EU Commission)
15The analytical framework and indicators that
can serve to study and benchmark different
levels/dimensions of systemic features
- Systemic maps (whole jigsaw)
- The overall aggregate picture EU level
- The national, regional, local picture
- The sectional picture
- The occupational picture
- Systemic strength and weakness, at a point in
time (cross-section analysis) and overtime
(longitudinal analysis) - Map transitions features, patterns and
evolution (New vs. Old) - Systemic change and its features, at the
aggregate/disaggregate levels
16Decomposable indicator how
- The GINI index measures the concentration of a
particular phenomenon (0no concentration,
100maximum concentration). It can be divided
into three elements -
17Distributions and Overlap
When overlap in the decomposition is high, it is
very hard to judge which group is the
best/worst, because distributions cross
18Distributions and Overlap
When overlap in the decomposition is low it is
easy to determine which group is the best/worst
performer
19Mono-dimensional Analysis of performance
(Overall)(GINI index calculated on exports)
20Mono-dimensional analysis of performance
(Exports) (By NACE Sectors)
21Multi-dimensionalanalysis of performance
- Decomposition for each dimension
Decomposition of the between component for each
dimension across K classes
22To make composite indicators easy to use and
interpret
Requires Aggregation
23Benchmarking performance levels or gaps?
- How far away from best performers enterprises
are?
24Weighted Composite Gap
- How far away from best performers enterprises
are?
25Composite Indicator (Breakdown by NACE sector)
- Three dimensions
? One Dimension
26NACE sectors
The Sectors in italic are defined ICT sectors
by OECD
27Composite Indicator (Breakdown by NACE sector)
- Three dimensions of enterprises performance
(1)Value Added (2)Employment (3)Exports -
- a) one dimension-three
areas
28Composite Indicator (Breakdown by NACE sector)
b) 3 dimensions into one indicator
Short fall from top or average
29Composite Indicator (Breakdown by regions)
Three dimensions of enterprises performance
(1)Value Added (2)Employment (3)Exports
a) 3 mono-dimensional
indicators
30Composite Indicator (Breakdown by regions)
b) from 3 to one 3-dimensional indicator
Short fall
31Composite Indicator (by size of firm)
- Three dimensions of enterprises performance
(1)Value Added (2)Employment (3)Exports
Short fall
32Conclusions
- The analysis which has been presented draws from
research results under two related FP5 projects
DIECOFIS and NESIS that deal with different
mapping and benchmarking aspects. - A strong investment in the design and development
of a complex and wide ranging system of
enterprise micro data which have been integrated
and systematised into one single hub. - The analysis is founded on micro-data drawn from
the integrated and systematised enterprise SIS,
which gives high flexibility and allows to
aggregate and disaggregate indicators a la carte.
33Thank You