Whenever the indicator reveals a high degree of distortions in particular sectors (as compared with
other countries or periods) there is indication of a stronger need to carry out reforms.30
In the following analysis, indicators for labour and product market reforms are constructed on the
basis of structural indexes measuring the degree of policy-induced distortions used in IMF (2004),
while pension reform indicators are built on information collected and processed by the Rodolfo
de Benedetti Foundation (FRDB) reporting the year of adoption and the main characteristics of
reforms.
Table 1 describes the sources of the original data and the methodology followed for constructing
the reform indicators used in the analysis that follows. The indicators take value 1 in countries and
years in which reforms took place and zero otherwise. Indicators constructed in this way permit to
better compare results across different types of reforms starting from data representing different
type of information (indexes summarizing the degree of distortions in the economy for labour and
product market and tax reforms, and dicotomic variables reporting when and where reforms took
place, and with which characteristics, for pension reforms).31 These indicators also account for the
discrete character of reforms, i.e., the fact that reforms are generally not evenly spread across time
and space.32 The indicators constructed cover EU-14 countries (except Greece for what concerns
labour and product market reforms). Data are available starting from the ‘70s and up to late ‘90s or
early 2000 for product and labour market reforms and for the 1985-2001 period for pension
reforms.
Table 2 reports the frequency across the sample of the type of reforms considered distinguishing
between different decades. It shows that labour and product market reforms have been more
frequent in the ‘90s than they were in the ‘80s and especially in the ‘70s. As for pension reforms,
they were considerably more frequent in the ‘90s than in the ‘80s (information on the ‘70s is not
30 A further method for measuring reforms is the use of structural indicators providing information on the
functioning of the economy. For instance, in the case of the measurement of the functioning of the labour market,
this approach would imply using a number of indicators concerning the magnitude and the characteristics of
unemployment, job creation and job destruction flows, etc. This approach has been followed at the EU level to
measure the progress towards the goals of the Lisbon strategy. Progress is benchmarked against indicators measuring
outcomes achieved in specific sectors of the economy in EU Member States.
31 Reforms in labour and product markets correspond to changes in the structural indexes indicating a sufficiently big
reduction in the degree of policy distortions. A similar approach is followed, for instance, in Heinemann (2004). By
convention, it is assumed that reforms need to induce a reduction in the degree of distortion greater than the median
reduction observed across the sample. The choice of the median value as a benchmark allows an easy interpretation
(reforms are events leading to a reduction in the degree of distortion belonging to the top 50%) and implies a
frequency of events classified as reforms in the order of 20-to-30% of the total, which permits using statistical
inference in the analysis of the links between reforms and fiscal variables across the sample.
32 However, the use of discrete reform indicators has the drawback of not permitting to take into account the
different intensity of the impact of policies in different countries and periods, while this can be captured by using
directly indexes summarizing the extent of policy-induced distortions.
105