36
8.3.2 Income Convergence Between the CAP Regions
To answer the question whether there is a convergence of per capita income between all the
EU regions, the measures of central tendency and variation for each of the per capita income series
1990 - 1997 is calculated. These statistics are listed in Table 7 below. A measure for the convergence
of per capita income is the distribution of observations around the mean value of a variable. The wider
the distribution around the mean, the more dispersed and the more dissimilar the observations will be.
The narrower the distributions around the mean, the more similar are the observations. The most
widely used measures of distribution around the mean are the standard deviation and the variance of
the observations. Hence, the smaller the standard deviation, the smaller the variance, and the more
similar the numerical values of the observations will be.
[Figure 9]
The statistics in Table 7 provide some evidence for the convergence of per capita incomes
between the regions. First, both the standard deviation and the variance are decreasing in each of the
years under consideration. The change in the variance over the period 1990-1997 is -21.7%. Further
evidence of convergence is provided by the coefficient of variation that declines by 14.5%. Second,
from 1995 to 1997 the value of the interquartile range declined from 31.0% to 28.2%. The
interquartile range consists of 50% of the ordered observations of the variable. Since this value range
is declining over the years, the statistic suggests that the mid-range of values of 50% of the
observations have declined. This means that there are more observations within that 50% range with a
similar value, and that regional per capita income convergence has taken place. It does not indicate,
however, in which CAP regions the largest convergence has occurred.
To assess which of the CAP regions have contributed the most to the convergence of per
capita incomes, the measures of central tendency and variance are calculated for each of the cluster of
regions in the model. The changes in four of the summary statistics are presented in Table 8.
[Figure 10]
The core region shows an increase of 15.7% in the size of the interquartile range, indicating an
increase in the diversity of per capita incomes. This increase is offset by the substantial decline in the