Search results

Relative bipolarisation indices are usually constructed making sure that they achieve their minimum value of bipolarisation if and only if distributions are perfectly egalitarian. However, the literature has neglected discussing the existence of a benchmark of maximum relative bipolarisation. Consequently there is no discussion as to the implications of maximum bipolarisation for the optimal normalisation of relative bipolarisation indices either. In this note we characterize the situation of maximum relative bipolarisation as the only one consistent with the key axioms of relative bipolarisation. We illustrate the usefulness of incorporating the concept of maximum relative bipolarisation in the design of bipolarisation indices by identifying, among the family of rank-dependent Wang–Tsui indices, the only subclass fulfilling a normalisation axiom that takes into account both benchmarks of minimum and maximum relative bipolarisation.

The relative bipolarisation literature features examples of indices which depend on the median of the distribution, including the renowned Foster–Wolfson index. This study shows that the use of the median in the design and computation of relative bipolarisation indices is both unnecessary and problematic. It is unnecessary because we can rely on existing well-behaved, median-independent indices. It is problematic because, as the study shows, median-dependent indices violate the basic transfer axioms of bipolarisation (defining spread and clustering properties), except when the median is unaffected by the transfers. The convenience of discarding the median from index computations is further illustrated with a numerical example in which median-independent indices rank distributions according to the basic transfer axioms while median-dependent indices do not.

This chapter shows that the algorithm recently proposed to decompose the Foster and Wolfson bipolarization index by income sources (see Bárcena-Martin, Deutsch, & Silber, forthcoming) may be extended to break down wage bipolarization by its determinants. The chapter gives an empirical illustration comparing the determinants of wage bipolarization and inequality in various European countries in 2011, with a special focus on Portugal. In Portugal higher levels of education are the main source of bipolarization and inequality. Gender and working in the public sector are important determinants of bipolarization while age and having a temporary job are important determinants of inequality.

This chapter proposes a definition of pro-middle class growth derived from the approach of Lasso de la Vega, Urrutia, and Diez (2010) to intermediate polarization. The authors show that a sufficient condition for growth to be pro-middle class is for the growth rate of what we define as the “intermediate median income” of the whole population to be higher than that of the weighted average of the growth rates of the rich and smaller than the weighted average growth rate of the poor, the “rich” and the “poor” being respectively those with an income higher and lower than the median income. An empirical illustration based on Israeli data for the period 1995–2018 indicates that in absolute terms growth was not pro-middle class for any income type. In contrast, growth was pro-middle class in relative terms for all market incomes (individual income from salaried work, individual wage per hour worked, household economic income, total household income and total equivalized income). But growth was not pro-middle class for net income and net equivalized income, even in relative terms. These conclusions appear to be related to the combined effect of developments in labor force participation, welfare policy changes and major modifications in income tax rates. The intermediate polarization measures indicate that in general there was no pro-middle class growth except in the case of specific market income types.