Abstract. The pessimistic flavour of the Human Development Reports appears to be in contradiction with their own numbers as developing countries fare comparatively better in human development than in per capita GDP terms. This paper attempts to bridge this gap by providing a new, 'improved' human development index (IHDI), informed by welfare economics. The IHDI is presented here alongside the United Nations Development Programme's (UNDP) HDI for the world and its main regions since the late 19th century. Social dimensions in the IHDI are derived, following Kakwani (Journal of Development Economics 41 (1993), pp. 307-336), with a convex achievement function, whereas a geometric average is employed to combine its dimensions (longevity, knowledge and income). Thus, the IHDI does not conceal the gap between rich and poor countries and casts a much less optimistic view than the conventional UNDP index, while it fits with the UNDP concern for international differences. The paper's findings highlight main weaknesses in human development dimensions of present-day developing countries.Keywords. Education; Human development; Life expectancy; Per capita GDPs Readers of the Human Development Report (HDR), published periodically by the United Nations Development Programme (UNDP), tend to have mixed feelings. The pessimistic rhetoric of the HDR seems to be contradicted by its own numbers. In fact, when weighed up in human development terms, developing countries tend to fare better relative to advanced countries than in per capita income terms. It is the purpose of this essay to bridge the gap between the empirical evidence on human development and the HDR rhetoric by providing a new, 'improved' index (IHDI, hereafter), informed by welfare economics. 1 The IHDI is presented here along the UNDP's HDI (UNHDI, henceforth) for the world and its main regions over the period 1870-2005. What defines the new human development index? In the first place, its social, non-income dimensions are derived using a convex achievement function as an alternative to the linear transformation employed in the UNHDI. Thus, in the new index, as a social indicator reaches higher levels, its increases represent higher achievements than if the same increase would take place at a lower level, whereas in the UNDP linear transformation the same change results regardless of the starting level. Second, in an attempt to reduce substitutability among the index components, 1 ESCOSURA its three dimensions (longevity, access to knowledge and average incomes) are combined into the new IHDI using a geometric average, rather than the arithmetic average used in the UNHDI. The final outcome is a new human development index which, by not concealing the gap between rich and poor countries, casts a much less optimistic view than the one provided by conventional UNDP index while satisfying the HDR concern for international differences.The paper is organized in four sections. Section 1 assesses the UNHDI and exposes its main shortcomings. In an attempt to provide a resp...
