Časopis pro politiku a mezinárodní vztahy

Global Politics

Časopis pro politiku a mezinárodní vztahy

Human Development Index: How to cope with its limitations?

The Human Development Index (HDI) is used to measure quality of life in countries across the world. This index has been published – on regular basis – since 1990, however its contribution and ability of covering up as sliding concept as human development in its own scope has been still highly discussed. Main aim of this paper is to focus at limitations of Human Development Index and try to answer questions, concerning how deeply this measurement is able to reflect the reality in particular countries. I will discuss the usefulness of the conceptual framework of “human development” and illustrate the significant sensitivity of this measurement and argue its limited insights.


1 Introduction

The Human Development Index (HDI) is used for measuring quality of life in countries across the world. When in 1990 United Nations Development Program (UNDP) issued a first Human Development Report (HDR) – which has been published on a regular basis every year – the most discussed part at that time was the indicator dealing with human development of particular countries in the world, which – logically and worth fully – received a great attention and encouraged many debates. By then well-being of a population has often been measured in terms of income, i.e. for more than four decades the indicators most used for analyzing development of countries were based on economic efficiency in principle as mainly gross national product per capita (GNP/N). However, very often has been insisted on the inadequacy of income as the sole indicator of welfare and augmented, that measuring income is losing its utility, becoming more puzzling and contributes only insignificantly to human development.

Main goal of this paper is not to participate in debate about this “shift”, instead it will focus at contemporary limitations of Human Development Index (HDI) and try to answer questions concerning how deep is this measuring enable to reflect the reality in particular countries. Also I will discuss the usefulness of the conceptual framework of “human development” and illustrate the significant sensitivity of the measuring approach and argue that it offers only limited insights.

In first part my main aim is to deal with used methodology for calculating HDI and subsequently illustrate its consequences and limitations at some specific examples. However it is important not only question how to deal with input data and transform them in the final HDI, but at the same time it is needful to analyze and realize how to deal with the final interpretation and what the limits are of year-on-year comparison. This issue will be covered with the second part. Final third part will focus at questions, how far contemporary dimensions are used for calculating HDI and enable to reflect reality and, at the same time, what role should human freedom, human rights and environmental sustainability play in its calculation, if any.

2 Applied methods and consequences of HDI

The HDI indicates level of achievement in living standard of a population in terms of attainment levels of different quality-of-live attributes (Mazumdar 2003: 537, Weiss 2008: 208), i.e. reflects three major dimensions of human live: (1) Life expectancy at birth, as an index of population health and longevity; (2) knowledge and education, as measured by the adult literacy rate (with two-thirds weighting) and the combined primary, secondary, and tertiary gross enrolment ratio (with one-third weighting); (3) standard of living, as measured by the natural logarithm of gross domestic product per capita at purchasing power parity.

All in all, within existence, the development of measuring human development had to undergo many changes. Until 1994, the minimum and maximum values were not absolutely fixed but were derived from the minimally and maximally achieved values worldwide in each year. Indeed this change is not fully visible at a first sign but it had far-reaching consequences. Because the result of ranking countries is changing from year to year, is obvious, that at the same time the minimum and maximum values were always changed. Changing in a maximal and minimal level doesn’t facilitate a year-on-year comparison and not automatically reflects changes which were achieved in particular countries.1 So far, this situation has been changed and cleared in 1994, when UNDP employed a new minimal and maximal threshold for different indicators.

Table no. 1: The values set for different indicators employed by the UNDP in 1994:

  Min Max
Life expectancy 25 years 85 years
Adult literacy 0 % 100 %
Years of schooling 0 % 15 years
Combined enrolment ratio 0 % 100 %
Per capita income PPP $200 PPP$40,000

Source: Mazumdar 2003: 538

Nevertheless, limits of HDI results calculation and consequential interpretation have not been fully diminished; it can be shown at example of USA and Japan, where different levels of different dimension can have – at the end – the same results. Both states are very close; Japan is ranked at the 10th position, USA is three positions below, at the 13th position, nevertheless both states have got up to those positions in a highly different ways. (Syrovatka 2008: 16). For Japan, very specific is high expectancy of live 82.2 years (1. position in total ranking), but on the other hand relatively low level of GNP/N (18. Position in total ranking), whereas USA is it clear opposite – life expectancy is there relatively low, with 75,5 years (30. – 31. in total ranking), compared with Japan, but has very high level of GNP/N (2. in total ranking).

What does this situation at those two examples illustrate? First of all, that difference of almost 5 years in life expectancy is significant, but at the same time, difference in GNP/N is also highly distinctive feature. But what in this situation under those circumstances should we consider as more predicative feature and what should we consider as more important for a human development? The conflict is evident because everyone has different preferences. The problem of HDI is that it allows a mutual substitutability within all three dimensions which are used and that similar countries which are very close in a ranking they can have significantly different development in a particular dimensions in a close look.

The HDI, as a combination of only four relatively simple indicators, doesn’t only raise a questions what other indicators should be included, but also how to ensure quality and comparable input data. It is logical that the UNDP try to collect their data from international organizations concentrating in collecting data in specific fields, nevertheless, as Syrovatka (2008: 18) noticed, we should not forget, that “all those organizations have to collect their data somewhere as well”. And if they don’t do own survey, they use information gained primarily in national statistic sources. But in this situation, quality and trustworthiness of those data is disputable, especially when we get the information from UN non-democratic members, as for example Cuba or China. Do they really provide adequate data or do they just follow they own interest for strengthening their credibility and legitimacy in the eyes of the western societies?

So far none of the statistical approaches is without difficulties in gathering as authentic data as possible. Nevertheless, in HDI this situation is visible especially in a dimension of literacy of adults, which is not conducted every year, and as HDI mentioned “…comparisons across countries and over time should be made with caution.” (HDI 2009: 184) The adult literacy rate is based on data estimated from censuses or surveys conducted within a period 1999 – 2007. The problem is that those indices of adult literacy can stay for many years unchangeable if government undergoes the population census only once in a while, every few years. In light of that fact, when the population census is not conducted regularly in each country, it is difficult to do a year-on-year comparison (Syrovatka 2008: 18 – 19). Moreover, dimension of literacy in HDI represents 20 %, so it represents a significant part which cannot be simply overlooked. All in all, cross-national comparisons generally draw upon official national census estimates that are not obtain through direct testing of literacy skills in a population, however every country has more or less own methodology based on: (1) who is classified as literate person, (2) who they consider as an adult person, and (2) how frequently those surveys are obtained.

Subsequently most of high-developed countries have already reached (theoretically) almost 100% literacy within adult population. As a result, many of them are not even publishing their statistics. In those cases the purpose of calculating the HDI a value of 99.0% is applied automatically. For example, within first twenty states in HDR 2009, only two states have not got “full literacy” – Spain and Italy at 15th and 18th position, although both countries are practically only percent point under the 99% level. In this situation, where the HDI of most developed countries are so close, indeed so close to each other, Lind (2004: 293) summing up, that in a case of highly-developed countries “…even a fraction on one per cent error renders the HDI ranking meaningless.”

What consequences does this situation have? Most significantly for many highly developed countries measuring of their literacy has lost value, because many of them are practically at the same level. In this situation, we can assume, that distinguishing for example between Norway, Australia or Iceland in an adult literacy basically doesn’t make any sense, because methodology used for HDR is not able to differentiate between them, even if the input data would be available.

Evidently, measuring literacy presents particular difficulties. Lind (2004: 292) made the same conclusion, when he claims that literacy cannot be measured reliably. Neither, if measured, we can compare between writing systems as different as Chinese and English. Literacy, by his words, “…is quickly brought close to 100% in the process of developing countries, but cannot be increased further regardless on the afford (…) literacy becomes a dummy variable for most countries, ranging from 90% to 99%.” It is true, that in a HDI those difficulties are explicitly mentioned and due to those differences all comparisons should be made with caution. But still, for results of HDI it has a significant negative effect and sets up a question, whether HDI should be measured every single year or published less frequently.

But this situation is not connected only specifically with adults literacy, but is a problem of all HDI dimensions. One country we can consider as an exception in a field, what are a limits of whole ranking in a cases of highly developed countries. In last report (HDR 2009) Luxemburg is ranked as no. 11, which at first sight does not show any significant exception. But when we have a look at the dimension of estimated earned income, Luxemburg is far beyond a maximum level (40 000 PPP USD) and because of that, any other shifts of the indicator would not bring any points. Subsequently, life expectancy at birth is very high and has a stabile growth. For third dimension – adult literacy rate – is used a 99.0% value automatically. In this situation, three indicators have lost their utility. When we put all four indicators together, only one which has some predicative value is an enrolment ratio in education. This example, or we can use exception, shows a limitation of measuring HDI in top-developed countries, where indicators are not able to distinguish among each other.

3 Comparison in time

As was already mentioned above, HDR has been issued every year since 1990. Since the first issue was published, methodology has been changed many times, especially within first years. Although the essence of the HDI calculation has remained constant, the details have changed significantly since the first HDR in 1990. In this situation, cross-national comparing base on HDR across different years is highly problematic and is legitimated questioning of UN member countries how to adequately explain their ranking every year, because at the end, how would a country be able to know how much of its move up in ranking is due to genuine progress in “development” or just a relative result towards changes in the way the figure has been calculated? The changes within methodology refer to all countries, so those “shifts” in absolute numbers are connected to all of them, nevertheless brings up misunderstanding and difficulties, when particular countries try to follow link of their own progress within a time.

As a very interesting correlation between used methodologies in particular years and changes in ranking of the countries is study by Morse (2003). In his study attained to conclusion, that within changes of methodology which were applied by UNDP, the movement on the list can be caused not only by a real development, but very easily by a change in used methodology as well. By his recalculation of HDI, grounded on various methodologies employed by the UNDP, the outcome resulted in substantial moves on the scale; moving between ±10–15 ranks. This recognition is not an unexpected eye-opener, we can assume, however Morse proclaims, i.e. that within this situation it is not possible to do a year-on-year and over countries comparisons without absence of continuity in indicator methodology. Subsequently he claims, that if a year-on-year comparisons should be taken after all, it is better to do that in a periods of relative time-consistency in the HDI calculation, which he calls “islands of relative calm”2 (Morse 2003, Syrovatka 2003: 20). Nevertheless, even in those “islands of relative calm”, the question is still up-to-date, because not only a methodology has a big influence in an absolute position of countries, but also a position of countries depends on ranking of other countries, when interpreting results.

All in all, it implies a question, if having results in a “league-of-nations way” of all the UN members (for which them were data available) is more useful than having only a relative “league-of-nations”, which would serve only for a comparison in particular geographical regions. To discuss this notion, it would be a mistake to forget about the original background, which has been set up by Mahbub Ul Haq, Amartya Sen and Paul Streeten, for helping answer the question. All of them were realising, that the whole concept is disputable. By Streeten, “such indexes are useful in focusing attention and simplifying problems. They are eye-catching. They have considerable political appeal.” (Mcneill 2007: 7). At the same time, they were fully conscious, that the concept of HDI was set up mainly for relative comparison of countries in one particular time. Therefore, the original notion was not to set up an absolute ranking, but let’s quite free hands in comparison of the results. But if we really accept the relativity of the whole concept, would not be better to use the second approach, which would probably bring more trustworthiness in to the whole ranking? As we can see (and as was already illustrated above), contemporary HDI is much better when distinguishing between countries with low and middle human development, instead of countries at the top of the ranking. Moreover, what we should not forget, that the countries from a top, as for example Switzerland, can emphasise a highly different goals. In Switzerland’s po­licy, higher levels of education are probably more emphasised, than dealing with illiteracy, which would be probably the main goal for many countries in Sub-Saharan Africa.

The truth is also, that within the last few years, when the methodology settled down, there have not been many movements, which could be seen in a first half of 90s. Because of that, every move on a HDI scale can be seen as relatively important. Moreover, every small movement year-on-year of particular country towards up or down is presented on a media as a big step and if we accept an instability and incomprehensiveness by Morse (2003) – emerging from a methodology employed (at that time) – every move on scale causes loose of its values at the end.

Nevertheless, this necessary does not have to be linked only with methodological difficulties and could have simple explanation. Information, that a country has worsen its human development and has fallen down (from for example 22 to 24) can have very simple explanation – the other countries, which were at the same level have just improved in their human development and just moved few positions up. Further, it can even mean in results, that particular country has improved in human development as well, but was not simply able to compete with the other countries which have moved up. This situation shows, besides the others, that the whole conception is still highly relative and is difficult to clearly and distinctly interpret its results.

4 What dimensions should be included?

HDR has been always disputable and has caught the public-eye, whenever it was published. It has many reasons. One of them is, “that the concept of human development is much deeper and richer than what can be caught in any index or set of indicators,” as Streetsen (1994: 232) mentioned. However, even when we will accept this, we can also agree with Syrovatka (2008), that all three dimensions used for calculation HDI would be approximately in a population approved as indicators of human development. Nevertheless, question is whether there are any other indicators next to them, which are also important for human development.

Very often there is a mention of human freedom and human rights. Both are discussed in first HDR 1990, where it is written, that “human development is incomplete without human freedom” and that “while the need for qualities judgement is clear, there is no simple quantitative measure available yet to capture the many aspects of human freedom.” (HDR 1990: 16). Those words were written in 1990 and even through the proclamations and debates within 90s, how to integrate other dimension in HDI has been never expanded. This situation has far-reaching consequences. How can we really compare countries like China with other democratic countries even through the fact that in many indications show they are on the same level like fully-democratic countries? Streetsen (1994: 236) accurately underlined, that “… quite high level of HDI could be achieved even in a prison, if the prisoner would live long life and had and access to the library.” Other scholars offering to go beyond those three basic indicators and suggest other plausible candidates as indicators of these categories, e. g. (Ranis, G – Steward, F. – Samman, E.: 2006) encompass 11 broad categories of human development based on different philosophical approaches and justifications towards a defining of a “good life”;3 Noorbakhsh (2008), likewise, suggesting a modification in human development index, by removing the same weights of its components in final calculation and accentuate role of national income. Sakiko (2001), on the other hand observing, that even if human rights and human development share common concerns and ends, it does not necessarily follow that they are identical in concepts and strategy with economical and social rights. In his words, “they overlap but are different.” (2001: 240)

Other discussed dimension is a resource exploitation and environmental degradation. Should be HDI “greener”? This question is answered by Neumayer (2001) in his recommendation, that resource input and pollution should be taken more seriously. This suggestion, however, how to integrate this dimension into calculation of HDI seems even more difficult than with human freedom and human rights. Neumayer however suggests not including these aspects directly to the HDI, but rather they should be assigned to a role of a fundamental qualification if a country’s achieved human development level appears to be unsustainable­.4 As a result, country would be considered by Neumayer as a sustainable or prospectively unsustainable and this information would be put as additional information to each country in HDI.

Scho Sagar – Najam (1998) are also suitability into account, when asking a question, which – by them – need to be asked: “Human Development, but at what cost?” An example, of how distribution of environmental performance of countries varies greatly, is shown on countries such as Brazil and Indonesia. Both improved their performance on the HDI in part of converting their natural capital to income. While the human development achievements of these countries may seem – at the first glance – impressive, the HDI indicators are not able to cover the sustainability of this development.5

5 Conclusion

So far, one more time there is a question, whether results of HDI should be presented in a “league-of-nations” or only as “islands of similar countries”, if we stay close to Morse’s termi­nology. Many examples above already showed, that whole concept of HDI has significant limits. One possible way of development of the whole concept to the future could be a separation of countries from HDR to at least three groups and use a different methodology for each different group. It is predictable, however, that with this step the whole concept would loose something from its simplicity and its “eye-catching abilities” for media and public in general, but at the end, what asset do we have from a measuring literacy in top twenty countries, when they get automatically 99%? How to deal with countries, as was shown in Luxemburg, where three of four dimensions have basically lost their value? What the whole concept tell us, when is able to distinguish between low and middle-developed countries, but not within a highly-developed ones?

But even if the HDI concept would be separated into three groups and would change its methodology directly to specifications of every group of countries, there will be lack of reflecting other necessary dimension of human development. As was shown above, human rights and human freedom are integral part of human development, but highly questionable how to reflect them into a mathematical concept. Similar situation is with environmental exploitation and pollution – both can show relevant information, about trends in particular countries – but facing the same obstacle. How to reflect them when calculating HDI?

Based on works of many scholars mentioned above, we can assume that efforts for integrating new dimensions reflecting human rights and human freedom in measuring HDI are on thin ice. Only realistic way is not try to integrate those dimensions into overall HDI and not to consider the incorporation of sustainability, human freedom and other indicators into the index, unless we want to destabilise, discredit and depreciate the whole concept of HDI. Instead of that, information about human rights and freedom and environmental un/sustainability should serve as additional information about each country in the index. At the end, human development and human rights are in fact two sides of the same coin – there is no development without human rights and human freedom.


  • Béla, J. P. (2009): China Since Tiananmen. The Massacre’s Long Shadow. Journal of democracy, vol. 20, no. 3., pp. 5 – 16.
  • Freedom House (www.freedomhouse.org).
  • Hicks, A. D. (1997): The Inequality – Adjusted Human Development Index: A Constructive Proposal. World Development, vol. 25, no. 8, pp. 1283 – 1298.
  • Human Development Report 2009. Overcoming Barriers: Human Mobility and Development (http://hdr.undp.org/…bal/hdr2009/).
  • Human Development Report 1990. Concept and Measurement of Human Development (http://hdr.undp.org/…bal/hdr1990/).
  • Human Development Report 2007/2008. Fighting Climate Change: Human Solidarity in a divided world (http://hdr.undp.org/…dr2007-2008/).
  • Human Development Report 1994. New Dimension of Human Security (http://hdr.undp.org/…bal/hdr1994/).
  • International Literacy Statistics: A review of Concepts, Methodology and Current Data (http://www.uis.unesco.org/…port2008.pdf).
  • Kelley, C. (1991): The Human Development Index: “Handle with care”. Population and Development Review, vol. 17, no. 2, pp. 315 – 324.
  • Lind, N. (2004): Values reflected in the Human Development Index. Social Indicators Research, vol. 66, no. 3, pp.283–293.
  • Mazumdar, K. (2003): A new Approach to Human Development Index. Review of Social Economy, vol. 64, no. 4, pp. 535 – 549.
  • Mcneill, D. (2007): Human Development: The Power of the Idea. Journal of Human Development, vol. 8, no. 1, pp. 5– 22.
  • Morse, S. (2003): For Better or for Worse, till the Human Development Index do us Part. Ecological Economics, vol. 45, no. 2, pp. 281 – 296.
  • Neumayer, E. (2001): The Human Development Index and suistainability – a constructive proposal. Ecological Economic, vol 39., no. 1, pp. 101 – 114.
  • Noorbakhsh, F. (1998): A Modified Human Development Index. World Development, vol. 36, no. 3, pp. 517 – 528.
  • Noorbakhsh, F. (1998b): The Human Development Index. Some technical Issues and alternative Indices. Journal of International Development, vol. 26, no. 10, pp. 589 – 605.
  • Ranis, G. – Stewart, F. – Samman, E. (2006): Human Development: Beyond the Human Development Index. Journal of Human Development, vol. 7, no. 3., pp. 323 – 358.
  • Rapports sur le Développement Humain. Human Development Index: Going beyond Income. Luxemburg (http://hdrstats.undp.org/…_fs_LUX.html).
  • Sagar, A. D. – Najam, A. (1998): The Human Development Index: A Critical Review. Ecological economics, vol. 25, no 3, pp. 249 – 264.
  • Sakiko Fukuda-Parr (2001): Indicators of Human Development and Human Rights – overlaps, differences…and what about the Human Development Index? Statistical Journal of the United Nations ECE, vol. 18., pp. 239 – 248.
  • Srinivasan. T. N. (1994): Human Development: A New Paradigm or Reinventing of the Wheel. The American Economic Review, vol. 84, no. 2, pp. 238 – 243.
  • Streeten, P. (1994): Human Development: Means and Ends. American economic Review, vol. 84, no. 2, pp. 232 – 237.
  • Sagar, A.D. – Najam. A. (1998): A Human Development Index: A Critical Review. Ecological Economics, Vol. 25, no. 3, pp. 249 – 264.
  • Syrovatka, M. (2008): Jak (ne)merit kvalitu zivota. Kriticke pohledy na index lidskeho rozvoje. Mezinarodni vztahy, vol. 43, no. 1., pp. 9 – 37.
  • The Next Generation of Literacy Statistics. Implementing the Literacy Assessment and Monitoring Programme (LAMP) (http://www.uis.unesco.org/…_2009_EN.pdf).
  • UNESCO Institut of Statistics (www.uis.unesco.org).
  • Weiss, T. G. (2008): What’s wrong with United Nations and How to fix it. Polity Press, United Kingdom.

Autor studuje politologii na Masarykově univerzitě.

Poznámky pod čarou

  1. Possibility of those outcomes was explicitly mentioned in HDR 1994: „This scaling could produce a frustrating outcome, since a country might improve its performance on life expectancy or educational attainment but see its HDI score fall because the top or bottom countries have done even better – in effect moving the goal posts.” (UNDP 1994: 92), i.e. if a country A (the lowest performing country in a life expectancy) has reached some achievements and raised the minimum level, the country B achievement level is changed, even if no change in the performance of country B has been reached. (Mazumdar 2003: 538)
  2. “Islands of relative Calm“ are periods between 1991 – 1993, 1995 – 1998 and since 1999.
  3. In light of this effort, they propose the following broad categories of Human Development: (1) The HDI itself, (2) Mental well-being, (3) Empowerment, (4) Political freedom, (5) Social relations, (6) Community well-being, (7) Inequalities, (8) Work condition, (9) Leisure conditions, (10) Political security, (11) Economic security. For each of these categories they identify a set of indicators (total of 39 indicators). (Ranis, G – Steward, F. – Samman, E. 2006: 328 – 329)
  4. According to his view, there are four major reasons why resources and environmental issues should not be integrated into the HDI itself: (1) there is no direct relationship between resource exploitation and environmental degradation on the one hand and the level of human development on the other hand. (2) While the variables included in the HDI so far are all clear on where improvements is to be gain, this is in general not the case for environmental variables. (3) The HDI already consists of variables that many regard as hardly commensurable and (4), we can say most importantly, any structural change of the HDI will make comparability over time with former years impossible. (Neumayer 2001: 114 – 115, Syrovatka 2008: 25)
  5. The similar situation, nevertheless, is concerning human rights also. Contemporary China shows a remarkable progress on human development, but without freedom so far. Even the fact, that in this post-totalitarian regime is economy rapidly growing up, the civil-rights and human freedom are still in a prenatal period. (see also Béla 2009)
blog comments powered by Disqus
Petr Pribyla
, ,
1. 4. 2010