ISBN 92-2-112579-3
First published 2001

Download the document in Adobe Acrobat's Portable Document Format (PDF)  

Contents

Tables and figures

Table 5.1. The size of working poor and working poverty rates (working poor/employed) in 1986 (actual and predicted and inflated) and 1997 (actual and predicted and inflated)

Table 5.2. Changes in union density (number of countries) 1985-1998

Table 6.1A Low income countries: Correlations across the decent work dimensions

Table 6.1B Middle income countries: Correlations across the decent work dimensions

Table 6.2. Significant correlations across indicators

Table 7.1. Classification of residuals

Table 7.2. Net country scores in ascending order

Table 7.3. Simultaneity in country performances on six indicators

Table 7.4. Analysis of Variance: Net good and bad simultaneity

Table 7.5. Analysis of Variance: Pure good and bad simultaneity

Decent work dimensions in work and society (drawing)

Figure 1. Formal employment and national income, 1995

Figure 2. Urban informal employment and national income, 1990s

Figure 3. Working poor rate and national income, 1997

Figure 4. Socially protected population and national income, 1996

Figure 5. Social security expenditure and national income, 1996

Figure 6. Adult illiteracy and national income, 1990s

Figure 7. Life expectancy gap and national, income, 1997

Figure 8. Gender inequality index

Figure 9. Gender differentials in primary education and national income, 1995

Figure 10. Child labour and national income , 1995

Figure 11. Freedom of organization and national income, 1985

Figure 12. Union density and national income, 1995

Figure 13. Civil liberties index and national income, 1997

Figure 14. Institutional quality and national income, 1997

Figure 15. Information quality and national income, 1985

Acknowledgements

Foreword

• What is the link between the decent work agenda and the theoretical welfare literature?
• Is there a scope for policy interventions on decent work dimensions in developing countries?
• Are the four decent work dimensions of decent work linked to each other with respect to levels of national income?
• Is there a case to be explored for complementarities in decent work dimensions?

In answering the first question, the paper begins with an interpretation of the "development objective" of decent work and explores how it maybe related to contemporary approaches in human development. It discusses two important conceptual links to the agenda. The first link is with the capability based human development framework which allows for the dimensions of decent work to be expressed as societal goals. The second link is with developments in economic theory that allow for efficiency and equity considerations to be addressed simultaneously in policy making.
In answering the second question, the paper empirically illustrates the relationship between particular dimensions of decent work and economic growth. The basic point of the illustrations in this section is that while the economic resource position of a country is critical with respect to what one may "expect" to achieve on a social dimension, there is a case to argue for some degrees of freedom in interventions on many of these dimensions even for countries at lower levels of wealth.
For the third question, the paper shows that there are discernible linkages that are observable between decent work dimensions in low and middle income countries respectively.
In answering the fourth question, the paper suggests that there is a case for simultaneity of achievement between different social dimensions of decent work. It then proposes some provisional explanations for the existence of simultaneity where it may be found.

Rashid Amjad
Officer-in-Charge
Employment Strategy Department

1. Introduction

The aim of this paper is to examine the International Labour Organisations's agenda, decent work ⁽¹⁾ for all, from a development policy perspective. The paper begins with an interpretation of decent work and how this maybe related to contemporary approaches in human development. This section also outlines the dimensions of decent work (sometimes referred to as strategic objectives), their broad expected relationship to each other and suggests how one can plausibly and very generally interpret these dimensions as goals in country specific circumstances. Having set the outline of the agenda two important links to it are discussed. The first link is conceptual and suggests how the decent work agenda may be connected to a framework that is fast gaining ground in the human development literature. This link is with the capability based human development framework which, it is argued, allows for the dimensions of decent work to be expressed as societal goals. The second link is both conceptual but also of operational value because it indicates why the decent work agenda and indeed other social development agendas are more "realisable" in a practical sense today because of changes emerging in mainstream economic policy making. These changes are driven by developments in economics that allow for efficiency and equity considerations to be addressed simultaneously in policy making. It is suggested that in order to join serious on-going policy debates, it is critical for an international institution with a social agenda to address, as a first step , the economic feasibility of its proposed social interventions, i.e. interventions with which its institutional agenda is associated. This is the applied focus of the present exercise with respect to the ILO agenda. Thus the next part of this paper empirically illustrates the relationship between particular dimensions of decent work and economic growth. The basic point made through the illustrations is that while the economic resource position of a country is critical with respect to what one may "expect" to achieve on a social dimension, there is a case to argue for degrees of freedom in interventions on many of these social dimensions even for countries at lower levels of economic development. Given the limitation of the data base and cross-country regression exercises as such, the illustrations are only used for the limited purpose of highlighting intervention possibility ⁽²⁾. The paper then goes on discuss if there are any discernible linkages that are observable between decent work dimensions in developing countries. The last section explores how one may examine the case for simultaneity of achievement between different social dimensions of decent work. We then propose some provisional explanations for the existence of simultaneity where it may be found. A summary of the preceding discussion and results concludes.

2. The development objective of decent work

2.1 The universality of the objective

Decent work argues for making interventions in enabling dimensions of living that affect the circumstances in which a working person, given her labour and personal characteristics, starts out to achieve a decent life ⁽³⁾. These dimensions have been expressed as the strategic objectives of decent work. These are employment, security, rights at work and representation ⁽⁴⁾.
The development objective of decent work can be characterised as being about improvements in the lived life of persons inhabiting the world of work. Since this is an agenda, expected to evolve in to a framework for policy advice in the future, its starting point is necessarily an evaluative one ⁽⁵⁾. It recognises that while solutions to problems in development do centrally involve economic and social policy interventions, that operate through society and her institutions in country-specific circumstances, the evaluation of these policy "efforts" in development, need to be ultimately seen in terms of what they do to actual lives of all people in society. In fact the evaluation of an achievement on a decent work "goal" (i.e a strategic objective) is, in this characterisation, primarily in terms of how this achievement improves the lives of all in society. This development objective is common to all societies, because all societies are in a state of transformation and development. It is probably also the simplest common ground that universalises the aims of the ILO's agenda to societies at different levels of development. An evaluative starting point that makes it applicable to all countries and to everyone who works in them.

2.2. Country specific conditions, institutional base and relativity of goals in decent work

There are a few important points of perspective that ought to be noted with respect to interventions for achieving decent work . The first point is concerned with the focus on all those who work and each of the four dimensions / strategic objectives of decent work; the second is of an institutional variety that is concerned with the specificity of society itself; and the third, which is derivative of the first two, addresses the relativity of decent work goals.
While it is true that the core of policy intervention in developing countries ought to concern itself more (or at least not less) with issues of well being of their largely poor working populations many of whom do not even have basic needs satisfied, this focus ought not to mean that objectives of decent work that are not directly focussed on employment generation (and therefore sometimes by implication not directly on growth) can be ignored. Typical institutional and resource circumstances tend to demarcate the set of pursuable policies in developing countries from those that are pursuable in developed countries. This is a matter of realism, insofar as ignoring this issue makes policy prescriptions infeasible. Given the importance of keeping in view each objective for all on the one hand and the specificity of institutional and resource circumstances for effective policy making on the other, it is important to recognise the issue of relativity of weights attached to decent work's universal principles in the real world. Decent work cannot argue for an absolute universal that all societies must adhere to, rather it argues for a given society to set its own targets on decent work objectives, and periodically re-set them at higher levels as and when it achieves these goals The role of international organisations and donors can and ought to be one to facilitate countries in achieving these goals, not in imposing them in any circumstance. Therefore it is crucial to understand that the weights attached to these objectives are best set by society itself ⁽⁶⁾, and that these can differ according the social and economic conditions in which a society finds itself. This does not preclude the idea of a critical minimum or floor, determined by society, that each of these objectives can be subject to.
Thus global goal relativity, country specific civil society determined priorities within objectives, and possible minimums that each objective may be subject to in any concrete situation are issues that are important to recognise. Practical and useful policy making that is meaningful in a real society must keep these features in mind.

2.3. Revisiting the strategic objectives of decent work

The objective of employment is important for realising decent returns from labour and ensuring its decent use. The objective of security in the workplace, covering physical working conditions (as well as protective security in society for all), gives those who possess labour, security in using it. The objective of work rights allows for social conditions of work to be improved and discrimination at work to be levelled, thereby enlarging the possibility of achievement from work for all. The objective of representation and dialogue is both important in itself and can be instrumental in achieving the other objectives in the domain of work. Without representation, consultative processes that allow dialogue in society are infeasible and the process of coalition forming seriously incomplete. The latter is a critical basis upon which un-attained rights in society can be secured. Further, if the strategic objectives of decent work are related to each other for achievement, then it is conceivable that the pursuit of some objectives can make others more realisable. Successfully advocating policies - for employment of quality, social security, rights at work, and better representation in favour of all those who work - can deepen the realisable attributes of labour for its possessors. They could also improve upon the adaptation of people's characteristics to realise the potential of the work they do, in the face of improving external social conditions.
The diagram below illustrates the issues that are being addressed as a chart. These are delineated as two connected spheres of "work" and "society". Under each are the associated dimensions of decent work represented in work and society respectively ⁽⁷⁾. Indicators in the sphere of work are more recognisable for the ILO as they are part of its immediate institutional concerns. For example, the parallel to the dimension of security at work is protective security in society as a whole. The broader equivalent of work rights is that of equality of opportunity in society as a whole. The associated dimension of freedom of association and dialogue at the workplace are civil liberties in society as a whole.
The arrows within and across these spheres suggest linkages. The process underlying arrows across the societal and work spheres are in large part reflective of the institutional structure and the distribution process. So for example how overall national income translates in to employment of a certain quality involves mechanisms of distribution ⁽⁸⁾ and the institutions existing in society. Similar arguments can be made for each of the other dimensions between spheres of work and society respectively. It can be inferred from the descriptive diagram that just as there is a limitation on the expansion of decent employment that is posed by national income, which in the short run is given because the structure of output and demand is given; the link from employment back to national income/growth itself can operate via human capital development, productivity and worker incentive mechanisms ⁽⁹⁾.
Decent work dimensions in work and society

The first methodological point to be made in this context is to state upfront the agenda's focus is on all those who work. This means that in an evaluative sense, costs and benefits of interventions need to be seen in terms of what they do to all those who work, and not just to those on whom they may be first directed.
Decent employment , security at work and protective security are however dimensions where costs can be directly evaluated. Consequently, there are policy concerns as to how affordable these are with respect to national income and its growth and whether or not they can in-themselves also be seen as enhancing growth. Most aspects of work rights and representation are institutional and legal in nature but that does not mean that they do not have costs ⁽¹⁰⁾. For example non-discrimination at work can mean transformations in the wage structure and can thus effect relative prices in the economy. Equality of opportunity may on the other hand also have more direct distributional effects on national income and growth. Dialogue (understood in the ILO as tripartite consultation) linked to civil liberties and democracy is an important mechanism for the achievement of any goals in society. The relationship of civil liberties to national income and growth arguably operates through the systemic confidence it induces in people to proceed on chosen priorities according to consensus. There are obvious broader social costs and benefits associated with these too. Consequently, most interventions in the improvement of social dimensions are to be understood as broadly "distributional" in nature and enhancing equity in society.
There is another important distinction that we need to keep in mind which relates to expectations with respect to historical time. This point is in principle applicable to all social agendas and not just that of decent work. We need to distinguish between indicators whose dynamic or process of change primarily has an individual focus and those indicators whose process of change is more societal, institutional or structural in nature. For example better employment or greater security are directly enjoyable by individuals, while it is less meaningful to talk of the achievement of democracy or gender equality or the freedom of association by an individual. The latter come about as society-wide changes and imply transformations in the social structures. Clearly individuals can gain from these societal changes, however the categories are different and there are reasons why we may find that they move differently during the process of growth ⁽¹¹⁾.
From the point of view of policy intervention there are two types of concerns we have. The first can be called an intrinsic set of concerns and the second an instrumental set. Intrinsic concerns are reflected in an observed deficiency with respect to a social dimension. So for example, if some form of gender inequality, an indicator of discrimination, is observed in a society (which arguably results in inequality of rights between sexes at work), this observation on its own ought to be sufficient to trigger some policy action. Such a policy act could also in itself be seen as an effort to reach an independent developmental goal (gender equality). In other words, intrinsic concerns only need the establishment of inequity in a given social dimension for society-wide policy action. Instrumental concerns on the other hand are about the effects of a deficiency/excess in an indicator in one development dimension on indicators in other development dimensions. There are also two types of questions to look at here. One concerns the possibility of virtuous and vicious circles resulting from a possible simultaneity or connectedness in achievement of social indicators. This is an area explored in this paper. The second question is that of trade-offs, between indicators of social dimensions on the one hand and between social groups affected (or not affected) by actual interventions in those social dimensions on the other ⁽¹²⁾. The latter are critical to appreciate but can be more appropriately examined at a country specific level. Although the discussion here does touch on the latter, it does not do so sufficiently. This should be a subject of further research.

3. Decent work, freedom and human development

Decent work is only an agenda of an international institution, while the capability approach, (of which development seen as freedom is an example), is more of a general theory of evaluation of human welfare. Decent work is connected to both as a point of departure. The connection between the notion of "development as freedom" due to the work of Sen ⁽¹³⁾ and the spirit of the decent work agenda is at one level an obvious one ⁽¹⁴⁾. Sen's conception of development consists of the idea of expanding freedoms in both social and economic spheres of life. For societies in which individuals or groups face obvious un-freedoms in these spheres, development means the removal of un-freedoms. An unfreedom must be removed in a sphere of life in a meaningful sense if freedom in it is to be expanded in it a positive sense. These un-freedoms can be expressed variously, for example as goals of: freedom from hunger, or the freedom from the lack of basic human rights, or the freedom from avoidable illness and so on ⁽¹⁵⁾. The expansion of freedoms/ removal of un-freedoms and the advancement towards these goals in one sphere of life (e.g freedom from hunger) are both important in themselves as goals, and they can have enhancing effects for overcoming other un-freedoms (e.g advancing on freedom from hunger arguably also advances one on the front of freedom from avoidable illness). Taken individually or together these social advances- that have a meaning in themselves- can have instrumental effects too. They can also furnish the building blocks on which consensus coalesces and provide platforms to society for determining priorities and act according to its expressed values (e.g in the present example, this could be an open debate on minimum acceptable standards of living and well being in that society). Decent work that is in part a rights based-agenda (which sees certain rights in the world of work as goals for societies to strive for) can be characterised as being close to such a conception of "development as freedom".
This brings us to the connection of decent work with the capability approach ⁽¹⁶⁾. Initially developed by Sen as an ethical investigation in to the issues of distribution and inequality, it has been developed and applied by him and others to the understanding of dimensions of well-being ranging from poverty, famines, gender, justice to development itself. In contrast to competing theories of welfare, the capability approach can deal with interpersonal differences between individuals, as it is focussed on the chosen ends for which the possession of commodities requires a "conversion". This essentially means that not only is the possession of an asset or commodity important in the approach, but also what a person is able to do with it ⁽¹⁷⁾. Secondly, for our purposes we need to note that the capability approach takes in to account external conditions under which individuals deem themselves to be "satisfied", such that the idea of satisfaction is not restricted to a socially disconnected metric of individual utility or happiness. This essentially means that not only is the possession of a commodity or asset and the intended aim of using it by the individual important in the approach, but also the fact that people must be considered on a universal basis. In other words the fact that some persons may be satisfied by less, simply because they take their un-freedom as a fait accompli, does not mean that in evaluating their position we should share their despondency. Consequently the importance of individual satisfaction does not allow one to take as given social and structural exclusion between persons and groups in society.
The capability approach is thus centrally concerned with what a person actually succeeds in doing with a commodity or asset given (1) its characteristics; (2) her own personal characteristics; and (3) the external (natural and social) circumstances faced. When we place the decent work agenda here, what we can immediately see is that it focuses on labour. It aims to enhance the characteristics and attributes of labour for all its individual possessors, taking in to account the actual external circumstances faced by them such that they succeed in living a decent life. These links of the ILO agenda with the capability approach need to be the subject of further conceptual work ⁽¹⁸⁾.

4. Economic performance and the development objective of decent work

4.1. The development policy implications of the new and old economics

The agenda of decent work recognises the importance of economic growth to development.
The first question that is to be addressed in this context is about the relationship of decent work dimensions to the level and growth of national income. The answer to this question is needed both for illustration and for persuading policy makers, who are agnostic on social agendas (for example in the ILO context- the decent work agenda) when confronted with the issue of growth, to change their minds. This may be less difficult to accomplish today than it may have been in the past, because of both an internal and an external reason. The external reason is that the need to reassess the impact of economic policy on the worst off has become imperative because of the slow progress in eradicating human misery in the world. The internal reason is to a significant extent theoretical, namely that economics as a discipline has undergone a change that can allow for the simultaneous examination of "social" and "economic" spheres of society more readily ⁽¹⁹⁾. The implication of this change for institutions with social agendas (agendas that were previously taken on board as auxiliary additions to the mainstream of economic policy making) is serious. The basis of this change is outlined in the present section.
A position still not uncommon in developmental policy making today is a legacy from the past that allows for the separation between efficiency (read economic) and equity (read social) issues. This position in its stricter theoretical versions argues that economic policy should focus on efficiency and need only be concerned with optimising resource use (including labour ⁽²⁰⁾) and social policy needs to focus on equity and this be geared towards effecting non-distortionary lump-sum transfers and improving public good access, that deal with inequities produced by the market mechanism .These powerful and important results of welfare economics were importantly based on special competitive market and costless information assumptions as well as the idea that instruments for lump-sum transfers existed ⁽²¹⁾. The crucial point however is that the separability of efficiency and equity in policy making is more amenable to justification if its underlying real world also corresponds to the one accompanying the traditional welfare results. These traditional assumptions have been relaxed over the years to yield results which show that resulting equilibria with the now relaxed assumptions need not be efficient ⁽²²⁾. While the non-existence of or imperfections in lump-sum transfer instruments in theory have revealed the "substituting" nature of efficiency and equity interventions (i.e. trade-offs), literature on institutional arrangements especially in developing country contexts (by relaxing costless information and other assumptions) have sometimes shown "complementarity" between equity and efficiency ⁽²³⁾. In short economic efficiency and equity are not separable in the real world, they can be shown to be either complementary or substituting, which also means that "re-distribution" (read: interventions in social spheres) in principle can either increase or decrease economic efficiency. There are therefore both complementarity emphasising results and those that emphasise trade offs.
Orthodox theoretical views of course translated easily in growth-exclusivist policy positions, which generally hold that policy for economic growth should proceed as if efficiency and equity issues are separable. Today in the context of theory with more realistic assumptions about the world newer versions of earlier growth exclusivist positions often assume that policy should proceed as if those decisions that are bound by efficiency-equity trade-offs ought to be decided in favour of efficiency. Quite apart from the issue of the legitimacy of inferring "policy positions" in this way from the original theory, these positions also tend to reproduce a base-line model of the world in which the welfare potential of human beings (and issues pertaining to collective and individual human behaviour that can enhance or block that potential) is separated from the heart of policy making or can be ignored in fundamental respects.
In the area of development, there is work that is beginning to challenge such policy positions more systematically. Of course, in order to do this these positions must rely on results that either suggest that some interventions for equity enhancement can lead to increased efficiency or not affect it, or argue that efficiency concerns under some conditions, where trade-offs do exist, could to be decided in favour of equity. As suggested some of this work is now systematically engaged in micro level analysis on the efficiency enhancing effects of inequality reduction. There is of course other more direct work, especially in non-economics disciplines, that tends to argue that policy instruments for growth are not always good for the poor, especially in the short-run and in the more "dis-aggregated" perspective of a community or in group-specific cases.
It is of course still not clear what an inclusive notion of human development on the one hand and economic policy making that does genuinely allow for efficiency and equity concerns to be addressed simultaneously (as opposed to holding to the primacy of one on the other on a priori grounds), will bring forth as an applicable policy framework in the future. There are important advances but we are still a distance from an alternative full fledged "human development policy framework". Movement towards the integration of the new economics and the capability based human development literature is possibly one of the areas from where this alternative can be expected to emerge in the future ⁽²⁴⁾.

4.2. The immediate challenge for policy makers

Historically the expansion of growth-exclusivist policy positions towards a more inclusive notion of development came about importantly in the 1980s when the inclusion of health and education as valid dimensions of poverty were accepted by many international and national policy making bodies. The focus on "institutions" as a part of making policy effective is going through a similar process of absorption today in the international system. Therefore differences between policy positions are increasingly being focussed not so much on ends but on means. Namely on the appropriate policy instruments.
For institutions espousing specific social agendas, however, one necessary step in this direction -i.e. of fully engaging in debates on policy instruments - is to try and first demonstrate the feasibility of intervening on their chosen social agendas with respect to national income, and exploring the wider characteristics of such agenda-specific interventions. This is in fact the central empirical aim of this paper, with respect to the ILO agenda. As suggested this could also be useful to have in place for taking the next step, i.e. to participate in debates on impact assessments of appropriate policy instruments ⁽²⁵⁾ - a direction in which the important policy debates are moving.
A challenge for policy makers is thus to resist conceptualising the development process in particular national economy contexts as either-or situations. If on the one hand we have growth exclusivist views which hold social dimensions and associated rights as "secondary" for countries at low income levels, we can also have social dimension centric-rights exclusivist positions that argue for the "primacy" of the latter at all levels of income. What is at stake is the assumption we make on the interaction between the two. These positions either suggest that growth definitely produces well-being and rights for all (for which economy-wide longer-run evidence is often cited) or that well-being and rights definitely produce growth (for which enterprise, firm or group level evidence of a shorter run and of a more case specific variety is cited) ⁽²⁶⁾ .
Since decent work is a social agenda we focus here on growth exclusivist positions as views to be qualified. At an historical level, as was suggested above, arguments based on the cases of sequencing of growth preceding rights, as in the histories of parts of present day Western Europe, are invoked to suggest what may have been the case in the past must be true for today ⁽²⁷⁾. Second, it is argued that interventions for equity which improve or alter distribution, important as they may be, are best avoided on a mix of grounds of liberty (negative freedoms) ⁽²⁸⁾ and generating disincentives for investment and therefore growth itself ⁽²⁹⁾ . Thirdly, it is suggested that effective interventions in rights-based spheres of society suffer from an agency problem. This means that implementation of direct social policy when attempted on-scale is both difficult and is un-monitorable ⁽³⁰⁾. Important as these arguments are, suffice is to say that there is more room for debate and empirical research on these matters. The position of policy makers agnostic to the simultaneous pursuit of human development with growth (on "practical" grounds) is summed up well by the rhetorical question:
Important as they are, can countries at low levels of "development" really afford protective and active freedoms before sufficient growth and accumulation have taken place ?
This question, as should be apparent from the foregoing, is equally about the way social policy is made as it is about what social policy is. In fact this is where one part of the debate relevant for policy makers is situated. Economic growth we know is good for reducing poverty, this is an empirical result that has rightly been emphasised in the literature and cannot be seriously questioned ⁽³¹⁾ . However in the last two decades it has also been true that the poverty problem has not been reduced to the extent desired. The emphasis on proportions (which have gone down in many instances) in preference to numbers (that have gone up in many instances) has sometimes been misleading and at other times simply disingenuous. The point is that good as it is, growth has often not been enough. It is also true in an historical sense that the sole focus on equity (in material terms at least) embodied in the experience of ex-socialist countries now "in transition", led to a growth failure. Here equity de-linked from growth may have become unsustainable ⁽³²⁾.
However, the matter at hand is that for those directly affected by poverty, policy designed to generate growth has brought with it levels of employment quality, health, education and rights that have not allowed entire populations to be freed of poverty. Correspondingly for national governments it has generated a fundamental concern for their citizens (levels of corruption and institutional imperfections notwithstanding). For other organisations, especially international institutions advising countries on policies of growth and its social dimensions respectively, there is a jurisdictional crisis. This is why most development agencies today have a common stated goal of "poverty reduction". The challenge, however, is not simply one of joining forces to generate "additive" programmes but one of conceiving policy advice simultaneously.
The question then is both one of policy integration and institutional coordination. Integration however has to be real, and there are some signs that institutions with policy leverage concerned with growth are recognising that "poverty reduction" needs be addressed and seen simultaneously with the formulation of growth strategies and not merely as an add-on ⁽³³⁾. The focus on growth and the centrality accorded to it is necessary but if we are to reach for an inclusive developmental objective, growth on its own is unlikely to be sufficient to deliver the internationally set goals on poverty reduction within an acceptable time period.
Although we have briefly put the issue of interventions in social dimensions in its wider development policy context, the aim of this paper is both a specific and a preliminary one. It is namely, to ask how economic growth (understood as an expansion of per capita national income), relates to indicators of social dimensions that an international institution focussing on work and labour is concerned with ⁽³⁴⁾. We explore three specific questions in this regard.

• First, whether there is some demonstrable scope for interventions in these dimensions, given the ultimate "constraints" that resource levels and their growth may impose on countries?
• Second, are there any linkages that are discernible between the four dimensions of ILO's agenda ?
• Third, if there is a basis to look for simultaneous movement on these dimensions.

5. How feasible are policy interventions for promoting decent work?

5.1. The dimension of employment

There is little doubt that the expansion of wealth in society must make a contribution to employment. Consequently the link between economic growth and the quality of employment has rightly been central to both the discipline of economic development and development policy making ⁽³⁵⁾. The positive link between the quality of employment achieved with higher level of resources, is not moot, it is however not straightforward to show. We use a few examples to illustrate the issue.

Formal employment

The incidence of formal employment is an indicator of employment quality for a good reason. Most formal sector jobs have some degree of protection associated with them with wages set well above the poverty line. Figure 1 shows the relationship between formal employment and per capita national income ⁽³⁶⁾. There is scarce information on the indicator of formal paid employment as a percentage of labour force. The reason for this information gap is that in many cases to enumerate paid employment, household surveys, as opposed to census or establishment-based surveys are used. However, only the latter give appropriate estimates of what can be called "formal and protected" employment. In fact we have such information for only 16 countries of which most are high income countries. The weighted average of this figure for high, medium, and low country classification is also reported ⁽³⁷⁾.

Figure 1

Although Figure 1 suffers from a lack of data and is over representative of high income countries, it is also definitely the case that there is limited amount of such employment in lower income countries. Given these qualifiers, Figure 1 indicates a few things. First, that on an ideal notion of formal employment, which in many ways suggests one route to achieving the employment objective of the ILO agenda (i.e employment in decent conditions), we can expect higher levels with the higher levels and growth of national income. Second, there are rather large percentages of populations that are not covered as we move down the national income ladder. Third, since the fit in the regression is so good (the unexplained variation being less than 14 per cent), the possibility of policy intervention to increase this kind of employment directly on-scale, i.e. increasing formal and protected jobs by policy fiat, may be problematic. Although, this is also intuitively obvious, it is particularly important to explicitly recognise.
We have focussed on a particular type of employment, the figure however points to general problems associated with the measurement of employment in developing economies. If employment is considered to mean the number of jobs in an economy it is in part a definitional matter, because societies always have some level of resources, most societies do grow, and a majority of people who can work do work in all societies. But if we are concerned with both the quantity and quality of all employment then we are concerned with not only those who enjoy decent employment but those who do not enjoy decent employment, as well as those who work but may not count as employed and about whose conditions of work little is directly known. Those who do not enjoy decent employment are a significant percentage of workforces in developing countries, and by that token in the world. The link between social well being and employment is at the core of this matter. The point is simple. Employment is supposed to generate income in a commoditized economy, and income is a critical means of attaining well being. In many developing economies, that at best possess an incomplete social floor ⁽³⁸⁾, the observed "fragmented" labour market is linked to deprivations in the space of well being. So we get employment categories that reflect varying shades of irregularity, insecurity, rightlessness and voicelessness ⁽³⁹⁾. By the same logic, it is also true that in many developed countries where minimum social floors exist, even though labour markets do get segmented a link to basic well being is maintained.
It would not be unreasonable to hold that since the growth experience of the last two decades, barring notable exceptions, has shown that the increases in this category of good employment have been slow at best, we need to ask ourselves another type of employment question that focuses on more of those who work, or at least larger parts of the labour force, excluded from the illustration given in Figure 1.

Informal employment

The category of the informal sector employment is amorphous and based on left-over or residual definitions. Moreover, the informal parts of a typical developing economy can also include relatively prosperous people. Many self-employed are part of informal sectors and many of these are non-poor. However, in an overall sense informal activities do imply, at least in net terms, ill being of the average person inhabiting its labour market relative to the rest of society. Figure 2 shows a scatter plot of the regression between an estimate of some those who work in the urban informal sector as a percentage of urban employment and national income.

Figure 2

The measure of informality taken here is an ILO estimate of the urban informal sector as percentage of the urban employed. This is a rough measure with some problems of uniformity of definition across countries Since data is based national definitions that vary from country to country it is not clear what or how much is left out. This is best illustrated by the magnitude of the percentage itself. A casual glace at this data shows that for a very large set of developing countries the figure is between 40 per cent to 50 per cent of urban employment. This suggests that around half of urban employment is formal. This indeed seems a high percentage. The suggestion that national data on "informal employment" often refer to employment in small and micro enterprises consequently makes sense ⁽⁴⁰⁾. The graph essentially shows the relationship between the parts of urban informal sector employment as a percentage of the urban employed and GNP per capita (PPP). The values have been averaged for each country on the two variables over the 1990s. The line on the graph is the simplest relationship that can be fitted between the variable and national income. It shows what on an average can be expected to happen to informal employment as national income increases. Clearly as national income grows urban informality declines. This is an interesting result given the data limitations discussed above. The points on the plot representing country observations (the data on which the regression line is based) suggests a wide enough spread of observations above and below the predicted line. Thus many countries can be said to "outperform" their expected value from the fitted line just as many countries "underperform". However observations at lower levels of income suggest more countries that are worse than the norm, so the spread is wider at slightly higher levels of income. The unexplained variation in this relationship (76 per cent in this case) confirms the visual observation, namely that there is policy scope to reduce informal employment, especially in countries where incomes are not too low ⁽⁴¹⁾. Not only is the level of informality high in developing countries but the experience of the last two decades has also shown, that the majority of countries, for which data on two points in time are available (16 out of 25 in the ILO sample), show increases in this category.

The working poor

Lastly, we examine the relationship between the working poor as a percentage of the employed and national income ⁽⁴²⁾. In developing countries where effective social insurance systems do not exist the poor work in order to support themselves and their families. In these countries it is also apparent that the long run unemployed largely cover those who in fact are not the "poor" in society. Employment policy tends to ignore the importance of these two stylised facts about poverty and long-run unemployment in developing countries ⁽⁴³⁾.
The advantage of using the working poor rate over the informal urban employment category is that it focuses on the poor and is applicable to both urban and rural areas. Its advantage over the formal employment indicator is that it has a greater coverage of countries and that it focuses on the employment issue in terms of its deficiencies and deficits. Figure 3 below shows the scatter-plot for this relationship. Once again it is fairly clear that although increases in national income reduce the size of the working poor, like poverty itself, countries at the same level of national income can have a less than expected or more than expected incidence of the working poor.

There is scope for policy intervention in this sphere as well. Although the spread of observations is greater at relatively higher income levels, there are cases to suggest that there is still room to intervene for improvement at the low end of national income. From an employment perspective trends over time in this category are alarming.
Some initial research at the ILO, as shown in Table 5.1, indicates that there were around 530 million persons who could be classified as the working poor in developing countries for 1997. The working poor constitute around 25 per cent of the employed labour force in developing countries. Around 95 per cent of these working poor of the developing world live in low income countries. Their populations in low income countries have increased, while these have declined in the middle income countries. The marginal decline that has taken place over time in the working poor populations in the all country classification is therefore driven by the declines middle income country category.
Table 5. 1 The size of working poor population and working poverty rates (working poor/employed) in 1986 and 1997

	Working poor in 1986	Working poor in 1997	Number of countries	% change annual
All low and middle income countries	536.67 (32.2)	534.22 (25.3)	139	-0.04
All low income countries	473.99 (35.1)	505.47 (29.8)	63	+0.60
All middle income countries	62.68 (19.2)	28.79 (6.7)	76	-4.91
Source : N. Majid(2001), The Size of the Working Poor Population in Developing Countries. Employment Paper 16/2001, ILO (2001).

At the group level, both low and middle income country groups include countries that show increases as well as declines in the working poor. However, while in middle income countries there is a preponderance of countries showing declines according to expectation; in the low income country category both declining and increasing working poor countries are significant in number. This suggests that a possible polarisation process may be taking place within low income countries between those that are reducing the working poor and those that are not ⁽⁴⁴⁾. This is a subject for further inquiry.
The issue is not so much about actual numbers but about proportions to target in an employment policy focus. It can be therefore be suggested that while the transformation of all the employed in to formal protected workers in the short term is bound by resource conditions, implying that the dominance of the best forms of employment in the labour market is closely tied to national income, the worst forms of employment can indeed be improved upon without waiting for higher levels of national income. While growth will improve overall well being and thus employment quality at least in some sense of the term, the extent of this improvement would depend on the distribution of income and assets, institutions as well as the effectiveness of policy itself, and it could entail advocacy, participatory human agency and struggle to realise this improvement. For this to happen economic growth policies would need to take an approach that integrates economic policy with its social outcomes. Important as it may be, the point is not so much about waiting for some elusive inter-disciplinary integration of the social sciences but rather one of policy timing, in which when economic policies (whose aim is economic growth and its stability) are formulated their postulated effects on at least all those who work and who belong to poor households in the economy are given priority attention. There are informational requirements for this as well that cannot be underestimated. This is a challenge for the future.
Although the bulk of the efforts in addressing the problem of the working poor are likely to come initially from the expenditure side of the budget, an element in the desired policy framework must be some real attempt at the inclusion of the working poor in the growth process itself as opposed to them just being second round beneficiaries of growth multipliers. Thus on the growth side some attempt must also be made to focus on improving the productivity of the working poor. Both redistributive credit with peer monitoring as well as direct enhancement of productivity are ways to do this at the micro and community level. The challenge here is to think through the issue of attacking the poverty of those who work both in terms of larger macro policy framework that is consistent with it (including the demand conditions facing an economy in which it is contemplated ⁽⁴⁵⁾) on the one hand, and efficiency enhancing inequality reducing interventions at the micro and institutional level on the other .

The employment challenge

The first employment challenge for growth in developing countries must be that income generating work, whatever form it takes, be at least equal to a standard set by society that is sufficient for earners and their families or households. The minimum of this standard of sufficiency can be seen as a nationally set poverty line. The aim is to produce conditions in which all those who want to work can find work; and those who do, live in households that are at least also non-poor ⁽⁴⁶⁾. It would be more feasible, sustainable as well as preferable if this goal is realised as a combination of enhancing security through expenditure and increasing the productivity of the working poor.

5.2. The dimension of security

The dimension of security is closely tied to employment. In fact a considerable difference between formal employment on the one hand, and informal, casual and irregular employment on the other, lies precisely in the dimension of security. Insecurity at work is an important part of the insecurity faced by those who work. The sources of work place insecurity are two-fold. The first derives from physical conditions of safety and hazard at work, and the second from the level and variability of income from work. The sources of insecurity for those who work, at a general societal level, pertain to social security and protective security systems operating in society. Indicators for security at work are often based on information on those places of work that have some security to start with. Economy wide security is reflected in public expenditure indicators on social security or wider outcome indicators of insecurity like life-expectancy gaps or basic illiteracy that disables learning.

Protection at work

One argument alluded to earlier, encountered in policy circles is that protective security is expensive and ought to be expected after the fruits of economic growth have produced a wealthier society. If high growth is achieved, it is argued, security will follow in some measure. Although strong versions of such views are less widespread, it nevertheless holds sway in some form with many policy makers.
The effectiveness of protection systems varies widely in the world. We first look at social protection coverage of a standard indicator which is that of old age, invalidity and survivors protection i.e. persons protected in this category as a percentage of the population ⁽⁴⁷⁾. This indicator of coverage in social protection does not have too many observations (35) although the observations for low (12), medium (12) and high (11) income countries are nearly equally divided. The results are however circumscribed by the lack of data. Given this, it is clear from the above that security coverage increases with national income but there is also room to intervene for increased security for countries at lower levels of income. One can argue that the pattern of socially protected populations tends to mirror that of employment illustrated in Figure 1. In fact it is not just the possibility of intervention but the limited coverage that is at issue. As reported only 6.5 per cent of the population in the low income countries sample can be said to have some standard form of protection. It also needs to be noted that more outliers at relatively higher levels of income perform better than expected, while those at lower levels of income perform worse that expected on the cross sectional fit.

Social expenditure

State expenditure on social security corresponds to a wider notion of security. We examine the relationship between national income and social security expenditure as percentage of GDP. We use the widest possible indicator that covers most social expenditures ⁽⁴⁸⁾. The scatter suggests that this category of expenditure improves with the wealth of a society. While the shares of GDP of this category show how seriously low it is low income countries, the dispersion of country observations at low levels of national income is wide enough to suggest that there is some policy space to increase this expenditure share.

This is not to say that an increased share will be sufficient. A basic point however needs to made in this regard, social benefits to the working poor and their families are likely to emanate significantly from the expenditure side of the budget, and the latter in lower income countries are likely to be very small in relation to security needs of society. Clearly in this sense good returns from employment itself is likely to be a critical source of security in low income countries.

Illiteracy

We need to look at some outcome indicators of insecurity that affect those who work. From the point of view of work a lack of basic education (which is funded largely through public provision in developing countries) fundamentally damages a person's ability to learn and therefore to adjust to changes in the labour market. This is critical to understand with respect to those discussions on labour market flexibility that assume labour can smoothly be re-trained and re-skilled with training. Basic illiteracy not only puts limits on many ways of enhancing labour productivity it also leads to insecurities both at a group and societal level as well at the level of the individual. Consequently there is a case to see how learning insecurities revealed in measures of illiteracy vary with national income. This is captured by a measure of basic adult illiteracy of those over 15 years of age.

Once again it is clear that countries perform on basic illiteracy with sufficient variation from what is expected of them on the basis of their national income alone. Interventions for improvement are thus possible for countries at moderately low levels of income ⁽⁴⁹⁾. It is in fact in the mid range of countries that illiteracy levels are mostly greater that what would be expected.

Life expectancy

Empirically life expectancy is an outcome indicator for security at the societal level. It is definitely linked to national income measures. The scatter plot below shows a measure of a life expectancy gap in the 1990s (82 years -life expectancy).

We find over the cross section, that the fit is fairly tight, and the unexplained variation in the relationship is low. This is of course to be expected since life expectancy for the whole population average is expected to change slowly over time. It also makes sense why the observations are bunched at very low levels of income, while the spread increases as income increases. Possibilities of intervention may be higher at relatively middling levels of income and are likely to be affected through universal access to components of basic health.

The security experience of developing economies

The general economy-wide security argument can be further elaborated by illustrative country cases of security experiences and rates of growth. Despite the long-run association between security and national income, country examples can be found of both high and low growth that have been successful or unsuccessful in the provision of security in recent history. We can have four possible classifications, there are the growth mediated economies like those of Taiwan (China) and South Korea (Republic of Korea), that have shown successes in both security and growth; there are the process led economies like that of Sri Lanka, or pre-reform China, that provided security but not growth. Thirdly there are the growth-first-security-later economies like that of Brazil ⁽⁵⁰⁾. There are of course many economies where as things stand there is both low growth and low security. Unfortunately it is the latter category of economy which is typically in a vulnerable position, and to which policy packages of growth-first -security-later variety have been offered in the past. The argument that insecurity for some is a necessary price to pay for prosperity and plenty in the future does not always hold up to the different experiences of countries.
Furthermore, the technical literature on the subject has a fairly robust finding, that the link between GNP per capita and life expectancy is largely driven by income distribution and public expenditure on heath and education ⁽⁵¹⁾. Although direct improvements in distribution may be a more difficult policy variable to handle, expenditure variables are economic policy variables. This brings us back to the resource question: how affordable can individual security be for a poor country? The issue to note in this regard is that both basic health and education are labour intensive activities and with low relative costs in typical developing country situations, so their costs may not be as high as imagined ⁽⁵²⁾.

The security challenge

Decent work's security objective in developing countries can be seen as an argument for the construction of a social floor for working people, chosen by society itself and appropriate to a society's conditions. This can be affected from both the workplace and through the larger collective of society. The broad empirical finding is that there is a possibility of intervention to reduce insecurity in many countries and specific cases highlight the growth-security paths that countries may take. Growth is definitely desirable when it is accompanied by the reduction of insecurity and this is the best way, but there are cases of high growth and low insecurity, and importantly of successes in insecurity reduction without high growth. Country experiences thus also suggest that we do not have to wait for security to come on-stream later.
From a point of view of policy the link between the employment goal and the security goal is critical. In the context of developing countries if basic security can be said to exist when the fall in work incomes below the national poverty line can be pre-empted then this requires three essential policy ingredients. First, provisions of basic security incomes must be in place. Second, institutional mechanisms- that can deliver basic health and protect poverty line incomes- need to be developed. These could include linking security to work through innovative insurance mechanisms. Third, a sufficient flow of information must be there such that where possible either warning systems for expected falls in poverty incomes can be developed ⁽⁵³⁾ or at least the time lapse between livelihood disasters and response is minimised. Like the case for the indicators of employment, information on security indicators especially the direct ones related to work, need to be improved.

5.3. The dimension of rights at work

The next dimension of decent work concerns basic rights at work. These have been expressed in the ILO's core labour standards namely as the freedom from discrimination, freedom from forced labour and freedom from child labour. We examine the issue of freedom of association in the next section on representation and voice.
What right is considered fundamental has varied historically. Some rights taken for granted today, would not have been universally acceptable say fifty to a hundred years ago. This is true for example of women's right to vote, and of both forced and child labour. Moreover global economic transactions with colonies for some parts of the world were structured on principles of unequal treatment and exchange. However, the general acceptance of some of these rights over time has given them greater global scope today. We examine two sets of rights relevant to work in this section . The first concerns gender and the second child labour.

Gender equality

One critical aspect of the work rights dimension is the equality of overall access to work for women. However, gender biases are reflected in all dimensions. They affect not only work rights but employment, security, and representation as well. An even larger issue related to gender inequality is reflected in the societal sphere where it pertains to the structure of the family itself. This lack of rights normally reinforces obstacles to equality of access for women.
The discrimination that women face is reflected strongly in the gender differences in mortality rates, especially in some developing countries. There is enough evidence to suggest that adjusting for features such as gender selection at birth and other demographic factors, gender ratios reflect biases essentially due to social factors. The size of such missing women may be moot, but the fact they are missing, is undeniable ⁽⁵⁴⁾.
Active social biases existing in education are revealed acutely in the freedom of women to seek employment outside the space of the family. This unfreedom has multiplicative effects on their empowerment as social beings, as well as the setting of welfare priorities within the household. The reason for such constraints is sometimes sought in custom or tradition. However the choice of which tradition or convention is to be observed needs to be a matter of public debate where all voices are equally heard. Much knowledge has accumulated in this important felid of inquiry today which can be drawn upon for effective policy making and advocacy.
In this context we first wish to ask how much effective scope there is for the removal of gender biases, given a country's level of resources. Secondly, it may be that these interventions positively impact on the growth process itself. We only have data at a society-wide level to examine this issue and usable data on gender at the level of work is not available on a comparable cross-sectional basis.

Some empirical research on the topic also tends to confirm a negative causal relation of gender inequality in education with growth. The argument is that gender inequality in education impedes growth by distorting incentives directly as well as indirectly via investment and population growth. There are possible effects suggested of gender-bias in employment on growth as well ⁽⁵⁶⁾. There is also research that shows that gender inequality in education is linked to fertility and child mortality ⁽⁵⁷⁾. These results suggests possible instrumentalities between the security dimensions of decent work and the work rights dimension.
There is however a general link between gender-bias in education and employment that needs to be recognised explicitly. It has both long and short-run implications. Discrimination at work is not only direct but is also indirectly predicated on other underlying biases. For example in a world where inequality in education (or say nutrition) is not removed, a possible non-discriminatory labour market environment, even if it were to be achieved, would still generate gender biases ⁽⁵⁸⁾.

Child labour

The issue child labour is located both in the realm of rights at work and is an independent human rights issue ⁽⁵⁹⁾. Working children tend to lose the opportunity for education and restrict building their capacities for a future as adults. They can also be subject to forms of work that are directly hazardous to their well being. Therefore, the larger societal aim in this context is appropriately the right of children to be educated unhindered. Child labour can obstruct that goal. Ultimately children ought not to be working, unless that work is non-damaging and demonstrably a part of their education itself.
The link of child labour and economic prosperity is an undeniable one. Both long run historical trends and more contemporary economic history can be used to argue that economic prosperity does reduce the use of child labour and there is much room for intervention in this field.

The figure above suggests that the middle income countries group has much lower percentages of working populations in the 10-14 age-group ⁽⁶⁰⁾, than the low income country group; the dispersion of values is higher in the middle income category, and the range of lower and upper values in each category wide enough to suggest that intervention is possible. In fact the declining trends in participation rates of child labour in the last decade or so, demonstrate this very clearly ⁽⁶¹⁾.
Any phenomenon that is sufficiently widespread in society has substantial reasons to exist. Unless the more fundamental of such reasons is addressed in policy "eradicating" the said phenomenon can easily have other negative effects in that society. This point has been partly made with respect to gender bias above. For child labour, for example, if the link with poverty is fundamental, then banning child labour in certain circumstances may not necessarily lead to the improvement of lives of children ⁽⁶²⁾. We need to be very careful in taking policy positions here ⁽⁶³⁾.
The immediate policy realms of child labour are essentially two-fold. They can be divided in to policies for legislative deterrence on the one hand and policies for education and family income enhancement on the other. For example, agreements and resolutions on elimination of certain forms child labour or national legal acts that restrict its use by fiat, fall in to the legislative dimensions. Educational policies, that are more inclusive of all young citizens, are examples of the other. There are possible policy mixes with varying shades of emphasis that may be appropriate in different situations.
Clearly educational incentives may provide good grounds for withdrawal of children from the labour market, especially if the physical well being of children is also accounted for (by the provision of basic health and nutrition). At the same time parents, even very poor parents, may decide that even at the effective loss of some income at the margin (if their child was earning more than what he or she consumed) that they would still prefer sending their children to schools within such incentive structures. Coupled with this if partial parental income enhancement is also feasible, the chances of withdrawal will be higher. One way to do this is to allow for controlled non-hazardous forms of work for children, that can be combined with education. There is some evidence to suggest that this may indeed be feasible ⁽⁶⁴⁾.
It needs recognition that grounds for the abolition of all child labour without the commensurate enhancement of family incomes and education facilities (other decent work concerns, relating to dimensions of employment and security) may make policy less effective or even counter productive. The gradation between types of child labour for policy action is important for an ethical reason as well. Some forms of child labour are particularly pernicious in damaging a child's right to develop as a full human being and usually such forms, that are often also wage-based, also tend to involve a binding for children in to such work. This is unacceptable on both grounds of hazard and lack of choice, and is ultimately ethically unacceptable today, whatever general historical practice may have been in early 19^th century or however weak or robust the negative link between child labour and growth of national income. Consequently, the emphasis on the most hazardous forms of child labour, is a realistic focus of the ILO. Other forms of child labour, often home based, are also more likely to be non-hazardous.
Policy work in this field has a very important advocacy role within societies, which is to encourage the development of children as fuller human beings. This advocacy work is meant to help governments programme and prioritise their goals on child labour, make effective the legislation against its worst forms, and provide educational incentives for voluntary withdrawals. But critically it is also to encourage an atmosphere of social activism that sets the targets for the removal of child labour internally, from within society itself.
Employment and security dimensions of decent work discussed above suggest that if the decent work "social floor" objectives in these dimensions are realised, then important gaps in work rights - e.g. of discrimination in gender and loss of freedoms due to work in child labour in so far as they are related to poverty line based social floors - could be diminished too. But beyond this much more has to be done, and the point here is general one, which is namely that wider societal inequalities that are symptomatic of the structure of society, may be relieved by but not necessarily eliminated by non-discriminatory and child labour free work. There is a connectedness between deeper biases operating at the level of society that reinforce loss of rights at and through work. Therefore these matters have to be tackled together.

5.4. The dimension of representation

The next dimension concerns representation and dialogue. This dimension is important in itself and it is often the mechanism through which consensus on the other three dimensions can be forged. It is important to recognise that the equivalents of the category of freedom of association at work pertain to civil and political liberties at a larger societal level. This issue was mentioned earlier where the work and societal spheres of each dimension were discussed.

Freedom of association

The broader notion of freedom of association entails the intuitive idea that individuals ought to be free to form groups and associations in order to pursue their chosen ends as groups in society. The applicability of these freedoms need not be restricted in principle to specific forms related to industrial organisation, like collective bargaining and unionisation, although at times this is in fact what is assumed. The latter are forms, albeit historically important ones, of freedom of association.
Empirically, the construction of an index of freedom of association is a mix of subjective and objective classifications that are proxies for different indicators on the respect for rights of freedom of association in a country. Countries are then often classified according to a judgement ⁽⁶⁵⁾.
The first indicator used here for illustration covers what its developers call freedom of organisation. The measure is a ranking based on the degree of harassment of trades unions and interest groups from the scale of 1 (least) to 4 (most) in 72 countries in 1985.

It is very clear here, that while improvement with national income takes place, a high dispersion, indeed a very wide range of variation in ranks is taken by countries in each of the three income classifications. This sustains the point about the possibility of achieving higher or lower levels of this freedom of organisation with respect to national income.
Freedom of association is arguably also reflected in rates of unionisation of the paid work force. Figure 12 below shows union densities rates for 58 countries. If we assume that there is some link between unionisation and the existence of collective bargaining, which is reasonable, than this is also what the relationship in this figure tells us. Once again the scatter is wide enough to suggest the possibility of high or low levels of union density associated with national income, though the relationship itself is fairly weak.

We need to be cautious in interpretation here on conceptual grounds, because freedom of association as suggested is a wider notion than either the harassment indicators related to trades unions and political groups in Figure 11, or the indicator of unionisation in Figure 12, can capture. This can be illustrated by looking at global trends in unionisation. Table 5.2 below shows trends in union densities in regions of the world.
Table 5.2 Changes in Union Density (number of countries) 1985-1998

Source: J. Visser, Trends in Unionization and Union Practice, (ILO mimeo) (2000).
The first point to note is that in a majority of countries in the 1980s and 1990s, for which estimates exist, union densities are going down. Secondly that these declines seem to be severest in the developed economies of Europe. The decline in unionization rates in developed economies where the freedom to associate is clearly high, suggests that there may be reasons other than "restriction" which explain trends. For example in highly developed economies, there is evidence that free-lancing and working from home for the self-employed is on the ascendence as a work form. Moreover new forms of associations are also coming in to being because of the IT revolution, some of which are based on vocational or professional criteria.

Similarly, there is a point to note about "coverage" with respect to the issue of unionization/ bargaining in a typical developing country. There are clearly many parts of the labour market in a developing economy which can enjoy freedoms of association at the work place, but which are restricted (for example, casual workers used in formal enterprises). However in many parts of the developing country labour market the underlying conditions, for unionization and collective bargaining, may be absent. These are situations where principals and agents do not exist in a manner as assumed by a typical labour contract. For example, many categories of the self employed especially in the informal sector shown in Figure 2, are often not any better off than the small workforces of one or two that may occasionally work for them. In such cases it is not clear who makes an association with whom. Casual daily wage workers, who change employers every day, can be another example both in rural and urban areas. In these instances there is also the issue of the instability of work categories. People in such work change occupations fairly regularly. Home work is another obvious case which takes work in to the space of the family. Such cases also make it difficult for one to say that the "absence" of an association is due to a "restriction". New forms of organizations that mobilise groups of agents (e.g. women's organizations) without specifically targeted principals are also coming about.
It needs to be appreciated that although the right to associate and organize is a general and fundamental one, the forms that this right takes are historically determined, and the existence of that right does not mean that specific forms of associations and organization are automatically formed for all. The lack of their formation may of course have something to do with restriction, but it may also have something to do with relevance of the "expected form" of the freedom of association to the situation at hand. The specific point is that there are assumptions - which are made with respect to the regularity of work, the change across categories of work, and the nature of contracting parties (principal and agent) involved in a labour contract - that underlie many of the ideas concerning the freedom of association as we "normally" understand it. The notion is a wider one and it is important to recognise the historicity of its form.
There are however three broad reasons why freedom of association, understood restrictively, is still important to the development process in developing countries. Firstly, parts of labour force that can enjoy freedom of association still do not enjoy it in developing countries. Secondly, there is an expectation that with economic expansion, forms of work in which the role of rights in association, bargaining and coalition forming are relevant will increase in developing countries. Third, there may be instrumentalities between rights, such that groups who have these rights may play instrumental and participatory roles in promoting other rights for their groups as well as getting the same rights for those who do not enjoy them in the labour force. This is of course an empirical question.

Civil and political rights

The literature on civil and political liberties and economic development is vast. It is largely based in an empirical tradition in political science. Arguments can be developed marshalling positions ranging from neutrality of societal freedoms to growth, to its being antagonistic or complementary. Figure 13 below uses data on an index of civil liberties produced by Freedom House ⁽⁶⁶⁾ for the 1990s. This shows characteristic improvement of rights with national income, as well as a wide enough range of values, to suggest that varying social climates of freedom are compatible within each bracket of economic prosperity.

The results with political rights and democracy indicators from the same data source are very similar.

The civil and political rights experience of developing economies

Research on the link between civil/political rights on the one hand and economic performance on the other has been used mostly to suggest either that more rights constrain economic performance or that they do not affect it all. It is clear that while there is reason to expect more civil and political rights and democracy in higher income economies, there is little reason to think that it is not possible to intervene to get these at lower levels of income. The issue of rights constraining economic performance involves a counterfactual and is clearly deeper.
This argument should be considered carefully, because historical experience, especially in the more advanced economies of today, tells us that formal civil and political rights in many respects followed or were "simultaneous with" economic growth, and at least that they did not precede it. Moreover groups drawn on the basis of ethnicity, race and gender were last to achieve these rights. The prescription for a fully functioning democracy in the developing world is sometimes seen as a reversal of this historical logic. This may be true in a sequential sense with respect to the past, but to infer from this that there is little scope and need for true democracy to exist in developing countries is not quite legitimate. There are three reasons for this. First since democracy at least tends to formally protect the basic rights of majorities in principle it is an independent goal that ought to be pursued on this count alone. Second, there is no conclusive counterfactual to suggest that had formal rights come in earlier than they did, say in Western Europe, there would have been a loss to growth. Thirdly, rights achieved in many of today's older industrial economies, were in part the result of people's struggles to achieve those rights. In today's environment there are other avenues like international consensus, that can arguably bring support to struggles for rights within developing countries, such that they may come sooner then they would have without such consensus ⁽⁶⁷⁾.
Coming to the present day developing countries, the matter economic performance and political regime has been subject to a more through examination. For specific periods in recent economic history just as economies of South East Asia like Republic of Korea, Taiwan (China) and Singapore registered phenomenally high growth rates under autocratic regimes, many other countries also under authoritarian regimes like Zaire, Uganda, Haiti, Mozambique or South Africa have not prospered. Similarly countries with open political regimes, as Botswana and Malta have done reasonably well on growth, just as Gambia and Jamaica have not done so well. There are cases of modest performance under both types of regimes as well. There is no clear cut evidence at a county case level to show that non democratic environments tend to help growth, just as there is little reason to believe that democracy invariably leads to lower long run growth. Evidence suggests agnosticism about implications of democracy on economic growth, just as it does on the economic implications of autocracy. The assessment also implies that growth, and employment, do not (in themselves) generate rights just as rights on their own do not generate growth and employment. This is the case despite the fact that there are proportionately more democratic regimes in high income economies ⁽⁶⁸⁾. There is however some evidence on the stability in economic performance with respect to authoritarian and democratic regimes. For example it has been argued with respect to the degree of de-centralisation in regimes, that short run volatility in economic performance is less in regimes where decision-making is dispersed ⁽⁶⁹⁾. More directly it has been argued that variation in long run income growth is definitely lower in democracies than it is in autocracies ⁽⁷⁰⁾.
These results suggest two broad points. First, the obvious point, that although growth and moving up the income ladder, increase the chances of a country enjoying greater civil, political and indeed social rights, this link with income is not necessary for enjoying rights. The assumption here is of course that there is some intrinsic value to having these rights to start with ⁽⁷¹⁾. Second, that even though definitive claims are difficult to sustain about the advantages of democratic political regimes for long-run growth, the autocratic country may be a riskier proposition for an ordinary citizen to live under on grounds of income stability.

Institutions and information

Economic development also produces market and civil institutions. The extent to which well being is delivered to all in a given economy apart from its structural features and social policy also depends on the both on quality of institutions it possesses and the accuracy of available information that it disseminates.
It is also important to note that the ability of a society to participate in decisions is fundamentally dependent upon both the quality of institutions and of information. Whatever the specific nature of these institutions, it is absolutely clear that if they exist, a lack of participation in them benefits minorities. To make institutions work for all, participation is necessary both at level of the work place as well as the level of the State. Institutional democracy is also central to making the knowledge about benefits of participation become known and acceptable as the norm in society. The make up of the State has therefore a role in institutional operation and knowledge dissemination. It is intuitively plausible then to suggest that for an effective set of rights to operate both informational and institutional conditions need to be satisfied in society.
The problem of using qualitative indicators is even more acute here than it is the case for indicators like civil liberties, but there is no option but to use these indices. Figure 14 below gives an index of institutional quality, where the lower the score the better the institutional quality ⁽⁷²⁾. It suggests that institutional quality increases with national income. However interestingly the scatter suggests that although there is some room for policy intervention at lower levels of income, the range in which this improvement can be expected is restricted at these levels ⁽⁷³⁾.
Therefore, while institutional quality definitely improves with the level of national income, the variation in values for low income countries covers a narrower range than that of medium and high income countries. To move beyond the best levels of institutional quality achieved in low income countries may require exogenous factors. However this pattern is somewhat less true for an index of information quality which ranks countries according to the degree of State control on information.

Figure 15, which ranks better information quality with a lower score, shows that access to balanced information obtains for a slightly wider range of countries across income and though we can expect to improve on this with economic growth as well, good and bad information quality is more compatible with a wider range of national income. Thus interventions are feasible for many countries. Although the data periods of the two figures are separated by a decade and the differences may be subtle, the results of Figures 14 and 15 are interesting taken together. They suggest a hypothesis to be explored : that institutional quality may tied up to the economic growth process more closely than information quality.

6. How are decent work dimensions related to each other at different levels of national income?

The question of what determines the indicators in each dimension is an issue for empirical research concerning causality. This requires careful empirical modelling as well as data of improved quality. However, correlations between different dimensions of decent work can be explored, to see if some dimensions are more closely connected to each other in low and middle income countries respectively. It was noted that recent empirical work does suggest that social dimensions improve in a certain sequence with individual life conditions improving first and fast and the structural or social organisational components changing more slowly and unevenly ⁽⁷⁴⁾. Table 6.1A and 6.1B, show correlation coefficients between selected indicators (all used before in Section 4) for low and middle income countries separately. These correlations are produced on the basis of pair wise observations.
For the choice of indicators in employment informal employment as well as the working poor rate have been chosen. The data as well as the coverage as we know for formal employment is restricted for low income developing countries . In this case then, while urban informality is likely to capture the employment situation in both low and middle income countries, the working poor are likely to be a slightly more relevant indicator for countries at lower incomes. On security, the social security expenditure variable chosen is the wider indicator covering all and is likely to be relevant in both low and middle income countries. The size of the socially protected population on the other hand, is likely to be more relevant in middle income countries. On representation, union densities are likely to be of relatively more importance in middle income countries, while civil liberties the index applicable to many more is likely to be relevant in both low and middle income countries. The gender primary enrollment bias and child labour are the two indicators taken for rights at work ⁽⁷⁵⁾. They are expected to generally applicable. Links between these variables are likely to be co-determined by many factors that cannot be assessed at this level. We already know that many of these variables show relationships mediated by national income.
The results are interesting and show a discernible pattern. If we look at the statistically significant correlations what is clear is that they increase from 11 out of 64 in the low income countries group to 32 out of 64 in the middle income countries group. Table 6.2 below gives a count of these statistically significant correlations.

Table 6.1A Low income countries: Correlations across the decent work dimensions

		Urban informal employment (as % urban emp)	Working poor rate	Socially protected population %	Social security expenditure (% of gdp)	Gender enrollment ratio	Child labour	Union density	(Lack of) civil liberty index
Urban informal employment (as % urban emp)	Pearson Correlation	1.000	.612(*)	.436	-0.107	0.062	0.358	0.437	-0.034
	Sig. (1-tailed)	.	.023	.194	.409	.414	.072	.090	.446
	N	18	11	6	7	15	18	11	18
Working poor rate	Pearson Correlation	.612(*)	1.000	.286	-.608(*)	.375	.487(*)	-.211	-.311
	Sig. (1-tailed)	.023	.	.321	.031	.052	.015	.255	.091
	N	11	20	5	10	20	20	12	20
Social protected population %	Pearson Correlation	.436	.286	1.000	.428	.005	-.239	.333	-.099
	Sig. (1-tailed)	.194	.321	.	.169	.495	.227	.233	.380
	N	6	5	12	7	8	12	7	12
Social security expenditure (% of GDP)	Pearson Correlation	-.107	-.608(*)	.428	1.000	-.206	-.570(*)	.344	-.353
	Sig. (1-tailed)	.409	.031	.169	.	.231	.011	.202	.090
	N	7	10	7	16	15	16	8	16
Gender enrollment ratio	Pearson Correlation	.062	.375	.005	-.206	1.000	.465(**)	-.381	-.018
	Sig. (1-tailed)	.414	.052	.495	.231	.	.001	.066	.454
	N	15	20	8	15	43	42	17	43
Child labour	Pearson Correlation	.358	.487(*)	-.239	-.570(*)	.465(**)	1.000	-.549(**)	-.074
	Sig. (1-tailed)	.072	.015	.227	.011	.001	.	.008	.285
	N	18	20	12	16	42	61	19	61
Union density	Pearson Correlation	.437	-.211	.333	.344	-.381	-.549(**)	1.000	.253
	Sig. (1-tailed)	.090	.255	.233	.202	.066	.008	.	.148
	N	11	12	7	8	17	19	19	19
(Lack of) civil liberty index	Pearson Correlation	-.034	-.311	-.099	-.353	-.018	-.074	.253	1.000
	Sig. (1-tailed)	.446	.091	.380	.090	.454	.285	.148	.
	N	18	20	12	16	43	61	19	63

* Correlation is significant at the 0.05 level (1-tailed)
** Correlation is significant at the 0.01 level (1-tailed)

Table 6.1B Middle income countries: Correlations across the decent work dimensions

		Urban informal employment (as % urban emp)	Working poor rate	Social protected population %	Social security expenditure (% of GDP)	Gender enrollment ratio	Child labour	Union density	(Lack of) civil liberty index
Urban informal employment (as % urban employment)	Pearson Correlation	1.000	.653(**)	-.897(*)	-.476(*)	.035	.359(*)	-.687(**)	-.073
	Sig. (1-tailed)	.	.006	.019	.015	.436	.030	.002	.356
	N	28	14	5	21	23	28	15	28
Working poor rate	Pearson Correlation	.653(**)	1.000	-.740(*)	-.405(*)	-.379(*)	.435(*)	-.433	-.049
	Sig. (1-tailed)	.006	.	.046	.028	.026	.012	.053	.403
	N	14	27	6	23	27	27	15	27
Social protected population %	Pearson Correlation	-.897(*)	-.740(*)	1.000	.629(*)	.140	-.733(**)	.948(**)	-.274
	Sig. (1-tailed)	.019	.046	.	.026	.341	.003	.007	.183
	N	5	6	13	10	11	12	5	13
Social security expenditure (% of GDP)	Pearson Correlation	-.476(*)	-.405(*)	.629(*)	1.000	-.067	-.507(**)	.672(**)	-.133
	Sig. (1-tailed)	.015	.028	.026	.	.346	.000	.000	.195
	N	21	23	10	45	38	43	21	44
Gender enrollment ratio	Pearson Correlation	.035	-.379(*)	.140	-.067	1.000	-.037	.318	.450(**)
	Sig. (1-tailed)	.436	.026	.341	.346	.	.390	.070	.000
	N	23	27	11	38	62	59	23	60
Child labour	Pearson Correlation	.359(*)	.435(*)	-.733(**)	-.507(**)	-.037	1.000	-.523(**)	-0.037
	Sig. (1-tailed)	.030	.012	.003	.000	.390	.	.002	.379
	N	28	27	12	43	59	74	28	71
Union density	Pearson Correlation	-.687(**)	-.433	.948(**)	.672(**)	.318	-.523(**)	1.000	.224
	Sig. (1-tailed)	.002	.053	.007	.000	.070	.002	.	.121
	N	15	15	5	21	23	28	29	29
(Lack of) civil liberty index	Pearson Correlation	-.073	-.049	-.274	-.133	.450(**)	-0.037	0.224	1.000
	Sig. (1-tailed)	.356	.403	.183	.195	.000	.379	.121	.
	N	28	27	13	44	60	71	29	86

* Correlation is significant at the 0.05 level (1-tailed)
** Correlation is significant at the 0.01 level (1-tailed)

Table 6.2. Statistically significant correlations across indicators

	Low Income Countries	Middle Income Countries
Working poor rate	3 / 8	5 / 8
Urban informal employment (as % urban emp..)	1 / 8	5 / 8
Social security expenditure (% of GDP)	2 / 8	5 / 8
Socially protected population %	0 / 8	5 / 8
Gender enrollment ratio (m : f)	0 / 8	2 / 8
Child labour	4 / 8	5 / 8
Lack of civil liberty index	0 / 8	1 / 8
Union density	1 / 8	4 / 8
Number of significant correlations	11 / 64 (17%)	32 / 64 (50%)
Note: The numerator count excludes the variable's correlation to itself. Source : Table 6.1A and 6.1B.

Looking at the linkages between dimensions in Table 6.1A, in low income countries the working poor, social security expenditures and child labour are the ones in which we find linkages in expected directions. So in a senses the employment-security link is there. Child labour insofar as it is also connected to poverty fits well into the picture here ⁽⁷⁶⁾.
In the case of middle income countries the linkages, in Table 6.1B, are wider across dimensions. Firstly, the employment security linkage becomes stronger compared to Table 6.1A and both the working poor indicator and the informality indicator get linked to each of the security indicators (protected persons and expenditure on social security). Second, the union density indicator reflecting freedom of association gets connected to both the security indicators. Child Labour, gets connected to both indicators of employment and security, as well as union density. The gender enrollment ratio gets connected to civil liberties ⁽⁷⁷⁾, which makes intuitive sense too.
There is not much that can be inferred either about causality between indicators or about the choices of policy instruments from the above analysis. However, correlations across dimensions do seem to indicate that there are patterns to the way these indicators get connected across income categories. In low income countries the linkages may dominantly be employment and security based, in middle income countries they are likely to be more generally connected. These findings are consistent with the aforementioned results on sequencing in the empirical literature. Future work on decent work dimensions can take these observations in to account.

7. Is there a case for expecting simultaneous achievement in decent work dimensions?

It was shown in Section 5 that there are grounds for intervening on the dimensions of decent work in developing countries, given the general positive expectation between economic prosperity and achieving more decent work. This was found to be plausible both at a societal level as well as the level of work, the distinction that we started with. In Section 6, there was an attempt to assess if there were any systematic interrelationships between dimensions of decent work. We found that there may be grounds for expecting some patterns here as well. The question that is asked in this section is whether achievements on some of these indicators may have simultaneity effects on achievement on other indicators.
Any assessment of simultaneity needs to control for the effect of national income, in other words isolating growth from indicator performance. This is because we know that national income and its growth are generally positively linked to improvements in decent work dimensions, therefore an indicator of country "performance" needs to be one of performance despite the national income effect. In other words, good performers are those countries that do better on an indicator than what would be expected simply on the basis of national income. Similarly bad performers are those countries that do worse on a given indicator than what would be expected on the basis of national income. Having classified countries in this way we need to examine if indicators move together in some sense.
Since the procedure we are using for examining simultaneity in country performance on indicators involves partitioning and analyzing converted residuals (converted in to binary values), we can call it a qualitative analysis of partitioned residuals. This test is described below.

7.1. Country and indicator selection

The first step is to select a few indicators of dimensions of decent work on which the information is complete for a large enough set of developing countries. Six indicators were chosen for forty-seven countries. In each of these indicators the greater value is indicative of worse performance.
These were:

• Working Poor Rate (Employment)
• Adult Illiteracy (Security)
• Life Expectancy (Security)
• Child Labour (Rights at work)
• Net Enrollment Ratio Males :Female (Rights at work)
• Civil Liberties (Representation)

The indicators correspond in definition to those used in earlier sections of this paper. The two indicators on security are damage indicators indicative of societal insecurity. One of the main determinants on the choice of indicators was that we needed our sample of countries to be of reasonable size and the data set to be full.

7.2. A qualitative analysis of partitioned residuals

The second step is to run two-variable simple OLS regressions with respect to national income for each of the indicators for the 47 countries. These are the same exercises as in section 5, except that these are done now for our full data set on the six variables for the reduced set of 47 countries. The next step is to list residuals from each of the regression equations according to value and sign. The argument is that negative residuals (actual minus predicted values) mean that a country has performed better than what would be expected purely on the basis of its national income, and a positive residual that it has performed worse than what is expected on the basis of its national income. One can then proceed to partition the residuals to assess degree of good and bad performance. The choice of partition is arbitrary though it needs to be consistently applied in each of the six variable series. Our partition of the residual data is done on the following basis. There are 47 countries, so we take the top 15 negative residuals as best performers and give them a value of -1, and the bottom 15 (positive) residuals as the worst performers and give them the value of + 1. The 17 countries in the middle are considered so close to the regression line that they are deemed as performing according to expectation. These are neutral performers and are given the value of 0. Table 7.1 gives the residual list transformed in to the -1, 0 and + 1 classification. Table 7.2 gives net scores on the six indicators for each country.
There are conflicts between indicators in countries as Table 7.1 shows, and doing well on one indicator does not automatically mean doing well on others. Table 7.2 suggests that there is a large number of countries that show some simultaneous movement on the indicators, good or bad, in net terms ⁽⁷⁸⁾.

7.3. Country performance

The next step is to examine the performance of good and deficiently performing countries on the six selected indicators according some criteria. We use the following definitions for grouping countries based on Table 7.1 and Table 7.2.

• A net good or bad performer can be any country that is either performing well or badly on at least one indicator in net terms. So for example if a country performs badly on three and well on three indicators it shows neither good nor bad performance. However if it performs well on two and badly on one indicator, it performs well in net terms. This is clearly a weak indicator of performance. In terms of Table 7.2, all non-zero positive score countries are bad performers and all negative scores are good performers.
• The second classification is that of net simultaneity. One way of defining the degree of good simultaneity would be to say that in the group of net good performers defined above, any country that performs better than expected on two or more indicators in net terms ⁽⁷⁹⁾ shows good net simultaneity. Similarly a way of defining the degree of net bad simultaneity would be to say that in the group of deficient performers, any country that in net terms, performs worse than expected on two or more indicators, shows bad net simultaneity. In terms of Table 7.2 all countries with values of -2 to -5 show net good simultaneity, and countries scoring +2 to +5 show net bad simultaneity.

Table 7.1. Classification of residuals

	Country	Child labour	Life expectancy gap	Gender enrollment bias	Working poor rate	Adult illiteracy	Civil liberties
Algeria	0	0	1	-1	1	1
Bangladesh	0	0	0	0	0	0
Brazil	1	1	1	0	1	1
Bulgaria	-1	-1	0	-1	-1	0
Burkina Faso	0	1	1	1	0	0
Chile	-1	-1	1	0	0	-1
China	1	-1	-1	1	0	1
Colombia	1	0	0	1	1	1
Cote d'Ivoire	0	1	1	-1	1	1
Czech Republic	-1	0	1	0	-1	-1
Dominican Republic	1	-1	-1	-1	1	0
Ecuador	0	-1	-1	1	0	1
Egypt, Arab Rep.	1	0	1	-1	1	1
El Salvador	1	0	-1	1	-1	0
Estonia	-1	1	0	0	-1	-1
Ethiopia	0	0	1	-1	-1	0
Honduras	0	-1	-1	1	0	-1
Hungary	-1	1	1	-1	-1	-1
India	0	0	1	1	1	1
Indonesia	1	0	-1	0	0	1
Jamaica	-1	-1	0	-1	0	0
Jordan	-1	-1	-1	-1	0	1
Kenya	0	0	-1	0	0	1
Latvia	-1	0	0	-1	-1	-1
Lesotho	1	1	-1	1	0	1
Madagascar	0	-1	-1	0	0	-1
Mali	0	0	1	1	0	-1
Mauritius	1	1	0	0	1	-1
Mongolia	-1	-1	-1	0	0	-1
Mozambique	0	1	1	0	-1	-1
Namibia	1	1	-1	1	1	0
Nepal	1	0	1	0	1	0
Nicaragua	0	-1	-1	-1	1	-1
Niger	0	1	1	0	0	1
Panama	1	-1	0	0	1	0
Paraguay	1	0	0	1	-1	0
Peru	0	0	0	1	0	0
Poland	-1	0	0	0	-1	-1
Romania	-1	1	0	-1	-1	0
Russian Federation	-1	1	0	0	-1	1
Senegal	0	1	1	0	1	0
Tanzania	-1	-1	-1	-1	-1	0
Thailand	1	-1	0	-1	-1	0
Trinidad and Tobago	-1	0	0	1	0	-1
Turkey	1	1	0	-1	1	1
Venezuela, RB	0	-1	0	1	1	0
Zambia	-1	1	-1	1	-1	-1

Table 7.2: Net country scores in ascending order

Country	Net scores
Tanzania	-5
Bulgaria	-4
Latvia	-4
Mongolia	-4
Jamaica	-3
Jordan	-3
Madagascar	-3
Nicaragua	-3
Poland	-3
Chile	-2
Czech Republic	-2
Estonia	-2
Honduras	-2
Hungary	-2
Romania	-2
Thailand	-2
Zambia	-2
Dominican Republic	-1
Ethiopia	-1
Trinidad and Tobago	-1
Bangladesh	0
Ecuador	0
El Salvador	0
Kenya	0
Mozambique	0
Russian Federation	0
China	1
Indonesia	1
Mali	1
Panama	1
Paraguay	1
Peru	1
Venezuela, RB	1
Algeria	2
Mauritius	2
Burkina Faso	3
Cote d'Ivoire	3
Egypt, Arab Rep.	3
Lesotho	3
Namibia	3
Nepal	3
Niger	3
Senegal	3
Turkey	3
Colombia	4
India	4
Brazil	5

Thirdly we can have a classification of pure simultaneity, of countries that perform well or badly, on the six indicators in non-conflicting ways. In other words countries that perform well at least on one indicator and well or neutrally but not badly on the rest of the indicators, would display pure good simultaneity. Similarly we can have a classification of countries that perform badly on at least one indicator and badly or neutrally but not well on the rest of the indicator, these would display pure bad simultaneity. These countries can be picked from Table 7.1.
Table 7.3. Simultaneity in country performances on six indicators

Net good performers	20
of which Net good simultaneity	17
of which Pure good simultaneity	7
Net bad performers	21
of which Net bad simultaneity	14
of which Pure bad simultaneity	7
Sample of neither good nor bad performers	6
Total sample	47

Table 7.3 above categorizes the results according to this group classification. We are more interested in the cases of simultaneity, especially of pure simultaneity. This seems to be high enough to warrant a further test.

7.4. A probability test for pure simultaneity ⁽⁸⁰⁾

A question we need to ask for cases of (- the 7 good and the 7 bad-) pure simultaneity cases is whether we would expect a number of countries greater or less than this number to emerge on probabilistic grounds given the same parameters of the problem?
In other words, in a sample of 47 countries, partitioned in 3 sections (one partition of 15 values of -1; one partition of 17 values of 0; one partition of 15 values of +1) on 6 indicators, what is the number of countries one would expect on probabilistic grounds to be in the following classifications:

• At least one value of -1 and no value of +1 for the six indicators (i.e only combinations of -1s and 0s).
• At least one value of +1 and no value of -1 for the six indicators (i.e only combinations of +1s and 0s)

We therefore need to calculate the expected number of pure simultaneity countries, S, which are the same for both good and bad complementarity categories since they are symmetrically defined. This can be expressed as:
S = n x u ⁱ x (1- v ⁱ)
Where n is the number of countries, u is the probability of a country getting -1 score or a 0 score. Similarly v is the probability of a country getting a score of +1 or a score of 0. The number of indicators is i.
If S < 7, which the number of countries in which pure simultaneity is observed then a case for simultaneity cannot be made. In our example,
n = Number of countries = 47
i = the number of indicators = 6
P ₁ = probability of achieving -1 = 15/ 47
P ₂ = probability of achieving 0 = 17/ 47
P ₃ = probability of achieving +1 = 15/47
P ₁ + P ₂ = u
P ₂ + P ₃ = v
v = u = 32/47 = 0.6808
u ⁶ = v ⁶= 0.0996
So
S = n x u ⁱ x (1- v ⁱ)
S = 47 x 0.0996 x (1 - 0.0996)
S = 47 x 0.0996 x 0.9003
S = 4.2
The number of pure simultaneity countries in our sample (good and bad respectively ) are 7, which is significantly greater than 4.2 (say 5), the number of countries we would expect at random to be displaying pure simultaneity. There is consequently a case to be made for the existence of pure simultaneity.

7.5. The reasons for the existence of simultaneity

We have found that there is a basis to make a case for expecting simultaneity in social dimensions of development related to decent work. The reason why simultaneity may obtain still needs to be explained. Although this is a subject of research in itself, it is quite likely that there are deeper features of societies that affect enabling conditions for achievement of the social indicators we have chosen. These are likely to be of both structural and institutional varieties. It is therefore useful to see if on a selection of structural and institutional proxy variables (not included in the indicators) these groups systematically vary.
To proxy a structural variable we use a gini-coefficient of the distribution of income, for institutional variables it is the institutional quality index used in Figure 14. Tables 7.4 and 7.5 below give results of a difference of means test on selected results of Table 7.3.
It is interesting to see that the structural and institutional proxies vary systematically with respect to the performance classifications. Generally speaking, the results are stronger in the pure tests in Table 7.5. While institutional quality makes a definite difference in performance

Table 7.4. Analysis of Variance: Net good and bad simultaneity

Groups	Count	Sum	Average	Variance
Institutions 14 +	10	36	3.6	0.71
Institutions 17-*	14	37	2.64	1.02
Gini 14 +	12	522.41	43.53	98.91
Gini 17 -***	17	656.68	38.63	100.32
*= difference significant at less than the 5% level **= difference significant at less than the 10% level ***= difference significant at less than the 20% level

Table 7.5. Analysis of Variance: Pure good and bad simultaneity

Groups	Count	Sum	Average	Variance
Institutions 7 +	6	22	3.67	1.02
Institutions 7-*	5	15	3	0.5
Gini 7 +	7	318.08	45.44	110.31
Gini 7 -*	7	240.07	34.3	34.97
*= difference significant at less than the 5% level **= difference significant at less than the 10% level ***= difference significant at less than the 20% level

in both Tables 7.4 and 7.5, income distribution clearly makes a more significant difference at the pure level in Table 7.5 than it does in Table 7.4. Thus countries that perform better in net and pure terms respectively in improving on social dimensions, abstracting from the level of national income, tend to have better income distributions and institutional quality than countries in the comparable categories that do badly. The processes involved here are a subject of future research.

8. Conclusions

Decent work's development objective is evaluative and focussed on the improvement in the actual lives lived by people. This well-being centred perspective means that both freedom and capability enhancement play a fundamental role in its conception. Decent work dimensions in which policy interventions can take place are located both at the workplace and in the society that contains it. Although today, developments in economics allow for justifying policy interventions for the realisation of such a human development goal more than before, there is much work ahead before a larger human development framework emerges, that can give direct support to agendas of some international development agencies and help them move forward.
As far as basic facts go, the deficiency of data on decent work indicators of employment and security is serious, but this information shortage does not mean that we cannot say anything factual about the agenda. Broadly for the period of the mid-1990s, we find that formal workforces (narrowly defined) are very limited in number in developing countries, and are on average likely to range between 7 to 16.7 per cent of the labour forces in low and middle income countries respectively. Correspondingly on average the range in which persons who enjoy some social protection in low and middle income countries accords with the figures on formal labour forces (i.e. populations that are socially protected range between 6.5 to 17.3 per cent of the population in low and middle income countries respectively). This increasing incompleteness of the social floor as we move down from middle to low income countries is directly reflected in the expenditure by government on social security (defined broadly) as a percentage of national income. This figure is very low in low income countries (less than 3 per cent of GDP) and reasonable (around 9.5 per cent) in middle income countries. It is this linked "employment" and "insecurity" situation in developing countries that explains in large part the phenomenon of the "working poor" . The working poor are likely to constitute around 30 per cent of employed persons in low income countries and around 7 per cent in middle income countries. Working children in the age bracket of 10 to14 are also around 18.1 per cent in low income countries and 8.7 per cent in middle income countries. Civil liberties and measures of freedoms to organise are also better in middle income countries compared to low income countries but most of these measures are qualitative.
That the four dimensions of the ILO agenda, by and large, improve with the economic growth of society is an undeniable fact. However, this relationship between the expansion of aggregate income in a society and decent work is not a deterministic one. Some of the preceding discussion was focussed on illustrating the point that there are grounds to doubt such determinism which still influences policy views amongst decision makers today. Essentially variations in the indicators with respect to national income, though suggesting a strong systematic relationship with it, also suggest that there are other factors that explain this variation. These other factors, it is argued, must include policy interventions that can directly improve (or indeed worsen) these indicators. There is consequently much that can be done to improve on the dimensions of decent work in many countries of world. The targets for this improvement need to be set within societies through consensus.
What is also discernible is that there are some grounds for expecting distinct patterns between dimensions of decent work for countries at low and middle income levels respectively. These show employment and security to be closely tied to each other at lower levels of income, while a general pattern of linkage is observable across all dimensions in middle income countries. This may be another useful area for future policy research.
Thirdly it can also be argued that good performance in the same direction on several indicators, independently of the national income effect, can be significantly shown for some groups of countries. There is consequently a case that can be explored for possible complementarities between dimensions of decent work. There may of course be deeper reasons why some indices of social well being, (in this case argued to be decent work related), display a tendency to move together. These may have something to do with both issues of distribution and governance, in other words with "equality" and "institutions", and if this is so then it has important implications for the policy framework to promote decent work in country specific contexts.
Conceptually the issue of causality, of what leads to what, at least at the level of ideas suggests that relationships are likely to be interactive. It is however clear that the proper establishment of directions of relationships discussed here would require the examination of country specific processes as well as of structure. This would entail looking more deeply and at details. There is much to be done in moving beyond associative illustrations and going on to establishing causal links to inform policy. Fortunately there is work going on in the academic world in this direction although empirical research is still in its early stages. This paper tries to set up some basic facts related to the decent work agenda in order to facilitate the on-going debate among decision makers.

>Annex

Table A1: Data set on indicators of decent work

Obs.	Class	Country	Formal employment as a % of LF 1995	Urban informal empl. as % of LF 1990s	Working poor rate 1997	Social security exp. (% of GDP 1996)	Adult illiteracy 1997	Life expectancy	Gender inequality Index 1992	Gender enrollment ratio 1995	Child labour 1995	Unfreedom of organisation	Union density 1995	(Lack of) civil liberties index	Institutional quality index 1997	Institutional quality index 1985
1	L	Afghanistan						45.49	0.26	2.01	25.25			7.00
2	M	Albania				10.85		71.68		0.97	1.11			4.00
3	M	Algeria			2.00		38.40	70.35	0.31	1.08	1.63	4.00		6.00	5.00	4.00
4	M	American Samoa
5	H	Andorra										1.00		1.00		1.00
6	L	Angola						46.46	0.02	1.04	27.09			6.00
7	M	Antigua and Barbuda						74.73					53.80	3.00
8	M	Argentina		42.41		12.40		73.11	0.13	1.00	4.53		25.40	3.00
9	L	Armenia						73.71			0.00			4.00
10	H	Aruba
11	H	Australia	72.50			15.70		78.18	0.03	1.00	0.00	1.00	33.00	1.00	1.00	1.00
12	H	Austria				26.20		77.39	0.05	1.00	0.00	1.00		1.00	1.00	1.00
13	L	Azerbaijan						70.91			0.00		75.40	5.00
14	H	Bahamas, The						73.72	0.07	1.00	0.00	1.00		2.00	2.00	2.00
15	M	Bahrain				4.21		72.79	0.20	1.00	0.00
16	L	Bangladesh		10.00	29.10		61.20	58.15	0.08	1.16	30.12	2.00	4.30	4.00		2.00
17	M	Barbados						76.14	0.02	1.06	0.00			1.00
18	M	Belarus				17.39	1.00	68.46			0.00		96.10	6.00
19	H	Belgium				27.13		77.79	0.08	1.00	0.00	1.00		2.00	1.00	1.00
20	M	Belize				3.52		74.72		1.00	2.38			1.00
21	L	Benin		47.90		2.24		53.41	0.05	1.71	27.52			2.00
22	H	Bermuda
23	L	Bhutan						60.72			55.10			7.00
24	M	Bolivia		52.57		7.05		61.46	0.12	1.04	14.36	1.00		3.00	3.00	2.00
25	M	Bosnia and Herzegovina						73.13			0.02			5.00
26	M	Botswana		19.30		2.68		47.27	0.09	0.95	16.92	1.00	11.50	2.00	2.00	2.00
27	M	Brazil		42.60	5.10	12.16	16.00	66.95	0.12	1.10	16.15	1.00		4.00	3.00	1.00
28	H	Brunei						75.69	0.07	0.99	0.00			5.00	2.00
29	M	Bulgaria			2.00	13.18	2.00	70.71		0.98	0.00	4.00		3.00	3.00	4.00
30	L	Burkina Faso			61.20		81.30	44.38	0.06	1.54	51.09			4.00
31	L	Burundi				2.20		42.37	0.04	1.16	48.97			7.00
32	L	Cambodia						53.97		1.00	24.66			6.00
33	L	Cameroon		57.30				54.46	0.08	1.09	25.25		14.70	5.00
34	H	Canada				17.70		78.90	0.06	1.00	0.00	1.00		1.00	1.00	1.00
35	M	Cape Verde						68.33	0.06	1.00	14.20		16.90	2.00
36	H	Cayman Islands
37	L	Central African Republic						44.85	0.03	1.46				5.00
38	L	Chad						48.56	0.12	1.83	38.32			5.00
39	H	Channel Islands						78.40
40	M	Chile		41.36	4.20	11.33	5.00	75.23	0.14	1.02	0.00	2.00		2.00	1.00	2.00
41	L	China	15.70		18.50	3.57	17.20	69.66	0.03	1.00	11.55	4.00	54.70	7.00	3.00	4.00
42	M	Colombia		51.61	11.00	6.09	9.10	70.23	0.14	1.00	6.62	1.00	7.00	4.00	4.00	1.00
43	L	Comoros						59.78	0.03	1.20	39.25			4.00
44	L	Congo, Dem. Rep.						50.71	0.03	1.43	29.56			6.00
45	L	Congo, Rep.				4.17		48.50		1.06	26.03			4.00
46	M	Costa Rica		38.12		13.03	4.90	76.54			5.48	1.00	13.10	2.00	2.00	1.00
47	L	Cote d'Ivoire		52.70	12.30		59.90	46.69	0.08	1.33	20.46	4.00	13.00	5.00	3.00	3.00
48	M	Croatia				22.32		72.50		1.00	0.00			4.00
49	M	Cuba						76.05	0.06	1.00	0.00			7.00
50	H	Cyprus				10.26		77.59			0.00		53.70	1.00	1.00
51	M	Czech Republic			2.00	18.84	1.00	73.91	0.02	1.00	0.00		42.80	2.00
52	H	Denmark	80.70			33.01		75.69	0.02	1.00	0.00	1.00		1.00	1.00	1.00
53	M	Djibouti						50.31	0.06	1.33				6.00
54	M	Dominica				4.82		75.95						1.00
55	M	Dominican Republic			3.20	2.49	17.40	70.95	0.16	0.96	16.06	1.00	17.30	3.00	3.00	2.00
56	M	Ecuador		45.19	20.20	2.00	9.90	70.25	0.18	1.00	5.43	1.00	9.80	4.00	4.00	2.00
57	M	Egypt, Arab Rep.			3.10	5.40	48.60	66.26	0.26	1.13	11.22	2.00	29.60	6.00	3.00	2.00
58	M	El Salvador			25.30	3.56	1.00	69.43	0.08	0.99	15.17		7.20	3.00
59	M	Equatorial Guinea						49.96		0.98	33.90
60	L	Eritrea						50.81		1.10	39.62		7.20	4.00
61	M	Estonia			4.90	17.11	1.00	70.08	0.03	1.00	0.00		26.40	2.00	2.00
62	L	Ethiopia		33.00	31.30	3.69	64.50	43.33	0.04	1.61	42.30	4.00	4.10	5.00	3.00	4.00
63	H	Faeroe Islands
64	M	Fiji		43.00				72.56	0.16	1.00	0.00			3.00
65	H	Finland	68.10			32.27		76.86	0.02	1.00	0.00	1.00		1.00	1.00	1.00
66	H	France				30.07		78.36	0.03	1.00	0.00	1.00		2.00	1.00	1.00
67	H	French Polynesia						71.89		1.00
68	M	Gabon						52.42			18.37		2.00	4.00
69	L	Gambia, The		67.25				53.16	0.07	1.26	36.99			6.00
70	M	Georgia						72.55		0.98	0.00			4.00
71	H	Germany				29.69		76.65		1.00	0.00	1.00		2.00	1.00	1.00
72	L	Ghana	79.00	78.50				60.01	0.05		13.27	3.00		4.00	3.00	3.00
73	H	Greece				22.67		77.73	0.09	1.00	0.00	1.00	24.30	3.00	2.00	2.00
74	H	Greenland
75	M	Grenada						71.77	0.19					2.00
76	H	Guam						77.15
77	M	Guatemala						64.23	0.19	1.10	16.22	2.00		4.00	4.00	2.00
78	L	Guinea						46.49	0.10	1.74	34.03	3.00	2.50	5.00	4.00	3.00
79	L	Guinea-Bissau						43.77	0.06	1.69	38.56	4.00		4.00	5.00	4.00
80	M	Guyana				5.79		64.42	0.06	0.99	0.00			2.00
81	L	Haiti						53.74	0.02		25.30			5.00
82	L	Honduras		44.87	40.50		29.30	69.44	0.09	0.98	8.52	1.00		3.00	4.00	1.00
83	H	Hong Kong, China						78.68	0.06	0.97	0.00		21.10		1.00
84	M	Hungary	78.40		2.00	22.30	1.00	70.64	0.02	1.01	0.17	4.00	52.50	2.00	1.00	3.00
85	H	Iceland				18.62		78.89		1.00	0.00	1.00		1.00	1.00	1.00
86	L	India		44.20	44.20	1.81	46.50	63.13	0.09	1.18	14.38	1.00	15.20	4.00	3.00	2.00
87	L	Indonesia	7.00	20.60	15.20	1.74	15.00	65.10	0.07	1.02	9.55	2.00	2.60	5.00	5.00	3.00
88	M	Iran, Islamic Rep.		17.90		6.07		70.23	0.21	1.04	4.71			7.00
89	M	Iraq						57.98	0.15	1.14	2.94			7.00
90	H	Ireland				17.81		75.94	0.11	1.00	0.00	1.00		1.00	1.00	1.00
91	M	Isle of Man	69.40
92	H	Israel				24.08		77.41			0.00		23.10	3.00	2.00
93	H	Italy				23.71		78.23	0.06	1.00	0.38	1.00		2.00	1.00	1.00
94	M	Jamaica		23.50	3.20	4.54	14.50	74.80	0.02	1.01	0.11	1.00		3.00	3.00	2.00
95	H	Japan				14.10		80.42	0.04	1.00	0.00		24.00	2.00	1.00
96	M	Jordan	79.20		2.00	8.88	12.80	70.85		0.99	0.68	3.00		4.00	2.00	2.00
97	M	Kazakhstan		11.30			1.00	64.46			0.00			5.00
98	L	Kenya		51.33	26.50	2.00	23.00	52.03	0.02	0.95	41.27	4.00	16.90	6.00	5.00	3.00
99	M	Kiribati						60.44						1.00
100	L	Korea, Dem. Rep.						62.95			0.00			7.00
101	M	Korea, Rep.				5.60		72.31	0.12	1.00	0.00		12.70	2.00
102	H	Kuwait						76.48	0.13	1.04	0.00	3.00		5.00	2.00	2.00
103	L	Kyrgyz Republic	53.00	11.90				66.89		1.00	0.00			4.00	4.00
104	L	Lao PDR						53.46		1.13	27.20			6.00
105	M	Latvia		9.10	2.00	19.17	1.00	69.91		1.02	0.00			2.00
106	M	Lebanon						69.86			0.00			5.00
107	L	Lesotho			43.10		30.50	55.97		0.84	22.11	2.00		4.00	3.00	3.00
108	L	Liberia						47.27			18.62			6.00
109	M	Libya						70.20		1.00	0.25			7.00
110	H	Liechtenstein										1.00		1.00		1.00
111	M	Lithuania		8.50		14.66	1.00	71.22	0.03		0.00			2.00
112	H	Luxembourg				25.20		76.48	0.12		0.00	1.00	43.40	1.00	1.00	1.00
113	H	Macao, China						77.54		0.96	0.00
114	M	Macedonia, FYR						72.35		1.02	0.04			3.00
115	L	Madagascar		57.50	60.20	1.29	54.20	57.46		0.96	35.83	2.00		4.00	4.00	3.00
116	L	Malawi						42.80	0.05	1.00	35.18			3.00
117	M	Malaysia				2.94		71.89	0.07	1.00	3.16	2.00	11.70	5.00	3.00	3.00
118	M	Maldives						66.98	0.06		5.73			6.00
119	L	Mali		71.00	72.80		70.70	50.35	0.12	1.52	54.53		13.70	2.00
120	H	Malta						77.03		1.00	0.00	2.00		1.00		2.00
121	M	Marshall Islands												1.00
122	L	Mauritania				0.76		53.46	0.14	1.16	23.97	4.00	2.70	6.00	5.00	3.00
123	M	Mauritius		24.00	3.80	5.95	17.10	70.40	0.12	1.00	3.00		25.90	2.00
124	M	Mayotte
125	M	Mexico		42.92		3.70		71.98	0.12	1.00	6.73	2.00		3.00	3.00	2.00
126	M	Micronesia, Fed. Sts.						66.95						1.00
127	L	Moldova				15.50	4.00	66.51			0.00			4.00	3.00
128	H	Monaco										1.00		1.00		1.00
129	L	Mongolia			13.90	8.84	17.10	65.76	0.01	0.96	1.90			3.00
130	M	Morocco				3.37	59.50	66.60	0.19	1.30	5.61	1.00		5.00	2.00	2.00
131	L	Mozambique			37.90	4.70	60.00	45.47	0.07	1.31	33.80			4.00
132	L	Myanmar						60.11	0.02	1.02	24.51			7.00
133	M	Namibia			34.90	3.87	72.50	55.63		0.95	21.68		22.00	3.00
134	L	Nepal			37.70		35.00	57.36	0.10	1.48	45.21	4.00		4.00	3.00	3.00
135	H	Netherlands				26.71		77.53	0.09	1.00	0.00	1.00		1.00	1.00	1.00
136	H	Netherlands Antilles						75.38			0.02
137	H	New Caledonia	82.90					72.56
138	H	New Zealand				19.22		77.13	0.06	1.00	0.00		21.70	1.00	1.00
139	L	Nicaragua			3.00	9.06	86.40	68.14	0.08	0.96	14.01	2.00	23.40	3.00	4.00	3.00
140	L	Niger			61.40		40.50	45.71	0.05	1.67	45.17	3.00		5.00	5.00	4.00
141	L	Nigeria					59.10	53.51	0.06		25.75	4.00		6.00	5.00	3.00
142	H	Northern Mariana Islands
143	H	Norway				28.50		78.14	0.02	1.00	0.00	1.00		1.00	1.00	1.00
144	M	Oman						72.86		1.03	0.51			6.00
145	L	Pakistan		67.10			8.90	61.68	0.26		17.67	2.00		5.00	4.00	3.00
146	M	Palau												2.00
147	M	Panama		32.92	10.30	11.29	7.90	74.04	0.11	1.01	3.51	2.00		3.00	2.00	1.00
148	M	Papua New Guinea						57.93	0.04		19.31			4.00
149	M	Paraguay		51.90	19.40		11.30	69.70	0.13	0.99	7.87		9.30	3.00
150	M	Peru		48.93	15.50		1.00	68.54	0.11	1.01	2.48	1.00		3.00	3.00	1.00
151	M	Philippines		17.00				68.30	0.08	1.00	8.04	2.00	22.80	3.00	2.00	2.00
152	M	Poland		12.80	5.40	25.10	3.10	72.65	0.02	1.00	0.00	3.00	27.00	2.00	2.00	3.00
153	H	Portugal				19.00		74.96	0.05	1.00	1.77	1.00		1.00	1.00	1.00
154	M	Puerto Rico						75.39			0.00
155	H	Qatar						74.35	0.24	0.97	0.00	3.00		6.00	2.00	4.00
156	M	Romania			2.80	12.45	0.50	69.00		1.00	0.17	4.00		3.00		4.00
157	M	Russian Federation			7.10	10.36	66.90	66.70	0.03	1.00	0.00			4.00
158	L	Rwanda						40.47			41.72	3.00		6.00	5.00	3.00
159	M	Samoa						68.57		1.00				2.00
160	L	Sao Tome and Principe						63.95						2.00
161	M	Saudi Arabia						71.61	0.33	1.11	0.00			7.00
162	L	Senegal			26.30		1.00	52.30	0.07	1.25	31.36	2.00	21.90	4.00	4.00	2.00
163	M	Seychelles				11.59		71.41						3.00
164	L	Sierra Leone						37.21	0.12		15.49			5.00
165	H	Singapore				3.33		77.05	0.06	1.02	0.00	2.00	13.50	5.00	1.00	2.00
166	M	Slovak Republic		18.07		20.92		72.70	0.02		0.00		52.30	4.00	2.00
167	H	Slovenia	98.20					74.72		1.00	0.00			2.00	1.00
168	L	Solomon Islands						70.50			28.90			2.00
169	L	Somalia						46.96			32.90			7.00
170	M	South Africa		17.40				65.23		1.00	0.00	2.00	42.70	2.00	2.00	2.00
171	H	Spain				21.99		77.91	0.15	1.00	0.00	1.00		2.00	1.00	1.00
172	M	Sri Lanka					9.80	73.10	0.06	1.00	2.43	2.00		5.00	4.00	2.00
173	M	St. Kitts and Nevis						70.04						2.00
174	M	St. Lucia						71.77						2.00
175	M	St. Vincent & the Grenadines						72.90						1.00
176	L	Sudan						54.97	0.12		29.38			7.00
177	M	Suriname						70.04	0.08		0.46			3.00
178	M	Swaziland						57.64	0.03	0.99	13.73		19.10	5.00
179	H	Sweden				34.66		79.09	0.01	1.00	0.00	1.00		1.00	1.00	1.00
180	H	Switzerland				25.92		79.08	0.08	1.00	0.00			1.00	1.00
181	M	Syrian Arab Republic						68.90	0.25	1.10	5.78			7.00
182	L	Tajikistan	94.00					68.38			0.00			7.00
183	L	Tanzania		61.70	19.90		34.50	47.92	0.01	0.98	39.48	4.00	17.40	5.00	3.00	4.00
184	M	Thailand		43.34	2.00	1.93	5.00	72.34	0.04	0.98	16.22	1.00	3.10	3.00	2.00	2.00
185	L	Togo				2.80		48.82	0.07	1.36	28.60			5.00
186	M	Tonga						70.45						3.00
187	M	Trinidad and Tobago			12.40	6.62	4.00	72.78	0.10	1.01	0.00			2.00
188	M	Tunisia				7.70		71.90	0.16	1.02	0.00	3.00		5.00	2.00	2.00
189	M	Turkey		15.00	2.40	6.80	18.40	69.04	0.06	1.03	11.89		21.90	5.00
190	L	Turkmenistan						65.71			0.00			7.00	5.00
191	L	Uganda	13.80	83.70			49.50	42.48	0.04		45.31	4.00	3.90	4.00	4.00	3.00
192	M	Ukraine		4.90		19.80	1.00	67.30			0.00		100.00	4.00
193	H	United Arab Emirates						75.17	0.22	1.04	0.00	4.00		5.00	2.00	3.00
194	H	United Kingdom				22.80		77.09	0.06	1.00	0.00	1.00		2.00	1.00	1.00
195	H	United States				16.49		76.13	0.04	1.00	0.00	1.00	14.90	1.00	1.00	1.00
196	M	Uruguay	76.60	30.71		22.42		73.66		0.99	2.08			2.00
197	M	Uzbekistan	82.10				2.80	69.17			0.00			6.00
198	M	Vanuatu						64.62		1.06				3.00
199	M	Venezuela, RB		39.24	14.70		8.00	72.78	0.11	0.97	0.95	1.00	14.90	3.00	4.00	1.00
200	L	Vietnam						68.17	0.00	1.02	9.12			7.00
201	H	Virgin Islands (U.S.)						76.72
202	M	West Bank and Gaza						71.31
203	L	Yemen, Rep.					55.90	55.04	0.28		20.15			6.00
204	M	Yugoslavia, FR (Serbia/Montenegro)						72.09			0.01	3.00		6.00	5.00	3.00
205	L	Zambia		80.70	72.60	2.45	24.90	43.15	0.05	1.02	16.27	3.00	12.50	4.00	3.00	3.00
206	L	Zimbabwe						52.36	0.05	1.02	29.44	2.00	13.90	5.00	4.00	3.00

Data Sources: 1. ILO Bureau of Statistics, Special Database (2000).2. Yearbook of Labour Statistics, ILO, 1998, 3.Majid. N, The Size of the Working Poor Population in Developing Countries, Employment Working Paper 2001/16, May, 2001.; 4 The 19^th International Inquiry in to the Cost of Social Security (1997) ; 5 World Labour Report 2000, ILO; 6. World Development Indicators, 1999 and 2000; 7. Human Development Reports, UNDP (various issues);8. Forsythe, Korzeniewicz and Durrant (2000).Gender Inequalities and Economic Growth. A Longitudnal Evaluation., Economic Development and Cultural Change, University of Chicago.; 9. Coppedge and Reinicke , Measuring Polyarchy, in Alex Inkles (ed) On Measuring Democracy : Its Consequences and Concomitants, Transaction Publishers, 1991; 10. Jelle Visser, Trends in Unionisation and Union Practice,(ILO mimeo) (2000);11. Freedom in the World, The Annual Survey of Political Rights and Civil Liberties, 1998-99, Freedom House,1999; 12.B.Weder , (2000), Institutional Reform in Transitional Economies: How far have they come? (unpublished manuscript), prepared for World Economic Outlook, IMF (2000).

Selected bibliography

Anand, S. and Ravallion, M. (1993): "Human Development in Poor Countries: On the Role of Private Incomes and Public Services", Journal of Economic Perspectives, Vol. 7, Nashville, Tennessee.
Anderson, E and H. White (2000): Growth versus Distribution : Does the Pattern of Growth Matter (mimeo); Institute of Development Studies, Sussex.
Anker, R. (1999): A Conceptual and Research Framework for the Economics of Child Labour and its Elimination (mimeo), ILO, Geneva.
Baland, J.M. and J.A. Robinson (2001): Is child labour inefficient? (mimeo); Network on the Effects of Inequality on Economic Performance, Institute of International Studies; University of California, Berkeley.
Basu, K. (1984): The Less Developed Economy, Blackwell Publishers, Oxford.
Basu, K. and Van, P. (1998): "The Economics of Child Labour", American Economic Review, Vol.88, Menasha, Wisconsin.
Bardhan, P. (2000): "Democracy and Development : A Complex Relationship", in I. Shapiro and C.Hacker-Cordon (eds.), Democracy's Value, Cambridge University Press.
Bardhan, P. and Samuel Bowles (1998): The Effects of Inequality on Economic Performance: A Progress Report and Proposal,1998, Network on the Effects of Inequality on Economic Performance, Institute of International Studies, University of California, Berkeley.
Bardhan, P. (2000): "Wealth Inequality, Wealth Constraints, and Economic Performance" (with Sam Bowles and Herb Gintis), in A. Atkinson and F. Bourguignon (eds.), Handbook of Income Distribution, North Holland, Amsterdam.
Bolen and Paxton (2000): "Subjective Measures of Liberal Democracy", Comparative Political Studies, Vol. 33, No. 1, February 2000, Beverly Hills.
Chong, A and C. Calderon (2000): Institutional Quality and Income Distribution, Economic Development and Cultural Change, University of Chicago Press.
Cohen, Gerry (2000): If You're an Egalitarian, How Come You're So Rich?, Harvard University Press, Cambridge, MA.
Inkles, Alex (ed.) (1991): On Measuring Democracy: Its Consequences and Concomitants, Transaction Publishers, Pascataway, NJ.
Dollar, David and Art Kraay (2000): Growth is Good for the Poor (mimeo), World Bank, Washington DC.
Drèze, J. and A. Sen, (1989): Hunger and Public Action, Clarendon Press, Oxford.
Emmerij, L., R.. Jolly and T.Weiss, 2001, Ahead of the Curve: UN ideas and Global Challenges Indiana University Press. 2001
Forsythe, N., Korzeniewicz, R. P. and Durrant, V. (2000): "Gender Inequalities and Economic Growth: A Longitudinal Evaluation" in Economic Development and Cultural Change, University of Chicago.
Freedom House (2000): Freedom in the World, The Annual Survey of Political Rights and Civil Liberties, 1998-99, Washington DC.
Ghose, A.K. (1999): Current Issues of Employment Policy in India, Economic and Political Weekly , September, 4 -10, 1999
Grootaert, C. (1998): "Child labour in Cote d'Ivoire: Incidence and Determinants", in C. Grootaert and H. Patrinos (eds.) The Policy Analysis of Child Labour: A Comparative Study, The World Bank., Washington DC.
Helliwell, J. (1994): "Empirical Linkages between Democracy and Economic Growth", British Journal of Political Science, No.24, Cambridge.
Hill, A. and King, E. (1995): Women's education in developing countries, Johns Hopkins Press, Baltimore, MD.
Hoff, K. and Stiglitz, J.(1993): "Imperfect Information and Rural Credit Markets", in K. Hoff, A. Braverman and J. Stiglitz (eds.), The Economics of Rural Organisation, World Bank, Oxford.
Hughes, B. (2001): "Global Social Transformation : The Sweet Spot, The Steady Slog and the Systemic Shift", Economic Development and Cultural Change, Vol 49. No.2, Chicago.
International Labour Organisation (Bureau of Statistics): Special Database (2000), Geneva.
International Labour Organisation: World Labour Report 2000, Geneva.
International Labour Organisation (1999): Decent Work, International Labour conference 87^th Session, 1999, Geneva.
International Labour Organisation (1999): Key Indicators of the Labour Market, 1999, Geneva.
International Labour Organisation (1998): Yearbook of Labour Statistics, 1998, Geneva.
International Labour Organisation (1997): The 19^th International Inquiry into the Cost of Social Security (1997), Geneva.

International Labour Organisation (1996): Economically Active Populations: Estimates and projections 1950-2010, Geneva.
International Monetary Fund (2000): World Economic Outlook, Washington DC.
Kanbur, R. and Lustig, N. (1999): Why is Inequality Back on the Agenda?, paper prepared for the Annual Bank Conference on Development Economics, World Bank, Washington DC, April 28-30, 1999.
Kanbur, R. (2000): "Income distribution and development", in Handbook of Income Distribution, Vol. 1, North Holland.
Klasen, S. (1994): "Missing Women Reconsidered", World Development No. 22, Oxford.
Klasen, S. (2000): Does Gender Inequality Reduce Growth and Development? Working paper #7. Development Research Group, Poverty Reduction and Economic Management Network, The World Bank, Washington DC.
Kucera, D. (2000), Decent Work and Rights at Work: New Measures of Freedom of Association and Collective Bargaining (ILO Geneva, mimeo)
Litchfield, J. and M. Lipton, (2001): Successes in Anti-Poverty: National Level Poverty Performance (ILO Geneva, mimeo).
Majid. N., (2001): The Size of the Working Poor Population in Developing Countries, Employment Paper 2001/16, International Labour Office, Geneva.
Nozick, R. (1974): Anarchy, State, and Utopia, Basic Books, New York.
Ravallion, M. (2001): Growth Inequality and Poverty: Looking beyond averages (mimeo), WIDER Development Conference, WIDER, Helsinki.
Rawls, J. (1971): A Theory of Justice, Oxford University Press.
Rodrik, D. (1998): Democracy and Economic Performance (1997), mimeo for Conference on Democratisation and Economic reform in South Africa, Cape Town, 1998.
Rodgers, Gerry (2000): Decent work as a development objective, Lecture to the 42^nd Annual Meeting of the Indian Society of Labour Economics, Madhya Pradesh, India.
Sah, R.K., (1991): "Feasibility, Human Organisations and Political Systems", Journal of Economic Perspectives, Vol. 5/2, 1991, Nashville, Tennessee.
Sen, A. (2000): "Work and Rights", International Labour Review, 2000, Vol 139, No. 2, Geneva.
Sen, A. (1999): Development as Freedom, Knopf, New York.
Sen, A. (1997): On Economic Inequality, expanded edition, Oxford, Clarendon.
Sen, A. (1995): "The Concept of Development", in Handbook of Development Economics, Vol.1, North Holland.
Sen, A.(1992): Inequality Re-examined, Oxford University Press.
Sen, A. (1989): "Women's Survival as a Development Problem", Bulletin of American Academy of Arts and Sciences.
Sen, A. (1987): Commodities and Capabilities, Oxford University Press.
Sen, A. (1981): "Rights and Agency", Philosophy and Public Affairs, Vol 11 no 1.1981, Princeton.
Shapiro, I. and C. Hacker-Cordon (eds.) (2000): Democracy's Value, Cambridge University Press.
Stiglitz, J. (2000): Democratic Development as the Fruits of Labour, Keynote address, IRRA, Boston, 2000.
Stiglitz, J. (1998): Towards a New Paradigm for Development: Strategies, Policies and Processes, Raul Prebisch Lecture, UNCTAD, Geneva, 1998.
Summers, L. (1994): Investing in All People, The World Bank, Washington DC.
United Nations Development Programme: Human Development Reports (various years), New York, NY.
Visser, J. (2000): Trends in Unionisation and Union Practice (ILO Geneva, mimeo).
Weder, B. (2000): Institutional Reform in Transitional Economies: How far have they come? (unpublished manuscript), prepared for World Economic Outlook (2000), International Monetary Fund, Washington.
World Bank, World Development Indicators, 1999 and 2000, Washington DC.

1. Decent Work, International Labour Conference 87^th Session 1999, Geneva, ILO.

2. Essentially relationships established from reduced-form regression equations with cross-country data cannot say very much either about causality between variables, or about what that relationship means at the level of particular policy instruments.

3. The question of the manner in which components of decent work are consistent with the idea of a "decent life" is a separate issue.

4. See Gerry Rodgers, (2000), "Decent work as a development objective". Lecture to the 42^nd Annual Meeting of the Indian Society of Labour Economics, Madhya Pradesh, India.

5. There is the related larger question (still unresolved) concerning the appropriate rights based framework in which one can place this new agenda. Since most international institutional agendas are being cast in terms of rights today, this is a general issue relevant to most development agencies. It is of importance because the phrase "rights based framework " is used, in this context, more in a rhetorical sense. The issue in question is one of framework and a separate one from the constitutive, instrumental and constructive roles that freedoms play in society according to Sen (A. Sen (1999), Development as Freedom, A. Knopf , New York).There are a few frameworks to chose from. The traditional approaches are of a rights constraints variety and rights consequentialist variety. In one rights are constraints on individual action (with no consequential justification) in the other rights are viewed as instrumental in achieving broads rights-independent goals. One is more legalistic in spirit, the other utilitarian. Neither approach can seriously accommodate rights-realisation in the evaluation of resulting states. A third framework (which is probably more suitable to a social agenda of a development agency) can be called a goal-rights- framework . It allows for the realisation of rights themselves to play a role in evaluation of states. According to Sen : "A moral system in which fulfilment and non-realisation of rights are included among the goals, incorporated in the state of affairs, and then applied to the choice of actions through consequential links will be called a goals rights system".(A.Sen, 1981, Rights and Agency, Philosophy and Public Affairs. Vol 11 no 1.1981).

6. Clearly "democracy" facilitates this setting of weights within society, although as discussed later the question is about the democratic process and not only about voting.

7. There is definitely a characterisation issue here. Security, work rights, and representation based dimensions of decent work are not only connected but can also be seen as 'attributes' of employment.

8. This space between society and work spheres is, in addition to distributional processes, also reflective of the extent to which non-market transactions obtain in society that affect both relative prices and well being.

9. There is some literature that shows that efficiency wages improve productivity, just as efficiency benefits can be reaped by worker involvement in decision-making. So workers' participation and better wages can in principle be demonstrated as a part of the profit motive itself. Whether this can be achieved through the market only or if there is a role for both advocacy (sharing knowledge) and a special role for government is a separate issue. The literature covers many domains but is essentially based on the idea that a workers productivity is linked to consumption (See Basu (1984) for a survey).

10. In fact in any commoditized economy all benefits have associated costs. Thus, for example, it is erroneous to suggest that some rights have no costs simply because they are legal or juridical in nature and expressed as such. Secondly in any case, if for example workers gain from certain rights, the argument for proposing those rights ought to be based on the benefit accruing to workers as a result of these rights, and not that since these rights appear to be "costless" while also "benefiting" workers, they might as well be adopted by employers.

11. Recent work suggest that social dimensions improve in a certain sequence with individual life conditions improving first and fast and the structural or what the author calls social organisational components changing more slowly and unevenly (B. Hughes (2001), Global Social Transformation: The Sweet Spot, The Steady Slog and the Systemic Shift, Economic Development and Cultural Change Vol 49. No.2 , 2001).

12. Trade-offs can be shown exist both between social dimensions and between groups. Any two social interventions that have costs can be expressed as competing claims for finite resources, and thus preference for one may be seen as a preference over the other. These posited trade-offs between interventions in social dimensions taken on their own are not very realistic, since they implicitly assume that all populations will equally benefit from any intervention. The issue here then becomes an abstract one of social choice. The other type of trade-off is more complex and also more relevant to social interventions in a typical unequal (more realistic) society. Here the benefits of social interventions by the State made in the name of society can become subject to "capture" by particular groups. Therefore in this case, the manner in which and the assumptions under which an intervention is made (however general and encompassing for all in society it may be in intent) becomes crucial. Here the trade-offs (from the point of view of the social intervention which needs to keep as true to its intent of maintaining the widest possible coverage as is possible) are between social groups who can or cannot respectively access the benefits of a social intervention or a societal right. There is also the important related issue here of whether we assume that "targeted" approaches to social interventions can work additively. If we do, then this would mean that groups initially un-targeted (or excluded) could smoothly and costlessly be brought in to the population benefiting from the intervention over time . On the other hand, it could be also be held that additivity cannot be assumed and that coalition-forming and exclusion become such endemic characteristics of benefiting and non-benefiting populations respectively , that expanding the targeted populations and the intervention's benefits over time to a wider set of people is neither a smooth nor a costless process.

13. A. Sen (1999), Development as Freedom, Knopf. New York

14. A. Sen, (2000), Work and Rights, International Labour Review, 2000, Vol 139, No. 2.

15. In fact the interesting link between Section 4.1 below and the discussion here is on the freedom to transact.

16. A. Sen (1987), Commodities and Capabilities, Oxford University Press; A. Sen (1992), Inequality Re-examined, Oxford University Press.

17. In the second part of this paper, we identify a policy position- that is arguably a representation of the commodity based view- that focuses on national income only.

18. Although the capability approach is presented normally in terms of the individual, it has important normative implications for distribution and equality in the (macro) world in which the individual may be deemed to have a greater chance of realising her capabilities. Another issue that needs to be the subject of further research to deepen the link between the capability approach and the individual in the decent work agenda concerns the type of application of the capability approach. There are two types of application of the capability approach. It can be used to examine the options a person has (given her capability set) or the choice of the actual functioning combination a person makes (given by her chosen functioning vector). See J. Foster and A. Sen, Annex in A. Sen, (1997) On Economic Inequality, expanded edition, Oxford, Clarendon.1997. It needs to be worked out which application is more appropriate for the individual focus of the decent work agenda.

19. There is a vast literature on this topic. For an overview of issues, See, Kanbur, R. and N. Lustig, 1999, "Why is Inequality Back on the Agenda, Paper prepared for the Annual Bank Conference on Development Economics, World Bank, Washington DC, April 28-30, 1999. Although it needs to be said that there are (political) reasons why in a globalizing (post-cold war) single world, with easy information access, the need for international institutions to at least "appear" to address social and economic policy issues simultaneously in developing countries also becomes politically expedient .

20. Since we are concerned with labour it may be useful follow through the implications for it in an orthodox framework, at the economy-wide and firm-level respectively. The core of orthodox macro policy for labour is regulatory, and only in so far as it ensures that market failure is avoided. Influencing the returns to labour directly (i.e. wages) is considered counter productive as it is an "equity-driven intervention" and in the framework leads to a loss of jobs at a macro level. However at a micro level despite its restrictive informational and market assumptions, the framework can yield results with different implications. For example it can recognise bargaining as a method of arriving at efficient solutions and can thus make a case for the improvement of workers conditions at the level of the firm. Therefore policy positions that allow for worker involvement in decision-making at wage setting at the enterprise level may nonetheless hold that interventions in wage formation at a macro level produces unemployment. See J. Stiglitz, Democratic Development as the Fruits of Labour, Keynote address, IRRA, Boston, 2000.

21. The first two fundamental theorems of welfare economics demonstrate that every equilibrium in a perfectly competitive economy is Pareto-efficient , and secondly under a set of conditions every Pareto-efficient allocation can be achieved by an equilibrium in a perfectly competitive economy.

22. Theoretically, it is well accepted now that with costly information (often in models with external uncertainty as well), a whole host of situations can ensue ranging from those involving externalities to market failure, and that informal institutional arrangements and implicit contracts are precisely the ways in which society tries to overcome these problems. If resource allocation is not necessarily optimal and efficient with free markets in any real word situation then the raison d'etre of separating economic and social policies becomes redundant as well. The technical literature is vast in this area, but for an accessible development application see Karla Hoff and J. Stiglitz,(1993) Imperfect Information and Rural Credit Markets, in The Economics of Rural Organisation, K.Hoff, A. Braverman and J. Stiglitz (ed) , World Bank, Oxford, 1993; and for a broader view on development policy implications of this see J. Stiglitz , Towards a New Paradigm for Development: Strategies, Policies and Processes, Raul Prebisch Lecture, UNCTAD, Geneva, 1998.

23. For an on-going research programme in this field, See Pranab Bardhan and Samuel Bowles (1998), The Effects of Inequality on Economic Performance: A Progress Report and Proposal,1998, Network on the Effects of Inequality on Economic Performance. Institute of International Studies; UC Berkeley. This is a MacArthur Foundation - funded Network on the Effects of Inequality on Economic Performance.

24. Following from footnote 21, the two theorems of welfare economics, (apart from being re-examined under more real world assumptions and generating results that show inefficient outcomes), have also been examined (mostly by Sen)
in terms of freedom. Amongst other things this has produced a distinction between two roles of the market, and brings us to the question of the freedom to transact. This may indeed be what is partly at stake in the linkage between the two directions that the re-examination of welfare theory has taken. The traditional welfare results established the "superiority" of the market mechanism in the sense that the market was shown to have instrumental value (when governed by perfectly competitive assumptions) in generating an efficient allocation of resources. While challenging these results affects the argument about the instrumentality of the market, it does not directly mean that the market mechanism insofar as it still gives constitutive freedoms to people is not potentially superior in this respect.

25. In terms of general acceptability while this is a relatively less difficult task for institutional agendas on health and education for reasons mentioned , it is more difficult for other agendas, e.g. as on employment or the environment.

26. These positions are also often compounded when one group of nations (e.g. often developed ) is seen by another (e.g often developing) to offer "advise" that can be seen in terms of such divides. The compounding factors are other related debates like the one on protectionism, trade and labour standards . For example a rights exclusivist position coming from developed nations as advice can be interpreted as diversionary and protectionist by developing countries (who may then take on a growth exclusivist view in reaction).

27. This view may not be as clear cut in a prescriptive sense in the present globalized world. The historical counter-factual, which such arguments ultimately entail, are nearly impossible to demonstrate.

28. Libertarian views of a stronger variety imply an absolute priority on grounds of liberty to certain libertarian rights, which cannot be violated even if they lead to violating substantive freedoms of others and entail morally horrific results. This kind of objection is linked to libertarian views on redistribution mechanisms themselves as these suggest that if there is a compelling moral reason for supporting the worst off in a society then it should be through private charity. (See Robert Nozick,(1974) Anarchy State and Utopia, Basic Books , New York). In contrast liberal views that allow for interventions for social redress, like Rawlsian ones tend to suggest that inequality is tolerable because it enhances motivations and incentives. In fact sometimes this position can be interpreted to suggest that inequality is "just" when and because it is neccessary to improve the position of the worst off.(J. Rawls (1971), A Theory of Justice , Oxford University press, 1971)

29. It has been suggested above that there is literature that shows that at a micro level inequality can lead to inefficiencies. At a macro level the views go either way (See R.Kanbur, 2000, Income Distribution and Development, in Handbook of Income Distribution , Vol. 1 , North Holland). It is however not clear whether removing inequalities at the micro level (or in the world of implicit contracts) always translates in to inequality reduction at the macro level, and that when it does whether it leads to greater economy-wide efficiency as well.

30. The point is that advocacy based on participation is a solution to problems of monitoring. Thus in an institutional context, the role of an advocacy organisation is about providing individuals and groups in society information to use and act upon. So the scaling-up of freedoms assumes the participation of citizens and their leading the process of change. Similarly this issue of monitoring and agency comes up seriously if a set of values is imposed, not if it is built up by consensus in society.

31. The results are based on the nation state taken as a unit of analysis, and are therefore valid at that level of aggregation. Recent literature on this topic concurs on the negative relationship between growth and poverty. (See for example E.Anderson and H. White (2000), Growth versus Distribution : Does the Pattern of Growth Matter, Institute of Development Studies, Sussex (mimeo); David Dollar and Aart Kraay, 2000, Growth is Good for the Poor , World Bank (mimeo), J.Litchfield and M. Lipton, 2001 , Successes in Anti-Poverty : National Level Poverty Performance (ILO , Employment Paper, 22, (forthcoming)). It has however also been suggested that low levels of initial inequality are also good for growth. If this is the case, then one policy challenge would be how to move from one steady state of high inequality to another steady state of lower inequality without sacrificing on growth. See M. Ravallion,2001, Growth Inequality and Poverty: Looking beyond averages, WIDER Development Conference (mimeo).

32. This is of course not to say that systemic changes alter past distributions immediately. This is a point we mention later in Section 7 in discussing country performance on decent work indicators in transition economies.

33. The joint responsibility of the IMF and World Bank in the Poverty Reduction Strategy Papers (PRSP) process, and the PRSP's consistency requirement with the IMF's lending facility (the Poverty Reduction Growth Facility (PRGF)) shows potential for this. However it remains to be seen whether there is real scope for countries proposing "alternative ways" of achieving growth in the PRGF. As suggested above, this would entail a discussion on policy instruments and their impact assessment on at least all the worst off in society.

34. Some points that are made in the empirical illustrations are more focussed on qualifying the views that are held within the institution, and which in some cases may be more obvious to those outside it.

35. The literature is vast as the discipline of economics itself. A. Sen (1988), The Concept of Development, in Handbook of Development Economics Vol.1 , North Holland (1988).

36. GNP per capita is gross national product divided by midyear population. PPP GNP is gross national product converted to international dollars using purchasing power parity rates. An international dollar has the same purchasing power over GNP as the U.S. dollar in the United States. For more information, see World Development Indicators Tables 1.1, 4.11, and 4.12.

37. The country classification is the World Bank's. The World Bank's main criterion for classifying economies is gross national product (GNP) per capita. An economy is classified as low-income, middle-income (subdivided into lower-middle and upper-middle), or high-income. Low-income and middle-income economies are referred to as developing economies. For example, for the 1990ss economies are divided among income groups according to 1998 GNP per capita, calculated using the Atlas method. The groups are: low-income, $760 or less; lower-middle-income,$761-3,030; upper-middle-income, $3,031-9,360; and high-income, $9,361 or more. See any recent World Development Report for details.

38. A social floor can be conceived of as a redistribution provision by the State to cover those whose incomes fall below what may be socially acceptable. This can be a combination of effective minimum reservation wage levels and a functioning social insurance system.

39. On top of this, because of definitional reasons, large parts of the workforce that are involved in work as home-workers are not even recognised as "employed". There has been much important work in the area of intra-household distribution in the last decade or so.

40. See ILO's Key Indicators of the Labour Market (1999) for a definitions.

41. It is important to note that the general argument that infers the "possibility of policy intervention" from the scatter plot and equation does not say anything about actual policy instruments needed for that intervention. Whenever a policy implication is discussed in this context it must be kept in mind that it is referred to as the possibility of reaching a particular policy goal and implies little about the instruments with which to reach that goal.

42. The idea of the working poor was also raised by the ILO's World Employment Programme in the 1970s as one component of the "employment problem" , though there was no comprehensive attempt to quantify it (See L. Emmerij, R. Jolly and T.Weiss, 2001, Ahead of the Curve: UN ideas and Global Challenges). The present work on the working poor at the ILO, which is still in its preliminary stages takes a more quantitative approach. See N. Majid, The Size of the Working Poor Population in Developing Countries, Employment Working Paper 2001/16, May, 2001.

43. This oversight is illustrated by the possibility of a negative relationship between the unemployment rate and the poverty rate in a cross section of developing economies where typically the unemployment rates are lower than poverty The relationship between the poverty rate (Pr) and the unemployment rate (Ur) for the 1990s for a set of developing countries for which both unemployment and poverty rates were available is:
P _r = 17.63* - 0.85*Ur (R-squared=0.13; N=25; t-values in parenthesis)
(3.94) (-1.91)
The relationship is of course driven by levels of national income, because unemployment rates are typically higher in middle income developing countries and poverty rates typically higher in low income developing countries. The relationship appears to be counter-intuitive only if the two stylised facts are ignored. The equation is however only illustrative because actual data on the unemployed typically not only include the long-run but also short run unemployed. Some of the latter can be in the poor. However the assumption made here, which is a reasonable one, is that long term unemployed are in the majority. In the country context of India the problems thrown up by the gap between poverty incidences on the one hand and unemployment on the other have been commented upon. (See A.K.Ghose (1999), Current Issues of Employment Policy in India, Economic and Political Weekly , September, 4 -10, 1999)

44. It is true that much depends on how poverty estimates turn out for China and India, and the particular period chosen. In the case of China for example while poverty rates dramatically increased during the second half of the 1980s, they have declined from 1990 onwards.

45. At a macro level there are at least two ways in which this inclusion can be affected. First increasing the productivity of the poor has importantly meant the transfer of labour from agriculture to the modern sectors in developing countries and the absorption of these persons in labour intensive but relatively more productive jobs. However productivity increases in modern sector jobs themselves have often come at the cost of a loss in employment. Therefore an expansionary demand context is necessary to accompany the transfers.

46. Apart from this it is important to note that the elimination of the working poor, however ambitious in actuality, as an employment goal still constitutes a floor.

47. See ILO World Labour Report, 2000, Table 11 for definitions.

48. Total social security expenditure covers expenditures on pensions, health care, employment injury, sickness, family, housing and social assistance benefits in cash and kind including administrative services.(See ILO World Labour Report, 2000. Table 14 Notes).

49. On going work with respect to these indicators, at the ILO, clearly shows that both low levels of national income as well as high levels of poverty are related to higher levels of illiteracy and infant mortality, but there does remain sufficient scope for policy interventions to reduce illiteracy and infant mortality. Michael Lipton and Julie Litchfield (2001): Successes in anti-poverty : National Level Poverty Performance (forthcoming). Successes in Anti-poverty: National level performance. ILO Discussion Paper.

50. J. Dreze and A. Sen, (1989), Hunger and Public Action, Oxford, Clarendon.

51. S. Anand and M. Ravallion (1993), Human Development in Poor Countries: On the Role of Private Incomes and Public Services, Journal of Economic Perspectives, Vol. 7.

52. The point is that while the level of national income and associated public revenue structures (and the allocation between expenditure heads by government) put limits on what can be afforded by an economy on security it is also true that the cost parameters of basic heath and educational activities are lower in countries at lower levels of national income. See Dreze and Sen (1989)

53. These can be because of economic slowdowns, natural disasters and health catastrophes affecting society.

54. A. Sen (1989) Women's Survival as a Development Problem, Bulletin of American Academy of Arts and Sciences; S. Klasen (1994), Missing Women Reconsidered, World Development 22.

55. This is suggestive of the process associated with income inequality where inequality is supposed to initially increase with growth in national income and then decline.

56. See S. Klasen (2000), Does Gender Inequality Reduce Growth and Development? Working paper 7. Development Research Group, Poverty Reduction and Economic Management Network, The World Bank. He estimates a 0.9 per cent per year faster growth, if South Asia and Sub Saharan Africa had a balanced educational achievement in 1960 and had active policies in place for gender balanced growth in education. The employment bias effect is less conclusive on measurement and causality grounds , but may have been responsible for another reduction of .3 per cent per annum for the above regions relative to East Asia.

57. See for example, L. Summers (1994) Investing in All People, The World Bank, on both fertility and child mortality effects; and on fertility effects A. Hill and E. King (1995), Women's Education in Developing Countries, Johns Hopkins Press..

58. The argument is an important one. The issue is the extent to which an all pervasive social bias involving gender is removable withing existing norms of society. Although the issue can be starkly illustrated in some developing country contexts, where laws can be anti-women, the feminist critique is a general and deeper one of a liberal theory of justice. The argument is that if choices not regulated by law are considered outside the purview of justice then biases will remain. For example in technical terms in a Rawlsian world, the fact that the space of the family arguably falls outside the so called "basic structure" of society where say Rawls' difference principle is applicable, means that gender biases remain and are perpetuated. (Gerry Cohen, 1996, Justice, Incentive and Selfishness, Gifford Lectures; No. 8 , University of Edinburgh, 1996 in If You're an Egalitarian, How Come You're So Rich? Harvard, 2000)

59. R. Anker (1999) A Conceptual and Research Framework for the Economics of Child Labour and its Elimination (ILO, mimeo); K. Basu and P.Van (1998), The Economics of Child Labour, American Economic Review, Vol.88.

60. In fact the weighted averages are 18.1 per cent for low income countries and 8.7 per cent for middle income countries with standard deviations of 15.1 and 7.0 respectively.

61. ILO (1996) Economically Active Populations : Estimates and projections 1950-2010. Geneva.

62. Theoretical results on models of child labour should be examined carefully with respect to the applicability of the assumptions under which they are derived. For example the use of child labour as a mechanism of transfer of resources across generations (from children to parents) or overcoming capital market imperfections and transferring resources across time (from future to present) have been shown; as well as the conditions under which banning child labour may be efficient (Pareto improving) in general equilibrium. See J.M. Baland and J.A. Robinson, 2000, Is Child Labor Ineffcient? Mimeo Network on the Effects of Inequality on Economic Performance. Institute of International Studies; UC Berkeley

63. Modelling the empirical determinants of each dimension of decent work is important. Because of direct policy implications, in the case of child labour, when determining the causes of its incidence, we need to be careful in drawing conclusions regarding the relative importance of its "causes" and the extent to which they are co-determined. For example, if educational/school attendance related variables and poverty are correlated themselves, as they often are in developing countries, simply ranking the importance of one explanatory variable over the other to determine policy priority would be misleading.

64. See for example C.Grootaert (1998) Child labour in Cote d'Ivoire: Incidence and Determinants in C. Grootaert and H. Patrinos eds. The Policy Analysis of Child Labour: A Comparative Study, The World Bank.

65. See Bolen and Paxton (2000), Subjective Measures of Liberal Democracy, Comparative Political Studies Vol. 33 No. 1 Feburary, 2000, on judgement and specification errors. The ILO is presently involved in constructing its own index see D. Kucera (2000), Decent Work and Rights at Work: New Measures of Freedom of Association and Collective Bargaining, (ILO mimeo).

66. Freedom House (1999), Freedom in the World, The Annual Survey of Political Rights and Civil Liberties, 1998-99, Freedom House.

67. It may be important to formulate these ideas in such a way that they do not get tied to the idea of conditionality or sanctions.

68. See for example A.K.Sen and Jean Dreze, Hunger and Public Action, part 3. Clarendon, Oxford, 1989.Dani Rodrik, Democracy and Economic Performance (1997)(mimeo) for Conference on democratisation and economic reform in South Africa, Cape Town, 1998; John Helliwell, Empirical Linkages between Democracy and Economic Growth, British journal of Political Science, No.24, 1994.

69. See R.K. Sah, Feasibility, Human Organisations and Political Systems (1991), Journal of Economic Perspectives. Vol. 5/2, 1991.

70. D. Rodrik (1997).

71. The debate however is much more interesting and meaningful if we go beyond simple definitions and deconstruct the notion of democracy in terms of its attributes, for example, the credibility of pre-commitments by the State, the degree of insulation of the civil bureaucracy, or the legislative process and the enactment of laws in democracy (which can be subject to capture). A good example of conflict of democracy with long run growth is the classic protected infant industry argument, where reliance on State subsidy by permanent infants can become difficult to discontinue precisely because of democratic processes. See P. Bardhan, 2000, Democracy and Development : A Complex Relationship , in I. Shapiro and C.Hacker-Cordon (eds.), Democracy's Value , Cambridge University Press

72. The measure of institutional quality for 1997 developed by B.Weder - (B. Weder, (2000), Institutional Reform in Transitional Economies: How far have they come? , prepared for World Economic Outlook, IMF (2000)) - is based on indicators of governance developed by Kaufman, Kraay and Ziodo-Lobaton for the World Bank (World Bank Policy Research Working Paper No. 2196, 1999). These cover indices of democracy, government effectiveness, extent of regulation, rule of law and graft.

73. There is also new research on the relationship between institutional quality (in the case of this data set the higher the score the better the institutional quality) and income inequality which suggests that for poor countries institutional quality is positively linked to income inequality and for rich countries it is negatively related to it. A quadratic relationship is found. This result is an interesting one since it suggests that initially institutional reform is likely to benefit the better off sections of society. A.Chong and C.Calderon, 2000, Institutional Quality and Income Distribution, Economic Development and Cultural Change, University of Chicago Press, July , 2000.

74. B. Hughes (2001) , Global Social Transformation : The Sweet Spot, The Steady Slog and the Systemic Shift, Economic Development and Cultural Change Vol 49. No.2 , 2001.

75. Child labour is of course also an indicator of employment.

76. The positive relationship between gender bias against women in education and child labour incidence needs to be explained and is a subject of further inquiry at a micro level. The gender net primary enrollment indicator which is a ratio of male to females (i.e the higher this indicator the worse the gender situation in primary education) seems to vary directly with child labour. It may have something to do with different reasons for which young boys and girls are put to work. There are both poverty and returns to education issues relevant here.

77. In middle income countries the gender enrollment indicator shows a (perverse) negative relation with the working poor rate. Although this needs further research, this may be a result of the relationship observed in Figure 10, which suggest that gender inequality increases when transiting from low to medium income countries. However there may be a degree of exclusivity to those who study and those who work.

78. What is striking is that in the list of net good scores, many countries are recognisably ones with "socialist" periods in their economic histories. This is an interesting area to explore in future research.

79. For example, net terms would mean if a country is a good performer on three indicators and bad performer on one indicator, it would be deemed to be a net good performer with a degree of simultaneity of two.

80. I am particularly grateful for Farhad Mehran discussing this part of the test with me.

Back to Employment Strategy Publications List