Select Committee on Social Security Minutes of Evidence


Memorandum submitted by Professor John Veit-Wilson[1] (ICC 15)

TRIANGULATION OF DATA FROM EXISTING SOURCES AS A BASIS FOR A GOVERNMENTAL MINIMUM INCOME STANDARD

SUMMARY

  1.  The construction of a politically credible and practicable governmental minimum income standard for the UK is feasible. It has two main stages. One is the collection of a wide range of scientific data from existing research and official sources about the deprivations and social exclusions with which the present government is actively concerned. Many, though not all, of these social problems are caused by low income. Statistical analysis of a wide range of sources of information about them is need to see which of them correlate with local incomes, and if so at what levels. Triangulation is the process of comparing the various low income levels which may be revealed to see if they differ, overlap or correspond and if evidence-based conclusions can reliably be drawn about poverty thresholds and minimally adequate incomes for social participation. This paper outlines this process and distinguishes it from the process of making political judgements.

  2.  The second main stage is not developed in this paper but has to be mentioned as it is an inseparable part of setting governmental minimum income standards. It it therefore outlined in appendices to this paper. Political judgements must be taken on which tier of the income maintenance system is to be used as the symbolic location of the standards—for instance, is it to be the minimum wage, the basic pension level or the level of social assistance? The political credibility of the standard is also crucial. It will derive from being based on the best evidence over a wide range of social conditions of the income levels at which the UK population can, on average, achieve the social participation and avoid those deprivations and social exclusions based on low income which rightly concern the government.

INTRODUCTION

  3.  This paper explores the process of "triangulating" or comparing a range of sources of relevant information about inadequate incomes, to help a UK government find a politically-credible basis for a governmental minimum income standard (referred to here as MIS) for use as a yardstick to assess the adequacy of various income maintenance provisions. There are many sources of information about what an inadequate income is, and a number of methods are available by which to calculate what an adequate income might be, at various different standards of adequacy. But no single source or research method may carry enough weight to be politically credible as a criterion or guide, as the DSS noted in a recent response to the Select Committee.

  4.  The government wishes to abolish child poverty, and it is concerned about a range of indicators of deprivations and social exclusions. As far as the range of indicators is concerned, some appear to reflect deprivations and exclusions caused by low income, while others are not necessarily based upon low purchasing power, inadequate personal disposable incomes, as such. But while the government admits that "low income is an important aspect of poverty",[2] at present it has no idea of what income level would be needed to abolish poverty in households containing children. Its measure of poverty for this purpose, the statistics of Households Below Average Income (HBAI) is acknowledged to be no more than a measure of unequal income and not of adequate income, and no one knows if half the mean or 60 per cent of the median is too high, too low or about right to achieve the intended objective of preventing deprivations and social exclusions. Similiarly, there is no systematic information about the whole range of the deprivations and exclusions so that the government can distinguish between those which are caused by low income and those which are not and plan policy accordingly. Nor is there information about the levels of low income at which the problems of deprivation and exclusion become intolerably severe.

  5.  In short, what we need is evidence to allow a clear distinction between household incomes which are on average too low to meet the government's objectives and those which are just high enough to do so, and we need this to be based on a range of reliable sources and a variety of approved methods so that it is scientifically robust and politically credible. Much of the necessary evidence alread exists and the methods are there. Now they need to be brought together in a way which satisfies the government's own legitimate needs for credibility, for evidence which is robust and not vulnerable to methodological and political attack, findings which are based upon methods and sources which are widely respected by experts and the public alike.

  6.  I have reported on the ways in which the governments of 10 other countries, seven in Europe, approached the question of discovering and using such politically-credible governmental minimum income standards.[3] What they found to be politically credible in their countries varied a good deal. Much depended on how far political consensus could be achieved over the establishment of such a standard. The chief guidance which that research has for the situation in the UK is to emphasise the importance of finding what is a respected basis for standards in this country.

  7.  This paper outlines the triangulation method and how it can be used to broaden the basis of reliable information on which a governmental minimum income standard can be based. However, in order to avoid misunderstandings, four preliminaries must be briefly mentioned before discussing triangulation. The first three are explained in more detail in the appendices, while I have written at length on the meanings of income adequacy in a chapter available to the Committee.[4]

FOUR PRELIMINARIES

I.  The distinction between the findings of scientific research, the prescriptions of government and actual benefit levels

  8.  First, the types of information about the income levels needed by households of varying size and composition to avoid poverty and achieve minimal adequacy or decency standards must be distinguished, between (a) those which are based on the findings of scientific research and statistical analysis, and (b) those which are based on political judgements, the governmental minimum income standards (MIS for short). Both of these must also be distinguished from (c) the actual levels of benefits, such as National Insurance pensions or Income Support, which are based on political expediency and not on independent evidence of income needs. These distinctions are set out in Appendix 1.

II.  Clarification of the four tiers of the income maintenance system

  9.  Second, the idea of having a standard of yardstick against which to measure the adequacy of household incomes itself raises the question of what level of living is to be considered adequate for comparative purposes. Should the standard against which we measure adequacy be based on the actual particpatory level of living achieved by full-time workers and their households? This is a kind of traditional norm for comparisions. Alternatively, when we talk about the standard of minimal adequacy for participation, do we mean to refer to the levels of living which ought to be achieved by those on the very lowest and often short-term incomes, those of households dependent on Income Support and means-tested Jobseekers' Allowance? Another possibility is to relate our adequacy standard to what would be needed for decency over the long-term by people not in work, for instance NI pensioners' households, and then to compare above and below this standard. In practice, the income maintenance system embraces at least four distinct tiers, set out in Appendix 2, and each of them is used in thinking about governmental minimum income standards in one country or another. We must therefore consider which is most appropriate for politically-credible use in the UK.

III.  Political credibility is indispensable

  10.  However methodologically elegant it is, a standard which lacks political credibility is worthless for government policy making. The idea of political credibility is discussed in Appendix 3. Both of the issues above deeply affect the question of what is deemed to be political credibility in the UK. That is why we cannot leap into technical discussion of method without taking account of these aspects of the policy context.

IV.  What is meant by the adequacy of personal or household incomes?

  11.  I am assuming the very complicated question of what income adequacy means[5] has already been answered in the current context by saying, adequacy of household incomes is the lowest level at which people on average have enough personal disposable incomes to avoid the deprivations and social exclusions which the population experience and define as intolerable and which concern the government, at any rate those which are caused by low incomes. There may well be other approaches to the question but they do not have this political salience and are not addressed here.

TRIANGULATION

  12.  Triangulation is a metaphor from surveying, meaning the use of multiple reference points to offer greater accuracy in locating an object's exact position. In social science, more than one method is often used in validation processes to ensure that any variance which is found reflects the trait to be observed (such as the inadequate income) and not the method being used. Ensuring that the data sources are comparing like with like may be a complex process. However, the credibility of the findings is enhanced when two or more distinct methods yield comparable findings.[6] Similarly, the political credibility of governmental minimum income standards in the UK requires the triangulation of the findings of scientific research data in as many of the fields above as possible.

  13.  The aims of triangulation in the poverty research field are:

    —  to build up a more complex, detailed and complete picture of the part which personal disposable incomes play in achieving the adequate level of living and participation of individuals and households and combating deprivations and social exclusions, over time;

    —  to give scientifically reliable grounds for assessing whether over this range one can find any broadly similar poverty boundaries—thresholds or bands of low income—which may then influence (not determine) governmental minimum income standard setting.

SOURCES OF INFORMATION ABOUT POVERTY, DEPRIVATIONS AND SOCIAL EXCLUSIONS FOR USE IN TRIANGULATION

  14.  Information about poverty, deprivations and social exclusions comes from a very wide range of sources. What is relevant here is those sources which tell us the income levels at which people on average experience them and the income levels at which they are avoided. These fall into broadly two groups. (a) One is the findings of scientific research studies which aim to answer the question directly, what income level (allowing for different sizes and compositions of households, net of taxes and deductions and either including or excluding housing costs) avoids these social problems and allows the population to live decently as society defines it? (b) The other is the statistical analysis of data collected for other purposes which contains within it information which may help to answer the question. Examples of recent UK data sources rich in potential evidence for analytical triangulation include the following sources. This list includes some of the relevant publications and databases and must not be taken as exhaustive.

Direct research studies:

    —  "Deprivation indicator" studies of what the population defines as the necessities which no-one should be without, in order to discover the income levels at which people on average are not deprived in these ways and the levels at which they suffer multiple deprivations. Recent published examples are Gordon et al 2000; Bradshaw et al 1998.

    —  Budget studies: highly detailed costings of the total conventional components of participatory levels of living according the the range of standards which the population itself defines, such as "Modest but Adequate" or "Low Cost but Acceptable" and which avoid the deprivations. Examples are found in the recent work of the Family Budget Unit (Parker et al 1998 and 2000).

    —  Focus group research, studying what small groups of typical people define as the indispensable necessities and decent levels of living and their associated costs; for instance Middleton et al 1994, 1998.

    —  Attitudinal research, studying what the population itself considers minimally adequate incomes for their household, sufficient to avoid what the United Nations calls "absolute" or "overall" poverty, or enough "to make ends meet". A UK example is Townsend et al 1997.

Indirect statistical secondary analysis

  15.  In addition to these direct social science research studies, there is a large collection of official and academic statistical data sources which may contain usable information on the income levels at which individuals and households suffer or avoid poverty, deprivations and exclusions. These surveys aim to collect information covering a wide range of patterns of living, of experiences of health, education and housing, as well as more directly of deprivations and social exclusions with which the government is concerned. Many (though not all) of the surveys also collect information in some form on the economic status of the individuals or the households in which they live. All of these data sources require analysis to discover if they reveal relationships between deprivations and low incomes—and if they do, what levels of income correlate with what levels of deprivation. Some examples of publications are mentioned but there are too many to include all of them there.[7]

    —  Household surveys such as Family Expenditure Surveys, General Household Surveys, Family Resources Surveys, British Household Panel Surveys.

    —  Nutritional studies and the National Food Survey.

    —  Health inequalities studies such as Shaw et al 1999 and the Acheson Report data; premature mortality at all ages; morbidities associated with deprivations and exclusions.

    —  Educational data on achievements, deprivations and social exclusions; National Child Development Surveys and other cohort studies.

    —  Data on variety of other deprivations and exclusions, naturally including those which are the object of both Government and Joseph Rowntree Foundation/New Policy Institute monitoring (eg Howarth et al 1998, 1999; Rahman et al 2000).

  16.  The procedure of collecting evidence for a politically-credible governmental minimum income standard is therefore complex and takes time. No single source at one time is enough.

PREVIOUS APPROACHES TO POVERTY TRIANGULATION

  17.  The idea of triangulation is not new. Even UK government officials have used it in the past for monitoring income maintenance policy. In 1964-65, officials of the National Assistance Board in a committee chaired by Robert Windsor used three distinct methods to examine the adequacey of the adult benefit levels. Like the DSS, they realised that no one method was sufficient to give a robust answer to the question of what was adequate, but they asserted that "all three taken together would provide a body of evidence on which a judgement could be made".[8]


  18.  The three methods the NAB officials used were (a) budget studies taking conventional consumption patterns and expert advice on diets into account; (b) consideration of the findings of the Family Expenditure Survey on what conventional consumption patterns were; and (c) special studies of how National Assistance claimants were or were not managing. Strictly speaking, this was not triangulation as it is now understood, since each of these perspectives focused on a different aspect of the question rather than being a distinctive methodological focus on one question. But what is important here is that it was an official recognition that different perspectives must be brought to bear, and that is what this paper is also arguing. This was the one and only occasion on which government officials used a systematic approach to examine the question of the adequacy of benefits.[9] It was long forgotten and needs replication in a methodologically updated manner.

  19.  The statistical analysis of social deprivation indicators and household or individual incomes recommended here to investigate causal correlations has not yet been used on any large scale for the purpose of triangulation. Yet as suggested above there is much statistical material which could be analysed to see if it can be done fruitfully. For instance, as long ago as 1954, Professor Peter Townsend reviewed the published evidence that, on average, households with children were failing to achieve recommended intakes of some nutrients. What was not published (but had apparently been calculated) were the lowest income levels at which households actually achieved the officially recommended nutritional intakes. Townsend commented that the National Food Survey could use this as a proxy poverty measure.[10] This is precisely the kind of information still needed, extended as appropriate to the many other deprivations and exclusions which are found to correlate with low income, and the findings then compared.

  20.  Dr David Gordon's research team has recently outlined the use of a range of the direct and indirect approaches to defining and measuring deprivations and associated poverty boundaries in their first report, but we must await the forthcoming book to see whether—and if so how far—they have triangulated their findings to arrive at robust conclusions.11[11] Professor Jonathan Bradshaw of the University of York and his colleagues have already carried out some analyses of income correlates of social problems affecting children, reported in their forthcoming study of Poverty: The Outcomes for Children. This research team has found that some of the social problems in question correlate directly and credibly with low income while others do not.[12] What we now need to know is which ones, and at what levels of income does the deprivation become more severe, and then how the inadequate income levels compare with each other, which is the essence of triangulation.




  21.  Such procedures can continue to be carried out periodically and on a programmed basis and thus give the governmental minimum income standards a dynamic base as better data is collected over time. The "political construct" should be periodically rebased, just as the Retail Price Index is; that is, the whole of what makes up the level of living which it reflects needs to be reviewed in the light of changing evidence about its components, and not simply updated according to inflation[13] This is what gives dynamism and thus continuing contemporary relevance to the MIS.

WARNINGS

  22.  But three important qualifications must be set out. They are not objections to setting governmental minimum income standards but they qualify the interpretation of the search for adequacy standards. First, the findings may not be clear-cut but instead may seem confusing and therefore need interpretation and judgement. Second, comparison may show that findings about minimum incomes do not coincide exactly (as Baroness Hollis has already remarked[14]) or that the bands of incomes at which deprivations are generally avoided seem at best to overlap if they coincide at all. Third, research has already shown there is a time-lag between changing people's material conditions and behavioural changes, and we do not know how long this lag is in some of the instances of deprived or excluded behaviours dependent on low incomes. A failure to change the situation in the short run even when income rises does not disprove the finding that in the long run low income is the cause of the problem. To take each of these in turn, there is no reason in principle to expect the following outcomes.

  23.  (1) There may be no distinct "poverty thresholds" at a clear cash cut-off point in any or every instance of such secondary analysis of statistical sources. Whether or not a threshold is found, or a band of income within which the relationship with deprivation is equivocal, are matters for research and statistical analysis and interpretation. Where any continuum of correlation between the problematic phenomenon (eg morbidity or educational attainment) and household incomes is found (as Professor Bradshaw has done), it is a matter for informed discussion, both by samples of the public (for instance by focus groups) and by experts, on where an "intolerability" boundary is to be drawn (intolerable to whom?—the poor, the better-off public and the experts may have different views from each other and from Government officials or Ministers). Guidance may be found in examining the clustering of deprivations or other adverse conditions at certain low income levels.

  24.  Some research has found that there is more than one threshold, a change of gradient on the curve of the relationship between income level and consumption or experience. This finding has been called the "S" curve. To take a relevant example, research by Dr Otto Blume on the social participation of a sample of elderly people in Stuttgart, Germany, in the 1960s,[15] found that below a low threshold of income old people could not afford to participate in social activities at all. Above this minimum income threshold there was a direct relationship between their increasing income and expenditure on social participation, but again, above a level of their apparently optimal participation no further addition of income led to greater participation. It is thus clearly a matter for judgement whether it is such minimum or optimum standards which the governmental minimum income standard should reflect.



  25.  (2) Professor Bradshaw's current research (personal communication) suggests that the critical bands of inadequate income vary from one deprivation or exclusion to another. Some of them overlap. While bands of income and statistical probabilities do not offer exact points as in the surveying metaphor, more analytical research may reveal how far these findings reflect varying positions on "S" curves which can then be compared with each other. In any case, one would not expect all deprivations to occur at the same low income level, since people make their own choices of which customary consumption or behavioural patterns to pursue at each stage as their income rises, and that is what such research is illustrating.

  26.  (3) One would not expect that all forms of cash-poverty induced deprivations and exclusions would soon be eliminated by demonstrably adequate household incomes in the short-run, even if having adequate household incomes in the first place is an indispensable pre-requisite for their eventual elimination, at any rate as far as the low income cause is concerned. There are lags in people's behaviour as their conditions change, and in addition, environmental and cultural depriving and excluding conditions also take time to change.16[16]

  27.  But these are not part of the governmental minimum income standards issue as much. The essential point here is that until low income can demonstrably be eliminated as the cause of the social evils, the cause should not be ascribed to voluntary behaviour patterns, nor the victims blamed for their deprivations and exclusions. Government rhetoric about good intentions cannot excuse inadequate incomes, but income adequacy must be demonstrated by scientifically credible means before it can become politically credible.

CONCLUSION

  28.  I agree with government ministers who have told us that they have examined "findings from a wide range of research on adequacy" but "those findings disagree",[17] and that no single research method is sufficient for assessing the adequacy of income maintenance benefits. The time has therefore come for triangulation of the existing data and the analysis and interpretation of the comparative material to enable the construction of a governmental minimum income standard for the UK. It will achieve the characteristics of political credibility only if it is constructed in an open and transparent manner.

  29.  The Secretary of State for Social Security had stated that "we need to work in partnership with those with expert knowledge of particular problems", and the government is "committed to tackling poverty and its causes".[18] Since adequate income is both a manifestation and a cause of poverty, depending on perspective, the question of discovering what an adequate income is can no longer be evaded. He should therefore set up machinery for the purpose of constructing and maintaining the governmental minimum income standard, in a manner and with a credibility similar to that of the Retail Price Index. Its first task must be to collect and analyse the mass of relevant material and triangulate it.






1   John Veit-Wilson is Emeritus Professor of Social Policy of the University of Northumbria and has been Visiting Professor in the Department of Sociology and Social Policy of the University of Newcastle upon Tyne since 1992. He is author of several papers on the Concepts of Poverty and Adequacy and Their Past and Present Uses in Income Maintenance Policy, and of Setting Adequacy Standards: How Governments Define Minimum Incomes (The Policy Press, 1998) which was researched following suggestions made at a joint meeting of the House of Commons Social Security Committee and the Social Security Advisory Committee in 1991. Back

2   Opportunity for All: Tackling poverty and social exclusion, 1999, p 23. Back

3   Veit-Wilson 1998: Setting adequacy standards: how governments define minimum incomes. Back

4   "Poverty and the adequacy of social security". Back

5   The meaning of adequacy or any other similar terms such as needs, enough or decency, can only be explained in terms of the answers to four questions-adequacy or what? Adequacy for how long? Adequacy for whom? And who says? These questions are discussed at length in Veit-Wilson 1999c. Back

6   Jick 1983 p 136 Back

7   Two recent volumes of material are relevant: Bradshaw and Sainsbury (eds) 2000a and 2000b. Back

8   The Windsor Report, vol 1, para 17, NAB 1965, quoted in Veit-Wilson 1999a p 134. Back

9   According to a former Permanent Secretary of DHSS-see Veit-Wilson 1999a p 118. The NAB study found that benefit rates were inadequate and recommended that they should be raised and regularly monitored for changing standards of adequacy. The only visible policy outcome of this secret study was the long-term addition to the Supplementary Benefit rates from 1966 until they were abolished in the 1980s. Back

10   Townsend 1954 pp 134-5. Back

11   Gordon et al 2000. Back

12   Bradshaw 2001. Back

13   Rebasing means reconstituting the components of the base to reflect the current and not the past situation. Updating means costing the given components to reflect current prices. Back

14   Lords, 11.10.99 col 165. Back

15   Blume 1979 reports on this research in general but not this finding in particular, which was in an unpublished research report to the City of Stuttgart. Back

16   Gordon et al 2000 appendix 2. Back

17   Baroness Hollis op cit. Back

18   Secretary of the State for Social Socurity (1999), Opportunity for All, pp viii and 1. Back


 
previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2001
Prepared 27 February 2001