back

Notes

Choices of low-income threshold

No single indicator or group of indicators can possibly capture the full complexity of income poverty in the UK.  But any quantitative presentation of trends in income poverty necessarily has to restrict itself to a limited number of indicators.  The challenge is to choose these indicators such that they best illustrate the essence of what has been happening.  The material below discusses some of the principles that we have used in making these choices.

Richard Dawkins in The ancestor's tale (a book on Darwinism) argues against the use of income poverty indicators on the grounds that they are drawing artificial lines in a continuum.  Whilst it is true that the income of the population forms a continuum, and that increasing someone's income by a small amount to move them from just below to just above any defined income poverty line in no way moves them from being "in poverty" to being "not in poverty", his argument misunderstands the main purpose of income poverty thresholds.  If their main purpose were to say that X people are in poverty, then his argument would have some validity.  But that is not their main purpose; rather, their main purpose is to say either that the numbers of people in poverty are rising/falling or that one group is more likely to be in poverty than another.  Such conclusions are actually reasonably robust to the precise indicators chosen because, although different indicators can give very different answers in terms of absolute numbers, they typically follow similar trends over time and show similar proportional differences between groups.

Individual income versus household income

Take a family of four, where the husband earns 30,000 a year, the wife earns nothing and the two children (obviously) earn nothing.  Clearly, the husband is not in income poverty.  What about the children given that their personal earnings are zero?  Pretty obviously, they are not in income poverty (unless one wants to argue that all children are, by definition, in income poverty).  What about the wife given that her personal earnings are also zero?  Again, pretty obviously she is not in income poverty as she shares income with her husband.  The conclusion is that what matters is the total income of the household rather than the income of particular individuals.

Principle: the main thresholds of income poverty in the UK should be defined in terms of household income rather than individual income.

top

'Relative thresholds' versus 'fixed thresholds'

Webster's dictionary defines poverty as "the state of one who lacks a usual or socially acceptable amount of money or material possessions".

Professor Peter Townsend, the leading authority of the last fifty years on UK poverty, defines poverty as when someone's "resources are so seriously below those commanded by the average individual or family that they are, in effect, excluded from ordinary living patterns, customs and activities."

The key words in these two definitions are "socially acceptable" (Webster's) and "ordinary" (Townsend).  The first implies that income poverty thresholds depend on the social norms of the society in which someone lives and the second implies that they depend on the incomes of the ordinary people in society.

So, what is considered 'socially acceptable' or 'ordinary' depends on the society in which the person lives.  For example, a widely accepted indicator of third world poverty is the numbers of people living on less than $1 per day, on the grounds that people on such incomes are literally in danger of starving to death and that no one should be in this situation.  This threshold is often termed 'absolute poverty'. Note that the term 'absolute poverty' is sometimes used in the media when referring to much higher fixed thresholds, but this is clearly erroneous as those thresholds can be fixed at any level and in no sense are measuring poverty directly.  But the use of such a threshold in the UK would obviously be completely inappropriate - no one in the UK lives on incomes anywhere near this low and its use in a UK setting would imply that all people with incomes above $1 per day did not suffer from serious deprivation.

What is considered socially acceptable' or 'ordinary' also changes over time.  Levels of income that would have been considered adequate in the UK 100 years' ago would certainly not be considered to be adequate nowadays.  Rather, as society becomes richer, so norms change and the levels of income and resources that are considered to be adequate rises.  Unless the poorest can keep up with growth in average incomes, they will progressively become more excluded from the opportunities that the rest of society enjoys.

The conclusion is that the main indicators of low income - and thus of income poverty - in the UK should be defined in terms of thresholds which rise or fall as average incomes rise or fall.  Such thresholds are often termed 'relative thresholds' or indicators of 'relative poverty'.

Principle: the main thresholds of income poverty in the UK should be defined in terms of thresholds which rise or fall as average UK incomes rise or fall.

This conclusion is generally accepted by most researchers, by the EU, by the UK government and by politicians of all hues across the UK political spectrum.

However, sole reliance on relative thresholds can become misleading if average incomes rise dramatically.  For example, incomes in Ireland have risen sharply over the last ten years or so – including incomes at the bottom end - whilst income inequalities have remained roughly constant.  Many researchers and politicians in Ireland believe that sole reliance on relative thresholds gives a misleading impression by suggesting that no progress has been made in reducing the extent of poverty.

Principle: the use of fixed thresholds combined with relative thresholds can help to provide a fuller picture of what is happening to the extent of income poverty.

Note that the numbers below any fixed threshold nearly always move downward over time.  For example, this happened during the 1980s even though it is widely agreed that the numbers in poverty grew substantially in this period.  That makes fixed thresholds very difficult to interpret (unless the numbers below them are going up, which is clearly very bad) as they require interpretation in terms of the rate of movement rather than its direction.

top

What thresholds?

The threshold of low income that was, until recently, most commonly used in the UK, by both government and independent experts, was 'half average (mean) income'.  The rationale for this was twofold: first, it represented a level of income which is of the same order of magnitude as independent experts' estimates of 'low, but acceptable' levels of income; and, second, it was arithmetically simple and relatively easy to understand.  Whilst these factors are not sufficient to qualify it as a measure of poverty or as a poverty line, they did suggest that movements in the numbers below half average income would usually provide a good indication of the way that real poverty levels are moving.

However, the most commonly used threshold of income poverty is now the numbers below 60% of median income rather than 50% of mean income.  One characteristic of the median measure, in comparison to the mean, is that it is less sensitive to changes in the incomes for groups of the population.  For example, if everybody below half mean income were given enough money to bring them up to half mean then, assuming all else equal, the mean itself would rise.  By contrast, if everybody below half of the median were given enough to bring them to that threshold, the median would still remain the same.  This gives the median a practical advantage in terms of setting targets and goals for the numbers below a certain threshold.  Furthermore, unlike the mean, the median is unaffected by changes in the incomes of the very rich and, in our view, this makes it a better indicator of what is considered normal in contemporary society.

Principle: 60% of median income is widely accepted at the primary threshold of income poverty that should be used.

This conclusion, like the other conclusions above, is generally accepted by most researchers, by the EU, by the UK government and by politicians of all hues across the political spectrum.  So, for example, when the UK government is talking about its targets to half or eradicate child poverty, the main threshold of income poverty that they are using to monitor success is 60% of median income.

Finally, exclusive use of any single threshold encourages a concentration of effort on those just below the threshold to the exclusion of those who are the very poorest, and thus there is a continuing need to use a variety of thresholds.  Rather, looking at a variety of thresholds potentially provides a fuller picture of what is happening and some of our indicators therefore use both 40% and 50% of median as well as 60%.

top

Duration of low income

Some commentators argue that an exclusive focus on income at a single point in time does not provide a full picture.  They point out that some people ('those at risk of poverty') can temporarily have low incomes but not be suffering serious deprivation, whilst others ('those emerging from poverty') can temporarily have higher incomes but still be suffering material deprivation. For example, see Gordon D. et al, Poverty and Social Exclusion In Britain, Joseph Rowntree Foundation, 2000.  Rather, they argue, an important aspect of poverty is the lack of essential goods and services which arises from prolonged periods on low income.

We are sympathetic with such arguments.  In response, some of our indicators look at incomes over time and, in particular, at the numbers of people who are persistently on low incomes.  Furthermore, some of our non-income indicators look at the numbers of people who lack certain essential goods and services, such as food, transport or bank accounts.

Principle: monitoring persistency of low income, as well as low income at a point in time, can provide a fuller picture of what is happening to the prevalence of income poverty.

Note that we have not attempted to construct more complex indicators which combine lack of income and lack of essential goods and services, using surveys and statistical techniques to define what is considered 'essential'.  Such indicators could not easily be constructed and updated on an annual basis from government surveys.  There are also certain methodological issues relating to the understandability of such indicators, how 'essential' is defined and the risk that the result is erroneously claimed to be a 'scientifically correct' measure of the numbers of people in the UK who are 'in poverty.

top

'Before' or 'after' deducting housing costs?

In common with most other commentators, most of our analysis is presented on an 'after deducting housing costs' basis, which is disposable income after housing costs have been deducted.  The reasons for this are twofold:

top

'Equivalisation'

Clearly, a lone adult does not require the same income as a family of four in order to have the same standard of living.  However, importantly but less obviously, economies of scale mean that the family of four does not require four times the level of income: many costs can be shared.  This means that achieving comparability is not simply a case of dividing household income by the number of people in the household.

'Equivalisation' is the process whereby disposable incomes are adjusted to reflect household composition and size and thus put them on a like-for-like basis.  It uses an internationally agreed set of scales called 'OECD equivalisation scales'.  For example, after deducting housing costs, the disposable income of a single adult is increased by dividing it by 0.58 to put it on the same basis as a couple with no children, whereas the disposable income of a family of two adults plus two children under the age of 14 is decreased by dividing it by 1.4 to put it on the same basis.  The table below is taken from the Department of Work and Pension's Households below average income publications, where further details of equivalisation are also discussed.

Household member Equivalisation factors
Before deducting housing costsAfter deducting housing costs
First adult0.67 0.58
Every subsequent adult0.33 0.42
Each child aged 14 years or over0.33 0.42
Each child aged under 140.20 0.20

Properly understanding the equivalisation process is vital if you wish to calculate actual incomes from equivalised incomes, or vice versa.  Some examples might help this.  Assume that a household has an actual disposable income of 120 per week before deducting housing costs and 100 per week after deducting housing costs.  The table below then sets out what this income is in equivalised terms for selected household types.  Note that the actual and equivalised amounts are always the same for a couple with no dependent children.

Household composition Equivalised household income per week
Before deducting housing costsAfter deducting housing costs
Couple with no dependent children120
(120/(0.67+0.3))
100
(100/(0.58+0.42))
Single adult with no dependent children 179
(120/0.67)
172
(100/0.58)
Couple with two dependent children aged under 1486
(120/(0.67+0.33+0.2+0.2))
71
(100/(0.58+0.42+0.2+0.2))
Single adult with two dependent children aged under 14112
(120/(0.67+0.2+0.2))
102
(100/(0.58+0.2+0.2))
Three adults, no children90
(120/(0.67+0.33+0.33))
70
(100/(0.58+0.42+0.42))

The table below does the reverse process, going from equivalised to actual incomes.  Assume that a household has an equivalised disposable income of 120 per week before deducting housing costs and 100 per week after deducting housing costs.

Household composition Actual household income per week
Before deducting housing costsAfter deducting housing costs
Couple with no dependent children120
(120*(0.67+0.3))
100
(100*(0.58+0.42))
Single adult with no dependent children 80
(120*0.67)
58
(100*0.58)
Couple with two dependent children aged under 14168
(120*(0.67+0.33+0.2+0.2))
140
(100*(0.58+0.42+0.2+0.2))
Single adult with two dependent children aged under 14128
(120*(0.67+0.2+0.2))
98
(100*(0.58+0.2+0.2))
Three adults, no children160
(120*(0.67+0.33+0.33))
142
(100*(0.58+0.42+0.42))

As per the table above, after deducting housing costs, a couple with no dependent children and a single adult with no dependent children have the same equivalised incomes if their actual disposable incomes are 100 per week and 58 pounds per week respectively.  In other words, the equivalisation scales assume that a disposable income of 58 per week for a single adult is sufficient to give him/her the same standard of living as that of a couple with an income of 100 per week.

top

Some final thoughts

All the points above are broadly accepted by most researchers, by the EU, by the UK government and by politicians of all hues across the political spectrum.  Whether they are either understood or accepted by the general public is, however, much less clear.  Two particular problems arise:

top