redistributive effects of tax and benefits, 47 was claimed by Q5, while Q1 secured only 10. Rowntree’s analysis was also limited because he was writing
as a social analyst of consumption, before Keynes in the General Theory 1937 moved economics towards consideration of the roundabout economic conse-
quences of consumption, savings and investment decisions. After Keynes, we need to consider the roundabout consequences of expenditure by richer and poor
households and consider their decision to save as well as to consume.
These roundabout consequences of household decisions need to be analysed in a new way which cannot be read from what Keynes wrote or what the
Keynesians subsequently said. Keynes himself always assumed men’s jobs at standard wages and conditions. Practically, in The Economic Consequences of
Mr Churchill in the 1920s, or theoretically, in The General Theory of the 1930s, his problem was the same: what would happen if one group of workers like the
miners was singled out for wage cuts in a deflationary environment? By the 1940s and How to Pay for the War, the focus was the mechanical relation between aggre-
gate demand and the number of jobs at standard wages and conditions; beyond full employment, the issue was about the diversion of demand into price infla-
tion. These assumptions were carried over into the prescriptions and practice of Keynesianism, which used demand management to hold the economy near full
employment and assumed that reflation could be used to curb unemployment. The practice of Keynesianism was then undermined by its perceived failure in
the 1970s under changing conditions as the rise of international trade under- mined Keynesianism in any one country. More immediately important from our
point of view, the assumption about standard jobs was weakened by the rise of part-time, lower paid women’s jobs plus the neo-liberal fix for unemployment
in the UK, which was to price people back into work by encouraging the creation of low quality jobs with poor wages, conditions and little security.
From this broader point of view, 100 years after Poverty and more than 50 years after the General Theory, our interest in household resources and choices finds a
broader object than Rowntree’s and a different object from Keynes’. Rowntree was interested in counting poor households after estimating what was basically
necessary through costing a physical efficiency diet. We are interested in under- standing the interaction of all households with labour and capital markets so that
we understand the dynamics of inequality and the experience of ordinary house- holds. We will do this by analysing two drivers of inequality, the cheap
goodscheap jobs nexus and retirement via funded pensions, where in each case we will consider macro dynamics and micro experience. After examining the
drivers, we will synthesise the argument by presenting a kind of circular flow dia- gram whose differences from Keynesianism are as important as its similarities.
2. T
HE CHEAP JOBS CHEAP GOODS NEXUS
The UK has uneasily combined US-style neo-liberalism and a more European- style of social protection: under very limited labour market regulation, large
numbers of new poorly-paid jobs have been created without dismantling com- prehensive systems of social welfare sustained by taxation. Under these circum-
stances, the state ends up subsidising low wages via child support, family credit
and housing benefit in a society where many are unwilling or unable to pay the necessary taxation, so that the logical next step for neo-liberals would be to
dismantle social protection. This would help the Treasury but solve nothing because the removal of income support would aggravate the fundamental dynamic
problem about economic trajectory when a deteriorating composition of employ- ment shapes the composition of demand and accelerates inequality.
This trajectory could be described as the opposite of flexible specialisation which, in the original Piore and Sabel 1984 statement, was largely about a
virtuous circle in manufacturing districts. Demand was assumed to be unprob- lematic and re-skilled workers could find ever more opportunity building dif-
ferentiated, up-market products. The only threat to the virtuous circle was pathological competition and sweating, which could be curbed by employer and
worker associations on the supply side preventing a race to the bottom. Present day capitalism, however, is increasingly about the demand for services, with
workers of varying skills and competences facing the general problem that con- sumption demand must be met at the level at which it is effective. Most con-
sumers require a basic bundle of goods and services including housing, food, transport and leisure and their expenditure on such objects depends most
obviously on income.
The complication is that consumption demand comes from households not individuals. The household plays a mediating role in shaping the composition of
demand by pooling income and resources from wages, benefits, investments and other sources and jointly consuming some goods and services like housing, fuel
and food. Larger households may not only have access to more resources, but they also benefit from some economies of consumption through pooling Ringen
1997. The individual’s consumption opportunities can therefore depend as much on the other members of the household as on hisher income. This mediation
effect has been increasingly important in a society like the UK which has moved towards universalised labour force participation for women over 16, so women
have became additional albeit often lower paid wage earners. Female partici- pation rates expressed as the percentage of those of working age in employment
rose from 52 in 1971 to 69 by 2000, while male participation rates fell from 93 to 80 over the same period Labour Force Survey, various years.
The mediating effect of the household is most obvious if we examine the links between household size and income and see how income pooling can put some
of those on modest incomes into middling and rich households. Table 1 presents data that relates household income and the number of economically active
people in different income bands. The occupation of the head of household makes a big difference to household income, particularly at the higher income levels:
almost half of Q5 the richest 20 of households are headed by professionals, employers or managers. But, significantly, households can also enter Q5 if they
have two working partners combining their moderate middle class salaries. Table 1 shows that the number of economically active people in the household i.e. those
who are of ‘working age’ and not in full time education rises from 0.3 in Q1 to 2.1 in Q5. Meanwhile, the most obvious feature of Q1 households is the absence
of any wage or income from employment: half of Q1 households are headed by a retired person, and a further quarter by those who are ‘unoccupied’.
One of the important trends in the 1980s and 1990s in the UK has been an increasing inequality in the distribution of work between work-poor and work-
rich households. The proportion of the working age population living in a job- less household doubled to 15 from the 1970s to the 1990s while, at the same
time, there has been an increase in the number of households in which all mem- bers are in work Gregg Wadsworth 1996. As Taylor 2000 observes, this
trend tends to be self-perpetuating in terms of patterns of household wealth ‘in that work-rich households almost exclusively live in their own homes and there-
fore benefit from an appreciating asset. Work-poor households are increasingly marginalised into social housing, from which it is difficult to escape’ p.93. The
most significant category of work-poor households is perhaps those headed by lone parents. Single parents are less likely to be in work than other kinds of
single adults heading households; furthermore, they are less likely to remain in
Table 1 UK household composition by income group, 1995–96
Q1 Q2
Q3 Q4
Q5 All
Number of people in the household:
1.4 2.2
2.6 2.9
3.2 2.4
adults 1.2
1.6 1.9
2.1 2.3
1.8 children
0.2 0.6
0.7 0.8
0.9 0.6
Number of economically active people per household
0.3 0.6
1.1 1.6
2.1 1.1
Household disposable income £
4814 8584
13 259 19 285
34 888 16 166
of which cash benefits £ 4200
5020 3410
2090 1530
3250 Disposable income per
economically active person before cash benefits £
2047 5940
8953 10 747
15 885 11 742
Occupational grouping of head of household of quintile Employees:
19.0 29.3
56.4 74.4
79.3 51.6
professionals, employers and managers
1.5 2.5
7.8 22.3
45.1 15.8
other non-manual 4.0
9.9 15.6
17.7 16.8
12.7 skilled manual
3.9 6.4
17.3 23.1
11.9 12.5
other manual 8.1
10.0 14.9
9.9 4.9
9.5 Self employed
2.7 6.5
12.3 11.6
12.7 9.2
Retired 50.6
45.9 19.5
7.8 5.2
25.8 Unoccupied
27.9 18.4
11.9 6.7
2.7 13.5
Source: Office for National Statistics 199596; 1998. Notes: Q1 is the poorest quintile measured using household income; Q5 is the richest. ‘Head of house-
hold’ as defined by the ONS is the member of the household who owns the residence, or to whom the residence is supplied by a third party. Where two members of different sex have an equal claim, the male
is taken as the head; where two members of the same sex have equal claim, the elder is taken as the head. ‘Economically active’ is defined as those between 18 and 6065, not in full-time education.
work if currently employed or to enter work if currently unemployed Taylor 2000: 94.
Much of the increasing inequality in the second half of the 1990s has been caused by rising incomes from work for those at the higher end of the scale. But
it is important to note that Q4 and Q5 in the UK are the privileged many, not the super-rich. The number of the rich has increased so that in its ‘Survey of the
New Rich’, The Economist explains how ‘the world now has 7.2 million people with investable assets of at least 1m, up from 5.2 million in 1997’ 16 June 2001:
3. ‘High net worth individuals’ interest the likes of Merrill Lynch-Cap Gemini in their ‘World Wealth Report’ and provide new opportunities for companies
like Volkswagen to sell luxury motor cars. But these individuals do not account for the 40 of households in Q4 or Q5. The analysis in Table 1 brings out the
point that Q4 and Q5 include many people like us, when the median income in Q4 is relatively low at £19 285. If, as the previous section suggested, Q4 and Q5
are just as much part of the problem of inequality as Q1 and Q2, it is important to be clear that the majority of Q5 and all those in Q4 are not rich.
After making these points about the distribution of household income and employment, the argument now turns back to the link between cheap jobs and
cheap goodsservices. The difficulty is that the existence of this link cannot be demonstrated from the available official statistics though the linkage can be
corroborated through analysis of household expenditure patterns. These patterns are considered below, before the argument goes on to consider some complex-
ities about the relation between demand and employment, which again reinforce the importance of the household as a mediating institution.
The low absolute spending power of the bottom three quintiles implies that cheap goods and services must be favoured by these groups as they meet their
basic needs, while Q4 and Q5 can consume a wider and more expensive range of products. Tables 2 and 3 present data on similarities and differences in con-
sumption patterns in various household quintiles. Table 2 presents a breakdown of expenditure for Q1 to Q5 by category. In terms of broad categories, the pro-
portions of household income spent on different objects are similar across the quintiles from top to bottom: housing, for example, takes 14 to 16; household
goods 7 to 8; personal goods and services 3.5. The major income-related difference is that poorer households spend a greater proportion of income on
fuel, power and food while richer households spend more on mobility, leisure and lifestyle, and savings and investments.
Table 3 presents expenditure patterns in a rather different way, illustrating the absolute amount spent rather than the percentage splits on various goods. This
provides a more revealing insight into the greater spending power of Q4 and Q5, especially on leisure and lifestyle services: for example, Q5 spend more on
leisurelifestyle or on savings and investment than Q1 spend in total. Q5 also spends nearly as much on leisure services as Q1 to Q4 put together. While house-
hold, lifestyle and leisure services are the fastest growing component of Q5 expen- diture, substitution effects within categories also exacerbate this trend. Q5 spend
a lower proportion of their income on food and non-alcoholic drink than Q1–4, but around one quarter of this much larger spend is on restaurant and café meals,
where a significant part of the bill is for labour services.
This raises one of the interesting and controversial issues about household expenditure. To what extent do the comfortable households in Q4 and Q5 directly
consume labour services provided by other households as substitutes for their own domestic labour? This ‘outsourcing’ of domestic activity can take place out-
side the home through purchase of restaurant meals or laundry services, or inside the home with purchase of domestic cleaning, childcare or gardening.
Household expenditure patterns should disclose the extent and growth of domestic outsourcing. But much of this activity takes place within the informal
economy and the evidence from some advanced economies does not confirm the popular image of a large and rapidly growing domestic service sector. Using
Table 2 UK share of household expenditure by quintile, 1995–96
Expenditure category Q1
Q2 Q3
Q4 Q5
Necessities: 51.6
46.1 42.8
39.1 35.6
Housing net of benefits 14.3
15.3 15.9
14.5 14.2
Fuel and power 9.1
6.0 4.2
3.4 2.6
Food and non-alcoholic drinks 23.9
20.2 17.3
15.6 13.4
Clothing and footware 4.4
4.5 5.3
5.5 5.4
Personal PreferenceLifestyle: 6.7
6.2 6.1
5.1 4.5
Alcoholic drink 2.8
3.2 3.8
3.5 3.6
Tobacco 3.9
2.9 2.3
1.6 0.9
Lifestyle: 17.7
16.3 14.9
14.8 15.1
Household goods 8.3
8.1 7.1
7.3 6.6
Household services 5.6
4.9 4.5
4.0 4.9
Personal goods and services 3.8
3.3 3.4
3.5 3.7
MobilityLifestyle: 8.4
12.3 13.5
14.3 13.4
Motoring 6.3
10.2 11.8
12.7 11.3
Fares and other travel costs 2.1
2.1 1.7
1.6 2.0
LeisureLifestyle: 11.9
13.7 13.9
14.4 16.1
Leisure goods 3.9
4.2 4.4
4.2 4.2
Leisure services 7.4
8.9 8.9
9.4 11.2
Other 0.6
0.6 0.7
0.8 0.8
Savings and Investments: 3.5
5.4 8.8
12.2 15.3
Insurance 1.9
2.6 4.7
6.3 7.9
Mortgage capital etc 1.4
2.7 2.5
4.4 5.3
Savings 0.2
0.2 1.4
1.2 1.9
Loan repayments 0.0
0.0 0.2
0.3 0.2
Total 100
.0 100
.0 100
.0 100
.0 100
.0
Source: Family Spending, ONS 1996; Economic Trends, ONS 1996. Notes: Households are divided into quintile groups according to their income, where Q5 and Q1 are the
richest and poorest one fifth respectively. Expenditure excludes income tax and national insurance charges.
Australian expenditure data for 1993–4, Bittman et al. 1999: 255–60 find that take-aways for the poor or eating-out for the comfortably off are nearly universal,
and that childcare, in the form of fees for pre-school care, is popular. But, only 13 of Australian households in the top income decile apparently purchase house
cleaning. Bittman et al. 1999: 249 conclude that discussion of domestic out- sourcing ‘has run ahead of the facts’ and we might add that the relevant facts are
not easily ascertained.
Whatever the level and trend of domestic outsourcing, the composition of demand for household services is shaped by the distribution of income and the
Table 3 UK levels of average annual household expenditure £ by quintile, 1995–96
Expenditure category Q1
Q2 Q3
Q4 Q5
Necessities: 2493
3958 5670
7543 12 408
Housing net of benefits 689
1317 2108
2792 4968
Fuel and power 439
517 563
664 905
Food and non-alcoholic drinks 1152
1738 2290
3017 4658
Clothing and footwear 213
386 709
1069 1876
Personal PreferenceLifestyle: 321
530 803
982 1553
Alcoholic drink 132
277 500
668 1243
Tobacco 189
253 303
316 310
Lifestyle: 850
1398 1979
2858 5276
Household goods 398
694 942
1416 2306
Household services 270
418 591
775 1697
Personal goods and services 183
285 446
667 1274
MobilityLifestyle: 406
1058 1785
2765 4664
Motoring 302
878 1559
2453 3,953
Fares and other travel costs 103
180 227
313 711
LeisureLifestyle: 574
1176 1849
2786 5634
Leisure goods 189
362 577
818 1449
Leisure services 356
764 1184
1820 3908
Other 28
50 88
148 276
Savings and investments: 170
465 1171
2351 5353
Insurance 90
219 626
1218 2744
Mortgage capital etc. 69
228 338
844 1850
Savings 11
21 185
230 675
Loan repayments -
3 21
59 84
Total 4814
8584 13 259
19 285 34 888
Source: Family Spending, ONS 1996; Economic Trends, ONS 1996. Notes: Households are divided into quintile groups according to their income, where Q5 and Q1 are the
richest and poorest one fifth respectively. Expenditure excludes income tax and national insurance charges.
fact that the vast majority of households in Q4 and 5 are comfortably off but not rich. Q4 and 5 households that want domestic services must stretch their income
so they can also afford aspirational expenditure on holidays and leisure or on branded clothes and household goods. Such households will not pay generously
for domestic services, which implies low wage rates and, given that much of this is in the informal economy, avoidance of social costs such as national insurance
contributions, pensions or holiday pay. Many households behave in much the same way as corporate employers who have used part-time workers to reduce
overall labour costs. Some Q4 or Q5 households will be good employers and domestic work may offer desirable flexibility for some workers. However, the
growth of unregulated domestic outsourcing with the current income distribution is likely to reflect, rather than reduce, household inequality.
It is important to avoid scape-goating Q4 and Q5 or caricaturing the service economy we all sustain. Even those in Q1 and Q2 buy take-away food, so that
low-wage service jobs benefit poor and rich households alike. And even in per- sonal services, some jobs in expensive restaurants or personal therapy provide
good levels of wages and benefits. It is equally important to remember the com- plication that household composition mediates the effects of low pay on con-
sumption patterns. A poorly paid individual can enjoy more expensive goods and services than their individual income would allow by pooling resources and shar-
ing consumption with another wage earner. Thus high rates of workforce par- ticipation in the UK partly compensate for and effectively subsidise the
proliferation of low wage jobs in the corporate and public sectors and, in doing so, complicate the linkage between cheap jobs and cheap goods. But, equally, those
with one poorly paid job or no employment must tailor their consumption to fit their limited resources and the number of Q1 to Q3 households in this group is
hardly trivial.
When all the qualifications have been registered, the basic point remains that, in a country like the UK, inequality accelerates inequality. The composition of
household demand shapes the composition of employment when demand meets supply in the labour market so as to create jobs with poor remuneration and
benefit profiles. In most advanced economies there is no shortage of individuals from Q2 and Q3 households who can and must earn what they can in a labour
market defined for them by the erosion of traditional well paid jobs for men. UK manufacturing employment is down from 7 to 4 million over the past
30 years and all of the growth in employment in the major advanced economies has been in services. As we argued in Social Settlement, Household Lottery Froud
et al. 1997, whether in firms or for households, service jobs bring lower remun- eration on average, not least because of lower social charges arising from the
spread of part-time and contingent work.
These macro dynamics arising from the cheap jobscheap goods nexus create a distinctive micro experience where, just as in Brecht’s Hollywood, heaven can serve
as hell for the unprosperous and unsuccessful. Q4 and Q5 include the many whose privilege is to employ Q2 in fulfilling their service needs as they eat out and buy
in cleaning and childcare; while grumbling that state health and education does not provide the standard of service they expect. Significantly, getting more out
of the National Health Service and state schools, without increasing taxation, is the aim of both the Labour government and the Conservative opposition in the
UK. The micro experience is also one of insecurity for the prosperous and suc- cessful because job tenure is uncertain and households can re-form. Households
or individuals who form new households can move up or down through the quintiles as a result of gaining or losing a job, the addition of a new household
member or because of death, divorce or other causes of household breakdown.
Thus households often move between quintiles, as a result of changes in income or household composition, although the extent of movement tends to be fairly
limited. Jenkins 2000 reports on the first six years’ results from the British Household Panel Survey, whose findings provide a new, dynamic view of inequal-
ity. On the official UK measure, ‘low income’ households are those with less than half the average income. By this standard, Jenkins finds that two-thirds of house-
holds were never in the low income bracket at the annual interview dates in each of the six years, 1991–96. At the other end of the scale, 2 were classed as low
income in every year, while 32 had low income at least once and 19 at least twice during the six years. A significant proportion of households therefore
experiences low incomes intermittently, with lone parents the most likely group to be on low incomes for a sustained period of three years Jenkins 2000: 115–22.
Many households which experience poverty move up the income scale though ‘most year-to-year mobility is short range’ Jenkins 2000: 114 with re-entry fairly
common and exit more difficult the longer the time spent in poverty.
The evidence is thus such that it is difficult to discuss the role of the house- hold in pooling resources, mediating risk and sharing consumption without intro-
ducing the idea of trajectory and life course. As O’Donnell and Hancock 2000 argue, ‘whereas postwar life courses tended to be few, predictable and linear, the
emerging life course appears to comprise transitions that are multiple, random and recurring’ p.14. Certainly, aggregate statistics disguise not only household
differences in experience, but also their life courses. One of the defining charac- teristics of current household experience is precariousness of income and expendi-
ture possibilities.
While all households are exposed to risk, those towards the bottom of the income distribution are most vulnerable. Q1, while poor, have some certainty of
income since most of it is derived from benefits such as state pensions. For those in Q2, where there is limited labour force participation, a particular kind of pre-
cariousness can be observed, characterised by constraints and difficult choices. This can be illustrated by looking at the experience of a group of people who
can be described as the ‘motoring poor’ in London. These are members of house- holds in Q1, Q2 and sometimes Q3 who run a car, while barely affording to do
so, usually to allow one of the members to work or to travel to a better paying job than those available in the immediate neighbourhood. For this group, motor-
ing has much the same significance as work boots for Rowntree’s labourers: the car is the price of their participation in the labour market, where local labour
markets are underdeveloped, employment is often at anti-social or irregular times and public transport provision is poor.
Our research in two London areas looks at the sacrifices some low income, working households make to run cars.
2
This is especially a problem of outer London, as Table 4 shows by comparing two areas. In the outer area, Epping,
car ownership rates are much higher than in Hackney in inner London, which has better public transport. Table 5 confirms that households are using their cars
to travel to work. In Epping, the car is more commonly used by all socio- economic groups except farmers and agricultural workers to travel to work than
it is in Hackney. And if we take London as a whole, then the average car mileage of poorer people is almost 3000 per year, which is an understatement given that
density of car use is much lower for such groups. Seventy-three per cent of all
Table 4 Car ownership of working households of households
Hackney Epping Forest
inner London outer London
No car 40
5 1 car
44 36
2 cars 15
40 3 or more cars
2 18
Source: Data derived from 1991 Census County Report Inner London and Essex. Note: Hackney is an inner London borough; Epping is in outer London.
Table 5 Household car usage in inner and outer London
Socio-economic group of socio-economic group that drives to work
Hackney Epping Forest
inner London outer London
Employers and managers 37
69 Professional workers
29 61
Intermediate non-manual workers 25
57 Junior non-manual workers
14 45
Manual workers foreman, supervisor, 40
67 skilled and own account
Personal service and semi-skilled manual workers 17
49 Unskilled manual workers
12 33
Farmers and agricultural workers 46
40 Armed forces and others
13 28
Total 25
56
Source: Derived from 1991 Census County Report Inner London and Essex. Note: Figures exclude those who receive a lift to work by car.
miles travelled by Q1 individuals are by car, not much less than the 85 of the richest individuals.
Poor households spend a large part of a limited income on the car because old cars may be cheap to buy, but all cars are expensive to run. As Table 2 showed,
for Q1 households 6.3 of all expenditure is on motoring, so that the one-third of Q1 who actually run cars spend a stunning 20 of their small incomes on
motoring. By contrast in Q5 households, 11.3 of all expenditure is on motor- ing, and 94 of Q5 households run at least one and quite often two cars per
household. Q1, Q2 and even Q3 can only run a car at significant sacrifice. Our interviewees from an estate in the outer London district of Epping Forest
confirmed this, when asked about what they had to forfeit to afford the costs of motoring:
‘cars are getting dearer to run. I’m cutting things out here and there to keep the car on the road’
‘I would spend the money saved on a car on the kids––they’re the next most expen- sive thing’
‘I would spend the money saved on clothes or holidays’ ‘I don’t eat so much as before. I don’t have luxuries because of the driving costs . . .
I would spend more money on food if I had no car’ ‘I would spend the money saved on beer’.
We have come all the way from Rowntree’s poor, who could not afford a decent meal because of drink, to the motoring poor who cannot afford a decent drink