Memorandum submitted by David Webster
(ES 11)
The argument of this memorandum is that the
Government's employment strategy places insufficient emphasis
on local labour market conditions, which vary widely across the
UK. Supply-side measures such as the New Deals need to be supplemented
by much more vigorous policies to boost demand for labour in the
most disadvantaged localities. This argument has been set out
previously in a number of publications which are referenced below.
As far as possible, this memorandum updates the research evidence.
How much has worklessness really fallen since
1997?
In order to judge what effect the current economic
slowdown might have on the labour market, it is necessary to examine
the effect of the economic expansion of the second half of the
1990s.
Judged in terms of unemployment as officially
measured, the record has been good. This applies whether the claimant
count (those receiving unemployment-related benefits) or the wider
ILO definition (looking and available for work) is used. Claimant
unemployment has fallen by about 600,000 since 1997 and ILO unemployment
by rather less. However, the record has been less good than it
appears, for two reasons:
(i) The New Deals themselves have a "register
effect", in other words they reduce the claimant count without
getting people into work. People on "options" are not
counted as unemployed, and those who are "sanctioned"
for non-compliance are also removed from the count. Calculating
the total effect is complex but the best estimate seems to be
around 50,000.The "options" effect is more important
than the "sanctions" effect because although the latter
are running at around 18,000 per year they do not last so long.
(ii) More important is the movement of people
out of unemployment but not into jobs, ie into economic inactivity.
The key labour market status from this point of view is long term
sickness, although there are other destinations also. At November
2001 there were 3.008m working age people claiming sickness and
disability benefits compared to only 2.803m at May 1997, an increase
of 205,000. On any sensible view, many of these people are really
unemployed. This issue is explored further below.
Overall, up to 40 per cent of the 600,000 apparent
reduction in claimant unemployment since 1997 is probably not
real. Nevertheless, there has still been substantial real improvement.
How much of the labour market improvement has
been due to the Government's employment programmes?
There have now been several studies of the economic
impact of the New Deals. The National Audit Office's estimate
(NAO 2002) is typical. This puts the impact on youth unemployment
of the New Deal for Young People at about 25,000-45,000, with
a positive effect on employment of other groups of about 10,000.
These are small numbers in relation to the total real fall in
claimant unemployment of about 350,000 since 1997. Indeed, the
real effect of the New Deals actually appears to be smaller than
their "register" effect.
Employment Zones have also had a very limited
impact. Up to September 2001, 54,389 participants had started
and claims had been made for 16,643 jobs.
The New Deal for Lone Parentsfew of whom
are found in the claimant counthas had a similarly small
effect. DWP figures to end-January show 146,000 participants entering
jobs, but most of these would have entered employment anyway.
The Working Families Tax Credit has increased
incentives to enter work, particularly for lone parents, and may
have a greater effect in the longer term. But to date take-up
has been clearly been largely by people already in work (Webster
2001). This was predicted by HM Treasury (2000b). Their maximum
estimate for the additional labour supply impact on lone parents
and couples of 122,900 is small in relation to their then projected
UK increase in claims of 583,000.
Total placements of people in work are not the
same as the net impact on unemployment. Comprehensive calculations
appear to be available only for the New Deal for Young People,
and estimates are therefore necessarily approximate. However,
the overall net impact on real claimant unemployment of all the
Government's employment programmes seems unlikely to have been
as much as 100,000 and is probably nearer 50,000. These programmes
have played a relatively small role in the labour market improvement
since 1997. Economic growth has been far more significant. The
current economic slowdown is therefore extremely important from
an employment point of view.
Long Term Sickness
The UK has the highest rate of working age sickness
of all 15 European Union (EU) countries. The UK had 7.0 per cent
of its working age population inactive due to long-term sickness
in 1999, compared to only 2.1 per cent in Germany and 0.3 per
cent in France. Britain compares favourably with the rest of the
EU in terms of ILO unemployment, with eight countries having a
higher rate. But if the working age sick were to be added to the
unemployed, Britain would become the third worst, after Finland
and Spain (Figure 1).
The Government acknowledges that the high level
of long-term sickness is a problem, and has attempted to address
it by tightening eligibility for Incapacity Benefit and through
specific labour market programmes which, though small, are being
expanded. However, it appears that the net flow into sickness
is continuing. The DWP's latest Client Group Analysis shows that
87,000 people who were claiming JSA in November 2000 were claiming
a disability benefit in November 2001. This is 9.1 per cent of
the group, and even after allowing for flows in the other direction,
the net flow is still 4.8 per cent. This is far greater than from
any of the other non-working groups. <mr40> A recent official
analysis based on the Labour Force Survey (LFS) showed that the
propensity of the unemployed to move into inactivity rose steadily
during the 1990s, from about 8 per cent to 13 per cent per quarter
(Young 2001). In other words, much of the improvement in unemployment
has been offset by a worsening in sickness.
Local authority level data on sickness back
to May 1995 have recently become available from DWP. Over the
whole period May 1995 to November 2001, the proportion of the
Great Britain working age population on sickness benefits rose
from 7.4 per cent to 8.4 per cent. In the areas with the worst
sickness rates, there was generally some improvement, but this
has been extremely slight for a period of such prolonged economic
expansion, underlining the seriousness of the problem. In Glasgow,
the proportion fell from 19.7 per cent to 18.7 per cent, Liverpool
from 18.2 per cent to 17.9 per cent, and Easington from 25.7 per
cent to 23.3 per cent, but it rose in Manchester from 13.6 per
cent to 15.4 per cent.
A full account of the evidence that many if
not most of the long-term sick are really unemployed is in Webster
(2002a). While the Government has yet to acknowledge officially
that the UK's true unemployment rate is far worse than shown in
the official statistics, this analysis is now widely shared. Since
the mid-1980s, British policy has been based on the assumption
that people can be pushed into work by a combination of low unemployment
benefits and direct pressure. However, as Beatty et al
(2000) explain in detail, there are such a large number of people
both employed and unemployed who can potentially qualify for sickness
benefits that the actual effect of this approach is to create
a large flow into sickness as well as into jobs.
A recent study showed that the large increase
in disability benefit claims in the USA since 1984, especially
among unqualified people, was strongly related to a rise in the
value of the benefits relative to relevant earnings, to relaxed
criteria for benefit award, and to declining employment prospects
(Autor & Duggan 2001). The processes involved in Britain appear
to have been similar, although the proportions of the labour force
involved here appear to be some 50 per cent greater than in the
USA (Autor & Duggan 2001, Nickell & Quintini 2001).
The "register" effect of the New Deals
has been particularly large in relation to long-term claimant
unemployment. In effect, it has become more or less impossible
for anyone to become long-term claimant unemployed. However, long-term
unemployment has been transferred on a substantial scale into
long-term sickness, where it is more intractable.
I am grateful to Paul Bivand of the Centre for Economic and
Social Inclusion for advising me of the size of the "register
effect" of the New Deals and for pointing out the data on
flows into sickness benefits.The Local authority and parliamentary
constituency claimant unemployment rates published in Labour Market
Trends use an invalid denominator. The Office for National Statistics
has already concluded that they should be withdrawn and an announcement
is expected soon. The House of Commons Library already publishes
different claimant unemployment rates for parliamentary constituencies
using a valid methodology. Claimant unemployment rates quoted
in this memorandum have been corrected to remove the distortion
caused by commuting between local authority areas. The methodology
is explained in Webster (2002b).
Geographical Variations in Worklessness
Despite the overall improvement, little progress
has been made in reducing labour market disparities between areas.
At January 2002 claimant unemployment varied across the 408 local
authorities in Great Britain by a factor of over 20, from 0.5
per cent in Hart, Hampshire, to 10.8 per cent in Tower Hamlets
(Webster 2002b). <mr40> These differentials are almost completely
static. The correlation across local authorities between Census
unemployment in April 1991 and claimant unemployment in January
2002 was 0.90 (Figure 2). In other words, if an area had high
unemployment a decade ago then it still has high unemployment
now. There has been little "convergence" in terms of
rankings.
Areas with high levels of unemployment also
have high levels of working age sickness. There is a strong correlation
across local authority areas (0.72 at May 1999) between the claimant
unemployment rate and the proportion of the working age population
claiming sickness or invalidity benefits. The areas with the worst
problems are cities and coalfields. Although in some former coal
and steel towns occupation-related illnesses are probably very
high, the main factor is the loss of manufacturing and mining
jobs, as has been shown by Turok & Edge (1999) and Beatty
& Fothergill (1997). While the sickness rates of 25 per cent
in Merthyr Tydfil and 23 per cent in Easington may have a substantial
occupational component, this could scarcely apply to the 15-19
per cent rates in the big cities of Glasgow, Liverpool and Manchester.
The 1981 Census data suggest that even in the former coal towns,
the occupational illness component would not exceed some 7 per
cent.
Official Denial of Local Job Shortages
The Treasury and DWP are still officially maintaining
a radical denial that there are any important labour market disparities
between areas. This is the fundamental obstacle to addressing
the important concentrations of worklessness effectively.
HM Treasury (2000) is the principal text arguing
this case. The weaknesses of its arguments are discussed in detail
in Webster (2000b). In summary, they are:
(i) In order to sidestep the evidence on
the geographical concentration of worklessness, the paper used
job vacancy data to suggest that there are enough vacancies to
employ all the unemployed and that these vacancies are sufficiently
evenly distributed across the country to deal with geographical
concentrations. The same type of argument was still being made,
in a more muted way, by the Treasury and DWP as recently as last
November (HM Treasury 2001). But ONS has now suspended the job
vacancy series, because of its distortion by the trial of the
Employment Direct scheme (Financial Times, 12/9/01). The
paper also neglected the different vacancy characteristics of
different economic sectors. Many services have higher labour turnover
than traditional industries, so that increases in the stock of
vacancies do not necessarily imply greater job availability.
(ii) The paper used the invalid "workforce"
claimant unemployment rates for local authority areas, thus producing
misleading conclusions (see Note 2).
(iii) The paper showed a lack of understanding
of the spatial structure of local labour markets. It argued that
because high levels of vacancies are found in cities alongside
high levels of unemployment, the unemployment cannot be due to
lack of labour demand. But for the most part these vacancies merely
represent turnover among the commuter workforce. Most British
cities have central business districts with numerous, mainly white
collar vacancies. This is why it is often true that "areas
of high unemployment lie within easy travelling distance of areas
where vacancies are plentiful". But the vacancies have to
be set against the labour force in the city's whole commuting
catchment area.
Lone Parenthood
Government employment policy has a strong focus
on lone parents. Lack of attention to the geography of worklessness
has particularly serious consequences in relation to them.
The Census showed that between 1981 and 1991
a very strong relationship emerged between lone parenthood and
unemployment across local authorities (Webster 2000a) (Figure
3). LFS data show that this relationship between did not weaken
between 1991 and 2000 (Figure 4). Taken together with other evidence,
this indicates that the huge rise of British lone parenthood since
the early 1980s has been caused by the huge increase in male unemployment.
This was a repeat of what had occurred earlier in the USA and
was well documented there. But the Government sees the rise of
lone parenthood as something which occurred in parallel with the
rise of unemployment but not because of it. HM Treasury (2001,
paragraph 3.20) described the doubling of lone parent families
between the late 1970s and mid 1990s as "demographic",
ie independent of the economy. But by using the invalid "workforce"
unemployment rates, the paper understated the strength of the
connection between lone parenthood and unemployment. It showed
a correlation of 0.73 across GB local authority areas of lone
parents on Income Support (expressed as a percentage of the working
age population) with the local unemployment rate, when the true
figure is much higher, at 0.83.
The close relationship between lone parenthood
and unemployment across areas means that employment policies based
simply on supply-side incentives to lone parents to take jobs
cannot work. It was loss of jobs from the affected areas which
caused family breakdown. This same loss of jobs means that local
labour demand is too weak to re-employ the large numbers of lone
parents involved. Performance data on WFTC is already demonstrating
this. While the overall rate of claim has increased in all areas,
the increase over Family Credit at August 2001 was only 10 per
cent of lone parent families in the highest unemployment areas,
where they are concentrated, compared to 20 per cent in the lowest
unemployment areas (Webster 2002a).
This relative position reflects the fact that
the proportion of lone parents in work was already twice as high
in the lowest as in the highest unemployment areas, again a position
which was unchanged between 1991 and 2000 even though the proportion
in work rose by around 10 per cent across the board over these
nine years (Figure 5). To attain the Government's target of a
70 per cent employment rate for lone parents in the next decade
would require negligible further increase in the most prosperous
areas, but a virtual doubling of the rate in the poorest areas.
This would be an increase three times greater than occurred in
the last decade.
The Theoretical Basis for Supply-Side Employment
Policies
The underlying problem is the theoretical basis
for the present package of Government employment policies. It
was based primarily on the ideas of Layard, Budd and the OECD
which derived from research in the mid-1980s (Webster 1997). The
dominant preoccupation was with long-term unemployment, which
appeared to have risen relative to short-term unemployment. It
was argued that being unemployed in itself makes people less employable
("state dependence"), so that the primary cause of continuing
high unemployment in Britain was the fact that unemployment had
been high before: the "pool" of long-term unemployed
people were difficult to get into work and, by exerting little
downward pressure on wages, made the "natural" rate
of unemployment higher than it need be. This conclusion was based
on a mistake in statistical analysis, as was confirmed by subsequent
experience which showed long-term unemployment coming down in
line with total unemployment exactly in the same way it had gone
up.
The form of the policies derived from the assumption
of "state dependence". Since the employability of every
unemployed person was assumed to be undermined by their unemployment,
all had to be put through a programme, irrespective of their individual
characteristics. Primary emphasis was placed on the young long-term
unemployed because they had the longest lives ahead of them and
were assumed to be the most impressionable. The more personal
treatment given to the unemployed under the New Deals has certainly
been widely appreciated, but the consequence has been that resources
have been diverted away from those with genuine employment handicaps.
Cost pressures have dictated that the programmes are inadequate
to meet the needs of those with more complex problems. The study
by Hoogvelt & France (2000) in Sheffield showed that the young
people who did best in the New Deal were those who were already
most advantaged to start with. Britton (2002) reported on the
basis of an Employment Service study of young people "sanctioned"
that "The support available on the New Deal, while it is
sufficient for the majority of the population, is not sufficiently
intensive, and comes too late, to make any real impact on this
group". Similar criticisms of the New Deal are often heard
in relation to homeless young people. Many in this group do not
even get picked up by the programme, because of their mobility
or because they are not claiming JSA. For those who do, the "Gateway"
is too short and inflexible to address their need for intensive
pre-vocational training.
The rise of working age sickness dates only
from the mid-1980s and had not been noticed by the time the Budd-Layard-OECD
research was carried out; all their arguments were cast in terms
of claimant or ILO unemployment. This was why the Government's
initial policy package placed very little emphasis on this vital
group.
The greatest weakness of the Budd-Layard-OECD analysis
was that it overlooked the geography of unemployment at the local
labour market level. As a result, it did not pick up the importance
of local declines in employment, particularly in manufacturing
and mining. This led to neglect of the demand side of the labour
market. Even where locally high levels of unemployment were recognised,
they were seen simply as concentrations of less employable people:
hence Employment Zones, which in spite of the name a supply-side
programme. This analysis also led to underestimation of the specific
importance of manufacturing employment decline in Britain's unemployment
problem. The decline has been far greater than in other OECD countries.
Robert Rowthorn notes that since 1973, British and USmanufacturing
output have increased by 14 per cent and 114 per cent respectively
(Rowthorn 2001; see also Rowthorn 2000). It is essentially because
of the poor performance of manufacturing that the south east,
London and the east regions increased their economic value by
24 per cent in 1995-2001 compared to 5 per cent in the north east
and 9 per cent in Scotland (Cambridge Econometrics, Financial
Times 4/3/2002).
The same underemphasis on manufacturing has
led to heavy reliance on consumer spending in producing the economic
expansion of the late 1990s, which is consequently now threatened.
Mervyn King, Deputy Governor of the Bank of England, commented
recently: "The trade deficit has widened sharply in recent
years, reaching some 3 per cent of GDP at the end of last year
. . . the increase . . . would have been even larger had it not
been for an improvement in the terms of trade" (ie rise in
value of the pound) ". . . The result . . . is that . . .
Over the past five years, domestic demand has risen by over 6
per cent more than output . . . In turn, net trade has made a
negative contribution to output growth for six years in succession.
A continuation of that trend would be unprecedented in Britain's
modern economic history. The need to rebalance the British economy
is clear." The unbalanced nature of the expansion is one
of the main reasons why it has done so little to reduce the disparities
in labour market prospects between high and low unemployment areas.
The high unemployment areas are generally those which were previously
most dependent on manufacturing, and the employment trends of
the later 1990s have generally continued to work against them.
Lone parents did not feature in the Budd-Layard-OECD
analysis. They were added to the New Deal on the basis of other
ideas which emphasised the role of social security benefits in
undermining their incentives to work. Official thinking in relation
to them however suffers from the same lack of awareness of the
geography of unemployment. At the time the key decisions were
being taken, policy makers were unaware of the evidence on the
local level lone parenthood-unemployment linkage which had emerged
from the 1991 Census. They were impressed by the much higher employment
rates of lone parents in Europe, not realising that no European
country has anything like the same kind of geographically concentrated
lone parenthood problem related to industrial decline. The UK
lone parenthood rate is twice as high as the next highest in the
EU.
Rebalancing Policy: The Role of Labour Demand
Britain has a severe localised problem of structural
unemployment. The main reason for the worklessness of the workless
groups is their location in weak local economies. As argued previously
(Turok & Webster 1998), labour supply-side programmes therefore
need to be complemented by effective local labour demand-side
programmes.
This means the following:
(i) Much greater emphasis on policies to
promote the growth of appropriate employment in the areas with
the lowest employment rates. This would mean stronger urban and
regional policies, and much more emphasis on derelict land reclamation
and transport infrastructure, with adequate funding and effective
instruments. Spending on all these policy areas is currently relatively
low, and the Chancellor's urban regeneration package of autumn
2000 is very limited. In the case of Scotland, Wales and Northern
Ireland, closer liaison would be needed between Whitehall and
the devolved administrations, since the latter control many of
the instruments of local labour demand side policy. In Glasgow,
for instance, approaching 10 per cent of the total land area is
vacant or derelict; but there is no dedicated land renewal budget;
there is no budget for transport projects required for regeneration;
and although the most important single project from a regeneration
viewpoint, the M74, has been approved, the earliest possible completion
date, which is not guaranteed, is 2008, 11 years after the present
Government came into office. At present, because Whitehall believes
that labour demand-side policies are not important, it takes little
interest in how they are used, and Government departments and
devolved administration are under little pressure to focus on
promoting employment in the areas of greatest labour market distress.
The voice of local Government in the high unemployment areas also
needs to be given far greater attention since local authorities
usually have by far the best appreciation of the local obstacles
to employment growth.
(ii) A more supportive policy towards manufacturing.
(iii) A greater focus on the long-term sick,
who are the largest workless group. This however needs to go beyond
the present purely supply-side approach to address the issue of
labour demand in the areas where the sick are concentrated.
(iv) Greater support to those groups who
have genuine labour market disadvantages, and less emphasis on
programmes lacking precise targeting.
(v) Greater local flexibility in policy,
so that the right combination of demand-side and supply-side programmes
can be implemented to meet the needs of the particular locality.
This author has previously suggested the adoption of a system
of Local Employment Plans, in which local authorities with high
unemployment would be invited to propose how relevant public expenditure,
both central and local, in the area should be best directed (Webster
2000c).
(vi) It would also be appropriate to review
policy on the level of unemployment benefits. As Nickell &
Quintini (2001) point out, other countries use effective combinations
of higher unemployment benefits with different approaches to getting
people into work. From every point of view, it would be better
to have more of the unemployed featuring in the claimant and ILO
figures.
(The views expressed in this memorandum are
not necessarily those of the Council.)
David Webster
Chief Housing Officer (Policy Review and Development)
Glasgow City Council
30 April 2002

|