Select Committee on Education and Employment Appendices to the Minutes of Evidence


Memorandum from Peter Dolton and Yvonne Balfour, University of Newcastle upon Tyne, Department of Economics[3]


  This paper briefly reviews the different forms of UK government intervention in the training of young people over the last 35 years. The New Deal scheme introduced in 1998 is described and its progress tracked up to the end of 1999. Evaluation of the New Deal, based on data relating to 144 small geographical Unit of Delivery areas, is presented. The results show that the policy has had strictly limited impact on the local level of unsubsidised jobs created and that policy effectiveness is constrained by demand conditions.


  The British Prime Minister, Tony Blair, claims that youth unemployment, and the crime and other social problems associated with it, cost taxpayers £10 billion a year. The Labour Party's proposals contained in the document "New Deal for a Lost Generation" suggests a way to tackle this problem. Are the training policies proposed new or simply a rehash of past failed attempts of government to solve the problem? The New Deal should be evaluated in the context of the recent history of the youth unemployment problem and the government training policies which have preceded it over the last 20 years.

  Many countries in recent years have introduced large scale state training schemes in an attempt to alleviate youth unemployment problems. It is unclear whether such training schemes are an adequate substitute for a co-ordinated system of private sector training which provide the appropriate training investment incentives for both individuals and firms. From a more positive perspective the education and training of a country's young people is vital for future economic growth.

  Training is a form of human capital investment which yields a rate of return via a higher marginal product of the trained labour. It is important to distinguish between "firm specific" human capital which results from the acquisition of skills which are of value only to a specific firm, and "general" human capital which is valuable for any firm. Economic theory suggests that individual firms would have no incentive to provide general training if workers trained in general skills would, once trained, leave the training firm to take a higher paying job elsewhere. In contrast the more "firm specific" the human capital that the individual acquires, the less incentive there is for that person to leave the firm. It has been suggested that the aggregate consequence of this "public good" dimension is private sector under-investment in training because firms are unable to appropriate the return on "general" rather than "firm specific" investment. In practice the acquisition of any human capital in unlikely to be entirely "firm specific" and there should always be some scope for joint investment in training.

  A further contributory factor to under-investment may be that individuals face capital market imperfections in the possibility of taking a loan against future potential earnings after they have trained. There is also the problem that most young people have imperfect information about the potential rate of return to such investments and are typically risk averse with a high rate of discount (and time preference).

  One radical solution to this market failure is intervention by government which may take the form of employment subsidies to target specific industries or regions. Alternatively the government could impose a system of taxes on firms which don't train and a subsidy to those firms who do. The difficulties with policing such a system abound. The alternative is to have a system of large scale government provision of training. Clearly the justification for a state training scheme relies partially on the argument that training (and education generally) is a "merit good" in the sense that its value may be greater for society than the individual and therefore optimal investment in it requires some state involvement. Formal evaluation of this proposition is difficult and it is possible that state training schemes are not necessarily the most appropriate way to spend any available subsidy.


  Historically Britain's system of training involved five to seven year apprenticeships served by the individual with a single employer. In 1964 the Industrial Training Act set up a national system of Industrial Training Boards (ITBs) which had the responsibility of improving the quantity and quality of training within their industry. These had the power to give grants and levy charges on firms according to whether they were training their young employees appropriately or not. The idea was to reward the companies who were training properly and penalise those who either trained badly or relied only on poaching trained workers from other employers. Since 1970 the involvement of the government in Britain's youth training has been more or less continuous but extremely complex. It has involved a plethora of new initiatives with a bewildering array of titles. The "time line" of these schemes is mapped in Table 1.

  Over the 1973-79 period there were a variety of special employment measures to try to introduce training for young people and foster retraining amongst the long-term unemployed eg the Training Opportunities Programme (TOPS). The most radical departure came in 1978 with the setting up of the Youth Opportunity Programme (YOP). This established the principle that all those between the ages of 16 and 18 who had left school, were not in full time education and were unable to get a job should have the opportunity of training or participating in a government funded programme. This scheme was extended with the setting up of the Youth Training Scheme (YTS) in 1983. It started as a one year, low level training scheme, but has gradually changed its nature and become a two year scheme.[4]

  In 1990, YTS was replaced with Youth Training (YT) which guarantees a place to all 16-18 year olds who are without a job. YT is modelled closely on the YTS Scheme, except that YTS did not promise 100 per cent coverage and did not have unemployment benefit penalties for non-participation. YT also offers more flexibility in the length and nature of the training schemes than did YTS. A common feature of both programmes is the offering of incentives for unemployed youths to join. For example, it is impossible for people aged between 16-18 to claim unemployment benefit if they refuse a YT place, but the training allowance they receive if they do participate is slightly higher than the unemployment benefit they would otherwise receive.

  Since 1995 YT has been gradually replaced with National Traineeships and Modern Apprenticeships. Both schemes provide employers the opportunity to employ and train young people up to NVQ[5] level 2[6] for National Traineeships (16-18 age group), and NVQ level 3[7] for Modern Apprenticeships (19-24 age group).

  The plethora of new initiatives and schemes associated with training, special employment measures and "enterprise" during the last 15 years is remarkable with the emphasis shifting to the provision of Employment Training (ET) through the Training and Enterprise Councils (TECs). These TECs have been set up on a regional basis and run by private enterprise employers to attempt to solve the problems of local labour markets. The idea is to deliver suitable training programmes though local colleges and tailor ET and YT to local demands. Therefore TECs amount to regionally based employer clubs for the provision of training.

  The one common thread through all these initiatives has been the need to tackle youth unemployment. Little objective evaluation of the training which has been received on these schemes has ever been undertaken. In addition, the extent to which schemes like ET constitute an optimal allocation of state subsidy to training is debatable. In addition it is still unclear whether the provision of training is more efficiently provided by the state than the private sector.


  The transition from school to work for young people in the late 1970s and 1980s became more difficult than at any time in post-war Britain. This period saw youth unemployment rise from 10 per cent in 1978 to 27 per cent by 1984 back down to 10 per cent in 1989 rising to 21 per cent by 1992. Subsequently this has fallen to 11 per cent currently. The period also saw the collapse of British manufacturing industry and with it the decline in the numbers entering manufacturing apprenticeships from 290,000 in 1975 to 45,000 in 1996. Successive governments have chosen to spend large sums, up to £1.3 billion by 1998, at 1985 prices, on youth training schemes which, at their height involved nearly half a million young people each year from 1981 to 1988.

  The true extent of government involvement in training schemes in their different guises has been massive. Numbers on YOP rose to a peak of over half a million by 1981. Numbers in YTS averaged 400,000 over the 1983 to 1989 period. YT has involved up to 300,000 young people annually over the 1990 to 1995 period. The average unit cost of training over the 1987 to 1995 period at 1995 prices, at around £2,500 per placement per year, which is nearly double the cost of having that person on the unemployment register. This prompts the question of whether such expenditure and government intervention can be justified on simple cost grounds?[8]

  Perhaps the clearest way of understanding what has happened to the youth labour market in the UK over the last 15 years is to consider the activity rates associated with each cohort of 16-18 year olds over the period 1975 to the present. School staying on rates have remained roughly the same at 16-17 per cent over this period and the proportion of the cohort entering further or higher education has risen slowly from 11 per cent in 1975 to around 35 per cent in 1996. The biggest changes have occurred in the split of the remainder between the destinations of work, unemployment and government training schemes. The proportion in a job has fallen from 61 per cent to 41 per cent whilst the proportion who are unemployed rose from 9 per cent in 1975 to 17 per cent in 1986 and 1987. Most dramatic of all has been the proportion involved in one of the government training schemes. These schemes did not exist prior to 1978, but by 1989 the proportion of young people entering YTS was as high as 16 per cent. These changes represent a huge difference in the prospects of young people. Prior to 1975 the pattern was one of entering employment or staying on in full time education, with only a minority being unemployed. By the late 1980s the job choices for 16-18 year olds had been curtailed with only a minority able to enter the labour market directly. For a sizeable minority, up to 25 per cent, the choice is unemployment or a government training scheme.


  The simple human capital interpretation of training schemes is that they constitute an investment which will yield a social and private rate of return. However if the investment in mass state training schemes is to be assessed a more detailed analysis of the costs and benefits of this investment need to be undertaken.

  Direct measurement of the effects of training is complex. The short-run effects may involve the increased probability of getting a job and enhanced earnings, although these effects may be quite small. What could be much more important are the potential long run effects. These could be in the form of enhanced labour force participation (possibly in more suitable jobs) over the whole work history and the long-run growth in earnings of the additional human capital.

  Another view of government led training schemes is that employers use them as a screening or signalling device to filter out the most promising trainees in order to decide which to employ. Many employers have a low regard to YTS schemes in terms of the value of the training but stress their role as helping in the selection of the most able and promising employees. This is partially possible as the person can be relatively costlessly appraised whilst working "on-the-job" which will invariably be better predictor of future productivity as an employee. The value of training schemes as screens or filters for the labour market is difficult to assess.

  One of the most widely held beliefs motivating the use of government led training schemes is that they can be used as a counter-cyclical policy to ameliorate the worst effects of recession on youth unemployment. More specifically one perceived advantage of youth training schemes in times of high youth unemployment is that they may contribute to downward market pressure on the relative wages of young people.

  Another potential benefit of training schemes is that they may act as a political palliative to help the economy and society cope with the adverse high consequences of youth unemployment. The existence of a scheme like YTS can deflect criticism away form a government and may also alleviate social unrest and criminal activity which have other additional public costs.

  One important negative component in the evaluation of training programmes are the scheme's "displacement effects". These occur because the presence of a scheme in which the government funds employers to train, will affect the incentives of the employers. Most notably the displacement effects can be divided into the "deadweight loss" where employers substitute programme trainees for, say, apprentices that would otherwise have been taken on, so that in effect the firm's training bill is paid for by the government with no net increase in training provision, and the "substitution effect"—where programme trainees are substituted for some other kind of workers (e.g. part timers or older workers) and unemployment is created elsewhere.

  A clear difficulty with the assessment of wage (or employment effects) of training programmes is that often the participation on a programme is not voluntary and hence random. For example, it is quite possible that individuals of lower ability, with fewer educational qualifications, poorer family backgrounds, fewer contacts, and lower self confidence and motivation find getting a job most difficult and are hence more likely to enter a training scheme. Therefore a straight comparison of participants with non-participants in terms of their labour market progress could well be invalid.

  A second qualitative assessment problem is whether the effect of all training schemes are the same. In the same way as there are different types of jobs there is also great heterogeneity in the types of training that people may experience in different schemes and with different employers. Some schemes offer formal apprenticeships and formal qualifications whereas others, in contrast, offer training with little or no formal training and no certificated qualifications. It is inevitable that a sorting process will develop which matches the more able young people with the best placements. Hence the economic assessment of such heterogeneous training programmes is extremely complex.


  In October 1997, the Labour government published its first initiative under its Welfare to Work Strategy. Entitled "New Deal for 18-24 year olds" it marks a change in the way in which the government will deliver employment-focused services to unemployed young people. This is not seen as a short-term measure but a long-term strategy to reduce unemployment, increase the flexibility, and raise the quality of the work force. The New Deal was launched in January 1998 in 12 Pathfinder Areas as a test before the national launch in April 1998.

  It was projected that the scheme would involve up to 250,000 people at one time and cost up to £3.1 billion annually. The scheme is designed for those who have been unemployed for six months or more although in some cases, such as those groups with particular disadvantages, they can join the New Deal programme before they have been unemployed for six months. The following four options are offered as a condition of continued receipt of Job Seeker's Allowance (JSA):

    (i)  Employment Option (EO)—employers receive a subsidy of £60 per week for up to six months and a training allowance of £750 towards the cost of training their New Deal employee. The equivalent of at least one day per week must be allowed for training towards a recognised qualification;[9]

    (ii)  Full-time Education or Training (FTET)—this can last for up to 12 months and is primarily designed to provide help for those without NVQ 2 or equivalent level qualifications;

    (iii)  Voluntary Section Option (VSO)—a placement for up to six months working within a voluntary sector organisation also including the equivalent of at least one day per week working towards an approved qualification; and

    (iv)  Environment Task Force (ETF)—a placement providing work for up to six months improving the environment also including the equivalent of at least one day per week working towards an approved qualification.[10]

  In the lead up to the introduction of this programme, the government have repeatedly emphasised that there is no fifth option in this mandatory programme and failure to take up one of the four options will mean the loss of the unemployed person's JSA. The Secretary of State for Education and Employment, Mr David Blunkett, stated at the launch of New Deal in January 1998: "Staying on benefit will no longer be an option".

  Given the objectives of New Deal to help young people find and keep jobs, employers' and other providers' reactions to the programme are a crucial factor in helping to determine its success. Employers have a vital role to play not only in providing unsubsidised employment opportunities and in their willingness to provide subsidised employment and training but also providing work experience placements for those clients who have chosen the FTET option and who must have the equivalent of 26 days work experience in a realistic work environment.

  The Gateway period is the first stage of the process of assisting people into employment and lasts for a period of four months. Each young person is assigned a "Personal Adviser" who identifies for them the help they may need in becoming more employable, eg preparing CVs, giving specialist career advice, perhaps help with basic skills and building self-confidence. An intensive job-search period begins and after four months, if an unsubsidised job is not found, one of the four above options must be chosen.

  Once participants have finished their option and they still have not found an unsubsidised job, they move on to Follow-Through. Once more they have further help and guidance in job search or they may move onto another option.

  Table 2[11] shows a month by month breakdown of New Deal starters and their destinations, whether to unsubsidised jobs, other benefits, other known destination (likely to be training and education) and an unknown destination—perhaps jobs, but not recorded as such. These figures are taken from the 10 Briefings published by the Unemployment Unit and Youthaid[12]. The first figures available start from July 1998 (after the nationwide launch of New Deal in April). The figures produced from January 1998, (the implementation of New Deal in the Pathfinder areas) are only broken down into the four options, ie Employment, FTET, Voluntary and EFT, and not any breakdown of destinations.

  These statistics need to be interpreted with care as the total starts and categorized leavers figures are reported cumulatively whereas the numbers in the options are reported as the number of people in that state in that month. These figures are net figures as they represent the stock of those in an option in that month, ie it subtracts the flow of new leavers but adds the flows of new entrants[13].

  Figure 1 and Figure 2 show the cumulative numbers and percentages on options, respectively, month by month since January 1998. Bearing in mind that the New Deal scheme did not go national until April 1998, it is interesting to note that the numbers on the options were almost equal from January to April 1998 and then there was a rise in the numbers in the Employed Option in June 1998. In September 1998 there was an increase from 41 per cent in the FTET option to 70 per cent. This is explained by September being the beginning of the academic year, and New Deal is not tied to this. FTET places were available on courses which could not accommodate a staggered entry throughout the year. At the beginning of the New Deal national launch 37 per cent had entered the FTET option with 40 per cent in the Employment option and 12 per cent and 11 per cent in the Voluntary and Environmental option respectively. The Employed option never again reached its peak of 40 per cent and it dropped to a low of 12 per cent in September 1998 when FTET was at its highest (when the autumn term for colleges begins) and then levelled off at 23 per cent and never going above 27 per cent. The FTET option on the other hand has become fairly stable at about 42 per cent, with its lowest at 38 per cent in July 1999 (when the academic year traditionally ends and most courses complete).

  Figure 3 and Figure 4 show the cumulative and percentage outflows from New Deal by destination, month by month. There was an increase of 14 per cent more leavers in October 1998, compared to a fairly steady increase of 2 per cent over the year. Of the leavers in October 1998, 55 per cent went to an unsubsidised job, 7 per cent went to known destinations and 28 per cent went to unknown destinations and 11 per cent to other benefits. In October 1998, 21 per cent more people were getting jobs outside the New Deal scheme and 20 per cent more were "disappearing" to unknown destinations. However, a survey of unknown leavers suggests that around half the unknown leavers were unemployed. Many of these had been picked up by the system and returned to New Deal. More than a quarter of the sample were working, around 5 per cent were in training, and the remainder were in "other", which includes those caring for families, sick, pregnant, doing voluntary work or in prison. Essentially, the one thing which is known about the leavers from New Deal who have left to "unknown destinations" is that they have not started to claim JSA within 13 weeks of leaving New Deal as they would have been picked up by the system and returned to New Deal. However, throughout the year from November 1998 to October 1999 there was a steady flow of starters and leavers with the exception of June 1999 when there was a higher increase in leavers and July 1999 when there was a higher increase in starters. This could be attributed to a seasonal variation including holidays and the traditional end to the college year for training and educational courses.

  Figure 5 shows, by month, the numbers of starters and leavers to the New Deal programme since the national launch in April 1998, Apart from the increase in leavers as described in Figure 4, the numbers have maintained a steady flow and output with there being more starters than leavers as described in Figure 4, the numbers have maintained a steady flow and output with there being more starters than leavers except in the months of October 1998, June 1999, August 1999, September 1999 and October 1999.

  The DfEE has claimed that over 169,000 young people have now found a job through New Deal. The published figures available suggest that around 120,000 of these are unsubsidised jobs. The government's pledge of helping 250,000 into a job is well on its way to being met. However, these figures include both sustained jobs and moves to jobs that have not been sustained. Out of the 169,000 young people who have left New Deal for jobs, nearly 43,000 have returned to claiming benefit within 13 weeks. In other words more than a quarter of those who got jobs became unemployed again within three months. These are the people who have once more entered the "revolving door" of benefits and short-term jobs.


    "I think evaluating these schemes is virtually impossible actually because of the fact there are so many other things going on in the economy at the same time that are affecting these market . . . I do not think it is at all easy to evaluate" (Professor Patrick Minford, Education and Employment Committee, Eighth Report).

  Until the scheme has been in operation for some time it is difficult to predict how it will operate. Critics claim that the most promising recruits to the scheme will be hand picked by the employers and self select themselves into the full time education and training option. This will leave the hard core of the less able and disaffected to opt for the Environmental and Voluntary Task forces. Such a hierarchy of destinations within the New Deal could then be divisive and mitigate against the possibility of overall success.

  What has induced the government to believe that the New Deal will be more successful than its predecessor schemes of TOPS, YOPS, YTS, YT and ET? The answer is that the scheme is slightly different and cannot be compared very easily to its predecessors. The first essential difference is that the New Deal is aimed at an older age group, i.e. 18-24, whereas most of the earlier schemes were for those aged 16-19. The second is that it is the long-term unemployed who are being targeted rather than those who have just left school. Clearly the rapidly rising participation rate in the 16-19 age group has left the government less concerned about the welfare of this age group and hence it has turned its attention to the long term unemployed who are slightly older.

  A third major difference between this scheme and some of its earlier predecessors is that it has the important element of compulsion to participate if future receipt of benefits are to be guaranteed. Rights and responsibilities must be accepted. Options are given and choices must be made. The option of not participating is not available as benefits will be cut if individuals do not take up one of the four options available to them.

  In this sense the scheme has an important similarity with the Restart[14] scheme and the Workfare schemes which have operated in the USA. There is evidence that such a mandatory element to a publicly funded system will create incentives to participate in a way which would be impossible to induce in a voluntary scheme. People will have to go back into the labour market and take the jobs which are available. However, the certainty of high participation does not ensure that the outcomes of the scheme will be any more valuable then previous schemes which have had voluntary participation.

  Critics of the New Deal have been quick to point out that unemployment in the UK today is falling rapidly and that there is no need to spend such vast sums of money on this scheme. When (as of January 1998) there are only 120,000 unemployed young people in the target age group this argument has a certain force.

  In common with many of the schemes introduced by previous governments it is also contended that this scheme will not create "real jobs" in the sense that the people who get offered placements will only be displacing older workers who do not come with such a subsidy from the government. The substitution and dead-weight loss effects of these schemes are hard to measure but one cannot help notice the large number of opportunities which have already been created by the large retail chains who employ large numbers of young people and casual labour to do fairly menial tasks. It is hard to argue that these employers will not just be taking the subsidy and cheating the system by not offering permanent jobs at the end of the statutory New Deal period. In this sense the concept of "welfare to work" is difficult to implement since it does not, in practice, provide any guarantee of actually getting a job.

  Another difference with the New Deal programme and that of past training schemes is the advertising campaign that led up to its launch. This included general media publicity such as television advertisements with employers backing the scheme and newly employed trainees taking part in the advertising campaign. The slogan "As more sign up, fewer sign on" was an attempt to alert the public to the benefits to everyone of the success of the New Deal policy. There was also a considerable amount of press coverage and marketing done by local Employment offices in an effort to portray the New Deal as something different to try to encourage as many people as possible to give it their positive support. According to David Blunkett, Secretary of State for Employment and Education "very few people have not seen the advertising campaign for the New Deal . . . It's working."

  A final crucial difference with this Welfare-to-Work scheme was the political climate at the time of its introduction. Most specifically there are elements of a crusading programme which was to be available to all and an opportunity to solve the problems of unemployment and was reflected in the Chancellor of the Exchequer's speech at the launch of the New Deal Pathfinder in January 1998:

    "I believe this is a national crusade . . . The government believes in helping thousands more from welfare into work in the years to come as part of the long haul towards our goal of full employment."

  At the national event launch in April 1998, Andrew Smith, Minister for Employment also suggested:

    "To succeed, New Deal has to be a national crusade, pulling together employers, trade unions, community groups and young people themselves. By working together, we will open up new hope, new opportunities for young people."

  With a change of government after a period of 17 years, the New Deal training scheme was optimistically viewed as the vehicle to overcome the problems of youth and long-term unemployment. At a time of major political change the New Deal policy was seen as a radical departure in labour market policy.


  The Measurement of the outcome of New Deal is not straightforward. Evaluating ALMP at the level of the individual in a non-experimental context is beset with problems (see Blundell and Dias (1998) for a summary). Most obviously, in individual cross section data, each individual can only be observed in one state, i.e. as a recipient of ALMP or not. The inability to observe the counterfactual when allocation to treatment is not random means evaluation at the individual level is therefore problematic. What we suggest in this paper is a method of aggregation to the level of the small geographical areas which administer the New Deal and use the variation in the administering proportion of different utilisation of the four New Deal options by these Units of Delivery (UoD) to capture the marginal effects of the different possible configurations of the use of the alternative options involved in New Deal. The data used is taken from the government's Department for Education and Employment (DfEE) website Core Performance Tables relating to the New Deal. The data refers to all 144 Units of Delivery in the country and relates either to the cohort who began the New Deal between April and June 1998 or the cohort who began New Deal between July and September 1998.

  Analysis and evaluation of unemployment measures and ALMP based on local, regional or state data is not new (see Barnow (1992), Friedlander and Gueron (1992), Coles and Smith (1996) and Gregg and Wadsworth (1996)) but it offers certain advantages. In particular such an assessment will enable us to get an estimate of the possible effects of each option on the treated by exploiting the different balance of treatments available in different UoD. This type of evaluation strategy is a particular type of assessment which uses a "pure" local average treatment effect[15] (see Imbens and Angrist (1996)) where the treatment itself can now be thought of as a vector of possible choices of the different option utilisation rates across the different UoD. As the government changes New Deal funding and the external local and national labour market conditions change as New Deal develops then the UoDs will choose different levels of involvement in the options accordingly. Hence the form of the treatment (different allocations to the four options) is chosen by the recipients (UoDs) as the government changes the financial and other conditions of the treatment (New Deal enforcement).

  There are a variety of ways of potentially measuring the impact of New Deal at the level of the Unit of Delivery. We explore just three such measures and statistically assess the explanatory power of the available regressors. These measures are: the number of people entering unsubsidised jobs, the number entering subsidised jobs and the number not returning to Job Seeker's Allowance within 13 weeks of entering a job. These three variables report slightly different measures of New Deal. The numbers into subsidised jobs refers to the number of jobs created by the local economy without any government financial involvement. The subsidised jobs number still incorporates an element of New Deal central funding and is therefore a less satisfactory outcome measure of New Deal. The final measure of those not returning to JSA is different again as it measures those not returning to registered unemployment. This could mean that they are leaving New Deal for full time education or training or leaving the labour market altogether.

  The range of possible theoretical models to explain the pattern of these variables by UoD is limited. We take as our stimulus the work from the theoretical matching and job creation and destruction literature (see Mortensen and Pissarides (1994), and Davies et al (1997)) and add to this the growing influence of the assessment of ALMP in this type of model (see Calmfors and Lang (1995), Bellman and Jackman (1996) and Richardson (1999)). The empirical investigations from this literature have used the number of matches of job seekers to vacancies as the dependent variable and regressed its log on the log of the number of vacancies and the log of the unemployment level. Various other controls are added to these regressors to account for observational heterogeneity. Our three candidates of the dependent variable can all be considered as detailed matching variables of varying quality. We adapt this model (see Coles and Smith (1996)) to the data available by UoD sites during the New Deal period. Our estimation model is

  The random effects Panel Instrumental Variables (IV) regression results are presented in Table 3. These results are IV estimations since the size of the New Deal Cohort and the Unit Cost variables must be treated as endogenous and are therefore instrumented by various other exogenous characteristics of the UoD. (Details of the instrumenting are provided in Dolton (2000)). The summary statistics for these regressions all justify the random effects specification in the sense of the significant r coefficient which represents the proportion of the variance in the dependent variable due to the unobserved heterogeneity term. In each equation the overall R2 indicates the justification for the random effects model over the simpler alternatives. These results present quite a distinct picture of the effect of New Deal which vary according to the dependent variable which is adopted. If we consider the results for the equation relating to the number of New Deal leavers to unsubsidised jobs or the number not returning to JSA within 13 weeks the results are very similar. These results are tabulated in the first and third columns of coefficients of Table 3. The results suggest that there is a "size of unemployed pool" effect in that the number of unemployed people in the UoD is positively related to the outcome. This result is found with all performance measures and simply captures the size effect that bigger UoDs (in terms of the number of New Dealers) have a larger potential number of matches or placements. We also find that the relative size of both long-term unemployment and youth unemployment also negatively influence the level of unsubsidised jobs and those lasting more than 13 weeks which are provided in the UoD. Intuitively this finding is plausible in the sense that the higher is the level of long-term unemployment and the higher is the level of youth unemployment the lower will be the level of permanent jobs provided.

  In addition the average amount spent by government at each UoD is positively related to the outcome. The results imply that there are no economies of scale in the provision of New Deal places. This suggests that the higher per capita expenditure on the programme the larger will be the number of private sector (unsubsidised) jobs which are created. This suggests that public sector funds are not "crowding out" non-government expenditure by the private sector in terms of the provision of jobs.

  The level of unsubsidised jobs is heavily influenced by regional demand side factors since the higher are regional earnings are the higher will be the number of unsubsidised jobs provided. In addition if the labour market is tight[16] then there is buoyant enough demand to increase the number of unsubsidised jobs. The good news for policy makers who advocated the New Deal, in terms of the creation of unsubsidised jobs, is in the significant coefficient on the relative size of the employment option. However, the coefficients on the relative size of the other options are all statistically insignificant—indicating that it is only the size of the employment option which matters in the determination of the level of unsubsidised jobs. The first column of Table 3 also shows that a higher proportion of men and a lower proportion of disabled are recipients of unsubsidised jobs.

  The second column of Table 3 presents the regression coefficients for the number of leavers of New Deal to subsidised jobs. This exit is clearly distinct from that to unsubsidised jobs as the employer is not bearing the full cost of the employment and there is less likelihood that the job will have a long tenure. The numbers of these leavers is highly correlated with the numbers on the Employment Option which is understandable since this is consistent with the New Deal placement being extended under a wage subsidy. The coefficient on the per cent Disabled term indicates that higher levels of take-up of job subsidies are consistent with a higher level of employment of registered disabled people. This may be because the subsidy reduces the risk to the employer of "trying out" a disabled person in a job in a temporary capacity. It also presents a convenient way of employers satisfying equal opportunity policies with respect to quotas of disabled people. In contrast the proportion of people of New Deal recipients who are from the ethnic minorities is negatively correlated with the number of subsidised jobs which are provided. This could be a result of either discriminatory practices on the part of subsidised job providers or it could be that people from the ethnic minorities are simply not accepting these openings. It is impossible to distinguish between these two competing explanations.

  The policy implications of our results must be tempered by the limitations of the estimation techniques used. However, taken at face value they are clear and of limited comfort to New Deal advocates. Firstly, on the positive side it would appear that the level of investment by the government in terms of the higher unit cost of providing New Deal places yields a return in terms of the number of unsubsidised or (semi) permanent jobs which are created. The other side of this coin is that there would appear to be little scope for economies of scale in this provision. The major positive funding is that the relative scale of the Employment option is highly positively significant in the provision of either unsubsidised or (semi) permanent jobs. This is of importance in that it suggests the focus of New Deal should be to get young people into job placements and it is only this option which really matters. This is of course bad news for the hopes which were placed in the value of the provision of placements into Voluntary Service and the EFT. More obviously it would appear that demand side factors and the structure of the unemployment in a UoD have a large effect on the potential for job creation in the New Deal scheme. This is reflected in the importance of the relative scale of youth and long-term unemployment in the determination of the effectiveness of the New Deal in creating jobs in the UoD.


  The assessment of training programmes and the policy implications of different training arrangements are very difficult to appraise rigorously. Do mass schemes like the New Deal and YT achieve much more than simply push down the unemployment figures and possibly lower relative youth wages? If the state intervenes it should provide good quality training which constitutes a rapid investment in the human capital of our young people. Throughout much of the debate about the role of state training schemes the focus has been the assessment of the effects of the supply side response to the introduction of a government training scheme. No mention has been made of the macro-economic demand side conditions which may impinge on the impact of a training programme. These aggregate macroeconomic conditions will dramatically affect the employment and earnings prospects of New Deal trainees as they attempt to enter the labour market.

  Notwithstanding the economic arguments we must remember that one of the main motivations for instituting national training schemes may be political.[17] The "get tough" policy in which the continued receipt of the Job Seeker's Allowance is dependent on participation in the New Deal could be attractive to floating voters. There may also political value in getting young people off the unemployment register with the added probability of the training being valuable and securing a future job.

  Our limited review of the available aggregate monthly data relating to inflow to and outflow from New Deal provides an insight into the complex pattern of entry into the different options and exit to the available destination states. The figures presented indicate the sheer scale of the New Deal programme and provide an insight into the complexity of any attempted evaluation.

  The last section of the paper analysed the matching process out of New Deal at the level of UoD. The analysis confirmed the importance of demand side factors in the impact of New Deal on matching and suggested only a limited role for the importance of the relative size of the Employment Option. There is a suggestion that the level of NVQ level 2 attainment could be important in the exit rate to both unsubsidised jobs and jobs which last more than 13 weeks.


  Barnow B (1992) "The effects of performance standards on state and local programs", Chapter 8 in "Evaluating Welfare and Training Programs", Manski C and Garfinkel I (eds), Cambridge, Harvard University Press.

  Bellman L and Jackman J (1996) "Aggregate impact analysis", Chapter 5 in International Handbook of Labour Market Policy and Evaluation, Schmid G, O'Reilly J and Schomann K (eds), Cheltenham, Edward Elgar.

  Blundell R and Dias M (1998) "Evaluation methods for non-experimental data", mimeo, Institute of Fiscal Studies.

  Calmfors L and Lang H (1995) "Macroeconomic effects of active labour market programmes in a union wage-setting model", Economic Journal, vol 105, pp 601-19.

  Coles M and Smith E (1996) "Cross section estimation of the matching function: Evidence from England and Wales", Economica, vol 63, pp 589-97.

  Danish Ministry of Labour (1993) Measuring Labour Market measures: Evaluating the Effects of Active Labour Market Policy Initiatives, Ministry of Labour, Denmark.

  Davies S, Haltiwanger J and Schub S (1997) Job Creation and Job Destruction, Cambridge, MIT Press.

  Dolton P (2000) "Does New Deal Match Up?" University of Newcastle, mimeo.

  Dolton P and O'Neill D (1996a) "Unemployment duration and the Restart effect: Some experimental evidence ", Economic Journal, vol 106, pp 387-400.

  Dolton P and O'Neill D (1996b) "The Restart effect and the return to full time stable employment", The Journal of the Royal Statistical Society A, vol 159, pp 275-88.

  Dolton P (1993) "The economics of youth training in Britain", Economic Journal, vol 103, pp 1261-78.

  Friedlander D and Gueron J (1992) "Are high-cost services more effective than low-cost services?" Chapter 4 in Evaluating Welfare and Training Programs, Manski C and Garfinkel I (eds) Cambridge, Harvard University Press.

  Gardiner K (1997) Bridges from Benefit to Work: A Review, York, Joseph Rowntree Foundation.

  Gregg P and Wadsworth J (1996) "How effective are state employment agencies? Jobcentre use and job matching in Britain" Oxford Bulletin of Economics and Statistics, vol 58, pp 443-67.

  Grubb D (1994) "Direct and indirect effects of active labour market policies in OECD countries", Chapter 4 in The UK Labour Market Barrel R (ed) Cambridge, Cambridge University Press.

  Haskel J and Jackman R (1988) "Long-term unemployment in Britain and the effects of the Community Programme", Oxford Bulletin of Economics and Statistics, vol 50, pp 379-404.

  Heckman J, Lalonde R and Smith J (1999) "The economics and econometrics of active labour market programs" Handbook of Labour Economics, Volume 4, Ashenfelter O and Card D (eds), North Holland.

  HMSO (1998) The New Deal/The New Deal Pathfinders Eighth Report by the Education and Employment Committee, House of Commons.

  Imbens G and Angrist J (1996) "The identification of local average treatment effects", Econometrica.

  King D (1995) Actively Seeking Work?, Chicago, Chicago University Press. Lancaster T and Chesher A (1981) "Stock and flow sampling", Economics Letters, vol 8, pp 63-5.

  Layard R, Jackman R and Nickell S (1991) Unemployment, Oxford, Oxford University Press.

  Mortensen D and Pissarides C (1994) "Job Creation and job destruction in the theory of unemployment", Review of Economic Studies, vol 61, pp 397-415.

  Richardson J (1999) "Targeted wage subsidies and long-term unemployment: Theory and policy evaluation", LSE PhD Thesis.

  Riddell C (1991) "Evaluation of manpower training programmes: The North American experience", in Evaluating Labour Market and Social programmes OECD.

  Zweimuller J and Winter-Ebmer R (1996) "Manpower training programmes and employment stability", Economica, vol 63, pp 113-30.

3   This research was funded by the ESRC Grant number L134301001126. Back

4   See Dolton (1993) for details. Back

5   National Vocational Qualifications (NVQs) are nationally recognised qualifications for work which people actually do in their jobs. They are not examinations or tests; they prove that you can carry out particular work-based activities to a national standard. These standards have been set by awarding bodies, employers and educationalists. Back

6   NVQ Level 2 equal in equivalence to four GCSEs (General Certificate of Secondary Education). These are the exams taken by most 16 year olds in secondary schools, at the end of compulsory schooling. Back

7   NVQ Level 3 in equivalence to two A Levels. Normal university entrance would be three A Levels. It is therefore not quite equivalent to university entrance qualifications. Back

8   Gardiner (1997) provides unit cost details of a large number of state ALMP in the UK which permits basic comparative analysis. Back

9   All qualifications must be validated by the Qualification and Curriculum Authority (QCA). Back

10   Qualifications are funded by the Further Education Funding Council (FEFC) and must therefore be approved by them. Back

11   Not all figures were published in the same format and for this reason the figures for June 1999 were taken from the Performance Tables published by the Unemployment Unit and Youthaid. These Performance figures were not available for every month and this creates difficulties when comparing month by month. Back

12   In order to make effective analysis of these figures the following had to be averaged: November 1998-to Unsubsidised jobs and Other Known; December 1998-Other Known; April 1999-left New Deal to Unsubsidised jobs, Unknown; May 1999-Unknown; June 1999-left New Deal, to Unsubsidised jobs, other known, Unknown. Back

13   Care is required in distinguishing between stock and flow samples (see Lancaster and Chester (1981)). Back

14   See Dolton and O'Neill (1996a and 1996b). Back

15   Unfortunately we cannot observe vacancy rates at the level of the UoD so we use vacancy outflow data at the regional level to approximate our missing ideal vacancy data. As a consequence it is not possible to make the usual inferences about whether there is constant returns to scale in the matching function from the estimated coefficients. This is not however the focus of this paper. Back

16   The tight labour market variable is one retrieved from the government website. Its measure is wider than unemployment since it takes account of the urban or rural nature of the UoD and its demographic structure. Back

17   There is a long history of the political character of ALMP. See King (1995) for a comparative review in the UK and the USA. 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 20 March 2001