UK Regional Policy: Does it Measure Up?
Abstract

There has been much interest in regional issues in recent years, but the regional disparities in unemployment continue to exist, and have worsened over the last decade, while expenditure on regional policy is falling and it is set to fall much further.  The paper reviews UK regional policy, focusing on the recent evidence for the employment effect and effectiveness of this policy.  It is found that regional policy created between 8,700 to 18,700 jobs per annum over the 1990s, but that this is small relative to the scale of the unemployment problem.  It is argued that the policy is an effective job-creation measure, so that some expansion is desirable. One way to do this is to reintroduce a distributional component to make the grants available for mobile UK-owned plants in the same way they are available to foreign direct investment.  Without this it is argued that current ‘regional policy’ is little more than a national policy applied at the level of the region.

JEL Classification Numbers

O2 (development planning and policy), H2 (taxation, subsidies and revenue) and R0 (urban, rural and regional economics: general).

Acknowledgements

The paper has benefited from presentations at the Alliance for Regional Aid Conference, Royal Institute for British Architects, London, April 2002; at a Department of Trade and Industry / Regional Studies Association seminar on Regional Investment, DTI, London, September 2003; and at an ESRC Urban and Regional Economic Seminar on Regional Policy at the London School of Economics, London, October 2003. The author gratefully acknowledges the comments of two anonymous referees, and that of one of the editors, Alissa Goodman.  Thanks also go to Owen Parker and Georgina Green of the Department of Trade and Industry for clarification on points of detail, but responsibility for the paper rests with the author.

Address for correspondence: Dr Colin Wren, as above.  Email: c.m.wren@ncl.ac.uk.

UK Regional Policy: Does it Measure Up?

Colin Wren

1:    Introduction

There has been much interest in UK regional issues in recent years, which has reflected several concerns.   The first is the belief that regional economic performance is vital to national growth, and that economic development policy is best delivered at the local level.  For this reason, many economic policy initiatives have been decentralised with the establishment of the devolved administrations and Regional Development Agencies in England.  The other concern is that despite strong national growth over the last decade the disparities in regional economic wellbeing have not closed, and may have widened.  The Government’s preferred measure of regional economic performance is per capita GDP, reflecting a focus of policy on productivity as a ‘key driver’ of national and regional economic growth (HM Treasury 2001).  However, while there are substantial and persistent differences in GDP between regions, it does not tell the full story.   This is because per capita GDP is calculated for regions as a whole and it masks the inequalities that occur within regions.  Other measures of the ‘regional problem’ capture some notion of the intra-regional inequality, of which there are several possibilities, including household poverty, the activity rate and unemployment rate (see Adams and Robinson, 2002).  The unemployment rate is the most widely used measure, which is highly correlated with other indicators of social deprivation (Armstrong and Taylor, 2000).  It makes recognition of the fact that the ‘regional problem’ is more than just a purely economic phenomenon.

The UK unemployment rate is shown in Figure 1 for the years 1960-2003, along with the dispersion of regional unemployment rates, and the difference in these between the ‘worst’ and best’ performing regions of the UK: the North East and South East of England.1  It shows the strong negative relationship between the national unemployment rate and the dispersion of regional unemployment rates (the correlation coefficient is - 0.81).  It also shows that while the difference in unemployment rates between the North East and South-East England follows the national unemployment rate, this relationship (which has existed since at least the 1920s) broke down in the recession of the early 1990s (Taylor and Bradley, 1994).  An explanation for this is that the housing-market crash of the early 1990s much more adversely affected the southern areas of England.  Since then, Figure 1 shows that the unemployment rate of the North East has increased relative to the South East, but that the relationship now seems to have returned to its long-run position.2  Overall, the ‘regional problem’ continues to exist, although with a possibly improving position since the year 2000.  Nevertheless, the position is similar to that of much of the 1960s and 1970s when regional policy was at its height.3

The programmes for regional development are broad, and as well as the activities of the devolved administrations and the Regional Development Agencies, they include the European Union Structural Funds, and initiatives in England such as the New Deal for Communities, the National Strategy for Neighbourhood Renewal and the Rural Development Programme (HM Treasury, 2003).   On a broader definition, they include regional aspects to ‘mainstream’ public expenditure, such as health and education.  However, regional policy is usually defined more narrowly than this, as measures to promote growth and employment in lagging areas (Regional Studies Association, 2001).  Regional policy has taken many forms (see McCrone, 1969), but a consistent feature over the last forty years are the investment grants available to firms for the purpose of job creation in the designated Assisted Areas.  The policy has undergone substantial changes over the years and further revisions are being planned.  These changes include an on-going reduction in expenditure; an abolition of the support for job transfers; a tightening of the rationale for grants that make it subject to a ‘market-failure test’; and a shift in support towards small firms and large projects.4   Further, the National Audit Office argues that regional grants are an ineffective job-creation measure (NAO, 2003).  Given all this, and against a background of a continuing ‘regional problem’, it seems reasonable to ask whether regional policy really measures up to the scale of the regional problem?

This paper addresses this question by focusing on current UK regional industrial policy. It considers the nature of regional policy over recent decades, and assesses the evidence on the employment effect and cost-effectiveness of this policy, considering it relative to the size of regional unemployment.  The mainstay of UK regional policy is Regional Selective Assistance, which is a discretionary assistance scheme, and the paper considers the changes to this scheme that have altered the nature of the policy.  It reviews the evidence contained in the Government evaluations, econometric studies and the National Audit Office report. It is argued that regional policy has a relatively small employment effect and that expenditure is small compared to that elsewhere, so that some expansion should be contemplated.  This can be achieved in a number of ways, including the introduction of a distributional component to make assistance available to domestic mobile investment on the same basis as foreign-owned firms.   Indeed, as currently constituted, it is argued that what is called ‘regional policy’ is not really a regional policy at all, but national policy that seeks to benefit all regions.

The paper is organised as follows. Section 2 examines the scale and composition of UK regional policy expenditure since 1960.   It is argued that the poor performance of UK regional policy in earlier decades was related to its automatic nature.  Section 3 describes the nature of Regional Selective Assistance, and the changes to this that have altered the nature of regional policy, while section 4 examines the potential advantages of discretionary assistance.  Section 5 considers if regional policy measures up to the scale of the regional unemployment problem, and makes recommendations for improvement.   Conclusions are drawn in section 6.

2:    UK Regional Industrial Policy

Expenditure on UK regional industrial assistance between 1960 and 2002 is shown in Figure 2.  This is taken from Wren (1996), but extended forward in time.  It gives expenditure on all forms of regional industrial assistance, including loans and investment tax allowances, expressing as ‘grant equivalents’.  This is the grant amount that if received at the same time as the subsidy has equivalent value to the firm in net present value terms.  Where a similar national measure exists the differential subsidy in favour of the Assisted Areas is calculated.  A three-fold breakdown of expenditure is given according to discretionary assistance, automatic investment support and employment premiums.  Discretionary assistance determines the grant rate as part of the grant-giving process, which is zero where the assistance is refused, whereas automatic support determines the grant rate automatically according to the published eligibility criteria.  The employment premiums directly subsidised labour, whereas the other categories mainly supported capital investment.5  A brief account of the scale and composition of regional policy is now given.  The rationale for many of these changes was the poor employment cost-effectiveness of UK regional policy, so that this evidence is also reviewed.

Pattern and Composition of Expenditure

In total, Figure 2 identifies £32 billion in regional assistance over 1960-2002, of which £18.8bn (59%) is on automatic investment support, £7.8bn (24%) is on employment premiums and £5.4bn (17%) is on discretionary assistance (all at 1995 prices).  At its peak regional policy assistance expenditure was £2,200 million in 1975/6 (1995 prices), but Figure 2 shows that it is now running at less than £200m per annum. Regional assistance consisted only of discretionary assistance in the early 1960s, but Figure 2 shows that spending built-up from this time with the introduction of the automatic investment support and employment premiums. 

In the case of the automatic investment support, the 1963 Budget introduced regionally-differentiated tax allowances for the first time (prior to this time discretionary loans and grants amounted to no more than £5m in total). These allowances were replaced by Investment Grants in 1966, and in one form or another the automatic investment support increased up to 1982, but with four dips in spending.  First, the introduction of Corporation Tax in 1965 diminished the real value of the investment allowances, leading to the introduction of the Investment Grants to restore the real value of the subsidy.  Second, the Investment Grants were viewed as expensive and replaced by regionally differentiated first-year writing-down allowances in 1970, although this was reversed by the introduction of Regional Development Grants (RDG) in 1972.  Third, spending fell in 1977 when the mining and construction activities were excluded from RDG to curb public expenditure.  Finally, spending on automatic investment support fell dramatically with the onset of recession in 1979, due to a depressed level of investment and take-up. 

The origin of the employment premiums was the Selective Employment Tax in 1966.  This surcharged employers’ National Insurance contributions to raise revenue for an expansion of public services, but to effect a structural change in the economy towards the industries with an export content it was refunded to those firms outside of the construction and service sectors.  In addition, a Selective Employment Premium (SEP) was paid to firms in these sectors, but it was withdrawn for firms outside of the Assisted Areas in the 1967 devaluation.  It coincided with the introduction of the Regional Employment Premium (REP), which was also payable at fixed amounts in respect of those employees eligible for SEP.6  The combined subsidy was put at 8 per cent of wages (McCrone, 1969).  In real terms, expenditure on the premiums decreased in every year after their introduction, but they were doubled in 1974 to restore the real value of the subsidy.  The premiums were cut in the first sterling crisis of 1976, and the subsidy was withdrawn altogether in the public expenditure cuts of December 1976.

Regional policy expenditure peaked in 1976, but the real turning point occurred in the early 1980s.  Revisions at this time included a phased reduction in the geographical coverage of the Assisted Areas, a reduction in the RDG grant rate and a four-month deferral on RDG payments.  The RDG was effectively scrapped in 1984, followed by a revised RDG scheme that related the grant amount to the number of jobs in projects backed by capital investment.  It was superseded by the Regional Enterprise Grant scheme from 1988 to 1997, which restricted the assistance to small and innovative projects.  Since the early 1980s, discretionary assistance has taken an increasing share of regional policy expenditure, and as Figure 2 shows virtually all assistance was in this form over the 1990s.  The mainstay of regional policy in this period is Regional Selective Assistance (RSA), which was introduced at the same time as RDG in 1972. RSA is made on a discretionary basis towards capital investment in projects that either create or retain employment in the Assisted Areas.  In 2000, smaller projects became ineligible for RSA, and these are now funded under an Enterprise Grant Scheme, which is available to small and medium-sized enterprises both inside and outside the Assisted Areas.  Annual expenditure on RSA is £200m per annum in Great Britain, and a further £12m is spent on the Enterprise Grant Scheme in England, but as we see below spending is set to fall much further.

Employment Cost-Effectiveness

The objectives of UK regional policy have been couched in many ways over the years, but the policy has always been judged in terms of its ability to create jobs.  This is the thrust of all Government evaluations and the recent National Audit Office report (NAO, 2003).  At the time that expenditure was cut in the 1980s a White Paper put the cost of a regional policy job at about £35,000 (DTI, 1983), which at today’s prices can be roughly doubled.  Undoubtedly, the high ‘cost per job’ was a factor in the scrapping of the automatic grants in the 1980s, and it explains the subsequent orientation of policy.  It is intriguing to note that employment cost-effectiveness was an important concern of early regional policy.  The 1960 Local Employment Act required the Board of Trade to pay attention to “the relationship between the expenditure involved and the employment likely to be provided”.  Treasury approval was required for cases where the ‘grant per job’ exceeded £1,000, which is about £10,000 at today’s prices.  However, the job link was dropped when the automatic Investment Grants were introduced in 1966, and it was only reintroduced after the RDG scheme had been terminated in the mid-1980s, so that throughout most of the experience of regional policy there was no direct link with job creation.  Nevertheless, it is important to understand the reasons why the policy was ineffective, and that this primarily related to the automatic nature of the assistance.

The automatic nature of much of regional assistance meant that a substantial proportion of it was taken up was taken up by firms in highly capital-intensive industries.  These firms were often using the grants to fund capital restructuring, creating few new jobs and possibly destroying them.  At the time, little was known about the operation of regional policy (HC 651, 1972 and HC 85, 1973), but when it emerged the evidence suggested very high grants in some capital-intensive projects, which could be as high as £900,000 per job (current prices).  Moore et al (1986) applied shift-share analysis to regional employment data and found a ‘cost per regional policy job’ of £47,750 over 1966-76 (1982 prices). However, when disaggregated, Table 1 shows the policy was ineffective in the capital-intensive industries.  These industries received £4,717m in regional assistance between 1966 and 1976, but according to Moore et al they created only 6,000 net new jobs, at an average cost of £786,000 each (1982 prices).

Table 1 shows that regional policy was more effective outside of the capital-intensive industries, with a ‘cost per job’ of just  £26,000.  This may still seem high, but it is primarily because the shift-share methodology gives an unfavourable measure of the ‘cost per job’. It is based on all those jobs created over 1966-76 but existing at 1976, so that it ignores the jobs that were created and lost by 1976.  Firm-based estimates that measure the jobs three years after the receipt of assistance give lower figures.  Wren (1994) obtains estimates of the ‘cost per job’ for regional incentives over 1975-81 of around £10,000 for large firms, and at about half this in the case of small firms (1985 prices).  The sample used excludes very large capital-intensive firms, and the estimates are similar to those found for other job-creation schemes.  Thus, the primary reason for the failure of regional policy over the 1960s and 1970s was because it was automatic in nature, and not directly related to job creation.  Similar problems have been found for the automatic investment incentives at the national level.7  Since this time automatic subsidies have been largely abandoned, while regional policy has focused on discretionary assistance in the form of Regional Selective Assistance, which became the main instrument over the 1990s.

3:    Regional Selective Assistance

       Regional Selective Assistance (RSA) was introduced under the 1972 Industry Act, and it has taken an increasing share of regional policy assistance since the early 1980s, so that it is now the mainstay of regional policy.  The advantage of discretionary assistance is that it can be tied directly to job creation, and refused to those projects that do not satisfy ex ante grant-per-job conditions (Swales, 1997).  It satisfies the requirement of the European Union regulations on state aids that projects must be supported by investment in fixed capital assets (CEC, 1998).  As well as capital investment and job creation, RSA is associated with a large number of other criteria that define the eligibility for the scheme.  These are important, as not only have they been progressively tightened over time, but they reveal the changing nature of regional policy itself.8   As such, they are worth considering in some detail.

       The guidelines for RSA initially placed qualifying projects into two categories: new projects and expansions that created additional employment (Category A); and projects that did not provide extra jobs but which maintained or safeguarded existing employment (Category B).  The assistance was related to the number of jobs, and it was given as a cheap loan, interest-relief grant or removal grant under Category A, but a loan at commercial rates under Category B.  In either case, the applicant had to demonstrate the viability of the firm and that the greater part of the project cost would be met from sources outside of the public sector.  In the case of Category B the assistance was offered where it could not be reasonably obtained from other sources, but no such condition operated in the case of job-creation projects.9

A review of industrial aid in July 1979 introduced two new important criteria for RSA.  The first was a ‘Proof of Need’ condition.  This required that the provision of assistance should lead to a significant change in the nature or scale of a project, or a significant advance in timing or a desirable change in location.  A similar condition had existed for Category B projects, but these cases were small in number and it now applied to all projects.  It reintroduced a condition under the 1945 Distribution of Industry Act that assistance could only be provided if the funds could not be obtained on requisite terms from elsewhere.10  It was dropped by the 1960 Local Employment Act, but it now applies to virtually every area of industrial support.  Second, a ‘Regional and National Benefit’ condition was introduced.  This required that the project should strengthen the regional and national economy, thereby providing more ‘productive and secure jobs’, such as through improved operating efficiency or new products.  To cut down on the administration costs, loans were given only in exceptional circumstances, so that after 1980 assistance was virtually always given in the form of a ‘project grant’, related to the fixed and working capital of the project, as well as the number of jobs.

More changes were made to RSA in 1984, at the time that RDG was abolished.  It was no longer payable to relocation projects where there was no net increase in jobs to the UK as a whole; a ‘claw-back’ clause was introduced; and an undisclosed ex ante grant per job limit was imposed.  The first of these was a major shift in regional policy, as plants that merely relocated within the UK were no longer eligible for assistance from this time. This was a break in policy, which for most of its existence had been about redistribution of industrial activity (see Ashcroft and Taylor, 1979).   The claw-back condition meant that the grant was repayable if the assisted jobs did not remain in place for at least eighteen months after the last asset had been provided.

The latest revised set of eligibility criteria was introduced in 2000, which represented a further major reorientation of RSA.  The current set of criteria is shown in Table 2 (it does not include the claw-back condition or ex ante grant per job limit, which do not define eligibility).  The first five conditions are essentially the same as before, but with changes in emphasis, while the other two eligibility criteria are new developments, which are likely to further reduce take-up.11  In the case of the ‘Quality’ condition the 1998 Competitiveness White Paper (DTI, 1998) focused RSA on ‘high-quality knowledge-based’ projects to improve regional competitiveness.  This was at the height of the ‘New Economy’, but there must be doubts about whether there are a sufficient of these knowledge-based projects in the regions (Regional Studies Association, 2001).  The ‘quality’ of projects is to be judged according to wages relative to the sector and region, the sustainability of employment and the levels of R&D and training (DTI, 2001).

The other new condition concerns the ‘Eligible Investment’ and means that RSA is now only available to projects where the capital expenditure exceeds £500,000.  For other projects there is an Enterprise Grant Scheme (EGS), which was introduced in January 2000.  This is a discretionary scheme for projects with up to £500,000 in capital investment, but it is available to small and medium-sized enterprises only.12   The EGS is available in Enterprise Grant Areas, which include the Assisted Areas (Tier 1 and 2 areas) and Tier 3 areas outside the Assisted Areas that are designated to include areas with particular social problems.13   It offers a capital grant up to a rate of 15% and a maximum award of £75,000 (and hence for capital investments up to £500,000), but at a reduced rate of 7.5% for medium-sized firms in Tier 3 areas.  RSA is available throughout the Assisted Areas, at grant rates of up to 35 per cent in Tier 1 areas and up to 20% in Tier 2 areas, but the minimum RSA grant amount is £75,000.  The changes mean that projects implemented by large firms in the Assisted Areas below £500,000 are no longer eligible for regional assistance.  The presumption must be that the larger firms can implement the smaller projects using their own resources.

The eligibility criteria for EGS are similar to those for RSA in Table 2, but with some important differences.  A Jobs condition does not explicitly rule out relocations and there is no Prior Commitment condition, while in each region the National and Regional Benefit is to be decided in consultation with the Regional Development Agency.  The Quality condition gives preference to ‘high-growth businesses seeking to maximise value-added projects with quality output’, which includes salary levels, training, R&D, supply chain improvements, innovation and environmental sustainability.  An important difference with RSA is that the EGS does not make it essential for a project to create jobs.  The scheme was initially administered by the Government Regional Offices, but it has since been transferred to the Small Business Service in each region, which are required to consult with the Regional Development Agency to decide on any additional scheme criteria to reflect local priorities.  However, spending on the scheme remains very small at only about £12 million per annum.

In the first financial year after the introduction of the EGS the number of applications in England for RSA fell by three-quarters.   Further, while expenditure on the schemes combined has held up - £103m was spent on RSA and EGS in England in 2002/03 compared with £112m on RSA in 2000/01 - the value of offers in England has fallen sharply from £205m for RSA in 2000/01 to £123m for the combined schemes in 2002/03.14   This indicates that regional policy expenditure is set to fall further.  However, the job condition on RSA suggests a more effective scheme than was the case for the automatic support, but before considering this evidence the potential benefits and drawbacks of discretionary assistance are first considered.

4:    The Advantages of Discretionary Assistance

The distinguishing feature of discretionary assistance is that it allows discretion to be exercised in the disbursement of assistance, so that in setting the assistance contract for each firm the terms of the contract can be varied in order to better fulfil the government’s objectives.  It differs from automatic assistance where a firm satisfying the published eligibility criteria is entitled to assistance on the specified terms, eg. at the published grant rate.  An advantage of discretionary assistance over automatic assistance stems from the fact that industrial assistance is a principal-agent relationship.  The government contracts a firm to carry out an investment project in return for assistance, but with asymmetric information over the nature of the firm and its behaviour.  This is problematic as the government and firm have differing objectives.  It gives discretionary assistance an advantage in dealing with the resulting adverse selection and moral hazard problems.  However, the drawback is that the government must appraise projects, which gives rise to errors in scrutiny.  We now examine three main advantages of discretionary assistance, and then consider the nature of these errors.  The discussion is based on an analysis of an industrial assistance scheme in Wren (2003a).  It is assumed that the assistance offers a proportionate job creation grant, so that attention focuses on the method of disbursement.

(i) Adverse Selection: When the government is uninformed over the nature of projects it is unable to distinguish those projects promising a relatively large number of jobs from those promising only a few jobs or destroying jobs.  The advantage of discretionary assistance is that the government can scrutinise projects in an effort to learn their nature, so that the assistance contract can be better tailored towards its objectives, eg. by offering lower grant rates for more capital-intensive projects.  The opportunity to vary the assistance contract does not exist for automatic assistance, unless there happens to be some an observable firm characteristic that is correlated with the relevant unobserved characteristic, which can then be incorporated into the published eligibility criteria.  In the absence of this, the government offers a single automatic grant rate to all firms, which is a pooling contract.  It means the capital-intensive projects will receive too high a grant rate, and explains the failure of the earlier regional policy efforts.

(ii) Moral Hazard: For assistance to be worthwhile it must change the firm’s behaviour, by, for example, encouraging it to implement a larger project with more jobs.  Moral hazard is when the firm in receipt of assistance implements a different project scale to that agreed in the assistance contract.  Optimally, it would seek to implement the without-subsidy scale, since if it can vary its funds after the contract has been agreed the marginal cost of its funds is the same as in the without-subsidy position, and therefore its decision is also the same.  In this case, the effective grant rate received by the firm is increased, but it is a lump-sum subsidy, and the firm earns an informational rent.  As the grant does not change the firm’s behaviour it is said to be wholly ‘non-additional’.15  Discretionary assistance enables the government to scrutinise a project to test the firm’s commitment to carry out the project that is to be agreed under the assistance contract.  Surprisingly, RSA operated this way prior to the introduction of a ‘claw-back’ condition in 1984.16 

(iii) Bargaining:  This advantage of discretionary assistance also relates to asymmetric information, but now because the government is better informed when there is negotiation over the assistance contract.  Bargaining occurs where the number of firms is small or the project scales are large, which increases the firm bargaining strength.  It can occur with both automatic and discretionary assistance, as it is independent of the method of disbursement.  Discretion gives the government an advantage in negotiation, as it does not have to reveal the minimum grant rate that it is prepared to offer.  This increases its bargaining hand, as in the disagreement position the grant rate is zero (ie. the assistance is refused), whereas the automatic grant rate is public knowledge.  Because of this advantage, in practice, governments tend to offer assistance in discretionary form where bargaining is more likely.17

Discretion is exercised in relation to the eligibility criteria, as well as any undisclosed conditions, so that the changes to the RSA eligibility criteria have also altered the way in which discretion is exercised.  In relation to job-creation projects (Category A), RSA discretion was initially exercised over the form and the rate of assistance in terms of the number of jobs, the firm’s viability and public-sector contribution.  In principle, these could be observed ex-post to the assistance and were contractible, but there was no ‘claw-back’ condition, so the purpose of discretion to mitigate moral hazard by testing the firm’s claims.  The introduction of the ‘claw-back’ condition to RSA changed the nature of discretion.  The purpose now is to test the firm’s claims about the ‘Proof of Need’ to alleviate adverse selection, in order to make an Exchequer cost saving and improve the cost-effectiveness of assistance.  The ‘Proof of Need’ concerns the prospective counter-factual position, which cannot even be observed ex-post. It is hypothetical, so that any conditions regarding it cannot be written into the contract. The drawback is that the government may make errors in scrutiny, of which there are two types:

Type II: The grant rate is higher than the minimum necessary.

Type I: The grant rate is lower than the minimum necessary.

The errors are analysed in Wren (2003b).  This assumes fixed project investment scales and non-negotiated assistance.18   First, a higher grant rate may be offered to a project than is necessary for it to proceed, so the firm extracts an informational rent (a Type II error).  This is non-additional assistance expenditure and it reduces the Exchequer cost saving.  However, if the maximum discretionary grant rate is the same as the single pooling grant rate offered under automatic assistance then it can be shown that the investment level is not affected.  The second type of error is where the grant rate is too low and a project does not go ahead (a Type I error).  In this case, investment may be lower than for the automatic scheme.  There are two special cases of the Type II and Type I errors.  These are when the project is either ‘non-additional’ or ‘additional’, which do not and do depend on assistance respectively, as follows:

Type IIa: The government wrongly offers a grant to a ‘non-additional’ project.

Type Ia: The government wrongly refuses a grant to an ‘additional’ project

Where a grant is given to a ‘non-additional’ project then there is no increase in project scale at all, so that the assistance is wholly non-additional (a Type IIa error).  Where a grant is refused to an ‘additional’ project then only in this case is investment lower (a Type Ia error).  As well as the errors in scrutiny, which have traditionally been characterised as a problem of ‘picking winners’ (see Burton, 1983), there are other drawbacks of discretionary assistance.  These are the higher firm compliance costs and government administration costs compared to automatic assistance (Swales, 1997), which combined are about ten per cent of the grant value (NAO, 2003).  There is also the possibility that firms do not take the assistance into account in their investment appraisal because of the uncertainty over the grant rate. We now consider the evidence on RSA. The experience of the automatic support was poor. Discretionary assistance has potential advantages that can lead to an Exchequer cost saving, but the errors that arise in scrutiny mean that the outcome is an empirical matter.

5:    The Evidence on Regional Selective Assistance

       Over recent years a steady body of evidence has accumulated on the effect of Regional Selective Assistance.  As the focus of national and regional economic policy has changed over time so has the nature of the studies, with a greater emphasis on efficiency and productivity.  However, RSA continues to be mainly about job creation, and the aim of the scheme is given as ‘encouraging sound projects, which will improve employment opportunities in the Assisted Areas’ (HC 852, 2003).  This is reflected in the nature of these studies, which have a strong emphasis on the employment effect of this scheme, and particularly its cost-effectiveness.  In this section the focus is on the way in which RSA changes the investment decision, including plant location, while in Section 6 the employment effect and effectiveness are considered.  There are two main sources of evidence (see Taylor, 2002).  The first are the government-funded evaluations based on a questionnaire survey of recipients of RSA, and the second are the econometric-based studies that have matched RSA payments to other data at the individual plant level. The survey-based approach extracts more qualitative evidence on both process and impacts, but it may be less reliable than the econometric studies.

Survey-based Evidence

       The Government has funded three evaluations of RSA, for the periods 1980-84 (King, 1990), 1985-88 (PACEC, 1993) and 1991-95 (Arup Economics and Planning, 2000).  These studies use structured interviews with senior personnel of randomly drawn samples of RSA-recipient firms.  The findings are based on the interviewees’ own subjective assessments some time after its receipt of assistance, so that caution should be exercised.  The survey approach is criticised by NERA (2003), who argue that it induces a strategic bias, whereby beneficiaries exaggerate the importance of grants to maintain their availability.19   However, the approach is much used by public bodies to evaluate the effects of schemes, while it is acknowledged that greater complexity does not necessarily imply greater accuracy (NAO, 2003).  Over the years, Regional Selective Assistance is perhaps the most evaluated scheme, while a similar pattern of results is obtained for RSA across the three evaluations.  This section focuses on the project ‘additionality’, which is the extent to which the provision of RSA alters the firm’s investment decision, whether in timing, scale or location, and on the wider effects of assistance.

       A concern with the Government evaluations is that the extent of a Type I error is not known.  This is because they interview successful applicants only, so that it is not assessed how many projects are foregone as a result of a grant offer that is too low (ie. a Type Ia error).  Moreover, it is not known what investment is foregone due to the more stringent eligibility criteria that have been introduced to RSA.  Over the 1990s, thirty per cent of RSA applicants were refused assistance, but many more may have been discouraged from applying.  Instead, the evaluations focus on the Type II error, and the ‘additionality’ of assistance.  These results are summarised in Table 3, for which there is a total of 526 observations, and which do not vary much across the three evaluations.  Arup Economics and Planning (2000) find that “53% of projects would have been able to go ahead at the same scale with a lower grant amount” (para 5.5), so that a Type II error occurs in half the projects.  This includes a Type IIa error in 18.5 per cent of cases (Table 3), so that a project goes ahead in exactly the same form in about a fifth of projects.   Since there is a claw-back condition throughout this period it is reasonable that the errors result from adverse selection rather than a moral hazard problem. 

In the remainder of cases, where RSA has some ‘additionality’, Table 3 shows that 21 per cent of projects would have been abandoned altogether, ie. wholly ‘additional’.  The other 60 per cent of projects were altered in some way, whether by timing, scale or location.  Of those changed in timing, the grant may have brought a project forward only by a year or two, while there is no information on the amount by which projects were changed in scale.  In the case of the projects changed in location, these are likely to be inward investors in the form of foreign direct investment, as throughout the period covered by the studies RSA was not given to firm relocations within the UK.   Overall, the evaluations suggest that RSA operated well in terms of the Exchequer cost savings, with an absence of a Type II error in roughly half the cases.  However, it cannot be judged if this was achieved at the cost of substantial Type I errors, since while thirty per cent of applications are turned down (and many more may have been discouraged), it is not known how many of these projects subsequently didn’t go ahead.

The results on ‘additionality’ highlight some important differences between projects of a different scale.  This is shown in Figure 3 for the period 1991-95.  Of the projects that were altered in some way, it shows that compared with the large projects (greater than £1 million) a much higher proportion of smaller projects (less than £25,000) were changed in timing (45% compared with 7%), whereas a much higher proportion of the large projects were changed in location (49% against 2%).  Between these two extremes there is a reasonably smooth pattern. While the evaluations do not reveal whether these large projects are foreign direct investment or not, there are reasons to believe so.  Taylor and Wren (1997) report that ninety per cent of UK-owned firms offered RSA were small and medium-sized enterprises (SMEs).  Further, Figure 4 shows the share of foreign direct investment in RSA over 1990-2003.  It reveals that the inward investors account for only about one-tenth of RSA offers, but were offered more than fifty per cent of the total grant.  Further, as much as eighty per cent of RSA-supported investment is in the form of foreign direct investment in some years.  Overall, it appears RSA alleviates a ‘soft’ budget constraint in small projects, enabling these to go ahead, but it alters the location of large projects.  Since RSA is now available to larger projects only it indicates that location is the main issue for future regional policy evaluation.  Figure 4 also shows the higher proportion of RSA offers going to inward investment in recent years.

The surveys also collect some evidence on the other characteristics of RSA-supported projects, and the extent of this information has gradually increased over time.  No attempt is made to relate this to project ‘additionality’, so that it refers to the project as a whole.  In the most recent survey, job quality was assessed in terms of earnings, but it was found that RSA-supported employees were less well paid than those in manufacturing nationally, but perhaps just reflecting regional wage differences.  Three-quarters of employees were recruited locally and two-thirds received some training, but 19 per cent of interviewees expected employment to decrease over the following three years.  In fact, little consideration is given to the duration of the jobs that are created and it is usually assumed that they are lost at the same rate as the physical decay of capital, as in Marquand (1980).  Evidence is collected on output effects, but it is not believed that this was always of good quality, while the evidence on the wider effects of RSA is that three-quarters of respondents believe they achieved a competitive advantage, mainly in the form of a lower cost base or through products that were either better quality, technically superior or new (Arup Economics and Planning, 2000, chapter 7).  These findings are difficult to quantify and they are mainly suggestive of the other benefits of RSA.

Econometric-based Evidence

Several recent studies have matched RSA details to the plant-based data in the Annual Respondents Database, and this has enabled some more rigorous assessments of the broader effects of regional assistance, such as location, survival and productivity.  As well as applying regression techniques to objective data, an advantage of these studies are the large number of observations on assisted and non-assisted plants.  The studies of interest are Devereux et al (2001), who manage to match 45 per cent of RSA offers to the plant-level data, and Harris and Robinson (2001), who match around 62 per cent of the relevant cases.  In addition, Jones and Wren (2003, 2004) match RSA offers to a dataset of 550 inward-investment projects in the North-East England, which permits the matching of RSA grant awards to individual projects.  These studies indicate qualitatively different effects for firms in receipt of Regional Selective Assistance, although they do not control for possible selection effects.

Devereux et al (2001) examine the location of new plants that are part of foreign-owned multinationals.  They examine location across the ten regions of Great Britain over 1992-96, and estimate the offered grant that a firm could expect to receive in each of the five Assisted Area regions.  This is based on the existing pattern of RSA offers according to the region, the plant activity and firm ownership structure.  The results suggest that at the margin an extra £5000 in RSA increases the probability of location in an Assisted Area by about one percentage point.  However, RSA becomes insignificant when regional terms are included into the regression equation, either for computer investment or industrial diversity, which perhaps suggests that the RSA variable is capturing a regional effect.  The expected amount of RSA is calculated using regional and industry variables, whereas plant-level variables are likely to be more important in determining the grant offer, such as the project scale and number of jobs.

Harris and Robinson (2001) examine the characteristics of RSA recipients over 1990-98.   Plants that are larger, older or foreign owned have a higher probability of receiving RSA, as do single-plant firms, but larger firms have a lower probability.  RSA-assisted plants have a significantly higher probability of survival, with a hazard rate for new plants that is 30 per cent lower.  An analysis of total factor productivity includes two RSA dummy variables.  The first takes a value of unity if the plant is at any time assisted by RSA over 1990-98, and the second has a value of unity for all the years after the receipt of RSA. When estimated across plants in Assisted and Non-Assisted Areas both are significant, but the first is negative and the second is positive.  Together they indicate that RSA-assisted plants have productivity that is 5 per cent lower than other plants, but that RSA improves this by 2.5 per cent.  However, when the equation is regressed for the Assisted Areas only, the RSA terms are insignificant, which may again indicate that they are picking-up a regional effect.

The study by Jones and Wren (2003) is for foreign-owned plants in a single Assisted Area region over the period 1985-98.  It finds that at start-up, RSA-supported plants are larger than other foreign-owned start-ups, they are more labour-intensive and they are more likely to reach their initial job targets.  In terms of plant performance, the receipt of RSA increases the probability of re-investment leading to a ‘significant’ expansion from 0.40 to 0.90, although it has no effect on the time duration to this first re-investment.  RSA-supported plants are more likely to survive, but only the smaller plants promising less than 50 jobs at start-up, where it doubles the life of a plant from about 7 to 14 years.  However, larger plants have high failure rates, whether in receipt of RSA or not, but only after a period of about ten years.  Putting this together with the earlier evidence on ‘additionality’, it again indicates that the main effect of RSA in the large inward investment projects is to determine location.  Nevertheless, recent evidence suggests that the attraction of foreign-owned plants can raise the productivity of other plants, with Haskel et al (2002) estimating a UK domestic productivity benefit that is valued in present value terms at around £2,500 per job in the inward investment plant.

6:    Does Regional Policy Measure Up?

       Notwithstanding the above evidence that has been accumulating on the positive effects of Regional Selective Assistance, UK regional policy has been thrown into some doubt by the recent National Audit Office report on regional grants.  This finds that the 3,845 RSA grants over the period 1991-95 resulted in just 21,000 jobs at a cost of £21,000 per job (NAO, 2003), and that probably fewer jobs than this were created than this and at a higher cost.  It has been interpreted as placing “a question mark over the value of regional intervention” (Financial Times, 2003), but there is great uncertainty about the regional policy employment effect.  The Arup Economics and Planning (2000) study, from which the National Audit Office draws its estimate, actually finds that 84,000 jobs were created by RSA over 1991-95.  However, a background study prepared for the National Audit Office finds that it is fewer than 6,000 jobs (NERA, 2003).  On these different views UK regional policy is either a reasonable strong or a very weak job-creation measure, so that the issue needs to be investigated.

       The Regional Policy Employment Effect

       The controversy over the size of the regional policy employment effect does not arise from differences in methodology, but from a disagreement over the value of the survey-based approach and the interpretation of the results found by Arup Economics and Planning (2000).  The results of this study are reproduced in Table 4.  This gives the employment effect of RSA in Great Britain and England, but since the latter is based on an allocation of jobs pro-rata to the distribution of grants, the ‘cost per job’ estimates are identical for these.  Table 4 reveals that a difference between the two reports is that the National Audit Office quotes a figure for England (of 21,000 jobs), but that Arup Economics and Planning do so for Great Britain (of 84,000 jobs).  However, this alone cannot reconcile the competing estimates.20  Table 4 shows there are four different estimates of the employment effect, ranging from 39,000 to 652,000 jobs for Great Britain, implying ‘costs per job’ of between £17,500 and £1,050 (1995 prices).  Of these, the National Audit Office reports the former only, but at 2002 prices, giving a figure of £21,000 for each job.21   To understand the different estimates it is necessary to understand the different job measures that are used.

       The gross employment effect shown in Table 4 refers to all 210,000 jobs associated with the projects in Great Britain offered and taking up RSA between January 1991 and June 1995.  However, since some jobs do not depend on assistance, while jobs may be displaced or created elsewhere as a result of the assistance it is necessary to adjust the gross figure to get a measure of the net employment effect.  This uses the survey results from a stratified sample of 165 RSA recipients.  It includes adjustments for additionality’, for the displacement of jobs in the Assisted Areas, for linkage effects in supplier firms and for multiplier effects.  Combined they mean that the net employment effect is forty per cent of the gross employment effect, at 84,000 jobs (Table 4).  A further adjustment is made because the jobs occur in different time periods and have different lifetimes. Table 3 shows that about a quarter of projects are brought forward through time by RSA, but otherwise it is assumed that the jobs have a lifetime equal to that of the asset provided by the project, which on average is about ten years.22  To express the jobs in present value terms the prevailing Treasury discount rate of 6 per cent is used and applied to the net jobs created.  The result is presented in two ways: as a present value net job year (PJY), which is a job year, and as a permanent net job equivalent (PJE), which is a job lifetime.  A job that lasts for 10 years is equal to 7.4 PJYs, and a job that lasts forever is equal to 16.7 PJYs or 1 PJE. Table 4 shows that the 84,000 net jobs created in Great Britain are equal to 652,000 PJYs or 39,000 PJEs, and this explains the different estimates.

       The study by NERA (2003) takes the PJE estimate for England, and reduces this figure for three reasons.  First, it argues that ‘additionality’ is over-estimated by the survey approach due to strategic bias.   Second, referring to the general equilibrium modelling of Gillespie et al (2001), it argues that real wages will adjust upwards in response to regional grants.  Finally, citing the empirical evidence on RSA in Wren (1994), it argues that the assisted jobs do not last as long as ten years on average.  Starting with an employment effect of 17,000 PJEs for England, they conclude that “a more accurate net number of jobs created in England might be no more than 5-6,000 permanent job equivalents” (NERA, 2003, p. 36).23  The National Audit Office voices similar concerns.  It is controversial, and at face value it suggests that regional policy is creating not much more than 1,000 jobs a year across the Assisted Areas.

       So, how is the employment effect of UK regional policy to be judged?  As a first point, the discounting of future jobs is highly unusual in evaluation work.  Since the job effects are occurring in different time periods it is reasonable that they are discounted to reflect different time values.24  However, it means that the PJYs and PJEs are not directly comparable with the estimates obtained for other assistance schemes, which focus on the number of jobs created. It suggests that for the purpose of comparison the appropriate estimate of the policy effect is the number of net jobs created shown in Table 4.  On this basis, RSA created about 18,700 jobs a year in the Assisted Areas of Great Britain over 1991-95. As a second point, to understand the significance the job effect it can be compared to the unemployment level.  While the Assisted Areas now extend to all regions, it compares with a combined unemployment claimant count of about half a million in Scotland, Wales and the three English regions with the highest rates (North East, North West and Yorkshire and Humberside).  Clearly, the effect is small relative to the scale of regional unemployment, while if permanent job equivalents are considered the effect is even smaller.  On a different interpretation, regional policy may have accounted for a substantial proportion of the fall in unemployment in the Assisted Areas over the 1990s.25  However, Figure 1 shows unemployment fell more sharply outside these areas.

       The Effectiveness of Policy

       The Arup Economics and Planning (2000) study suggests that regional policy created about 18,700 jobs per annum in the Assisted Areas of Great Britain, which at 1995 values is equal to about 8,700 permanent jobs (PJEs) a year.  This is small relative to the scale of the unemployment problem, but it seems that RSA is highly cost-effective.  The headline figure of the National Audit Office is £21,000 per job, while NERA put it at around £50,000 per job, but these are permanent net job equivalents.  Most studies express the ‘cost per job’ in terms of the net jobs created, and on this basis Table 4 shows that the ‘cost per job’ is much smaller at £8,150 (1995 prices).  This expresses the cost in terms of the present value net Exchequer expenditure on grants, and it deducts the flow-backs to the Treasury in the form of higher tax revenue.  Clearly, RSA is much more cost-effective than the automatic support.

       To judge the efficiency of regional policy the cost of a job created can be compared to the annual amount paid in benefits to an unemployed person.  Table 4 shows that the ‘cost per job year’ is £1,075 (1995 prices and values), which compares with the annual Jobseeker’s Allowance of just under £3,000 paid to an unemployed person aged 25 years or more.  Even allowing for the front-end costs of job creation, and the possibility that not all the jobs created will lead to a reduction in unemployment, it suggests RSA is relatively efficient.  Moreover, it ignores the other positive benefits that assistance may have, such as on firm productivity in recipient and other firms.  RSA is also cost-effective when compared with measures designed specifically to reduce unemployment.  In a back-of-the-envelope calculation, Layard (2001) puts the short-run average cost of a person in employment from the New Deal at about £7,000, but in gross expenditure terms.26   Thus, not only is RSA much more cost-effective compared with earlier UK regional policy efforts, but it appears relatively efficient when compared with other job-creation measures and with the transfers paid to the unemployed.

       Suggestions for an Expanded Regional Policy

       While regional policy may have made a sizeable contribution to the recent reduction in regional unemployment, it is small in relation to the scale of the problem, creating only about 18,700 jobs per annum in the Assisted Areas of Great Britain.  Spending on regional grants is set to fall further, with the combined annual value of offers on RSA and EGS decreasing over recent years from £205m to £123m in England.  Moreover, expenditure on regional grants is one of the smallest in the European Union.  Figure 5 shows expenditure on regional assistance in the Objective 1 areas of the EU, corresponding to the Tier 1 Assisted Areas.  It expresses assistance as an annual average over 1994-96 relative to the GDP of each area, showing the national state aids separately from the assistance under the EU Structural Fund.  It reveals that UK spending on regional grants is less than a quarter of the EU average, and smaller than in countries such as France or Germany.  Further, expenditure on Structural Fund assistance in the UK is half the EU average.  The latter is likely to fall substantially when the membership of the EU increases to twenty-five states (see HM Treasury, 2003).  

       The Department of Trade and Industry (DTI) is currently reviewing and rationalising its grant schemes, with a view to simplifying its industrial support and making it subject to ‘market-failure tests’.27  This includes RSA and EGS, in order to increase their contribution to improving productivity and competitiveness.   The National Audit Office reports that the DTI expect “to create a new capital investment support instrument to replace the existing schemes” (p. 7, NAO, 2003).  However, notwithstanding the value of changing the EGS scheme so soon after introduction, it was argued above that RSA is an effective job-creation measure.  Further, the scale of the regional policy employment effect is relatively small and expenditure is set to fall, so that if anything regional policy should be expanded.  This can be achieved in many ways (Taylor and Wren, 1997), but here we focus on two main ways in which the RSA grants can be extended.  One is to change the operation of the scheme, including its promotion, grant rates, exercise of discretion and so on.  The other is to relax the eligibility criteria in Table 2, including some of the more recent changes.

       (i) Changes to Scheme Operation.  The first change is to increase the promotion of the scheme, not just to increase the overall take-up, but to increase the take-up by certain kinds of firm.  Currently, ninety percent of projects are in manufacturing, which is serving to support the existing industrial structure of the regions rather than diversify it.  The demand-led nature of regional policy may explain the poor take-up of EGS, and the high degree of repeat awards to firms in the case of RSA, amounting to 31 per cent of the value of offers accepted (NAO, 2003).  A potential difficulty with this are European Unions rules governing the promotion of schemes, although the government is still able focus the aid on certain categories of project, such as in services or those improving productivity and competitiveness. 

A second change is to offer higher grant rates to encourage more marginal projects.  It involves trading a larger employment effect for some loss of effectiveness, but the evidence presented above suggests that RSA is cost-effective, and that there is slack for this to occur.  In fact, the average grant rate on capital investment is in the order of 15 to 20 per cent (Figure 4), but the limits on the grant rate set by the European Commission offer considerable scope for discretion. They permit grant rates of up to 40 per cent in Tier 1 areas (and 55% for SMEs) and between 10 and 30 per cent in Tier 2 areas (with a 10% supplement for SMEs).  A danger in setting the grant rate too low is that only projects that do no need assistance come forward.  Finally, around a third of RSA applicants are not assisted, and the stringent eligibility criteria and lack of scheme promotion may mean that many potential applicants are either discouraged or not aware of the scheme.  The Government evaluations collect information only on the projects that are in receipt of assistance, so that the extent of the Type I errors and the reasons for non take-up are not explored.  This kind of information is vital in order to understand the feasibility for an expansion of regional grant assistance.

       (ii) Changes to Scheme Eligibility.  As well as changing the operation of the scheme it is possible to change the eligibility criteria. It was mentioned above that two of the recently-introduced eligibility criteria are likely to lead to a substantial reduction in spending on RSA. These are the Eligible Investment and Quality conditions shown in Table 2. However, it is two of the other criteria that go to heart of what defines current regional policy.  These are the Jobs condition, first introduced in 1984, that rules out assistance for employment diversion, and the National and Regional Benefit condition, first introduced in 1979. 

Regional economic policy is defined as the promotion of growth and employment in lagging areas (Regional Studies Association, 2001).  However, in recent years there has been confusion, and there are three reasons for believing that ‘regional policy’ is little more than a national economic policy that is applied at the regional level.  First, the Government regional productivity paper defines its objective as improving the performance of every region in order to achieve high and stable levels of growth and employment (HM Treasury, 2001).  There is no intention to favour lagging regions.  Second, the Regional Development Agencies exist in every region of England and they essentially exploit the comparative advantage of these areas.  Again, there is no presumption to favour lagging regions, and it is conceivable that regional specialisation will actually favour the more prosperous regions of the UK.  Finally, the only explicit industrial policy measure to favour the regions is RSA, but it is biased towards inward investment.  Since it is the only real opportunity to offer financial assistance to large projects, and hence for competing internationally for inward investment, it may be best viewed as an instrument of national policy.  Thus, ‘regional policy’ is primarily about the National Benefit rather than the Regional Benefit.

       More recently, a regional economic growth Public Service Agreement target has been agreed for England by the DTI, the Office of the Deputy Prime Minister and HM Treasury.  It aims to “make sustainable improvements in the economic performance of all English regions and over the long term reduce the persistent gap in growth rates between regions” (HM Treasury, 2003).  Performance is measured by gross value added per head, and detailed plans are currently being drawn-up (November 2003).  The new target makes a commitment to reduce the ‘regional gap’, but in order that this is achieved it seems likely that it must embody a substantial distributional component to policy (Adams et al, 2003).  The RSA Job condition rules out job diversion, whether by plant relocation or employment displacement, but this contrasts with earlier efforts where the transfer of jobs was one of the main aims of regional policy.  Over the 1960s Ashcroft and Taylor (1979) find that regional policy shifted around 500 establishments to the Assisted Areas, and that it raised their employment by about 100,000.28   The time may now be ripe to reintroduce a distributional component to RSA to encourage the relocation of UK-owned plants.  There are four reasons for this.

The first is that the survey-based evidence suggests that the grants have a significant effect on plant location and it is the main effect in large projects.  Figure 3 shows that it alters the location in half the projects.   Second, it is reasonable to make the grants available to UK-owned plants on the same basis that they are available to foreign-owned plants.  An inward investor may be given a grant to help it choose between an Assisted and Non-Assisted Area location, but the assistance is not available to existing UK plants to relocate between regions.  Again, it reflects the over-riding national objective of regional policy.  In the absence of RSA, Figure 3 shows a third of the large projects would have chosen a location elsewhere in the UK, and presumably a Non-Assisted Area.  Third, foreign direct investment has fallen sharply in recent years, and it is unlikely to continue underpinning regional policy (Invest UK, 2003). Finally, redistribution involves efficiency losses, but it can relieve inflationary and congestion pressures in the prosperous regions, leading to an overall gain.  It need not conflict with EU state aid policy, as Besley and Seabright (1999) find that the European Commission approves about 98 per cent of the applications for state aid, often for social or distributional reasons.  Further, it need not lead to competition at the regional level, as the DTI handles large RSA cases in London.  In the case of Scotland and Wales, financial assistance is a devolved matter, and there is consultation on relocations under a Memorandum of Understanding.

7:    Conclusions

Regional disparities in unemployment continue to exist, and have been worsening over the last decade, while expenditure on regional policy has been falling and it is set to fall much further.  The paper has reviewed the UK regional industrial policy experience, focusing on the recent evidence for the effect and effectiveness of this policy.  Regional policy continues to be primarily about job creation, but there is considerable controversy about the employment effect, with research putting it between 8,700 to 18,700 jobs per annum, or even 1,000 jobs a year, depending on the study and the particular job measure that is chosen.  The difficulty is that even on the larger estimate the effect is small relative to the scale of the unemployment problem.  The paper argues for an expansion of regional policy to boost the overall effect, but what is currently called ‘regional policy’ it finds is little more than a national policy applied at the level of the region.   The regional grants are a highly cost-effective job-creation measure, that do not share the problems of earlier regional policy efforts, so that one way to expand the policy is to make the grants available to mobile UK investment in the same way that they are available to foreign direct investment. Indeed, given the recent substantial reduction in inward investment it may be the only way to sustain a substantial job creation policy that is biased towards the regions, and which may genuinely be called a ‘regional policy’.

Notes

1.  The North-East of England is now the worst performing UK region, with an unemployment rate above that of Northern Ireland.  The South-East region, including Greater London, has a higher unemployment rate than some other UK regions. Greater London has a higher unemployment rate than the North-East, but this is not a proper comparison.  The claimant count of unemployment is used, as a consistent time series is available, although there have been many changes to claimant eligibility.  The claimant count understates the true scale of the unemployment problem (Regional Studies Association, 2001; and Beatty et al, 2002), so that the regional differences are probably greater than is indicated here.

2.  The relationship may have returned to its long-run position, but the change in the sectoral composition of the economy does not rule out the possibility that it could breakdown again at a time of recession.  This is because the size of the service sector is such that the south of England may now be more adversely affected by recessions.  However, counter to this, the sectoral shift in employment from manufacturing to services has occurred most rapidly in the traditionally Assisted Area regions (Wren and Taylor, 1999).

3.  Taylor and Wren (1997) note that the regional disparities in unemployment rates virtually disappeared in the mid-1990s, but that the regional aggregates concealed immense spatial disparities in unemployment within regions.  However, since relatively high unemployment persisted in certain areas, it was this that justified spatially discriminating policies.  Since this time unemployment has fallen nationally, but disparities in unemployment at a regional level have worsened, so that a ‘regional problem’ has re-emerged.

4.      Other regional development programmes are subject to revision. The increase in European Union membership to twenty-five states will mean that the Structural Funds will be much reduced, with EU policy focusing on the Cohesion Funds (Armstrong, 2001). The House of Commons Treasury Committee is investigating the regional pattern of ‘mainstream’ government expenditure.

5.  Details of the schemes making up each component can be found in Wren (1996) or below, while the payment of assistance can lag the life of a scheme by some period, so that the Regional Development Grant (see below) was still being paid ten years after its abolition.

6.  SEP was paid at 37.5p per male worker, and smaller sums for women and juveniles. REP was initially paid at £1.50 per week for full-time male workers, but 75p a week for women and boys and 47.5p a week for girls (at half these rates for part-time employees).

7. Sumner (1999) finds that “it is no easier to reach firm and positive conclusions about the effects of fiscal policy on investment than it was twenty years ago” (p. 298).

8.  The changes have had four motivations. First, they sought greater effectiveness by making projects depend on the least amount of assistance.  Second, they sought to reduce spending by restricting take-up. Third, they sought greater efficiency by reducing the administration costs.  Finally, they sought to promote national as well as regional objectives.

9.  Service firms that were not primarily serving a local market were eligible, but the bulk of assistance went to manufacturing industry under Category A.  Service firms relocating to the Assisted Areas could obtain additional assistance, but this was scrapped in 1984.

10. The 1945 Distribution of Industry Act set the framework for regional industrial policy in the immediate post-war period, but in fact it was little used.

11. The changes introduced to the first five conditions of Table 2 were that there should be no prior commitment to the project, the project and firm must both be viable, a payback period of three years was required and a ‘National and Regional Benefit’ was sought, whereas previously it was a ‘Regional and National Benefit’.

12. These are firms with up to 250 employees.  Similar schemes exist in Scotland and Wales.

13. The designation of the Assisted Areas is subject to EU guidelines, including a ceiling on population coverage.  The latest Assisted Areas are discussed in Armstrong (2001).  Tier 1 areas have per capita GDP below 75% of the EU average, and are designated for Objective 1 support under the EU Structural Funds.  Tier 2 areas are defined by the UK Government and capture ‘acute labour market need’.  In England the Government has designated Tier 3 areas for small business support.  These are areas of ‘identified special need’, including high unemployment, low employment, coalfield closure and rural areas. Since they include areas outside the Assisted Areas they are designated as part of national assistance under Section 8 of the 1982 Industrial Development Act.

14. After the introduction of EGS, applications for RSA in England fell from 1,075 in 1999/0 to 269 in 2000/1, and in Scotland from 348 to 76. Wales lags these developments.

15. Spending of this kind is referred to as ‘deadweight spending’ in the evaluation literature, but in a strict economic sense it is a transfer rather than a deadweight welfare loss, although it still has implications for efficiency.

16. Since moral hazard can induce an adverse selection of firms, a ‘claw-back’ condition can also dissuade some poor types from applying for assistance.

17. This is where the number of firms and projects is small and / or the project investment scales are large. Bargaining can occur with automatic schemes where a firm with sufficient bargaining strength can extract extra assistance even though it is nominally available at a specified rate.  In the case of discretionary assistance, the maximum available grant rates are determined by EU guidelines, but the Government does not reveal the maximum grant it is prepared to offer. As well as uncertainty over the grant rate, the eligibility criteria for discretionary assistance are usually couched in a way that allows discretion to be exercised in conjunction with undisclosed conditions, so that a firm could be refused assistance even though it believes it has satisfied the eligibility criteria. A variable grant rate or the refusal of assistance does not define discretionary assistance, as the grant rate could be varied in the case of an automatic scheme by making it depend on objective eligibility criteria.

18. Fixed investment scales are reasonable, as Table 3 below shows that less than 20 per cent of projects were changed in scale by the receipt of assistance. Similar outcomes to that described below occur where the grant rate is bargained.

19. In its place, NERA suggest an assessment based on contingent valuation techniques, but this can really only tell us the about the preferences of individuals and how Government should operate, but not about the impacts of its programmes.  Swales (1997) argues for a totally different evaluation approach based on a full-blown Cost-Benefit Analysis.  Owing to the informational problems that make evaluation difficult, in practice evaluations focus on the measurement of the direct employment effect (McVittie and Swales, 2003)

20. Over the period 1991-95 43.4 per cent of RSA grant offers were in England, which pro-rata gives an employment effect of around 17,000 jobs, but these are at 1995 values.  To produce the estimate of 21,000 jobs the National Audit Office appears to have valued the jobs at 2002 values using a discount rate of 6 per cent, but this is not explained.

21. Similar estimates were obtained in the other RSA evaluation studies.  Based on the first of these (at 1995 prices) they are 1980-84: £15,800; and 1985-88: £18,000.

22. Figure 3 suggests that these tend to be smaller in scale, where up to 40 per cent of projects are moved forward in time, so that the assisted jobs in larger projects have longer lifetimes.   If a project brings a job forward by a year or two, then it lasts only one or two years, as it is no longer attributable to the assistance after this time.

23. For strategic bias and ‘additionality’ NERA (2003) make an adjustment of 0.74, and for flexible labour markets they make an adjustment of 0.44, giving 17,000 jobs x 0.74 x 0.44 = 5,500 jobs.  They use the figure of 17,000 jobs for England, which is at 1995 values, and not the 21,000 jobs reported by the National Audit Office, which is at 2002 values.

24. In Cost-Benefit Analysis the benefits and cost are expressed in present value terms, where these are in monetary units, such as willingness-to-pay prices.  However, regardless of how the jobs are valued, it is seems plausible that the jobs should be discounted to the present to reflect the different time values, even though they are not expressed in monetary units.

25. NERA (2003) report that the reduction in unemployment in English travel-to-work areas with unemployment rates more than two per cent above the national average was 154,000 over 1996-2002.  If RSA continued creating jobs in England at the same rate as previously, then it created 60,000 jobs over this period, which is a sizeable contribution.

26. This is the New Deal for Young People, which is designed to prevent young people from entering long-term unemployment.  The gross Exchequer cost of the scheme was £350 million by about the end of 2000, but less flow-backs to the Exchequer of around half this amount.  It is estimated that claimant unemployment was reduced by about 50,000 by this time and that regular employment increased by half this number. The average cost of a person in employment from the New Deal at about £7,000, while the average cost of a person off the unemployment register is put at about half this amount.

27. The National Audit Office reports that, while RSA aims to combat the effects of multiple market weaknesses, these weaknesses are not identified by the Government (NAO, 2003).

28. The EGS does not rule out plant relocations, but it is available only to SMEs and the way the grants operate seems likely to be left to local discretion.  Rees and Miall (1981) analyse the effect of regional policy on manufacturing investment and capital stock over 1951-78.  They find a shift of investment towards the Development Area regions from 1960, and a shift of UK capital stock to these regions equal to 1.6 % of UK capital stock by 1973.

Figure 1: Disparities in Regional Unemployment

Per cent
 
Sources: Abstract of Regional Statistics, Regional Statistics and Regional Trends.

Notes: Figures are generally the annual averages of unemployment based either on the numbers of unemployed registered or the unemployment claimant count, but later years give the position at mid-year.  The regional variation is calculated as the coefficient of variation of regional unemployment rates multiplied by ten, but excluding Northern Ireland.  ‘NE - SE difference’ is the North East unemployment rate minus that of the South East, including London. Prior to 1965 the statistics make no distinction between the South East and East Anglia regions, and between the regions of Yorkshire and the West and East Midlands. The regions are based on Government Office regions from 1997, but the North West includes the Merseyside Office area and the South East includes the London Office area.  Prior to this time the North-East region includes Cumbria.
 
Sources: Abstract of Regional Statistics, Regional Statistics and Regional Trends.

Notes: Figures are generally the annual averages of unemployment based either on the numbers of unemployed registered or the unemployment claimant count, but later years give the position at mid-year.  The regional variation is calculated as the coefficient of variation of regional unemployment rates multiplied by ten, but excluding Northern Ireland.  ‘NE - SE difference’ is the North East unemployment rate minus that of the South East, including London. Prior to 1965 the statistics make no distinction between the South East and East Anglia regions, and between the regions of Yorkshire and the West and East Midlands. The regions are based on Government Office regions from 1997, but the North West includes the Merseyside Office area and the South East includes the London Office area.  Prior to this time the North-East region includes Cumbria.
 

Figure 2: Expenditure on Regional Industrial Assistance
£’ billions

1995

prices
 

 
Source: Wren, C., 1996a., and Annual Reports on the 1982 Industrial Development Act.

Note: Figures are for actual grant payments at constant prices for Great Britain.  Data are for financial years, so that 1960 refers to the financial year 1960-61 and so on.

 

Table 1: The ‘Cost per Job’ of Regional Policy

            Manufacturing           Gross Assistance       Net Jobs            ‘Cost per Job’

            Industry                      Expenditure                Created

            Capital-intensive                 £4,717 m                     6,000                  £786,000

            Labour-intensive                 £5,311 m                 204,000                    £26,000

            All industries                     £10,028 m                 210,000                    £47,750

Source: Moore et al, 1986 and author’s own calculations.

Notes: Regional assistance expenditure over 1966-76 at 1982 prices.  The net jobs created include the direct and indirect jobs created over 1966-76 and existing at 1976.  Capital-intensive industries comprise coal and petroleum, chemicals, mechanical engineering, metal manufacture, shipbuilding and marine engineering. Labour-intensive industries comprise all other manufacturing industries.

Table 2: The Current Eligibility Criteria for RSA

Jobs. Projects that create overcapacity, simply displace jobs elsewhere in the UK, or aim to relocate jobs from one part of the country to another, are not eligible.

Viability: Businesses and projects should be viable, and the project will normally be expected to be profitable within three years.

Other Funding. The greater part of the project cost should be funded from private-sector sources.

Proof of Need. Applicants must demonstrate that a grant is necessary to enable the project to proceed.

National and Regional Benefit. All projects should contribute positive benefits to both the national and regional economies.

Prior Commitment. Project appraisal must have been completed and a formal offer of assistance issued, before the applicant enters into a commitment to proceed with the project.

Quality. Assistance is focused more on high-quality knowledge-based projects providing skilled jobs.  Four key factors are used to determine the quality of projects: wage levels, sustainable employment, R&D and training.

Eligible Investment. The project must involve capital expenditure of more than £500,000 on fixed assets, such as property, plant or machinery. Expenditure can relate to expansion, modernising or the establishment of a new company.

Source: HC 83, 2001.

Table 3: Regional Selective Assistance and Project ‘Additionality’

Response:                                                  All              1991-95           1985-88           1980-84

Gone Ahead Unchanged                            18.5                 19.3                 15.1                 22.1

Gone Ahead, but:

            Later in time                                  23.8                 24.0                 29.2                 16.8

            On a smaller scale                          13.6                 12.9                 13.2                 14.8

            Elsewhere outside UK                   13.0                 18.1                   9.0                 13.4

            Elsewhere inside UK                       5.0                   8.2                   1.9                   6.0

Some combination                           5.3                   4.1                   6.1                   5.4

Abandoned Altogether                               20.8                 13.5                 25.5                 21.5

Number of sample firms                          526                  165                  212                  149

Source: Arup Economics and Planning, 2000, and author’s own calculations.

Notes: The table reports the percentage of firms in each category responding to a question on what would have occurred to the project in the absence of RSA, given the alternatives shown.  These are the interviewers’ assessments based on the responses given by senior personnel at the recipient firms.

Figure 3: The ‘Additionality’ of RSA
Source: Arup Economics and Planning, 2000.

Notes: Adapted from survey results of 165 RSA recipients over 1991-95.  It excludes the small number of cases where there is some combination of effects.
 

Figure 4: The FDI Share of RSA
 
Source: Annual Reports on the 1982 Industrial Development Act.

Note: The figure shows the percentage taken by foreign-owned plants of the total RSA-supported investment, the total RSA offered grant amount and the total number of RSA awards.
 

Table 4: Regional Policy Employment Effect, 1991-95

                                                                                   Employment effect       ’Cost per job’

                                                                                         England        Great Britain
Gross jobs created                                                          111,000           210,000             £3,250   
            Less non-additional jobs (x 0.45)
                                                                                           50,000             94,000
            Less displaced jobs (x 0.76)
                                                                                           38,000             71,500

            Plus linkage and multiplier effects (x 1.18)

Net jobs created                                                                45,000             84,000             £8,150

Present value net job years (PJY)                                  351,000           652,000             £1,050

Permanent net job equivalents (PJE)                               21,000             39,000           £17,500

Sources: Arup Economics and Planning (2000) and NAO (2003).

Notes: The table shows the estimated employment effect of RSA over the period 1991-95.  The figures for England are those implied by the National Audit Office report.  This gives an estimated PJE for England of 21,000 at 2002 values, which is equal to 17,000 PJEs at 1995 values.  The ‘costs per job’ are based on net RSA expenditure, at 1995 prices.

                                                                                                                                                  

                                                                                                                                               

Figure 5: Regional Policy Expenditure in Europe
 

Per cent

of GDP
 
Source: DG Competition and DG Regional Policy, European Commission, Brussels.

Notes: Direct financial support to firms in Objective 1 / Tier 1 Assisted Areas. Annual averages 1996-98.
 

References

Adams, J. and Robinson, P. (2002), A New Regional Policy for the United Kingdom, Institute for Public Policy Research, London, Interim Report, December.

Adams, J., Robinson, P. and Vigor, A. (2003), A New Regional Policy for the UK, Institute for Public Policy Research, London, October.

Armstrong, H. (2001), ‘Regional Selective Assistance: Is the Spend Enough and Is It Targeting the Right Places?’, Regional Studies, 35.3, 247-57.

Armstrong, H. and Taylor, J. (1993), Regional Economics and Policy, Philip Allan, Oxford.

Arup Economics and Planning (2000), Evaluation of Regional Selective Assistance 1991-1995, Arup Economics and Planning, London.

Ashcroft, B. and Taylor, J. (1979), ‘The Effect of Regional Policy on the Movement of Manufacturing Industry in Great Britain’, in D. Maclennan and J. Parr (eds), Regional Policy: Past Experience and New Directions, Martin Robertson, Oxford.

Beatty, C., Fothergill, S., Gore, T. and Greene, A. (2002), The Real Level of Unemployment 2002, Centre for Regional Economic and Social Research, Sheffield Hallam University.

Besley, T. and Seabright, P. (1999), ‘State Aids: Making EU Policy Properly Reflect Geography and Subsidiarity’, Economic Policy, April, 15-53.

Burton, J. (1983), Picking Losers? … The Political Economy of Industrial Policy, Hobart Paper 99, Institute of Economic Affairs, London.

CEC (1998), ‘Guidelines on National Regional Aid’, Official Journal of the European Communities, C74, 9-31, Commission of the European Communities, Brussels.

CEC (2001), State Aid Scoreboard, Second Edition, Commission of the European Communities, Brussels, COM 782.

Devereux, M. P., Griffith, R. and Simpson, H. (2001), The Geography of Firm Formation, paper presented to the Royal Economic Society Conference, Durham University.

DTI (1983), Regional Industrial Development, Department of Trade and Industry, Cmnd 9111.

DTI (1998), Our Competitive Future: Building the Knowledge Driven Future, Department of Trade and Industry, Cm 4176.

DTI (2001), The Government’s Expenditure Plans 2001-02 to 2003-04, Department of Trade and Industry, Cm 5112.

Financial Times (2003), ‘Job-creation ‘over-estimated’’, Financial Times, London, 17 June.

Gillespie, G., McGregor, P., Swales, J. K. and Yin, Y. P. (2001), ‘The Displacement and Multiplier Effects of Regional Selective Assistance: A Computable General Equilibrium Analysis’, Regional Studies, 35, 125-39.

Harris, R. and Robinson, C. (2001), DTI Industrial Support Policies: The Impact of Regional Selective Assistance and SMART / SPUR Schemes, Final Report to the DTI, December.

Haskel, J., Pereira, S. and Slaughter, M. (2002), ‘Does Inward Foreign Direct Investment Boost the Productivity of Domestic Firms?’, CEPR Discussion Paper No. 3384, Centre for Economic Policy Research, London.

HC 651 (1972), Public Money in the Private Sector, Expenditure Committee, Trade and Industry Sub-Committee, Sixth Report, House of Commons Session 1971/2, HMSO, London.

HC 85 (1973), Regional Development Incentives, Expenditure Committee, Second Report, House of Commons Session 1973/4, HMSO, London.

HC 536 (1977), Committee of Public Accounts Tenth Report, House of Commons Session 1976/7, HMSO, London.

HC 852 (2003), Annual Report on the 1982 Industrial Development Act, House of Commons, House of Commons Session 2002/3, London.

HM Treasury (2001), Productivity in the UK: 3 - The Regional Dimension, HM Treasury and Department of Trade and Industry, London, November.

HM Treasury (2003), A Modern Regional Policy for the United Kingdom, HM Treasury, Department of Trade and Industry and Office of the Deputy Prime Minister, London, March.

Invest UK (2003), Invest UK: Operations Review 2003, Invest UK, London.

Jones, J. and Wren, C. (2003), ‘Inward Investment and the Regional Economy’, unpublished manuscript.

Jones, J. and Wren, C. (2004), ‘Inward Foreign Direct Investment and Employment: A Project-Based Analysis in North-East England’, Journal of Economic Geography, forthcoming.

King, J. (1990), Regional Selective Assistance, 1980-84: An Evaluation by DTI, IDS and WOID, HMSO, London.

Layard, R. (2001), Welfare-to-Work and the New Deal, Centre for Economic Performance, London School of Economics and Political Science, London.

McCrone, G. (1969), Regional Policy in Britain, George Allen and Unwin, London.

Marquand, J. (1980), Measuring the Costs and Effects of Regional Incentives, Department of Industry, London.

McVittie, E. and Swales, J. K. (2003), ‘Regional Policy Evaluation: Ignorance, Evidence and Influence’, Discussion Paper 03-06, Department of Economics, University of Strathclyde.

Moore, B., Rhodes, J. and Tyler, P. (1986), The Effect of Government Regional Economic Policy, HMSO, London.

NAO (2003), The Department of Trade and Industry: Regional Grants in England, Report by the Comptroller and Auditor General, HC 702, Session 2002-03.

NERA (2003), Review of Economic Rationale and Evaluation Methodology for DTI Regional Grants, National Economic Research Associates, London, February.

PACEC (1993), Regional Selective Assistance, 1985-88: An Evaluation by PA Cambridge Economic Consultants, PA Cambridge Economic Consultants, HMSO, London.

Rees, R. D. and Miall, R. H. C. (1981), ‘The Effect of Regional Policy on Manufacturing Investment and Capital Stock within the UK between 1959 and 1978’, Regional Studies, 15.6, 413-24.

Regional Studies Association (2001), Labour’s New Regional Policy: An Assessment, Regional Studies Association, Seaford, England.

Sumner, M. (1999), ‘Long-run Effects of Investment Incentives’, in Driver, C. and Temple, P. (eds) Investment, Growth and Employment: Perspectives for Policy, Routledge, London.

Swales, J. K. (1997), ‘A Cost-Benefit Approach to the Evaluation of Regional Selective Assistance’, Fiscal Studies, 18.1, 73-85.

Taylor, J. (2002), ‘The Evaluation of UK Regional Policy: How Much Progress has been Made?’, in B. Johansson, C. Karlsson and R. Stough (eds) Regional Policies and Comparative Advantage, Edward Elgar, Cheltenham.

Taylor, J. and Bradley, S. (1994), ‘Spatial Disparities in the Impact of the 1990-92 Recession: An Analysis of UK Counties’, Oxford Bulletin of Economics and Statistics, 56, 367-82.

Taylor, J. and Wren, C. (1997), ‘UK Regional Policy: An Evaluation’, Regional Studies, 31.9, 835-48.

Wren, C. (1994), ‘The Build-up and Duration of Subsidy-induced Employment: Evidence from UK Regional Policy’, Journal of Regional Science, 34.3, 387-410.

Wren, C. (1996), ‘Grant Equivalent Expenditure on Industrial Subsidies in the Post-war United Kingdom’, Oxford Bulletin of Economics and Statistics, 58.2, 317-53.

Wren, C. (2003a), ‘Investment Scale as a Signal in Industrial Assistance Schemes with Employment Objectives’, Economica, 70, 331-52.

Wren, C. (2003b), ‘Informational Rents and Discretionary Industrial Assistance’, paper presented to the Royal Economic Society Conference, University of Warwick, April.

Wren, C. and Taylor, J. (1999), ‘Industrial Restructuring and Regional Policy’, Oxford Economic Papers, 51.3, 487-516.

[此贴子已经被angelboy于2008-3-4 14:18:01编辑过]