Welfare reform in the UK 2010-16: Expectations, outcomes and local impacts

Welfare reform has been central to UK government policy since 2010. This article compares initial expectations with key outcomes by 2016. The article shows that although the financial savings to the Treasury have been large, they have been rather less than the government first anticipated, mainly because the reduction in spending on incapacity‐related benefits has proved far smaller than 
expected. The financial losses have also been spread highly unevenly across the country, and the evidence from a pilot study in Scotland suggests that the reforms have had little impact on levels of worklessness. The article concludes that whilst forecasting the financial savings from welfare reform is an inherently uncertain activity, the United Kingdom's reforms should be understood first and foremost 
as about reducing public spending in the poorest places.

The article draws on quantitative data assembled in a number of studies by the authors. Many of the figures in the report are estimates, but in every case they are deeply rooted in official statistics-for example, in the Treasury's own estimates of the financial savings, the government's Impact Assessments, and benefit claimant data.
Welfare reform is a contentious issue. Throughout the period since 2010 the dominant motivation of government has been to reduce the United Kingdom's budget deficit, but an ideological agenda that is based on the need to reduce welfare dependency and incentivize paid employment has also been central to the aims of the reforms. This has contributed to political, policy, and contemporary debates which have sought to stigmatize those in receipt of benefits, blame individuals for their predicament, and ignore the structural factors at play. In documenting the impacts, the article does not attempt to comment on the merits of the overall strategy or of each of the reforms. Nor is it possible in a short article to examine all the impacts of the reforms, including for example the financial losses to different income groups and household types. The article does, however, make no apology for considering the impact on different places because this is a dimension that is too often overlooked, and is a key tool in monitoring labor market impacts.

| THE WELFARE REFORMS
The ten major welfare reforms implemented in the United Kingdom between 2010 and 2015 and covered by this article are: • Housing Benefit-Local Housing Allowance Changes to the rules governing assistance with the cost of housing for low-income households in the private rented sector.
• Housing Benefit-under-occupation in the social rented sector New rules governing the size of properties for which payments are made to working-age claimants (widely known and also referred to in this article as the "bedroom tax").

• Non-dependant deductions
Increases in the deductions from income-based benefits to reflect the contribution that non-dependant household members are expected to make towards the household's housing costs.

• Benefit Cap
A new ceiling on total payments per household, applying to the sum of a wide range of benefits for out-of-work claimants of working age.

• Council Tax Support
Reductions in the entitlement of working age claimants arising from a 10% reduction in total payments to local authorities.

• Personal Independence Payment
Phased replacement of Disability Living Allowance (DLA) by Personal Independence Payment (PIP) for workingage claimants, including more stringent and frequent medical tests.

• Employment and Support Allowance
Replacement of Incapacity Benefit (IB) and related benefits by Employment and Support Allowance (ESA), with more stringent medical tests, greater conditionality, and time-limiting of non-means tested entitlement for claimants in the Work-Related Activity Group.

• Child Benefit
Three-year freeze and withdrawal of benefit from households with a higher earner.

• Tax Credits
Reductions in payment rates and eligibility for Child Tax Credit and Working Tax Credit paid to lower and middle income households, including increase in working hours requirement for Working Tax Credit.
• 1% up-rating Limit in annual up-rating of value, for three years for most working-age benefits and for two years for Child Benefit and the Local Housing Allowance element of Housing Benefit.
A fuller description of each of the reforms, including the timing of implementation, is included in the Appendix.
The majority of these reforms were initiated by the Conservative-led coalition Government, but the introduction of ESA was a Labour measure that pre-dated 2010 and only took full effect later, whereas the time-limiting of non-means-tested ESA entitlement was a coalition innovation. The full impact of both the introduction and time-limiting of ESA is included here to provide a comprehensive view of the impact of the reforms implemented from 2010 onwards. By March 2016, nearly all these reforms had come into full effect. The important exception is the changeover from DLA to PIP, which the government does not expect to be completed until at least 2018.
An astute observer will note the omission of Universal Credit, which is scheduled to replace just about all means-tested working-age benefits, and is administratively the most ambitious welfare reform of all. However, the introduction of Universal Credit differs from the other reforms listed here. Universal Credit is best understood as a repackaging of existing benefits that for the first time introduces a consistent withdrawal rate, with the rules governing eligibility carried over from the existing benefits it replaces (Department for Work and Pensions, 2010).
In the original plan, Universal Credit modified the exact value of the entitlement of most individuals and households, but was intended to reduce overall spending only by making it financially worthwhile to move into work. In practice, revisions to the "Work Allowances" within Universal Credit, announced in July 2015 to take effect for new claimants from April 2016, mean that for many individuals and households Universal Credit will now be paid at a lower rate than the benefits it replaces. More importantly in the present context, the introduction of Universal Credit has proved slow.
By March 2016, there were only 230,000 Universal Credit claims, nearly all unemployed single people.
The list of welfare reforms also omits benefit sanctions. What needs to be remembered here is that the power to impose benefit sanctions is not new. What happened after 2010 is that sanctions were more widely applied, especially to those claiming Jobseeker's Allowance (JSA), although the numbers and the rate now appear to have peaked (Webster, 2016). The National Audit Office (2016) estimated that during 2015 the application of sanctions reduced welfare spending by £97 million. A further omission from the list is Income Support for lone parents. The qualifying age of the youngest child was reduced in 2011 from under seven to under five, but the effect is to transfer the lone parent onto JSA at the same payment rate.
The list also excludes the changeover from RPI to CPI 1 for benefits uprating, which was introduced in 2011 but is part of a much wider accounting reform, including for example all public sector pensions. In contrast to the Treasury, we have therefore refrained from adding the impact of this particular changeover to the list of savings to the Exchequer (and losses to claimants) arising specifically from welfare reform. Additionally, the RPI to CPI changeover was superseded for most working-age benefits by the 1% uprating cap introduced in 2013. 2

| THE EXISTING EVIDENCE BASE
Much of the previous research on the United Kingdom's welfare reforms has focused on the impact on particular groups or the impact of individual measures. Distributional analyses have found that those in the lowest income decile have been affected most, even when positive changes in tax allowances and the minimum wage are taken into consideration (Finch & Whittaker, 2016;Hood & Johnson, 2016). Analyses by gender highlight the disproportionate impact on women (MacLeavy, 2011;Scottish Government, 2013). Families with dependent children are also hit hard (Browne & Elming, 2015). Unsurprisingly, research confirms that, in most cases, the reforms increase the financial incentive to work (Adams & Browne, 2013;Hirsch & Beckhelling, 2011). The UK government's Social Security Advisory Committee (2014) reviewed the evidence on the cumulative impact of the reforms, but not in a systematic, quantitative way Turning to the impact of individual measures, there has been extensive research into the impact of the reforms to Housing Benefit. This includes the impact on landlords and tenants in the social rented sector (Power, Provan, Herden, & Serle, 2014;Williams, Clarke, & Whitehead, 2013) and in the private rented sector , notably in London and amongst young single people (Beatty, Cole, Foden, & Powell, 2014). Specific attention has also been paid to the impacts of the "bedroom tax" (Clarke et al., 2015) and the Benefit Cap (Department for Work and Pensions, 2014).
The anticipated impact of the welfare reforms on local areas across the country was initially documented by Beatty and Fothergill (2013), and subsequently by Wilson, Morgan, Rahman, and Lovedeep (2013). Both studies pointed toward larger financial losses in poorer areas and in the parts of London hit hard by the Housing Benefit reforms. Figures for Wales (Welsh Government, 2014) confirmed the impact on poorer areas. Studies in Scotland (Beatty & Fothergill, 2014a) and in Sheffield  drove down the estimated impacts to neighborhood level.
In contrast, the potential divergence between expectations and outturns has so far received little attention. The UK government itself has steered clear of looking back at how its welfare reforms are matching up to expectations. It publishes forecast savings when policy changes are first announced, and from time to time it publishes revised forecasts, but setting one against the other is not normal practice.
The figures published by Beatty and Fothergill (2016), which are the starting point for the present article, were the first to compare expectations and outcomes across the package as a whole. The Institute for Fiscal Studies (Emmerson, Johnson, & Joyce, 2017) subsequently argued that, by 2015-16, welfare spending in the United Kingdom was lower than it would have been otherwise, but shared Beatty and Fothergill's assessment that the savings have been less than expected. The Office for Budget Responsibility (OBR)-the official watchdog on Treasury spending and forecasts-also subsequently published estimates of the reductions in welfare spending through to 2015-16 that differ in detail, but not in the broader conclusions (OBR, 2016). Table 1 shows the annual financial loss to claimants, by March 2016, arising from each element of the welfare reforms listed above. Details of the data sources and methods are included in the Appendix, and set out in greater detail in Beatty and Fothergill (2016).

| THE FINANCIAL LOSS TO CLAIMANTS
The first column shows the financial losses to claimants originally anticipated. These are taken from Fothergill (2013, 2014b), and are mostly theTreasury's own figures, published when the reforms were first announced.
TheTreasury's initial forecasts were generally for financial savings through to 2014-15. The initial forecasts have therefore been adjusted where further increases in the number of affected claimants were anticipated by 2016 (e.g., from the ESA reforms) and where it became clear that there would be a divergence between the savings to the Treasury and the loss to claimants (e.g., in Council Tax Support). The initial forecasts have also been adjusted to reflect exemptions or modifications (e.g., to the withdrawal of Child Benefit from higher earners) announced prior to implementation. Statements. These Treasury figures are all revised forecasts rather than outturn savings, and necessarily, therefore, still subject to a margin of error. In a number of other cases the losses have been estimated using outturn data, for example, on the number of ESA claimants, on the numbers affected by the Benefit Cap and the "bedroom tax", and (in the context of up-rating) on the outturn rate of inflation.

| Overall losses
There are important differences between what was originally anticipated and what by March 2016 had proved to be the outturn. At just under £14.5 billion a year, the loss to welfare claimants is very large, but down on the original forecast. Overall, the loss to claimants by March 2016 is estimated to have been £4.3 billion a year less than forecast.
The subsequent estimates by the OBR (2016) confirm the magnitude of the outturn loss to claimants. It is difficult to draw precise comparisons with the figures in Table 1 because there are differences in methods and scope. However, the OBR estimates that the reduction in spending on working-age welfare and on child benefits sums to £18.1 billion a year by 2015-16. Deducting the impact of the changeover to CPI for uprating (an estimated £4.2 billion a year excluding the impact on the second state pension) and adding back in the £370 million a year reduction attributable to Council Tax Support (which the OBR excludes as a local scheme) brings to OBR outturn figure to £14.3 billion a year, quite close to the total in Table 1. The omission of the changeover to CPI for uprating also accounts for most of the divergence from the £16.7 billion estimate published by the Institute for Fiscal Studies (Emmerson et al., 2017).
Two of the reforms have delivered bigger savings to the Treasury than the loss to claimants. In Scotland and Wales, the reduction in Council Tax Support has not been passed on to claimants, and in some English local authorities as well the reduction has not been passed on in full or in part. This has reduced the loss to claimants by £120 million a year. 3 In Scotland, the impact of the "bedroom tax" has been fully averted using Discretionary Housing Payments-a saving to claimants of around £50 million a year. 4 In both cases, the reduction in welfare spending by the Treasury has resulted in reductions in other budgets within the devolved administrations and/or local authorities.
What is also apparent from Table 1 is that the biggest financial losses to claimants have not been those that have attracted the greatest publicity or controversy. Arguably, the "bedroom tax" and the Benefit Cap are the two post-2010 welfare reforms that have generated greatest attention, partly because they were entirely new and partly because they have resulted in substantial losses for specific households. But in the overall jigsaw of welfare reform they account for relatively modest sums-£360 million and £100 million a year, respectively. The big reductions have arisen from changes to Tax Credits (£4,210 million a year), Child Benefit (£3,030 million) and the 1% uprating (£2,700 million), all of which have affected very large numbers of individuals and households.

| Employment and Support Allowance
The reduction in spending on incapacity-related benefits-these days, ESA-accounts for by far the largest shortfall between expectations and outturn. The original estimated loss to incapacity claimants of £4,350 million a year was the largest single reduction in welfare spending and comprised two elements-an estimated £2,600 million a year reduction in spending due to the extension of means testing and a further £1,750 million a year arising from the measures initiated by the pre-2010 Labour Government. The Treasury never published forecasts of the financial losses arising from the pre-2010 Labour measures, so these were estimated on the basis of the numbers expected to be affected. In practice, the outturn financial loss to claimants arising from the reforms to incapacity-related benefits is estimated to be £650 million a year, far below what was originally expected. The OBR (2016) confirms this assessment. It estimated that the changeover to ESA reduced spending by £400 million a year by 2015-16 and the introduction of means-testing by just £200 million a year-a very similar combined total.
Four factors appear to explain why the savings to the Treasury arising from the reform of incapacity-related benefits have fallen so far short of expectations. First, the new medical test (the Work Capability Assessment) has reduced incapacity numbers by far less than was anticipated, in part because following sustained criticism the test itself has undergone successive revisions. By March 2016, the headline number of incapacity claimants was only down on the 2010 level by around 100,000 (at 2.5 million).
Second, following the medical test a much smaller proportion of ESA claimants have been placed in the "Work-Related Activity Group" than was originally expected. Instead, a higher proportion have been placed in the "Support Group", intended for those with higher levels of ill-health or disability, which attracts a higher payment rate.
Again, this may be attributable in part to revisions to the Work Capability Assessment.
Third, because the numbers in the Work-Related Activity Group are much smaller than the government first anticipated, the savings to the Treasury arising from the time-limiting of non-means-tested entitlement, which applies only to this group, have been much less than expected.

| Other divergences
Four other divergences between expectations and outturns are worth highlighting. First, the financial loss to claimants arising from the introduction of the Benefit Cap proved less than expected-£100 million a year by March 2016 compared to an initial Treasury forecast of £270 million a year. This is mainly because fewer than expected households were affected by the cap-an average of just 28,000 between November 2013 and April 2014. 5 There may be evidence here of behavioral change: faced with a reduction in benefits, some households may have moved into work or reported a change in circumstances. If this has not led to other offsetting welfare claims, for example by displacing someone else onto benefits, the Treasury may in practice have made greater savings than the financial loss to claimants.
Second, the 1% uprating of benefits did not lead to quite the savings theTreasury expected. This is because the annual rate of inflation fell by more than expected so that, in the absence of the 1% limit, benefits would not have been uprated by a great deal more. Indeed, by the final year in which the 1% uprating applied the inflation rate was actually below 1%.
Third, the reduction in Tax Credits delivered greater savings than expected-£550 million a year more by March 2016. The initial forecast and the outturn are both the Treasury's own figures. What the increase would suggest is that higher employment, the slow growth of earnings and the spread of low pay, all of which characterized the United Kingdom's recovery from recession, increased the number of households affected by reductions in Tax Credits.
Fourth, the introduction of PIP has so far failed to deliver the scale of savings to the Treasury that the government initially expected. Indeed, figures subsequently published by the OBR (2016) indicate that this element of the welfare reform package is actually much further off track than the figures presented here suggest. The OBR notes that, by 2015-16, 75% of the working-age caseload remained on PIP's predecessor, DLA, and the savings to the Treasury among those moved to PIP were only around 5% compared to an initially anticipated 20%. As a result, the OBR estimated that in 2015-16 the reduction in spending associated with the introduction of PIP was as little as £100 million a year, and that only a fraction of the originally expected savings to the Treasury will be realized as the new benefit is gradually rolled out in the second half of the decade.

| THE FINANCIAL LOSS TO LOCAL AREAS
Although from time to time the UK government has published local and regional data on individual elements of its welfare reform package, it has never produced estimates of the overall local impact. It is to be expected, however, that the financial losses will vary from place to place, not least because benefit claimants are so unevenly spread across the country. Official statistics do, however, allow the local impact to be estimated. These include the Impact Assessments the government publishes for most elements of the reforms and the claimant numbers and expenditures, by local authority, published by the Department for Work and Pensions (DWP) and Her Majesty's Revenue and Customs (HMRC). Figure 1 shows the estimated financial loss, by local authority across Britain, in March 2016. In order to compare the impact on different places this map shows the annual financial loss per adult of working age. This is the best measure of the intensity of the "hit" in each area and is different to the loss facing each affected individual or household because it also reflects the number of claimants and non-claimants in each area. The focus on adults of working age (16-64) is appropriate because the welfare reforms impact almost exclusively on this group and, by contrast, benefit claimants of pensionable age are largely unaffected. The methods underpinning the local estimates are set out in the Appendix, and maps on the impact of each of the individual reforms can be found in Beatty and Fothergill (2016).
The overall impact by March 2016 presents a seemingly complex picture, but there are clear patterns. Three types of area have been hit hardest: • The older industrial areas of England, Scotland and Wales. These include substantial parts of North West and North East England, Yorkshire, the South Wales Valleys and the Glasgow area in Scotland. Older industrial areas tend to have high numbers on out-of-work benefits and on low wages, which triggers Tax Credits and Housing Benefit as income top-ups.
• A number of less prosperous seaside towns. These too often have high numbers on out-of-work benefits and on low wages, and a large private-rented housing market which Housing Benefit reforms have hit hard.
• Some London boroughs. Some of these are also relatively deprived, but exceptionally high housing costs have inflated the losses arising from Housing Benefit reforms in the private-rented sector.
At the other end of the spectrum, a substantial part of southern England outside London has been much less acutely affected by the reforms. A number of rural areas in northern England, including most of North Yorkshire and parts of Cumbria, plus the Aberdeen area in Scotland, also escaped relatively lightly. It has a high proportion of adults of working age out-of-work on benefits, including one of the highest incapacity claimant rates in the country. It also has a high proportion of households living in the private-rented sector, who have been badly exposed to the reductions in the Local Housing Allowance element of Housing Benefit, and low wages in the local economy have inflated the numbers hit by reductions in Tax Credits. Westminster, at number two on the list, has been hard hit because extremely high rents mean that, more than anywhere else in Britain, the Housing Benefit reforms and the Benefit Cap have led to exceptionally large financial losses for some households. Nevertheless, in all the worst affected local authorities the financial losses by March 2016 are less than first anticipated (Beatty & Fothergill, 2013, 2014b, often by £150-£200 per adult of working age. This is principally because the reduction in ESA numbers and spending has been far less than expected. The reforms to incapacityrelated benefits were always going to cause the largest financial losses in the places where incapacity claimants are concentrated. These are principally Britain's older industrial areas, where a diversion from unemployment to disability benefits has for many years hidden the true scale of worklessness (Beatty & Fothergill, 2005). In the South Wales Valleys, in particular, the "failure" of the incapacity-related benefits reforms eased the financial losses by around £200 per working-age adult.
The worst affected districts have nevertheless still experienced losses that are typically two-and-a-half to three times higher, per adult of working age, than the least affected districts. The "failure" of the incapacity reforms has mattered less in the latter because the claimant rate there has always been relatively low. Of the ten districts least affected by the welfare reforms up to March 2016, seven are in the prosperous south and east of England outside London, and the remaining three are in the parts of Scotland that benefit strongly from the oil industry.

| IMPACT ON THE LABOR MARKET
Welfare reform can be expected to influence labor market behavior by changing financial incentives. For an unemployed person, a reduction in the value of out-of-work benefits increases the incentive to work. For a person in work who receives benefits or tax credits, a reduction increases the incentive to take on extra hours or to find higher paid work in order to maintain their income. The UK government has been quite explicit that its welfare reforms are intended to encourage claimants to move off benefits and into work (HM Treasury, 2010).
But not all welfare reforms work in this way. A reduction in Working Tax Credits, for example, can actually make taking a job financially less attractive. The scale of the financial incentive also varies a great deal. The loss of entitlement to non-means tested incapacity benefit (ESA) can be worth £5,000 a year, whereas the three-year freeze in the value of Child Benefit had a far smaller impact, around £100 a year. More significantly, the idea that welfare reform will trigger additional employment is rooted in a particular view of how an economy works. It assumes that extra labor supply will lead to additional labor demand-that firms will expand to take on the extra workers. At times and in places where an economy is operating at or close to full employment, this argument has some validity. At times of recession, or in places where there is a substantial surplus of labor-a fair description of many of Britain's weaker local economies-the notion that extra labor supply will be smoothly absorbed by additional labor demand is more problematic.
Whether in practice welfare reform leads to lower numbers on benefit and higher numbers in work is therefore essentially an empirical question. The complication is that welfare reform is only one of several things happening simultaneously. In the United Kingdom, economic growth accelerated from mid-2012 onwards after a period of stagnation in the wake of the 2008/09 recession, at much the same time as several of the welfare reforms took effect.
Compared with 2012, employment is now higher and unemployment lower. However, it would be wrong to assume that these improvements can be attributed in whole or in part to welfare reform. Benefit changes may (or may not) have played a role, but other factors such as a revival of consumer spending and borrowing, particularly around the housing market, have also contributed to the upturn.
The big variations from place to place in the scale of the financial losses (see Figure 1) do, however, offer a way forward. Because the United Kingdom's welfare reforms impact so much more in some places than others, if there is a positive impact on the labor market it should be much greater in these places. In each individual town or district, local factors will of course help to shape local benefit numbers-the closure of a major employer for example, or improvements in transport links to a neighboring center of employment. However, by pooling data from a number of places it should be possible to begin to detect the impact of welfare reform. If welfare reform is an important factor, we should expect to observe bigger reductions in benefit numbers, and bigger increases in employment, in the places where the financial losses arising from welfare reform are greatest.
In a pilot study in Scotland, commissioned by the Scottish Parliament, we explored these relationships (Beatty, Fothergill, & Houston, 2015). It is appropriate to begin by looking at the relationship between welfare reform and the out-of-work benefit claimant rate. Changes in the out-of-work benefit claimant rate will reflect revisions to the rules governing entitlement, which since 2010 have applied to incapacity-related claims in particular, though with a rather smaller impact on the headline total than was originally anticipated, as we noted earlier. The out-of-work benefit claimant rate should also reflect moves into work prompted by the new financial incentives created by the welfare reforms. It will also, of course, reflect wider trends in the national and local economy.  That all the fall has been in JSA rather than ESA numbers is in some respects surprising. The welfare reforms have not in themselves reduced entitlement to JSA, though the tougher stance by Jobcentre Plus on sanctions may have had this effect, and many JSA claimants will have experienced a reduction in Housing Benefit and/or tax credits.
The Work Capability Assessment, in contrast, was intended to make access to ESA more difficult for new claimants and to disqualify some former claimants, and the wider application of means-testing to make ESA less generous.
The financial losses attributable to ESA are also substantial-an estimated £85 million a year in Scotland by March 2016.
The potential impacts on economic activity rates and employment rates are harder to monitor because, unlike the benefits data, which is an administrative count, the local authority data beyond 2011 comes from a sample survey and is subject to an important margin of error. The data on the number of jobs in each authority, from the Business Register and Employment Survey (BRES) is more reliable though still affected by errors and discontinuities. Figure 4 shows the absence of a relationship between the change in the number of employee jobs in each Scottish local authority (from BRES) and the financial losses arising from welfare reform.
Labor market trends in Scotland, at least, therefore provide only limited evidence that welfare reform is an important causal factor. The important exception concerns the trend in JSA claimant numbers, which in turn underpins the trend in overall out-of-work claimant numbers. The reduction in JSA numbers is higher in the places where the financial loss from welfare reform is larger. This could be interpreted as evidence that, for JSA claimants at least, welfare reform is working.
The problem with this interpretation is that when unemployment falls during an economic upturn it generally does so more in some places than others, irrespective of welfare reform. In particular, it is always likely that in an upturn the reduction will be greater in the places with initially high unemployment than in more prosperous, low unemployment areas. This is partly because a reduction of, say, 4 percentage points is possible where the unemployment rate starts at 8%, but not where the rate starts at just 3% or 4%. The point here is that economic upturns are normally associated with convergence in unemployment rates and the high unemployment areas are the ones most affected by welfare reform. The more rapid fall in JSA numbers that can be observed in the areas hit hardest by welfare reform may, in practice, have little to do with welfare reform.
Two upturns are never exactly the same, so finding the perfect match for the recent post-recession period is impossible. Figure 5  • February 1998 to November 2004. This is a rather longer period but one in which claimant unemployment fell from broadly similar levels and by similar amounts, though unemployment was already well down on peak levels in the early 1990s. 7 • August 1993 to August 1996. This is a more comparable period in the economic cycle, covering the immediate recovery from recession, but one in which claimant unemployment was falling from a far higher starting point.
Unemployment Benefit had also not yet been replaced by JSA, with its significantly more restrictive entitlement for claimants of more than six months.
To maintain comparability, the horizontal axis on all three diagrams-the financial loss arising from the welfare reforms by March 2016-is held constant. Left to right, each Scottish local authority is therefore at the same place  on the scale in each diagram. The point here is not to suggest that these welfare reforms had any impact on earlier events but simply to examine whether the pattern of change by local authority that can be observed in the 2011-16 period is similar to that in previous upturns.
As will be immediately apparent, there is very strong similarity between the geography of the reduction in unemployment in all three periods. The places in Scotland that have experienced the biggest percentage point reduction in JSA unemployment during the recent upturn, which were also the places hit hardest by welfare reform, are the same places that experienced the biggest reduction in claimant unemployment during previous upturns. This is a key observation. What it suggests is that the reductions in JSA unemployment cannot be attributed to welfare reform. Rather, the big reductions in areas of high unemployment are a normal feature of economic upturns.

| CONCLUSIONS
In the context of the United Kingdom's welfare reform, what emerges is a divergence between expectations and outcomes. The shortfall in financial savings to the Treasury, to March 2016, is arguably the least concerning of these.
Even though welfare rules and regulations are under the direct control of government, the actual level of spending is always going to depend in part on factors such as economic growth and inflation which cannot be accurately predicted. There is also unavoidable uncertainty about how welfare reforms will work out in practice, especially where there is the possibility of behavioral change. Forecasting the financial savings from welfare reform is an inherently uncertain activity and we should not be surprised that government sometimes get things wrong. Even so, the margin by which the reforms to incapacity-related benefits failed to deliver the expected savings is remarkable.
The uneven impact of welfare reform across the country is more concerning. The suspicion here is that because the same benefit rules apply everywhere the UK government never gave much thought to how the reforms would play out in different places. If this was indeed the thinking, it proved to be a serious blind-spot because in practice the financial loss in some places has been far greater than in others. Indeed, the evidence indicates that by and large it is the poorest places that have been hit hardest. There is a certain inevitability in this because if welfare benefits are reduced, the losses will nearly always tend to be concentrated in the places where claimant rates are highest.
That the poorest places are nearly always hit hardest by welfare reform is an uncomfortable reality that needs to be logged by government. Welfare reform, an essentially "national" policy, has distinctly uneven regional and local impacts and tends to work directly against other policy objectives such as the convergence in prosperity and well-being across the country. Indeed, because in the poorest places the financial losses from welfare reform are so large they are often likely to exceed the spending on policies to strengthen local and regional economies. If more out-of-work claimants in disadvantaged areas are to be encouraged to look for work, for example, there is a good case for boosting regeneration efforts to generate more jobs in these places.
The impact on the labor market points to the biggest gap between expectations and outcomes. We need to be clear that the evidence here comes from an exploratory study covering Scotland and that the impact of welfare reform on the labor market is something that deserves much more research. However, if the tentative conclusions from Scotland can be generalized to the rest of Britain there is little evidence that the welfare reforms have delivered lower numbers on benefit and higher numbers in employment. If this is indeed the case and positive labor market outcomes have not in practice been delivered, it indicates that the United Kingdom's welfare reforms need to be understood first and foremost as being about reducing public spending.

CONFLICT OF INTEREST
None declared. • Rents set at 30th percentile of local rents, rather than 50th percentile, from 2011-12 • Caps on maximum rents by property size, with 4-bed limit, from 2011-12 • Abolition of £15 excess for tenants below maximum rent, from 2011-12 Council Tax Support • 10% reduction in Treasury grant compared to previous scheme, from 2013-14.
• Reduction in entitlement only permitted for working-age households

Allocation to local authorities
Based on DWP and HMRC expenditure data by local authority