In the first of a multi-parter on the Open Data Sweet Shop, I looked at data for school admissions to help parents make choices - http://www.ocsi.co.uk/news/2011/09/01/three-picks-from-the-open-data-sweet-shop-%E2%80%93-part-1/. In this second post, I cover ‘how good are services’.

Open Data Pick 2. How good are services?

From a non-random and non-scientific sample of friends who don’t work in public services, most people take government talk about transparency and open data to mean spending data. The Guardian’s lovely data visualisation of public spending in the UK has ended up on people’s walls, and it’s difficult to avoid the constant media snippets on how councils or other public sector organisations are allegedly pouring money down the drain on things like days out at the races (somehow the actual facts of using Newmarket’s reasonable conference facilities on non-race days didn’t seem to get as much press).

But I don’t just want to feel the quantity, I want to feel the quality too. Simply knowing how much something costs doesn’t help us decide how good that thing is.  And simply publishing how much something costs is just fuel for media articles about how much is being wasted - articles that don’t want to go into all that boring detail about whether it was money well spent or not.

If the public sector is real about using open data to improve quality of services, then it needs to be publishing how good services are. A “scores on the doors” rating for services.  Useful to local voters, users of services, media and indeed councillors/ MPs and officers in local and national organisations. Next time I’m voting, or talking to my councillor, or hearing someone moan about the bleedin’ council and how much money they are wasting, this is the sort of information I’d want at my fingertips.

Surely there’s something out there already, I hear you cry. Well, kind of. The Audit Commission, set up by Michael Heseltine in the 80s “to protect the public purse“, carried out evaluations of local authorities and services including a “Use of Resources” element (how effectively was money spent). The previous government also set out a series of National Indicators (and earlier Best Value Performance Indicators - love the jargon) by which local services and agencies were assessed. (And of course, there is already information on quality for some services such as schools, where pupil exam results and value-added measures are routinely published. But this isn’t the case for most local services.)

But both the Audit Commission and National Indicators are on their way out. Although many won’t mourn their passing, it does leave a gap in knowing how good local services are - and how they compare with services in other areas. And that makes it harder for us (and councillors and senior managers and local media and …) to understand where our council is doing really well, and where it could do better. And that’s a bad thing.

So here’s the suggestion. Local councils and other agencies to open-up the information which they use to decide how good services are. Whether services are run internally or commissioned externally, there should be measures of performance.  These measures tell us both the yardstick by which local services are judged by the council, and how well they are doing against that yardstick.

Publishing this information would be a powerful way of moving the debate on from simply “how much”, to “how good”. And understanding the yardstick by which services are currently evaluated is an important step in widening the debate on how to improve services - and even perhaps involving us all in agreeing the yardstick to judge services.

These sound like topics for next month’s Brighton CityForum. Come along and get involved.

Tom Smith, OCSI.

In this blog, we flag-up three suggestions for the Big Lottery Fund to help ensure maximum impact for the fund, using lessons from previous programmes.

Context – growth and localism

The underlying context for the Coastal Regeneration Fund is the ongoing “coastal debate” which has rumbled on in various forms since the 2007 House of Commons DCLG Select Committee report. The last Government got a bloody nose from the original (somewhat glib) suggestion that coastal areas were pretty much like everywhere else. The Committee respectfully replied that a) this wasn’t really true, and b) that the Government really needed better information on coastal areas.

A new Secretary of State tried to defuse the row by reissuing the Government’s response in a more conciliatory tone, and the announcement of coastal regeneration funding could be seen as the long-delayed victory for the pro-coastal side of that particular argument. But the central questions remain unanswered: is there a ‘coastal factor’ that distinguishes coastal areas from those inland? And if so, what should we do different ?

These questions are particularly pressing in the context of the Coalition’s Growth Agenda. The Government wants private sector employment to grow in areas with a high proportion of public sector employment. That’s a big issue for coastal areas, because work shows that of the top ten cities likely to be worst affected by public sector job cuts, nine are coastal. Inland places fare dramatically better: of the ten least affected cities, only one is coastal (assuming that we do not consider London to be a coastal city).

The new Coastal Regeneration Fund will have to thread a way through these issues if we are going to get really effective spending that helps make a difference in coastal areas. In this blog, we have pulled out three lessons from past experience that might act as a guide.

1)    Set clear objectives

Set out early and clearly what the fund is, and isn’t, for. Keep the goalposts fixed - don’t change what the programme is all about half way through.

Lessons learnt from community-level regeneration programmes run by Regional Development Agencies such as the East of England Investing in Communities showed it was important to establish core principles early, and stick to them.  The Coastal Regeneration Fund will need to understand whether it intends to maximise potential economic growth, ameliorate conditions where deprivation is most concentrated, or spend money in places with the greatest absolute number of deprived people.   This isn’t just splitting hairs: as we have found in developing funding frameworks, decisions here can lead in very different directions, particularly around the spending split between rural and urban areas.

The good news is that there’s a lot of help for those strategic decisions.  On the “what to do?” questions, the Coastal Communities Alliance handbook manages to pull off the trick of being both informative and beautifully written.  Research by Sheffield Hallam very neatly pulls out the key problems in coastal areas.  And new work for the Marine Management Organisation (MMO) by us (Roger Tym & Partners RTP and OCSI) provides a conceptual framework on why coastal areas really might be different from those inland, how to best respond, and how funders can “back success” by building on what could be a renaissance in the prospects for coastal economies.

On the “where to spend?” questions, the Index of Multiple Deprivation can be partnered up with our work for the MMO, which pulls together a typology of coastal areas.  This gives a powerful headline view not just of deprivation, but of coastal environments and populations. The work also plugs the gap in the availability of coastal specific data, by breaking out specifically coastal performance in key data areas.

2.  Really devolve responsibility for delivery

Put local areas in absolute charge of the practical delivery.

Management gurus talk about the importance of a “tight/loose” management style.  That means knowing when to exercise tight control, and when to adopt a “looser” approach that allows for decentralised decision making.   We’ve suggested above that RDAs were frequently quite unfocused (”loose”) on setting out the overarching objectives.  But the flipside was that they were often too controlling (”tight”) on local delivery processes.  We suggest that programme managers for the Coastal Regeneration Fund might want to turn that particular “tight/loose” management approach on its head.

The Big Lottery Fund needs to be clear about the Coastal Regeneration Fund’s objectives, but also match the localism agenda by putting local areas in charge of practical delivery.  This allows local areas to pick up the local context of what it is that they need to do and maximise impact on the ground.

3)    Provide the right central support to reduce overheads

Reduce management costs by cutting out duplicated work.

Whilst local delivery needs to be the order of the day, there are still some “overhead” functions that it is most efficient to provide centrally.  For example, if local areas are going to bid for funding, this there is potential for a lot of duplication and wasted effort around collecting, analysing and mapping contextual and performance data.

To avert this, Fund managers should start discussions with data publishers (such as Nomis and Neighbourhood Statistics) and providers, e.g. to centrally provide datasets and information on local coastal communities that bidders can use to evidence their applications. That way, local communities bidding will waste as little time as possible when applying to the fund and managing local delivery.

The bottom line

So, we have three suggestions for the Coastal Regeneration Fund:

1. Set clear objectives: Set out early and clearly what the fund is, and isn’t, for.  Use the work that’s already there to help you.
2. Really devolve responsibility for delivery: Put local areas in absolute charge of the practical delivery.
3. Provide central support: cut duplication and wasted effort by providing the right information centrally.

It’s going to be a tall order, but if the Coastal Regeneration Fund can use these lessons to deliver against the Localism and Growth Agendas, then it could be a really useful tool.  There are good reasons to think that some coastal areas might be in for an economic renaissance, and the Coastal Regeneration Fund could be very much part of that success.

Andy Clarke (RTP) and Tom Smith (OCSI)

It has been a key government ambition in recent years to help the most deprived areas ‘close the gap’ against national averages. However, there has been a lack of good labour market data to identify the success of these programmes.

Recently-published labour market data can be used to identify compare labour market patterns by level of deprivation, including analysis of how these areas are changing over time. A quick overview of the data shows:

• People living in the most deprived areas are still more than twice as likely to be unemployed as the average across England (17% compared to 8%), and more than 4 times more likely than those in the least deprived areas. Employment rates in the most deprived areas are one-third lower than in the least deprived areas (52% compared to 78%). Qualification levels also vary significantly, with those in the most deprived areas more than twice as likely as the average to hold no qualifications at all (and one-half as likely to have a degree).
• Unemployment increased rapidly since the recession in deprived and non-deprived areas alike. Although the least deprived areas have experienced a larger percentage increase in unemployment, the most deprived areas have seen a larger increase in raw numbers (ie, numbers of people). And the gap between the most deprived and least deprived areas has got larger since the start of the recession.

Background

The excellent Nomis (www.nomisweb.co.uk) recently started publishing labour market data by level of deprivation, and I have been having a brief look at what this data shows.

The Annual Population Survey (APS), or Labour Force Survey as was, provides data on employment, skills and unemployment over time, and is usually published to Local Authority level. The sample size is too small to produce neighbourhood level data (eg at Super Output Area), so it is not possible to use the published data to analyse how unemployment or employment levels are changing in the most deprived areas. As a result, workless benefits data (Jobseeker’s Allowance and Incapacity Benefit/ Employment Support Allowance) are often  used instead, but these only count those people eligible for benefit - so underestimate the local level of unemployment (for example, in July 2011 there were just over 1.2 million on unemployment benefit, but the most recent APS data shows nearly 2 million unemployed).

The Annual Population Survey data published on Nomis has been aggregated by the level of deprivation, based on the Index of Multiple Deprivation 2010 [1].  It is now possible to compare how deprived areas fare on key labour market indicators compared with the rest of the country. Important APS datasets include:

• Employment and unemployment rates
• Skill levels
• Occupation groups
• Employment status (part time, full time, self employment).

Labour market indicators by level of deprivation

The charts below compare levels of employment, unemployment, degree-level qualifications, and no qualifications, by the level of deprivation for all areas across England. In each case, there is a strong gradient showing that the most deprived areas fall well behind the average, with the least deprived areas performing well above average. (data is taken from APS 2010 data, linked to the Index of Multiple Deprivation 2010).

Employment rate by level of deprivation

Unemployment rate by level of deprivation

No-qualifications rate by level of deprivation

Degree-level qualifications rate by level of deprivation

Working-age people living in the most deprived areas are more than twice as likely to be unemployed as on average: The unemployment rate in the most deprived 10% of areas is 17%, compared with 8% across England as a whole (and 4% in the least deprived areas). By contrast, the proportion of people in employment is lower in the most deprived areas with just over half (52%) in employment, compared with 70% across England as a whole (and 78% in the least deprived areas).

People living in the most deprived areas are more than twice as likely to hold no qualifications at all, and half as likely to hold degree level qualifications: Skill levels are generally lower in more deprived areas, with one in four people in the most deprived 10% of areas holding no qualifications (more than double the 11% average across England as a whole). Only 15% in the most deprived areas hold a degree level qualification, compared with the national average of 31% (and 43% in the least deprived areas).

Labour market trends by level of deprivation

The APS is updated quarterly (with the most recent data covering the period January to December 2010), so it is possible to track change over time. This time series analysis can be used to see how unemployment levels are changing in deprived areas and whether these areas are closing the gap or falling further behind the rest of the country, particularly in the context of the recession.

The two charts below show how the unemployment rate changed over the period 2004 to 2010. The three trend lines show data for a) the most deprived 10% of areas (dark solid line); b) England average (dark dashed line); c) the 10% least deprived areas (light solid line). The first chart shows the unemployment rate, and the second shows the unemployment rate ‘baselined’ to the 2004 values (where a value of 200% would indicate that unemployment has doubled since 2004, and a value of 50% would indicate that unemployment had halved).

Unemployment trends for most and least deprived areas in England

Changes in unemployment since 2004 for most and least deprived areas in England

The charts show that unemployment has risen across deprived and non-deprived areas alike between 2004 and 2009. This rise was fairly gradual until the end of 2007, but then rose rapidly in deprived and non-deprived areas as a result of the recession. Since 2009, unemployment levels have fallen somewhat (particularly in less deprived areas).

The APS data is smoothed over 4 quarterly time points, so it is difficult to use this data to pinpoint exactly when the 2007/8 downturn started affecting unemployment levels. However, it is clear from the charts that the first sharp increases appeared in the January 2008 to December 2008 dataset, which would be the first time that data from the Oct-Dec 2008 quarter was included (as a reference point, Lehman’s Bank collapsed in September 2008).

Key points from the data are:

The most deprived areas saw smaller-than-average increases (in relative terms) in unemployment rates during the recession. The rise in unemployment levels was larger across the least deprived areas during this period (albeit from a lower base); with the unemployment rate in 2010 65% higher than in 2004. The most deprived areas saw a smaller relative rise in unemployment over the period, with the unemployment rate in 2010 40% higher than in 2004.

However, deprived areas saw the greatest absolute increases in unemployment during this period, leading to a larger gap between the most deprived areas and the average. In 2004 the unemployment rate in the most deprived areas was 11.8%, compared with 2.3% in the least deprived areas (a gap of 9.5%). By 2010 the unemployment rate in the most deprived areas was 16.6%, compared with 3.8% in the least deprived areas (a gap of 12.8%). This gap has been increasing steadily since 2005, with a sharper increase at the start of the recession (2007/08).

This shows that while unemployment increases have been greatest in relative terms in the least deprived areas, they have been greatest in absolute terms in the most deprived areas.  In other words, while the least deprived areas have experienced a larger percentage increase in unemployment, the most deprived areas have seen a larger increase in raw numbers (people). And the gap between the most deprived and least deprived areas has got larger since the start of the recession.

Stefan Noble, OCSI.

---

I’ve tried hard and managed to restrict myself to only 3 picks from the sweet counter. I’ve also avoided the easy answer “open up everything and let us decide”, and gone for things that (I hope) are reasonably plausible. And finally, I’ve gone for things that I think will make a difference. Data that we can use across the city to change things, not just sit in a data store gathering digital dust. And I hope that some of these could be useful starters-for-10 for the October CityForum meeting in Brighton.

In this first of a 2-part blog, I’m looking at data for school admissions to help parents make choices.  A follow-up part will look at information on value-for-money, and understanding detailed patterns of deprivation across the city.

Open Data Pick 1. School choices - helping parents

It’s exam results season again, with the usual excitement and scramble for university places. As a parent of a 3 year-old, that all seems a long way off. By contrast, the application for school places is frighteningly real. From today (1st September) onwards, the parents of 6,000 children across the city will be wading through documents galore as they try to work out what schools they have any chance of getting into, and how those schools compare on Ofsted inspections and exam results.

So why do I think open data could help? Because currently the system is a maze. Without going into the gory details, getting hold of the available information involves phone-calls with the (very helpful) council school admissions team, reading many PDF documents, copying and pasting between Excel spreadsheets, following-up on rumours from friends that “such-and-such information” might be available, estimating probabilities over time, and numerous Google searches.

For a researcher this is bread & butter stuff (although I still have an uneasy feeling I’ve missed something critical) but not all parents will be so used to working with spreadsheets. As one possible future scenario is that open data could actually increase exclusion for those groups with less time or knowhow to find out what they need (Michael Gurstein has a couple of good posts on the ‘data divide’ here and here), it’s important we make sure that published data can easily be digested by all groups.

The information needed by parents includes the following list, which is all currently available to parents but from multiple sources.

1. Home-school distance for all schools: For admission to primary schools, the walking distance (not straightline distance) from your home to each school is the critical piece of data in Brighton & Hove. Currently you either ring up the Council school admissions, or fill in this web-form. As the admissions team presumably has a bit of GIS kit that gives them the answers (rather than working it all out from maps and bits of string), this would need to be opened up via some kind of API, or perhaps pre-calculated for all postcodes and schools in Brighton. [Note, the School Map website has location of all schools across England, with estimated straight-line distance from your postcode -  but Brighton uses walking distance not straight-line.]
2. School application information, including location, application forms, open day dates (or visitor details) etc. This is currently held on school websites and PDF documents.
3. Previous year applications and acceptances data, including number of places; number of applications (and whether the school was 1^st, 2^nd or 3^rd choice); number of places granted; number of applications by status (basically whether they got a place through having brother/ sister at the school, religious status in case of the religious schools,  being close to the school, or other reason).  Currently this is spread across PDF documents, with data for religious schools confusingly in a different location.
4. Home-school distance for previous successful applicants: This is the key bit that tells you whether your child would likely get in to a particular school (assuming they don’t have siblings at the school or other special status). Having this data available by year would allow the website to identify whether an application from your address would have likely got in on each of the previous years.
5. School exam results over time: Including any value-added (progress) scores, and breakdowns by free-school meals or other standard indicators. (currently available on the Department for Education website as a postcode-based search).
6. Ofsted school inspections reports: Currently available on the Ofsted website from a postcode-based search for Ofsted reports - http://www.ofsted.gov.uk/inspection-reports/find-inspection-report.

So what could make this information easier to navigate? Basically I want a single place to go to get all this information, where the information is well explained and easy-to-use. And critically I want to be able to input my address (or postcode) and have the website show me the appropriate schools - and whether or not I would have been successful in previous years.

If each of these datasets were published and referenced by school code, they could be linked-up relatively straightforwardly, put on a nice interactive map front-end, and updated annually from easily published info.

There are commercial organisations out there that aim to do this sort of thing for local authorities, but they’re fairly expensive and have fairly low take-up (I’ve only come across a handful of LAs using these). There are also some free services to parents such as the Hertfordshire School Guru. I’d take a punt that something similar in Brighton could work very well.

Tom Smith, OCSI.

And if you’re looking for commercial applications of open (freely available) data, census data is one of the biggest successes. Lots of commercial datasets and services rely heavily on census data, including the ACORN and MOSAIC user segmentations.

Which is why I’m excited about the arrival of Census 2011 datasets next year. The latest news is that the raw data has arrived from Lockheed Martin (presumably in big shiny steel briefcases padlocked to the wrists of very burly statisticians), and ONS are hard at work validating it all - with Census Director Glen Watson identifying coverage of 94% of the UK population (although some local authorities may reach as low as 80%, which suggests that coverage for particular groups will again be well below average).

So what is coming out and when? The provisional running order (download from the Census output page) is:

• Data table definitions: September 2011
• Final timetable for outputs: March 2012
• First data release (July 2012): UK, region and LA population totals by age and gender.
• Second release (not dated): Key Statistics and Univariate Tables from the smallest area (Output Areas) upwards. These correspond to the 2001 KS and UV outputs. [UPDATE: Rumours are that KS and UV tables will start to surface around October 2012]
• Third release (not dated): Multivariate data from Output Area upwards. These correspond to the 2001 CS and CT outputs. [UPDATE: Rumours are that CS and CT tables will not appear until 2013]
• Fourth release (not dated): Multivariate Standard Tables from ward level upwards. These correspond to the 2001 ST outputs.
• Subsequent releases: Migration tables,  workplace tables (ie data based on workplace rather than residential), detailed ethnic/ religious group outputs.

In other words, second half of 2012 is when it all gets going (I’ll update the dates when these are published) - just in time for the Olympics.

(Part of this is re-used from comments I made on a blog from Emma Mulqueeny).

On a first skim, the consultation has lots about getting more data out - proposing an enhanced right to data, and a presumption that public bodies (and public service providers) will need to publish data. All well and good, but there appears to be less consultation on what is needed to help ensure that the data is used for service improvement or transparency (although there is a useful appendix of UK and international examples).

As local government and others have found, it’s not just a matter of getting the data out there. And we know from collaborating on DataBridge that more support is needed for voluntary sector and local communities to make good use of the data. We will be going through the consultation properly soonish, and will post up our  thoughts here.

And responses are due October 27th, so get typing!

Visuals:

• Guardian map at Super Output Area level: http://www.guardian.co.uk/news/datablog/2011/mar/29/indices-multiple-deprivation-poverty-england#zoomed-picture
• Maps of England from Alasdair Rae, including cartograms: http://undertheraedar.blogspot.com/2011/03/indices-of-deprivation-2010.html
• Coventry: http://ias.facts-about-coventry.com/IAS/dataviews/report/fullpage?viewId=236&reportId=234&geoId=7&geoSubsetId=199&geoReportId=7774
• Cornwall: www.cornwall.gov.uk/deprivationmap
• Nottingham: http://www.nottinghaminsight.org.uk/IAS/dataviews/report?reportId=383&viewId=992&geoReportId=26806&geoId=5&geoSubsetId=
• Change over time in London: http://spatial.ly/eEGcEO

Comments:

• Guardian DataBlog: http://www.guardian.co.uk/news/datablog/2011/mar/29/indices-multiple-deprivation-poverty-england
• OCSI: http://www.ocsi.co.uk/news/2011/03/24/headline-results-from-the-indices-of-deprivation-2010/ and http://www.ocsi.co.uk/news/2011/03/24/why-the-imd-is-still-important-in-the-open-data-age/
• Centre for Cities: http://centreforcities.typepad.com/centre_for_cities/2011/03/the-index-of-multiple-deprivation.html
• Regeneration & Renewal: http://regenandrenewal.blogspot.com/2011/03/digested-2010-indices-of-multiple.html

Datasets

• Report and download datasets: http://www.communities.gov.uk/communities/research/indicesdeprivation/deprivation10/
• Open data: http://opendatacommunities.org/

Overall, the IMD 2010 shows broadly similar results to the older IMD 2007 - areas that were deprived in 2007 are still in the main those that are highly deprived in 2010 (with a correlation of 0.986 between the 2 timepoints). Overall 66% of areas in England are in the same ‘decile’ (10% band) of the IMD 2010 as they were in 2007.

Similarly, there is little change at the top of the rankings, with the top 2 most deprived Local Authorities on IMD 2010 the same as in 2007. The top 5 most deprived LAs are (2007 ranks in brackets):

1. Liverpool (1)
2. Hackney (2)
3. Newham (6)
4. Manchester (4)
5. Knowsley (5)

However the general agreement between the 2010 and 2007 datasets masks a dynamic picture of deprivation patterns for local areas, including:

Highly deprived areas that have got significantly more deprived (relative to other LAs)

• Blackpool moved from 12^th most deprived on IMD 2007, to 6^th on IMD 2010
• Burnley (21 to 11)
• Waltham Forest (27 to 15)
• Hastings (31to 19)

Large increases in relative deprivation for coastal areas (major seaside resorts and ports)

• 54 of the most deprived 100 small areas (Lower Super Output Areas or LSOAs) are in coastal LAs, including the most deprived LSOA in England, in Jaywick in Essex
• Blackpool moved from 12^th most deprived LA on IMD 2007, to 6^th on IMD 2010
• Hastings (from 31 to 19)
• Thanet (from 65 to 49)
• Torbay (71 to 61)
• Brighton and Hove (79 to 66)
• East Lindsey (88 to 73)
• Portsmouth (93 to 76)
• Eastbourne (104 to 84)
• Scarborough (97 to 85)
• Tendring (103 to 86)
• West Somerset (103 to 90)
• Weymouth and Portland (127 to 94)
• Shepway (123 to 97)

Relative improvement in Inner London:

• At regional level, London saw the biggest fall in numbers of most deprived LSOAs, with 80 fewer LSOAs in 2010 identified as highly deprived than in 2007
• Tower Hamlets moved from 3^rd most deprived LA on IMD 2007, to 7^th on IMD 2010
• Islington (6 to 12)
• Lambeth (19 to 29)
• Southwark (26 to 41)
• Camden (57 to 74)
• Westminster (72 to 87)

We’ve been hard at work providing the data and visualisations to our Data Packs users, exploring why the IMD2010 is still important, and identifying the headline national and local messages. For further analysis of the Indices and other key social and economic data, please get in touch!

What is the IMD 2010?

The Indices of Deprivation attempt to measure a broad concept of ‘multiple deprivation’, made up of several distinct dimensions, or domains, of deprivation. The data is based on 38 separate indicators across seven domains: Income, Employment, Health and Disability, Education Skills and Training, Barriers to Housing and Other Services, Crime and Living Environment. As well as the overall IMD 2010 and domains, data for the underlying indicators has also been published.

Commissioned by CLG from the Social Disadvantage Research Centre at Oxford University, the Indices of Deprivation are an update of previous datasets in 2007, 2004 and 2000. The full datasets and technical report are available, see links at the end of this article.

There is lots of data out there. Why should we be interested in the IMD 2010?

It is true that the data landscape has changed massively since the last IMD was published in 2007, with thousands of new datasets available through data.gov.uk, the London Data Store , our own Data4nr and many others.

However, along with the Census, the IMD is a key dataset for targeting services to help tackle deprivation. The IMD provides a single overview indicator of how all English areas compare on levels of deprivation (separate measures are available for Wales, Scotland and Northern Ireland. Also, the same methodology has been used in South and Southern Africa by the Oxford University team responsible for the IMD).

Government programmes - both national and local - aiming to tackle deprivation often use the IMD to target funding to the most deprived areas - put simply, there is no other robust indicator that captures the wealth of issues covered by the IMD. The Oxford University team has identified that as much 1% of all government spending was allocated using the IMD. Although this analysis is now a few years old (I’d be very interested if anyone has looked at this more recently), there are some striking examples of where the IMD has been used to target government resources:

• Targeting regeneration programmes: Including the Neighbourhood Renewal Fund (NRF), Single Regeneration Budget (SRB), Neighbourhood Management, and programmes to increase enterprise in disadvantaged areas. In addition, the 2000 Spending Review funding for all domestic regeneration programmes used the IMD (eg, in 2002-03 £430 million of the £2.55 billion single pot budget to RDAs was allocated based upon the IMD)
• Sure Start and Children’s Centres: The location of the initial waves of Sure Start centres were based on the most deprived areas according to the IMD, as was funding for the Neighbourhood Nurseries Initiative and other programmes aimed at supporting vulnerable children and families
• Many of the National Lottery grants are explicitly targeted at the most deprived areas based on the IMD, as are other funds eg Bill and Melissa Gates Foundation gifts for the provision of information technology learning centres.
• Stamp duty on property and land transactions were reduced in deprived areas , based on the IMD.

The bottom-line is that the IMD is an important measure of how local areas compare with others on a comprehensive basket of deprivation indicators - so provides a key input to understanding “need” for service commissioning.

What are the limitations of the Indices?

Although the IMD 2010 and other datasets published today are an important piece of the data puzzle, there are limitations:

• Not a direct measure of deprivation: The IMD score is used to compare (rank) areas, but is not a direct or meaningful measure of deprivation, in the way that eg the ‘unemployment rate’ is a real measure of the proportion of people out-of-work. An area with a score of 50 is not twice as deprived as an area with a score of 25.
• Change over time is relative to other areas: The IMD 2010 can be compared with older datasets to identify how areas are changing over time relative to other areas across England. However, it is not a direct measure of whether areas are “improving” or “closing the gap” against the average - so we need to be a bit cautious about using to identify trajectories over time. The difference in 2007 and 2010 measures is down to how an area has fared compared to areas across England with similar levels of deprivation. A local neighbourhood could well have improved in real-terms (eg lower levels of unemployment, higher incomes, higher skill levels, lower crime rates, better environment and so on),  and may have improved faster than the average. However if other areas with similar levels of deprivation have done slightly better, the local neighbourhood will score as more deprived in 2010 than 2007. The message is you need to use direct measures alongside the IMD to understand how local areas are changing.
• Older data: The IMD2010 is based on data from 2008 (with one indicator still based on Census data). In many cases this is the most recent data available at neighbourhood level - timeliness of data is still (and always was, always will be?) an issue.
• Area-based measures are only one part of the story: Many groups argue that the IMD2010 - and other area-based measures which essentially measure concentration - are biased against rural areas, where deprivation is more scattered. Our own analysis has highlighted that although only 2% of the most deprived areas in England are rural, 17% of all households in poverty (earning less than 60% of the median income) in England are in rural areas.

Links to data, comments and analysis:

• Datasets and technical report: http://www.communities.gov.uk/communities/research/indicesdeprivation/deprivation10/
• Twitter news: http://twitter.com/search?q=%23IMD and http://twitter.com/search?q=%23IMD2010
• Headline findings from OCSI: http://www.ocsi.co.uk/news/2011/03/24/headline-results-from-the-indices-of-deprivation-2010/
• Let us know about additional analysis and we’ll add it here.

We’ve been hard at work providing the data and visualisations to our Data Packs users, exploring why the IMD2010 is still important, and identifying the headline national and local messages. For further analysis of the Indices and other key social and economic data, please get in touch!