This stems from health secretary Andrew Lansley’s recent comment that “age is the principal determinant of health need” and that distribution of the £100bn budget for the NHS in England should “get progressively to a greater focus on what are the actual determinants of health need.”

Somewhere along the line, those comments were interpreted by a generally cheesed-off medical profession that Mr Lansley intends to introduce an “age-only” NHS allocation formula, switching substantial NHS funds from, generally younger, Labour-voting constituencies in north to the octogenarians who thrive in the Conservative-voting villages of the south.

It’s a good story, which might even contain elements of the truth, but the reality, as ever, is a little more complicated.

At present, five separate allocation formulae are used to divvy up different bits of the £100bn NHS pot to different areas of England. The largest share – the hospital care budget – is divided up using one formula, while four others – mental health, GP prescribing, health inequalities (more on that in a later post) and maternity – are each allocated using their own separate formula. (Think for a second about the demographics driving the demand for maternity services as opposed to, say, hip replacements, and you will grasp why this makes sense.)

Health economists and statisticians frequently tweak and argue over these formula in order to move, hopefully, ever closer to the Holy Grail: a distribution of health resources which is fairly distributed on the basis of health need.

To that end, the Department of Health is currently considering is a move to a so-called Person Based Resource Allocation formula. The “person based” element here relates to the formula’s use of real patient-level data about an individual’s previous treatments and diagnoses – extracted from GP and hospital records – to predict how much cash that patient is likely to need.

That marks a change from previous formulae which have predominantly distributed funds on the basis of proxy indicators of health need – such the deprivation and age profile of the geographic area, as opposed to data extracted from the actual medical records of each patient in that area.

In theory, the new formula calculates a NHS budget for every patient registered at each NHS GP practice which can then be aggregated up to create a target allocation for geographic areas or groups of GP practices.

So, what impact would this new formula have on the spending of different areas?

Luckily for us, every GP practices’ need assessment under the new formula has already been published by the Department of Health, although it is not yet used for official allocations. We can compare the weighting the new formula gives to each practice to what it got under the current, or “old” formula (for the geeks: the acute element of CARAN formula).

NOTE: Some caution is needed when assessing each of the below charts, as the overall figures do not quite balance to zero and instead show an overall increase in need weighting under the new formula of 0.3 per cent. That is likely due to changes in population make up in between the calculation of the two different formula weightings, as well as the need to exclude some practices where data is not available.

Chart one takes just under 8,000 of England’s 8,300 GP practices (covering over 54m patients) and divides them into quintiles based on the proportion of their patients who are aged 75 or more.

At the left, quintile 1 represents GP practices with the lowest proportion of over-75 year olds. At the right, quintile 5 represents those with the highest proportion of elderly patients. The bars then measure the percentage change in the relative needs attributed to each quintile’s patients between the old and the new formula.

A negative percentage change indicates that the new formula regards patients in those practices as less needy than indicated by the old formula. A positive change indicates the new formula regards those patients as needing additional funds.

Happily for Mr Lansley then, the new formula does do a bit of what he wants – it reduces the need weighting for GP practices with the lowest proportion of over-75s, while increasing the weighting of the rest.

But this clearly is not a wholesale redistribution to elderly areas. As we can see, the quintile benefiting the most is slap bang in the middle.

It’s not big bucks either. This bit of the formula only distributes around two thirds of the NHS commissioning budget – about £38bn. This is the chunk that is spent on hospital and community services. We also need to keep in mind the 0.3 per cent general increase in the sample between the old and new formula – which would almost wipe out the “gain” for the eldest two quintiles.

The difference between the two formulas for GP practices with the lowest proportion of elderly patients would be the equivalent to losing between £18 and £20 per patient a year, whereas those in the middle would gain in the region of between £7 and £9 per patient.

That does add up, however. The total annual loss to a GP practice in the bottom quintile with the average 6,800 patients, would be around £125,000 – the cost of around 22 hip replacements. That of course might be a good thing, if those in the bottom quintile do not actually need those hip placements.

But are these changes anti-poor? We might expect so, given the negative correlation between life expectancy and deprivation.

The next chart shows the same changes between the two formulae, only this time divides GP practices into deprivation quintiles, as measured by the 2010 Index of Multiple Deprivation, with quintile one representing the richest 20 per cent and quintile five the poorest. (The IMD score here is based on where the practices’ patients live, rather than the location of the practice itself.)

There is a small shift away from GP practices in the poorest areas. But here the numbers are even smaller, and again we need to allow for that 0.3 percent discrepancy in the sample. The figures therefore suggest that GP practices in the poorest fifth of England could lose between £5.50 and £7 per patient a year, while those in the richest could gain by between £4 and £5.60.

What about the thing everyone wants to know: is Mr Lansley engaging in some pork-barrel politics: shifting NHS cash to Tory voting areas?

Here are what the changes look like by parliamentary seat:

The percentage changes are now even smaller – and our 0.3 percent discrepancy all but swallows up the Conservative “gain”. Allowing for that, GP practices in Labour constituencies could lose between 10p and £2.10 a year for each patient, while those in Conservative constituencies could gain between £1 and £3, while the lucky few in Lib Dem constituencies could win by between £2.50 and £4.50.

Nationally, that would see Labour constituencies lose a total of around £40m, with Conservative constituencies gaining around two thirds of that and the Lib Dems the rest – a redistribution of around 0.04 per cent of the £100bn NHS budget and, on the face of it, perhaps not as controversial as elsewhere suggested.

The proposed move to the new formula does have its other controversies, however: With not insignificant additional tweaks, it could form the basis of portable individual NHS allocations, which a future government might exploit to introduce an insurance-style system of NHS funding. Patients would receive a basic NHS personal budget to lodge with a private insurer and top up with private funds as per their fancy and ability to do so.

Furthermore, the last public iteration of the formula includes a variable based on whether or not the patient has recently received privately funded care.  If they have, all things being equal, their personal allocation will be smaller – a simple pragmatic calculation that some will worry conflicts with the principle of equal resources for equal healthcare need.

More immediately, if implemented in one step, the change could lead to swings of up to 70 per cent in individual GP practice budgets for hospital and community care. The data behind those implementation issues will be explored in the next, hopefully shorter, post.

The excellent Mary Ann Sieghart says in today’s Indie:

…if you are bright but poor and you live in Kent, Essex, Buckinghamshire or Northern Ireland, your parentage doesn’t have to dictate your progress. You have nearly the same chance of becoming a cabinet minister, a judge, a newspaper editor or a top rower as your privately educated neighbour. Why is that? Because these areas still have grammar schools, those turbo-chargers of social mobility.

But I am afraid that is straightforwardedly untrue.

Here is the grade distribution of schools in Kent, London and England at large at GCSE. As ever, the poorest children, as defined by their postcodes, start at left. The richest children sit at the right. (I chose Kent just for its size and range of population.)

The metric is GCSE points – 8  points for an A* with no caps. (This is a choice of measure that flatters grammar school systems, incidentally. Lots of researchers “cap” attainment, or use point scores that are less geared towards high attainment.)

Zero on the y axis is the national average (37 points). The units are standard deviations (22 points). The population I have used for each line is the set of children educated by state schools in each region.

This graph makes a few things clear:

• The slope of the lines tells you how equitable the systems are. Kent has a gradient of about 0.015. England as a whole is a little over 0.012 and London is lower at about  0.10. Kent is less equitable than the others.
• The level of the line tells you the level at which the schools are performing. So a line which is higher up, on average, is an area where average grades are higher. On this, Kent is at the national average.

What we are showing here is the same thing others have found. Taken together, what that means is that, on average, poor children do markedly worse in Kent than in the rest of the country. Kent is less socially mobile than the rest of England – and much less mobile than London.

A poor child in Kent, using the usual definition of a child eligible for free school meals,  has a 55 per cent chance of getting results that put it in the bottom fifth of results- that means results weaker than around 3 Cs at GCSE.

To look at the other end of the spectrum, only 4 per cent of FSM-eligible children in Kent get results in the top fifth nationally – that means the equivalent of eight As. If Kent were overcoming disadvantage totally, this would approach 20 per cent.

It’s doable: in the Royal Borough of Kensington and Chelsea, home of the Indie’s offices, 24 per cent of FSM-eligible children finish in the bottom fifth and 15 per cent finish in the top fifth. K&C is a good borough and – let me remind you – in a good city.

Indeed, at the risk of sounding monomaniacal, we do have a region with excellent state education. This may be the only time a British journalist has ever had cause to say this, so cherish this moment: “We aren’t talking enough about London.”

Most notable among them is the steady erosion over the past year in executives’ perceptions of the “business friendliness” three of the world’s biggest developing economies, India, China and Brazil.

Each quarter, the Economist Intelligence Unit, which conducts the online survey from a global panel of business executives, has asked respondents to rate the “business-friendliness” of the G7 and Brics nations on a five-point scale. This chart ignores the neutral respondents and groups the rest into broad classes of respondents rating each country on the “friendly” or “unfriendly” side of the scale.

The case of India has been particularly stark. While 45 per cent of respondents rated India business-friendly a year ago, that figure has slumped to 28.6 per cent in the most recent survey, published on Thursday. Perhaps reflecting scandals and controversies over the scrapped plan to liberalise of foreign investment in retail, respondents saying the Indian government is not business-friendly, has quickly grown from 24.8 per cent a year ago to 35.2 per cent today.

A year ago, views on China’s government were split, with slightly more executives rating its business friendliness positively than negatively. Now, just 26.3 per cent say China is friendly towards business, while nearly 40 per cent say it is unfriendly. The change in Brazil’s business-friendliness rating is less pronounced, but also on a gradual downward trajectory.

The fourth member of the Brics club, Russia, was consistently rated as unfriendly towards business throughout the year. By contrast, executives’ perception of the governments of Britain, Canada, Germany, and the US have remained consistently positive over the course of the year.

Two eurozone countries, France and Italy, have also seen significant changes in their business-friendliness rating. Italy’s business friendliness rating has been recovering from a sharp decline a year ago. Respondents’ views on the French government, meanwhile, have grown more polarised, with a steadily growing proportion of respondents rating it as unfriendly towards business. Given business wariness of new president François Hollande, it be interesting to see whether France’s business-friendliness rating tips into net-negative territory when our next Barometer is published in August.

Shale gas developments in the US have sparked a wave of euphoria about the opportunity for a renaissance of its domestic manufacturing base.  Petrochemicals should be one of the main beneficiaries, as the ethane produced from shale gas discoveries now provides the US with some of the cheapest feedstock in the world.

Major producers including Dow Chemicals, Shell and Chevron Phillips have already announced plans to build new ethane-based capacity.  Others are likely to join them.  Current estimates suggest total US ethylene capacity could therefore increase by 25 to 30 percent from today’s 27 million tonnes.

However, one key factor has the potential to spoil the story – much of this new capacity will need to be exported in the form of polyethylene (PE) and other major plastics.  Yet as the chart shows, based on data from Global Trade Information Services, US net PE exports have actually been declining since 2010, even though its cost advantage from shale gas was increasing.

This is quite a different picture from the 2006-09 period, when US exports rose from 700KT to 2.6 million tonnes (MT).   China’s demand surged in 2009, due to its major stimulus programmes, enabling US exports to jump from 267KT to 900KT.  But since then, US exports to China have dropped back to 2006’s level, and its total exports have fallen 39 percent to 1.6MT.

Analysis of recent market developments in China highlights the root cause of this dramatic shift. As the second chart shows, based on Q1 data for 2010-12:

• China’s own demand has fallen 4 percent, as stimulus programmes have ended
• Meanwhile, its own production has increased 7 percent

This, of course, makes no sense in economic terms.  China has some of the highest cost production in the world, being based largely on imported oil.  But it values social stability above economics.  So it is unlikely to shut plants and increase unemployment in order to access cheaper US imports.  Instead, it is already planning further increases in its own capacity.

In addition, China is now taking a strategic view of its remaining import needs.  Middle East imports were up 34 percent over the period, due to the ‘strategic corridor’ under which China provides access to markets in exchange for energy supplies.  South East Asia imports were also up by 22 percent, due to the free trade agreement now in place.  Imports from other regions thus suffered major falls.  North East Asia imports were down 38 percent, with US imports down 62 percent and EU volumes down 68 percent.

Thus PE highlights the major challenge facing the USA as it seeks to use the shale gas bonanza to reinvigorate its manufacturing base.  It has the feedstock, and the cost advantage.  But it needs to increase its exports, if it is to move the additional volume.

The PE market however suggests that potential customers, such as China, are no longer playing by the old rules.  Social and political factors are instead becoming increasingly important in dictating global trade patterns, at the expense of pure economics.

Paul Hodges is chairman of International eChem

Part of the principle of this idea is that it is difficult to recruit in areas of high “outside” wages, if you have a national pay-scale. Conversely, the theory goes, when the state “overpays”, it can price out local enterprise.

Some economists from the Centre for Economic Performance at the LSE found a scary natural experiment to test this idea: survival rates from heart attacks.

We predict that areas with higher outside wages should suffer from problems of recruiting, retaining and motivating high quality workers and this should harm hospital performance. We construct hospital-level panel data on both quality – as measured by death rates (within hospital deaths within thirty days of emergency admission for acute myocardial infarction, AMI) – and productivity. We present evidence that stronger local labor markets significantly worsen hospital outcomes in terms of quality and productivity.

A 10 per cent increase in the outside wage is associated with a 4% to 8% increase in AMI death rates. We find that an important part of this effect operates through hospitals in high outside wage areas having to rely more on temporary “agency staff” as they are unable to increase (regulated) wages in order to attract permanent employees.

Others – notably Chris Giles – have discussed some of the broader themes around this issue elsewhere. But there is a curious wrinkle with this approach to English schools.

I have, quickly and dirtily, worked out the average GCSE performance for each 16 year-old in England. I have then ranked them by deprivation, using their postcodes. I then stripped out the middle fifth of kids, so I have a sample of pupils from similar backgrounds for each LA.

I then worked out average GCSE points for the middle fifth of children in each LA in a very simple manner: I awarded 8 for an A* down to 1 for G, then added them all up. Guess what: London wins again! This first table shows the best local authorities on this GCSE measure.

I have also included the local teachers’ pay ratio in the table, as calculated by the DfE: that is the relationship between classroom teachers’ pay to that of other local professionals. You can read numbers above 1.00 as meaning “overpaid”,  below is “underpaid”.

Now, here are the worst local authorities on that crude GCSE measure. I have, again, included the ratio of teachers’ pay to that of local professionals.

In general, it appears that if we were to regionalise pay, we would cut it in the regions with our worst schools and raise it in regions with the best (particularly London).

It would, in particular, be tougher to attract good teachers to the urban north, where schools are weakest, if we regionalise pay according to local labour market conditions. Many areas have problems recruiting good teachers with the payscales as they are.

That might be why the government is considering ditching pay-scales altogether for teaching.

UPDATE: 22:00

Rebecca Allen, guru on all things schooly from the Institute of Education, and co-author of studies on teachers’ pay, writes in to say:

According to the report, currently only 15 percent of the world’s population lives in a country where more than 90 percent of births and deaths are registered – and unsurprisingly most of these 34 countries are in Europe and the Americas.

It’s not surprising that war torn countries like Afghanistan might have had other concerns than registration data. But the list of countries without comprehensive data include major economic and population centres like China and India – both of whom use sample registration approaches. The full country by country list is on this pdf.

Unsurprising, this means the WHO has to use extensive statistical modelling to produce internationally comparable mortality estimates. The report says:

These estimates, in particular for cause-specific mortality (for example, maternal, HIV/AIDS or malaria mortality), are subject to considerable uncertainty and variation due to the different assumptions and methods used. The substantial efforts and attention devoted to the regular updating of these estimates reflect the global need to improve assessment of cause-specific mortality trends.

Something to consider next time you read about the prevalence of deaths from a specific disease in a low income country.

The cause of good quality data might not grab the heart strings, but a recent academic spat over the number of deaths caused by malaria indicates how important it is.

Research from the Institute for Health Metrics and Evaluation (IHME) in Seattle suggested the real global death toll from malaria could be double the WHO estimate, 1.24 million instead of 655,000. Their controversial methodology used ‘verbal autopsies` – interviews with relatives and friends of the deceased – to complete the estimate.

Whilst the number of countries reporting causes of death to the WHO tripled between 1950 and 2000, it has only increased by 13 countries in the last decade. Efforts to improve the data continue, but at the current rate it will be a long time before we really know how many people die of malaria.

The rapid growth in greener, more fuel efficient vehicles will leave the Treasury with a £13bn shortfall in motoring taxes by the end of next decade, despite an anticipated 44 per cent jump in traffic.

The predictions from the Institute for Fiscal Studies, an independent think tank, are based on an analysis of the government’s own long-term forecasts, which show that by 2029 fuel duty will contribute 1.1 per cent of GDP, down from 1.7 per cent today. Vehicle excise duty will drop to 0.1 per cent from 0.3 per cent.

Receipts from the taxes total £38bn a year, equivalent to 7 per cent of all Treasury income.

The lost revenue is equivalent to increasing the basic rate of income tax from 20p to 23.4p, VAT from 20 per cent to 22.7 per cent or raising fuel duty by more than 50 per cent, according to the IFS report, published today.

Region	Comprehensives	High performers on EEF measure
North West	399	52
London	383	164
South East	368	56
West Midlands	330	39
East of England	301	48
Yorkshire and the Humber	282	17
South West	262	29
East Midlands	230	30
North East	152	11

What are they putting in the eels? I’m not sure if London’s success is depressing or not. On one hand, we have a sizeable area of the country which has massive inequality and poverty where the school system is speeding ahead of other regions.

It shows schools can be turned around: London’s school system, once a renowned catastrophe, is now excellent. It now comfortably beats its hinterland. There is, quite simply, no better place in England for a poor child to get an education.

But London also has advantages. It’s easier to raise aspirations when your school looks onto the City or Canary Wharf. London also has the best academy chains, had the London Challenge, enjoys easier recruitment rounds and reaps the benefits of immigrant children.

So, for a start, the EEF number suggests we really need to work out what on earth has happened in London in general. It also suggests that we need to have a close look inside the 282 schools outside London which pull off the feat of beating the averages.

UPDATE  - 15th May, 7:30pm – As it happens, after further poking around, only 75 of the EEF schools outside London have more than 10 per cent of children eligible for free school meals (FSM), a widely used measure of poverty. (The national average for secondary school pupils is 16 per cent.) The secret of much of these schools’ success may be “have very few poor children”. The non-London schools that really repay attention are the 20 which have more than 20 per cent FSM.

Further analysis makes the London schools look even better. There are 128 London EEF schools where poor kids beat the national average with FSM rates of more than 10 per cent and 77 with more than 20 per cent. In 25 cases, they break 40 per cent.

In response to various requests, here is the information in  a more systematic fashion. First, the list of schools. This is the list of all English comprehensives. The variable “highperformer” marks out schools which are in the EEF’s list.

Second, this graph shows the distribution of FSM children in comprehensives nationally:

This graph shows the same figures for the 446 EEF schools, which tend to be a little richer.

This is particularly true of the EEF schools outside London…

Lots in London, however, are very poor on the FSM measure.

This all just reconfirms London’s exceptionalism. Maybe it’s the pie. Or mash. Or liquor. Or kettle-destroying chalky water. Whatever London is doing, we need to bottle it.

He identified 446 schools where poor children, defined as those eligible for free school meals (FSM), were beating the national average for all children. This is a real core of excellent schools who are market-beaters, and excludes grammar schools. (For people who want to replicate the results in Stata, the necessary file is here.)

Here’s how that population of children within those 440 schools look like, alongside the whole school system graph, on what one civil servant affectionately calls my “Graph of Doom“ for 2010-11:

These are average GCSE points attained by each child – 8 points for an A* down to 1 point for a G. Poorest children, defined by the multiple deprivation score for their postcodes, are at left. Richest are at right. Zero on the y-axis is the national average (37.3 pts) and the units are standard deviations (22.0 pts).

These schools are getting more equal results at a higher scale than others. Whereas moving to a postcode ten national percentile ranks poorer means you fall 3.3 percentiles (on this measure) on average, the effect is 2.2 percentiles in these schools. Poverty explains half as much of the attainment variation in these schools than nationally.

We can quibble about some of these schools and whether they represent replicable, scalable models for schools. You’ll see the poor kids are still being left for dust by their classmates. But it’s certainly an interesting group of schools to consider. These schools do better on efficiency (better results) and equity (more evenly spread).

Having been nice about that, let me kick one statistical measure down a flight of stairs: the attainment gap. The government is publishing statistics showing how far poor children are behind their more affluent classmates. These stats show how much worse poor children (FSM kids) fare than other pupils at the same school.

This is not a terribly useful statistic since it tells us more about the school’s intake than its performance.

As Rebecca Allen, a guru at the Institute of Education, has already written:

…I believe free school meals (FSM) attainment gaps should not be used as a measure of success at a school. Here the problem is that the background characteristics of pupils who are not FSM will vary considerably across schools. So, calculating the group average attainment of non-FSM students at a school tells us a lot about what these non-FSM students are like on entry, and little about how well the school serves them once they arrive…

I first noticed how problematic attainment gaps were in practice as a governor of a school that was struggling to produce strong academic results but was very proud that its FSM gap was zero. All the students at the school came from low income families living on a very large and universally deprived council estate. Some of the families happened to claim benefits that made them eligible for free school meals (they probably weren’t the poorest), others didn’t or couldn’t. Not surprisingly, the GCSE performance of the FSM and non-FSM pupils in this school were no different, on average, because these pupils were no different in their social or educational background. Nothing the school was doing was contributing to this supposed ‘success’.

Attainment gaps compare groups within a school, whereas we should be comparing a group across schools. What matters to FSM pupils is that a school enables them to achieve qualifications to get on in life. If a low income student gets a low quality education from a school, it is little consolation or use for them to learn that the higher income students were equally poorly served by that school.

If you want evidence of that, here is the Graph of Doom™ for the best 30 per cent of schools in the country at closing the attainment gap, measured using the DfE’s GCSE points measure.   (NB - I ranked the schools by measuring the size of the gap in proportion to the average number of points achieved, which improves the measure.)

Adjusting for income, children at these schools do 1 percentile rank better than they would at an average school. Not nothing. But overall, these schools are no better at helping the poor catch up to the rich. Given that this is what the “attainment gap” measure is supposed to capture, it’s a pretty unhelpful measure.

Immigration has helped to raise average wages of most UK-born workers but held back wage growth for those at the most poorly paid end of the labour market, new study has shown.

Research published in the Review of Economic Studies examined the period between 1997 and 2005, when there was an increase in the foreign-born population equal to 3 per cent of the native population.

The authors – economists Christian Dustmann, Tommaso Frattini and Ian Preston – estimate that immigration depressed wages by 0.7p per hour at the 10th percentile, or the bottom layer, of UK-born workers.

But immigration contributed about 1.5p per hour to wage growth at the median and slightly more than 2p per hour at the 90th percentile.