How GDP varies between Regions: Inequality, Life Expectancy, Happiness, Austerity, Referendums, Experimental Statistics, and lots more graphs

I hope that people aren’t getting too bored with discussions of inequality, because this is another blog which touches on 2 familiar themes; how income is distributed; and how governments produce the statistics that we use to inform political and economic debates.

There is a common assumption that over last few decades the economic performance of different parts of the UK has diverged, with London and the Home Counties out performing the rest of the Country.   I wrote recently about income inequality and I wondered if the picture would look different if we thought about it geographically – rich areas and poor areas rather than rich people and poor people.

Lets start with Gross Domestic Product (GDP).  It is the most commonly used measure of how the economy is performing and allows international and historical comparisons, although it does have it’s critics.

David Cameron flirted with the idea of measuring happiness, but abandoned it when it turned out that no-one was happy, and lots of people blamed their unhappiness on his Government:

Screen Shot 2018-05-16 at 12.01.57

I like the way the Guardian’s photo editor picked a photo of Cameron with a sad face to illustrate the story.

Most people will have gathered from the Media that growth in GDP has been slowing since the Brexit vote.   This from the BBC recently is typical:

Screen Shot 2018-05-16 at 22.25.20

In fact while it is true that GDP growth is lower than it was before June 16 it has actually been slowing since David Cameron announced there was going to be a referendum:

Screen Shot 2018-05-01 at 10.16.41

This isn’t a surprise.

The referendum increased uncertainty for businesses, which in turn reduced investment, which impacts on the rate of growth.   This is however part of a longer term trend in GDP, as this graph from the ONS shows:


Screen Shot 2018-05-16 at 12.04.48

GDP growth has been slowing for decades, although it has become more consistent with smaller peaks, but fewer troughs.  The cycles of boom and bust may not have actually ended (as Gordon Brown foolishly claimed) but the busts have got much less frequent.

If we try and make sense of the squiggly graph we can break up the last 60+ years into a number of periods and see some patterns.

Eden-MacMillan-Hume 1955 – 1964

Nearly 10 years of Conservative rule, and GDP grew by an average of 3.1% per year. Wage rises for working class jobs were still low, but if you were middle class you hadn’t had it so good for a very long time.  Growth was patchy as this rather odd set of results from 1958 illustrate:

Screen Shot 2018-05-16 at 13.09.43

Patchy growth was the consequence of the Government trying to grow the economy without letting the trade deficit get too big.   At one point exchange controls were so tight that the Treasury had to approve the transfer of Jimmy Greaves from AC Milan to Spurs because the fee – a colossal £100,000  – might affect the delicate balance of trade.

Wilson – Heath – Wilson – Callaghan 1964 – 1979

A mix of Labour and Conservative administrations on either side of the oil price shock and the economic turmoil of the 1970s. Despite all of this GDP grew by an average of 2.96% a year.   GDP growth is very bumpy, with big swings from growth to contraction; 1965, 66, 68, 69, 70 and 71 all saw 3 quarters of growth and 1 quarter of contraction.   Both Labour and Tory followed variants on “stop/go” policies trying to balance growth, inflation and a trade deficit.

The Thatcher years 1979 – 1992

I am afraid that this is more bad news for the remaining Thatcher fans.   GDP grown was markedly lower in her years in office.  Despite the boom years under Lawson in the mid-80s GDP growth averaged 1.83%.  Given that Thatcher reaped the benefits of the oil boom this is a dreadful set of results.

Major and Blair 1992 – 2007

Major comes into office with continuity Thatcherite policies, but a sharp recession and Black Wednesday soon change his mind.  From Black Wednesday to the Credit Crunch GDP growth is back to 3% a year on average.  Unemployment falls during this period, and inflation is low and stable.  Blair even manages to reduce Government debt as well.  This is the era of NICE economics (Non-Inflationary Consistent Expansion).

Brown-Cameron-May 2007 onwards

GDP Growth is sharply negative in the immediate aftermath of the Credit Crunch, and recovers only slowly.   Average GDP growth is a 1.9% a year, only slightly better than the Thatcher era.  Inflation is low, but unemployment also low.   If we split out data from after the Referendum we growth is even lower – a meagre 1.5% –  like the early 80s but with worse pop music.

This, however only really tells us about information at a National level, it doesn’t tell us much about what is happening in different parts of the Country.

To work out what is happening regionally we have to use the ONS Gross Value Added (GVA) dataset.  This is a new way of looking at Regional economics and until a couple of years ago was treated as an experimental statistical method, although the data exists right back to 1999.  It takes a long time to produce and we won’t see the 2017 data until the end of 2018.

Experimental statistical methods – does it get more exciting than this?

We can start by looking at GDP growth by Nation:

Screen Shot 2018-05-16 at 13.20.52

England and Scotland are growing much faster than Scotland and Northern Ireland.

We can break down England further into Regions:

Screen Shot 2018-05-16 at 13.29.08

Screen Shot 2018-05-16 at 13.30.27

Since the records started all English regions have below average GVA growth except London. If we look at the post Credit Crunch years the variation is even larger:

Screen Shot 2018-05-16 at 13.33.04

Screen Shot 2018-05-16 at 13.36.25

This is total GVA not per capita, and it is also a chained value series so inflation effects have been adjusted out to show actual change.    The results for the North East and for Yorkshire and Humber are shocking.

There are 2 other things we can do with the data.  We can look at GVA per capita:

Screen Shot 2018-05-16 at 13.57.11

This shows that disparities in wages between London and the rest of the country reflect that the GVA per capita is  much higher in London than the rest of the country.  In fact the gap between London and everywhere else is great than the gap between the areas of the UK excluding London with the highest and lowest levels of GVA.

The second interesting thing that we can do is  look at changes within Regions by zooming in on data at Local Authority level.  At this point you should notice that the data changes, and the differences between London and the rest the UK doesn’t look as stark.   That’s because we are no longer dealing with a chained value series, so inflation effects are included, and because we are looking at GVA per capita, so population changes have an effect.  The growth in population in London balances out some of the total increase in GVA.   This is the Income based approach to GVA.

This is the data for Yorkshire and Humberside which shows how GVA per capita growth fell after the Credit Crunch.  The period 1998 to 2007 showed healthy levels of growth well above inflation right across the region.   Post Credit Crunch growth levels are well below inflation which means a real terms fall in GVA, and a real terms fall in average income.

Screen Shot 2018-05-16 at 14.17.29

The numbers for areas like North Lincolnshire are shocking – a fall in GVA before inflation adjustments.  I think of York as a pretty place full of tourists, but in terms of GVA it has more in common with the stagnant Northern Cities than the more dynamic bits of the North.   Or maybe an economy based on tea shops doesn’t create as much wealth as we thought?

These are the same numbers for the North East:

Screen Shot 2018-05-16 at 14.12.10

It’s not hard to see a pattern of big falls in growth in places like Newcastle, Redcar, Barnsley, Sheffield, and Doncaster.

This is the data for London:

Screen Shot 2018-05-16 at 14.21.16

Not only are the affluent parts of London a lot richer to start with, but the falls in GVA growth overall aren’t as stark as their Northern Cousins.  Overall however London’s average is held down by the huge disparities between the more prosperous and fast growing Boroughs and the worst.   Which leaves the obvious question:

What the fuck is happening in Croydon?

Now we have the Local Authority level GVA data some interesting correlations start and emerge.

There is a clear correlation between GVA growth at Local Authority level and the EU Referendum results.  I can’t claim credit for this analysis as it came from the Financial Times.  This is their rather nifty graph:

Screen Shot 2018-05-16 at 11.53.22

I am sure that there are cultural aspects which underpin the Leave vote too, but the link between Regional GVA and the EU referendum is pretty stark.

There is however another link that stood out to me – the link with life expectancy.  While life expectancy across the UK continues to improve the rate of improvement has slowed noticeably, and in some parts of the country it is starting to reverse.

I blogged about these falls in life expectancy and the relationship with austerity economics a while back:

Does Austerity really kill? If so how many? How can it if we are all living longer?

It isn’t an exact correlation but there are clear signs that the areas which show the biggest growth in GVA are among the areas where life expectancy is increasing the fastest, and the areas where GVA growth is lowest are the ones with the slowest growth in life expectancy or even falls.

These are the tables for the same regions showing changes in life expectancy.

Screen Shot 2018-05-17 at 15.24.46

Life expectancy across the region is still increasing,  but the rate of increase is slowing, and in some areas is starting to decline.

The North East is even worse:

Screen Shot 2018-05-17 at 15.15.27

Life expectancy at birth is falling, but is only increasing for women at age 65.

Where there are declines it is the familiar list of names: Middlesbrough, Newcastle, Redcar, Barnsley, Doncaster.

The data for London is predictably a lot better, with increases in all catagories

Screen Shot 2018-05-17 at 15.44.44

If you squint carefully you will see that the awful GVA data for Croydon is reflected in a declining male life expectancy at birth.

We need to be a bit careful in describing a rigid relationship between economic change and life expectancy; while Scotland’s economy is growing faster than Wales and Northern Ireland it’s life expectancy is the lowest of all the Nations, and the gap is getting bigger:

Screen Shot 2018-05-16 at 22.47.37Too many deep fried Mars Bars.

It is also clear that it takes a long time for economic changes to feed through into changes in life expectancy.  Sadly this probably means that even if things start to improve over the next few years it will take a long time for this to reverse the trend in health status.

There are some pretty easy conclusions that we can draw from all of this.

The period up to the credit crunch saw growth in GVA and life expectancy across the UK.  London did better than parts of the North, but not at the expense of the North.  Since the credit crunch not only has growth been lower, but the balance of growth has been much more uneven, and the gap between London and the rest of the UK has been much more severe.

Some of these changes in GVA have clear links to the austerity policies of the Cameron and May Governments.   I haven’t gone through the Industry by Industry data here but the cuts in Government spending were a major cause of changes in GVA in the North.  It is harder to map the impact of QE and low interest rates through the data, but it does look rather like lots of the extra money that the Bank of England printed didn’t get much further than the M25.

It is also clear that the prolonged slump in GVA and GDP growth in the North is having an impact on life expectancy.   These are small changes, and they are forecasts, but they are driven by an increase in mortality rates over the last few years.

We have been used to being told that we are all living longer.   For some of us this is no longer true.

What all of this reminds me of is this paper by on mortality and morbidity in the USA by Deayton and Case:

I’m not too bothered about inequalities in income.  I am bothered more by poverty.  I would rather live in a country where no-one is poor but someone people have gold hats, than a country where no-one has a gold hat but there is lots of poverty.

But these regional datasets really bother me.  It looks like parts of the North of England are starting to follow the pattern of economic decline, falling life expectancy and political volatility which we have seen in big chunks of the States.  Trump was driven to victory by white voters in areas with declining economic fortunes and sharply declining life expectancies.   

We are creating the exact same explosive mix in the UK too. 

And just like America the people living in these areas are largely absent from the national debates about politics and economics, and are cultural invisible to lots of people.

“Two nations between whom there is no intercourse and no sympathy; who are as ignorant of each other’s habits, thoughts, and feelings, as if they were ….. inhabitants of different planets.”

Thanks to the Health and GVA team at ONS for pointing me in the right direction for some of the data.





Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.