Data

Sources

Sentiment towards EU at regional level

Cohesify => http://www.cohesify.eu/data/ (http://www.cohesify.eu/wp-content/uploads/2018/05/EU-Identification-typology-COHESIFY-Output-2.4a.xlsx) This data is derived from Eurobarometer survey data from 2015.

Time frame for analysis

There is a lot of historical data on GDP and cohesion fund payments, but in order to keep things simple, it is probably best to limit the analysis to the 2007-2016 period. This is 10 years of data, starting with the 2007-2013 funding period and ending halfway the 2014-2020 funding period. This is a time frame the EU is also using in many of its reports. This period contains the economic crisis from 2008 that many regions are still recovering from.

One issue is that the sentiment data is using the old regional boundaries (NUTS 2010), while the other data is using the new NUTS 2013 boundaries. So at the moment we have some NA’s when joining data together. Some manual mapping or filtering still needs to be done.

Relevance

After some more research, some elements came up that make this story very relevant at this moment:

Story outline

1. Country GDP’s

The base numbers of the story are regional gdp numbers. Following chart shows the gdp per inhabitant for the countries of the EU.

But the cost of living is not the same in each country: with the same amount of money you could comfortably live in 1 country, but struggle in another. That’s why gdp numbers are corrected for differences in price levels.

To make comparisons between countries easier, and especially compare countries with the EU average, we scale the numbers, so that the EU average is equal to 100.

Luxembourg is clear outlier: it has the highest GDP/capita in the world (not in ppp, though): it has many banks and many foreigners commute from nearby countries (Belgium, France and Germany). Ireland is also an outlier (and alos infamous as a tax haven). The poorest countries in the EU are Bulgaria, Romania and Croatia.

2. Country map

On the map, we see that the Nordic countries and the countries in center of Europe are the rich countries. Poorer countries are in the east and the south of Europe. Also note Iceland, Norway and Switzerland: they are not part of the EU.

3. Regions map

But countries are not homogenous economic entities: country averages hide big regional differences. The following map shows the 276 regions of Europe as used by the EU. Each region contains 800.000 to 3 million inhabitants.

The poorer regions are mostly in eastern and southern Europe, but there are also poor regions in the UK and France, for example. Pockets of prosperity in Eastern Europe: these are the capital regions (TODO: highlight them on the map).

Countries with an internal gap: Belgium, Spain, Italy, Germany.

4. Regions dot plot

Let’s return to the chart, to be able to better compare regions. [ANIMATE MAP TO CHART?]

Let’s split the countries up into regions. [ANIMATE REGIONS IN INTO THE CHART]

  • Richest regions (London, Luxembourg, Ireland-Dublin, Brussels, Hamburt), poorest regions (…)
  • In Belgium, the poorest region (the province of Hainaut), is below 70 % of the EU average, while the richest region (the capital of Brussels) is at 200 % of the EU average.
  • Capital regions tend to be much richer than other regions. See Belgium, France, UK, Czechia, Slovakia, …
  • … (https://ec.europa.eu/eurostat/statistics-explained/index.php/GDP_at_regional_level)

EXPLICATION ABOUT EU OBJECTIVE OF COHESION => GOAL IS TO SHIFT THE DOTS ON THE LEFT TO THE RIGHT HOW DOES THE EU DO THIS? WITH INVESTMENT FUNDS, THAT POOR REGIONS HAVE MORE ACCESS TO: explication about 75% and 90% eligibility

5. Suspicious histograms

So: regions have a motivation to keep their gdp under the 75% and 90% of the EU average gdp threshold: then they are eligible to get more cohesion fund money and they have to cofinance projects at a lower rate. If regions manipulate their gdp’s to stay under these thresholds, the histograms should show peaks just to the left of the black vertical lines. Clear peaks in 2012, 2015 and 2016 => In what years are the gdp’s evaluated to calculate eligibility? For the eligibility for the ERDF and Social Fund for 2014-2020 the average gdp for 2007-2009 was taken, for the Cohesion fund, it was the average GNI/capita for 2008-2010.

The map on this page suggests that the 2014, 2015 and 2016 data are used to calculate the eligibilty for the 2021-2017 period.

[ANIMATE DOT PLOT WITH REGIONAL GDP’S TO UNIT HISTOGRAM (with 2016 data)][FADE IN HISTOGRAMS FOR EVERY YEAR IN 2007-2015 => HIGHLIGHT PEAKS]

6. Convergence?

The whole point of the EU Cohesion Policy is convergence: the gap between rich and poor regions should be made smaller. Do we see that in the data?

[OVERLAY DOTPLOT FOR 2016 OVER 2007]

Alternative plot: density plot. Blue is density plot for 2007 region gdp, red is 2016 region gdp. We see some convergence: the very poor regions have shifted to the right. But: the economic crisis made a lot of regions poorer (peak of the distributions shifted left).

7. Which regions converged?

Another view on the convergence: scatterplot of 2007 gdp vs gdp growth 2007-2016. Clearly visible: the Greek tragedy (labels starting with “EL”), and the big growth of the poor regions of Romania (RO), Poland (PL), Bulgaria (BG) and Slovakia (SK).

## Warning: Removed 1 rows containing missing values (geom_label).

8. Money distributed vs. perception of EU

Plot of how much money each region received and its gdp in 2016 + color mapped to sentiment towards the EU. Regions that received much money have a more positive attitude towards the EU, regions that received much money but still see the EU as negative are some German regions, Czech regions and a region from the UK.

## Warning: Removed 12 rows containing missing values (geom_label).

Sentiment map of the EU regions: the Poles, the Romanians and the people from Brussels, Luxembourg and London like the EU.

9. Wrap up