Analysing migration between Germany and other EU countries
Reza Hosseini
18 Juli 2016
Synopsis
The aim of this study is to analyse the data of migration between Germany and other EU countries in the time frame between 1974 to 2014. We will explore the data analyse change in time, and relative to countries, use the data to forecast mogration in years 2015 and 2016.
To summarize the result, we see a incresing pattern in the number of migrations both to and from Germany, specially from 2000’s. The highest total number of departure from Germany to another country took place in 2014 to Poland with 138 thousand migrants. And the greatest number of people came to Germnay from another cuntry were also from Poland in 2014 with around 197 thousand. Finally we concluded a meanigful difference between the number of Female and Male migrants from Germany both in the years 2013 and 2014.
Data exploration
The dataset contains the number of people arrived to or departed from Germany for living from 1947 to 2014. It also gives these data based on the gender of migrants.
To have an understanding of the pattern here you can see the total number of arrivals and departures based on gender.
It can be seen that the trend in both arrivals and departures is roughly increasing specially from mid 2000’s. Also till mid 2000’s the number of departures is roughly equal and sometimes more that the number of arrivals. But from then the number of arrivals started increasing faster and gets higher than departures.
The maximum number of departures/Arrivals
We want to explore how many people came and left when and where the most. These can be seen the best in the plots below
Based on the plots, in the yeras 2008 and 2014 respectively 132 and 138 thousand people have left Germany to Poland which are the most number of departures to a single country. Also in each of years 2013 and 2014, 197 thousand people came to Germany from again Poland.
Forecasting
As no seasonal pattern could be seen in the data but there are local trends in them, I decided to use the linear exponetial smoothing method for forecasting. The results of forecasting including 95 percent confidence intervals can be seen in the following charts and table.
| State | Arrival/Departure | 2015 | Conf. Interval | 2016 | Conf. Interval |
|---|---|---|---|---|---|
| Austria | Arrivals | 19,557 | (16806,22307) | 19,821 | (16806,22307) |
| Austria | Departures | 21,293 | (18661,23924) | 21,148 | (18661,23924) |
| Belgium | Arrivals | 6,307 | (6053,6560) | 6,515 | (6053,6560) |
| Belgium | Departures | 5,344 | (4953,5735) | 5,359 | (4953,5735) |
| Bulgaria | Arrivals | 82,296 | (77638,86954) | 90,381 | (77638,86954) |
| Bulgaria | Departures | 49,162 | (48179,50145) | 54,110 | (48179,50145) |
| Croatia | Arrivals | 44,240 | (NA,NA) | 44,240 | (NA,NA) |
| Croatia | Departures | 17,327 | (NA,NA) | 17,327 | (NA,NA) |
| Cyprus | Arrivals | 1,182 | (1036,1329) | 1,307 | (1036,1329) |
| Cyprus | Departures | 521 | (455,586) | 545 | (455,586) |
| Czech Republic | Arrivals | 13,777 | (12632,14923) | 14,650 | (12632,14923) |
| Czech Republic | Departures | 9,086 | (7833,10339) | 9,342 | (7833,10339) |
| Denmark | Arrivals | 3,368 | (3022,3714) | 3,187 | (3022,3714) |
| Denmark | Departures | 3,668 | (3153,4184) | 3,695 | (3153,4184) |
| Estonia | Arrivals | 1,244 | (988,1500) | 1,309 | (988,1500) |
| Estonia | Departures | 991 | (922,1060) | 1,037 | (922,1060) |
| Finland | Arrivals | 2,620 | (2137,3103) | 2,636 | (2137,3103) |
| Finland | Departures | 2,416 | (2044,2788) | 2,411 | (2044,2788) |
| France | Arrivals | 23,399 | (20383,26416) | 23,492 | (20383,26416) |
| France | Departures | 19,551 | (17806,21295) | 19,584 | (17806,21295) |
| Greece | Arrivals | 31,273 | (25057,37489) | 30,859 | (25057,37489) |
| Greece | Departures | 16,273 | (9933,22612) | 15,483 | (9933,22612) |
| Hungary | Arrivals | 66,013 | (60333,71692) | 71,774 | (60333,71692) |
| Hungary | Departures | 43,965 | (40658,47271) | 46,906 | (40658,47271) |
| Ireland | Arrivals | 2,967 | (2063,3872) | 3,016 | (2063,3872) |
| Ireland | Departures | 2,388 | (1510,3265) | 2,429 | (1510,3265) |
| Italy | Arrivals | 75,500 | (63036,87964) | 77,640 | (63036,87964) |
| Italy | Departures | 36,700 | (29475,43925) | 37,098 | (29475,43925) |
| Latvia | Arrivals | 8,270 | (6184,10356) | 9,093 | (6184,10356) |
| Latvia | Departures | 6,520 | (5804,7237) | 7,215 | (5804,7237) |
| Lithuania | Arrivals | 11,279 | (9517,13041) | 12,232 | (9517,13041) |
| Lithuania | Departures | 6,700 | (6368,7032) | 7,204 | (6368,7032) |
| Luxembourg | Arrivals | 3,773 | (3561,3985) | 3,895 | (3561,3985) |
| Luxembourg | Departures | 2,884 | (2725,3042) | 2,986 | (2725,3042) |
| Malta | Arrivals | 317 | (262,372) | 339 | (262,372) |
| Malta | Departures | 300 | (269,332) | 322 | (269,332) |
| Netherlands | Arrivals | 14,384 | (13389,15378) | 14,472 | (13389,15378) |
| Netherlands | Departures | 11,724 | (10793,12656) | 11,772 | (10793,12656) |
| Poland | Arrivals | 211,756 | (186892,236620) | 222,109 | (186892,236620) |
| Poland | Departures | 139,631 | (125660,153601) | 140,583 | (125660,153601) |
| Portugal | Arrivals | 11,507 | (6661,16353) | 11,053 | (6661,16353) |
| Portugal | Departures | 7,990 | (2532,13448) | 7,377 | (2532,13448) |
| Romania | Arrivals | 187,227 | (169502,204951) | 207,227 | (169502,204951) |
| Romania | Departures | 128,557 | (118760,138354) | 140,387 | (118760,138354) |
| Slovakia | Arrivals | 15,942 | (14340,17543) | 16,899 | (14340,17543) |
| Slovakia | Departures | 11,568 | (10375,12762) | 11,851 | (10375,12762) |
| Slovenia | Arrivals | 7,685 | (6751,8619) | 8,493 | (6751,8619) |
| Slovenia | Departures | 4,377 | (3976,4778) | 4,751 | (3976,4778) |
| Spain | Arrivals | 43,862 | (40892,46832) | 46,633 | (40892,46832) |
| Spain | Departures | 27,393 | (24462,30325) | 30,636 | (24462,30325) |
| Sweden | Arrivals | 4,370 | (3972,4768) | 4,405 | (3972,4768) |
| Sweden | Departures | 4,609 | (4208,5010) | 4,644 | (4208,5010) |
| United Kingdom | Arrivals | 18,682 | (16123,21240) | 18,786 | (16123,21240) |
| United Kingdom | Departures | 19,506 | (17450,21562) | 19,780 | (17450,21562) |
Difference between the number of male and female travelers
To test the significance of difference I tested the null hypothesis that the number of male and female travelers are equal against the alternative hyphothesis that they are different. As we are dealing with a contingency table of count data, I decided to use the chi squared test.
I developed contingency tables of states vs sex for both years. Below is a chart of the table for 2013.
And for 2014
Then I ran a chi squared test on the contingency tables assuming 50% probability for each sex. The P-value of the twat for both years are almost zero. So the null hypothesis is rejected and the number of male and female migrants are significantly different.
Reference
Statistisches Bundesamt, Wiesbaden 2016