The year 2020 is a special year for the PECBMS because the general data set revision has been running. Therefore, we will spend time dedicated to the calculation of European indices and indicators on the tool and database update. The deep revision will enable us to implement new tools, clean up the database, and run various analyses. This way, we aim to improve data quality controls and speed up all the processes needed to produce indices and indicators updates in the next years.
Therefore, the 2020 Indicators update is not computed classically, but an estimate of the indicator value for the year 2018 is used instead. Indicators for all the regions are estimated using a moving average (also called running mean).
For the 2020 update, we used the file “indicators-europeeu-till2017.xls” from the former year. We added the extrapolated value for the year 2018 to the data. It is possible to download both files, “indicators-europeeu-till2017.xls” and “indicators-europeeu-till2017&estimate2018.xls”.
In the table below, the way of calculation is shown. For illustrative reasons, there are the last 10 years shown in the table. If the estimate (Estimate2018) is compared to the real data (Indicator), a minimal difference in tenths of a per cent is visible. The estimates are actually a bit higher than real values because the indicator descends. The mean is calculated from former years is therefore approximately 0,5% higher than the real value, which corresponds with inter-annual differences in the real values.
For each year, we calculated the mean of year-before-last and recent year. For the estimate, we used values from the last four years: mean for the year 2016 was calculated (value2014+value2016)/2, mean for the year 2017 was calculated (value2015+value2017)/2. For the estimated indicator value for the year 2018, we added the difference between the means of the last two years to the mean of the year 2017: mean2017+(mean2017-mean2016).
Year | Indicator | Estimate2018 | Mean name | Mean calculation | |||
2008 | 71,71 | ||||||
2009 | 70,89 | ||||||
2010 | 70,13 | 70,91 | mean1 | = | (values for 2008+2010)/2 | ||
2011 | 69,45 | 70,15666667 | mean2 | = | (values for 2009+2011)/2 | ||
2012 | 68,84 | 69,47333333 | mean3 | = | (values for 2010+2012)/2 | ||
2013 | 68,30 | 68,86333333 | mean4 | = | (values for 2011+2013)/2 | ||
2014 | 67,83 | 68,32333333 | mean5 | = | (values for 2012+2014)/2 | ||
2015 | 67,42 | 67,85 | mean6 | = | (values for 2013+2015)/2 | ||
2016 | 67,07 | 67,44 | mean7 | = | (values for 2014+2016)/2 | ||
2017 | 66,78 | 67,09 | mean8 | = | (values for 2015+2017)/2 | ||
estimate | 66,74 | estimate | = | mean8+(mean8-mean7) |
The estimated values are properly marked and described in all the graphs on our website.
Be aware that the estimates cannot substitute the real indicators. The estimates in the 2020 Indicators update were used exceptionally this year to gain time for general data revision. We plan to run the usual computations using standard PECBMS methods next year.
We thank all national coordinators and volunteer observers for their efforts both in the field and in the office!