Preliminary GDP flash estimate in 30 days for Europe

From 29 April 2016 onwards, Eurostat will publish preliminary flash estimates of quarterly GDP for the euro area (EA) and for the European Union (EU) at about 30 days after the end of the quarter (t+30). This earlier publication is an achievement of the European Statistical System, as Member States contribute by providing their national estimates to Eurostat two weeks earlier than before.

The methodology mainly follows the current methodology for the GDP flash estimates published 45 days after the end of the quarter (t+45)[1]. The main difference is the use of more preliminary country estimates. Because of the earlier timing, data for the third month of the quarter often have to be (at least partially) estimated by Member States.

How is the preliminary flash produced?

In brief, a majority of the EU Member States transmit to Eurostat their GDP growth estimates one day before the agreed publication date. This group includes those Member States already publishing their t+30 estimates as well as a number of Member States that transmit confidential estimates to Eurostat. Eurostat then aggregates countries’ quarter-on-quarter growth rates by using the weights of countries’ annual GDP to obtain estimates for the EU and euro area GDP growth rates.

The decision to start publishing the EA/EU GDP flash t+30 estimates is based upon the results of testing phase for the quarters of 2012–2015. In the last test estimates for the quarter 2015Q4, the country coverage reached 91 % of the EU-28 GDP and 94 % of the EA-19 GDP.

Is the preliminary estimate reliable?

As flash estimates are available earlier than the traditional estimates, there is a trade-off between timeliness and accuracy. Based on the 16 test estimates, the expected typical revision at t+45 for the GDP growth rates published at t+30 (with one decimal rounding) would be -0.1, 0.0 or +0.1 percentage points for both the euro area and the EU. In the case of a sudden dramatic change in the economic development, as in 2008Q2 with the financial crisis, larger revisions than these may be expected.

The expected revisions have been tested against predefined quality criteria and passed these tests [2]. Figure 1 below shows the differences in the (non-rounded) growth rates at t+30 (test estimates), at t+45 and at t+65 days for 2012Q1–2015Q4. The upper graph gives the growth rates for the EA and the lower graph for the EU. Figure 1 below shows visually the differences in the growth rates at t+30 test estimates, at t+45 and at t+65 days after the quarter for 2012Q1–2015Q4. The upper graph gives the growth rates for the EA and the lower graph for the EU.

Figure 1: The growth rates (in %) of EA (upper graph) and the EU (lower graph) at t+30 (test), t+45 and t+65 days

Notes

  1. Further details are available in this link.
  2. See Eurostat statistical working paper 'Euro area and European Union GDP flash estimates at 30 days' published on 29 April 2016 on the Eurostat website.