Improvements of the Swedish estimated Intrastat data by adding estimations based on VIES data Nordic meeting 16 September 2014 Jennie Bergman & Ari Mansikkaviita
Introduction VIES – VAT Information Exchange System Agreement with the Swedish Tax Authorities since November 2010 Monthly VIES-data deliveries The aim of the project: To evaluate and provide better reliability in the Swedish external trade statistics in goods To develop new and improved estimation methods for non-collected data To investigate the reliability in the VIES data
Estimation methods Estimates of non-response are done in two stages: Step 1: Estimation of the total flow per company Step 2: Distribution on commodities (CN8) combined with country codes - where history is available - where history is missing Estimation methods: Imputation with VAT value Manual imputation Exponential smoothing
Swedish Intrastat system Threshold value SEK 4.5 million for both arrivals and dispatches Coverage 2012: dispatches 98.2%, arrivals 96.3% Number of companies reporting 2012: dispatches 5,200, arrivals 8,800
VIES data Monthly/quarterly reported dispatch data to Swedish Tax Authorities No threshold Number of companies (on average per month): 16,000 Arrival data obtained from other member states’ reported dispatch data Number of companies (on average per month): 58,000 Dispatch data: reporting period, VAT-number of the reporter, VAT-number (incl. country code) of the receiver of the goods, value of goods, value of triangular trade and value of services Arrivals data: reporting period, VAT-number of the receiver of the goods, country code of dispatching MS, value of goods, value of triangular trade and value of services
Differences in definitions between Intrastat, VAT and VIES Reporting periods Invoice or goods flow Inclusion/exclusion of certain transactions; deliveries without an invoice, deliveries sold to/purchased by private persons, goods for processing Staggered consignment Triangular trade
Comparing Intrastat, VAT and VIES Total sum of reported values in Intrastat, VAT and VIES for the same companies, arrivals. Corrected values for April and December in brackets. Period Number of Intrastat reportersIntrastatVATVIESIntrastat-VATIntrastat-VIESVAT-VIESIntrastat-VATIntrastat-VIESVAT-VIES Bill. of EUR Diff. bill. of EUR Diff. percent , %6.5%4.9% , %12.1%10.8% , %11.0%9.7% , (6.2) (0.2)-3.5 (0.0)3.1%-51.6% (3.1%)-56.5% (0.0%) , %10.1%8.8% , %16.2%14.9% , %15.8%14.3% , %11.5%10.0% , %9.8%11.3% , %10.1%8.8% , %15.9%17.1% , (5.8) (0.1) 0.0%44.1% (1.7) (69.7) (8.1)6.3 (7.3)1.0%9.1% (10.4%)8.2% (9.5%)
Comparing Intrastat, VAT and VIES Total sum of reported values in Intrastat, VAT and VIES for the same companies, dispatches. Period Number of Intrastat reportersIntrastatVATVIESIntrastat-VATIntrastat-VIESVAT-VIESIntrastat-VATIntrastat-VIESVAT-VIES Bill. of EUR Diff. bill. of EUR Diff. percent , %10.7%2.0% , %3.7%-2.0% , % 0.0% , %1.9% , % 0.0% , %12.1%1.9% , % 0.0% , % 0.0% , % 0.0% , % 0.0% , % 0.0% , %9.1%4.8% %7.2%0.7%
Comparisons on country level Number of companies and sum of value per country in Intrastat and VIES 2012, arrivals. (*Trade not reported by Swedish companies) Country code Number or Intrastat reporters Number of VIES companies* Intrastat value bill. of EUR VIES value bill. of EUR Diff. bill of EURDiff. percent AT1, BE2,6592, BG CY CZ1,2781, DE6,5826, DK5,5136, EE1,1901, ES2,1052, FI3,7094, FR3,3924, GB4,4655, GR HU IE9511, IT3,9524, LT8071, LU LV MT NL4,3455, PL2,3362, PT RO SI SK
Comparisons on country level Number of companies and sum of value per country in Intrastat and VIES 2012, dispatches. Country code Number or Intrastat reportersNumber of VIES reporters Intrastat value bill. of EURVIES value bill. of EURDiff. bill of EURDif. percent AT1,7821, BE2,4182, BG CY CZ1,7601, DE3,9934, DK4,2454, EE2,2862, ES2,2202, FI4,0604, FR2,8243, GB3,2133, GR9801, HU1,3461, IE1,1191, IT2,3832, LT1,5491, LU LV1,4701, MT NL2,9613, PL2,8843, PT1,1811, RO9541, SI SK9321,
Possibilities with VIES data Greater opportunities to improve the quality of the estimated data More estimation based on companies’ own historical data on country level – both for non-respondents and companies below the threshold Three step approach instead of two (for companies with VIES data); Step 1: Estimation of the total flow per company Step 2: Distribution on country level by using VIES data Step 3: Distribution on CN8 level by using Intrastat values
Proposals for future development Implement VIES data in the new Intrastat system that will be build Testing and implementing VIES data in the estimations Improve the checking between Intrastat and VAT by implementing VIES data Investigate large discrepancies between Intrastat and VAT/VIES for dispatch data