NTTS 2015 9 13 March 2015, Brussels Session 13C - Survey - PowerPoint PPT Presentation

NTTS 2015 9 – 13 March 2015, Brussels Session 13C - Survey integration, coordination and alignment Avoiding duplicate collection of flow data: estimating intra-EU inbound tourism using partner data Christophe Demunter, Krista Dimitrakopoulou EUROSTAT – Unit G-3 "Short-term business statistics and tourism"

Quick guide to tourism statistics  3 main types of tourism Domestic tourism = visits within a country by residents of that country  Outbound tourism = visits by residents of a country outside that country  Inbound tourism = visits to a country by visitors who are not residents   2 sources for European statistics on tourism Household surveys  cover domestic and outbound trips by EU residents  Business surveys  cover domestic and inbound trips of tourists who  stayed in rented accommodation  1 main gap in measuring tourism flows in Europe Total inbound  due to missing inbound flows of tourists not  staying at rented accommodation 2

Inbound tourism  User needs Physical flows  number of visitors, number of nights, purpose, etc.  Monetary flows  tourism expenditure by foreigners   How to adress the user needs Border surveys  difficult (free movement, Schengen) and very costly  Using existing data  e partner data from other countries  Outbound trip to Italy by a Finnish tourist  Statistics Finland via the Finnish household survey Inbound trip in Italy by a Finnish tourist  ISTAT via border survey?  or… simply re -using the Finnish information on the same flow?

Conditions for exchanging partner data  Harmonisation Common EU regulatory framework  Regulation 692/2011 concerning  European statistics on tourism Harmonised scope, concepts, definitions, variables and breakdowns,  reference period and file formats (micro-data transmitted to Eurostat)  Trust Using data not produced in country A to say something about country A  Quality assurance  e.g. data validation by Eurostat   Access Sufficiently detailed data  Summary tables based on Eurostat micro-data   Coverage  Sample size, completeness, availability and reliability 4

Coverage  Sample size Total sample of 490 000 tourism trips   Trips with at least 1 overnight stay by residents from the EU (-SE) and CH  of which 154 000 outbound trips (31%)  of which 114 000 outbound trips with a destination within the EU (74%)  In other words: a sample of 114 000 intra-EU inbound trips !  But big differences in 'inbound' sample size Sample size for inbound trips from EU-28 (excl. SE) and CH > 10000 ES, FR, DE, IT 5000 - 9999 UK 4000 - 4999 EL, PT, NL, BE 3000 - 3999 HR, AT, PL 2000 - 2999 SE, IE, RO 1000 - 1999 CZ, HU, LV, DK, BG, EE, SK, CY 5 < 1000 MT, FI, LT, SI, LU,

Coverage  Completeness 28 reporting countries x 29 destinations (28 EU, EFTA) = 784 flows  (excluding domestic flows)  94% are included in the sample  6% were not observed (e.g EL to EE, EE to RO, CZ to LU) 6

Coverage  Availability and reliability of data 28 reporting countries x 29 destinations (28 EU, EFTA) = 784 flows  (excluding domestic flows) Share of intra-EU flows Share of observed trips 3.5% 6% 5.5% 36% 43% Available (> 50 obs) Lower reliability (20-49 obs) 91% 15% Not published (< 20 obs) Not observed 7

Coverage  Availability and reliability of data Number of partners countries broken down by sample size for this destination country 2 4 4 5 4 5 9 10 11 11 11 12 12 10 9 16 16 17 17 21 22 23 20 21 21 21 23 23 22 26 27 3 3 2 4 3 5 4 5 6 5 6 6 5 9 11 6 4 5 4 4 2 4 1 3 3 3 1 2 4 23 22 21 19 19 18 18 13 11 11 10 10 10 9 8 7 7 6 6 6 4 4 4 3 3 3 3 3 2 IT DE ES FR UK EL AT HR NL CH BE SE CZ PT PL HU NO BG DK IE SK CY EE FI MT RO SI LV LT LU IS LI <20 obs (not used) 20-49 obs (unreliable) 50+ obs (reliable) 8 Note: not including trips by residents of Sweden

Coverage  Availability and reliability of data Share of inbound trips accounted for by flows with more than 50 observations 100 99 99 99 99 98 95 95 93 93 92 90 88 86 80 84 81 79 73 72 69 60 62 58 52 50 46 46 46 40 42 33 31 20 19 10 0 0 IT DE ES FR UK EL AT HR NL CH BE SE CZ PT PL HU NO BG DK IE SK CY EE FI MT RO SI LV LT LU IS LI 9 Note: not including trips by residents of Sweden

Coverage  Availability and reliability of data Share of inbound trips accounted for by flows with 20 or more observations 100 99 7 99 99 99 9 98 Flows with 20 to 49 observations 23 8 17 95 26 95 13 93 4 93 92 4 Flows with more than 50 observations 45 90 42 88 86 80 84 81 8 79 14 33 73 72 59 69 14 20 60 62 58 26 22 52 50 46 46 46 40 42 33 31 20 19 10 0 0 0 IT DE ES FR UK EL AT HR NL CH BE SE CZ PT PL HU NO BG DK IE SK CY EE FI MT RO SI LV LT LU IS LI 10 Note: not including trips by residents of Sweden

Some preliminary results Average expenditure per inbound trip by residents from the EU Inbound visitors into the EU coming from other EU Member States  spent on average 575 EUR per trip Ranging from 205 EUR in Slovakia (many short trips – average is  4.7 nights – from neighboring countries) to 793 EUR in Cyprus (longer trips – average is 12.1 nights) and 834 EUR in Finland Average expenditure per inbound trip by residents from the EU, in euro (2012) 1500 1272 966 1000 834 793 789 733 683 660 650 623 619 593 575 547 535 530 523 501 500 495 482 475 475 451 427 396 394 346 326 320 545 500 205 156 0 FI CY EL ES IT PT UK BG MT IE EU DK FR AT HR SE HU LT EE SI RO DE PL NL CZ BE LV LU SK CH IS LI NO Note: not including trips by residents of Sweden

Some preliminary results Average expenditure per inbound night by residents from the EU Inbound visitors from the EU spent on average 79 EUR per overnight stay  Ranging from 37 EUR in Romania to 166 EUR in Estonia (the country  with the shortest average length of stay – 2.9 nights – for inbound trips, mainly made by Finnish – 84% of all Estonian inbound tourism trips from other EU countries) Average expenditure per inbound night spent by residents from the EU, in euro (2012) 200 166 140 150 124 113 101 105 94 93 93 92 92 87 100 85 83 83 82 80 80 80 79 78 77 76 74 74 68 65 60 60 54 53 43 37 50 0 EE FI UK AT IT LU SE BE NL CZ LT IE SI DE EL EU ES HU DK PT FR MT CY BG LV PL HR SK RO CH IS LI NO Note: not including trips by residents of Sweden

Coherence Coherence check of the data source with accommodation data Nights spent at accommodation establishments < estimate for inbound nights ?  OK for all countries where data is available, differences can among others be  explained by differences in the share of stays at non-rented accommodation (in particular stays at friends or relatives by visiting migrants) or by data collection thresholds in accommodation statistics (e.g. DK) Ratio of nights spent at tourist accommodation establishments by non- residents from EU-28 and estimated number of nights on inbound trips by residents from EU Member States 100% 80% 60% 40% 20% 0% 13 AT BE BG CY CZ DE DK EE EL ES FI FR HR HU IE IT LT LU LV MT NL PL PT RO SE SI SK UK IS NO

Coherence Coherence check of the data source with accommodation data The gap between supply side and inbound demand side data can partially be  explained by considering the inbound trips spent at non-rented accommodation: about 25% of intra-EU trips was spent at non-rented accommodation in 2012 RO: lowest ratio, but non-rented accommodation (68%) explains most of the gap  DK: biggest unexplained gap due to under-coverage of accommodation data  Ratio of nights spent at tourist accommodation establishments by non- residents from EU-28 and estimated number of nights on inbound trips by residents from EU Member States 100% 80% 60% 40% 20% 0% 14 AT BE BG CY CZ DE DK EE EL ES FI FR HR HU IE IT LT LU LV MT NL PL PT RO SE SI SK UK IS NO

Coherence Coherence check of the data source with existing inbound data Six Member States kindly provided data from their border surveys for 2012  (CZ, IE, ES, HU, MT, UK), this preliminary analysis has been anonymised Demand side data systematically lower than border survey (recall bias?)  Deviations are bigger for trips than for nights (less recall for longer trips), as a  consequence also average duration is longer on the basis of demand side data Ratio of inbound demand side data and inbound border survey data (2012) 120% 119% 115% 120% 110% 99% 94% 100% 84% 82% 78% 72% 71% 71% 80% 59% 60% Number of inbound trips 40% Number of inbound nights 20% Average duration 15 0%

Conclusions  Burden & cost neutral  High potential for bigger countries and/or popular destination countries, less for other countries Sample size in the generating countries?  biggest cost driver…  Correlated with size and attractiveness  more observed trips   Some methodological issues inherent to the data source Non-response in household surveys  possible systematic bias  Recall bias regarding tourism trips  up to 20 to 25% underestimation   Coherence with other sources is satisfactory and promising  Mixed enthousiasm regarding dissemination of this data 16