1. TOURISM SURVEYS IN BELARUS
Natallia Bokun,
Jelgava, 2018

2. Content
1) tourism in Belarus: main indicators and trends; 2) tourism satellite account and possible information sources; 3) tourism households surveys; 4) tourism establishment surveys; 5) accommodation surveys; 6) Border surveys.

3. ReFERENCES
Bokun, N. (2013). Sample survey of households in Belarus: state and perspectives. Statistics in transition, Warsaw,
Bokun, N. (2016). Micro-entities and small enterprises survey in Belarus. In Proc. of the XI Intern. Conf. “Compute Data and Modeling”, Minsk, Sept. 6-10, – P
International Recommendation for Tourism Statistics Complication Guide (2010).
European Implementation Manual on TSA (2014). UN WTO. Tourism Highlights, 2017 Edition (2017).
Metodologicheskie polozheniya po postroeniyu vspomogatelnogo scheta turizma Respubliki Belarus (2017). Minsk, Belstat.
Metodika po raschetu vjezdnogo turisticheskogo potoka (2018). Minsk, Belstat. Metodika po raschetu obschego objema vvoza i vyvoza fizicheskimi litsami, peresekauschimi Gosudarstvennuyu granitsu Respubliki Belarus tovarov, ne uchityvaemyh tamozhennoy statistikoy (2015). Minsk, Belstat.

4. Tourism in the world
International tourist arrivals: million million million International tourism receipts, US $: billion billion billion International tourism represents 7 % of the world’s export of goods and services. Tourism has grown faster than world trade for the past five years.

5. Tourism in belarus: main indicators and trends
In tourism contributes directly around 2.5 % of GDP, using to 6 % if indirect impacts are also included (Travel and Tourism Economic Impact 2017 Belarus) Export revenues from tourism amount to approximately US $ million annually, equivalent to % of total exports of goods and services. In million people visited Belarus. 40 % of the clients in accommodation establishments were foreign tourists.

6. Table 1 - Tourism in belarus: main indicators
2010
2013
2014
2015
2016
2017
Tourism flows, thousand
Total domestic trips
79.0
655.1
703.7
836.8
1001.8
976.8
Inbound tourism
Total international arrivals (organized tourists)
120.1
136.8
137.4
276.3
217.4
282.7
Top markets:
Russian Federation
80.9
111.3
113.2
243.9
171.1
191.5
Poland
4.0
3.1
1.7
5.9
8.0
22.7
Lithuania
4.4
2.1
2.0
5.8
26.2
Latvia
1.4
1.0
2.3
3.0
6.5
Ukraine
1.9
1.8
7.0
5.2
China
0.6
0.7
0.3
1.6
Germany
2.2
2.5
Estonia
0.8
2.8

7. Table 1 - Tourism in belarus: main indicators
2010
2013
2014
2015
2016
2017
Trips of foreign tourists
5673.8
6240.4
5350.0
4357.2
Outbound tourism
Total international departures (organized tourists)
414.7
708.4
740.5
736.7
495.8
727.5
Top destinations:
Turkey
83.4
99.0
116.0
111.8
57.7
140.2
Egypt
50.8
52.9
94.0
85.3
41.9
123.2
Ukraine
132.3
175.8
5.6
18.5
45.9
91.2
Bulgaria
30.8
89.2
102.8
79.2
58.7
62.4
Russian Federation
25.8
37.3
73.9
130.5
71.5
46.7
Spain
4.0
16.3
26.4
25.3
24.8
31.4
Greece
4.3
26.8
39.1
35.4
22.1
28.0
Montenegro
4.1
12.6
18.1
17.8
19.4

8. Table 1 - Tourism in belarus: main indicators
2010
2013
2014
2015
2016
2017
Trips of Belarus citizens
7464.2
8840.8
7236.3
6962.4
8339.6
9208.6
Tourism receipts, billion rubles
156.7
735.5
935.4
1129.6
136.6
165.9
Export and import by “trips”, million US $
Export
440.4
791.4
867.6
728.7
710.6
789.8
Import
621.5
1153.3
1158.7
901.1
806.1
992.2
Tourism industries, number of establishment
accommodation services for visitors (hotel and similar establishments)
693
945
996
1014
1052
1072

9. Table 1 - Tourism in belarus: main indicators
2010
2013
2014
2015
2016
2017
Travel agencies and other reservation services industry
763
1085
1254
1364
1376
1444
Main tourism resources
museums
158
162
157
156
159
theatres 27 28 29
reservations 2
national parks 4
sports facilities
26173
23171
22790
23278
23167
23291

10. Tourism in belarus: main indicators and trends
the tourism industry has been growing consistently;
international tourist arrivals in 2017 grew by 30 % to reach a total of thousand; it was the eight consecutive year of above-average growth in international tourism following the 2009 global economic crisis;
travel for holidays, recreation and other forms of leisure accounted for just over half of all international tourist arrivals and departures; some 10 % of all international tourist reported travelling for business and professional purposes;
one of the major tourism spheres is supporting international conferences, cultural and sport events, in order to increase the number of overnight visits by foreign tourists;

11. Tourism in belarus: main indicators and trends
in 2016 redirection of tourism flows was observed;
in the negative export-import balance by “trips” was observed; values of import exceed those of export;
low level of service, low quality of tourism industries; low development rates of alternative kinds of tourism, low tourism efficiency;
careless, partial and discordant set of information, connected with different aspects of tourism;
important direction of the tourism estimation is development of tourism satellite account

12. Tourism satellite account and possible information sources
The multiplicity of stakeholders involved in the tourism system (international organizations, national, regional, local administrations) implies different needs in terms of typologies of information: from tourism demand to the economic role and impacts of tourism; from statistical data to quantitative analyses.
The final result is an enormous and growing request for information which requires different methodologies. That is why it is necessary to increase efforts to harmonize methodologies, develop tourism satellite account.

13. Tourism satellite account and possible information sources
Nowadays the National Statistical Committee of the Republic of Belarus does preparatory work on development of tourism satellite account.
In 2016 Methodological Recommendations for construction of Tourism satellite account were adopted.
Since , the first tables of this account are calculated.

14. Tourism satellite account (TSA):
The main aims are:
to identify and measure the contribution of tourism to the national economy, in line with the National Accounts framework, and thus allowing comparisons with other economic domains (output and value added by tourism industry, employment, consumption of tourist commodities, demand for commodities)
to achieve total coverage in terms of visitors, their expenditures, and the industries serving visitors, as well as reasoned reconciliation of different statistical sources involved, in order to ensure consistency among the data derived from them

15. Tourism satellite account (TSA):
to become part of the system of information in which individual sources are interconnected (for instance annual data are consistent with monthly or quarterly data; demand data match information on supply)
to develop structural relationships with other macroeconomic frameworks, particularly Balance of Payments, National Accounts.

16. two aspects of tourism are measured:
Main principles:
two aspects of tourism are measured:
consumption of commodities and services by visitors (demand);
production of tourist commodities and services by industries (supply)
tourism expenditures are accounted by kinds of activity (accommodation of visitors, services of public catering entities, transports, tourism industry entities, sporting and others)

17. domestic, inbound, outbound kinds of tourism are considered
Main principles:
supply indicators are indicators tourist product by kinds of activity, observed unit is institution (entity)
domestic, inbound, outbound kinds of tourism are considered
consumption indicators include visitors consumption, gross fixed capital formation related to tourism and consumption of collective non marked services (education, museums, public health)
TSA structure:
domestic, inbound, outbound, national tourism indicators (tables 1-4);
production account by tourism industries (table 5);

18. total domestic demand and supply (table 6); employment (table 7);
Main principles:
total domestic demand and supply (table 6);
employment (table 7);
gross fixed capital formations related to tourism (table 8);
collective non marked services (table 9);
non-monetary indicators (number of hotels, accommodation capacity, arrivals, departures and others)
TSA is formed once a two years (since 2016)

19. Recommended Methodological Frameworks (RMF)
Existing standards:
Recommended Methodological Frameworks (RMF)
International Recommendations for Tourism Statistics 2008 (UN, UNWTO, 2010)
WTO technical manuals on TSA
National methodological recommendations on TSA
Legislation: Lows of Republic of Belarus are “ About state statistics”, “ About tourism”

20. The main sources and instruments for formation TSA:
tourism industry enterprises reports (tourism supply indicators)
system of different surveys, which include establishment samples, households samples (tourism demand indicators)

21. Tourism households surveys
The main aim is asking residents in their usual environment (usually at their home) about tourism expenditures, trips they have taken, after reference period, e.g. the past month.
The following procedures to measure domestic tourism can be used:
Specifically designed surveys to estimate tourism activity of the resident population through comprehensive questionnaires or light telephone surveys (CATI); questions in the latter case need to be simpler and direct
The inclusion of a “tourism module” – a set of interconnected questions to learn more about certain characteristics of visitor behavior – as part of multipurpose surveys.
In Belarus the second procedure is used.

22. Sample survey of households (since 1995)
Survey object is household. Survey is carried out at all country regions and separately in Minsk. It’s annually is covered 0,2 % or 6000 HH.
In this kind of sampling is used territorial probabilistic three- stage sample:
at the first step sampling units are cities and rural soviets;
on the second step – local-polling districts in city and data of the soviet account in rural soviets,
on the third – HH.
Procedure of selection of administrative and territorial units repeats once a 10 years. Selection of polling districts and HH carries out annually.

23. Sample survey of households (since 1995)
The methodology of weighing and raising of the selective data on a general population is based on assignment of each finite unit (HH) the corresponding weight (Вi):
(1)
where p1 – probability of selection of each city and rural soviet;
p2 – probability of selection of each polling district in cities, zones in rural soviet;
p3 – probability of selection of everyone HH within polling
district or a zone.

24. Sample survey of households (since 1995)
Base HH weights are corrected on uninhabited apartments and non-responses by using mathematics methods.
The sample program assumes filling of some questionnaires (living conditions, personal subsidiary plots, education, health, employment) and additional tourism module (trips, duration, tourism expenditures).
Daily and quarterly questionnaires: expenses on food and unfood, payment of services etc.

25. Labour Force Survey (since 2012)
Purposes are:
to obtain empirical statistics on the labour force, employed by kinds of activity, including tourism;
employed and unemployed by sex, regions, rural, urban;
to determine real labour force demand and supply.
Frequency of the results: quarterly and annual
Forecast response rate – 80 %.

26. Sampling Frame is based on the 2009 Census and included:
set of cities in each region
set of village councils in each region
census enumeration districts in each selected city
villages in each selected village council
the household totality in each census enumeration district and village

27. Sampling Design:
survey object is the private households in urban and rural areas for each region, resident persons aged years
observed units:
primary unit – city or village council
secondary unit – census enumeration district or village (zone)
final sampling unit – household
at each stage units are selected with systematic selection with probability, that proportional population number or household number
variables used for the stratification:
administrative districts
urban/rural

28. Weighting procedure: Household weight
1. Design weights are calculated as inverse of overall sampling probabilities:
(2)
where p1 – probability of selection of each city and village council;
p2 – probability of selection of each census enumeration district in cities, zones in rural soviet;
p3 – probability of selection of everyone HH within or a zone.
2. Non-response weights are calculated using the weighting classic method. There are 25 census enumeration districts in cities and 16 village councils (zones)

29. Individual survey person weights is based on iterative weighting:
Weighting procedure:
Individual survey person weights is based on iterative weighting:
Iteration I
a) weights are calculated by sex, five-years group design
b) the first correct coefficient (k1) is calculated; variables of weighting are region, sex, rural/urban
c) the second correct coefficient (k2) is calculated; variables are region, sex, 11 five-years groups.
Weight is equal within each region, five-years group in urban or rural area
Iteration II: follow adjustment of weights
Final individual weight for each five-years group:
(3)
where:

30. Tourism establishment sample surveys
Purposes are:
to obtain detailed information on the tourism, production, employment in the informal sector
to obtain information on the structure of tourism industries, employment
Multipurpose survey is micro-entities sample survey which provides information on the tourism micro-entities

31. MICRO-ENTITIES SAMPLE SURVEY
Sampling frames are:
micro-entities, represented the state statistical reports on the financial results for basic years (report 1-MP (micro));
set of the private farms. The first file is 80 thousands units, sample fraction depends on a character of the initial information, namely: the size of total population, kind of economic activity, region. The second array includes more than 2 thousands farms; it is observed completely

32. Sample design
The combination of univariate and multivariate (multidimensional) sample is used.
To receive optimal sample size for i-th kind of activity and j-th region the next algorithm is used:
1. The set of observed variables is allocated (for example, the wages fund, average number of employees, volume of production, revenues, profitability). Average, total values, variability of indicators are calculated.
2. Statistician chooses sampling method: univariate or multivariate. Univariate stratified samples with simple, proportional and optimal allocation are most often used.
3. It should be executed one of three conditions for applying multidimensional sampling:
- variation coefficient is more than 100 %;
- survey objects are non-uniform on many variables;
- the small size of total population (top limit – 30 – 40 units).
Otherwise univariate sampling should be used: random selection without allocation, simple random sample, proportional and optimal allocation.

33. Sample design
4. It is expediently to use univariate stratified sample, total population is divided by rather homogenous groups. Then different variants of the sample size are executed (minimal is 0.05N, maximal is 0.8N). Minimal error is a main criteria of the determination of sample size. The choice of an optimum way of selection for carrying out of particular survey depends on a survey object and character of the auxiliary information, namely: degrees of uniformity, the sample size, presence of the natural isolated groups, availability of the correlated auxiliary information. It is expedient to approve several sampling designs for the same survey and to choose that from them which gives more precise and unbiased and unbiased estimates.

34. Sample design
5. It is expediently to use multidimensional sample, selection is carried out by cluster analysis:
- total population is partitioned using cluster analysis on homogenous groups to k-variables, i.e. clustering;
- in each received group the leading (basic) variable is determined and subsequent random selection of units is performed.
Optimal sample population is chosen for each cluster, where standard errors of k-variables are criterias of productivity. If the error exceeds admissible bounds, three methods of its reduction may be applied: a) increasing sample population in cluster; b) additional stratification of the enterprises in cluster to a leading variable; c) repetition of clustering, but with larger number of steps.
Sample population is formed once a three-four years, Sample size is entities, sampling fraction is %

35. Statistical weighting
To extrapolate sample data on the total population methods have been used:
Traditional group raising factors (weights), HT
Calibration (GREG-estimators)
Calibration (SYN – estimator)

36. Accommodation surveys
In Belarus visitor survey at collective accommodation establishment (VS) is conducted (since 2015). The purpose is estimation of visitor expenditures. Average expenditures per day are measured Expenditures are: accommodation, food and beverages, transport, car rental, recreation services, culture, others Observed units are visitors and accommodation establishments (hotels and similar) Sampling frames are collective accommodation establishments, represented in state statistical reports ((4-tur (accommodation))

37. Accommodation surveys
Sample design:
univariate stratified random sample is used
establishment stratification is carried out by the main variable – average value of bed-place,
weighting procedure: HT-estimates are used
frequency: quarterly (week)

38. Sample size
Several types of the sample size calculations were executed:
random selection without stratification
random selection with stratification
combined selection (large entities are observed completely)

39. Table 2 – Initial data for the determination of sample size
Accommodation establishments (by average value of bed-place, rubles)
Number of establishments, Ni
average, thousand rubles, xi
Variance
Variation coefficient, %
below 256
225
168.4
32.5
100
349.6
19.0
503 – 750 26
615.3
9.8
750 and more 28
1227.0
42.4
including: 750 – 1000 13
834.8
22.5
1000 – 2000
1415.8
2000 and more 2
2550
26244
6.4
Total
379
325.5
98.6

40. Table 3– Sample size for Vs
Variants
Predicted limited errors, % 8 6 5 3 2
Variant I (without stratification): n
234
280
305
348
365 -
sampling fraction , d % 61 74 80 92 96
Variant II (with stratification):
101
148
182
273
323 27 39 48 72 85
Variant III (combined): n1 31 51 69
142
213 9
14,5 20 40 60 n2 46 58 12 15 18 22 24
2/100
Total sampling fraction, d % 23 44 63

41. Sample size
Variant III is more efficient:
Sampling fraction is 17.9%, error is – 5.9%,

42. Border surveys
In Belarus a Sample survey of individuals at automobile roads checkpoints across the State border is conducted (since 2015)
The main aims are:
to obtain statistical data on the volumes of commodities, imported and exported by individuals crossing the State border;
to obtain information on the tourism expenditures and tourism trips

43. Border surveys
Observed units are individuals (residents and non-residents)
Two forms of blanks for inbound and outbound tourism are used
Automobile roads checkpoints across the State border, used in sample are:
“Brest”, “Bruzgi” (Belarus – Poland)
“Kamenny Log”, “Benyakoni” (Belarus – Lithuania)
“Urbany” (Belarus – Latvia)
“Novaya Guta”, “Mokrany” (Belarus – Ukraine)

44. Border surveys
Frequency: twice a year in the II and IV quarters;
It was first held on April 10th, 2015, from 5 PM to 7 PM, and on April 11th, 2015, from 9 AM to 11 AM
In April total number of responses is 3121 filled questionnaires, or 72.9 %, in October 2016 – 7437, or 51.9 %.
Non-probability sample is used

45. Concluding remarks
The experience of construction of samples in Belarus, has shown:
Sample survey of households:
the priority is given to inclusion of a “tourism module” as part of multipurpose survey (Sample Survey of Households Living Standards, Labour Force Survey)
main problems are: localization of the sample, non- responses (30-40%), the enough high level of errors, sample and non-sample errors, discrepancies between data from these surveys
it is necessary to extend “tourism modules” (to detail expenditures, employment)
Establishment Sample surveys :
the priority is given to the multipurpose survey with module questionnaires (for example, Micro-entities Sample Survey)
It is important to present tourism employment, tourism products by kinds of activity

46. Concluding remarks Accommodation surveys :
main problems are: non- responses, the enough high level of errors, real sampling fraction of visitors is 0.2 % instead of predicted 1.82 % (pilot survey)
Border survey :
frequency and duration are not enough
conducted survey does not solve the problem of hidden and informal tourism; reasons are voluntary nature of the questionnaire, openness of the border with the Russian Federation
Perspectives:
the use of combination of univariate and multivariate samples, quasicausal samples, expert estimates, tertiary sources, increase of sample size of Border surveys, updating existing questionnaires
following surveys are planned: special establishment surveys (tran-sport, recreation), surveys in National Airport, international bus trips

47. Thank you Very much!