EUROPEAN COMMISSION EUROSTAT. Directorate C: National and European Accounts Unit C-3: Statistics for administrative purposes

Transcription

1 EUROPEAN COMMISSION EUROSTAT Directorate C: National and European Accounts Unit C-3: Statistics for administrative purposes Eurostat/C3/GNIC/230 - Final EN Orig: EN BP-13-43a1 24 TH MEETING OF THE GNI COMMITTEE JULY 2012 LUXEMBOURG, BECH BUILDING ROOM AMPÈRE ILLEGAL ACTIVITIES IN NATIONAL ACCOUNTS Point VII.7 on the agenda Document available on CIRCABC at the following address: 1

2 ILLEGAL ACTIVITIES IN NATIONAL ACCOUNTS 1. BACKGROUND 1. Illegal activities in national accounts with a focus on prostitution, the production and trafficking of drugs and smuggling have been dealt with separately in GNIC documents GNIC/108 (April 2008), GNIC/139 (April 2009) and GNIC/181 (April 2010). This document tries to recap these documents and the conclusions of the discussions on these papers in the GNIC. As far as available new findings will be included. The final aim is to agree within the GNIC common recommendations on how to measure and estimate illegal activities for the purpose of national accounts. This is done on the basis of ESA95 and on the background of the use of GNI for EU own resource purposes. 2. The last documents summarised and classified country studies on possible methods to estimate illegal activities as far as available at that time. In the meantime a number of countries have made new studies, developed or improved methods or already implemented estimates of illegal activities in their national accounts. Due to resource restrictions this time-intensive work has not been continued at this stage. However, in case new country studies are available the delegates are invited to provide them to Eurostat in order to be loaded on the restricted GNIC Circa Site in the section on illegal activities. In that way other Member States may consult these studies for further development of their own estimates. In case one country has developed a new method or found a new solution, beyond what was presented in the past, this should specifically be indicated to Eurostat and it will be taken into account. The delegates are invited to provide new country studies to Eurostat in order to have them loaded on the restricted GNIC Circa Site. In case a new method or a new solution has been developed, beyond what was presented in the past, this should specifically be indicated to Eurostat. 2. TRANSVERSAL RESERVATION 3. On 27 January 2012 the Commission notified some transversal reservations on the GNI calculation of all Member States as used for own resource purposes. Reservation VI is on the inclusion of illegal activities in national accounts: VI. Inclusion of illegal activities in national accounts The ESA 95 provides that illegal activities (e.g. prostitution and production of drugs) fall within the production boundary of national accounts (see paragraphs 1.13(g) and 3.08, last paragraph). The sources and methods used need to be reviewed in order to ensure that illegal activities are properly included in the national accounts. A GNP Committee Task Force concluded its work and presented its results to the GNP Committee in December The Task Force encouraged Member States to develop studies and research, particularly concerning the three activities which should have the most significant impact on GDP/GNI: drugs, prostitution, and smuggling of alcohol and tobacco. 2

3 The GNI Committee discussed the issue of the inclusion of prostitution, drugs and smuggling between 2008 and 2010 (see in particular GNIC/108, GNIC/139 and GNIC/181). Following the conclusions of the Task Force and the subsequent discussions mentioned above, the coverage of the relevant illegal activities and the appropriateness of sources and methods used need to be reviewed. 4. In order to resolve this reservation a detailed comparative analysis of the solutions adopted by Member States is necessary. The results from this comparative analysis might also identify the need for changes in the calculation of GNI. Our aim is to discuss and agree within the GNIC some minimum standards for the inclusion of illegal activities in national accounts which should be fulfilled before lifting the reservation. This includes also the problem of possible double counting (see sections 6 and 7). 3. THE NEED TO INCLUDE ILLEGAL ACTIVITIES IN NATIONAL ACCOUNTS 5. ESA g and 3.08 (last paragraph) clearly require that illegal activities, which fulfil the characteristics of an economic transaction (see ESA and 1.42), should be included in the measure of GDP/GNI. Illegal activities are defined as those transactions which are generally forbidden by national law. The delineation between what is considered legal or illegal can vary between different countries. 6. Following the exhaustiveness decision (Commission Decision 168/1994, article 2) all legal economic activities have to be included in an exhaustive estimate of GDP/GNI. This includes also all economic activities which are legal themselves but hidden from national authorities, which may be an illegal activity in itself. We may take the example of prostitution in a country where prostitution in brothels is legal but street prostitution is illegal. The legal part of prostitution should fully be included in the estimates, even if partly not disclosed to authorities or statistics. The street prostitution part should be estimated as an illegal activity. Very important is also that the criterion of an economic transaction is fulfilled, i.e. an interaction between units by mutual agreement. This excludes from the scope of our consideration illegal activities like trafficking in human beings or women forced to prostitution. 7. There is no general provision in ESA95 that gives the possibility to exclude certain elements from the production boundary because they are difficult to measure or estimate. On the other hand, the GNIC should assess reliability, comparability and exhaustiveness of GNI and its components taking account of the cost-benefit principle. The application of this principle should avoid committing disproportionate resources to calculate insignificant items. Results from country studies indicate that the impact of illegal activities on GDP/GNI in the EU is not extremely big (less than 1%, see table 2 below) but it cannot be considered insignificant. Do the delegates agree with this interpretation? Are there differing views, e.g. a specific uncertainty margin of an estimator at which data should generally not be included anymore in national accounts estimates? 3

4 4. WHICH ILLEGAL ACTIVITIES TO INCLUDE 8. At the GNP Committee meeting of 7th November 2001 a "Task Force of the GNP Committee on Illegal Activities" was set up. In its final report (see minutes of 11th December 2002) the task force came to the conclusion that for all Member States trafficking of drugs, prostitution and smuggling of alcohol and tobacco will have the most significant impact. At least these three areas should be investigated in more detail. As a consequence the studies undertaken by countries concentrated on these three areas of illegal activities. 9. Does the GNIC agree to the following recommendation: Due to their significance in the EU the following illegal activities should be included in national accounts estimates: Prostitution; Production and trafficking of drugs; Smuggling of alcohol and tobacco products. If any of these three is considered insignificant and not estimated for one Member State, this needs to be justified properly. In case any other illegal activity is significant in relation to the total level of GDP/GNI, this activity should also be included in the estimates. 5. CURRENT SITUATION 10. An overview of the current situation of recording illegal activities in national accounts is provided in the following table: 4

5 Table 1: Overview - recording of illegal activities Current situation as of July 2012: Already includes estimates of Does currently not include estimates of Included where? illegal activities in NA illegal activities but plans to do so as of In national Only for EU own Country ProstitutionDrugs Smuggling Other Prostitution publication resources delivery Drugs Smuggling BE N N N N N N 30/09/ /09/2014 N BG Y Y N N CZ DK Y Y Y N* Y* Sept Sept Sept DE Y N N Y N Possibly in September 2014 EE Y Y Y Y N EL ES FR Y N Y Y N N IE N Y Y Y N? Does currently not plan to include estimates of illegal activities Does not plan to include drugs. IT N N N N N N Does not plan to inclde illegal activities. CY LV N N N N N N Sept Sept Sept LT Y Y Y N Y LU N N N N N N Waiting Waiting not relevant Waiting for GNIC decision. HU Y Y Y Not yet decided MT N N N N N N Currently no plan to include i.a. NL N N N N N N Does not plan to include the estimates AT Y Y Y Y PL N N N N N N Part concerning smuggling is under preparation. PT RO N N N N N N SI Y Y N Y N SK Y Y Y N Y N FI Y Y Y N Y N SE Y Y Y amblin Y N UK N N Y Y** Y N N N IS NO Y Y Y Y CH N N N Y 30/06/ /06/ /06/2012 Does not plan to include illegal prostitution and drugs. * Inclusion in national publ. from Sept ** Other adjustments such as Missing Trader Intra Community VAT fraud, evasion, etc., each varying between %, and in total %. 11. In table 2 a summary is given of results on the estimates of illegal activities in relation to GDP (situation as of July 2012). The data must be interpreted carefully as some of the studies have a preliminary character. Comparing the data between countries is difficult because sometimes different categories of illegal activities were covered and the delineation of what is considered legal and illegal varies between countries. It should also be noted that even if prostitution might generally be legal in a country, certain forms might be considered illegal (e.g. street prostitution or 5

6 brothels). The estimate provided may refer to the illegal part only or to the total value. 12. The overview is based on the summary results available from country studies on the inclusion of illegal activities in national accounts. The underlying studies are listed in Annex I. More details on some country studies are also available in documents GNIC/108 (April 2008), GNIC/139 (April 2009) and GNIC/181 (April 2010). In the meantime some countries have already included estimates on illegal activities in their national accounts. This has also been included in the table as far as available to Eurostat. Table 2: Share of illegal activities in GDP - first estimates Country Source or Study Legal situation Reference Share of GVA in GDP Comment L=legal, I=illegal period Prostitution Drugs Smuggling Prostitution Drugs Smuggling Total AT NA estimates Sep L I I % 0.06% 0.01% 0.15% BE L I I % 0.28% 0.37% BG NA estimates March 2012 L I I % 0.03% CZ CZ-2005 L I I % 0.23% 0.10% 0.56% DE DE-2005 L I I % n.a. * 0.02% 0.33% NA estimates % * Very week data sources: estimate not reliable DK DK-2005 L I I % 0.06% 0.02% 0.16% EE EE-2008 L I I % 0.21% 0.15% 0.38% EL EL-2006 L I I % 0.09% 0.04% 0.89% FI NA estimates March 2012 L I I % 0.06% 0.01% 0.10% FR L I I % 0.02% HU NA estimates Sep L I I % 0.33% 0.82% IE NA estimates 2011 L I I 2011 n.a. 0.60% 0.26% 0.86% IT IT-2010 L I I % 0.37% 0.04% 0.99% LU LU-2005 L I I %* 0.11% 0.25% % 0.08% LT 2011 NA estimates I I I % 0.07% 0.19%* 0.34% LV LV-2010 L I I % 0.02% 0.05% 0.24% * Very week data sources: estimate not reliable * Including illegal production of alcoholic beverages inside the country NL NL-2012 L I I % 0.22% 0.02% 0.32% Share of other illegal act. is 0.07% NO NO-2011 I I I % 0.04% 0.00% 0.04% SE SE-2008 I I I % 0.07% 0.05%* 0.13% * Including illegal production of alcoholic beverages inside the country % 0.07% 0.04% 0.13% Adding Illegal gambling: total of 0.16% SI 2012 NA estimates L I I % 0.23% 0.36% SK 2012 NA estimates L I I % 0.33% 0.05% 0.56% UK NA estimates L I I % % n.a. = not available 6

7 6. HOW TO MEASURE AND ESTIMATE ILLEGAL ACTIVITIES 13. National accountants face three major problems when trying to include illegal activities: 1) availability of reliable sources and measurement problems; 2) best estimation methods and accuracy of the estimates; 3) the risk of double counting of production, expenditure or income when adding explicit estimates of illegal activities. It has to be noted that national accounts estimation and balancing procedures can result in implicitly including a part of illegal activities in GDP/GNI. However, this fraction is difficult to identify and cannot be estimated exactly. 14. Experience from studies undertaken by countries show that the results can vary strongly, depending on the approach taken and the assumptions used. Therefore, below Eurostat presents proposals on possible best practices on how to measure and estimate relevant illegal activities. This assessment by Eurostat is based on an analysis of country pilot studies as presented in documents GNIC/108 (April 2008), GNIC/139 (April 2009) and GNIC/181 (April 2010) and discussions in the GNIC. An agreement on the methods would assure at least a minimum of harmonisation and consistency between EU countries from that side. 15. This chapter concentrates on best practices to measure and estimate the value of illegal activities before balancing the different approaches to GDP and irrespective of possible strategies to avoid double counting (see chapter 7) Prostitution 16. For prostitution services we face a particularly blurred situation as prostitution, at least in certain forms, is legal in many countries. As such there is a clear requirement to include these legal prostitution services in national accounts estimates, also comprising the concealed part. At the same time all those prostitution services that do not fulfil the characteristics of an economic transaction, i.e. are performed without mutual agreement, have to be excluded from the measurement. 17. A number of countries have already developed estimation methods and some already include estimates of prostitution services in their accounts. Some of them are described in country studies (see Annex I). More details on the country studies are also available in document GNIC/108 (April 2008) and in particular from the Excel file Illegal Activities - annex (not for printing) on Circa. Countries are invited to consult this file in order to compare their own approach with other countries in a similar situation 18. It is obvious that standard data sources, such as administrative records (e.g. tax, social security) or business surveys do not provide reliable data on the activity, even for the legal part. However, in all countries at least some data sources are available. In most cases these are ad hoc studies on the subject or irregular reports by police, government institutions or welfare organisations. In some countries regular reports are available. Generally, the data sources available do not fit national accounts purposes as such but can be used as basis for further estimates. 19. Demand side studies do not seem to be suitable for estimates of services rendered by prostitutes. Either the methodology is not suitable for national accounts purposes or 7

8 the results are not representative or reliable enough. Estimates based on the supply of prostitution services generally seem to be more suitable for national accounts estimates. 20. Basic model for estimates of prostitution services 21. Starting point is an estimate on the number of prostitutes in the country. Depending on the specific situation in the country and available data sources this should be broken down by different types of prostitution services. In some countries it could be useful to further differentiate between legal and illegal prostitution. 22. For each type of prostitution the average number of contacts per prostitute per period should be estimated in order to achieve the total number of contacts. 23. Estimates on the average prices for the different types of prostitution seem to be relatively easily available. With this information Sales/Output can be calculated. GO (P.1) = Sales = Number prostitutes x Number contacts per prostitute x Price 24. Intermediate consumption is relevant for some types of the prostitution activity. It should therefore be estimated and deducted in order to estimate GVA. It is likely that certain products recorded as IC for prostitution services are included in HFCE. Therefore, if relevant, a corresponding counter-booking needs to be done in HFCE for the products concerned. Furthermore, if rental payments are recorded as relevant IC for prostitutes (e.g. payments to owners of brothels, clubs, apartments) consistent recording in the output for rental activities needs to be assured. If this transaction is not or not fully included, it needs to be added separately as rental output. GVA = Sales IC 25. Prostitutes are generally considered as "self-employed", i.e. no employer employee relation is assumed (also for clubs, brothels, etc.). 26. The prostitution services could be offered by residents and non-residents, i.e. prostitutes working for less than one year and not repeatedly in the country (no centre of economic interest). This share of non-resident prostitutes needs to be estimated. Only the output and GVA of the residents has to be recorded. Resident Sales = Share of residents x Sales Resident GVA = Share of residents x GVA 27. The resident GVA created is at the same time considered as gross operating surplus. 28. We assume that all prostitution services are consumed by households. No relevant amounts appear in IC of enterprises. 29. The Resident Sales can be purchases by resident households or non-resident households. The share consumed by non-resident households (tourists) needs to be estimated. This estimate can be based on expert estimates or specific studies. Resident Sales x Ratio = Sales to non-resident HH = EX (P.6) Resident Sales x (1-Ratio) = Sales to resident HH 8

9 = HFCE (P.3) 30. One further element that could be relevant for some countries are purchases of prostitution serviced by resident households from non-resident prostitutes on the domestic territory but also as tourists abroad. The element of expenditure as tourists abroad is in theory already now included in the tourism balance. The corresponding amount should be included in HFCE and imports (IM) and as such does not impact GDP. Therefore, it is not necessary to separately identify this item for the national accounts. However, this element could be of interest for cross-country comparisons and consistency checks. 31. For the transition from GDP to GNI the crucial question is which part of the generated GVA and income is finally attributed to non-residents and transferred abroad. The part transferred abroad should be recorded as property income paid to the rest of the world (and vice versa for the destination country). An estimate of the transferred share can be based on expert estimates or specific studies. A reasonable approximation might be done with the share of foreign prostitutes. Furthermore, it needs to be investigated carefully how far these flows are already included in existing estimates (balance of payments flows). 32. As consumption patterns are relatively stable an extrapolation of a benchmark year result for a number of years is generally acceptable. 33. Questions to the GNIC: Do the GNIC members agree with the basic model for estimates of prostitution services? 34. Recommendations for estimates of prostitution services 35. A supply side based approach is recommended. If another approach is preferred this needs to be justified carefully. (R1) 36. An estimate of the number of prostitutes needs to be done in a suitable breakdown for different types of prostitution (e.g. street, brothel, apartment, club). Similarly it could be useful to further differentiate between legal and illegal prostitution, but this is not obligatory. At that point also prostitution services clearly performed without mutual agreement should be excluded from the measurement. (R2) 37. Estimates on the average number of contacts per prostitute per period should be based on expert estimates or specific studies. (R3) 38. Average prices for the different types of prostitution can be taken from expert estimates or specific studies. (R4) 39. Estimates on the share of IC for certain prostitution activities, as far as considered relevant, need to be added. This can be based on expert estimates or specific studies in the field. Consistency of the recording of IC with HFCE and with rental output should be verified if relevant amounts have been recorded as IC. (R5) 40. The share of non-resident prostitutes needs to be estimated, if considered relevant. It needs to be noted that these non-residents are only those operating less than one year 9

10 in the country (without a centre of economic interest). This is different from foreigners working as prostitutes in the country. This estimate can only be based on expert estimates or specific studies. Care must be taken if these non-resident prostitutes had been included in the initial estimate of the total number of prostitutes or not. (R6) 41. The share consumed by non-resident households (tourists) should be estimated, basing on expert estimates or specific studies. (R7) 42. It should be verified that purchases of prostitution serviced by resident households from non-residents, in particular as tourists abroad, are properly included in the tourism balance and recorded as HFCE and IM. (R8) 43. The share of income generated and transferred abroad should be estimated. This estimate can be based on expert estimates or specific studies. A reasonable approximation might be done with the share of foreign prostitutes, possibly in a break down by types of prostitution. It should be investigated how far income generated from prostitution and presumably transferred abroad is already included in the existing estimates on the transition from GDP to GNI. (R9) 44. Extrapolations from a benchmark year should be done separately for the number of prostitutes and prices. Prices might be updated more frequently. If possible this should be done in a breakdown by different types of prostitution. (R10) 45. Questions to the GNIC: Do the GNIC members agree with the recommendations above? Is this considered an acceptable basis for harmonised estimates of prostitution services of sufficient quality? 6.2. Production and trafficking of drugs 46. The production and trafficking of drugs is illegal in all EU/EFTA countries. All country pilot studies covered the following types of drugs: cannabis, ecstasy, amphetamines (speed), cocaine and heroin (opiates). In some countries in addition the following seem to be relevant: LSD (hallucinogens), Subutex (burprenorphine, illegal use) and narcotic medicines. 47. Based on available pilot studies (see Annex I) the following country types in respect of transactions in drugs seem to be relevant: Type I: only imports, no (significant) production, no (significant) export; Type II: only imports, no (significant) production, no (significant) export, transit country for certain drugs; Type III: imports, production of some drugs for domestic market only, no relevant exports, transit country; Type IV: imports, domestic production of some drugs for domestic market and for export, re-export of imported drugs, transit country. 10

11 48. More details on the country studies are available in document GNIC/139 (April 2009) and in particular from the Excel file Illegal Activities - annex (not for printing) on Circa. Countries are invited to consult this file in order to compare their own approach with other countries in a similar situation. 49. Due to the nature of these activities the data situation for trafficking and production of drugs is particularly difficult. Estimates must rely on the following sources: administrative information from police, customs or ministries (in form of reports or expert opinions), reports from health/welfare or specific drug institutions, research reports, information from drug assistance or other non-profit organisations, or international research projects. The problem with these sources is that they don t fit national accounts purposes as such, they differ in numerous aspects (e.g. population covered, periods covered, type of drug users covered) and are difficult to compare or to combine, some are only one-off studies or not available regularly. On the other hand, the positive aspect is that in all countries some information on production, trafficking or consumption of drugs is available and experts seem to have a general picture of the situation. 50. Important information can also be deducted from two international sources, i.e. the annual report by the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) and the annual report by the United Nations Office on Drugs and Crime (UNODC). Both reports are available on the respective websites. They provide some information on individual countries but also give some recent background information on the drugs market and enable a valuable comparison of national estimates with other countries. 51. The studies GENACIS (Gender, Alcohol, and Culture: An International Study) for 2002 and ESPAD (European School Survey Project on Alcohol and Other Drugs) for 2003 were international studies in which also a number of EU/EFTA countries participated. These may therefore also serve as reference for some countries. 52. Due to available data sources, in some studies it has proven useful to differentiate the analysis between 'heavy abuse' (or 'hard drug users', or 'problem drug users') and 'other abuse'. Though the definitions are not 100% identical in the different studies, the first group tends to comprise daily/almost daily users and the second group from brief/experimental to more regular users (but not daily/almost daily). 53. Basic model for trafficking of drugs 54. The case of imports of drugs is analysed. 55. In a first step total domestic consumption of the drugs needs to be estimated; in a breakdown by type of drug, type of user or other factors, as best suitable to the situation in the country and available data sources. 56. We assume that all drugs are consumed by households. 57. Direct imports of drugs by households are not taken into account or considered negligible. 58. The value of expenditure on drugs can be estimated with average street prices: Expenditure = Quantity consumed x average street price 11

12 = HFCE (P.3) 59. Due to dilution the imported quantity can be different from the quantity consumed for certain drugs. A purity factor indicates the average dilution of "pure" drugs before putting them on the market: Import Quantity = Quantity consumed / purity factor 60. Direct imports of drugs by households are included in this estimate. 61. The import value can be estimated with average "wholesale" prices of the drugs: IM (P.7) = Import Quantity x average wholesale price 62. Output is the trade margin achieved by the trafficker for drugs sold to households: GO (P.1) = Trade Margin = Sales IM 63. The trade margin implicitly also includes a profit margin achieved by diluting the drug, if done so. 64. Intermediate consumption of traffickers consists of costs for transport and storage. This needs to be estimated on basis of expert estimates, if not considered negligible. GVA = Trade Margin IC 65. The drug traffickers are considered to be self-employed (resident units or notional resident units). GVA generated on the domestic territory is therefore identical to gross operating surplus for these units. 66. For the transition from GDP to GNI the curial question is which part of the generated GVA is attributed to resident and which to non-resident units. The non-resident part will be transferred abroad and should be recorded as property income paid to the rest of the world (and vice versa for the destination country). An estimate of the nonresident share can only be based on expert estimates. Furthermore, it needs to be investigated carefully how far these flows are already included in existing estimates (balance of payments flows). 67. Basic model for domestic production of drugs 68. The quantity produced on the domestic territory of a specific drug needs to be estimated, e.g. as share of total domestic consumption or independently. 69. Output is estimated with average street prices: GO (P.1) = Quantity produced x average street price. 70. Output is sold to households only; in some cases a part of domestic production may be exported. 71. The intermediate consumption for drug production needs to be estimated as a share in output; based on expert estimates. GVA = GO IC. 12

13 72. Total GVA would be the sum of GVA from trafficking and domestic production. 73. Questions to the GNIC: Do the GNIC members agree to the basic models for trafficking and production of drugs? Are there any needs for corrections or amendments? 74. Recommendations for estimates of the production and trafficking of drugs 75. A demand side based approach is recommended. If another approach is preferred this needs to be justified carefully. (R1) 76. A supply side based approach, based on quantities seized, is considered as potentially too instable. The seizure quota can vary sharply and is dependent on many external factors (e.g. customs strategy and priorities, customs equipment and staff numbers, strategy of drug traffickers). Particularly for countries with a high degree of transit trade destination quotas (i.e. the share destined for the country) need to be assumed and drug routes may change quickly. In case seizures are used it is recommended to use a longer time series and statistical methods to determine average seizure quotas under similar conditions (moving average). However, at least for benchmark years it is recommended to confront the results from the demand side with supply side estimates. (R2) 77. Quantities and relevant prices should be estimated separately. (R3) 78. First step in the consumption based approach is the estimate of the quantity of drugs consumed. This is at the same time also the most difficult step and can be split into estimates of the number of users, frequency of use and average quantity consumed per use per year. Depending on the specific situation in a country and available data sources the estimate should be stratified by type of drug, by age groups, by type of user (regular, brief or experimental user), etc. (R4) 79. Results for a benchmark year can be extrapolated for a number of years with suitable indicators (e.g. expert opinion on the development of number of users in a certain strata), as long as stable consumption patterns can be expected. (R5) 80. Information on average street prices and import prices (wholesale prices) seem generally to be available from studies or experts working in the field. This can reasonably be used for the estimate. (R6) 81. Extrapolation of drug prices for a number of years must be justified. Generally no suitable price indices are available and prices for some drugs can vary sharply. (R7) 82. The share of IC (for transport, storage or production) in output needs to be estimated on basis of expert estimates and information from similar activities, if not considered negligible. (R8) 83. The share of non-residents active in trafficking of drugs within the country needs to be assumed based on expert estimates. (R9) 84. It should be investigated how far income generated from trafficking of drugs by nonresident units and presumably transferred abroad is already included in the existing estimates on the transition from GDP to GNI. (R10) 13

14 85. Questions to the GNIC: Do the GNIC members agree with the recommendations above? What needs to be added, what should be more specific? Is this considered an acceptable basis for harmonised estimates of trafficking and production of drugs of sufficient quality? 6.3. Smuggling of alcohol and tobacco products 86. Smuggling is defined as the illegal import or export of goods which are legal products in themselves. The incentive for smuggling stems from legal settings, i.e. the prohibition or restrictions on trading the good in a country, or unusually high price differences, very often induced by import duties or excise taxes. This definition could cover a large variety of goods. However, following the recommendation of the GNP Committee Task Force on illegal activities (see minutes of the meeting on 11 December 2002) the analysis is restricted to the most relevant cases identified as the smuggling of alcohol and tobacco products. 87. Even more than for other illegal activities, smuggling points to the blurred borderline between exhaustive estimates for national accounts and the inclusion of illegal activities. In fact exhaustive estimates of imports, exports or consumption should already now cover also smuggled goods, e.g. textiles mis-declared to customs in order to avoid taxes and duties. The only reason to treat the smuggling of alcohol and tobacco differently is the quantitative relevance. And this again could be very different from country to country, as well as the degree to which existing estimation procedures already now cover parts of smuggled products. 88. Based on the country studies (see Annex I) the following country types in respect to smuggling seem to be relevant for the EU/EFTA: Type I: only illegal imports, no exports, no illegal domestic production, direct imports by households (travel expenditure); Type II: illegal imports, no exports, no illegal domestic production, direct imports by households (travel expenditure), significant transit trade; Type III: no significant illegal imports, some illegal production, some illegal exports, no relevant direct imports by households (travel expenditure); Type IV: only illegal imports, no exports, some illegal domestic production, direct imports by households (travel expenditure); Type V: illegal imports, no significant exports, some illegal domestic production, direct imports by households (travel expenditure), significant transit trade. 89. More details on each country study are available from document GNIC/181 (April 2010) and in particular from the Excel file Illegal Activities - annex (not for 14

15 printing) on Circa. Countries are invited to consult this file in order to compare their own approach with other countries in a similar situation. 90. The illegal domestic production concerns the brewing of strong alcoholic drinks or beer by households. This activity has to be covered anyway and is independent of the smuggling investigated here. Therefore ways to estimate illegal domestic production will not be investigated further here. 91. Estimates on direct imports by households (travel expenditure) already now have to be produced for the BoP and tourism expenditure. Mostly these estimates are based on questionnaires. This item of course may need to be taken into account when investigating smuggling of alcohol and tobacco products. However, the estimation method for this item will not be investigated here. 92. Basic model for smuggling of alcohol and tobacco products: 93. The contraband goods illegally cross the border, i.e. without or with wrong declarations to the authorities. This always involves a change of ownership for the smuggled goods between non-residents and residents (see ESA (k)). Up to the border this is exports by the exporting country. However, export of smuggled goods seems not be relevant in most EU/EFTA countries (see country types above). EX (P.6) = Quantity x Price(fob); IM (P.7) = Quantity x Price(fob) 94. The export/import price can be approximated with the wholesale price of the relevant good in the exporting country. 95. The transit of contraband through a country is excluded from export/import figures. 96. We assume that the smuggled goods are sold directly to resident households and to households only. With this simplifying assumption we exclude that a part of the smuggled goods enter the legal distribution chain (retail sales or hotels and restaurants). No country study indicated that this is particularly relevant. 97. Counterfeit products for alcohol and tobacco products are not identified separately; even so they might partly be included in the estimated quantities for smuggled goods. No price difference between original and possible counterfeit products is assumed. 98. The sales value of smuggled goods is: Sales = Quantity x Price(sale) Sales = HFCE (P.3) 99. The average sales prices of smuggled alcohol or tobacco products need to be estimated. Expert estimates can be used to determine this average price The output consists only of the trade margin achieved for the smuggled goods: Trade Margin (P.1) = Sales IM 101. Intermediate consumption of smugglers consists of transport and storage. The share of IC needs to be estimated on basis of expert estimates and information from similar activities, if not considered negligible. 15

16 GVA = Trade Margin IC 102. The smugglers are considered as self-employed (resident units or notional resident units). GVA generated on the domestic territory is therefore identical to gross operating surplus for these units For the transition from GDP to GNI the curial question is which part of the generated GVA is attributed to resident and which to non-resident units. The nonresident part will be transferred abroad and should be recorded as property income paid to the rest of the world (and vice versa for the originating country). It needs to be investigated carefully how far these flows are already included in existing estimates For countries with illegal production of alcohol or tobacco products the standard requirements, as for other hidden production, apply. GVA generated by this production and possibly a trade margin need to be estimated For countries with relevant export of contraband the related trade margin (up to the border) needs to be estimated Questions to the GNIC: Do the GNIC members agree with this basic model? Are there any needs for corrections or amendments? 107. Recommendations for the estimates on smuggling of alcohol and tobacco products: 108. A demand side based approach is recommended. If another approach is preferred this needs to be justified carefully. (R1) 109. A supply side based approach, based on quantities seized, is considered as potentially too instable. The seizure quota can vary sharply and is dependent on many external factors (e.g. customs strategy and priorities, customs equipment and staff numbers, strategy of smugglers). Particularly for countries with a high degree of transit trade, this needs to be taken into account carefully. In case seizures are used it is recommended to use a longer time series and statistical methods to determine average seizure quotas under similar conditions. However, quantities determined from the supply side can be very useful to confront these, at least for benchmark years, with the results from the demand side. (R2) 110. Quantities smuggled and relevant prices should be estimated separately. (R3) 111. For countries exporting relevant quantities of contraband, the export value should be calculated as EX (P.6) = Quantity x Price(fob) (R4) 112. Imports of contraband should be calculated as IM (P.7) = Quantity x Price(fob) (R5) 113. Quantity smuggled: The first step is to estimate total consumption on the domestic territory by households of the goods in questions. This estimate can be based on available information on consumption patterns in relevant breakdowns (e.g. by age, 16

17 by sex). Also judgements by experts working in this field can be useful. An alternative or additional approach is the analysis of waste in order to identify the share of non-declared cigarettes in total consumption. 1 The outcome is an estimate of total domestic consumption by households of alcohol and tobacco products in a relevant product breakdown. Results for a benchmark year can be extrapolated for a number of years with suitable indicators (e.g. population development stratified by age or sex) as long as stable consumption patterns can be expected Generally good statistics on legal sales of alcohol and tobacco products are available. After deduction of quantities directly imported by households (travel expenditure) the quantity smuggled can be estimated: Quantity smuggled = total domestic consumption legal sales direct imports by households (travel expenditure) (R6) 115. Import price: The export/import price can be approximated with the average wholesale price of the relevant good in the main exporting country or countries (countries of origin). If this is not available estimates on the basis of domestic prices and reasonable assumptions can be used. (R7) 116. The estimate of average sales prices for the smuggled goods can be based on available statistics or expert estimates. (R8) 117. Price estimates for a benchmark year can be extrapolated for a number of years with suitable price indices, as long as no structural change can be expected. (R9) 118. The share of IC (transport, storage) needs to be estimated on basis of expert estimates and information from similar activities. (R10) 119. It should be investigated how far income generated from smuggling by non-resident units and presumably transferred abroad is already included in the existing estimates on the transition from GDP to GNI. (R11) 120. Questions to the GNIC: Do the GNIC members agree with the recommendations above? Is this considered an acceptable basis for harmonised estimates of smuggling of alcohol and tobacco products of sufficient quality? 7. THE RISK OF DOUBLE COUNTING 7.1. Dimension of the problem 121. The first major problem for including estimates of illegal activities in national accounts is the reliability of the estimates. Due to the data sources, methods and 1 For more details on the approach see on Circa: DE 2005(EN1) - EN - Amendment Germany Report 2005 Drugs-Tobacco-Alcohol, or AT Report Illegal Production (EN) 17

18 numerous assumptions or expert opinions used the uncertainty margin is particularly high. Precise estimate on illegal activities will never be possible On the other hand, taking into account the results of the pilot studies (see section 5), the estimates on individual illegal activities are not particularly high. This consequently means that a possible mistake is also limited in relation to the total GDP/GNI. Of course the objective still is to achieve the best possible estimator and to reduce the error margin as much as possible. But the dimension of the mistake should be taken into account when discussing the inclusion of illegal activities into national accounts The second major problem is a possible double counting. Due to the sources and methods used for national accounts on the one hand and possible strategies by criminals, particularly money laundering, on the other hand it is very likely that illegal activities are at least partly already included in the estimates. Therefore we are facing the dilemma that a part might be missing if we leave out illegal activities or we add too much when including illegal activities. The first best solution is of course to add only the actually missing part. Some ideas on how to approach this are presented in the following. However, the dimension of the problem should also here be kept in mind Business register problem 124. One argument is that units engaged in illegal activities may already be included in the business register, either in the correct NACE or hiding their activity in a different NACE. This is of course also linked with the money laundering problem discussed below However, national accounts face a similar problem also for concealed legal activities. If for example units engaged in legal prostitution or massage clubs are included in the business register, their output etc. needs to be taken into account when adding estimates for legal activities. The aim is to reach exhaustive estimates False recording problem 126. Double counting may appear because illegal activities are recorded falsely in the source statistics: IC wrongly recorded as HFCE (for prostitution services, for 'home' production of drugs, for transport activities in smuggling or trafficking). Households' expenditure on prostitutes, drugs or smuggled goods may be included in other categories of the HBS. Imports or Exports of drugs or contraband may be declared as different goods and as such enter the statistics wrongly (e.g. drugs in fruit containers) All these problems are not unique to illegal activities, though they may be particularly relevant there. The wrong recording of IC as HFCE needs to be corrected. This could be done by defining the major activities concerned, e.g. 18

19 prostitution in apartments or home growers of cannabis, and rebooking relevant IC identified in the estimates False reporting of the types of expenditure under consideration in the household budget survey is a well-known problem, also for legal consumption. A careful analysis of under- or wrong reporting in the HBS is anyway necessary for these activities A wrong declaration in export or import of goods is difficult to capture in statistics. However, it could be argued that the values involved are not relevant to significantly distort the aggregates concerned and would contribute to a considerable double counting It is difficult to give general recommendations on how to tackle these problems because it depends very much on the sources and methods implemented in each country. The best tool to verify at least the overall consistency of the estimates and adjustments is a detailed supply/use framework Money laundering problem 131. One important argument against the inclusion of illegal activities in national accounts is that the money earned needs to be laundered and therefore at least partly appears, under wrong declaration, in the legal economic circuit. None of the country studies tackled this problem in more detail Firstly, it needs to be stated that the money laundering problem is broader and is not restricted to the illegal activities discussed in this paper. However, prostitution, drugs and smuggling are profitable business and may contribute to a significant fraction of the money laundered. And the risk of possible double counting remains an open issue. The question to the GNIC remains if the money laundering problem should generally be analysed in more detail, or only restricted to drugs, smuggling and prostitution or not at all At the moment Eurostat does not have available more detailed information on the money laundering problem in any of the EU/EFTA countries, particularly which routes the illegal money takes within and outside the country, which activities are concerned, is it productive activities or purely financial transactions A first step could be an analysis of actions at national and EU level to fight money laundering. 2 Apparently the following activities seem to be prone to money laundering: The gambling sector, casinos, Financial agents, particularly where they provide cross-border financial services, Real estate/letting agents, 2 See e.g.: "Report from the Commission to the European Parliament and the Council on the application of Directive 2005/60/EC on the prevention of the use of the financial system for the purpose of money laundering and terrorist financing", COM(2012) 168 final, from

20 Dealers in precious stones and metals Countries should investigate the statistics on these activities in more detail, also in a supply/use matrix. This could give first hints on the relevance and the possible size of the problem An open question not analysed by now is a possible impact of money laundering in international financial flows recorded in the balance of payments Recommendations as concerns the possible double counting of illegal activities 138. The EU/EFTA countries are invited to report on their approach to avoid possible double counting when including illegal activities in national accounts, in particular: Identifying units likely engaged in illegal activities included in certain activities in the business register. Adjustments for possible false recording of IC in HFCE. Analysis of statistics on activities prone to money laundering Questions to the GNIC: Does the GNIC agree with the above recommendations as concerns the possible double counting of illegal activities? Should the money laundering problem be analysed in more detail and with which scope? Can we expect any benefit from this analysis? Would it be possible/advisable to analyse the activities prone to money laundering in more detail? Cross-country comparisons may help to improve the quality of the estimates of illegal activities in general, but also in respect of the double counting problem. On the other hand this would increase the requirements on the statistics to be produced. 20

21 Annex I: Background Documents and Studies (available on Circa) Minutes of the meeting on 11 December 2002 by the Task Force of the GNP Committee on Illegal Activities. GNIC/024: Document presented at the 2 nd GNI Committee meeting in March GNIC/091: Document presented at the 10 th GNI Committee meeting in July GNIC/108: Document presented at the 13 th GNI Committee meeting in April GNIC/139: Document presented at the 16 th GNI Committee meeting in April GNIC/181: Document presented at the 19 th GNI Committee meeting in April Country Studies AT-2006: " Report on estimating illegal production in the Austrian National Accounts in accordance with ESA 95", by Ferdinand Kaßberger, Statistics Austria, 20 December CZ-2005: " Data Collection: Illegal Activities", Report by the Czech Statistical Office, April DE-2005: "Möglichkeiten und Grenzen der Quantifizierung von Drogenwirtschaft sowie Tabak- und Alkoholschmuggel in den VGR", Statistisches Bundesamt, Wiesbaden, September English translation and an amendment document are available on Circa. DK-2005: "Estimating Illegal Activities in Denmark"; study by Statistics Denmark, September EL-2006: "Sources and methodology for estimating illegal activities"; study by the National Statistical Service of Greece, National Accounts Division. FI-2008: "Finland's non-observed ecnonomy"; report by Statistics Finland, 30 June HU-2009: "The estimation for illegal activities"; extract from the GNI Inventory, Chapter 7, June IT-2006: "Estimation of illegal economic aggregates according to ESA95"; study by the Instituto Nazionale di Statistica, Italy, September [only Italian original]. IT-2010: "Non-observed economy in National Accounts. Integration of illegal activities data by industry and institutional sectors in the main aggregates of national accounts", study by the Instituto Nazionale di Statistica, Italy, LU-2005: "Etude de faisabilité sur l'introduction des activités illégales dans l'évaluation du PIB et du RNB du Luxembourg"; Statec, Luxembourg, 30 September LV-2010: "Improvement of quality of National Accounts"; study by the Central Statistical Bureau of Latvia, March NL-2005: "The illegal economy in the Netherlands"; by Marret Smekens and Marleen Verbruggen; Statistics Netherlands, 8 June NL-2007: "Linking the Illegal Economy to National Accounts"; by Ron de Heij, Statistics Netherlands, 12 July