Exhaustiveness, part 1 - Main issues 1

Exhaustiveness, part 1 - Main issues 1 Introduction This paper continues the series dedicated to extending the contents of the Handbook Essential SNA: Building the Basics 2. One of the main themes in this series has been GDP compilation since GDP (Gross Domestic Product) is the most important single indicator measuring economic activity in a country. Earlier in this series we reviewed the three main approaches to GDP compilation the production approach, the expenditure approach and the income approach. Since these separate approaches all aim to measure the same GDP, an important goal for national accountants is to minimize the discrepancies between these different compilations. By comparing different compilations of the same GDP one has a powerful method to detect errors and omissions in any individual approach. As we have seen the supply and use table (SUT) integrates the three approaches so that a balanced SUT guarantees that the three approaches all yield the same GDP number. But although less likely, it is still possible that errors and omissions affect the different approaches in such a way that they will remain after balancing. How can we be sure that this does not happen? To answer this question we want to explore in this paper the closely related question (related to omissions rather than errors) how we can ensure that the compiled GDP is exhaustive. According to the general dictionary this term means testing all possibilities or considering all elements 3. Synonyms for exhaustive are: complete, comprehensive, thorough. In the first section of this paper we will see what exhaustive means in the context of GDP compilation. Next, we will give some examples of issues in GDP compilation that could cause the estimates to be non-exhaustive. In the remainder of the paper we will outline a framework that the national accountant can use to assess the exhaustiveness of the GDP estimates. What is meant by exhaustiveness? Rather than focusing on the exhaustiveness of GDP let us expand our scope to the whole of national accounts. A first meaning of exhaustiveness is related to the institutional units for which the accounts are prepared. Accounts may be regarded as exhaustive when all units, both registered and nonregistered (including individual persons), are included in the estimates, either by full enumeration or by sampling. This is difficult to achieve in practice and an important source of non-exhaustiveness is the (partial) omission of particular classes of units, such as individual entrepreneurs. Making the accounts exhaustive with respect to units means for example including estimates for output, intermediate consumption and value added of individual entrepreneurs. But there is more to the meaning of exhaustive. Also, all kinds of economic activities should be included in the compilation. Not all activities are included in the accounts, SNA specifies a production boundary, with some activities falling outside the boundary, such as some non-market household activities within the household sector. The fact that such activities are not included in the accounts 1 This paper was initiated and financed by EUROSTAT through the Project Essential SNA: Building the Basics, implemented by DevStat Servicios de Consultoría Estadística in consortium with ICON Institute, for which information can be found at the following link:http://circa.europa.eu/irc/dsis/snabuildingthebasics/info/data/website/index.html 2 Henceforth called the Handbook ; this paper is based on the second (2012) edition; it can be found at the following link: http://epp.eurostat.ec.europa.eu/portal/page/portal/product_details/publication?p_product_code=ks-ra-12-001 3 http://www.merriam-webster.com/dictionary/ 1

do not make the accounts less exhaustive. On the other hand not including activities in the accounts that fall within the SNA production boundary do make the accounts less exhaustive. An example is narcotics production and distribution, an illegal activity that is difficult to measure but falls within the production boundary. Making the accounts exhaustive in this respect means including in the accounts estimates for output, consumption and trade in narcotics. But there is even more to exhaustiveness than the comprehensive inclusion of units and activities. Even when values are reported for particular units on certain activities these reported values may not be the true values. It is well known that some economic units, often of small and medium size, under-report the value of their activities (e.g. for purposes of tax evasion). Is the difference between the true value of an activity and the reported one by a registered unit to be treated as measurement error, or as a value created in the hidden economy which should be included in the estimates of GDP in the same way as the value created by an unregistered unit which does not report anything? Our exhaustiveness framework which we will discuss below will allow for the latter possibility. Thus, National accounts can be said to be exhaustive if all economic units are included, all kind of economic activities falling within the SNA production boundary are included and full values of such activities are taken into account. Looking at exhaustiveness from another point of view one could also say that national accounts are non-exhaustive if certain accounts are omitted from compilation, such as financial accounts and balance sheets, or if quarterly or regional estimates of GDP are missing, or if input-output tables and social accounting matrices are not compiled. However, this is usually not meant when national accountants speak of exhaustiveness. Some examples of non-exhaustiveness in GDP compilation What are some of the typical elements which make national accounts non-exhaustive? To find out we can benefit from the comparative studies done within the framework if the International Comparison Programme (ICP) 4. The main objective of the ICP is to compare the real GDP (and per capita real GDP) of participating countries. The starting point is to have reliable and consistent estimates of the level of GDP in the national currency of each participating country. Comparability of the national accounts estimates is of crucial importance to the ICP, so the estimates of GDP must be exhaustive in the sense of the previous section. The ICP experience has shown that the following areas are often under estimated in (or even omitted from) countries accounts: Consumption of crops and livestock products that households have produced for themselves; Food, clothing and household goods sold by street traders or in village markets; Goods and services that are provided to households by informal or unregistered producers, such as food and drinks sold on the street by itinerant vendors and services of unregistered taxi drivers; Plumbing, electrical, repairing vehicles, and other household maintenance services; 4 See the website of IPC: http://siteresources.worldbank.org/icpext/resources/icp_2011.html; in this section and the next we have used the following documents found at this site: Ensuring the Consistency of National Accounts for the ICP, 1st ICP National Accounts Working Group Meeting February 16-17, 2011 National Accounts Framework in the ICP - Operational Material 2

Private teaching lessons; Personal services such as hair-cutting and shoe-cleaning; Goods that have been smuggled into the country from abroad without payment of customs duties and without being recorded in the international trade statistics; Government expenditures, including expenditures on the military forces, expenditures by municipal and local authorities, and expenditures by the head of state; Dwellings that people build for themselves; Illegal activities that contribute significantly to production, such as prostitution and drugs. Non-exhaustiveness because of defective imputations Some goods and services are acquired without any payment involved. For national accounting purposes it is necessary to impute values to all these types of transactions to ensure that GDP measures the value of all the production in the economy. Defective or missing imputations can be important reasons for non-exhaustiveness. Some of the main imputations the national accountants make are the following: Consumption of own-produced goods; Income in kind; Rents of owner-occupiers; Financial intermediation services indirectly measured (FISIM); Barter transactions; Consumption of fixed capital for government; Net expenditure by residents abroad. We will look at each of these imputations separately. Consumption of own-produced goods (SNA 9.53 9.55) The production of goods and services for own account can be split into those that need to be included (via imputed values) in the national accounts and those that are excluded. The production of household services that are consumed within the household is specifically excluded. Examples are services such as cooking food (although the cost of the food items themselves must be included), washing etc. Housing services provided to the owner-occupiers of a dwelling are included via an imputed value, as are services produced by paid domestic staff. Household final consumption expenditure also includes the values of goods and services produced by unincorporated enterprises owned by households and which are consumed by members of the household that owns the unincorporated enterprise (e.g. food produced for own consumption by farmers). Income in kind (SNA 7.48 7.54) Employers may remunerate their employees in kind for tax advantages or for disposing of outputs that are periodically in excess supply. Employees receive these goods and services free or at very low prices as part of their compensation. Examples are free train travel for railway workers, holiday facilities for employees and their families, and free meals for factory workers. In the national accounts, goods and services provided as income-in-kind are recorded as part of compensation of employees and the same amount is included in individual consumption expenditure by households. 3

Rents of owner-occupiers (SNA 9.65) Owners of houses provide housing services to themselves, similar to the services consumed when renting similar property. When dwellings are rented, rentals are recorded as output of housing services by owners and final consumption expenditure by tenants. When dwellings are occupied by their owners, imputed value of the housing services are recorded as both output and final consumption expenditure of the owners. Purchases of materials used for repairs are also included in final consumption expenditures. However, expenditures on major improvements (reconstructions or enlargements) to dwellings are excluded from household consumption expenditure and are recorded as gross fixed capital formation. The output of dwelling services of owner-occupiers at current prices is in many countries estimated by linking the actual rents paid by those renting similar properties in the rented sector to those of owner-occupiers. Similarity in the case of dwellings is usually judged by considering the type of dwelling (single-family or multi-family), location (urban or rural), and facilities (floor-space, running water, indoor toilet, electricity, central heating, etc.). This allows the imputation of a notional rent for the service that owner-occupiers receive from their property. This calculation is usually only carried out for a benchmark year, as few countries have the necessary data to apply the method on an annual basis. The benchmark is then extrapolated using indicators that reflect the change over time in the price of rent and the volume of dwellings 5. Financial intermediation services indirectly measured (FISIM) (SNA 6.163 to 6.165; Annex 3 - C2, A3.24 A3.27) Valuing the output of the financial sector is particularly difficult because some financial intermediaries provide services for which they do not charge explicitly. Instead, they generate income by charging different rates of interest to borrowers and lenders. They pay lower rates of interest than would otherwise be the case to those who lend them money and charge higher rates of interest to those who borrow from them. In this situation, the SNA must use an indirect measure (financial intermediation services indirectly measured, or FISIM) of the value of the services for which the intermediaries do not charge explicitly. FISIM is measured by comparing the actual interest rates on loans (rl) and deposits (rd) with a reference rate (rr) and multiplying these interest differentials with the stocks of loans (L) and deposits (D): FISIM = : (rl - rr) L + (rr - rd) D All loans and deposits are included, not just those made from intermediated funds. The reference rate should contain no service element and reflect the risk and maturity structure of deposits and loans. The rate prevailing for inter-bank borrowing and lending may be a suitable choice as a reference rate. SNA recommends that the consumption of FISIM should be allocated between users (lenders as well as borrowers) treating the allocated amounts either as intermediate consumption by enterprises or as final consumption or exports. 5 In case such an imputation is not possible the services of owner-occupiers are often estimated by the cost approach, as being equal to the sum of its production costs, including intermediate consumption, compensation of employees, operating surplus, consumption of fixed capital (reflecting the decline in the value of the fixed assets) and taxes on production (if applicable). Value added is then simply the total sum excluding intermediate consumption. 4

Barter transactions (SNA 9.49 9.50) Barter is the exchange of goods or services for other goods and services without money changing hands. Final consumption expenditure by households should include the value of barter transactions, which should be valued at the purchasers prices of the goods or services exchanged. Barter transactions involves two parties only and neither taxes on products nor transportation costs apply. Consumption of fixed capital for government (SNA 10.155 10.157) SNA requires that consumption of fixed capital be estimated for all fixed assets including railways, roads, bridges, tunnels, airports, harbours, pipelines, dams, etc. and for every year the asset is in use in production. According to the ICP experience the national accounts of some countries contain such estimates while those of others do not. Not including an estimate is equivalent to making an assumption that government consumption of fixed capital is zero, which may not be true. According to ICP the accounts of a number of different countries shows that it ranges from about 5% of government final consumption expenditure in developing countries to around 10% in some OECD countries. When direct estimation of consumption of fixed capital for government is not possible, it is recommended to include an estimate of the order or 5-6% of government final consumption expenditure. Net expenditure by residents abroad (SNA 9.79 9.80) Resident households make expenditures while travelling abroad, and non-resident households may make expenditures inside the economic territory of a country. Household final consumption expenditure should include the former, but not the latter. In order to calculate household final consumption expenditure it usually is convenient to calculate the total expenditure made by all households, whether resident or not, within the economic territory and to adjust this figure by adding expenditures by residents abroad and subtracting expenditures by non-residents within the economy territory. Expenditures by residents abroad constitute imports, while expenditures by nonresidents are exports. Non-exhaustiveness due to data deficiencies Another reason for non-exhaustiveness may be due to deficiencies of the regular data sources 6. We will follow the Handbook here (p. 119), which lists the following possibilities: Under-coverage of enterprises: enterprises, or parts of them, are excluded from the data collection programme, though in principle they should have been included. This can happen for a variety of reasons: o an enterprise is new and has not yet been included in the survey frames; o an enterprise falls below the size cut-off for surveys; o an enterprise has been incorrectly classified by type of activity or by region and thus improperly excluded from the survey frame; o an enterprise has not been entered in the statistical register, due to shortcomings of the register. 6 As described in Chapter IV of the Handbook: Statistical infrastructure for national accounts 5

Non-response by enterprises: enterprises are included in the sample but no data are collected from them and no imputations are made for the missing observations, because: o the survey questionnaire was wrongly addressed; o the enterprise, or part of it, did not return the questionnaire. Under-reporting by enterprises: data is obtained from enterprises, but is misreported by the respondent, or correct data is received but is inappropriately entered into the IT system. Non-observed Economy A next reason for non-exhaustiveness may be that units are not observed as part of the regular data collection programme. The totality of all such missing units make up the so called non-observed economy (NOE). This refers to all productive activities that may not be captured in the basic data sources used for compiling national accounts. The OECD Handbook Measuring the Non-Observed Economy is the authoritative reference work for the NOE and its measurement. The Handbook gives the following list of NOE activities (p.120): Underground activities: these are activities that are productive and legal but are deliberately concealed from public authorities to avoid: o payment of income, value added or other taxes; o payment of social security contributions; o having to meet certain legal standards such as minimum wages, maximum hours, safety or health standards, etc.; o complying with certain administrative procedures, such as completing statistical questionnaires or other administrative forms. Illegal activities: these are activities falling within the SNA production boundary that: o generate goods and services forbidden by law (e.g. production and distribution of illegal drugs); o are unlawful when carried out by unauthorized producers (e.g. unlicensed practice of medicine). Production of households for own final use: these are activities that result in goods or services consumed or capitalized by the households that produced them, such as: o production of crops and livestock; o production of other goods for their own end use; o construction of houses and other own-account fixed capital formation; o imputed rents of owner-occupiers, and services of paid domestic servants. Non-observed informal activities: these activities are part of the informal sector also covering observed activities undertaken informally; in general, informal activities are those productive activities conducted by unincorporated enterprises in the household sector that are unregistered and/or that have less employees than a particular threshold size, and that have some market production. Eurostat Tabular Approach to Exhaustiveness In the late 1990s Eurostat developed a systematic method (the Eurostat tabular approach to exhaustiveness) that was designed to identify potential sources of non-exhaustiveness in the national accounts. The tabular approach was developed in the context of several projects with European Union Candidate Countries and tested by nearly all these countries. Over recent years, there has 6

been a growing interest by different countries (including non-eu) in the application of the tabular approach. The Eurostat tabular approach was also recommended by the 2011 International Comparison Programme (ICP) to ensure exhaustiveness in the estimation of the gross domestic product (GDP). The tabular approach lists seven types of non-exhaustiveness, labeled N1 to N7. The distinction between the seven different N-types is not important in the sense that some things could potentially be classified under one heading rather than under another. Their importance lies in the systematic approach that ensures that all potential sources of omission from the accounts are identified and included in one of the categories and that there is no duplication across categories. We will follow the Handbook, which presents the seven N-types in four groups (Not registered, not surveyed, misreporting, other) as follows (Box VI.1, p. 121): 1) Not registered N1 - Producer deliberately not registering - underground The producer deliberately does not register to avoid tax and social security obligations. Most often this refers to small producers with turnovers that exceed threshold levels above which they should register. Producers that do not register because they are engaged in illegal activities fall under type N2. Type N1 does not include all underground activities, some are associated with type N6. N2 - Producers deliberately not registering - illegal The producer deliberately does not register as a legal entity or as an entrepreneur because it is involved in illegal activities. Type N2 excludes illegal activities by registered legal entities or entrepreneurs that report (or misreport) their activities under legal activity codes. N3 - Producers not required to register The producer is not required to register because it has no market output. Typically these are nonmarket household producers that engage in production of goods for own consumption, for own fixed capital formation, and for construction of and repairs to dwellings. Or, the producer has some market output but it is below the level at which the producer is obliged to register as an entrepreneur. 2) Not surveyed N4 - Legal persons not surveyed Legal persons not surveyed due to several reasons such as: the business register is out of date or updating procedures are inadequate; the classification data (activity, size or geographic codes) are incorrect; the legal person is excluded from the survey frame because its size is below a certain threshold etc. This leads to (systematic) exclusion of the legal person from surveys when in principle they should be included. 7

N5 - Registered entrepreneurs not surveyed Registered entrepreneurs that may not be surveyed for a variety of reasons: the statistical office does not conduct a survey of registered entrepreneurs; the registered entrepreneur is not in the list of registered entrepreneurs available to the statistical office, or if available, is systematically excluded from it; the registered entrepreneur is not in the survey frame because the classification data (activity code, size code, geographic code) are incorrect. 3) Misreporting N6 - Producers deliberately misreporting Gross output is under-reported and/or intermediate consumption is overstated, in order to evade income tax, value added tax (VAT), other taxes, or social security contributions. Misreporting often involves maintenance of two sets of books, payments of envelope salaries which are recorded as intermediate consumption; payments in cash without receipts, and VAT fraud. 4) Other N7 - Other statistical deficiencies Type N7 is subdivided into: N7a: data that are incomplete, not collected or not directly collectable; N7b: data that are incorrectly handled, processed or compiled by statisticians. The following areas should be investigated: handling of nonresponse; production for own final use by market producers; tips; wages and salaries in kind; and secondary activities. The usual starting point in identifying the seven types of non-exhaustiveness is the production approach, due to the more uniform data sources that generally underlie these estimates. The adjustments are usually tabulated against activities, e.g. ISIC sections. In theory, exactly the same data sources as for the production approach can be used for the income approach, making the income-based and the production-based estimates of GDP completely consistent, so the adjustments have to be linked. However, if separate surveys are used to collect data for these two approaches (e.g. an independent survey of profits being used to estimate gross operating surplus on the income side) then it should be possible to confront the data from the different sources to check their consistency. The expenditure approach is based on data that generally come from different sources than those underlying the production and income data sets. However, the conceptual consistency between the three different approaches to measuring GDP provides a framework for checking the accuracy of data from different sources. For example, domestic production plus imports (less exports) of a particular type of equipment should be equal to gross capital expenditure on that type of equipment (excluding any adjustments for second-hand purchases and sales). 8

Concluding remarks This paper set out to provide some background to the role of exhaustiveness in GDP compilation. We have highlighted some of the meanings of the term exhaustiveness, as used by national accountants. We saw that non-exhaustiveness may be caused by defective imputations, by missing units, by missing activities or by under-reported values. Finally, we introduced the Eurostat tabular approach to exhaustiveness which was designed to identify potential sources of non-exhaustiveness in the national accounts. In the next papers we will focus on some particular areas of non-exhaustiveness, such as illegal activities and informal sector non-observed activities.. To find out more, Essential SNA: Building the Basics (2012 edition), Chapter 6 The 2008 SNA, European Commission, IMF, OECD, UN, World Bank, 2009, Chapter 25 Measuring the Non-Observed Economy, OECD 2002 9