STANDARD METHODOLOGIES FOR CONDUCTING HOUSEHOLD BUDGET SURVEYS LABOUR FORCE SURVEYS

STANDARD METHODOLOGIES FOR CONDUCTING HOUSEHOLD BUDGET SURVEYS, LABOUR FORCE SURVEYS AND CALCULATING CPI-THE COMPARISON WITH THE METHODOLOGIES USED IN BOSNIA AND HERZEGOVINA HOUSEHOLD BUDGET SURVEY (HBS) 1. INTRODUCTION 1.1. Historical notes and overview The Household Budget Surveys (HBS) in the European Union are sample surveys of private households carried out more or less on the regular basis and under the responsibility of the National Statistical Offices (NSIs). They provide information about household consumption expenditure on goods and services, information on income, possession of consumer durable goods and cars, basic information on housing and many demographic and socioeconomic characteristics. Contrarily to other European statistical domains, the transmission of HBS data to EUROSTAT was voluntary and no EU regulation existed. HBS are national surveys, and there is therefore a great degree of freedom for each Member State to decide on the objectives, methodology, programming and resource assignment for their respective HBS. The initiative for compiling the existing information in the Member States on household budget surveys appeared in order to make all this information available at European level as well as to improve harmonization of surveys, in terms of concepts used, classification of variables, data collection and data processing methods. It helped each country to keep the targets, the uses and the programming of its national HBS and, at the same time, to collaborate with EUROSTAT in order to compile a European-wide data set on household budgets. One of the sources of the problems was the wide variety of uses and users. Traditionally, their main use has been to collect information on household consumption expenditure for updating the ‘weights’ for the basket of products used in the Consumer Price Indices. Following that major use, many other uses have arisen either at national or European level: to estimate the household consumption accounts for National Accounts purposes, to carry out a wide variety of analyses on consumers and consumption, to provide complementary information for studies on poverty diagnostics and social exclusion, to conduct research on economic and consumption issues, etc. 1 The key concept of the data collected by the HBS was “household final consumption expenditure”. For this purpose the COICOP-HBS classification was used to disaggregate these data. Together with these data, the HBS collect numerous cross-sectional variables regarding households and household members. These variables allowed the HBS results to be used in many different ways. In the very beginning of this process, the methodologies used by the Member States to carry out the HBS were very far from being harmonized. Since then, all the countries participating in this project and EUROSTAT have made great efforts to harmonize their HBS and to improve data comparability. However, there is still some room for improvement. In order to allow EUROSTAT to process the data received so as to perform ex-postharmonisation and to answer specific requests of the users, countries delivered micro-data to EUROSTAT. It allowed the EUROSTAT disclosing aggregated tables or indicators in the initial phase of the harmonization process. In the further phases, this process was became more ambitious and was improved by the common efforts of the EUROSTAT and Member and Candidate States. 1.2. Objectives and scope As mentioned in the previous section, there were fairly complete data sets collected from the HBS of the present Member States and some EFTA countries. However, the available data from some countries were very scarce. In order to fill this gap, EUROSTAT launched a project in mid 2002 to collect some aggregate HBS data from the Candidate Countries for the reference year 1999. The aim was to have a picture of the “household final consumption expenditure” in the Candidate Countries. All the Candidate Countries agreed to participate in this project and all of them sent two types of information: • a set of tables with aggregate data following a common format proposed by EUROSTAT. • a document giving methodological information about how these data were collected. This document put together all the methodological information sent by the countries and analyzed the most significant differences from the methodology proposed by EUROSTAT. Because this document was not an exhaustive methodological guide the new publication was created: “Household Budget Surveys in the EU: Methodology and Recommendations for Harmonization. 2003”. This document focused on the identification of the methodological differences among the CC and the methodological recommendations of EUROSTAT. The conceptual scope of this document was the description of the main methodological features regarding the Household Budget Surveys in the Candidate Countries. It helped also other 2 countries to see how far from the European standards regarding the household budget surveys they are, and to see the evolution of this process among the developed and developing countries in the Europe. In the following paragraphs the HBS methodological issues will be analyzed and the recent situation in Bosnia and Herzegovina regarding the HBS will be compared. Also, some recommendation will be proposed in order to improve the survey and its harmonization with the EU Regulations. 1.3. National aim and users of the HBS The calculation of weights for consumer price indices is the main use of the HBS results in most EU countries. Besides this, there are a number of other uses with variable importance depending on the country: estimation of household national accounts, data for social policies, poverty measurement, etc. The HBS is a very important source for the CPI in most EU countries; it is an important source of information for the NA in about half of the EU countries. The main users of the HBS data at national level in the EU are mostly the government and statistical offices (NA and price statistics), universities and private companies. B&H current situation: The main aim of the HBS was to calculate weights for the CPI. Actually, the HBS 2004 introduced the new CPI methodology into B&H statistical system and new weights were calculated in order to replace the old ones produced by the HBS 1989 which were totally out of date. This project made possible to replace the existing Retail Price Index with the Consumer Price Index harmonized with EU regulations. The second aim of the HBS was to give the data for the GDP estimationexpenditure approach. The HBS 2004 was used for the poverty analysis in order to continue with providing the data for poverty reduction strategies established few years ago with LSMS (2001) and “Living in B&H” (2002-2004) surveys. Recommendations: Because the CPI basket was much bigger then the list of the products and services covered by HBS, there were some problems in weights estimation which should be minimized in the further HBS. The solution is in the better design of the HBS and in the introduction of new surveys in domain of the trade statistics, which could improve weights estimation on the level of groups of product and services. The use of the HBS for National Account purposes should be improved in order to provide more data for expenditure approach in the GDP estimation. The additional modules and variables should be introduced in the HBS questionnaires. 3 Probably, there is no more possibility to conduct the LSMS and the HBS is becoming one of the basic statistical surveys. It should be the basis for the poverty analysis and diagnostics and for this purposes the HBS must be improved in order to create the opportunity for the similar country poverty profile as was created on the LSMS basis. The HBS is scheduled for 2007 and above mentioned recommendations were introduced in its methodology (the better correspondence of the HBS and CPI list of items was made, the design of the variables important for NA was improved, the Investment Module was created and the Income Module was expanded and improved in order to get detailed data on social transfers. 1.4. Timing and frequency The HBS are annual or continuous in the majority of the EU countries. Only Cyprus and Malta have five-yearly surveys. Historically, the most of these countries began to carry out their HBS in the middle of the last century. Generally, these surveys are therefore well established and have a long tradition. B&H current situation: In the pre-war situation the HBS was conducted on the regular basis within the statistical system of former Yugoslavia. After the war, the HBS was totally sponsored by donors and there is still no possibility to be founded by the B&H government. This was the reason that the HBS was conducted ad hoc depending of the available resources and technical assistances. Recommendations: The idea is to have the household budget survey annually using the calendar year approach. This could improve the HBS methodology and to introduce the continuity and tradition. 2. SAMPLE DESIGN This chapter describes the main sampling characteristics of Household Budget Surveys in the EU countries. From the perspective of comparability, it is by no means necessary for the national sample designs to be similar, or even comparable. The sampling designs were determined by specific national circumstances, constraints and survey objectives. Any sample should, nevertheless, meet the basic requirements of representative or probability sampling both in design and in implementation, and should be of a sufficient size to permit the type of tabulation and analysis envisaged for the survey. These requirements were met to varying degrees in the national Household Budget Surveys. This and the following sections discuss various features of sample design and implementation from a comparative angle, covering aspects such as survey coverage, sample size and allocation and sampling stages. 2.1. Unit of measurement and coverage 4 All Household Budget Surveys were restricted to the population residing in private households. Collective or institutional households (old persons' homes, hospitals, hostels, boarding houses, prisons, military barracks, etc.) were excluded, as were generally persons without a fixed place of residence. In most cases the population excluded in this way was no more than 2% of the total population, though the effect was more significant for particular groups such as old persons, and certainly the homeless. As to geographical coverage, most Household Budget Surveys covered the entire population residing in private households in the national territory. In the EU countries, there were a few small differences in the exact definition of coverage: some countries included foreign households living in the country while others excluded them; certain countries excluded households whose head was a student or has been unemployed for a long time, etc. B&H current situation: The unit of measurement in the HBS 2004 were private households in the national territory. The collective, institutional and foreign households were excluded from the survey. The same approach was used for the HBS 2007. and will be kept in the future. Recommendations: None. 2.2. Sampling frame There are three main possibilities for obtaining the HBS sample: 1. Population registers. Where available, population registers could provide upto-date lists of households or individuals, with many relevant characteristics useful for stratification and efficient selection of the sample. 2. Use of an area frame. Another common arrangement is to obtain a sample of area units from a suitable source such as the population census (Greece, Spain and Ireland), or a master sample of areas (Portugal), and then to prepare or update lists of households or dwellings in the selected areas for the final sample. 3. Electoral register. Although less suitable for HBS than population registers, poll registers may also be used for building HBS sampling frames. In the first arrangement, the frame is in the form of lists of the ultimate units (dwellings, households, individuals) from which the sample for the HBS can be drawn directly. In practice, the sample selection may, of course, involve multiple stages and/or phases. In the second arrangement, the frame is used to draw a sample of area units. The areas may be drawn from the whole frame, as for example from the population census, or from a 'master sample' specially constructed for the purpose. In the areas selected, lists of addresses, households or persons may be prepared or updated from other sources to complete the process of sample selection. In most cases lists exist which can be used after 5 appropriate updating. Supplements are often added to the main frame to improve coverage. The sampling frames in the majority of the EU countries were taken either from the last census or from population registers. Only Malta created the sample from the electoral database. B&H current situation: The HBS 2004 sample was taken from the 2003 Master sample created as an area frame containing 710 enumeration areas. They were totally enumerated using the semi-intrusive method and the list of 40000 dwelling units was created. From this list the HBS sample was selected. For the purposes of the HBS 2007 and further household based surveys in Bosnia and Herzegovina, the new sample frame was made at the beginning of 2006. It consisted of 1500 Enumeration Areas which were totally enumerated using the intrusive method and the list of 80000 households was created containing additional variables on socio-economic characteristics of the households and their heads. Recommendations: Having in mind the changes in the population characteristics in Bosnia and Herzegovina, it is strongly recommended to improve the sample frame at least every three years. This should solve problems till population census was conduct and its results were available. 2.3. Population and sample sizes In the EU countries sample sizes varied between 1 276 households in Slovenia and 36 072 households in Romania. Few countries had sample sizes of fewer than 5 000 households, while some of them had sample sizes of more than 25 000 households. B&H current situation: The HBS 2004 sample size was 7413. The size was decided on the basis of budget constraints and experiences of other countries rather than on the basis of the statistical principles and calculations. Recommendations: To keep the same sample size in the forthcoming HBS. For the HBS 2007 the final sample size will be 7600. 2.4. Sampling methods Probability sampling means a sampling scheme in which each and every unit in the study population is given a known, non-zero chance of being selected in the sample. Probability sampling requires that (i) all units in the population be represented in the frame used for selecting the sample; (ii) the selections are done by applying a proper randomized procedure which gives definite selection 6 probabilities to all eligible units; and (iii) the sample as selected is successfully enumerated in the survey. Although the most common method was stratified random probability selection (as recommended by EUROSTAT), there were also some non-random methods, such as quota sampling (Czech Republic and Slovakia) or combinations of random and systematic sampling (Cyprus and Estonia). Most countries designed their samples in multiple stages and stratified them according to certain criteria during the sampling procedure in order to obtain samples that are representative of the population. There were a number of advantages in using clustered, multi-stage sampling. By concentrating the units to be enumerated, it reduced travel costs and other costs of data collection. The work involved in sample selection could also be reduced. The major disadvantage was the loss of efficiency of the sample due to clustering. The complexity of design and analysis was also increased. The most common practice was to use a two-stage design. First, a stratified sample of suitable area units was selected, typically with probabilities proportional to size after stratification by geographical and other variables. The second stage consisted of the selection, within each sample area, of households or addresses for inclusion in the survey. Over-sampling of small population groups was carried out in Czech Republic, Hungary (pensioners and poor households), Slovakia (farmers and selfemployed) and Turkey (small population groups). Five countries (Bulgaria, Hungary, Latvia, Poland and Turkey) allowed household substitution in order to improve the response rates. B&H current situation: The HBS 2004 sample was two-stage systematic sample providing the implicit and explicit stratification. The Primary Sampling Units (PSUs) were the enumeration areas selected systematically from the Master Sample. After the enumeration of all dwelling units/households within the 710 PSUs, 9570 dwelling units/households as Secondary Sampling Units (SSUs) were selected in the second stage. The HBS sample was created to be representative on the state and entity level. The same design was applied for HBS 2007: in the first stage 1500 Enumeration Areas (PSUs) were selected and enumerated and 9274 households (SSUs) were selected in the second stage in order to obtain the final sample size of 7600 households (17,8% the nonresponse rate was considered). Recommendations: To improve the sampling method depending of the quality of the current sampling frames. The goal is to design the samples on the basis of the population census data in the future. 2.5. Response rates 7 High rates of non-response are a common and major problem in Household Budget Surveys. In a number of surveys the sample initially selected is substantially larger than the total number finally required. At worst, the sample may become essentially self-selected and hence quite unrepresentative of the population of private households. It is therefore important to keep track of the response rates achieved. Analyzing the experiences of few European countries from 1999 data, the response rates varied from 38% in Malta to 93.9% in Cyprus. B&H current situation: The response rate in HBS 2004 was 82.2% which could be considered as a satisfying result respecting the survey instruments and socioeconomic situation in the country. Recommendations: The response rate could be improved in at least three manners: a) adjusting the survey instruments which introduce fewer burdens for households (open diaries, shorter recording period etc.); b) providing better survey publicity by media; c) introducing presents for household in form of lottery etc. 3. SURVEY STRUCTURE AND CONTENT 3.1. Reference and recording periods In all Household Budget Surveys, data collection involves a combination of (a) one or more interviews, and (b) diaries maintained by households and/or individuals, generally on a daily basis. The period for which a diary is maintained is called the recording period and its duration and distribution over time is the primary determinant of the structure of the survey. On the other hand, the reference periods mean the periods of time for which the household consumption expenditure is calculated. In the European countries the reference periods varied from 14 days to one year depending of the nature of survey variables. The reference periods are usually one month or one year, rarely one quarter (Poland). B&H current situation: In HBS the recording periods were as follows: HBS 2004: -14 days for diaries food and beverages - 1 month for expenditure for housing, furniture, garments and footwear, health, etc. 8 - 3 months for durable goods, vehicles, regular maintenance of the house, communication equipment etc. - 12 months for income. HBS 2007: -14 days for food and beverages - 1 month for expenditure for housing, garments and footwear, health, etc. - 3 months for regular maintenance of the house, transport and communication, spare time etc. - 12 months for durable goods, furniture, vehicles, education, communication equipment, investment, income. The reference periods used in HBS 2004 were one month and one year. The same is scheduled for the HBS 2007. Recommendations: The recording periods should be closely connected not only to the recalling capability of the respondents, but also to the method of the survey data analysis in order to get reliable monthly and annually estimates. So, they depend inter alia on the analytical capabilities of the survey analysts. 3.2. Survey instruments The HBS in the EU used two main instruments: interviews and diaries. The way each country used these instruments is slightly different among countries. There are no strong rules in the design of the survey instruments. Each country designed the survey instruments according their consumption habits. B&H current situation: In both HBS 2004 and HBS 2007 were used the same survey instruments: a) diary of purchase b) self-consumption booklet c) final interview First two instruments were filled in by households and were designed as “closed” diaries. The last one was filled in by the interviewer in face-to-face interview. Recommendations: To discuss the possibility of the “open” diaries implementation in order to facilitate households keeping them during the recoding period. 4. MAIN CONCEPTS AND DEFINITIONS This chapter gives descriptions of the concepts and definitions actually used by the EU countries. This information is essential to assess the comparability of the 9 figures supplied by each country. Whenever possible, these descriptions are compared with the EUROSTATrecommendations, so that the changes that would be required in order to harmonize the HBS in other countries with the current Member States can be assessed. 4.1. Main concepts and definitions The basic unit of data collection and analysis in Household Budget Surveys is the household. How the household is defined is important for two reasons. Firstly, as a unit for selection of the sample, the definition as adopted influences the survey’s coverage of the population. The objective is to define the household in such a way that each person in the study population belongs to one and only one household. Definition of the household EUROSTAT recommends that the definition of the household for the purpose of HBS is based on the following two criteria: co-residence and sharing of expenditure. Bulgaria, Cyprus, Czech Republic, Malta and Slovakia follow EUROSTAT’s recommendation. All the other countries mention other conditions, such as sharing of income, existence of personal ties, etc. B&H current situation: The EUROSTAT recommendations were strictly followed. Recommendations: None Members of a household In practice these definitions need to be amplified to specify exactly what categories of persons are included and excluded from the definition. The definition adopted has a bearing, for instance, on whether or not resident employees, lodgers, boarders and other unrelated persons living at the sample address are included in the same or as a separate household. This affects the average household size and composition, as well as the coverage achieved in the survey. Furthermore, the definition is often extended to include certain categories of persons who are absent from the household for some specified reason, such as full-time education or military service. The EU countries differ in the exact rules applied for this purpose. The EUROSTAT recommended using for the HBS the definition of household membership proposed for the EU-SILC project, which covered all the categories mentioned above, provided that they fulfill certain conditions specified by the EU-SILC regulation. B&H current situation: In both HBS 2004 and HBS 2007 the household members were all individuals who satisfied two main criteria of the household definition; students regardless to the absence from the household; members who 10 were absent less then 12 months and who intent to return to the household; members with absence longer than 12 months but they were excluded from the survey and the estimation; members who worked abroad but returned regularly to the household and new born babies. Recommendations: To strictly follow the EU-SILC household members definition1. Head of household and reference person It is necessary to identify a particular individual in a household as its head (or reference person) whose personal characteristics can be used in the classification and analysis of the information on the household. The social group, occupation and employment status, income, sex and age, etc. of the head are often used to classify the sample households for presentation of the results and for weighting classifications used in the derivation of the survey estimates. In order to clarify the terminology, the head of household is the person designated in each original national survey and reference person is the person complying with the harmonized criteria recommended by EUROSTAT and designated for the European Household Budget Survey statistics (main income earner). The EUROSTAT recommended that the designation of the reference person should be based on objective criteria. For the tabulation of consumption patterns in the Household Budget Surveys, the appropriate criterion was the contribution to household income. Preferably, the person to be chosen should be the adult with the highest income in the household (main income earner). But a lot of countries do not fulfill this recommendation. B&H current situation: The households themselves delegated their heads. It was recommended to identify the head as a person contributing most to the total income of the household. Recommendations: To strictly define the head of the household as a person who contributes most to the household income in order to avoid the damage of the data comparability. Child-adult definition For various purposes it is also necessary to classify household members into adults and children. Different age limits or other criteria may be used in this classification for different purposes. One of them is the calculation of household size using the OECD scale of adult equivalents. Another one is the requirement 1 Household Budget Surveys in the EU, Methodology and Recommendations for harmonization2003, p. 18-19. 11 to identify persons who have to complete the individual diary in surveys where this type of diary is used. The definitions differed significantly among the EU countries. The age limit for distinguishing the children and adults varied from 12 to 20 and sometimes combined with school attending. Consequently, all the variables depending on child headcount cannot be harmonized. B&H current situation: The OECD scale was used. Recommendations: To record the age of each household member that allows applying any age limits for different purposes. 4.2. The concept of household consumption expenditure There are two relevant conceptual bases in ESA for household consumption expenditures: “household actual final consumption” (ESA 95, 3.81–84) and “household final consumption expenditure” (ESA 95, 3.75–3.77). Household actual final consumption consists of the acquisitions households obtain through their spending on consumption goods and services in their own country or abroad (“household final consumption expenditure”) and acquisitions from the government and non-profit institutions serving households, which are essentially provisions in kind to the households (“social transfers in kind”, such as some medical devices). Taking into consideration the practical difficulties for the measurement of the “household actual final consumption” in many Member States, EUROSTAT recommended that the “household final consumption expenditure” continue to be the conceptual base of the Household Budget surveys. The EUROSTAT recognized that the concept of “household actual final consumption expenditure” would be more suitable, since it was based on the “acquisition” idea. Member States’ practical difficulties with implementing it prevented this concept being used in the 1999 round of the HBS. In particular with health and education items, actual use could not even be observed at the individual household level: in almost all the EU countries expenditures are mainly public and are financed through taxes and/or premiums. In other fields such as transport, recreation and culture, which are also partly financed by the government, comparability was probably only marginally affected. B&H current situation: The main concepts of the “household final consumption expenditure” were applied. 12 Recommendations: The exact contents of the “household final consumption expenditure” should be analyzed in more depth. 4.3. Borderline cases Goods or services retained for own final consumption The goods and services retained by the household for own final consumption are part of final consumption expenditure. The EUROSTAT recommends to use purchasers’ prices and to record the consumption expenditure at the moment when the product is retained for own final consumption by the household. In this case the recommendation for HBS is therefore slightly different from the recommendation of ESA 95. Imputed rent In ESA 95, which is the reference for the HBS, the purchase of the dwelling as such was regarded primarily as capital formation (investment) and not consumption expenditure. However, the ownership of a dwelling was considered to produce a service – a shelter –, which was actually consumed over time by the households. As a consequence, ESA required the estimation of the price of the shelter, by imputation of a rent, since no monetary transaction was involved. This imputed rent is part of household consumption expenditure. So, for the HBS to be consistent with the ESA principles, it was recommended to exclude the acquisition of dwellings, whereas the consumption of the service of the dwelling should be included. B&H current situation: ESA recommendation were followed. Recommendations: None. Health and education expenditure, and other social benefits in kind The most important social benefits in kind are health and education goods and services provided as transfers in kind to individual households by government units. Consequently, only a relatively small part of health and education goods and services received by the individual households were actually paid for by them. This fact created an important problem of incomparability for the recording of these items because of the great differences between the social protection systems of each country. EUROSTAT’s recommendation on this point was not good from a theoretical point of view but took into account the huge practical difficulties of evaluating the actual consumption of this type of goods and services. EUROSTAT recommended that the concept of actual use of health and education services 13 will not be included in the conceptual base of household consumption expenditure of HBS; only the part of these services actually paid for by the individual households will be recorded as consumption expenditure. The same will be applicable for all the other social benefits in kind except housing. B&H current situation: EUROSTAT’s recommendation were followed. Recommendations: None. Wages and salaries in kind EUROSTAT recommended including the wages and salaries in kind supplied to households in the form of goods or services by employers for free or at a reduced rate, to the extent that they were believed to be quantitatively significant to the household. B&H current situation: EUROSTAT’s recommendation were followed. Recommendations: None. 5. CLASSIFICATIONS AND VARIABLES EUROSTAT recommended using the most recent versions of the standard nomenclatures and classifications wherever applicable. The following table presents the most important nomenclatures and classifications for HBS: Table 1 : Main classifications proposed by EUROSTAT CLASSIFICATION CONCEPT TO BE BROKEN DOWN BY THE PROPOSED CLASSIFICATIONS Consumption expenditure of households Territorial units and regions Codes for the representation of the names of the countries Education level Occupation Status in employment COICOP-HBS 2003 (i.e. including the modifications approved by the HBS Working Party of May 2003) NUTS-2003, level 1 ISO 3166 ISCED-1997 ISCO-1988 (COM ICSE-93 14 B&H current situation: The greatest attention was paid on the implementation of the COICOP-HBS. It was strictly followed in both HBS 2004 and HBS 2007. The use of other classification was marginal. Recommendations: To EUROSTAT proposals. 6. DATA TREATMENT 6.1. Grossing up and weighting The need to weight data from the Household Budget Survey is generally recognized. For some variables, double weighting is required: a) Spatial weighting aims to improve the representativeness of the sample in relation to the size, distribution and characteristics of the population under investigation. Methods for calculating coefficients may differ, and here we will look at the gradual approach recommended by EUROSTAT. b) The temporal weighting of data stems from the fact that the household observation period is often different from the reference period. Grossing up is a concept strongly linked to the weighting concept because it involves the estimation of values for the whole surveyed population. The practices of the grossing up and weighting differed among the countries2 and were depending on the current situation in each of them. B&H current situation: The data were grossed up into national monthly and yearly files. The method applied was too simple and introduced some doubt about the estimates obtained. The weighting was based on the calculation of inclusion probabilities and design weights and on the non-response analysis and calculation of the final weights for households. Recommendations: To improve the grossing up procedure in order to get more reliable monthly and yearly estimates. Sophisticated methods of the survey data analysis are needed. 6.2. Data processing There were different softwares used in the HBS data processing among the EU countries (Blaise, SPSS, SAS, SUDAN etc.) depending on the available resources. use the most recent classifications according 2 Household Budget Surveys in the EU, Methodology and Recommendations for harmonization2003, p. 40. 15 B&H current situation: Blaise software was used for data entry and SAS was used for data management and sample design. Data editing and imputation were made using SAS and CONCORD package. Data analysis was made using the SPSS. All softwares except the SPSS were available through the cooperation with Italian NSI-ISTAT. Recommendations: To provide the SAS or similar software in order not to be depended on donors. 6.3. Control procedures The quality of the results obtained from the Household Budget Surveys might be affected by various types of error. One such error that affects the overall quality of the survey is the observation (or collection) error (for example, be due to the omission of enumerated households, to the recording of expenditure incurred outside the reference period, or to false declarations). Other errors arised during input, encryption or data processing. Most countries compared the final aggregate results with other sources in order to check globally the coherence of the collected data. B&H current situation: The control procedures in HBS 2004 were scheduled as checks made by supervisors, data entry controls and data editing procedures. Results were compared with the final household consumption figures from NA. Recommendations: To improve controls made by supervisors because they were estimated not to be of the highest quality and did not start on time. The controls activities for HBS 2007 were better supported from the organizational and financial point of view and it is expected to yield more results. 6.4. Production time and dissemination The length of the period from the end of the survey until the first results were very different among the countries and varied from few month to two years. B&H current situation: The first results were available six month after the survey end and publication with final results was finished a year after. Recommendations: To have the final results six month after the survey end. The publication should be finalized nine months after the survey end. 7. METHODOLOGY CHANGES Some countries have reported plans for updating their HBS methodologies depending on the existing level of the survey adjustment to EUROSTAT recommendations. 16 B&H current situation: On the basis of experiences from the HBS 2004 the following changes were planned: a) b) c) d) improvement of the sample design based on the new Master sample introducing of new variables and modules for the NA purposes improvement of the Income module better design of the survey instruments in order to reach more correspondence between HBS and CPI basket e) better grossing up procedures The above-mentioned changes were processed for the HBS 2007. Recommendations: To continue with the adjustments according EUROSTAT recommendations in order to improve the final survey results. the LABOUR FORCE SURVEY Introduction The first attempt for the collection of the comparable data on employment and unemployment from all six Member States of the then European Community was made in 1960. Since that date, the number of Member States rose to twenty five, and the character of the European labour market has been transformed by the significant changes. Throughout this period, the institutions of the European Union included the issues of employment and unemployment among their highest priorities. The demand for accurate and comparable information on the labour market became progressively more urgent. In this context, the role of the EU Labour Force Survey has gained in importance, and was universally recognized as an indispensable tool for observing labour market developments and for taking the appropriate policy measures. The LFS is the only source of information in these areas to provide truly comparable data. The general methodology used in the LFS, together with a host of details concerning the definitions used and the practical implementation, was the subject to continual evolution. EUROSTAT, which is responsible for the dissemination of the results of the survey at European Union level, was conscious that accurate and up-to-date information on these aspects was indispensable to interpretation of the results. This information was therefore published on a regular basis under the title Labour Force Survey: 17 Labour force surveys in the European Union The purpose of labour force surveys In general a labour force survey is directed towards households, designed to obtain information on the labour market and related issues by means of interviews. Labour force surveys were usually confined to a sample of households, the actual size of which depends primarily on the level of detail required in the survey estimates. The principal advantages associated with labour force surveys are related to: (1) the opportunity of obtaining comprehensive information (at less cost than a census) across the entire economy, which could be assessed in a global setting embracing society as a whole; (2) the inherent flexibility of such surveys, which made it possible to define or conceptualize not only employment and unemployment, but also the circumstances surrounding other groups outside or on the margins of the labour force. This latter feature (i.e. the facility to conceptualize or define) assumed greater importance in recent years because of the manner in which labour markets and society generally evolved, and in view of the growing need to view labour market phenomena in an international context. The history of labour force surveys The notion of obtaining information on the work force by means of householdbased inquiries is not in any sense new. Questions on labour force were introduced in censuses of population in some countries during the latter half of the 19th century. The advancing trend of industrialization and the resultant restructuring of society created a need for new approaches, and for more sophistication in measuring labour market phenomena. The situation became particularly urgent with the advent of mass unemployment in the 1930's following the Great Depression. The first labour force survey was introduced in the United States in 1940 (on a monthly basis) with a new conceptual framework designed to provide information on relevant labour market characteristics. The movement towards the use of labour force surveys was somewhat slower in Europe. All Western European countries maintained comprehensive unemployment registers which, despite their disadvantages, provided a rudimentary basis for monitoring unemployment trends. The first European country to carry out a labour force survey was France in 1950, then Germany 1957 and Sweden in 19593. 3 The European Labour force survey, Methods and definitions 2001, p.5. 18 The development of the EU Labour Force Survey The first attempt to carry out a labour force survey covering the then European Community was made in 1960 with the six original Member States (Belgium, Germany, France, Italy, Luxembourg and the Netherlands). The Member States of the European Community agreed to apply the common recommendations in a new series of Community Labour Force Surveys which would be conducted annually. During the course of this series a substantial and coherent collection of labour market data was built up. The survey continued to be conducted annually, but for the first time a criterion of statistical reliability at regional level was introduced. The list of variables covered was revised and the ILO recommendations were applied. The continued commitment to the ILO recommendations ensured a high degree of comparability between the results obtained from the surveys. A number of Member States themselves felt the need for the improvements and looked into ways of reforming their surveys by possibly conducting them at more frequent intervals. It was likely that these national initiatives would not always be taken in the same direction or at the same time. The problem could only be solved by laying down a common reference framework and it was the source for the development of the European Regulation related to the LFS in order to get harmonized and comparable results. The organisation of the EU Labour Force Survey The earliest Community Labour Force Surveys were not official, but, from 1973 onwards, a Regulation was passed by the Council of Ministers governing the operation of the survey. Each Regulation applied only to a single year's survey, until the surveys of 1990 and 1991 were included in one Regulation (Council Regulation (EEC) No 3044/89) to remain in force until explicitly replaced by new legislation. In 1998 a new Regulation was adopted (Council Regulation (EEC) No 577/98). The national statistical institutes were responsible for selecting the sample, preparing the questionnaires, conducting the direct interviews among households, and forwarding the results to EUROSTAT in accordance with the common coding scheme. EUROSTAT devised the programme for analysing the results and was responsible for processing and disseminating the information sent by the national statistical institutes. Data collection and diffusion of results After becoming the data from each Member State, the EUROSTAT checked the data for errors according to its own programme of controls. When the data are 19 considered to be errorfree they were converted into a SAS dataset, which can be easily accessed to produce reports. Most Member States also produced regular publications setting out the results of their national surveys. The survey results were completely integrated into the EUROSTAT statistical system. Users with specific requirements which were not met by the existing publications and databases might also ask for customized tables to be produced for a fee. Technical features of the EU Labour Force Survey Field of the Survey The survey is intended to cover the whole of the resident population, i.e. all persons whose usual place of residence is in the territory of the country. For technical and methodological reasons, however, it was not possible in all countries to include the population living in collective households and, for the purpose of harmonizing the field of survey; results were compiled for the population of private households only. This comprised all persons living in the households surveyed during the reference week, and those persons absent from the household for short periods due to studies, holidays, illness, business trips, etc. It did not cover persons who, although having links with the household under survey: (a) usually live in another household; (b) live in collective households; (c) have emigrated. B&H current situation: The EUROSTAT recommendations were strictly followed. Recommendations: None. Reference period The labour force characteristics of each person interviewed referred to their situation in a particular week. B&H current situation: The reference week in LFS 2006 was the week from 3rd to 9th of April 2006. Recommendations: None. Units of measurement The main units of measurement for which results were obtained from the survey are individuals and households. The definition of a household varied somewhat from country to country but these differences were unlikely in the majority of cases to have a significant effect on the comparability of the results. 20 B&H current situation: The unit of measurement in the LFS 2006 were households and their individuals aged 15 and over. Recommendations: None. Reliability of the results As with any sample survey, the results of the Labour Force Survey were subject to sampling and non-sampling errors. Experience shown that at national level the survey information provided sufficiently accurate estimates for the levels and structures of the various aggregates into which the labour force was divided, provided that analyses of this type were confined to levels of a certain size. Survey results at regional level might, however, be affected by considerable sampling errors, even for relatively large groups of the population. Reliability of the results was assured by the size of the samples and the sampling methods used, in addition to careful and thorough planning of the various survey operations and rigorous administration of all phases of the survey. B&H current situation: The following table shows the coefficient of variation for three main LFS 2006 results at the B&H level: Table 2. Relative errors of main estimates in LFS 2006 Indicator Unemployment rate Employment rate Activity rate CV 2.51 1.70 1.17 Higher coefficient of variation were present in Brcko District because of the smaller sample size (although the District was over sampled) and in some specific subpopulation (for ex. people aged 65 and over). The non-response rate in LFS 2006 was as follows: Table 3. Non-response rates in LFS 2006 NonNonUrban/Rural response Responding responding Entity EA type Rate (%) HH HH Urban 12,75 2333 341 Federation of BiH Rural 4,16 3133 136 Total 2674 3269 21 Republic Srpska Brčko B&H Total Urban Rural Total Urban Rural Total Urban Rural Total 8,03 10,77 6,55 8,07 6,49 9,45 8,00 11,73 5,36 8,04 5466 1110 2068 3178 274 278 552 3717 5479 9196 477 134 145 279 19 29 48 494 310 804 5943 1244 2213 3457 293 307 600 4211 5789 10000 Recommendations: To keep the same overall sample size and to increase the subsample for Brcko District in order to get better estimates for the subpopulation of interest. Comparability of results between countries Perfect comparability among countries was difficult to achieve, even were it to be by means of a single direct survey, i.e. a survey carried out at the same time, using the same questionnaire and a single method of recording. Nevertheless, the degree of comparability of the EU Labour Force Survey results was considerably higher than that of any other existing set of statistics on employment or unemployment available different countries. B&H current situation: The LFS 2006 results are comparable with other countries owing to the high level of the harmonization to the EU methodology. Recommendations: To conduct LFS on the quarterly level in order to get results of higher quality and to reach the frequency of the survey as in developed countries. Comparability of results between successive surveys Since 1983 improved comparability between results of successive surveys was achieved, mainly due to the greater stability of content and the higher frequency of surveys. B&H current situation: The only comparison of the results were possible between the LFS 2006 and LFS-Pilot 2005 results. Doing that it is necessary to pay attention on the different timing of this survey and to respect the impact of the seasonality to the employment situation in the country. Recommendations: To conduct the LFS on the regular basis, i.e. quarterly based on smaller samples. This approach will provide more reliable results and and to show the effects of the seasonality. 22 Basic concepts and definitions4 The main statistical objectives of the Labour Force Survey was to divide the population of working age (15 years and above) into three mutually exclusive and exhaustive groups - persons in employment, unemployed persons and inactive persons - and to provide descriptive and explanatory data on each of these categories. Respondents were assigned to one of these groups on the basis of the most objective information possible obtained through the survey questionnaire, which principally related to their actual activity within a particular reference week. In order to get planned answers, most questions applied to selected groups only. A filter based on information already obtained specified who should answer a particular question. The definitions of employment and unemployment used in the Community Labour Force Survey closely followed those adopted by the 13th International Conference of Labour Statisticians. Employment A person is considered as having an employment if he or she did any work for pay or profit during the reference week "Work" means any work for pay or profit during the reference week, even for as little as one hour. Pay includes cash payments or "payment in kind" (payment in goods or services rather than money), whether payment was received in the week the work was done or not. Also counted as working is anyone who receives wages for on-the-job training which involves the production of goods or services (ESA 11.13 f). Self-employed persons with a business, farm or professional practice are also considered to be working if one of the following applies : (1) A person works in his own business, professional practice or farm for the purpose of earning a profit, even if the enterprise is failing to make a profit. (2) A person spends time on the operation of a business, professional practice or farm even if no sales were made, no professional services were rendered, or nothing was actually produced (for example, a farmer who engages in farm maintenance activities; an architect who spends time waiting for clients in his/her office; a fisherman who repairs his boat or nets for future operations; a person who attends a convention or seminar). (3) A person is in the process of setting up a business, farm or professional practice; this includes the buying or installing of equipment, and ordering of supplies in preparation for opening a new business. An unpaid family worker is said to be working if the work contributes directly to a business, farm or 4 The European labour force survey, Methods and definitions 2001, p.11-14. 23 professional practice owned or operated by a related member of the same household. B&H current situation: The definition of the employment was strictly implemented in both LFS 2006 and LFS-Pilot 2005. Recommendations: None. Self-employed persons If self-employed persons were absent from work, then they were regarded as in employment only if they could be said to have a business, farm or professional practice. The classification as employment of persons who works on their own small agriculture farm, who do not sell their products, but produce only for their own consumption depends on whether it falls within the production boundaries. When this production was included in national accounts, underlying employment must be identified. This depends on the relative quantitative importance of the production of agricultural products for own consumption in relation to the total supply of these products in a country (ESA 3.08) B&H current situation: The definition of the self-employment was strictly implemented in both LFS 2006 and LFS-Pilot 2005. Recommendations: None. Seasonal workers During the off-season, seasonal workers cannot be considered as having a formal attachment to their high season job— because they do not continue to receive a wage or salary from their employer although they may have an assurance of return to work. B&H current situation: The definition of the seasonal workers was correctly implemented in both LFS 2006 and LFS-Pilot 2005. Recommendations: None. Maternity leave Maternity leave is first given to the mother (but may include the leave of the father in the case of a transfer of the entitlements) and corresponds to the compulsory period of the leave stipulated by national legislation to ensure that 24 mothers before and after childbirth have sufficient rest, or for a period to be specified according to national circumstances. People in maternity leave should always be considered in employment. B&H current situation: The definition of the maternity leave was correctly implemented in both LFS 2006 and LFS-Pilot 2005. Recommendations: None. Unpaid family workers The unpaid family worker can be said to have a job but not be at work if there is a definite commitment by the employer (a related household member) to accept his/her return to work and the total absence does not exceed a period of 3 months. In this point EUROSTAT diverges from the ILO recommendation. B&H current situation: The definition of the unpaid family workers was correctly implemented in both LFS 2006 and LFS-Pilot 2005. Recommendations: None. Lay-offs A person on lay-off is one whose written or unwritten contract of employment, or activity, has been suspended by the employer for a specified or unspecified period at the end of which the person concerned has a recognized right or recognized expectation to recover employment with that employer. Lay–offs are classified as employed if they receive 50% of their wage or salary from their employer or have an assurance of return to work within a period of 3 months. B&H current situation: The lay-offs were not treated in LFS 2006 and LFS-Pilot 2005. Recommendations: To include lay-offs in the LFS. Long-term absence from work. If the total absence from work (measured from the last day of work to the day on which the paid worker will return) exceeded three months then a person was considered to have a job only if he/she continued to receive 50% of the wage or salary from their employer (ESA 11.14a). B&H current situation: The definition of the long-term absence from work was correctly implemented in both LFS 2006 and LFS-Pilot 2005. Recommendations: None. 25 Unemployment In accordance with the ILO standards adopted by the 13th and 14th International Conference of Labour Statisticians (ICLS), for the purposes of the Community labour force sample survey, unemployed persons comprised persons aged 15 to 74 who were: (a) without work during the reference week, i.e. neither had a job nor were at work (for one hour or more) in paid employment or self-employment; (b) currently available for work, i.e. were available for paid employment or self employment before the end of the two weeks following the reference week; (c) actively seeking work, i.e. had taken specific steps in the four week period ending with the reference week to seek paid employment or self-employment or who found a job to start later, i.e. within a period of at most three months. For the purposes of point 1(c), the following were considered as specific steps: — having been in contact with a public employment office to find work, whoever took the initiative (renewing registration for administrative reasons only is not an active step), — having been in contact with a private agency (temporary work agency, firm specialising in recruitment, etc.) to find work, — applying to employers directly, — asking among friends, relatives, unions, etc., to find work, — placing or answering job advertisements, — studying job advertisements, — taking a recruitment test or examination or being interviewed, — looking for land, premises or equipment, — applying for permits, licences or financial resources. Education and training are considered as ways of improving employability but not as methods of seeking work. Persons without work and in education or training will only be classified as unemployed if they are ‘currently available for work’ and ‘seeking work’, as defined in points 1(b) and (c). Lay-offs are classified as unemployed if they do not receive any significant wage or salary (significant is set at = 50%) from their employer and if they are ‘currently available for work’ and ‘seeking work’. Lay-offs are treated as a case of unpaid leave initiated by the employer — including leave paid out of government budget or by funds (16th ICLS). In this case, lay-offs are classified as employed if they have an agreed date of return to work and if this date falls within a period of three months. During the off-season, seasonal workers cannot be considered as having a formal attachment to their high-season job because they do not continue to receive a wage or salary from their employer although they may have an assurance of return to work. If they are not at work during the off-season, they 26 are classified as unemployed only if they are ‘currently available for work’ and ‘seeking work’, as defined in points 1(b) and (c). B&H current situation: The definition of the unemployment was strictly implemented in both LFS 2006 and LFS-Pilot 2005. Recommendations: None. CONSUMER PRICE INDEX 1. Introduction to the CPI Consumer Price Index (CPI) is economic indicator constructed to measure the changes over time in the prices of consumer goods and services acquired, used or paid for by households. CPI is used for a wide variety of purposes, including: as a guide for monetary policy; for the indexation of commercial contracts, wages, social protection benefits or financial instruments; as a tool for deflating the national accounts or calculating changes in national consumption or living standards. 2. Conceptual basis and coverage of the CPI The aim of the CPI is to cover the full range of final consumption expenditure for all types of households in order to give a timely and relevant picture of inflation. According to the European approach, the CPI is a Laspeyres-type pure price index and not a cost of living index (COLI). It reflects the price change between the current and the reference period eliminating the influences on price movements due to other factors. Taking into account the European System of Accounts (ESA 95), the coverage of the Consumer Price Index is the Household Final Monetary Consumption Expenditure (HFMCE). Some practical consequences of the use of ‘household final monetary consumption expenditure’ are: • The geographical and population coverage is of all purchases by households within the territory of a country, those by both resident and non-resident households (the so-called ‘domestic concept’). The CPI covers the prices paid for goods and services in monetary transactions. So for example some special fees and taxes paid to government for licenses will be excluded (when there is no equivalent good or service received in return). • 27 • • The prices measured are those actually faced by consumers, so for example they include sales taxes on products, such as Value Added Tax, and they reflect end-of-season sales prices. The CPI excludes interest and credit charges, regarding them as financing costs rather than consumption expenditure. 3. The Harmonized Index of Consumer Prices (HICP) The Harmonized Index of Consumer Price (HICP) is a set of EU Consumer Price Indices calculated according to a harmonized approach and a single set of definitions. The key HICPs are5: • The Monetary Union Index of Consumer Prices (MUICP) – aggregate indices covering the countries within the euro-zone. • The European Index of Consumer Prices (EICP) – for the euro-zone plus the other EU countries. • The national HICPs – for each of the EU Member States. Beyond these, there is also the European Economic Area Index of Consumer Prices (EEAICP), HICPs for the EEA countries and interim HICPs for the Acceding and Candidate Countries (1). The MUICP and EICP were calculated by EUROSTAT using statistics provided by the Member States on price changes and the consumption patterns of consumers within their economic territories. As explained further below, the aggregation across countries used country weights for ‘household final monetary consumption expenditure’. The HICPs are in principle open to revision, in particular when new or improved information becomes available. 4. Differences between the HICP and national CPI The differences between HICP and individual national CPI can sometimes be significant in practice. The differences have in general been diminishing, although national CPI uses their own national methodologies. In many countries national CPI was set up to serve different purposes, for example as ‘cost of living indices’ or ‘compensation indices’, and some of the underlying concepts and methods of national CPI are inappropriate for the HICP as a ‘pure’ inflation measure (of the impact of inflation on purchasing power). Some examples of differences between the HICP and national CPI are: • The treatment of subsidised healthcare and education. The HICP includes the net price paid by consumers (after reimbursements), while some national CPIs exclude these purchases or record the gross price. • The treatment of owner-occupied housing. In the HICP, the imputed prices for the services provided by owner-occupied housing are currently excluded. However, an index based on housing acquisition costs is being piloted for 5 Harmonized Indices of Consumer Prices (HICPs), European Commission, 2004., p.3. 28 possible inclusion in future. It will be compiled separately from the HICP on an experimental basis before any decision is made to incorporate it within the HICP. National CPI uses a variety of methods – for example some use an approach involving imputed rents, some include mortgage interest in their CPI, while others entirely exclude the shelter costs of owner-occupiers. • The aggregation formulae used at the most detailed level of stratification in the index calculations to produce the so-called elementary aggregates. The HICP uses ratios of arithmetic mean prices or of geometric means, while some national CPI uses other formulae. • The geographical and population coverage. The HICP covers all expenditures within the territory, whether by residents or visitors, while some national CPIs aim is to cover expenditures by domestic residents both within and outside the country. 5. Basic points on the calculation of CPI This section gives some brief information on how selected calculation issues are treated in the CPI. Collection of price data Price collection in the Member States is typically carried out by a combination of visits to local retailers and service providers and central collection (via mail, telephone, email and the internet). The consumer price survey is typically a sample survey. European Regulation 1749/96 establishes the Minimum standards for sampling in article 8: “HICPs constructed from target samples which, for each category of COICOP/HICP and taking into account the weight of the category, have sufficient elementary aggregates to represent the diversity of items within the category and sufficient prices within each elementary aggregate to take account of the variation of price movements in the population shall be deemed reliable and comparable.” It is clear that the Regulation requirement is very general and it has represented the reference for sampling in very country. EUROSTAT has set up two task forces that have produced a first attempt to identify the possible source of errors and a theoretical framework for sampling in the field of consumer price statistics. B&H current situation: In Bosnia Herzegovina, as in the majority of EU Member States, the sample selected for the consumer price survey is not a random sample but a purposive one. Generally speaking, it is possible to identify four stages in the selection of the sample for Bosnia Herzegovina: a. The selection of the geographical areas; b. The selection of the outlets; c. The selection of the products (basket of products); 29 d. The selection of the elementary items. The geographical areas chosen for the data collection in Bosnia Herzegovina are selected taking into account their weight in terms of population and their role with respect to the geographical areas they belong to (Canton or Region capital or main centre). The other relevant aspect at this stage of selection is the presence of a statistical Local Office that is able to carry out the collection of the elementary quotations in the field and check them before sending the data to the statistical Central Offices in Sarajevo and Banja Luka or to BHAS. On the basis of these criteria, the localities where the elementary prices are collected are 11 and they are represented by the main towns of the Federation of Bosnia Herzegovina (FB&H) that are also the capitals or the main towns of 5 cantons, by the main towns of Republika Srpska (RS), that are also the capitals or the main towns of 6 regions and by Brcko for the homonymous District (Table 1.1). Table 4 The geographical areas in B&H where consumer price survey was carried out and for which elementary prices were used to calculate CPI. Years 2005 – 2006 FB&H Bihac Tuzla Zenica Sarajevo Mostar RS Banja Luka Bijeljina Trebinje Doboj Prijedor East Sarajevo Brcko Brcko District The data collection is carried out in outlets, service providers, hospitals, physicians, dentists, cinemas, theatres, etc. The present design of consumer price survey is conceived as a non-probabilistic sample design. Therefore the sample of collection units is selected by the canton/regional offices according to a non-probability sampling procedure. The basket of products for Bosnian CPI in 2005 consisted of 642 and in 2006 of 646 products. They were selected in order to represent all the different typologies of products available in the market, to cover the entire spectrum of consumption that is summarised in COICOP classification and to take into account, as required by European Regulation 1749/96, the weights of the COICOP categories (the greater the weight of the category, the wider the spectrum of products belonging to that category). 30 The elementary items for which prices have to be collected monthly or bimonthly were identified on the basis of the mix of information regarding product, outlet, variety, brand and quantity collected. The collectors had to identify the elementary item according to the criterion of the more sold item, that means that in a specific outlet for a specific product, he has to select the variety, the brand and package more sold. In order to select the item he/she can ask information about the quantities sold for each product to the sales director in the case of big chains of distribution or to the single retailer in the case of traditional distribution. The identification of each single item has to be carried out at least once a year during the change of base and it has to be monitored monthly in order to maintain the representativeness of the items in terms of consumer behaviour. Monitoring the representativeness of each item selected means checking that it remains the more sold; if it loses this requisite, it has to be substituted. Recommendations: a. To improve the outlet selection. This could be done by: • the selection on the basis of “cut-off “ approach (using an auxiliary variable, such as the turnover, the units above the cut off value were included) and the selection on the basis of the weight of each typology in the local distribution. These weights should be provided from the HBS or specific trade survey • b. To improve the selection of the products in basket by the deeper revision of the basket c. To improve data collection by using methods of central data collection for items which prices are the same on the entity level (electricity, telecommunication services, etc.). 5. Quality adjustment The need for quality adjustment arises because the nature of the goods and services on the market changes over time. Quality adjustment is widely accepted by price index experts to be one of the most important and intractable problems in consumer price index construction. For the CPIs there are minimum standards for quality adjustment – explicit quality adjustments must be made whenever possible and the whole of a price change should never be ascribed to quality differences without justification. For quality adjustment European Regulations fix only some general criteria (Regulation 1749/96 in article 5). A task force has been set up by EUROSTAT in order to deal with the issue of quality adjustment. For the time being, the choice suggested has been the case by case approach, that is to say for different groups of products a ranking of the quality adjustment methods is done (from A methods, i.e. the reference ones, to C methods, that 31 should not be used). In general, methods for quality adjustment can be divided in explicit and implicit: European Regulation 1749/96 recommends the adoption of explicit estimates but also the implicit methods are acceptable with the exception of attributing the whole of the difference between the two prices (the price of the previous elementary item and the price of the replacing one) to the quality change (price change taken as quality change). B&H current situation: The replacement of an elementary item is due to a change in at least one of the four aspects that identify each elementary item: brand, variety, package or outlet. In B&H CPI the approach adopted is the overlap: the value of the quality change between the elementary item and the replacing one is assessed in terms of difference in price between the two items in a period when both items are available. Therefore, whatever is the case of substitution, the collector has to record the price of the replacing item in the previous month. The price is recorded in the dedicated cell of the questionnaire. On the basis of the availability of the prices respectively of the replaced and the replacing item for the previous month, the procedure will recalculate a new base for the micro index adopting the following proportion: PN : PO = BN : BO where PN = previous month price for the new item; PO = previous month price for the old item; BN = calculation base for the new item (unknown term); Bo = calculation base for the old item; Recommendations: None. 7. The basket of goods and services and weighting In reality the distribution of purchases of goods and services, and the precise nature of some of the goods and services themselves, varies from country to country – there is no uniform basket applying to all countries. The CPIs reflect this reality by being based on the prices and expenditures, which are representative in each country and not on an average ‘euro-basket’. The weights used for computing CPIs within a country may relate to a period up to seven years prior to the current year. However, to minimize any incomparability this might cause, adjustments must be made each year for any especially large changes in expenditure patterns. It is required that the CPIs should cover all newly significant goods and services. Special rules cover situations where prices are newly introduced for goods or services that were previously free, and other rules concern the situation when markets are opened to new suppliers – which may in practice deliver price benefits for consumers. 32 B&H current situation: The B&H CPI basket constants of 646 items grouped in 580 representative positions, 124 voices of products, 92 classes, 40 groups and 12 categories. Weights for 2005 CPI basket were estimated with data on household expenditure coming from the Household Budget Survey carried out in 2004. The procedure for the estimation has been driven through the following steps: a) HBS headings that are out of the coverage of CPI have been excluded. Therefore self consumption, expenditures for games for chance, expenditure for extraordinary maintenance of the house and imputed rentals for owner-occupied houses have not been considered; b) the HBS heading for which correspondence in the CPI basket does not exist were attributed to the HBS headings belonging to the same CPI voices of product in proportion to the weight of each HBS heading with respect to the voice of product or higher aggregate; c) the issue of splitting the HBS headings among the 642 products was carried out taking account of the lack of sources able to provide recent reliable information. Therefore, when possible and reasonable, the data used to split HBS data among the products have been: • • the weights used in the Federation in 2003; Italian data, where the assumption that the consumer behaviour are no so different between BiH and Italy was reasonable. • Recommendations: To deeply revise the basket in order to make it smaller and much representative. In the HBS 2007 the better correspondence between the HBS and CPI basket should be reached in order to facilitate weights estimation. 8. Computation and aggregation In order to produce comparable results, each country’s CPIs must be compiled using specified formulae (the ratio of either arithmetic or geometric means, but not the arithmetic mean of price relatives). In the EU as a whole, well over one million price observations are used to calculate the CPIs each month. B&H current situation: For 580 representative positions indices are calculated as geometric mean of micro indices (calculated for each elementary item in each outlet). Recommendations: None. 9. The HICPs – price stability and international comparisons 33 Consumer price indices have a variety of potential uses (for example, for indexing social benefits or contracts, and as inputs to other economic analyses), but the driver for this harmonization project has been their use as the main measure for monitoring price stability in the euro-zone. The HICPs have been set up to provide the best measure for international comparisons of household inflation within the euro-zone and the EU. B&H current situation: The HICP was calculated by taking into account price reductions. Recommendations: To improve the HICP calculation by strict following of the EU recommendations and definitions. 10. Future steps on harmonization The considerable progress made on the harmonization of CPIs does not mean that this work is now at an end. There are several major issues where further harmonization will still be necessary. The two major technical issues currently on the agenda are: • quality adjustment and sampling. EUROSTAT and the Member States are actively following-up an Action Plan concerning this subject. The aim is to agree some more concrete best practices for a range of specific goods and services (in particular for cars, consumer durables, books and CDs, clothing, computers and telecommunications services). The Regulation which addressed this issue in 1996 was only a first step – it is not in itself a sufficient guarantee of full comparability. • owner-occupied housing. The imputed prices for the consumption of the service provided by owner-occupied housing are currently excluded from the HICPs. Pilot calculations are being carried out using an approach based on the acquisition costs of housing which is new to the household sector (mainly this concerns newly-constructed dwellings). Indices will be compiled separately from the HICPs on an experimental basis before any decision is made to incorporate them within the HICPs. Beyond these, there are some other technical issues which may be of less fundamental importance for the HICPs but which will nevertheless bear on comparability, such as the treatment of seasonal items and the level of elementary aggregation. Some other essential tasks to be taken forward are: • The development of more comprehensive systems to assess Member States’ compliance with the existing Regulations and other guidance • The support of the national statistical offices to ensure that their HICPs are also comparable where this is not already the case. 34 • The consolidation of the now very extensive legal framework for HICPs, and the production in due course of a methodological manual to assist both compilers and users. 35 THE COMPARISON OF THE HBS RESULTS WITH THE RESULTS FROM SELECTED COUNTRIES By average monthly expenditure consumption of 667 € B&H with Croatia (762 €) is in between (Slovenia 1263 €) and new EU members as Bulgaria (167 €), Romania (186 €) and Estonia (172 €). Table 5: Structure of household consumption in selected countries in 2004 ( € ) Consumption expenditure items FOOD, NON-ALCOHOLIC BEVERAGES ALCOHOLIC BEVERAGES, TOBACCO CLOTHING, FOOTWEAR HOUSING, ELECTRICITY, GAS AND OTHER FUELS FURNISHING, HOUSEHOLD EQUIPMENT HEALTH TRANSPORT COMMUNICATION RECREATION, CULTURE EDUCATION CAFE, RESTAURANT, HOTEL MISCELLANEOUS GOODS AND SERVICES Total in € Austria 330,2 71,1 142,2 566,4 157,5 78,7 408,9 66,0 320,0 20,3 139,7 236,2 2537,5 Italy 407,2* 61,9 157,1 738,1 150,0 90,5 338,1 50,0 114,3 28,6 259,5 1988,1 Slovenia 239,7 35,3 102,2 151,4 84,5 21,4 224,6 58,0 135,0 12,6 63,1 135,0 1262,9 Estonia 50,6 6,1 10,8 27,6 10,0 6,1 19,7 10,5 12,0 2,9 6,2 9,5 172,0 Croatia 240,0 30,9 61,4 99,2 40,8 18,3 90,3 40,8 49,8 5,6 26,0 58,8 761,8 BiH 175,8 56,6 34,0 156,5 46,0 24,6 66,0 16,9 25,2 6,4 12,0 46,6 666,7 Serbia 141,9 17,6 23,0 66,9 17,6 16,1 40,9 10,3 15,7 7,7 5,7 19,1 382,5 Bulgaria 61,0 8,1 7,1 29,3 6,9 9,7 11,3 10,3 6,3 1,1 8,4 7,1 166,8 Romania 86,1 10,9 11,7 27,3 7,1 6,7 11,3 8,3 7,2 1,7 2,0 5,2 185,5 Source: Statistical agencies of selected countries *Food, non-alcoholic beverages, café, restaurant and hotel in one item 36 By structure of household consumption expenditure average Bosnian household spend about one third of its budget on food, beverages and tobacco which is characteristics of countries in transition unlike developed countries that spend roughly less than one fifth of its budget. Due to geographical position of B&H in mountainous region of Balkans Bosnian households spend about one forth of its budgets on housing and energy which is the highest value in comparison to other selected countries. Table 6: Structure of household consumption in selected countries in 2004 (%) Consumption expenditure items FOOD, NON-ALCOHOLIC BEVERAGES ALCOHOLIC BEVERAGES, TOBACCO CLOTHING, FOOTWEAR HOUSING, ELECTRICITY, GAS AND OTHER FUELS FURNISHING, HOUSEHOLD EQUIPMENT HEALTH TRANSPORT COMMUNICATION RECREATION, CULTURE EDUCATION CAFE, RESTAURANT, HOTEL MISCELLANEOUS GOODS AND SERVICES Total Austria 13,0 2,8 5,6 22,3 6,2 3,1 16,1 2,6 12,6 0,8 5,5 9,3 99,9 Italy 17,1* 2,6 6,6 31,0 6,3 3,8 14,2 2,1 4,8 1,2 10,9 100,6 Slovenia 19,0 2,8 8,1 12,0 6,7 1,7 17,8 4,6 10,7 1,0 5,0 10,7 100,1 Estonia 29,4 3,6 6,3 16,1 5,8 3,5 11,4 6,1 7,0 1,7 3,6 5,5 100,0 Croatia 31,5 4,1 8,1 13,0 5,4 2,4 11,9 5,4 6,5 0,7 3,4 7,7 100,0 BiH 26,4 8,5 5,1 23,5 6,9 3,7 9,9 2,5 3,8 1,0 1,8 7,0 100,1 Serbia 37,1 4,6 6,0 17,5 4,6 4,2 10,7 2,7 4,1 2,0 1,5 5,0 100,0 Bulgaria 36,6 4,9 4,3 17,6 4,2 5,8 6,8 6,2 3,8 0,6 5,0 4,3 99,9 Romania 46,4 5,9 6,3 14,7 3,8 3,6 6,1 4,5 3,9 0,9 1,1 2,8 100,0 Source: Statistical agencies of selected countries *Food, non-alcoholic beverages, café, restaurant and hotel in one item 37 THE COMPARISON OF THE LFS RESULTS WITH THE RESULTS FROM SELECTED COUNTRIES AND KOSOVO Labor force indicators in BiH are similar to indicators from Serbia, Montenegro, FYRM and Kosovo characterized by low Activity and Employment rates, high Unemployment rates and very high Long term unemployment. Table 7: Labour force indicators in selected countries Period Country BiH* Austria Italia Slovenia Estonia Croatia Bulgaria Romania Serbia* Montenegro* Kosovo FYRM* April 2006 I Q 2006 I Q 2006 I Q 2006 I Q 2006 IV Q 2005 I Q 2006 I Q 2006 IV Q 2005 IV Q 2005 I Q 2006 I Q 2006 Active population in thousands 1177 3995 24271 993 658 na 3223 9357 3453 257 na 878 Activity rates Total 43,1 72,2 62,7 70,9 72,1 63,2 61,5 62,3 53,5 49,9 46,2 54,3 Males 56,2 78,5 74,6 74,7 75,8 70,3 65,9 68,9 63,0 57,4 68,1 66,6 Females 30,8 66,0 50,9 66,9 68,8 56,3 57,1 55,8 44,6 42,9 25,3 42,0 Emplyment rates Total 29,7 68,2 57,9 65,9 67,4 55,2 55,5 57,2 42,3 34,8 27,9 34,6 Males 39,9 74,2 69,9 70,6 70,0 62,5 59,6 62,8 52,4 42,4 46,8 42,5 Females 20,0 62,3 45,8 61,1 64,9 48,2 51,5 51,7 32,9 27,6 9,9 26,8 Unemployment rates Total 31,1 5,6 7,7 7,0 6,6 12,6 9,8 8,1 20,8 30,3 39,7 36,2 Males 28,9 5,6 6,2 5,5 7,6 11,2 9,6 8,8 16,8 26,2 31,5 36,3 Females 34,9 5,6 10,0 8,8 5,7 14,4 9,9 7,3 26,2 35,5 60,7 36,2 Total 85,9 1,4 3,6 3,2 3,3 7,3 5,5 4,7 79,1 85,4 83,5 na Long term unemployed Male 85,5 1,5 2,8 2,3 3,7 6,4 5,3 5,1 78,4 80,5 83,7 na Female 86,6 1,2 4,8 4,2 3,0 8,5 5,8 4,2 79,6 90,0 84,0 na Source: EUROSTAT, Statistical agencies of respective countries * Calculation of Activity/Employment rates calculated by ILO recommendations that refer to population age 15+, EUROSTAT methodology refer to the population 15-64. 38 THE COMPARISON OF THE CPI WITH THE CPI FROM SELECTED COUNTRIES CPI in B&H in 2005 was changing in similar way as CPI in neighborhood countries. CPI in B&H has highest correlation is with CPI in Croatia (0.90). In January 2006 B&H introduces VAT and CPI registered prices change (Table 8). Table 8: CPI in selected countries ø2005=100 CPI in Selected countries 107 105 103 CPI 101 99 97 95 BiH Slovenia Croatia I. 98,9 97,9 98,3 II. III. IV. 99,3 99,6 99,9 V. 99,3 99,9 99,9 VI. 99,4 99,8 VII. 2005 99,4 100,1 98,5 99,8 99,6 99,7 99,4 100,1 99,5 99,6 99,1 99,7 99,9 101,4 101,7 102 VIII. IX. X. XI. XII. I. 2006 106 100,0 100,7 100,1 101,1 101,3 100,8 100,8 100,3 100,2 100,9 101,1 101,6 102,2 99,6 100,7 101,2 100,5 102,2 Macedonia 99,3 99,6 100,0 100,2 100,1 99,6 Surce: Statistical agencies of respective countries 39 LABOR FORCE SURVEY: Overall purpose and its use in terms of policy-design Purpose and objectives The overall purpose of the LFS is to divide the working-age population into three mutually exclusive classifications - employed, unemployed, and not in the labor force and to provide descriptive and explanatory data on each of these categories. The main objective of the LFS is to estimate the number and the characteristics of the labor force of the country and to elucidate the current state of employment and unemployment in country on a monthly/quarterly basis. Data applicability Data from the survey provide information on labor market such as employment across industrial sectors, hours worked, labor force participation and unemployment rates, the ratio of full time, temporary and part time workers, the profile of the unemployed and employed population by gender, age, educational attainment, occupation, income etc... Above all, the LFS data are used to capture trends on the labor market such as shifts in employment across industrial sectors, changes in labor demand/supply by educational attainment, age, gender. This microdata file contains most Labor Force Survey variables, including both demographic and labor related variables which allows to conduct detailed labor market analysis on a regular basis. Trends on the labor market are essential indication of economic structural changes and are important for the assessment of countries economic performance as well as for the analysis of future economic prospects. Further, detailed characteristics of the labor force and its respective changes are important for the assessment of the labor market cohesion with the current and future economic needs and the economic direction of an individual country. Common policies based on LFS More specifically, the data collected through LFS are of outmost importance for the a) analysis of the correspondence of the labor market trends and the characteristic of the labor force in particular with the desired economic trends, b) analysis of economic restructuring by changes in industry employment c) design of an effective employment policy and; d) specific (i.e. target group) employment policies e) the assessment of an appropriateness of education system and; f) the design of an effective educational policy, g) the design of a (minimum) wage policy h) the design of an employment benefit policy and; i) unemployment benefit policy impact assessment 40 j) the design of a choice of a specific socio-economic policies aiming at, i.e. 1) increasing employment of particular population groups such as the low income groups, rural unemployed, women etc; 2)qualification or prequalification (structural unemployment) of the unemployed; 3) support policies with respect to shifts in employment by education, occupation, and industry HOUSEHOLD BUDGET SURVEY - Overall purpose and its use in terms of policy-design HBS is specific research of population expenditure and income. The basic purpose of HBS is to determine a detailed pattern of expenditure. HBS also reflects household income and sources of income. HBS represents a useful instrument for measuring the monetary dimensions of poverty and wellbeing expressed in monetary values, given that HBS provides valid information about household incomes and expenditures. Nonetheless the HBS is often of multi-thematic nature with i.e. health and education modules added often so to capture the complex and multidimensional nature of poverty and to reflect the wellbeing of households in relation to other aspects of living. This is particularly the case for less developed or the developing economies. As such the survey provides numerous indicators for the formulation, monitoring and evaluation of key social and economic indicators. (e.g., measures of poverty and wellbeing, aspects of national accounts) and supplying specific data that would allow examination and evaluation of some specific programs and policies (e.g., impact of free primary education). Data applicability More specifically HBS is designed to provide necessary information for updating CPI and measuring and monitoring poverty and living standards, compiling national accounts statistics and updating employment statistics. The survey often aims at providing data on socio-economic aspects of the population such as education and health (adding short modules to gather basic info on health and education. Common policies based on HBS (extended) The data collected by HBS are essential for evidence based decision making with respect to numerous socio-economic policies; a) Measuring and monitoring of poverty on regular and methodologically coherent basis b) Analyzing living standards and changes in living standards and poverty trends on a consumption expenditure but also a multidimensional basis c) measure and monitor income inequality and analyse changes in inequality trends – allow for comprehensive analysis in terms of level of disaggregation as well as by demographic and socio-economic dimension of the population 41 d) analyse trends in household consumption/ income and changes in the structure of consumption by i.e. population categories, household profile, rural urban etc. d) provides data on type and cost consumptions for computing Gross Domestic Product (GDP) by expenditure methodology e) allows for assessment of pensions and social benefit payments, and the impact of other social transfers f) setting tax allowances/tax thresholds, using consumption (structure) trends g) providing a point of reference for evaluating changes in wages/salaries (minimum wage policy) h) providing information for monitoring the steps towards meeting the Millennium Development Goals i) allow for comparison of consumption and expenditure trends with other countries ,and the EU countries in particular J) allows for the variety of social impact assessment policies relying on the effects on changes in living standards of for instance poverty trends Data applicability of Consumer Price Index (CPI) CPI has the basic objective of: a) Measuring inflation and designing inflation targeting policies b) Comparing price movements and analyzing inflation trends and forcasting price changes e) Deflating incomes and consumption/expenditure for estimating changes in real incomes f) designing of an effective macroeconomic policy based on the inflation trends THE COMPARISON OF THE HBS AND LFS RESULTS WITH THE RESULTS OF THE EXISTING STUDIES The survey practice in Bosnia and Herzegovina started in 2001 with the implementation of the first survey-Leaving Standard Measurement Survey (LSMS). This survey introduced a new approach in the activities of the statistical institutions and adverted the importance of the standardization and harmonization with other European countries. The LSMS was also the starting point for the panel surveys “Living in B&H” which were conduced from 2002-2004 in order to get longitudinal data. Owing to these four surveys the authorities in B&H were able to do additional analyses based on household data in order to create different policies at all administrative level. This experience had a very important impact on the improvement of the “statistical culture” in the country which resulted not only with the practice of wider use of statistical data, but also with the establishing of new institutions producing statistics and analysis for the government (for. ex. EPPU). 42 The initial set of household surveys in B&H was continued with HBS in 2004 and LFS in 2005 and 2006 which created the occasion to compare their results. In this section the results of above mentioned surveys will be presented and compared. The attention will be paid only on the main results and indicators that we usually check and compare between different sources. These results are presented in the following table: Table 9 Main survey results No. Indicator 1 Population size 2 Population structure by gender (%) Male Female 3 Population structure by age groups (%) -14 15-24 25-49 50-64 65+ 4 Population structure by entity/district (%) FB&H RS Brcko 5 Population structure by area type (%) Urban Rural 5 Number of the households 6 Average household size 7 Education level (%) No education Primary education Secondary education High school, first university level University education 8 Legal status of use of the dwelling (%) Owner Rent or sublet Other 9 Durable goods ownership (%) TV PC Car 10 Household actual average annual income (KM) HBS 2004 3507868 49 51 17.3 15.3 5.6 17.8 14.0 62.0 36.0 1.9 39.9 60.1 1067120 3.29 28.2 29.5 37.6 2.5 2.4 87.1 4.6 8.3 94.9 12.7 47.5 7568,05 LSMS 2004 48 52 14,1 LFS 2006 3397328 49 51 18.8 14.7 34.9 17.1 14.5 63.5 34.7 1.8 39.8 60.2 1034538 3.28 29.4 28.7 36.6 2.6 2.7 n.a. n.a. n.a. n.a. n.a. n.a. 5744.28 15,7 3,05 24,4 25,3 44,3 3,0 3,1 36,9 43 11 Activity rate 12 Employment rate 13 Unemployment rate 14 Poverty Headcount Poverty gap Squared poverty gap Number of poor 60.2 39.9 33.8 17.9 5.0 1.0 627908 64.1 42.6 21.5 17.8 3.7 1.2 681455 n.a. n.a. n.a. n.a. 43.1 29.7 31.1 The basic demographic estimates from HBS and LFS are very close. There are small differences in the population size and number of households, but the population structures by gender, age, entity/district, area type and the average household size are almost the same. The same situation is present regarding the education level. There are differences in the household income averages, employment and poverty indicators between the analyzed surveys. One of the possible reasons for income differences was the difference in the income question design. In the LFS, the income was asked to be reported in income ranges, but in the HBS the households reported income in absolute values. The sensitivity of this question could also have an impact to the income values reported in both surveys. The differences in employment variables were caused by the significant differences in designing the employment questions. The LFS strictly designed these questions according the ILO definitions and recommendations in order to get the harmonized indicators of the employment. These questions in the HBS were designed on the basis of the self-evaluation of the employment status and the ILO definitions were not strictly applied. The poverty headcounts from the HBS and the LSMS were almost identical and the gaps very similar; the only difference was present in the estimated number of poor and was caused by the different population sizes which were the base for this calculation. For the LSMS poverty analysis the official estimates of the population size from the statistical institutions were used, and for the HBS the population estimate based on HBS data was used. These two estimates differed for the amount of approximately 300000. POTENTIAL IMPLICATIONS OF HBS, LFS AND CPI ON THE BIH MTDS Immediate implication on BiH MTDS will be low or medium basically because of: 44 • • • • Lack of time series Lack of analytical capabilities in B&H to use exiting data Methodological comparability with previous surveys is problematic until further exploration on data comparability would be done Results are similar to the findings of previous studies Long term impact will depend on regular repetition of these surveys. From the end of the war 1996 since 2006 surveys in B&H were financed by International Donor Community. As this type of finance is not regular, regular conduct of B&H LFS and HBS is under question mark. Ideally B&H governments should finance surveys from its own Budget. Capacity building of statisticians in future DEP and in B&H statistical system as well as building of working connections between these two institutions will be one other way to straighten more complex use of data from those surveys. Data dissemination policy will play important role in impact of these surveys. B&H is good example of transparent dissemination policy. Anonymous data bases of surveys as LSMS and LiBiH are available for free on the web pages of B&H statistical institutions. Anonymous data bases of HBS 2004 and LFS 2006 are not anymore publicly available on internet that disables brother research community to use those data. CPI as the measure of the inflation and the input for Real GDP calculation probably will have the highest immediate impact on B&H MTDS. Growth of Real GDP in B&H is used as main indicator for evaluation of MTDS. CPI already captured change in prices in January 2006 after introduction single rate of VAT that had implication on social policy in B&H. LFS was done for the first time in 2006. It allowed for the first time international comparisons of B&H labor indicators. International comparisons are important step toward EU integration process. Same as the LSMS 2004 even if the numbers are different LFS shows high unemployment rate and very high long term unemployment rate. Even if it is not possible to compare the change of unemployment rate using LSMS 2004 and LFS 2006 both surveys shows that situation in B&H is not good and at least is not changing over the time. These findings will emphasize the need of revision of employment policy in B&H. HBS 2004 is already included in update of B&H MTDS. In combination with LSMS 2004 it was basis for calculation of the Laeken indicators. Second wave of HBS in 2007 will be basis for poverty monitoring in period 2004 – 2007. However this result will be available by the end of 2008. 45 REFERENCES 1. Household Budget Surveys in the EU, Methodology and Recommendations for harmonization, EC, 2003 2. Household Budget Survey in the Candidate Countries, Methodological analysis, EUROSTAT, 2003 3. Labour market policy data base, Methodology, EC, Revision of June 2006 4. Labour force survey in the EU, Candidate and EFTA countries, Main characteristics of the national surveys, EC, 2004 5. The European Union labour force survey: main characteristics of the national surveys, EC, EUROSTAT, 2005 6. Labour force survey in acceding countries, Methods and definitions-2002, EC, EUROSTAT, 2004 7. Harmonized Indices of Consumer Prices (HICPs), A short guide for Users, EC, EUROSTAT, March 2004 8. HBS 2004-Main results, Agency for Statistics of B&H, 2004 9. CPI, First release, Peliminary data, Agency for Statistics of B&H, Jun 2006. 10. Statistics in focus: Labour Market Latest Trends 1st quarter 2006 data 11. Republic of Serbia, Office of statistics of Republic of Serbia, Report No 100, year LVI, 20.04.2006 Labour force survey October 2005 12. Republic of Macedonia, State Statistical Office, Report 2.1.6.22 Labour Market ,July 2006 13. Republic of Montenegro, Office of Statistics, LFS October 2005. 14. Provisional Institutions of Self Government, Ministry of Public Services, Statistical Office of Kosovo, Series No 5, Statistics of Labour Market 2005 15. DFID, BOSNIA AND HERZEGOVINA, LABOUR AND SOCIAL POLICY IN BOSNIA AND HERZEGOVINA:THE DEVELOPMENT OF POLICIES AND MEASURES FOR SOCIAL MITIGATION Contract Number CNTR 00 1368A Living in BiH, Panel Study WAVE 4 Report, Draft for discussion 46