Sector Review of the Implementation of the Labour Force Survey in Egypt Final report October2014

Transcription

1 Sector Review of the Implementation of the Labour Force Survey in Egypt Final report October2014 Prepared by Antonio R. Discenza Etienne Caruana

2 TABLE OF CONTENTS 1. INTRODUCTION The Labour Force Survey European Statistics Code of Practice Sector Reviews Sector Review of the LFS in Egypt The Self-Assessment Questionnaire Disclaimer General Overview of CAPMAS General Overview of the Labour Force Survey in Egypt FINDINGS Institutional Environment Statistical Processes Statistical Outputs RECOMMENDATIONS Recommendations on Part 1 Institutional Environment Recommendations on Part 2 - Statistical Processes Recommendations on Part 3 - Statistical Outputs ANNEXES ANNEX 1 Self-assessment questionnaire ANNEX 2 SR Agenda ANNEX 3 Long Questionnaire ANNEX 4 Metadata IMF Sector Review of the Implementation of the Labour Force Survey in Egypt i

3 LIST OF ABBREVIATIONS AGA Global Assessments CAPI Computer-Assisted Personal Interviewing CAPMAS Central Agency for Public Mobilization and Statistics CATI Computer-Assisted Telephone Interviewing EA Enumeration Areas ENP European Neighbourhood Policy ESA 2010 European System of Accounts 2010 ESCoP European Statistics Code of Practice ESS European Statistical System EU European Union EU-LFS European Labour Force Survey GDP Gross Domestic Product ILO International Labour Office IMF International Monetary Fund ISCED International Standard Classification of Education ISCO International Standard Classification of Occupations ISIC International Standard Industrial Classification LFS Labour Force Survey LPR Light Peer Review MS Master Sample Statistical classification of economic activities in the European NACE Community NSI National Statistics Institute PAPI Paper and Pencil Interviewing PSU Primary Sampling Units SAQ Self-assessment Questionnaire SDDS Special Data Dissemination Standards SR Sector Review UNECE United Nations Economic Commission for Europe Sector Review of the Implementation of the Labour Force Survey in Egypt ii

4 1. INTRODUCTION (1) The use of statistics has today become of paramount importance for research and policy making purposes. Today, statistics has many interested users, ranging from national governments, international organisations, the business community, non-governmental organisations, the media and the research community. In a time when all countries deem as an absolute necessity to use their own resources as effectively as possible, there is a strong need to engage in statistical enquiries which can provide relevant, comparable, and high quality results The Labour Force Survey (2) The Labour Force Survey (LFS) is a large household survey, designed to produce timely statistics on participation in the labour market, as well as on persons outside the labour force. In many countries, these statistics are collected on a continuous basis, and compiled on a quarterly basis. On the other hand, there are still a number of countries that carry out this survey only on an annual basis. A number of developed countries have been carrying out this enquiry for a long period of time and now possess long time series covering over two decades. (3) Within the European Union, this survey is governed by a regulation (Council Regulation (EEC) No. 577/98 of 9 March 1998) which ensures the highest possible quality results and full harmonisation across all participating countries. Definitions of employment and unemployment in the European Labour Force Survey (EU-LFS) follow strictly the guidelines stipulated by the International Labour Office (ILO), which establishes principles upon which measures of employment and unemployment should be collected and compiled. The strict adaptation of these principles within the ESS is a clear proof of Eurostat s active cooperation with international agencies and organisations, aimed at providing harmonised statistics at international level. (4) In practice, this means that by following strictly the ILO guidelines in the EU-LFS, Eurostat can compile harmonised labour market statistics which are comparable across the EU as well as with the data from non-eu member countries that are also carrying out this survey. It is this harmonised methodology used across many countries that makes the labour market indicators produced from the LFS often more relevant to users than the statistical reports derived from national employment/unemployment administrative registers European Statistics Code of Practice (5) Since 2004, the ESS has been working towards a consistent strategy for strengthening statistical systems across the European Union (EU). Good governance across the EU demands for high quality information on the economy and the society in general. In addition, measures need to be in place in order to guarantee the professional independence of the NSIs across the ESS, and to ensure that user needs are the basis for the production of statistics by of the NSIs. (6) The European Statistics Code of Practice (ESCoP), last revised in 2011, is a self-regulatory instrument which contains the principles and standards to produce and disseminate high quality statistics on the European Union. The ESCoP is based on 15 principles covering three major areas, as follows: Sector Review of the Implementation of the Labour Force Survey in Egypt 1

5 Institutional Environment: 1. Professional independence 2. Mandate for data collection 3. Adequacy of resources 4. Commitment to quality 5. Statistical confidentiality 6. Impartiality and objectivity Statistical Processes: 7. Sound methodology 8. Appropriate statistical procedures 9. Non-excessive burden of respondents 10. Cost effectiveness Statistical Output: 11. Relevance 12. Accuracy and reliability 13. Timeliness and punctuality 14. Coherence and comparability 15. Accessibility and clarity. (8) All European Statistical Authorities, comprising the Commission (Eurostat), National Statistical Institutes and other national authorities responsible for the development, production and dissemination of European Statistics, commit themselves to adhere to the Code. Eurostat also plays a key role in ensuring that ESCoP is fully respected within the EU Sector Reviews (9) As part of its role in cooperating with international statistical institutions and promoting statistics at international level, Eurostat also assists enlargement countries (Albania, Bosnia and Herzegovina, the former Yugoslav Republic of Macedonia, Iceland, Kosovo1, Montenegro, Serbia and Turkey), as well as countries covered by the European Neighbourhood Policy (ENP)2 in projects aimed at developing the production of statistics in line with EU and international standards. The ENP countries are: Armenia, Azerbaijan, Belarus, Georgia, Moldova, Ukraine, Algeria, Egypt, Israel, Jordan, Lebanon, Libya, Morocco, Palestine, Syria and Tunisia. (10) As part of this strategy, during the period Eurostat carried out a series of global assessments (AGA) and light peer reviews (LPR) among enlargement and ENP countries. While AGA s were held in order to assess the compliance of the NSIs with EU and international standards for the production of statistics, LPR were more geared towards the assessment of compliance with the ESCoP. The approach used during these reviews was very much similar to that used during the reviews carried out among the EU Member States during This designation is without prejudice to positions on status, and is in line with UNSCR 1244 and the ICJ Opinion on the Kosovo declaration of independence. Sector Review of the Implementation of the Labour Force Survey in Egypt 2

6 (11) Although these reviews have proved very efficient in determining strengths and weaknesses in statistical systems within these countries, the need for more detailed information on specific statistical sectors was felt. This led to the introduction of sector reviews (SR s). Contrary to AGA s and LPR s, SR s focus mostly on statistical processes in a particular sector, rather than on the whole statistical system. (12) In view of this, SR s are expected to provide very detailed information regarding the statistical sector and more specifically, the scope of the SR comprises: An assessment of the statistical production of the relevant sector; An assessment of the statistical production against the acquis; Reviewing of medium and long-term plans within the sector; Proposal of a list of actions to be undertaken in order to improve the data delivery and functioning of the sector under review Sector Review of the LFS in Egypt (13) The scope of this SR is to assess the Labour Force Survey (LFS) carried out by the Egyptian National Statistical Office (CAPMAS). For the purpose of this SR, two independent reviewers, namely Mr Antonio R. Discenza (Project Leader) and Mr Etienne Caruana, were engaged to collect and assess all the information required for the scope of this review. Both reviewers have extensive experience with the LFS and ESCoP, and are fully aware of Eurostat s recommendations for best practice in LFS. (14) The two main objectives of this SR were to assess the administrative and technical capacity of CAPMAS to produce high quality statistics on the labour force, as well as to assess the statistical production of this survey against the acquis. In order to carry out this task, reviewers used the ESCoP as their main guidelines for collection of information and analysis. (15) By using this approach, the reviewers were able to fully assess coherence of results produced from the Egyptian LFS with those produced at EU level. Significant attention was also given to quality, dissemination standards, as well as to documentation of work processes. In addition, the reviewers were requested to provide a number of recommendations based on their findings, for the improvement of the LFS in Egypt. (16) This SR was carried out in three distinct phases. The preliminary phase of this project entailed the selection of the reviewers by Eurostat. In addition, reviewers were briefed thoroughly and provided with necessary material in preparation for this task. The contractor also ensured that both reviewers were familiar with Eurostat s expectations for this SR and that all specific Eurostat requirements were taken on board during the planning stage of this SR. (17) During the second stage, the reviewers prepared a detailed self-assessment questionnaire (SAQ, please refer to Annex 1) covering nearly all the principles of the ESCoP, which was approved by Eurostat and sent to CAPMAS in preparation for the main review. The scope of the SAQ was to enable the reviewers to collect preliminary information regarding the Egyptian LFS. This questionnaire also served for CAPMAS to prepare for the reviewers mission. (18) The reviewers mission was carried out during the period April An agenda is provided in Annex 2covering all activities carried out during this mission. During the preliminary stage of the mission, the reviewers gave a clear overview of the objectives of the SR, as Sector Review of the Implementation of the Labour Force Survey in Egypt 3

7 instructed by Eurostat. Reviewers also requested CAPMAS to send a list of documents that were necessary for the SR. (19) Unfortunately, it was not possible for CAPMAS to fill the SAQ and send it to the reviewers before the mission. However, during the mission, the reviewers were given full support and cooperation from CAPMAS, in order to collect all necessary information for the SR as accurately as possible. The experts discussed all matters related to the LFS with all members of CAPMAS staff in an open and transparent way, which was indispensable for the success of the mission. Reviewers were also given the opportunity to visit CAPMAS training centre, the computer laboratory, as well as the GIS laboratory and discuss technical issues pertaining to the LFS with the people responsible for these Units. (20) The reviewers very much appreciated the openness of discussion by the CAPMAS-LFS team on all matters pertaining to their LFS, and the willingness to provide information for the purpose of this SR as accurately as possible. Reviewers also found exemplary the way by which the staff discussed the limitations in their statistical system and the willingness to improve their work. It is very clear that all CAPMAS staff members are highly enthusiastic about their work, and keen towards further improving the statistical capacity of this Office The Self-Assessment Questionnaire (21) The SAQ for LFS was based on the ESS Quality Assurance Framework 2012, and most of the questions were selected from the Eurostat s documents Joint Standard Quality Report for Labour Force Survey and Regional Labour Market Statistics and National Action Plan for the implementation of the recommendations of the task force on quality of LFS. The SAQ was divided in three parts, namely: Institutional environment: the first part of the questionnaire focussed on institutional and organisational factors that have a significant impact on the effectiveness and credibility of a statistical authority for developing, producing and disseminating LFS Statistics. Topics addressed in this part of the SAQ covered principles 2, 3 and 6 of the ESCoP. Detailed questions covering the principles 1, 4 and 5 were removed from the questionnaire as they were not deemed relevant for this SR. Statistical Processes: the second part of the SAQ assessed the main processes for the organisation, collection, compilation and dissemination of the Egyptian LFS against international standards, guidelines and good practices. Specific questions covering ESCoP principles from 7 to 10 were included in this section. Statistical Outputs: the last part of the questionnaire covered principles 11 to 15 of the ESCoP. In this part the Egyptian Labour Force statistical outputs were assessed against the ESS quality framework criteria, relevance, accuracy, reliability, timeliness, punctuality, coherency, comparability, accessibility to users and clarity. It aimed at verifying consistency with EU standards and comparability of national outputs with those from European countries Disclaimer (22) The Egyptian LFS was assessed against almost all of the ESCoP principles. For the purpose of this SR, reviewers assessed the LFS based on standards stipulated by European regulations governing this survey, as well as the ILO recommendations. Eurostat s LFS-explanatory notes were constantly referred to in order to assess coherence of the Egyptian LFS with the European Sector Review of the Implementation of the Labour Force Survey in Egypt 4

8 survey. Reviewers also referred to the recommendations made by the Eurostat task force on the quality of the LFS in order to assess the quality of this survey and make recommendations for improvement. (23) It is worth noting that, while the reviewers did their utmost to collect all information required for the purpose of this SR in an impartial and objective manner, limitations in the review process might still exist, due to the fact that: Observations and conclusions from this review depend primarily on information provided by CAPMAS, which despite being highly comprehensive, might still not be fully exhaustive. Moreover, observations, conclusions and recommendations made by the reviewers are based mainly on their experience in this line of work and their familiarity with the EU-LFS. It is possible, that some of the recommendations made by the experts might not be feasible to be implemented in a medium term by CAPMAS (24) In view of these limitations, the experts strongly recommend CAPMAS to invest in further research before implementing any of the recommendations that are being made by the reviewers in this report General Overview of CAPMAS (25) CAPMAS was established in 1964, and is the official source for national statistics in Egypt, under Presidential Decree (No. 2915). Employing more than 4,000 personnel and managing more than 60 regional offices, CAPMAS is the only governmental agency responsible for publishing official statistics in all demographic, social and economic fields. The office is also responsible for implementing the population and housing census and the economic census. CAPMAS has the role of regulating statistical work carried out by other Government Agencies. In fact, according to Egyptian legislation, no public or private entity may publish statistical data for national purposes without CAPMAS prior authorisation. (26) CAPMAS mission is to produce high quality, and timely statistics that match all user needs within the economic, social and environmental spheres, making it available to all users. In order to achieve these important goals, during the past years, CAPMAS has invested considerably in its statistical systems and collaborated with many national and international agencies by organising conferences, data exchange programmes, seminars and twinning programmes. CAPMAS also contributes effectively to research by providing data to researchers and organising regular conferences to promote statistical research initiatives. (27) CAPMAS has a relatively flat organisation. Apart from the statistical production units there is also the President s Office Central Administration, an Advisory Committee for Statistical Coordination and a Legal Office (which also fall under the responsibility of the President of CAPMAS). The statistical production falls under the responsibility of five departments, namely: Secretariat sector; Population statistics and censuses sector; Information technology sector; Economic statistics and mobilisation sector; and, Regional offices sector. Sector Review of the Implementation of the Labour Force Survey in Egypt 5

9 (28) CAPMAS has joined the Special Data Dissemination Standards (SDDS) thus fulfilling the requirements of the International Monetary Fund with a view of enhancing the availability of timely and comprehensive statistics in order to support macroeconomic policies3. As a member of the SDDS, CAPMAS satisfies a number of criteria (e.g. advanced news release calendar), which are also of interest for this SR General Overview of the Labour Force Survey in Egypt (29) The Egyptian LFS is the main source of labour market statistics in Egypt, and is considered one of the most important surveys in the country, targeting more than 90,000householdsevery year. Primary indicators from the LFS are published on a quarterly basis, while a publication containing more detailed labour market statistics is published every year. In order to carry out the LFS, CAPMAS follows the methodology and harmonised definitions as stipulated by the International Labour Office (ILO). (30) The first survey of this kind was carried out in Egypt in 1957, and has been developed ever since. One important development in the Egyptian LFS occurred in 2006, when CAPMAS revised its questionnaire to align its methodology with ILO s guidelines and recommendations. In 2008, CAPMAS started implementing this survey on a continuous basis, by spreading the surveys more evenly over the year. (31) During the past years, CAPMAS has also cooperated with international organisations like the United Nations and Eurostat in order to further improve this survey. One recent initiative to enhance the LFS entailed the introduction of panel sampling in order to improve the quality and consistency of results. CAPMAS invests considerable human and financial resources in carrying out the LFS. This survey is highly promoted among users, including researchers, who are encouraged to use this data for research purposes. (32) Work on LFS is mainly carried out by a dedicated unit which is responsible for the collection and analysis of this survey. This unit is supported by two other units, which are: the IT unit (responsible for providing the necessary IT tools for the collection and compilation of data), and the sampling unit (responsible for the sampling design of LFS as well as the calculation of sampling errors). 3 Please refer to the book The Special Data Dissemination Standard: Guide for Subscribers and Users Sector Review of the Implementation of the Labour Force Survey in Egypt 6

10 2. FINDINGS 2.1. Institutional Environment PRINCIPLE 2: MANDATE FOR DATA COLLECTION (33) All statistical enquiries carried out and published by CAPMAS, including the LFS, are governed by national legislation, Act No. 35, of 1960, which was later amended in In particular, this Act stipulates that CAPMAS is responsible for the collection and compilation of national statistics as well as for census taking. This Act also empowers CAPMAS to establish the methodology by which official statistics are collected, compiled and disseminated. (34) This Act empowers CAPMAS to collect individual data, as long as it is used for statistical purposes, and imposes obligations on this office to ensure non-disclosure of individual information. This Act also allows CAPMAS to access administrative registers for statistical purposes. However, for the purpose of LFS, CAPMAS still relies on households only to provide all the information. (35) This Act authorises CAPMAS to enforce participation in all household and enterprise surveys, including the LFS, and also stipulates legal actions which can be taken by CAPMAS, against respondents who refuse to provide accurate information for the purpose of this survey. Based on the non-response rates provided by CAPMAS for the Egyptian LFS, it is very clear that this Act is very effective in enforcing high participation in this survey. (36) Despite these strengths there are new and emerging statistical issues, also involving LFS, which are not tackled by this legislation. For example, the Act fails to stipulate rules and clear lines of action which should be taken by CAPMAS when anonymised micro-data is made available to researchers. (37) Moreover, although the Act empowers CAPMAS to access administrative registers for statistical purposes; it fails to empower CAPMAS to demand changes in existing registers mainly for statistical purposes. For example, administrative data held within the ministry responsible for employment and unemployment benefits still proves insufficient regarding registered employment and unemployment data, despite CAPMAS efforts to access this information to complement the LFS. PRINCIPLE 3: ADEQUACY OF RESOURCES (38) In general, CAPMAS personnel feel that the current level of human and financial resources assigned to the LFS is adequate. The management is clearly aware of the high levels of human and financial resources that need to be invested in order to ensure that this survey is carried out in accordance to high quality standards. CAPMAS also feels that LFS is currently satisfying all major user needs for labour market statistics, and this justifies all the resources that are currently being invested in this survey. (39) During 2011, CAPMAS assigned 217 full-time employees for the purpose of LFS. This number grew to 220 workers in 2012 and then to 250 employees in In addition, CAPMAS also employs more than 350 personnel on part-time basis, every year, to assist during the data collection. These figures do not include interviewers. When it comes to spending, CAPMAS Sector Review of the Implementation of the Labour Force Survey in Egypt 7

11 allocates 2.5 million Egyptian pounds (approximately 0.27 million) every year specifically for this survey. Table 1. Survey annual budget for each 100,000 population. Year Annual Budget Million LE Population Millions Budget per 100,000 population GDP Million LE % GDP ,104 1,475, ,037 1,508, ,954 1,677, (40) A large share of staff working on LFS is qualified in statistics or in a related field of study, and also has a long experience working in social surveys and compiling social statistics. There is a high level of commitment at all levels towards producing high quality statistics, and all staff understands the importance of their work. (41) CAPMAS is also highly committed towards developing the skills of all personnel further. In fact, during the past years, the office has developed an internal training centre which caters for many of the training needs, offering both general training programs and intensive statistical or IT programs for specific tasks. In addition, CAPMAS personnel are constantly encouraged to attend externally and internally organised training courses, as well as conferences. Moreover, technical personnel are encouraged to collaborate with researchers in order to further develop their skills. CAPMAS also organises regular conferences jointly with the research community where technical papers, often based on statistics collected by CAPMAS, are presented and discussed. (42) CAPMAS has a very strong IT infrastructure which supports collection, analysis and dissemination of data, including that of the LFS. During the past years, CAPMAS has developed its own data warehouse, which considerably improved the Office s statistical capacity and efficiency. At the moment, CAPMAS is working on a new website which is expected to improve CAPMAS dissemination standards. Furthermore, the IT unit is collaborating with the LFS team on a project aimed at replacing the existent Paper-and-Pencil Interviewing (PAPI) method used for data collection by a Computer-Assisted Personal Interviewing (CAPI). CAPMAS has also invested in a GIS system in preparation for the forthcoming Egyptian census. (43) There is however still need for more technical assistance on statistical techniques which are indispensable for smooth running of the LFS. It clearly emerged from this review, that although the CAPMAS-LFS team is highly familiar with the advantages of probabilistic weights in order to reduce the effect of non-response bias, they are still unfamiliar with the advanced weighting and calibration techniques that have been introduced in many countries during the past years. Sector Review of the Implementation of the Labour Force Survey in Egypt 8

12 PRINCIPLE 6: IMPARTIALITY AND OBJECTIVITY (44) CAPMAS does not have a written dissemination policy specifying methods and processes ensuring impartiality and objectivity when disseminating results. CAPMAS holds documents on work processes employed in LFS. However, these documents are not exhaustive and need to be enhanced. Moreover, there is no sufficient information provided to advanced users on the work processes carried out by the LFS team on this survey. (45) This does not mean that these principles are not upheld within this office. In fact, from feedback provided to the reviewers, it is very clear that this principle is taken very seriously by CAPMAS President and all the management team. The legislative status of CAPMAS allows it to determine the choices of sources and methods used in the compilation of LFS solely based on statistical considerations. Moreover, the Statistics Act does not allow for any external influence in the compilation of statistical reports. As a member of the SDDS, this office also abides to special dissemination standards set by the International Monetary Fund, stipulating impartiality and objectivity. (46).CAPMAS is committed towards correcting any erroneous data as soon as they are detected when it comes to correction of published statistics. However, there is still no publicly available written policy which establishes the principles and procedures by which these corrections are made. On the other hand, CAPMAS has a revisions policy which also stipulates that users should be informed in advance on major revisions and changes in the methodology of the LFS Statistical Processes PRINCIPLE 7: SOUND METHODOLOGY Infrastructural issues (47) CAPMAS has a methodological and technical infrastructure in place which defines, validates and monitors the implementation of the work processes of LFS. The responsibility for monitoring the work processes of the LFS falls on CAPMAS Central Administration Department, which reports on the quality of work and work processes directly to the President of CAPMAS. This unit has the responsibility of evaluating the quality of all CAPMAS projects, including LFS. It is also responsible for reviewing annual and periodical bulletins before publication stage. (48) In addition, sampling for LFS falls under the responsibility of the sampling unit, which establishes the methodology for maximising the LFS sample s efficiency. This unit is also responsible for maintaining and updating the LFS sampling frame. More details on the methodology by which these updates are made are provided below. (49) The IT unit supports the LFS team by providing the necessary IT tools for the collection, coding, data entry and compilation of results. It also computes the LFS weights based on instructions provided by the LFS team and the sampling unit. (50) Coding for this survey is carried out by an internal classifications unit. Sector Review of the Implementation of the Labour Force Survey in Egypt 9

13 Technical Staff (51) During the past years, CAPMAS has changed its recruitment policies in order to ensure that the persons selected for technical posts indeed possess the technical knowledge and abilities to carry out their work as efficiently as possible. In parallel, CAPMAS is implementing a series of initiatives, which mainly comprise training, aimed at improving the knowledge of existing staff. LFS questionnaire (52) The LFS questionnaire4 follows many of the ILO recommendations for measuring labour market statistics. In particular, it contains all the questions required for calculating the unemployment rate according to the ILO definition, although differences from the recommended ILOmethodology exist. Reviewers also noted that the Egyptian LFS questionnaire does not fully comply with the twelve principles stipulated in Annex 2 of Commission Regulation 1897/2000 in order to ensure full harmonisation of results. More details regarding these divergences are provided in the section entitled Divergences of national concept from European concepts and requirements. (53) CAPMAS currently uses two versions of the questionnaire - a long and short version. The main difference between the two versions is that the long questionnaire contains two ad-hoc modules addressing health and housing, while the short version of the questionnaire contains only labour market-oriented questions. Both questionnaires contain all the annual and quarterly variables5, which means that full consistency between the LFS annual and quarterly datasets can be achieved. However, this goal has not been achieved yet, due to the fact that annual weights are computed independently from the quarterly weights. Coding (54) Coding of occupations and economic activity is carried out internally by trained coders, based on a description of the job as reported by the respondents. Coding of the main economic activity of the enterprises is similarly recorded based on description of the main function of the enterprise as reported by the interviewers. Coding is carried out at the end of the data entry but only the codes are entered into the database for data analysis. Although the practice of entering only codes is acceptable due to the fact that entering the descriptions would prolong the data entry process, it limits considerably the ability of the analysts to verify the quality of the work carried out by the coders. (55) CAPMAS currently uses ISCO-88 for classifying occupations in the LFS and is planning to adopt the new ISCO-08 classification by In order to minimise errors during coding, CAPMAS has developed a 6-digit coding system for internal purposes, which classifies occupations in a much more detailed manner than the 4-digit ISCO. Although this fine-tuned classification has been designed on ISCO-88, it can still be effectively linked to ISCO-08. In view of this, revisions on past data using the new coding system are still possible. (56) CAPMAS uses ISIC Rev. 4 for classification of economic activity. Although many of the codes in ISIC Rev. 4 can be linked to NACE Rev 2, reviewers have identified a number of cases where 4 Annex 3 contains the LFS questionnaire which has been used in As defined in Com. Regulation n 377/2008 Sector Review of the Implementation of the Labour Force Survey in Egypt 10

14 one-to-one mapping (at full digit level) between these two classifications is not possible. Still the minimum European requirements for coding in LFS are satisfied since this linking suffices to provide reliable data at NACE division and section level. (57) In order to classify current and highest achieved level of education, CAPMAS uses a pre-defined list of education levels. The current level of disaggregation of these items does not suffice to enable complete one-to-one mapping to ISCED for both 1997 and 2011 versions6.however by the end of 2014, CAPMAS intends to start applying the new ISCED2011 classification. CAPMAS is also currently studying the possibility to construct a mapping table in order to recode the past data into ISCED 1997 and ISCED LFS target population (58) The Egyptian LFS is a household survey, and therefore does not include persons living in institutional households. A household is considered to be a group of persons, generally family members, sharing expenses. However, non-family members, like servants, who also share the household s resources are also considered as members of the household. (59) In the Egyptian LFS, persons who leave the household for a period exceeding 6 months are considered as non-members of the household for the purpose of LFS. This definition is not in line with the UN definition of usual residence, which stipulates that the duration criteria should be of 12 months. CAPMAS insists of using the 6-month threshold for LFS to ensure full consistency with the latest population and housing census, which was carried out in (60) Moreover, surveys carried out by CAPMAS address only persons of Egyptian nationality, which means that foreign residents are excluded from the Egyptian LFS. Also excluded from this survey are persons working in the Armed Forces. Sampling Frame for LFS (61) For sampling purposes, CAPMAS uses a Master Sample (MS) which is used as sampling frame for all social surveys, including the LFS. The MS can be regarded as a sample of 5,024 Enumeration Areas (EA s) with over 1 million households, taken from CAPMAS register of households and persons, produced from the latest population and housing census. On average, every enumeration area has 200 households. However, EA s with much less than 200 households do exist, especially in rural areas. In some cases, CAPMAS merges adjacent small EA s prior to sampling so that all Primary Sampling Units (PSU s) have very similar sizes prior to selection. (62) Selection of PSU s for the MS is made through stratified random sampling based on the 27 governorates and on the degree of urbanisation (urban/rural), where equal proportions of PSU s are selected within every strata. Prior to the selection of these units, the frame is arranged in such a way as to enable implicit stratification with regard to the geographic location. In practice, this means that each sub-governorate (kism/markaz) is ordered in a serpentine fashion according to the geographic location within the strata based on its location. This approach ensures that an adequate representation of every strata is obtained in the MS, thus improving the efficiency by which sampling is carried out. Implicit stratification is possible since CAPMAS has developed a system by which the location of all EA s could be identified from their census-based register of 6 It seems that at the moment is not possible to distinguish Masters from Ph.D. as required by the new classification. Sector Review of the Implementation of the Labour Force Survey in Egypt 11

15 households. The MS is designed to cover all parts of Egypt. However, no estimates of under-or over-coverage are currently available for LFS. (63) Every five years, the MS is replaced by a new, and independent, set of 5,024PSU s. Whenever possible, extraction of the MS is carried out from a new census-based register, in order to ensure that information regarding all the households in the MS is as updated as possible. During the inter-censual periods, when a fresh register of households is not available, CAPMAS updates the selected list of EA s through a specific enquiry. The current version of MS was extracted from the 2006 population and housing census and updated in (64) Apart from the contact address, the MS also contains auxiliary information regarding the households which can be used for sampling purposes, including: governorate name and code, city name and code, shiakha/village name, degree of urbanisation groups, PSU, road name, road number, block number, building number, organising number, building owner name, location and number of the housing unit in building, name of head household. The LFS Sample (65) Every year, the MS is split into four equally sized sub-samples of 1,256 PSU s, i.e. one for each quarter. A gross sample of 18 households is then selected randomly from each of the selected PSU and contacted for the LFS. Two additional households are selected per PSU to enable substitution of non-responding households. In this regard, reviewers would like to remark that the use of substitution of non-responding households is not widespread within the ESS and its effect is still controversial. The sampling scheme used by CAPMAS also incorporates the use of a panel, as described in the following section. (66) In the effort to limit the burden on households as much as possible, CAPMAS eliminates households that already participated in previous surveys from the MS prior to selection. (67) The Egyptian LFS is carried out on a continuous basis; however, not all the weeks of the year are covered. More specifically, every quarter the selected EA s are split randomly into 5 equal subsamples and assigned for interviewing for specific weeks of the quarter as follows: 1st month: week 1, week 3; 2nd month: week 1, week 3; 3rd month: week 1. The last month of every quarter is covered by only one week, while the remaining two months are represented by two weeks. (68) At the moment, it is not possible for CAPMAS to cover more weeks from the last month due to a number of operational constraints which are limiting the efficiency of data collection and data analysis. In fact, according to CAPMAS personnel, under the current circumstances, if CAPMAS decides to prolong the period of data collection to cover all the months equally, then CAPMAS will risks not meeting its LFS publication deadlines. (69) Although this approach suffices to compile reliable quarterly and annual estimates, it still fails to provide full coverage of all quarters, since more than half of the survey weeks are skipped. Moreover, the months are not covered equally and this may introduce biases to the sample. (70) CAPMAS is considering ways by which this problem can be solved. At the moment, CAPMAS is planning to introduce Computer Assisted Personal Interviewing (CAPI) in the LFS. It is hoped that the improved efficiency in data collection that will result from this significant investment, will suffice to allow a better coverage of all the reference months. Sector Review of the Implementation of the Labour Force Survey in Egypt 12

16 Documentation (71) It is CAPMAS policy to make available relatively detailed information regarding the methodology used for the compilation of statistics carried out within this Office. The LFS is no exception. In this regard, CAPMAS produces metadata files highlighting the methodology and principles used for the compilation of the employment and unemployment rate from the LFS. These metadata files can also be downloaded from CAPMAS website. (72) Information on the LFS that can be accessed from these files includes details of the sampling methodology used, mode of data collection, etc.(examples of these metadata files are provided in Annex 4.) More detailed information for the users is provided in the LFS annual publication. (73) Reviewers noted that these methodological notes are very readable and clear, and are surely useful for the general public to understand the basic concepts and technical issues behind the published figures. However, the level of detail does not suffice for researchers who need to use LFS for more advanced research. (74) The reviewers also noted a number of inaccuracies in the metadata files which might lead to misleading conclusions regarding the quality of the survey itself. One example is the following: metadata reported that the LFS is carried out in four surveys in January, April, July and October but actually the surveys cover all the three months of each quarter. CAPMAS is committed to correcting these inaccuracies immediately. Methodological Improvements (75) CAPMAS is clearly very open to learn from best practices in order to improve its statistical system. This is also evident for the LFS. In fact, during the past years, CAPMAS cooperated with international organisations (e.g. ILO, Eurostat, etc.) and with the research community (e.g. Universities) on new initiatives aimed for improving the survey. (76) During the past years, CAPMAS has also proved to be very keen to introduce new methodologies and change past practice to improve the survey. A clear example of this is a new project that is currently being implemented by CAPMAS, on the introduction of CAPI in the LFS. This initiative is in line with Eurostat s recommendations. PRINCIPLE 8: APPROPRIATE STATISTICAL PROCEDURES LFS questionnaire and data collection (77) The LFS team consults the field-workers and supervisors on a regular basis in order to assess the validity of the questionnaire. Feedback is also sought from the responding households themselves, as well as from the main users of data. Basing on this feedback, CAPMAS makes regular updates in its questionnaires. (78) All revisions to the questionnaire are thoroughly tested prior to data collection, through pilot surveys that are carried out among small samples of households. No written procedures are currently in place explaining the processes by which these tests are made. (79) Data collection is currently carried out by face-to-face interviewing using paper questionnaires (PAPI).Small groups of EA s are assigned to supervisors who monitor the work and progress made by the interviewers working in those areas. Supervisors are responsible for the quality of Sector Review of the Implementation of the Labour Force Survey in Egypt 13

17 work carried out by the interviewers and are expected to assist them in case they experience any difficulties during their work. (80) All interviewers and supervisors are briefed prior to the data collection, and are provided with regular training on basic skills required for field work. In particular, training sessions are carried out every quarter among all interviewers and supervisors, which entail: Delivering presentations on general issues related to the LFS (e.g. survey importance, objectives, definitions, methodology according to the international standards); Detailed briefing of the LFS questionnaire; and, On-the-job training using case studies. (81) Interviewers are also provided with detailed manuals explaining all concepts and definitions used in the questionnaire. Interviewers are given a letter signed by CAPMAS President explaining the purpose of LFS, as well as the legal obligation of the selected households to participate in the survey. Interviewers are obliged to show this letter to the interviewed households prior to data collection. (82) In order to minimise the rate of non-contacts, interviewers are requested to carry out a minimum of three visits in every household before classifying them as non-contacts. Proxy interviews are allowed, although highly discouraged. Interviewers are requested to indicate proxy interviews in the questionnaire so that the rate of proxy interviews can be monitored on a regular basis. (83) In order to monitor the quality of LFS data collection, CAPMAS has a quality control unit which regularly carries out audit checks among a highly representative sample of responding households. These audit checks entail re-collecting of partial information from the LFS which has already been collected by the interviewers. This information is then compared with the data provided by the interviewers, in order to assess the quality of work carried out by the interviewers and supervisors. Every month, CAPMAS produces reports on interviewers and supervisors performances based on these audit checks and immediate corrective measures are taken whenever required. The LFS Sample: (84) Based on the information provided in the previous section, it can be deduced that the sampling methodology used by CAPMAS is based on a two-stage cluster design, whereby every year: 5,024 PSU s (or EA s) are selected by stratified random sampling within urban and rural governorates, as explained above, and, A gross sample of 18 (+2) households per PSU is selected. (85) The selected PSU s are then split into four equal and independent sub-samples to be interviewed over the four consecutive quarters. On average, CAPMAS contacts approximately 90,432 households for the purpose of LFS, and a sampling rate of 0.45% is achieved. (86) During the past years, CAPMAS has also explored the possibility of introducing panel sampling in order to improve sampling methodology efficiency. In 2007, CAPMAS initiated a rotational design 2-(2)-1 which comprised a system where one third of the interviewed households (approximately 6 households per PSU) were carried forward to the following quarter, while Sector Review of the Implementation of the Labour Force Survey in Egypt 14

18 another third were carried forward to the following year. With this system, only a third of the interviewed households per quarter were new panel households. (87) This methodology was stopped in 2013, and then replaced by a more simplistic design in The new panel design comprises a 50% overlap of the households interviewed during the third quarter which are forwarded to the consecutive year. This approach is clearly less efficient than the one which was used during the period Administrative Registers (88) CAPMAS does not use administrative data for the purpose of LFS. Moreover, CAPMAS still has not explored the possibility of linking information obtained from households with administrative registers. This means that all information pertaining to this survey is currently being collected directly from households. Classifications (89) CAPMAS has an internal classifications unit responsible for coding (mainly ISIC and ISCO). Coding is done manually by trained personnel who use specifically designed software/programs to make the coding process as efficient as possible. This software is based on a search engine which helps the coders to select the appropriate code. Automated programs for coding have not been implemented yet, although CAPMAS is currently exploring this option, as part of the CAPI project. Calculation of weights for LFS (90) The computation of probabilistic weights for the Egyptian LFS is currently being made using post-stratification weighting, at persons level by sex, degree of urbanisation and governorate level. Quarterly household population estimates provided by the demographic statistics unit, are used for the computation of these weights. These figures result from updates on household population statistics from the latest population and housing census. Updates are made on quarterly basis by taking into account only the natural increase. The effect of migration on household population counts (and distributions) is ignored since it is considered insignificant. (91) Reviewers noted that age is not being taken into account for post-stratification weighting, which in practice means that inconsistencies between consecutive quarterly surveys can be expected, especially in age-dependent variables. Moreover, this methodology does not comprise corrections for unit non-response bias and panel attrition (where applicable). Some inconsistencies also exist between the surveyed population and figures used for weighting purposes (e.g. members of the army are excluded from the LFS sample but included in demographic estimates). (92) Moreover, since the data is not being calibrated at household level7, the integrated design of the LFS is not being taken into account for the computation of the final set of cross-sectional weights. As a consequence of this, this survey cannot be used to produce high quality and 7 Calibration at household level would imply that all the household members have the same individual final weight, which can be used to produce estimates from the household perspective. Sector Review of the Implementation of the Labour Force Survey in Egypt 15

19 coherent estimates at household level (e.g. number of reference persons who are unemployed is not coherent with the number of households whose reference person is unemployed). Treatment of item non-response (93) CAPMAS relies considerably on the data collection team in order to keep item non-response at minimal levels. Any missing information identified by supervisors, or the central office, is verified with the interviewers and the households. Missing information retained further to the data entry process is not imputed during data analysis and reported as missing. In this regard, reviewers have noted that CAPMAS still has not developed proper statistical imputation procedures to tackle item-non-response. Revisions of past data (94) So far, CAPMAS have never carried out any revisions of past time-series data, even when significant changes in the data collection methods were introduced. Reviewers have also noted that CAPMAS does not have written work procedures to be implemented, should revisions be required. PRINCIPLE 9: NON-EXCESSIVE BURDEN ON RESPONDENTS (95) The CAPMAS Advisory Committee has the responsibility for monitoring the use of CAPMAS published statistics. This Advisory Committee advices the President s Office on measures to be implemented in order to improve the relationship between CAPMAS and users. (96) So far, CAPMAS had never carried out user surveys in order to measure the level of satisfaction of users with the data produced from specific surveys, like the LFS. However, processes are in place in order to assess the usability of statistical data and whether the information that is being disseminated justifies the associated costs and burden. Measures implemented by CAPMAS in order to monitor data usage include specific meetings with data users and opinion polls of the general public. (97) From the feedback obtained from users, CAPMAS feels that the LFS is meeting most of the user demands. This survey is also regarded as the main source of labour market statistics in Egypt, which makes it highly relevant for policy makers. Consequently, CAPMAS feels that ensuring highest quality standards for this survey should be prioritised over all other principles, including household response burden. This is the main reason why CAPMAS feels that there should be no distinction between annual and quarterly ad-hoc modules for labour market oriented questions in the LFS and all questions in the LFS should be treated as if they were core questions. One important measure that has been taken by CAPMAS so far, in order to limit the burden on respondents, was the introduction of a long and a short questionnaire so that ad-hoc modules designed to collect non-labour market statistics are introduced only during the third quarter8. (98) In addition, the sampling unit also flags households which participated in past social surveys so as to ensure that these are not burdened again with a new survey. The reduction in quality due to this measure, and its impact on the final results, can be ignored. 8 The short questionnaire takes 15 minutes while the long questionnaire takes 25 minutes per household to compile Sector Review of the Implementation of the Labour Force Survey in Egypt 16

20 PRINCIPLE 10: COST EFFECTIVENESS (99) During the past three years, annual spending on this survey amounted to 2.5 million Egyptian Pounds, as shown in Table 1. Spending patterns are expected to change significantly in the coming years further to the introduction of CAPI. (100) On average, CAPMAS employs a total of 4,050 employees9 in the central and regional offices. From these, nearly 900 employees are statisticians, while approximately 2,300 are other professional and managerial staff (e.g. supervisors). An additional workforce of 291 interviewers is engaged every month for data collection. (101) CAPMAS has an internal mechanism in order to monitor the use of financial and human resources for the production of statistics. In case of the LFS, this responsibility falls directly under the Head of the Population Statistics and Census Unit, who is also responsible for estimating costs and allocating resources for new initiatives related to this survey. (102) The Head of the Population Statistics and Census Unit reports to the President s Head Office. The latter carries out separate audits on overall use of resources. No externally managed mechanisms are yet in place in order to monitor the use of CAPMAS financial and human resources Statistical Outputs PRINCIPLE 11: RELEVANCE (103) As a member of the SDDS, CAPMAS is obliged to give equal access to data to all its users. This entails the dissemination of an advanced news release calendar, at least three months before the dissemination date, and simultaneous access of data to all its users. Within the SDDS there are also specific time schedules, set by the IMF, within which LFS data has to be published. The news release calendar can be accessed freely on CAPMAS website. (104) CAPMAS also issues yearly publications with detailed results from the LFS apart from news bulletins highlighting salient results from the LFS. Other information available on CAPMAS website includes very old data with labour market indicators, going back to So far, seasonally adjusted data is not being compiled by CAPMAS. (105) CAPMAS also makes available anonymized10lfs micro-data to all users for free. All this information can be downloaded from CAPMAS website, but is subject to registration. (106) CAPMAS is currently working on a new website to enhance the Office s dissemination standards. Apart from being more efficient, this new website will contain an online dissemination tool which will enable users to extract custom-made tables directly from the data warehouse. (107) CAPMAS personnel indicated policy makers, media, researchers and businesses as main users of LFS statistics. CAPMAS invests considerable resources in order to ensure that the LFS remains highly relevant to these users. For example, this year, CAPMAS introduced more questions 9 Full-time equivalent, excluding interviewers 10 However, the reviewers noted that CAPMAS is not using any appropriate software for this task thus there are concerns that the LFS are actually fully anonymized. Sector Review of the Implementation of the Labour Force Survey in Egypt 17

21 addressing under-utilisation of the labour force and potential increases in the labour market, since it was highly demanded by policy makers. (108) Information mostly sought by users includes statistics on employment and unemployment. This information is mainly sought at governorate level, and disaggregated by core demographic variables like age and sex. Statistics on educational attainment and participation in lifelong learning are also sought by LFS users, although important education-related indicators such as the early school leavers rate and tertiary educational attainment rate are still not being computed. PRINCIPLE 12: ACCURACY AND RELIABILITY Data collection (109) A detailed overview of all the procedures that have been adopted by CAPMAS so far in order to minimise errors during data collection has already been provided in this document. Reference was also made to the audit processes that are currently in place in order to monitor the work carried out during the data collection. (110) Very strict checks are also made on the filled questionnaires in order to ensure that complete information is provided by all households. The data entry program that is used by CAPMAS also has a number of in-built validation rules to identify inconsistencies during the data entry stage. (111) CAPMAS also carries out double data entry on 5% of its questionnaires in order to verify the quality of the data entry process. (112) The following table provides measures of non-sampling error rates measured by CAPMAS prior to the data analysis stage: Table 2 - Non-sampling errors Errors Error rate in % Errors found by supervisors 2 Errors found during data editing stage, prior to data entry 8 Errors found during data entry 3 Number of inconsistencies illogical errors identified before coding 10 (113) CAPMAS also compiles figures on non-response rates at household level. These rates are underestimated, since they are computed further to substitutions. Non-responding households also include ineligible households and non-contacts. Sector Review of the Implementation of the Labour Force Survey in Egypt 18

22 Table 3 - Household non-response rates in LFS Survey year Quarter Non Reason for non-response (%) response Noncontacts reasons Other rate (%) Refusals Quarter Quarter Quarter Quarter Chart 1. Household Non-response rates by Governorate (%) Cairo Alexandria Port Said Suez Damietta Dakahlia Sharkia Kalyoubia Kafr El sheik Gharbia Menoufia Behira Ismaelia Giza Beni Suef Fayoum Menia Asyout Suhag Qena Aswan Luxor Red Sea El wadi El Gidid Matrouh North Sinai South Sinai Data analysis (114) Quality checks are made during the analysis stage in order to identify misleading information. Still, these checks do not seem to be carried out in a standardised and systematic manner. In fact, processes outlining the rules which need to be followed by statisticians in order to identify inconsistencies in the LFS datasets are not documented. This means that, with the current system, quality checks are being implemented manually by the analysts, thus leaving room for inconsistencies in the approaches used. (115) Moreover, no procedures seem to be in place in order to determine the effect of weights and imputations on the main estimates. Statistics on possible under/over-reporting are also not compiled, making it very difficult to reduce their effect using appropriate statistical methods. Sampling errors are also not computed, not even on the main LFS indicators. Reviewers also noted that the CAPMAS team has not yet adopted a common statistical package which can be used to validate and analyse appropriately LFS data. Sector Review of the Implementation of the Labour Force Survey in Egypt 19

23 (116) Quality checks are in place in order to determine the level of consistency of results, by comparing LFS estimates with past time series data. No comparisons are made with auxiliary information in order to determine the level of coherence of LFS with other statistical sources. Item non-response (117) CAPMAS invests considerable resources for the data collection in order to minimise item nonresponse, as explained in previous sections of this report. However, information that is still missing after the data collection is not imputed during data analysis, and is reported as missing data. Sampling frame (118) The processes carried out by CAPMAS for updating the LFS sampling frame have already been explained in previous sections of this report. The fact that the MS is not updated on a yearly basis is of most relevance to the quality of LFS, leaving room for possible under/over-coverage problems in LFS results. PRINCIPLE 13: TIMELINESS AND PUNCTUALITY (119) CAPMAS respects SDDS timeframes for the publication of LFS quarterly results. In addition, CAPMAS produces a publication with detailed annual LFS statistics, four months after the end of the reference year. Anonymised micro-data for researchers is made available just one month after this publication. According to CAPMAS team, there have never been instances when these deadlines were not met. PRINCIPLE 14: COHERENCE AND COMPARABILITY Inconsistencies in time-series data (120) CAPMAS has been compiling LFS statistics since A long time series on labour market data is therefore expected. Although this data is available, it might not be fully suitable for timeseries analysis due to inconsistencies caused by methodological changes that occurred during the past years. (121) Major changes have been made during the past years in the LFS questionnaire in order to align the Egyptian LFS with ILO standards, which have surely accentuated these inconsistencies. For example, one of the major changes that were carried out recently on the LFS questionnaire, entailed the addition of auxiliary questions to the LFS questionnaire in order to determine more accurately whether persons who declare Not Having worked during the reference week, not even for one hour and Were absent from a job/business should indeed be classified as unemployed. This change, which was implemented in 2007, involved adding a number of specific questions in order to identify those persons which have been involved in minor employment activities (e.g. agriculture production, vegetables or food preparation, sewing and knitting, paid handicraft, fishing, etc.). This caused a significant break in series, with a consequent increase in the employment and a decrease in the unemployment. Another significant change which also occurred during the past entailed a modification in the panel design, i.e., from a rotational pattern 2-(2)-1 to a simpler design comprised of one overlapping quarter only. Sector Review of the Implementation of the Labour Force Survey in Egypt 20

24 (122) In each of these cases, no revisions have been made by CAPMAS for historical data in order to avoid breaks in series. In addition, these breaks in series were not flagged to warn users about potential inconsistencies in the main LFS indicators. Inconsistencies caused by weighting methodology (123) After assessing the methodology used for weighting purposes, reviewers concluded that annual data might not be consistent with the quarterly data at regional level, due to the fact that annual weights are computed independently from the quarterly weights. (124) In addition, as already explained above, weights are computed at individual level, which means that consistency between household estimates and individual estimates by household type cannot be ensured. For this reason household estimates are not disseminated. CAPMAS tried in recent years to produce household weights but did not succeed, probably due to lack of experience and lack of knowledge about suitable methods and software. Divergences of national concept from European concepts and requirements (125) Although several adjustments have been recently made to better classify the population according to ILO recommendations, there seem to be room for further improvements. The following is a list of divergences from international recommendations that were noted by the reviewers. Definition of resident population for the scope of the LFS is different from the UNECE and Eurostat (ESA 2010) concept of usual residence a 6 months criteria is currently used instead of one year (or more) one. The age range for unemployed person is 15-64, instead of the as recommended by ILO. The unemployment classification used by CAPMAS uses a reference period of three months for the duration of seeking work, instead of four weeks (or one month) as recommended by ILO and Eurostat. The wording in the questions used to classify persons by main economic status according to the ILO definition (see for example Annex 3, questions 117, 118 and 119) could be improved, in order to improve harmonisation of results. For example, it must be made more clear that: the category of persons who are employed includes all persons who were engaged in any kind of work done for pay or profit for at least one hour during the reference week, where the pay includes cash payments or "payment in kind" (payment in goods or services rather than money); employment includes also the unpaid family workers, i.e. those persons working in a family business or on a family farm without pay, living in the same household as the owner of the business or farm. Moreover, The specification of the reference week from Monday X to Sunday Y should substitute the more general term past week. The criteria of formal attachment to the job in order to consider a person which is absent from work employed, should also be assessed in the questionnaire. Reviewers also noted that the order of the questions in the LFS questionnaire is not according to the European and ILO recommendations. In this regard, Commission Regulation (EC) No 1897/2000 concerning the operational definition of unemployment (Principle 1 in Annex II) states that questions on the labour status according to the ILO definition should not be Sector Review of the Implementation of the Labour Force Survey in Egypt 21

25 preceded by questions on the main or the usual activity (student, housekeeping, retired, etc.).this is envisaged to avoid a prejudice in the response to the questions on the ILO labour status. This principle has been widely interpreted in Europe by Member States in the sense that questions concerning participation in education and training during last 4 weeks cannot precede those regarding ILO status in the personal/individual questionnaire. At the same time, the highest level of education completed (which is considered a core social variable) is often asked before the ILO status, in the household section, in order to be able to introduce consistency rules between education attainment and occupational level. This specific order of questions is not being respected in the Egyptian LFS questionnaire. Moreover, an additional module about The Housing Conditions of the Household is added at the beginning of the long questionnaire and collected during the third quarter of the year. This contrasts with the international recommendations because of its possible influence on the answers to the standard LFS questionnaire. Target population is the civilian population excluding persons working with armed forces; however the results are grossed up to the total population, thus ignoring inconsistencies between the sample and target population. Coherence of LFS data with population statistics is assured by gender and at regional levels only, since age is not used in the post-stratification weighting process. In practice, this means that inconsistencies in variables which are highly dependent on ages of respondents are expected. PRINCIPLE 15: ACCESSIBILITY AND CLARITY (126) LFS data are disseminated using both traditional means and modern web technology. Quarterly bulletins are made available to the public in printed copies, as well as in electronic format. The quarterly bulletins contain a wide set of tables and useful metadata that facilitate proper interpretation of results to the users. CAPMAS also offers assistance by telephone or by to assist users to interpret these results. Custom-made reports are compiled by CAPMAS upon request, against a cost recovery charge. (127) A specific web page containing information on the national LFS exists. It is translated in English and displays only partial data and metadata. Full access is guaranteed only to enrolled users who have paid a small enrolment fee, as previously explained. CAPMAS is currently working on a new and more efficient website, which will contain more detailed metadata on all the data produced by CAPMAS, including the LFS: (128) When it comes to dissemination of micro-data, CAPMAS already provides anonymised microdata from the LFS on its website. CAPMAS is also currently working on a policy establishing clear rules, protocols, restrictions and anonymisation criteria by which anonymized data can be disseminated for research purposes. Sector Review of the Implementation of the Labour Force Survey in Egypt 22

26 3. RECOMMENDATIONS 3.1. Recommendations on Part 1 Institutional Environment Rec.1. There is strong need for more communication and collaboration between different units and divisions, in order to allow for more cross-cutting initiatives to further improve the LFS. a. Such collaboration requires more detailed and organised documentation of all the internal work processes and their changes in time (methodology, organization, actors involved, phases, timing, necessary inputs, expected outputs, methods, software, etc.). b. There needs to be a clearer division of roles and responsibilities within the directorates in order to avoid work duplication. When a process has to be run jointly by two separate/independent units, it is important to define who the stakeholders are. Rec.2. Further methodological assistance/training is required on more advanced statistical techniques and best practice which have been successfully implemented in the LFS by more experienced NSI s. Training is also required on statistical packages in order to improve efficiency during data analysis. Areas where training is required include: a. complex sampling design (stratification, rotation pattern, space-time allocation of the sample, oversampling, replacement/substitutions of final units), in order to explore the possibility to get more efficient designs; b. analysis of non-response and methods for non-response adjustment; c. verification of data and imputation of item non-response; d. weighting methods, moving from simple stratification to calibration, exploiting the possibility to get household weights; e. seasonal adjustment of time series; f. calculation of standard errors and coefficient of variations, and their dissemination; g. anonymisation techniques of micro-data; and, h. linking longitudinal micro-data and longitudinal estimates. Rec.3. Proposals for changes in the LFS should be thoroughly assessed before they are implemented. a. Any proposed change should be evaluated in detailed and comprehensive technical papers, in which the impacts of all the aspects of the survey, are carefully assessed (organization, methodology, results, break in time series, etc.). b. Reviewers also recommend the setting-up of an internal cross-unit task force to analyse new methodological proposals and their impact on processes and outcomes. c. Reviewers strongly advice to seek more than one expert s opinion on important technical issues. d. CAPMAS should give more importance to the comparability over time when introducing major modifications to the LFS. In cases when these modifications cannot be avoided, then CAPMAS should consider revising historical data in order to enable analysis of LFS statistics over time. e. CAPMAS should carefully monitor possible breaks in the time series caused due to changes which shall shortly be implemented in the rotational structure of the LFS sample (as the sample has no more overlap, the estimates of changes from quarter to quarter could be less precise than in the past). Sector Review of the Implementation of the Labour Force Survey in Egypt 23

27 Rec.4. Rec.5. Carry out coherence tests on LFS data and ensure that differences between labour market statistics derived from LFS, National Accounts and Business Statistics are kept to a minimum. In addition, CAPMAS should provide reasons to the users on why these differences exist. Investigate the possibility of using more administrative sources for statistical purposes, even for the purpose of LFS Recommendations on Part 2 - Statistical Processes Rec.6. Enhance the LFS questionnaire 11 for better compliance with ILO and Eurostat recommendations: a. Change the reference week for unemployment, which should be four weeks rather than three months; b. Apply the international recommendation for the age-range of unemployed persons (a range is currently used instead of 15-74); c. Apply the ESA2010 concept of usual residence as stated by UNECE/Eurostat Recommendations for the 2011 round of census (6 months criteria is currently used instead of one year or more); d. improve the set of questions used to identify employed persons clarifying that: any kind of work done for pay or profit should be taken into account; mode of payment can be both in cash or " in kind" (payment in goods or services rather than money); unpaid family workers should also be considered, i.e. those persons working in a family business or on a family farm without pay, living in the same household as the owner of the business or farm; the criteria of formal attachment to the job in order to consider employed a person which is absent from work. e. Improve the specification of the reference periods for employment using the words from Monday X to Sunday Y instead of the more general past week. f. Move the two modules on housing and health, which are present in the long questionnaire, from the beginning of the questionnaire to the end of the questionnaire (as an additional module), in order to avoid possible influence of these questions on the answers to the standard LFS questionnaire; g. verify, with experts from the Educational Sector, that the questions about education variables allow recoding into ISCED2011; h. evaluate the option to move the questions on the current enrolment and participation in formal education and training after the sections Table 2 and Table 3 of the LFS questionnaire (see Annex 3). Moreover, both current education and general/vocational training should be asked to all the persons aged over 14 years, not only to those people who are not enrolled in formal education. 11 With reference to the LFS Questionnaire, version 2010, in Annex 3. Sector Review of the Implementation of the Labour Force Survey in Egypt 24

28 Rec.7. Rec.8. Test other approaches to update the master sampling file more frequently, even if only for some turbulent regions, with high internal or international migration/movements. a. The technical unit responsible for sampling should use the feedback obtained from the production units to make amendments in the PSU s, on regular basis. b. Make more use of the information provided in the sampling frame for sampling, as well as analysis of non-response and non-response adjustment. c. Explore the possibility of using other administrative registers to update the frame (e.g. water and power suppliers database). Improvements in sampling design. a. Spread the number of survey weeks more evenly during the all the quarters of the year. CAPMAS should aim to cover all the weeks of the year. If this is not possible in the short run, then CAPMAS should strive towards covering at least a sixth week during the last month of the every quarter, so that all the survey months are equally represented. b. Take accurate stock of, and possibly measure, all advantages and disadvantages of panel design. Try to identify strengths of the former panel structure compared to the new panel structure. In this regard, reviewers strongly recommend that CAPMAS should re-consider the old panel structure, and possibly, enhance it to a panel structure 2-(2)-2 recommended by Eurostat. c. Explore the potential of the rotation scheme, in combination with mixed mode CAPI- CATI data collection, to improve efficiency and reduce costs and burden associated with data collection. d. Explore the potential of dependent interviewing, whereby respondents interviewed for the second or third time, are asked to confirm information that they already provided during the previous interview. This interviewing method should be limited only to those variables which are generally very stable over time (e.g. sex, date of birth, nationality, etc.) And never applied to those variables needed for classification of ILO status. The benefits of dependent interviewing can be fully reaped if applied in combination with an efficient panel structure. e. Try to find feasible alternatives to increase response rates, studying the effect on the estimates of over-sampling and substitution. Rec.9. Avoid substitutions of non-respondents as this practice introduces biases in LFS estimates which can be hardly catered for. Rec.10. Consider the use of effective automated methods for validity checks and probabilistic models for imputation against item non-response. Rec.11. Improvements in weighting procedures are required. a. Consider re-designing the weighting methodology, taking into account the sampling design, using calibration techniques instead of post-stratification, and trying to compute household weights. b. Add age groups in the weighting procedure, 5-years age-groups at national level and wider age-groups at regional or governorate level. c. There is strong need for more information on under-coverage and over-coverage in the sampling frame, so that they can be counteracted effectively during the weighting stage. Sector Review of the Implementation of the Labour Force Survey in Egypt 25

29 d. Exploit more the information available in the MS file for weighting purposes and try to make use only of variables which remain unchanged over time (e.g. year of birth, governorate). e. Add an additional step in the weighting procedure to adjust for the non-response and panel attrition verifying if these are correlated to individual characteristics only or any household characteristics. f. Ensure full consistency in the LFS main variables between annual estimates and quarterly estimates, and if possible, avoid computing annual weights based on a methodology which ignores the quarterly weights. Rec.12. There is a need for more detailed quality-related information on fieldwork in order to improve management during data collection. a. Analysis of visit time is important to reduce the number of attempted contacts, thus making fieldwork more effective (e.g. identify the time periods during which people are more likely to be at home). b. Interviewers should be able to distinguish between different types of failed surveys, e.g. ineligible, non-contacts, refusals, etc. This information is indispensable for proper computation of non-response adjustments. Data collection units should be able to compute more refined rates on over-coverage, non-contact, refusals, etc. to monitor fieldwork, before substitution. c. Find ways to measure non-sampling errors caused by interviewers and respondents, in particular in case of proxy respondents. d. Exploit analytical capabilities of fieldwork indicators, even using GIS applications, to identify limitations and increase efficiency in data collection. Rec.13. Explore the possibility to implement, directly on the CAPI software, the electronic tool already used for coding (e.g. databases containing dictionary of descriptions of jobs or economic sectors, associated with the exact codes, better if containing old and new classifications) to allow interviewers to code directly ISIC/NACE and ISCO. a. The full descriptions given by the respondents should be always registered in the micro-data to verify the quality of the coding carried out by the interviewers. Rec.14. CAPMAS is planning to pass to a CAPI data collection system using tablets. The following are some of the actions that CAPMAS should implement during the design stage of this system. a. The layout of the LFS questionnaire should be modified in CAPI form. b. An efficient IT platform should be identified for the design of the CAPI software which permits future enhancements of the questionnaire in the future. c. Carrying out pilot surveys or simulation of interviews using CAPI. d. Intensive testing of the CAPI system should be carried out before launch of CAPI survey. These tests include: testing of in-built validations for identification of hard and soft errors; testing file transfer from client and server; and, carry out tests against possible loss of data. e. CAPMAS should carry out study visits, or ask for assistance, from countries where CAPI technique is proved to be well implemented and effective. Sector Review of the Implementation of the Labour Force Survey in Egypt 26

30 f. The CAPI program should have a series of in-built validations to reduce errors during data collection. g. Editing mechanism must become more efficient with the introduction of CAPI, although the principles of checking and editing should not change. h. Thorough checking is required on the CAPI program before implementation stage, thus there is urgent need to acquire more technical competence to detect and solve data inconsistencies directly on a centralized database. i. A more suitable system for a continuous and automated monitoring of fieldwork and interviewers is needed Recommendations on Part 3 - Statistical Outputs Rec.15. Metadata are available to generic users with sufficient level of detail. a. More detailed information is needed for advanced internal and international users (e.g. mapping between the National Education Level Classification collected from the questionnaire and the recoded ISCED International classification). b. More documentation should be made available in English for international users. Rec.16. A better explanation of some aspects of the LFS is required for the SDDS metadata system. In some cases, there are inaccuracies in the information that has been provided. Rec.17. Increase the information about quality of statistics which is available to the users. Rec.18. Pay attention to a possible lack of consistency when producing estimates by household characteristics, or by relation to the Head of the Household. These quality limitations are expected since the weights are computed at individual level only and not at household level. Rec.19. Do not publish anonymized micro-data files to the general public on the website, if these data have not been fully anonymized, using appropriate software. Should any of the information provided to users be disclosed, the credibility of CAPMAS among its users could be seriously compromised. a. CAPMAS could work on a standard contract which stipulates clearly the conditions under which micro-data can be shared (using as reference the Eurostat guidelines _for_microdata_access_1.pdf ). Rec.20. Make arrangements to revise data backwards when major breaks occur. Rec.21. Take into account time-related inconsistencies when introducing modifications, and estimate the effect of these changes on the consistency of results. Rec.22. Explore the possibility of time series analysis on aggregated data, which also comprises the production of: a. monthly estimates of employment and unemployment; b. seasonally Adjusted time series; and, c. possible forecasts on employment and unemployment. Sector Review of the Implementation of the Labour Force Survey in Egypt 27

31 Rec.23. CAPMAS should carry out regular coherence tests between estimates of LFS and other national sources (e.g. National Accounts and Business Statistics) so as to identify possible limitations in the survey, and explain differences to the users. Rec.24. Dissemination of figures, bulletins and historical data, should be free on the website, and subscription should be asked only on a voluntary basis. Sector Review of the Implementation of the Labour Force Survey in Egypt 28

32 4. ANNEXES Sector Review of the Implementation of the Labour Force Survey in Egypt 29

33 ANNEX 1 Self-assessment questionnaire

34 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Sector Review - Labour Force Survey Egypt Self-assessment questionnaire This questionnaire has been drawn to assess the compliance of the Labour Force Survey (LFS) carried on by the Egyptian National Statistical Office (CAPMAS) with the principles of the European Statistics Code of Practice for the national and community statistical authorities and with the Regulations on the organisation of a labour force sample survey in the European Community. The self-assessment questionnaire for LFS is based on the ESS Quality Assurance Framework 2012, and most of questions have been selected from the Eurostat s document Joint Standard Quality Report for Labour Force Survey and Regional Labour Market Statistics and National Action Plan for the implementations of the recommendations of the task force on quality of LFS. It is divided in four parts; part 1 to 3 are part of the assessment as such, while part 4 is suggested for orientation of future action aimed at improving the general quality of the survey and its outputs. PART 1 - INSTITUTIONAL ENVIRONMENT In this part Egypt Labour Force sector/division is assessed against institutional and organisational factors that have a significant influence on the effectiveness and creditability of a statistical authority developing, producing and disseminating LFS Statistics. The relevant issues are mandate for data collection, adequacy of resources, quality commitment, impartiality and objectivity. PART 2 - STATISTICAL PROCESSES In this part Egypt Labour Force statistical processes are assessed against the European and other international standards, guidelines and good practices to organise, collect, process and disseminate LFS statistics. The relevant aspects are sound methodology, appropriate statistical procedures, non-excessive burden on respondents and cost effectiveness. PART 3 - STATISTICAL OUTPUTS In this part Egypt Labour Force statistical outputs are assessed against the ESS Quality Framework criteria, relevance, accuracy, reliability, timeliness, punctuality, coherency, comparability across regions, and accessibility by users and clarity. It aims at verifying consistency with EU standards and comparability of national outputs with those from European Countries. PART 4 - RECOMMENDATIONS OF THE TASK FORCE ON THE QUALITY OF THE LABOUR FORCE SURVEY This part could be used (is CAPMAS own decision) to assess against the recommendations contained in the Final Report of the Task Force on the quality of the Labour Force Survey which was set up by the LAMAS Working Group at Eurostat in March The goal of the Task Force was to review the quality of the LFS along the dimensions of the quality framework for statistical output of the European Statistical System (ESS), detect weaknesses and recommend improvements. The focus of the review was on the estimates of employment and unemployment, as these are the most relevant and largely used indicators produced by the LFS. CAPMAS could use this questionnaire for orientation, in order to indicate some directions for further work, or to take future actions aimed at improving the general data quality of LFS. Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 1

35 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Part 1 - Institutional environment Principle 2 Mandate for data collection 1.1 Is the mandate to collect information for the production and dissemination of official Labour Force statistics specified in law? Please specify. 1.2 Is CAPMAS allowed to use administrative sources for Labour Force statistics by legislation or by other forms of agreement? Please specify. 1.3 Does CAPMAS use administrative data sources for the production of Labour Force statistics? If yes, please specify, which source for which indicator/variable. 1.4 Does the legal act allow CAPMAS to penalise units that do not respond to LFS? Is there a system of sanctions in place? If yes, how often do you practice it? 1.5 What actions you would take to strengthen your mandate for LFS data collection? Principle 3 Adequacy of resources 1.6 To what extent are your LFS resources adequate, both in magnitude and quality, to meet current statistical needs as regards to the production of Labour Force statistics? Please indicate on a scale from 1 = completely adequate to 5 = completely inadequate. Magnitude of staff resources:... Quality of staff resources:... Magnitude of financial resources:... Magnitude of computing resources:... Quality of computing resources: If you think about future challenges that CAPMAS will have to face in order improve the quality of the Labour Force statistics, how much do you think that your current resources are able to meet these new requirements? Please indicate on a scale from 1 = completely adequate to 5 = completely inadequate. Magnitude of staff resources:... Quality of staff resources:... Magnitude of financial resources:... Magnitude of computing resources:... Quality of computing resources:... Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 2

36 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Please provide the data for the last three years: a) Total number of full time staff (excluding interviewers), working on Labour Force Survey data collection, compilation and dissemination; of which total number of full time IT staff; b) Total annual budget (including administrative expenses) on Labour Force Survey; of which IT expenditures; c) Total annual budget (including administrative expenses) on Labour Force Survey per 100,000 population and in percentage of GDP. 1.9 How would you assess the scope and detail of your Labour Force statistics as compared to user needs (too low/adequate/excessive). Please provide any available user feedback on Labour Force survey Do you have an external or internal training programmes aimed at improving the expertise of staff working on Labour Force statistics? Please explain your training activities. Principle 6 Impartiality and Objectivity 1.11 Are choices of sources and methods used in the compilation of Labour Force statistics only informed by statistical considerations? If they are published, please provide a copy Are errors discovered in published LFS statistics corrected at the earliest opportunity? 1.13 Do you have a published revisions policy for LFS statistics? If so, please provide a copy Do you provide notice in advance on major revisions and changes in methodologies in LFS? Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 3

37 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Part 2 - Statistical Processes In this part of the questionnaire, the Labour Force Survey statistical processes are assessed against international standards, guidelines and good practices adopted within the ESS. It aims at verifying consistency with EU standards when it comes to the ability of the Statistical Office to apply adequate statistical methodology and statistical procedures, whilst maximising cost effectiveness and burden reduction. Principle 7 - Sound methodology This section assesses whether the overall methodological framework used for the Egypt Labour Force Survey statistics follows European and other international standards, guidelines, good practices, concepts, definitions and classifications. Moreover it assess the existence of adequate expertise, continuous vocational training of the staff and good co-operation with the scientific community to improve methodology and tools. 2.1 Do your LFS data comply with the EU definition of unemployment as described in Annex 1 of Commission Regulation (EC) No 1897/2000? If not, please explain divergences. 2.2 Is the order of the sequence of questions addressing the ILO variables on unemployment is in line with requirements highlighted by implementing regulation EC 1897/2000? 2.3 Does the LFS questionnaire complies with the twelve principles in Annex 2 of Commission Regulation (EC) No 1897/2000 recommended 2000 in order to ensure that the rate of unemployment is estimated in a harmonised manner? If not, please explain divergence. 2.4 Is the national classification of industry consistent with the European classification systems NACE Rev 2? If not, please say what industrial classification is used and if it would be possible to convert the collected data into the NACE Rev Is the national classification of occupation consistent with International Standard Classification of Occupations (ISCO-08)? If not, please say what classification is used and if it would be possible to convert the collected data into the ISCO Is the national classification of educational level consistent with ISCED classification? If not, please say what classification is used and if it would be possible to convert the collected data into the ISCED. 2.7 Would you be able to accurately re-code your national variables into the EU list of variables laid down in Commission Regulation (EC) No 377/2008 of 25 April 2008? If not, please indicate which variables cannot be recoded and explain the divergence. 2.8 Is there in place a methodological/technical infrastructure (e.g. units, nets, committees) which defines, monitors the implementation and validates the results of the following processes: - statistical methods for sampling - statistica methods for data collection, - statistical methods for data editing and imputation - statistical methods for weighting/calibration - statistical methods for codification Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 4

38 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Are the Staff of the statistical authority recruited openly and with appropriate qualifications from relevant disciplines Is there an adequate structure that ensure continuous vocational training of the staff involved in LFS? 2.11 Do your Staff attend courses, seminars and workshops at a national or international level on the application of standards, classifications, good practices, etc.? 2.12 Do you make public any methodological documentation containing all pertinent information on metadata, namely concepts, methods, classifications of the Labour Force Survey process? 2.13 Is there a regular contact, e.g. through conferences, workshops, task forces, with the scientific community to discuss methodological, IT and innovation developments How often do you cooperate with the scientific community in order to improve the methodology by which you compile your labour market (mainly LFS) statistics? 2.15 How often do you introduce new measures to improve the quality of the data that is produced from the LFS? 2.16 Please describe the sampling frame used for the purpose of LFS (e.g. household register, etc). Describe also the sources used to update this database; indicate how often and how this database is updated in order to ensure high quality of samples Is there any quality problem in your frame, especially in coverage? Is yes, do you have any plan to improve it? 2.18 Eligibility criteria: Already implemented No, but will soon be implemented. Kindly indicate the year when this measure will be implemented. Not yet planned Use the UNECE-Eurostat recommended definition for usual residence Full coverage of all persons living in conventional households Full coverage of all persons living in conventional households Optional coverage of institutional households using a separate data collection mechanism 2.19 Are there some inclusion/exclusion criteria for members of the household? Are these criteria in line with EU- LFS regulation s requirements? If no, then explain the divergence Participation is voluntary/compulsory for all household members? If participation is obligatory by law, then provide the reference and a copy of the Legal Act. Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 5

39 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Sampling issues: LFS is carried out on a continuous basis In case of continuous survey, then the reference weeks are spread evenly throughout the year The reference week is defined in accordance to Eurostat regulations Full geographical coverage The sampling frame used for the scope of this survey is updated regularly in order to minimise over/under-coverage problems Do you ensure consistency between LFS sub-samples and annual sample as required by Implementing Regulation 430/2005? Is there overlap between consecutive surveys (panel) Already implemented No, but will soon be implemented. Kindly indicate the year when this measure will be implemented. Not yet planned 2.22 Data collection issues: LFS data is collected through personal interviews, although follow-up calls can be collected by telephone surveys A number of built-in validations are implemented into data collection software tool in order to minimise errors during data collection Do you use administrative data in order to collect at least part of the required information? Already implemented No, but will be implemented. Kindly indicate the year when this measure will be implemented. Not yet planned 2.23 Do you compute standard errors for LFS estimates? Please provide an overview of the methodology used for the compilation of sampling errors. Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 6

40 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Principle 8 Appropriate Statistical Procedures This section assesses whether appropriate statistical procedures, from data collection to data validation, are implemented in Labour Force Survey 2.24 Do you integrate LFS with administrative data? If yes, please specify which administrative sources are used and specify which LFS variables are collected from these registers. Please explain the method used for linking this administrative data with the data collected from surveys Where administrative data are used, are they a good approximation to what is required for statistical purposes? Please, explain if differences between administrative and statistical processes exist in terms of definitions, concepts, coverage, etc Is there in CAPMAS an appropriate organizational structure which provides guidelines, recommends appropriate methodologies and periodically examines the methods used in the statistical processes? 2.27 Is there in CAPMAS an appropriate group of experts which are in charge to code ISCO and NACE? Is the coding made directly by the interviewers? 2.28 Is your survey questionnaire assessed and validated by experts in the statistical domain and in questionnaire design? 2.29 Is your survey questionnaire tested prior to data collection with pilot in real situation, in depth - interviews, focus groups, interviewer support, etc.? Describe the testing procedures 2.30 Do you use electronic questionnaires with an optimized mix of PAPI, CAPI, CATI, CAWI techniques to exploit their advantages (e.g. to reduce cost, burden, non-response and non-sampling errors; to improve quality, accuracy and timeliness)? Describe the methodology used for data collection at CAPMAS Do you use dependent interviewing, asking the respondent to confirm some questions given in a previous wave? 2.32 Questions relating to employment status are put to all persons aged Are proxy interviews allowed during the data collection? What measures are taken (during field work and data analysis) in order to minimise proxy interviews and reduce biases associated with proxy surveys? 2.34 Are respondents provided with all necessary documents prior and during the data collection (i.e. letters, questionnaires, leaflets). Are these documents reviewed regularly? 2.35 Are data collection techniques monitored on a regular basis? Describe the monitoring processes used by CAPMAS Are interviewers provided with an LFS manual/handbook to aid them during data collection? 2.37 Are there regular training courses to be attended by interviewers? Provide a brief overview of the material taught during these courses Are respondents given support with filling in the questionnaires (help on-line, free toll number, support from interviewers or statisticians)? 2.39 Do you have procedures in place to answer to respondents' requests and complaints? Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 7

41 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Are follow-up procedures in place and implemented in the case of non-response and unreachable units? 2.41 Are automatic coding methods periodically reviewed and revised if necessary? 2.42 Are quality indicators related to data collection and coding regularly produced and analysed? 2.43 Describe the procedures used by CAPMAS in order to validate information collected by interviewers prior to data entry 2.44 Are the methods used for survey sampling, sample selections and estimation recommended by an appropriate organizational structure? Are they developed according to standard methods? Are they periodically renewed? 2.45 Please give details about the following aspects of the survey design? a) Sampling design b) Base used for the sample (sampling frame) c) Last update of the sampling frame d) Primary sampling unit (PSU) e) Final sampling unit (FSU) f) First (and intermediate) stage sampling method g) Final stage sampling method h) Overall theoretical yearly sampling rate (i.e. including non-response) i) Size of the theoretical yearly sample (i.e. including non-response) j) Stratification k) Rotation scheme 2.46 Do you calculate household and individual weights to produce estimates pertaining to the total population? 2.47 Gender is used in weighting (Y/N) 2.48 Nationality is used in weighting (Y/N) 2.49 Which age-groups are used in the weighting (e.g., 0-14, 15-19,..., 70-74, 75+)? 2.50 Which regional breakdown is used in the weighting? 2.51 Do you use monthly constraints in the weighting? 2.52 Please give details about the following aspects of the weighting process? Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 8

42 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no a) method of calculating the base weights b) method of calculating the final weights (e.g. post-stratification, weighing, calibration, etc.) c) adjustments for total non-response of individuals and/or households d) adjustments for wave non-response or panel attrition e) adjustment for rotation groups f) other weighting dimensions 2.53 Are the methods used for survey sampling, sample selections and estimation reported to public? 2.54 Do the methods used for estimation, including the correction of non-response, data calibration and seasonal adjustment follow transparent methodological rules? 2.55 Are appropriate editing and imputation methods used and regularly reviewed, revised or updated as required? 2.56 Is there in CAPMAS an appropriate organizational structure which provides guidelines, recommends appropriate methodologies and periodically examines editing and imputation methods? 2.57 Is the analysis of the effect of editing and imputation regularly undertaken as part of assessing quality of the data collection? 2.58 Describe the methodology used for the carrying out the following tasks for the purpose of LFS: a) Cross-sectional and longitudinal checks on data b) Imputation of missing information c) Linking of dataset with administrative information 2.59 Describe the main statistical procedures used by CAPMAS in order to minimise non-sampling errors: a) non-response bias; b) under-reporting; c) under-coverage problems; d) item non-response; e) attrition; 2.60 Does CAPMAS carry out regular revisions on past data? If yes, explain why revisions on past data are generally required Do revisions follow standard guidelines and principles, well-established procedures, and are transparent to users? Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 9

43 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Are revisions accompanied by all necessary explanations and made available simultaneously to all users? 2.63 Does LFS make use of subsamples to survey structural variables (wave approach)? 2.64 Does LFS make use of a subsample for yearly variables? (Y/N) 2.65 If yes, are the 30 totals for ILO labour status (employment, unemployment and inactivity) by sex (males and females) and age groups (15-24, 25-34, 35-44, 45-54, 55+) between the annual average of quarterly estimates and the yearly estimates from the subsample all consistent? (Ref.: Commission Reg. 430/2005, Annex I) 2.66 If not, please list deviations 2.67 List of yearly variables for which the wave approach is used (Ref.: Commission Reg. 377/2008, Annex II) Principle 9 - Non-excessive burden on respondents This section assess if the reporting burden is proportionate to the needs of the users and is not excessive for respondents, and targets for its reduction over time are set by statistical authorities What procedures are already in place at CAPMAS in order to ensure that data collected from surveys are indeed required by users? How much of the data collected from the LFS is indeed published or used for policy making purposes? 2.69 What measures has CAPMAS taken during the past years in order to minimise response burden on respondents as much as possible? 2.70 Is the response burden measured periodically? (The reporting burden is calculated for the time needed: to answer the questionnaire, to retrieve the required information, etc.) 2.71 Questions used to collect information which will not be published are limited and justified.? 2.72 Is the reporting burden spread as widely as possible over survey populations? How is this done in Egypt? 2.73 Duration of the interview (Note: This table should only show the burden on the respondents. Not time spent in the field to contact the household or fill out administrative forms.) Average time spent in the household during PAPI or CAPI interview Average time spent with the household of the telephone during a CATI interview Core questionnaire (pr person) Ad hoc questionnaire (pr person) 2.74 Average number of units Households visited over the year Persons interviewed over the year Persons interviewed for the ad hoc module over the year Minutes Total First wave Later waves Number Total First wave Later waves Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 10

44 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no To minimize data collection there is explicit consideration of alternative data sources, including existing survey and administrative data? 2.76 Administrative sources are used whenever possible to avoid duplicating requests for information? 2.77 Legal access to the administrative sources is granted and the administrative authorities have the obligation to provide the administrative data if requested? 2.78 Guidance on the availability and quality of administrative sources is available to survey managers. Describe the procedure how this information is made available to the survey managers. Principle 10 Cost Effectiveness This session assess whether statistical authorities uses the potential of information and communications technology is being optimised, and implement standardized solutions that increase effectiveness and efficiency of the resources Does CAPMAS have an internal or external mechanism in order to monitor the use of resources for the production of LFS? 2.80 What measures have been taken in recent years in order to improve efficiency of data collection and data analysis? 2.81 Does CAPMAS use automated tools in order to reduce manual interventions with consequent improvement in timeliness and cost-reduction? 2.82 Are administrative data used where possible? 2.83 Do policies, procedures and tools exist to promote automatic techniques for data capture, data coding and validation Number of staff involved in central and regional offices, ( in Full-time equivalent; consider only staff directly employed by the NSI, excluding interviewers) Full-time equivalents Total - of which statisticians - of which other professional and managerial staff 2.85 Interviewers Interviewers Number of interviews per interviewer (weekly average) Cost of each interview PAPI CAPI CATI 2.86 What is the total cost of the survey in the last year? Is it increasing or decreasing? Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 11

45 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Part 3 - Statistical Outputs In this part Labour Force statistical outputs are assessed against the ESS Quality Framework criteria, relevance, accuracy, reliability, timeliness, punctuality, coherency, comparability across regions, accessibility by users and clarity. It aims at verifying consistency with EU standards and comparability of national outputs with those from European Countries. Principle 11 Relevance This section assesses whether labour force statistics meet and serve the existing needs of governments, research institutions, business concerns and the public generally; if labour force statistics consider emerging needs and priorities; whether user satisfaction is monitored on a regular basis and is systematically followed up. 3.1 Is a Labour Force Survey release calendar for quarterly and annual data made available well in advance to users on the website? 3.2 Are quarterly and annual indicators available on your national online database? 3.3 Is the set of your indicators consistent with the Eurostat LFS main indicators available on: (If not, please specify what indicators are not currently available. Please set out any plans to start producing data for the missing indicator) 3.4 Are not seasonally adjusted quarterly time series available? 3.5 Are seasonally adjusted quarterly time series available? 3.6 Please specify relevance of the main LFS outputs For policy makers (a) For social actors (b) For the media (c) For researchers and students (d) For enterprises (e) Total Employment (levels) Total Employment rate Part-time employment Temporary employment Employment by socio-demographic breakdown (age, sex, education) Employment by governorate breakdown Employment by governorate and urban/rural breakdown Employment by economic activity (NACE) breakdown Hours actually worked Total Unemployment (levels) Total Unemployment rate Long-term unemployment Youth unemployment Unemployment by socio-demographic breakdown (age, sex, education) Unemployment by governorate breakdown Unemployment by governorate and urban/rural breakdown (a) i.e. if the statistic is used as policy target or if it is used in official studies influencing policy making Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 12

46 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no (b) i.e. if the statistic is used by employers associations, trade unions, lobbies, at national or regional level, for their decision making (c) i.e. if the statistic is mentioned in national or regional media, specialised or for general public (d) i.e. if the statistic is used by academics for scientific research (e.g. appears in scientific publications) or by students (e) i.e. if the statistic is used by enterprises for own market research activities or for consultancy services in the information sector 3.7 What is the periodicity of your results? monthly quarterly annual Periodicity of the LFS statistics at national level Periodicity of the LFS statistics at governorates level Periodicity of the LFS statistics at governorates*urban/rural level Principle 12 Accuracy and Reliability This section assesses whether labour force statistics accurately and reliably portray reality, if results and outputs are regularly assessed and validated, if sampling errors and non-sampling errors are measured and systematically documented according to the European standards; if revisions are regularly analysed in order to improve statistical processes. 3.8 Are source data, intermediate results and outputs regularly assessed and validated? If yes, please explain what is done. 3.9 Is the coefficient of variation (CV) for quarterly and annual estimates computed? If yes, please explain if design effect is taken into account, and provide references about the method and the software used Please complete the following table of Coefficient of variation (CV) for estimates at national level if the information about CV is available CV of national aggregates (in %) Quarter Number of employed Number of parttime employed Number of unemployed Rate of unemployment Average number of hours actually worked per week Annual 3.11 Please complete the following table of Coefficient of variation (CV) for estimates at governorate level. Only if the corresponding estimate is made available to users and if the information about CVs is available. CV of regional annual aggregates (in %) Governorate Please Add rows Number of employed Number of parttime employed Number of unemployed Rate of unemployment Average number of hours actually worked per week Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 13

47 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Give quantitative information and brief comments on the main problems affecting frame quality. Mention specifically which regions / population groups are not or badly represented in the sample. Under-coverage Rates Description of over-coverage, under-coverage and classification errors and their impact on estimates Over-coverage Classification errors of the statistical units* Reference on frame errors *Misclassification refers to statistical units having an erroneous classification where both the wrong and the correct one are within the target population Do you measure or assess the measurement errors due to the respondents? If yes, please mention the methods you have used or intend to use to reduce such errors Do you measure or assess the measurement errors due to the interviewers? If yes, please mention the methods you have used or intend to use to reduce such errors Do you measure or assess the measurement errors due to questionnaire? If yes, please mention the methods you have used or intend to use to reduce such errors Do you measure or assess the measurement errors due to survey mode? If yes, please mention the methods you have used or intend to use to reduce such errors Please provide comments on methods you have used or intend to use to reduce other possible measurement errors When do you carried on the last pilot survey in order to test the questionnaire? 3.19 Is there a report from cognitive laboratory available? If yes, give details Is there any methodological notes on the measurement errors? 3.21 Is information available about processing errors and the error rates? Please, complete the table below. Is information available? About data capture errors which occur when information on a questionnaire is converted to a computer format About data codification errors Yes / No / Not Avaiable Error rate in % Comments About errors which occur during the editing phase About other processing errors ((for example due to the use of computers, electronic questionnaires, wrong files e) Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 14

48 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Is information available on non-response rates by survey mode and by categories? If yes, please complete the tables below and provide comments if these rates are increasing over the last 5 years or have high variability across quarters. Non-response rates by survey mode and by categories (%) Survey mode Annual average 2013 Annual average 2012 Annual average 2011 Annual average 2010 Annual average 2009 TOTAL by Survey Mode PAPI POSTAL CAPI CATI CAWI Refusals By Non-Response Categories Other reasons Survey mode 2013 Quarte Quarte Quarte Quarte4 TOTAL Non-response rates by survey mode and by categories (%) by Survey Mode PAPI POSTAL CAPI CATI CAWI Refusals By Non-Response Categories Noncontacts Noncontacts Other reasons 3.22 Is information available on non-response rates by survey wave? Please, complete the table below and provide comments if these rates have high variability across waves and years. Annual average 2013 Annual average 2012 Annual average 2011 Annual average 2010 Annual average 2009 Non-response rates by survey waves (%) Is information available on non-response rates by regions? Please, complete the table below and provide comments if these rates have high variability across regions. Regions (Governorates) Please Add rows Non response rate (%) 3.24 Are the non-response rates provided in the above tables weighted? If weighted, please state the definition of the weights Is the non-response calculated at household level or person level? Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 15

49 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Is there an assessment of the underestimation and overestimation bias of main characteristics due to nonresponse? If yes, please complete the table below. Total employment Part-time employment Unemployment Numbers of hours actually worked Other characteristic Other characteristic Other characteristic Underestimation assessment Overestimation assessment Quantitative Descriptive Quantitative Descriptive 3.27 Is there any method you use or intend to use in the future to adjust for unit non-response (e.g. post-stratification by ad hoc auxiliary information for non-response)? If yes, please provide information about the method and provide details about the auxiliary variables used Is substitution of non-responding units allowed? If yes, please provide details about the criteria and substitution rate by wave Do you use any other method (e.g. results of non-response surveys, copying information from previous quarters, etc.) to adjust for unit non-response? If yes, please provide details Do you use statistical methods for imputation of item non-response? If yes, please provide references to methodological notes, explain if these information are made available to users and how; answer next question Please provide the imputation rate and describe the method used for imputation of statistical item non-response in the following table. The imputation rate expresses the percentage of records in which at least one variable has been imputed or when the original values, missing or not, have been modified) Question/Variable/Characteristic Full questionnaire Add most relevant Question/Variable/Characteristic Imputation rate (%) Describe method used, mentioning which auxiliary information or stratification is used 3.32 Are revisions regularly analysed to improve statistical processes? If yes, please explain what is done Is the quality of the LFS, in all its dimensions, systematically documented and made available to users? If yes, please explain what is done. Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 16

50 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Principle 13 Timeliness and Punctuality This section assesses whether labour force statistics are released to public in a timely and punctual manner, in a specific day, at a standard daily time; if the periodicity of statistics takes into account user requirements as much as possible; if divergence from the dissemination time schedule is publicised in advance Please complete the table below with the reference period and dissemination dates Quarterly and Annual result for the year Annual average Date of data collection beginning Date of end of data collection Main dates in the national production process Date of end of quality check and weighting for statistics published in press release Date of national publication of press release Date of national dissemination of microdata and metadata Reasons for late delivery if this has occurred 3.35 If data are not currently available in twelve weeks after the end of the reference period, are there any plans to accelerate production to meet this deadline? Please explain ways for improving timeliness Is the dissemination of LFS estimates scheduled in only one day or in more days (e.g. preliminary results of acceptable aggregate accuracy are released earlier than the final results, or when considered useful)? 3.37 In case of delay in the survey process, is the divergence from the dissemination time schedule publicised in advance, explained and a new release date set. Principle 14 - Coherence and Comparability This section assesses whether labour force statistics are internally coherent and consistent between quarterly and annual (i.e. arithmetic and accounting identities observed), are comparable over a reasonable period of time and between regions. Whether statistics from different surveys and sources are comparable and are compiled on the basis of common standards, definitions and classifications. Whether comparability of national data is possible with other countries of the European Statistical System Please specify the year from which LFS data are available. Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 17

51 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Please enumerate and describe in the table below all the changes at the concept level that would affect comparability across years? Add rows if needed. Changes in Description of the impact of the changes on the statistics. Estimation of effect for aggregates (%) Estimation of adjustment F/P for aggregate Are statistics revised backwards (Y/N) If statistics are revised, give brief comment on the method of revision Concepts and definition Coverage (i.e. target population) Legislation Classifications Geographical boundaries Other concept 3.40 and describe in the table below all the changes at the measurement level (for example changes in data collection, weighting scheme, new design, use of auxiliary information) that would affect comparability across years. Add rows if needed. Changes to Description of the impact of the changes on the statistics. Estimation of effect for aggregates (%) Estimation of adjustment F/P for aggregate Are statistics revised backwards (Y/N) If statistics are revised, give brief comment on the method of revision Sampling frame Sample design Rotation pattern Questionnaire Instruction to interviewers Survey mode Weighting scheme Use of auxiliary information Others Others Others 3.41 Please specify if annual data are consistent with quarterly data, and at which level of aggregation Are household estimates produced and disseminated? If yes, are these consistent with individual estimates by household type? (For example, the number of employed persons which are single is equal to the number of one-personhouseholds in which the person is employed) Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 18

52 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Is there any divergence of national concepts from European concepts. Please list all concepts where any divergences can be found; add rows as necessary. Is there a divergence between the national and European concepts for the following characteristics? Definition of resident population (*) Identification of the main job (*) Employment Unemployment Others (Y/N) Give a description of difference and provide an assessment of the impact of the divergence on the statistics 3.44 Are national data comparable with European data? Please explain any other divergence with European Statistical System requirements (if not explained in other parts of the questionnaire) Is the coherence of LFS data with population statistics assured? If not, complete the table below Overall effect Men 25 years and over Men 25 years and over Women under 25 years Women 25 years and over Regional distribution Give an assessment of the effects of the differences 3.46 Are the LFS data coherent with National Accounts data? Give details in the table below Description of difference in concept Description of difference in measurement Give an assessment of the effects of the differences Give references to description of differences Total employment Total employment by NACE Number of hours worked 3.47 Are the LFS data coherent with Business Statistics data? Give details in the table below Description of difference in concept Description of difference in measurement Give an assessment of the effects of the differences Give references to description of differences Total employment Total employment by NACE Number of hours worked 3.48 Are the LFS unemployment data coherent with registered unemployment? Give details about differences in concept, purpose, measurement; provide references if exist. Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 19

53 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Is possible to have an assessment of the effect of differences of LFS unemployment and registered unemployment? Please, complete the table below. Overall effect Men under 25 years Men 25 years and over Women under 25 years Women 25 years and over Regional distribution Give an assessment of the effects of the differences Principle 15 Accessibility and clarity This section assesses whether labour force statistics are presented in a clear and understandable form, released in a suitable and convenient manner, available and accessible on an impartial basis with supporting metadata and guidance Are LFS data dissemination using traditional information and communication means? 3.51 Are LFS data dissemination using modern information and communication web technology? 3.52 Are statistic and the corresponding metadata presented, and archived, in a form that facilitates proper interpretation and meaningful comparisons? 3.53 Are custom-designed analyses provided when feasible and the public informed? 3.54 Is the access to microdata allowed for research purposes and subject to specific rules or protocols? Provide details about conditions of access to data, means, support, marketing conditions, anonimisation, possible restrictions, existing service-level agreement, etc Are metadata documented and published according to a standardised metadata system? Please provide details, references and web link Are users kept informed about the methodology of statistical processes (including the use of administrative data)? Please provide details, references and web link Are users kept informed about the quality of statistical outputs with respect to the national quality criteria and to the international quality standards? Please provide details, references and web link Is assistance easily available to users by telephone and ? 3.59 Does a specific web page containing information on the national LFS exist? Is it translated in English? Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 20

54 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no Part 4 - Recommendations of the Task Force on the Quality of the Labour Force Survey In this part Egypt Labour Force Survey is assessed against the recommendations contained in the Final Report of the Task Force on the quality of the Labour Force Survey which was set up by the LAMAS Working Group at Eurostat in March The goal of the Task Force was to review the quality of the LFS along the dimensions of the quality framework for statistical output of the European Statistical System (ESS), detect weaknesses and recommend improvements. The focus of the review was on the estimates of employment and unemployment, as these are the most relevant and largely used indicators produced by the LFS. Following this review, the Task Force formulated forty-three recommendations on sampling design and sampling errors, weighting schemes, non-response; some of these recommendations are excluded from this questionnaire because are recommendations to Eurostat only. The questionnaire below has already been used by each National Statistical Institute in the European Statistical System as a self-assessment against these recommendations. CAPMAS could use the results to draw a national action plan for the coming years, in order to try to fulfill these recommendations. N. RECOMMENDATIONS R2 R3 R4 R5 R6 R7 Already fulfilled Yes/No DESCRIPTION OF THE PRESENT SITUATION Sampling design and sampling errors Achieve consistency between target population, sampling frame and population estimates Balance samples over geographical areas and reference weeks, in order to both improve the national quarterly and yearly estimates, and to enable the production of good European monthly estimates Regularly review the efficiency of the sample design and adapt it when necessary Fully apply the concept of resident population as defined by art. 2(d) of EP and Council Regulation (EC) No 763/2008 Use sample designs with intra-annual rotation patterns to allow calculation of quarter-to-quarter labour market flows estimates Examine options for further harmonization of rotational patterns to enhance comparable longitudinal analysis at European level DESCRIPTION OF PLANNED IMPROVEMENTS THAT WOULD MEET THE RECOMMENDATIONS (IF ANY) Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 21

55 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no N. RECOMMENDATIONS R8 R9 R10 Treat collective households separately from private households, for instance when applying weights. This is needed to ensure crosscountry comparability and produce meaningful EU estimates, as not all Member States cover collective households Apply weighting schemes which ensure the coherence between household and individual estimates Optimise weighting schemes to correct for non-response bias, taking into account specific characteristics of non-respondents Already fulfilled Yes/No DESCRIPTION OF THE PRESENT SITUATION Weighting schemes DESCRIPTION OF PLANNED IMPROVEMENTS THAT WOULD MEET THE RECOMMENDATIONS (IF ANY) R11 R12 R13 Regularly collect information on the characteristics of non-respondents (e.g. by retrieving structural data from the sampling frame or from low-level geographical databases that can be linked to it, or by conducting non-response surveys) to assess and adjust for non-response bias and to improve fieldwork strategies. Specific non-response surveys should be used primarily for improving fieldwork strategies, and care must be taken when using them to adjust the results of the LFS Make participation in the LFS compulsory, wherever legally feasible, in order to reduce nonresponse Make use of tools to inform or reassure respondents, such as presentation letters, free-toll numbers and visible identification cards for interviewers Non-response Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 22

56 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no N. RECOMMENDATIONS R14 R15 R16 R18 R19 Already fulfilled Yes/No DESCRIPTION OF THE PRESENT SITUATION Arrange for special tools to increase participation of non-nationals, such as translated questionnaires and presentation letters, and free-toll numbers for non-nationals where information is provided in other languages Consider implementing the wave approach, as it helps to reduce response burden Make use of dependent interviewing to reduce response burden and thus increase participation, but only for variables that tend to be stable from quarter to quarter. Dependent interviewing must not adversely impact on the measurement of the ILO labour status. In particular, the question on having worked for at least one hour during the reference week should be asked by default Interviewers and fieldwork organization Provide specific training to interviewers to avoid refusals by respondents and to limit noncontacts. In particular, contact attempts should be carried out at different times of the day and interviewing time should also cover periods after working hours (evenings and weekends), to reduce cases of non-response. Collecting preferred contact time during the first interview may facilitate contacts at further waves Make use of permanent professional interviewers. Minimise turnover of interviewers to ensure continuity. Interviewer remuneration should be adequate with respect to the interviewers' crucial role for the quality of the survey DESCRIPTION OF PLANNED IMPROVEMENTS THAT WOULD MEET THE RECOMMENDATIONS (IF ANY) Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 23

57 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no N. RECOMMENDATIONS R20 R21 R22 R23 R24 R25 Provide interviewers with appropriate training on how to conduct the interview, covering both communication skills and survey content. Training should be provided on a continuous basis Organise periodical focus groups and debriefing for interviewers to review and tackle critical issues Instruct the interviewers in general not to adapt questions to the respondent but just to read them out as they are written in the questionnaire, as personal interpretations may lead to biased results Continuously monitor and systematically assess the quality of the fieldwork. In the case of telephone interviewing, centralized fieldwork supervision should be used in a telephone facility. In the case of face-to-face interviewing, examples of suitable monitoring methods are regular quality surveys but also other methods like the systematic analysis of interview protocols and behaviour coding of audio recorded interviews Keep interviewing periods as close as possible to the reference period, to avoid recall problems and to support timely production of results. Exceptions may be made in particular periods such as holiday seasons, when shorter interviewing periods would result in a low response rate Stress, both in the presentation letter and from the interviewers side, that the information collected from individual respondents is treated as confidential and will be used for statistical purposes only Already fulfilled Yes/No DESCRIPTION OF THE PRESENT SITUATION DESCRIPTION OF PLANNED IMPROVEMENTS THAT WOULD MEET THE RECOMMENDATIONS (IF ANY) Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 24

58 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no N. RECOMMENDATIONS R26 R27 R28 R29 R30 R31 Already fulfilled Yes/No DESCRIPTION OF THE PRESENT SITUATION Survey modes and questionnaire Consider the use of mixed modes, in the light of possible gains in response rates, burden and costs versus the mode effects. The impact of selfadministered electronic data collection, including web-based modes, on the measurement of ILO labour status should be carefully investigated Carefully test any changes to modes, questionnaires and other explanatory survey material and assess their impact before introduction Always carry out the LFS by computer-assisted questionnaires, given that the traditional paper interviewing mode is no longer suitable to cope with the complexity of the survey Information for users Accompany published estimates by information about their accuracy, covering in principle both sampling and non-sampling errors Make available specific documents to explain the differences between LFS and National Accounts employment estimates and provide guidance to users on which source fits which purpose. The aforementioned documents should be easily accessible to domestic and international users Coherence Produce, at least annually, reconciliation tables between National Accounts and LFS estimates of employed persons to enable targeted improvement measures for the LFS and National Accounts and for communication to users. A suitable template should be developed DESCRIPTION OF PLANNED IMPROVEMENTS THAT WOULD MEET THE RECOMMENDATIONS (IF ANY) Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 25

59 Assessment of the statistical systems and selected statistical areas of the ENP countries Contract no N. RECOMMENDATIONS R32 R33 R35 R37 Foster close communication between LFS experts and National Accounts labour market experts, to enhance cooperation and increase the mutual understanding of the methods used in each domain and their respective strengths and weaknesses Conduct further research in the areas of marginal employment, employment in black economy activities, employment in private households, illegal immigrants, the influence of non-response and proxy interviews, as these were identified as areas with potential accuracy problems in the LFS that could result in inconsistency with National Accounts estimates Already fulfilled Yes/No DESCRIPTION OF THE PRESENT SITUATION Comparability of employment and unemployment statistics Adequately plan and monitor, at national level, all changes initiated either by Eurostat or by countries in order to assess the statistical effect on time series Produce and disseminate consistent time series at least for the headline employment and unemployment indicators DESCRIPTION OF PLANNED IMPROVEMENTS THAT WOULD MEET THE RECOMMENDATIONS (IF ANY) R40 R42 Improve significantly the timeliness of the EU-LFS in order to further enhance its relevance for short-term economic analysis Comply with the twelve-week deadline in the Regulation as the one for final, not first data transmission Timeliness and Punctuality Sector Review Labour Force Survey Egypt Self-assessment questionnaire (prepared by Antonio R. Discenza and Etienne Caruana) 26

60 ANNEX 2 SR Agenda

61 Sector Review of the Labour Force Survey Egypt Sector Review of Labour Force Survey CAPMAS Central Agency for Public Mobilization And Statistics Cairo - Egypt Agenda Draft, March 21 Date and place of the SR Date April 14-17, 2014 Place CAPMAS premises - Cairo - Egypt Assessment team GOPA-expert: Mr. Antonio R. Discenza (leading expert) Mr. Etienne Caruana Agenda SR Labour Force Survey CAPMAS 14 to 17 April

62 Sector Review of the Labour Force Survey Egypt Day 1: April 14 (Introductory meeting and SAQ part 1 Institutional Environment) 09:30 10:00 Welcome and introductory meeting with the top management of the Central Agency for Public Mobilization And Statistics 10:00 11:00 Meeting with managers of official LFS statistics Review of principles 2, 3 and 6 for LFS. Assessment Team Top Management LFS Management Assessment Team LFS Management 11:00 11:15 Coffee Break 11:15 13:00 Meeting with managers of official LFS statistics Cont.: Review of principles 2, 3 and 6 for LFS Assessment Team LFS Management 13:00 14:00 Lunch Break 14:00 15:30 Presentation of Egypt LFS by the LFS team Overview of organization, processes and outputs Assessment Team LFS Team 15:30 15:45 Coffee Break 15:45 17:00 Cont: Presentation of Egypt LFS Overview of organization, processes and outputs Assessment Team LFS Team Day 2: April 15 (SAQ part 2 - Statistical Processes) 09:00 10:15 Meeting with producers of official LFS statistics Principle 7 - Sound methodology Assessment Team LFS Team and Methodology Unit 10:15 10:30 Coffee Break 10:30 12:00 Meeting with producers of official LFS statistics Principle 8 - Appropriate Statistical Procedures Assessment Team LFS Team and Methodology Unit 12:00 13:00 Lunch Break 13:00 14:30 Meeting with producers of official LFS statistics Principle 9 - Non-excessive burden on respondents Assessment Team LFS Team and Methodology Unit 14:30 14:45 Coffee Break 14:45 16:30 Meeting with producers of official LFS statistics Principle 10 - Cost-effectiveness Assessment Team LFS Team Agenda SR Labour Force Survey CAPMAS 14 to 17 April

63 Sector Review of the Labour Force Survey Egypt Day 3: April 16 (SAQ part 3 - Statistical Outputs) 09:00 10:15 Meeting with producers of official LFS statistics and Communications Unit Assessment Team LFS Team 10:15 10:30 Coffee Break Principle 11 Relevance Principle 12 Accuracy and Reliability 10:30 12:00 Meeting with producers of official LFS statistics Principle 12 Accuracy and Reliability Assessment Team LFS Team 12:00 13:00 Lunch Break 13:00 14:30 Meeting with producers of official LFS statistics Principle 13 Timeliness and Punctuality Principle 14 - Coherence and Comparability Assessment Team LFS Team 14:30 14:45 Coffee Break 14:45 16:30 Meeting with producers of official LFS statistics Principle 14 - Coherence and Comparability Principle 15 Accessibility and clarity Assessment Team LFS Team Day 4: April 17 (SAQ part 4 - Statistical Outputs and Final meeting) 09:00 10:15 Meeting with management and producers of official LFS statistics. Action Plan of future actions 10:15 10:30 Coffee Break 10:30 12:00 Meeting with management and producers of official LFS statistics. Cont. Action Plan of future actions 12:00 13:00 Lunch Break 13:00 14:30 Meeting with Management and Senior Staff about a) medium and long-term planning within the sector; b) proposal of a list of actions to be undertaken in order to improve the data delivery and functioning of the sector. Assessment Team LFS Management LFS Team Assessment Team LFS Management LFS Team Assessment Team LFS Management Top Management Agenda SR Labour Force Survey CAPMAS 14 to 17 April

64 ANNEX 3 Long Questionnaire

65 Arab Republic of Egypt Central Agency for Public Mobilization & Statistics Labor Force Sample Survey Long questionnaire This survey data is confidential and used only for research purposes