Methodologies: Sources of information on labour statistics (focus on the economically active population)statistics (focus on the economically active population)
Monica D. CastilloChief Decent Work Data Production UnitChief, Decent Work Data Production UnitILO Department of Statistics – Geneva
National Labour Market Information Training Programme Port of Spain, Trinidad and Tobago
31 October – 11 November 2011
ILO Department of Statistics
Contents
• Overview of key official sources of labour statistics:– Advantages, disadvantages of each type
• Integrated labour statistics systemg y• Integrated national statistics system• National programme for statistics of the • National programme for statistics of the
economically active population (EAP)L b f k i t t f i • Labour force survey: key instrument for measuring the EAP
ILO Department of Statistics 2
Official sources of labour statistics• Population and housing census
• Household surveys (LFS, HIES, …)
• Economic census
• Establishment surveys (current and long-term needs)
Ad i i i d • Administrative records
• Agricultural census and surveys
ILO Department of Statistics 3
• Agricultural census and surveys
Official sources:Differences in coverageDifferences in coverage
Differences in geographic, worker/enterprise, job coverageP&H Census Universal coverage of civilian population living in
households
Household Surveys Sample of the population living in households; coverage depends on objective & sample design
Economic Census Census of establishments, coverage depends on objective and frame
Establishment Surveys Sample of establishments, coverage depends on objective, frame, sample design
Ad i i t ti S D fi d b l l d i i t ti bli ti
ILO Department of Statistics 4
Administrative Sources Defined by legal or administrative obligations: Registered/contributing population
Official sources:Differences in periodicityDifferences in periodicity
P&H Census Every 5-10 years
Recommended every 10 yearsRecommended every 10 years
Household Surveys Usually more frequent than censusesDepending on priorities and resources: For LFS could be Depending on priorities and resources: For LFS could be continuous, monthly, quarterly, or annually
Economic Census Every 5-10 years
R d d 5 10 Recommended every 5-10 years
Establishment Surveys Monthly, quarterly, annually
Recommendation depends on purpose/resources: monthly or quarterly for time series; annually or 3-5 years structural dataC ti
ILO Department of Statistics 5
Administrative Sources Continuous-updating as per administrative procedures
Population censusp
• Population census: is the official procedure of systematically acquiring d di i f ti b t th b f th l ti ( ll ) t and recording information about the members of the population, (normally) at
regular, long-term intervals (e.g. every 10 years)
Importance of collecting data on the economic characteristics of the population through a population census:– In some countries is the only source– Produces broad-based estimates on selected economic characteristics– Provides benchmark data on economic characteristics to which statistics
from other sources can be related – Complements the other sources of data on economic characteristics
Department of Statistics 6
Population and housing census:Strengths and LimitationsStrengths and Limitations
Advantages as data source for economic characteristics:• National scope and consistent geography• Inclusion of all persons
– Permits extensive and detailed cross-tabulationsProduce summary statistics for all civil or administrative divisions – Produce summary statistics for all civil or administrative divisions
– Generate finely detailed statistics for small geographical areas, small population groups, or finely classified groups of industries or occupations
• Permits cross-tabulation of economic characteristics with other demographic and socio-economic variables
Limitations as data source for economic characteristics:C d t d l 10 • Conducted only once every 10 years
• High cost and sometimes heavy reporting burden• Compete with demands for coverage of other topics; placement of questions on the questionnaire.• Some of the concepts related to economic characteristics are very complex for censuses to handle
A t i f ti t i h t i ti t k t d t (t l f HH )
ILO Department of Statistics 7
• Accurate information on certain characteristics not known to proxy respondent (true also of HH surveys)
Labour Force Surveys
Labour Force Surveys:• Surveys carried out using a sample of households• Surveys carried out using a sample of households• Constitute the main data collection instrument for statistics on employment and
unemployment worldwide P it th ll ti f h t i f ti th t t l ki l ti • Permit the collection of coherent information on the total working age population and its components, in particular the economically active population.
• Concept of employment in household surveys refers to persons employed rather th t j b i h l j b d k i diff t than to jobs, since a person may have several jobs and work in different establishments
• Large range of topics can be included. Examples:E l t l t t i l b f– Employment, unemployment, not in labour force
– Time-related underemployment – Informal employment– Occupational segregation by sex (also captured by establishment surveys)
ILO Department of Statistics 8
– Youth not in education and not in employment, 15-24 years– Hours actually worked
Labour force surveys:Strengths and LimitationsStrengths and Limitations
Advantages:– Flexible can cover a large set of labour topicsFlexible, can cover a large set of labour topics– Allows personal interviews– Detailed questioning permits precise measurement of concepts – If frequent, can provide data to monitor short-term labour force trendsq , p– Provides data on persons & their households– Can be designed to provide both stock and flows estimates– Covers all workers, including all self-employed– Allows disaggregations by demographic factors: by sex, age group, ethnic group
Limitations:– Sampling often prevents reliable estimates for small groups/areasp g p g p– Sometimes have limited geographic coverage (cost considerations)– Sampling and non-sampling errors– Reliability of estimates regarding data by industry or occupation; and
Still: is considered the best source for
sensitive information (e.g., employment-related income)
9
labour force statistics
ILO Department of Statistics
Establishment surveysyEmployment-based establishment surveys:
• Surveys based on a sample of establishments that seek to obtain statistical information on • Surveys based on a sample of establishments that seek to obtain statistical information on employment, hours and earnings.
• They may have a short-term (current statistics, trends) or longer term purpose (structural statistics, distribution of earnings, etc), g , )
• Employment measures the number of jobs held by persons working in establishments, not the number of employed persons
• Generally cover employees (i.e., not all workers are covered)• Examples of indicators and topics:
– Average hourly earnings in selected occupations– Manufacturing wage index– Employment by branch of economic activity*– Wage/earnings inequality*– Labour cost– Hours paid for– Hours paid for
*Can also be covered in a LF survey
ILO Department of Statistics 10
Advantages and Limitations of Employment-Related Establishment SurveysRelated Establishment Surveys
Advantages:– Permit the collection of coherent data on employees who work in sampled establishmentsPermit the collection of coherent data on employees who work in sampled establishments– Often best source of periodic data on wages & hours statistics– Good coverage of medium sized and larger establishments – Estimates disaggregated by economic activity are generally considered more reliable than
those from household surveys
• Limitations: – May exclude small establishments from sampling frame– Difficult to obtain reliable estimates for small industries– May not allow data disaggregation by demographic variables (age, sex, ethnic group, etc.)
S li d li – Sampling and non-sampling errors– Excludes self-employed workers not covered in sampled establishments
• In developing countries can represent a high percentage of total employment– Generally exclude the agricultural sectorGenerally exclude the agricultural sector
• A sector with a high concentration of employment in most developing countries
ILO Department of Statistics 11
Administrative Records• Administrative Records:
Systematic registers related to administrative procedures maintained by institutions of – Systematic registers related to administrative procedures maintained by institutions of the public sector
– Used to carry out the administration or operationalization of Government programmes– Serve to control verify and monitor compliance with legal and / or administrative Serve to control, verify and monitor compliance with legal and / or administrative
obligations– Examples of topics and indicators:
• Minimum wageg• Occupational injury rate, fatal and non-fatal• Union density rate *• Number of collective bargaining conventions and worker coverageg g g• Share of population covered by (basic) health care provision*
*Topics that may also be included in household surveys
ILO Department of Statistics 12
Advantages and limitations of Administrative Records Administrative Records
• Advantages:– A very low-cost source of dataA very low cost source of data– Allow a broad disaggregation of information and coverage of different topics– Allow for more frequent production and dissemination of data (continuous collection)– Not subject to sampling errors that are inherent to surveys– In some cases can be used to benchmark sample survey estimates – Good source of information on small groups– Population/business registers serve as sampling frames for sample surveys
• Limitations:– Generally are structured for administrative purposes rather than statistical purposes– Classification systems can differ from those used in statistical surveys– Availability of data (access) – Time and geographic coverage of the data– Coverage of target population defined by legal/administrative procedures
Quality may depend on incentives and penalties for registration– Quality may depend on incentives and penalties for registration– Concepts & definitions classifications often based on legislation, not always consistent with statistical standards
ILO Department of Statistics 13
Integrated labour statistics system
No single data source can meet all needs• All sources contribute to an overall system of national labour statistics• Need to recognize relative strengths, limitations & complementarities
Census data for benchmarking & to develop sample frames for surveys– Census data for benchmarking & to develop sample frames for surveys– Population/business registers to develop sample frames for surveys
S d t f i t l ti t & t it h t t t d– Survey data for inter-censal estimates & to monitor short term trends– Administrative data to complement or benchmark certain survey data– Survey data to gauge under-registration in administrative sources– Census and administrative data for small areas/groups
ILO Department of Statistics 14
Integrated labour statistics system
Requires coherence between sources• Concepts
• Definitions
• Classifications Labour statistics
frame ork & standards• Reference periods
• Methods
framework & standards
ILO Department of Statistics 15
Example: Labour Statistics System, Statistics Canada
Standards &Methods
Sources
Statistics
Users
ILO Department of Statistics 16
Users
16
Integrated national statistics systemg y• Labour statistics are a part of national statistical systems• Labour statistics concepts & methods need to be related &
consistent with other official statisticsEducation statistics– Education statistics
– National accounts– Demographic statistics– Industrial production statistics– Agricultural statistics– Health statistics
• Labour statistics to be included as part of the overall national statistical plan & data collection programme
ILO Department of Statistics 17
p p g
Example: South Africap
ILO Department of Statistics 18
National programme for statistics of the Economically Active Population (EAP)Economically Active Population (EAP)
Minimum needs defined in 13th ICLS Resolution I, paragraph 2:Short term needs:• Short-term needs:– Statistics of the EAP and its components to monitor trends and
seasonal variations– At minimum once a year (preferably at shorter intervals)
• Long-term needs:Comprehensive data on the EAP for structural in depth analysis and as – Comprehensive data on the EAP for structural in depth-analysis and as benchmark data
– In-depth statistics on relation between employment, income & other social and economic characteristics
– Data on particular topics (children, youth, women) – At minimum every 10 years
ILO Department of Statistics 19
At minimum every 10 years
LFS programme for current and long-term data needslong-term data needs
For long term decent work (LFS based) statistics programme: For long-term decent work (LFS based) statistics programme: – Occasional survey (e.g. one-off multipurpose survey)– Regular survey (e.g. inter-censal stand-alone or multipurpose survey)
For current decent work (LFS based) statistics programme: – Annual survey with periodic data collection for annual point-estimates
Depending on resources & capacity
– Continuous survey system with sample rotation for monthly/ quarterly/ annual average estimates
ILO Department of Statistics 20
Some strategies to improve the LF statistics programmeLF statistics programme
Objective: To build a regular LFS programme to provide for j g p g pcurrent & long-term LF data needs
Strategy: Progressive development & implementationStrategy: Progressive development & implementation
ILO Department of Statistics 21
Annual LFS with continuous data collection
Advantages:S l i i i d d i d ff li i d– Seasonal variations over time are captured and period effects eliminated
– Estimates reflect the average situation during a month, quarter or year
Flexibility in periodicity of data dissemination (depending on sample design)– Flexibility in periodicity of data dissemination (depending on sample design)
– No longer need to use concepts based on long reference periods (e.g. usual activity, annual income), which are prone to recall errors
– Improved timeliness in dissemination due to continuous data entry & processing
– Smaller dedicated field interviewers teams = better quality & lower staff turnover
ILO Department of Statistics 22
Considerations in developing a LFSp gConsiderations:– Topics that need to be covered; length of questionnaire & special modules
– Geographic coverage: total national best
– Survey periodicity
– Capacity of the institution conducting the survey (survey design; sampling; interviewer training & data collection; processing, coding and validation; estimation; and publication)
– Cost-sample size tradeoff (does funding match the desired degree of reliability for key indicator estimates?) indicator estimates?)
– Preserving public goodwill (reporting burden)
L l f d t lit ibl– Level of data quality possible
ILO Department of Statistics 23
Main objectivej• To facilitate the collection of the desired information with
– Accuracy– Efficiency– Simplicity
• All elements in a questionnaire should contribute to ensure survey lit quality
– Relevant– Contribute to reduce measurement errors
ILO Department of Statistics 24
Labour force surveyrelevance & topic coveragerelevance & topic coverage
Current activity status (in the reference week) of the Working Age Population
ILO Department of Statistics 25
Basic questionnaire structure to apply labour force framework priority ruleslabour force framework priority rules
Persons above minimum age to define the working age population
Current employment• Any work (paid/unpaid) for at least 1+ hr in reference period?1+ hr in reference period?• Temporary absence from work?
J b h/ il bilitJob search/availability• Job search in recent past (eg. last 4 weeks)?• Active search method in recent past (eg. last 4 weeks)?
A il bilit f k i f i d?
Employed
• Reasons did not search/was not available• Availability for work in reference period?
Characteristics Main Job
26
UnemployedNot in labour force-discouraged, …ILO Department of Statistics 26
Translation of standards into questions for use in household surveysfor use in household surveys
• Measurement of employment & conditions of employment of persons depends on identification of persons as employed in the persons depends on identification of persons as employed in the first place
• Questions should be appropriate for persons in paid-employment & in self-employment
• Useful to include examples in the questionnaire of most common types of jobs/activities likely to go unreported casual work unpaid types of jobs/activities likely to go unreported –casual work, unpaid work, production for own-consumption
• Important to probe!p p• Degree of probing needed depends on country circumstances
27ILO Department of Statistics
Types of measurement errorsyp
• Questionnaire– Specification problems (inappropriate measurement of
desired concept)– Design effects –question wording, length, order,
questionnaire layout• Interviewer
– Variability in conduct of interview, errors of yinterpretation/recording, deliberate manipulation
• Respondentp– Misunderstanding, memory recall problems,
omissions, concerns of social desirability ILO Department of Statistics28
Model sequence of LFS questions to identify persons
employed
Probes, as per country circumstances:Li t f ti iti t b i l d d k
p y
-List of activities to be included as work-Engagement in production for own-household consumptionp-Temporary absences from work
Essential question elements:Essential question elements:-Reference period-Reference to work for at least 1 hourSuitability for paid & self employment
29
-Suitability for paid- & self-employment-Explicit mention of paid & unpaid work
ILO Department of Statistics
Model sequence of LFS questions to identify
unemployed persons & p y p &persons not in the LF
Essential question elements:-Reference periods for job search and p javailability for work-Suitable methods of job search for paid- & self-employmentpaid- & self-employment-Identification of discouragement in reasons for not seeking work
ILO Department of Statistics 30
ILO Department of Statistics
Conclusions• Sources of labour statistics include: population and housing census, household
surveys (e.g. LFS, HIES), economic censuses, establishment surveys, administrative y ( g , ), , y ,records, and agricultural censuses and surveys
• No single data source can meet all needs• All sources contribute to an overall system of national labour statisticsAll sources contribute to an overall system of national labour statistics• Need to recognize relative strengths, limitations and complementarities between
sources• Labour statistics concepts and methods should be related and consistent with other • Labour statistics concepts and methods should be related and consistent with other
official statistics in an Integrated national statistics system• National programme for statistics of the economically active population (EAP) should
aim to develop statistics to monitor short-term trends and for in depth-analysisaim to develop statistics to monitor short term trends and for in depth analysis• Labour force surveys should seek to facilitate the collection of the desired information
with accuracy, efficiency and simplicity
31ILO Department of Statistics
References• Benes, Elisa. Power Point presentation, “Introduction to labour and labour force statistics”. ILO-
ITC-Turin Training Course: LMI2: Designing labour force surveys and LF modules for household surveys to monitor progress towards decent work, ILO Department of Statistics, Geneva. April 2011
• Benes, Elisa. Power Point presentation, “Putting it all together” ILO-ITC-Turin Training Course: LMI2: Designing labour force surveys and LF modules for household surveys to monitor progress LMI2: Designing labour force surveys and LF modules for household surveys to monitor progress towards decent work, ILO Department of Statistics, Geneva. April 2011
• Castillo, Monica. Power Point presentation, “Data Sources for Collecting Decent Work Information”, World Bank / AusAid / ILO Regional Labor Markets Analysis Workshop Bangkok, Thailand. April o d a / us d / O eg o a abo a e s a ys s o s op a g o , a a d p2011
• Hussmanns, Ralf, Farhad Mehran and Vijay Verma. Surveys of economically active population, employment, unemployment and underemployment: An ILO manual on concepts and methods. Geneva: International Labour Office. 1990.
• Resolution concerning statistics of the economically active population, employment, unemployment and underemployment, adopted by the Thirteenth International Conference of Labour Statisticians (October 1982); web link:(October 1982); web link:
• http://www.ilo.org/wcmsp5/groups/public/---dgreports/---stat/documents/normativeinstrument/wcms_087481.pdf
ILO Department of Statistics 32
Questions• Are labour statistics concepts and methods, classifications, etc. consistent across
national labour statistics sources? Are they consistent with other official statistics in yTrinidad and Tobago? Does a national statistics programme coordination mechanism exist? If so, how well does it function and what are the obstacles to improving it?
• What are the current statistical needs of data users in Trinidad and Tobago regarding labour force/employment statistics? g g p y
– What are the resources available?– What are the existing statistical sources for labour force/employment statistics?
• Establishment surveysE i • Economic census
• Household survey• Administrative records• Agricultural census/survey
– Do data sources include instruments for short-term and long-term needs?– Are the existing statistical sources meeting the needs of data users? Why?
ILO Department of Statistics 33