Republic of Rwanda

LABOUR STATISTICS
FRAMEWORK OF
RWANDA

NATIONAL INSTITUTE OF STATISTICS OF RWANDA

 Republic of Rwanda

LABOUR STATISTICS FRAMEWORK OF RWANDA

Additional information about the Labour Statistics Framework of Rwanda may be

obtained from the NISR:

P.O. Box 6139, Kigali, Rwanda; Telephone: (250) 252 571 035
E-mail: [email protected]; Website: http://www.statistics.gov.rw.

Recommended citation:

National Institute of Statistics of Rwanda (NISR), Ministry of Finance and Economic

Planning (MINECOFIN) [Rwanda], Labour Statistics Framework of Rwanda
Contents

Chapter 1: Background ......................................................................................................... 1

1.1 Scope ........................................................................................................................ 1
1.2 Purpose .................................................................................................................... 1
1.3 Uses .......................................................................................................................... 1
1.4 Outline of document ............................................................................................... 2
1. Conceptual labour statistics framework for Rwanda ................................................ 3
1.1. Labour supply .......................................................................................................... 3
2.2 Labour demand ....................................................................................................... 3
2.3 Labour market ......................................................................................................... 4
2.4 The Rwanda Labour Statistics Framework (RLSF) ........................................... 5
2. Operational framework for Rwanda Labour Statistics ........................................... 11
3.1 Structure for organizing Rwanda labour statistics ........................................... 11
3.1.1 Labour supply (labour force) ............................................................................ 11
3.1.2 Labour supply (Other work activities) ............................................................. 11
3.1.3 Labour demand .................................................................................................. 11
3.1.4 Context ................................................................................................................ 11
3.2 Sources of Labour supply statistics ................................................................... 12
3.2.1 Labour force surveys ......................................................................................... 12
3.2.2 Other household surveys ................................................................................. 13
3.2.3 Population censuses ......................................................................................... 14
3.2.4 Other surveys ..................................................................................................... 14
3.2.5 Administrative records ...................................................................................... 15
3.3 Sources of labour demand statistics .................................................................. 15
3.3.1 Establishment surveys ...................................................................................... 16
3.3.2 Establishment censuses ................................................................................... 17
3.3.3 Administrative records ...................................................................................... 17
3.4 Conclusion ............................................................................................................. 18
3. Available labour statistics system in Rwanda ......................................................... 20
4.1 Major producers of labour statistics ................................................................... 20
4.2 Assessment of sources of labour market statistics in Rwanda ..................... 29
4.2.1 Source institution: NISR ....................................................................................... 29

i
 4.2.2 Source institution: Rwanda Social Security Board (RSSB) ............................ 35
4.2.3 Source institution: Rwanda Revenue Authority ................................................ 35
4.2.4 Source institution: Ministry of public service and labour (MIFOTRA) ............ 36
4.2.6 Source institution: Rwanda Development Board (RDB) .................................. 37
4.2.7 Source institution: Ministry of Education (MINEDUC) ..................................... 37
4.2.8 Source institution: Workforce Development Agency (WDA)........................... 38
4.2.9 Other institutions .................................................................................................... 38
4. Bridging the Gap ................................................................................................... 39
5.1 NISR ........................................................................................................................ 39
5.1.1 The Annual Labour Force Survey ................................................................... 39
5.1.2 The annual integrated business survey ......................................................... 41
5.1.3 3-yearly establishment census ........................................................................ 42
5.1.4 3-yearly EICV ..................................................................................................... 42
5.1.5 Dissemination ..................................................................................................... 43
5.2 RSSB ...................................................................................................................... 43
5.3 RRA ......................................................................................................................... 44
5.4 MIFOTRA ............................................................................................................... 44
5.5 RDB ......................................................................................................................... 45
5.6 MINEDUC............................................................................................................... 45
5.7 WDA ........................................................................................................................ 45
5.8 Other Institutions/sources .................................................................................... 46
5.9 National Statistics System (NSS) ....................................................................... 46
5. Programme for the Production of labour market statistics for the RLSF (2014/15
– 2019/20) ............................................................................................................................. 47
Part I: Situation Analysis ............................................................................................... 47
Part II: Sub-Programme A – Production of labour statistics from Surveys and
Censuses (2014/15 – 2019/2020)................................................................................. 48
Part III: Sub-Programme B – Production of labour statistics from Administrative
Sources (2014/15 – 2016/17) ........................................................................................ 53
Part IV: Conclusion ...................................................................................................... 55
Annex 1: Work Plan and budget ....................................................................................... 56
Annex 2: Labour Market Indicators* .................................................................................. 60

ii
Chapter 1: Background

1.1 Scope

Labour statistics are used to describe and analyse the size and structure of labour
markets and how these change over time and space, including differences between
groups. They relate to persons and businesses, describing their characteristics and
behaviours in and around the world of work, and their interactions with government.
In its widest sense, work references work done for pay or profit (employment work),
work done in production of goods and services for own consumption (own-use
production), work done by unpaid trainees and apprentices and volunteer work. In
the context of the labour market, work done in own-use production of services,
which is excluded from the System of National Accounts (SNA), is not covered.

1.2 Purpose

The labour market in Rwanda is becoming more and more complex over time,
raising many questions. To what extent is employment provided less and less by
government and public service and more and more by the private sector? Are there
less rural subsistence agricultural jobs and more urban own-account low quality
jobs? What effect does volatile global markets and the existence of interactive
regional markets and economies have on the economy and on the labour market?
The population is increasing as well as the number of educated persons. What
pressure are they putting on the labour market? Are sufficient new jobs being
created to absorb the increase? Are these good jobs? How fast is production
shifting from the agriculture, manufacturing sectors to the service sectors? Are there
skill shortages and skill mismatches in the labour market? Do shortages of skilled
workers in priority sectors exist? If so, which ones and which occupations? Are they
hard technical skills or soft skills such as initiative, flexibility, communication, group
interaction, etc.? At the same time, are there persons working in jobs lower than
their skill level? If so, who are they and what are their characteristics (age, sex,
location, areas of study, etc.)?

These questions and many more are posing challenges for producing relevant
labour statistics. It is therefore imperative to develop an organizing framework for
their production, analysis and dissemination in response. Whilst the framework
recognizes the needs of EDPRS2 and NEP, it takes an all-encompassing approach
to organizing labour statistics in general.

1.3 Uses

Labour statistics are used for policy formulation as well as for development,
evaluation and monitoring of programmes in various areas such as employment and
active labour market policies, vocational training, human resource development,
poverty and social exclusion.
1
In particular, they are needed for:
 employment and active labour market policies and programmes (e.g. NEP);
 analysis of the dynamics of the labour market with respect to (a) the size,
distribution and characteristics of the labour force; and (b) the evolution in
labour demand, including job creation and vacancies (e.g. EDPRS2 targets);
 assessing mismatch between labour supply and demand, instrumental in the
formulation of policy regarding the role and performance of the educational
system, training and labour adjustment policies (one of the key objectives of
EDPRS2);
 macro-economic policies and measurement of the performance of the
economy;
 analysis of terms and conditions of work including wages policies and
protection of workers (health and safety);
 labour market regulatory policies, industrial relations policies, monitoring the
volatility of labour management relations;
 welfare policies, poverty concerns, social inclusion and other assessment of
hardship of the population;
 monitoring the employment situation of special groups such as women and
youths;
 providing signals on undesirable economic and social conditions of work
such as child labour, gender biases in work places, etc.

All of the above are required at sub-national, national as well as supra-national

levels.

1.4 Outline of document

The conceptual framework is introduced in Chapter 2. It is an adaptation of the

classical model of a labour accounting system. Chapter 3 describes the operational
framework, in terms of identifying a structure for organizing labour statistics in
Rwanda and describing the best sources for producing these statistics. In Chapter
4, the labour statistics that are available in Rwanda are presented and assessed,
along with the producing institution, vis-a-vis the identified best sources. Chapter 5
presents recommendations for moving towards the ideal sources through the
realisable ones. It also makes suggestions with respect to institutional
arrangements, analysis and dissemination strategies. Chapter 6 is about the
program of production of labour market statistics for the Rwanda Labour Statistics
Framework

2
Chapter 2: Conceptual labour statistics framework for Rwanda

2.1 Labour supply

Persons (living in Rwanda or elsewhere) actually supply their labour or are potential
suppliers of their labour to employers in Rwanda – the supply side. These persons
live in households, broadly defined. These households have characteristics such as
geographical location, size and so which can then be associated with persons living
in them. The persons themselves have their own individual characteristics, indeed
very many, such as age, sex, educational attainment, migration status, etc. These
household characteristics and/or person characteristics can then be used to
generate statistics relating to the numbers of persons and/or households having
each of these characteristics. (See Diagram 1). For example we can produce the
distribution of persons by sex. These are the initial labour (supply) statistics.

2.2 Labour demand

Employers in Rwanda (defined in a wide sense to include own-account workers and

households employing staff) require labour to carry out their production activities –
demand side. They parcel out the work that needs to be done into individual posts.
These employers have characteristics (e.g. economic activity sector, institutional
sector, geographical location, ownership status, etc.) which then pass on to the
posts. The posts themselves have their own characteristics such as occupation,
contractual hours of work, working time arrangements (e.g. part-time, full-time),
educational qualifications, skills, experience, etc. associated with the post. Counting

3
the number of posts having any of these employer and /or post characteristics can
generate statistics on labour demand. (See Diagram 2.) For example the distribution
of posts by level of qualifications required can be generated. These are the initial
labour (demand) statistics.

2.3 Labour market

Most posts created by employers have financial returns attached to them in the form
of monetary gains in cash or kind. When a person and an employer agree on terms
for such a post, the person fills the post, which then becomes the person’s job. So
jobs are the point where labour supply meets labour demand. A person that has
secured a post is labelled ‘employed’ and is considered to be in the ‘labour force’. It
is important to note that a person can in fact have more than one job negotiated
with one or more employers. Some persons however potentially want to supply their
labour by securing any post but are unable to do so even though they want to and
are looking for one. They are labelled as being ‘unemployed’ but are still considered
as being in the ‘labour force’. Some others who also potentially want to supply their
labour by securing a post are unable to do so even though they want to do so but
are not looking. They are considered as ‘out of the labour force’ and as being in the
‘potential labour force’. Hence the supply side can be categorized as “employed”,
“unemployed” and “out of the labour force”. (Diagram 3)

A job is then a post that has been filled, a filled post. However employers may look
for persons to take up their posts without success. Such posts are then considered
to be ‘vacant posts’. A few posts do not have any financial returns attached to
them, for example posts occupied by volunteers or by persons working for their own
4
consumption. When these are filled, they are not considered as jobs but are
referred to as ‘other filled posts’. Thus the demand side can be categorized as jobs,
vacant posts and other filled posts. It is possible, though rare, that a post can be
filled by two or more persons and so become two or more jobs, e.g. two persons,
one working mornings (as a part-time job) and the other working afternoons (as
another part-time job) filling the same post! (Diagram 3).

So the above statistics of labour supply can now be extended to the sub-groups of
‘employed’, unemployed’ and ‘out of the labour force’ whilst those of labour demand
can be extended to the sub-groups ‘filled posts (jobs)’, other filled posts and ‘vacant
posts’.

2.4 The Rwanda Labour Statistics Framework (RLSF)

When a job is created by the agreement between an employer and a person

through an implicit or explicit contract, it includes the characteristics agreed on by
both parties (set of tasks and the conditions of work) – the duties and
responsibilities, the pay and method of payment, the normal and or overtime hours
of work expected, the entitlement to benefits such as paid leave, training, to
insurance against occupational injuries and diseases and so on. Others come in
with the implementation of the agreement such as hours actually worked in the job
or actual overtime hours. These then translate through the job to both the employed
5
person and to the filled post. Statistics relating to these characteristics as well as
the household and person characteristics can then be produced for each employed
persons or their households, for example the average number of hours worked by
employed persons in Kigali. These are then extensions to the labour (supply)
statistics. In a similar way the labour (demand) statistics are extended to take on
these acquired characteristics from the job, for example the average pay for
manufacturing jobs in the private sector.

Another extension to both labour (supply) statistics and labour (demand) statistics
comes through the linking of the supply side characteristics of the employed person
to the demand side characteristics of filled post through the job. Thus, these
statistics can now be produced by combining any of those of the employed person,
their household, their job, the filled post and employer characteristics. For example
the well-known distributions of employed persons by occupation, industry,
institutional sector or status in employment come from this. The average income,
hours worked in all jobs of employees by industry is another example. (Diagram 4)

6
Diagram 5: Operational frame work for Rwanda Labour Statistics

7
Diagram 5 presents an alternative to the above framework and its wide range of
statistics that is more familiar and closer to the way these statistics can be
generated. This alternative assigns the inherited characteristics from labour demand
to the labour force categories and vice-versa. As a result, they can be seen by the
way the data are usually collected through household surveys, on the one hand,
and establishment surveys on the other. The presentation thus leads to the
operational framework discussed in the next chapter.

A major actor in the labour market is government, through its regulatory framework
and its social protection framework to ensure equity and protection for all, both
employers, those in the labour force and in the potential labour force. Government
also influences the labour market through its macro-economic, fiscal and micro-
economic policies which enhance the environment for employers to create jobs.
Statistics relating to industrial relations, labour and factory inspection services,
social security directly impact on the labour market and those relating to inflation,
economic growth, infrastructure and productivity do so indirectly (Diagram 6).

Diagram 6: Extension of labour statistics framework for Rwanda

Linkages to other statistical systems come through the positioning of labour

statistics as a bridge between social and economic statistics. (Diagram 7) In
particular, the linkages with social statistics are as a result of persons being the
basic units on the labour supply side and in all these systems. The interaction is
both in terms of statistics relating to persons in the labour force and their situation
vis-a-vis these other systems as well as statistics relating to the work force required
in these other systems. For example, links with health statistics come from the need
to follow up the health status of workers, in particular occupational injuries and
diseases, as well as to describe and analyse the health work force. Education
statistics are required to assess the current and future skilling of the labour force
whilst labour statistics contribute to the planning and analysis of the education work
8
force. Population statistics is key in following up the evolution of the labour force
and its categories. Migration statistics helps in understanding skill needs and
planning of training programmes. The demand side links up with economic
statistics mainly through employers. For example, statistics relating to sectoral
growth and productivity are important in analysing employment and in the
development of appropriate training programmes to meet the expected demand
from employers. All these systems also interlink through the common use and
exploitation of data collection instruments, common definitions and classifications.

Diagram 7: Linkages of the Rwanda Labour Statistics Framework to Other

Statistical Systems

The labour statistics framework is thus a core element of the national statistical
system, as enunciated in NSDS2.

The advantage in utilizing a labour accounting framework is that it enables an

exercise in reconciling statistics from both the supply and demand sides and
improving all aspects of the quality of labour market statistics. Statistics obtained
from different sources for the same characteristic could differ for various reasons.
There may be differences due to:

a) Definitions (e.g. unemployment and job-seekers) and classifications (age

groups) of variables;
b) Coverage - scope of units (e.g. all workers and paid employees), of variables
(income and taxable income), types of units (enterprises and
establishments);
c) Precision - censuses and samples; national and district;

9
d) Errors - types and levels of errors.

Possible actions that can be taken in such cases are: simply noting the differences;
harmonising definitions etc.; reconciling the differences (i.e. account for
differences), adjusting one or other of the estimates. This coherence between
sources, an important quality measure, can be assessed through using the
framework. For example, the conceptual model described above implies that the
total number of people employed (in their only or main job, their second job and so
on, counting them as often as they have jobs) as measured from a labour force
survey should be consistent with the total number of jobs as measured by an
establishment survey, everything else being equal.

10
Chapter 3: Operational framework for Rwanda Labour Statistics

3.1 Structure for organizing Rwanda labour statistics

The structure proposed for the framework is the one adopted in the Metadata
Handbook (Annex 1). It flows naturally from the above conceptual framework,
grouping the statistics according to their relevance for labour supply and labour
demand as well as introducing context indicators for the interactions between
government’s interventions and the labour market. It consists of four major
categories: labour supply (labour force), labour supply (other work activities), labour
demand and context. Each category is further sub-divided into groups representing
major themes.

The structure is as follows:

3.1.1 Labour supply (labour force)

1) Population
2) Labour force
3) Employment
4) Labour underutilization
5) Wages
6) Skills
7) Employment equity & Industrial relations
8) Safe work & Social protection
9) Poverty of employed population

3.1.2 Labour supply (Other work activities)

10)Other labour input into SNA

11)Other non-SNA labour input

3.1.3 Labour demand

12)Vacancies & Jobs

13)Labour as costs to employers
14)Establishments

3.1.4 Context

The above structure fits in with other alternatives such as:

 The ILO Convention 160 and Recommendation 170,

 ILOSTAT, the ILO’s flagship database,
 ILO’s Key Indicators of the Labour Market (KILM),
 ILO’s Decent Work Indicators (DWI); and
 Millennium Development Goals (MDGs)
11
  East African Community database (EAC).

3.2 Sources of Labour supply statistics

These statistics, as described on the right-hand side of Diagram 5, can be obtained

from a variety of the following sources:

 Directly from households, through labour force surveys, other household

surveys or population censuses;
 Indirectly from households, through reporting by persons to some institutions,
e.g. Workers’ Organization, Employment services.

3.2.1 Labour force surveys

Labour force surveys are the best source for these statistics in terms of the
coverage of the population, the wide range of topics relating to the work activities of
the population and the high quality of data collected on these topics. Labour force
surveys, appropriately designed, can cover:

Virtually the entire population of a country and all categories of workers,

including the self-employed, unpaid family workers, casual workers and
multiple jobholders;

Socio-demographic details, geographical location, education, training and

other person and household characteristics (as listed in Diagram 1);

All branches of economic activity and all sectors of the economy;

Employment, employment in the informal sector and informal employment;

Unemployment, under-employment, the potential labour force and other

forms of inadequate employment;

Characteristics of jobs of employed persons - occupation, industry,

employment status, institutional sector, hours of work, wages and
employment related income, methods used to find the job, work experience,
etc. – equivalent to characteristics of filled posts, if all jobs (main, secondary,
etc.) are counted;

Characteristics of previous jobs of the unemployed and persons out of the

labour force;

Other work activities: production of goods and services for own use, unpaid
trainee/apprentice work, volunteer work;

Population out of the labour force;

Others: labour migration, social security & social protection, trade union
membership, children’s activities & child labour, occupational injuries,
12
 vocational training, labour market experience of recent school leavers, etc.
through special modules

In addition, labour force surveys allow joint and mutually exclusive measurement of
the employed, unemployed and those out of the labour force in a manner that is
proper (i.e. according to international standards1) and produces good quality data.

Using appropriate sampling schemes, labour force surveys can be designed to

measure individual changes between and within labour force categories and thus
provide not only stock data for a given point or period of time, but also data on flows
and gross changes over time.

However, labour force surveys

 are relatively costly to conduct, especially if conducted with high frequency;

 introduce imprecision through sampling, especially for small area statistics of
interest (e.g. Number of female unemployed in Kiyovu), and so have limited
application for statistics at detailed geographic and other disaggregation
levels;
 are subject to recall errors;
 do not always cover all jobs of employed persons;
 may not produce good data on income from employment, especially of the
self-employed;
 exclude persons employed who are not domicile;
 do not collect information on vacancies and on training needs; and
 require good statistical infrastructure.

Labour force surveys should preferably be regular (monthly or quarterly in some

countries, at least annually), generally comprehensive (covers the whole population)
and provide high quality estimates meeting international standards.

3.2.2 Other household surveys

Data on employment, unemployment and persons out of the labour force at less
detailed level are also produced as part of other household surveys such as living
standards surveys, household income and expenditure surveys, child labour
surveys, demographic and health surveys. They can sometimes by classified by
some of the same characteristics as in the labour force survey. They can enrich the
analysis of the employment situation of persons by cross-classifying the labour
force status of persons with the main survey variables, for example to produce
estimates of working poor, of employed persons living with HIV/AIDS, etc.

13
They usually have good coverage of the population. However the coverage of
household/person categories would depend on main purpose of survey. Also, the
classification of persons by their labour force status and related data on the
characteristics of these groups may not be of the best quality due to small sample
sizes and the demands on enumerators to cover so many topics. They also have
limited coverage of employer, posts and jobs categories relative to the labour force
survey. These surveys also have the same challenges as those for labour force
surveys.

3.2.3 Population censuses

The importance of this source of labour supply statistics is in the

comprehensiveness of its coverage of the population. It can thus provide estimates
of basic labour force statistics at very detailed geographical levels (e.g. village
unemployment rates) and at detailed levels of occupational and industrial
classifications (e.g. number of plumbers in the farming sector in a given district).
The population census can also relate the labour force characteristics of the
population to other census variables such as education and migration.

The main challenge in using this source is the likely poor quality of the employment
data due to the use of fewer questions compared to the previous sources. As a
result, the classification of persons by their labour force characteristics is less
precise and there is limited scope for topics covered with respect to the person,
employer, post and job categories. It is also an expensive exercise, hence its 10-
yearly frequency and excludes persons employed who are not domicile.

The population census is a very good sampling frame for household surveys.

3.2.4 Other surveys

Statistics relating to training can be derived from special surveys or censuses of

training/education institutions. Data is collected on staffing of these institutions as
well as on final outputs, other exits and enrolment at different levels of the courses.
Such data is valuable in understanding the skills availability of the future cohorts
entering the labour force and could guide to further training that may be required.
The challenge facing such surveys is the low response rates due to school systems
that are already over-burdened with administrative and academic tasks. Ideally,
such surveys should be annual.

A preferred method for obtaining this data is through the administrative system of
the ministry of education. Provided the system is efficient and effective, the data
collected will be comprehensive and complete. The cost will be relatively cheap and
the data will be available annually. A computerized system is best placed to do this.

Also tracer type surveys provide information on the adequacy and efficiency of the
training system by assessing the labour market performance of past graduates from
14
the system and the satisfaction of their employers with their entry skills. It is
necessary to allow a sufficient time lapse between the exit of the graduates from the
training system and the survey so that enough appreciation of the use of the
available skills of the graduates would have been gained by both the graduates and
the employers. The challenge is ensuring a good sampling frame of these
graduates with up-to-date contact information. Otherwise, the rate of persons not
traced could be high. As the time since graduation increases, the higher this rate
will be. A good balance is required between these two apparently conflicting time
constraints. Such surveys are carried out about once every 3 to 5 years.

A particular type of survey relating to training is the school-to-work transition survey

which is directed at ascertaining the experiences young school-leavers are having
relating to their entry or otherwise into the labour market. This is largely an ad hoc
survey.

3.2.5 Administrative records

Statistics on some aspects of labour supply can be obtained from the administrative
records of institutions to which persons report in relation to their work activities such
as their trade unions and employment agencies. These sources are useful for some
labour force characteristics such as trade union membership and the job-seeking of
persons and can also include some person categories such as sex. They are low
cost, reference the whole group covered across the country without any additional
reporting burden on respondents. Timeliness is another good characteristic of these
sources due to their reporting mechanism. There are however many drawbacks
usually associated with administrative sources such as the possibility of a change in
their current legislation which could affect the data collected, the scope and
coverage issues (e.g. not all unemployed register as job-seekers and not all who
registered job-seekers are unemployed), false/inadequate reporting and issues of
confidentiality.

3.3 Sources of labour demand statistics

Statistics on labour demand are produced from various sources:

 Directly from enterprises, through establishment censuses; establishment

surveys; notification system.
 Indirectly from enterprises, through reporting to other institutions, e.g. Tax
authorities, Social Security, Employment Ministry, Immigration, Employers’
Organization and Employment Services

15
3.3.1 Establishment surveys

Establishment surveys with employment data provide an excellent source of data on

income from wage employment, earnings, labour costs, hours of work (hours
actually worked, hours paid for, normal hours and overtime hours) and labour
turnover (based on payroll records). They are a good source for statistics on paid
employment (including jobs created), labour demand (vacancies), training needs,
conditions of work and labour productivity. They can be employment specific: (a) the
establishment surveys of employment, earnings and hours of work conducted
monthly, quarterly, half-yearly or annually; (b) manpower (labour demand) surveys
of establishments and their individual employees conducted every 3 to 5 years.
Sometimes the employment data is generated as part of a production survey of
establishments.

Manpower surveys collect total employment, paid employees (occupation, sex,

educational level, nationality, etc.), labour turnover (main causes, characteristics of
persons), vacancies, future job openings (employers’ expectations), training
provided, skills needed, etc. from the establishments. From the employees data
collected include socio-demographic characteristics, educational attainment, job
characteristics, training received, previous job history and information on their other
jobs.

Establishment surveys produce good data for population of establishments covered

and the variables available as the information is usually based on establishment
records. Also, the data collected is consistent with those for production and factor
income (SNA) and is usually at detailed industry and occupation levels (3-digit
codes). They are easier to implement than household surveys and not as
expensive.

A major challenge faced by such surveys is that their coverage depends on the
quality of the population frame of businesses. Often, only registered businesses or
formal sector establishments of a certain size are included in the frame. Sometimes
the frame excludes very small businesses, certain economic activities and many
informal sector enterprises, especially those operating from home or which are
itinerant. As the birth and death rates of establishments are high, the frame quickly
becomes out of date with incomplete coverage and/or dead elements. The surveys
have limited information on person/household categories for those in filled posts
and suffer from poor response rates. Moreover the data submitted is only summary
information on filled posts (paid employment) and total wages classified by sex. It
should be noted that establishment surveys provide information on jobs and
persons in those jobs. So for multiple jobholders, there is some risk of double
counting if the statistics are interpreted as number of employed persons. Also the
data may include non-domicile persons.

16
3.3.2 Establishment censuses

Establishment censuses are conducted less often than establishment surveys,

sometimes 5-yearly, and collect limited information about establishments. They
produce a register of establishments which can act as a good frame for the
establishment surveys provided it is maintained and kept up-to-date.

3.3.3 Administrative records

Labour statistics may also be compiled using administrative records of agencies

such as the ministry of labour, immigration authorities, the social security
organization, the revenue authority, the employers’ organizations and employment
service agencies. Examples of such records with statistical potential are:
Employment services registers; Unemployment insurance records; Social security
files; Revenue authority files; Reports of disputes resolutions; Reports of
labour/factory inspections; Public sector payrolls; Employers’ organization records,
Immigration office records.

These records can be used to provide estimates of

 employment (actually jobs), employers (establishments ) and their

characteristics (from social security files, revenue authority files, employers
organizations records),
 wages, social security beneficiaries and benefits paid (social security files,
revenue authority files),
 public sector employment and wages (public sector payrolls);
 migrant workers (immigration office records),
 occupational injuries (compensation payment records),
 registered jobseekers and notified vacancies (employment services
registers);
 industrial relations matters (reports of dispute resolutions, union registrations,
collective agreement registrations), safety and health, observance of labour
laws (reports of labour/factory inspections),
 membership of employers’ organization (employers’ organization records),
and so on.

Administrative data are the easiest and cheapest source of labour statistics. They
are timely and current, as they are frequently produced, and can be obtained at
detailed levels of disaggregation. They have good coverage, depending on
compliance of establishments concerned, and good data on variables of interest to
administrative process. They are frequent and timely and not too expensive,
relatively. They provide a full count of their relevant units and impose no reporting
burden on them. It is sometimes possible to provide statistics from them at deep
geographical and other disaggregation levels.

17
However as they are produced as a by-product of the administrative functions of an
agency, (i) their definitions etc depend on the legislation and the administrative
system of the agency; and (ii) they are subject to the whims of the administrative
processes. As such they rarely conform to accepted statistical standards. Also, they
may have low coverage (not all units are registered or required to be registered)
and may not be timely, as priority may not be given to statistics. Data on variables
not of direct concern to the administrative process may be unreliable (data checking
may be restricted to items of strong interest to the admin system) and so may be of
poor quality, e.g. sex to tax authorities. There is also possibility of inaccurate
reporting to maximise benefits or to minimise penalties. Statistics may vary over
time and between locations because of differences in procedure, resources, etc.

It is important to note that the statistics produced are on jobs and not necessarily on
employed persons. Also the data may include non-domicile persons. Finally, the
data may not be in a format convenient for producing statistics and so there could
be conversion costs.

3.4 Conclusion

Statistics on the extended labour force, employment, labour underutilization, the

informal economy and their characteristics are best obtained from labour force
surveys. To the extent that all jobs of the employed population are included, these
surveys also produce good data on job characteristics. This is particularly useful
when establishment surveys do not cover small and informal sector enterprises.
These statistics can also be produced by other household surveys but with less
rigour and less scope.

Statistics on jobs and their characteristics, wages, hours of work, vacancies, training
needs and skills mismatches are best obtained from establishment surveys,
especially manpower surveys, provided coverage is complete. Again, provided the
incidence of multiple jobs is low, the statistics on jobs can also refer to the
employed population.

Administrative systems such as social security institutions and revenue authorities

can be used to produce statistics on jobs and their characteristics as well as wages.
However the extent to which they could be rigorous and complete would depend on
take-up by establishments.

Income from employment statistics can be produced from labour force surveys but
accuracy and rigour, particularly with self-employment income, could be
challenging. To the extent that it can be done accurately, income from self-
employment is best obtained from household surveys on income and expenditure
with an enterprise module in which it can be measured as mixed income.

Other sources can variously provide these statistics but with less scope, rigour,
coverage, etc. than the above.
18
Thus no single data source can meet all needs of the labour statistics framework.
All available sources should be seen as components of overall system of labour
statistics. The choice of any one source for a particular need must be done
carefully, taking into consideration their relative strengths and limitations.

As many types of labour statistics can be generated using different sources, in

disseminating the results it is necessary to make users aware that each source has
differing strengths and limitations. Dissemination of labour statistics should follow
the guidelines approved by the 16th International Conference of Labour Statisticians
(ICLS)2, the UN Fundamental Principles of Official Statistics3 and the IMF GDDS
framework4.

2
http://www.ilo.org/global/statistics-and-databases/standards-and-guidelines/guidelines-adopted-by-international-
conferences-of-labour-statisticians/WCMS_087614/lang--en/index.htm
3
http://unstats.un.org/unsd/dnss/gp/fundprinciples.aspx
4
http://dsbb.imf.org/pages/gdds/home.aspx
19
Chapter 4: Available labour statistics system in Rwanda

4.1 Major producers of labour statistics

The table below presents the various data sources from all major producers of
labour statistics according to the above structure of the labour statistics framework
for Rwanda. For each group in the structure, the best source available is indicated
with a star. The contents of the table are based on the published reports,
questionnaires and websites for each source.

The major producers of labour statistics in Rwanda are identified as follows:

 National Institute of Statistics for Rwanda (NISR)

 Rwanda Social Security Board (RSSB)
 Rwanda Revenue Authority (RRA)
 Ministry of public service and labour (MIFOTRA)
 Rwanda Development Board (RDB)
 Ministry of Education (MINEDUC)
 Workforce Development Authority (WDA)

Some of these producers disseminate their statistics through the website of the
Labour Market Information System (LMIS) of the Rwanda Development Board. In
some instances, LMIS adds value to the statistics by adjusting some of the values.
The data collection methodologies used by these institutions are surveys, censuses
and exploitation of their administrative records.

20
Table 1: Available data sources for the labour statistics framework for Rwanda

Data
Group Source Disaggregation (as published) Frequency Agency Notes
type
Labour supply (labour force)
Population 10-yearly (last Working age population,
HB Age, Sex, Province, Residence NISR
census* 2012) Dependency ratios
5-yearly (last
Working age population,
EICV HB Age, sex, province, residence EICV 3 NISR
Population Dependency ratios
2010/2011)
National Projections of working age
10 yearly (Last NISR /
Population AB Age, sex, literacy population and dependency ratios
2014) LMIS
Projections above (2007 – 2022)
5-yearly (last
Age, Sex, Province, Residence,
EICV* HB EICV 3 NISR
Educational attainment, Nationality
Labour force 2010/2011)
Population 10-yearly (last
HB Age, Sex, Province, Residence NISR
census 2012)

Age, Sex, Province, Residence,

Data to measure informal
Educational attainment, Nationality, 5-yearly (last
employment collected but not used,
Employment EICV* HB Industry, Occupation, Status in EICV 3 NISR
Does not include employment in
employment, Hours worked, 2010/2011)
informal sector.
Institutional sector

21
 Data
Group Source Disaggregation (as published) Frequency Agency Notes
type

Age, Sex, Province, Residence,

Population Educational attainment, Nationality, 10-yearly (last Does not disaggregate by
HB NISR
census Industry, Occupation, Status in 2012) formal/informal sectors.
employment, Institutional sector

Informal sector
HB Once, 2007 NISR
survey

Sex, Province, Residence, Actually number of persons in jobs,

Establishment Educational attainment, Nationality, with possible multiple counting of
EB Once 2011 NISR
census Industry, Institutional sector, Size of employed persons with multiple
establishment, jobs. Not all employed covered.

Employers’ Module: Actually number

of persons in jobs, with possible
multiple counting of employed
persons with multiple jobs.
Age, Sex, Province, Nationality, Employers’& Employees modules:
Industry, Occupation + Marital status, Formal sector employees & persons
Manpower survey EB Once 2012 NISR
Educational attainment, Training, working in informal sector
Disability status, etc. enterprises, Formal sector
employers & Informal sector
employers, Not all employed
covered. Did not cover
establishment with only one worker.

22
 Data
Group Source Disaggregation (as published) Frequency Agency Notes
type

All registered employers and

contributors, Contributors only, Not
Annual Statistical Province, Industry, Institutional sector, Annual (last all employed covered. Also possible
AB RSSB
Bulletin Age, sex 2010) multiple counting of persons with
multiple jobs, if all jobs registered
with RSSB.

All registered employers and

contributors, Contributors only, Not
Province, Industry, Institutional sector, Quarterly (last
LMIS / all employed covered. Also possible
RSSB records AB Age, sex, migration, marital status (at November
RSSB multiple counting of persons with
affiliation) 2013)
multiple jobs, if all jobs registered
with RSSB.

Annual (Last LMIS / All registered TPR and income tax

RRA records AB Province, Industry
January 2014) RRA payers, LMIS adjusted data
Age, Sex, Province, Industry,
Occupation, Nationality, Hours of
Labour work, Status in employment, MIFOTRA / All employed in inspected
AB Variable
inspections Educational attainment, Duration of LMIS establishments
employment, work accidents, labour
disputes , labour injuries
Age, Field of education, Level of MIFOTRA /
IPPS AB Continuous Public sector employees only.
education, Contract, Location LMIS
ORG/RDB RDB / All registered enterprises. Number
AB Province, Institutional Sector
records LMIS of intended employees
Statistics on time-related
Age, Sex, Province, Residence, 5-yearly (last
Labour underemployment as defined
EICV* HB Educational attainment, Previous job EICV 3 NISR
underutilization internationally not published but
characteristics 2010/2011)
data collected.
23
 Data
Group Source Disaggregation (as published) Frequency Agency Notes
type

Population Age, Sex, Province, Residence, 10-yearly (last

HB NISR Only unemployment.
census Marital status, Nationality 2012)
Manpower Employees only. No self-
EB Industry, Occupation Once 2012 NISR
survey* employment income.
5-yearly (last
EICV HB EICV 3 NISR Data collected but not analysed.
2010/2011)

Annual Statistical Annual (last Monthly income distribution:

EB None RSSB
Bulletin 2010) Contributors only
Wages
Labour Age, Sex, Province, Industry, MIFOTRA / Wages of all employed in inspected
AB Variable
inspections Occupation LMIS establishments
Age, Field of education, Level of MIFOTRA / Wages of public sector employees
IPPS AB Continuous
education, Contract, Location LMIS only.
Quarterly (last
Age, sex, migration, marital status (at LMIS /
RSSB records AB November Salary of Contributors only
affiliation) RSSB
2013)
5-yearly (last
EICV* HB Occupation, Labour force status EICV 3 NISR Measured by educational attainment
2010/2011)
Population 10-yearly (last
HB Occupation, Labour force status NISR Measured by educational attainment
Skills census 2012)

Manpower survey EB Occupation, Industry Once 2011 NISR Skills mismatch by Self assessment

Statistical Annual (last

AB Level of education MINEDUC Outputs, Enrolled persons
yearbook 2012)

24
 Data
Group Source Disaggregation (as published) Frequency Agency Notes
type

Reports from
Immigration AB Industry MIFOTRA Number of foreigners
Department
Sector surveys ORG /
(ICT, Mining, EB Once 2012 RDB / Skills availability, Skills mismatch
Energy) LMIS
5-yearly (last
Data available to compute some
EICV* HB EICV 3 NISR
indicators on employment equity.
2010/2011)
Population 10-yearly (last Data available to compute some
HB NISR
Census 2012) indicators on employment equity.

Employment Employees’ Module: Formal sector

equity & Industrial employees & persons working in
relations informal sector enterprises - Trade
Manpower survey EB Occupation Once 2012 NISR union membership and collective
bargaining agreement, Not all
employed covered. Did not cover
establishment with only one worker.

Labour MIFOTRA /
AB Province, Industry Variable Industrial disputes
inspections LMIS

Annual Statistical Benefits & Beneficiaries, Pensions,

AB None, Sex, Province, Sex, Province RSSB
Safe work & Bulletin Occupational injuries
Social protection
Labour Age, Sex, Province, Industry, MIFOTRA /
AB Variable Occupational injuries
inspections Occupation LMIS

25
 Data
Group Source Disaggregation (as published) Frequency Agency Notes
type

5-yearly (last
Type of work, Industry, Institutional
EICV* HB EICV 3 NISR All employed, Employees only
sector, Frequency of payment
Poverty of 2010/2011)
employed persons
Annual Statistical Annual (last Monthly income distribution:
EB None RSSB
Bulletin 2010) Contributors only

Labour supply (Other work activities)

Other labour input
into SNA
5-yearly (last
EICV* HB Sex, Age, Sex, Province EICV 3 NISR Domestic work, Inactivity rate
Other non-SNA 2010/2011)
labour input
Population
HB Age, Sex, Residence, Reasons Inactive population
census
Labour demand

Occupation, Industry, Duration,

Manpower Educational Requirements, Education
EB Once 2012 NISR Vacancies, Jobs, Skills
survey* field, Recruitment method, Skills
Vacancies & Jobs required, Jobs created, Provision, Gap

Status in employment, Industry, 5-yearly (last

Jobs created, Multiple jobs, VUP
EICV HB Occupation, Sex, Age, Province, EICV 3 NISR
participants (not analyzed)
Residence 2010/2011)

26
 Data
Group Source Disaggregation (as published) Frequency Agency Notes
type

Sector surveys
(ICT, Mining,
ORG / RDB
Energy, EB Once 2012 Skills gaps in selected sectors
/ LMIS
Manufacturing,
Finance, ...)

Status
Manpower
Costs to employer EB (Permanent/Temporary/Casual), Once 2012 NISR Labour costs
survey*
Industry, Type of expenditures

Establishment characteristics:
Manpower Ownership, Size, Location, Province,
EB Once 2012 NISR All establishments,
survey* Training provided, Formal/Informal
Sectors & Education, etc.

Skills provision, Outputs and

enrolments
Education establishments
Establishment characteristics:
Establishments Establishment Ownership, Size, Location, Province,
EB Once 2011 NISR All establishments
census Registration, Keeping accounts,
Capital, etc.
LMIS EB From MIFOTRA, RSSB & RRA
Quarterly (last
Province, Industry, Institutional sector, LMIS /
RSSB records AB November All registered employers
Size, RSSB
2013)
Size classes different from
Establishment census
Not all establishments covered.
27
 Data
Group Source Disaggregation (as published) Frequency Agency Notes
type

Province, Industry, Institutional sector, Annual (Last All registered establishments, LMIS
RRA records AB LMIS / RRA
Size January 2014) adjusted data
Context
National accounts estimates, CPI,
Economic
EB Industry Annual NISR Expenditure shares, Investment
statistics
share

Sex, Age, Province, School

Child labour Children’s economic activities &
HB attendance, Characteristics of work Once, 2008 NISR
survey Child labour
activities & of Head of household

DHS HB

28
4.2 Assessment of sources of labour market statistics in Rwanda

4.2.1 Source institution: NISR

Data sources: Surveys and Censuses

EICV

The primary survey source for labour market statistics is the Integrated Living
Standards Surveys (Enquête Integrale sur les Conditions de Vie des ménages –
EICV). These surveys have been conducted about every 5 years the first one in
2000/01 (EICV1) to the most recent on in 2010/11 (EICV3). The fourth in the series is
in the process of being conducted in 2014, following the change in periodicity to every
3 years after EICV3. The assessment in this section is based largely on EICV3,
however. Each of these surveys included an extensive employment module covering
aspects relating to

 what used to be called “the economically active population”,

 the categories of this population - employed, unemployed and not active, and
 the main characteristics of these categories.

The economically active population consisted of those of working age who were either
employed or unemployed. In EICV3, the working age population was taken as all
those 16 years old and above, although data was collected for all those 6 years old
and above. As data collected directly from children aged 6 to 9 years as well as data
collected through proxies are known to be of poor quality, the analysis of data for
those below 16 years should be done cautiously. Also, although recommended by
international standards, the use of an open-ended upper limit for age has
repercussions on the values of some estimates based on the working age population,
for example the labour force participation rate.

The identification of persons as employed was based on two reference periods:

 Any work during the last 12 months (with no clear specification of what was
meant by ‘any work’);
 Any work, even for one hour, during the last 7 days.

The analysis of the data that was done made use of both reference periods (EICV3
Thematic Report – Economic Activity) but it gave pre-eminence to the long reference
period. A similar definition was used to collect data in the previous two surveys, but
analysis was done using only the 12 month reference period. It would seem that the
national preference for defining employment is that based on the long reference
period.

The above definition of employment, based on any work during the last 12 months, is
not one of the measurement frameworks in the international standards for measuring

29
this population. It is an all-encompassing definition that gives recognition to any work
that was done for the production of goods and services, as defined in the System of
National Accounts (SNA), over the period of 12 months. So the employed are those
who ‘ever’ did some work during the long reference period of 12 months.

It should be noted that though in the publication of the results this type of employment
is referred to as ‘usual employment’; this is not consistent with the internationally
recognized definition of ‘usual employment’. The latter, which is no longer recognized
by the 19th International Conference of Labour Statisticians (ICLS) in October 2013, is
based on the length of period of employment over the 12 months. A person is ‘usually
employed’ if this length is not less than that of unemployment and the total length of
employment or unemployment is not less than that of being not active. So many of
those classified as ‘usually employed’ in these surveys would not be so classified
according to the international standards which existed then.

The national definition is however particularly useful for analysis of labour input into
production (GDP), as it measures total labour input over the same period (12 months)
used for compiling GDP. Its challenge is in the fact that, as so defined, unemployment,
the flip side of employment, measures ‘no work’, absolutely zero work, over the 12
month period. Consequently the unemployed population is relatively small and the
unemployment rate low.

The 7-day reference period is recognized in international standards and widely used
by countries when measuring employment. The definition of the ‘labour force’ is only in
relation to this period. Moreover it is now the only reference period recognized by the
19th ICLS for measurement of employment. Apart from this variable, albeit a very
important one, the concepts, definitions and classifications used in the EICVs conform
to international standards.

EICV3 captured information on all jobs done by household members over the 12
month period. This is very useful as many characteristics associated with employed
persons in fact relate to their jobs: for example occupation, industry, wages, hours of
work, status in employment and so on relate to the job. In some instances, surveys
limit data collection to what is referred to as the ‘main job’ or extend data collection to
also the secondary and/or tertiary jobs. The EICV is constructed in a way that the
‘main job’ can still be identified, when necessary. The approach used also allows for
analysis of both jobs and persons employed. It is however important not to mix up the
two in the analysis. In some instances in the published report, it is not clear whether
the tables are referring to jobs, to persons employed or to persons employed in their
main job.

The survey is a rich source of data on many aspects of the labour market:
employment, labour underutilization and other forms of inadequate employment, hours
of work, wages and other entitlements and benefits of employees, the job
characteristics mentioned earlier, job search methods for the unemployed and

30
methods used to secure jobs for the employed, special employment programmes like
the VUP and other non-SNA activities of the population. As always, some data
collected were not analysed in the publication, and some data needed were not
collected. Examples of the former are the wage data, the informal employment data
and the VUP data, which were collected but not analysed in the published report. Data
on employment in the informal sector was not collected. As seen from the above table,
EICV3 is useful for producing most of the statistics of labour supply, in particular and is
often the best source.

The results of EICV3 are disseminated in a printed publication using 71 main tables
and 4 diagrams with descriptive analysis of each of these. There are also 11 Annex
tables. An e-copy of the report is available online through the NISR data portal and
public-use micro-data files are available through the National Data Archive (NADA).
No other dissemination mechanisms were used.

Population census 2002

The population census collected data on the economic activities of all those 6 years
old and above. It should be noted also that collecting data from children aged 6 to 9
years through an administered questionnaire does not usually produce good quality
data. It is in fact quite likely that the responses were not given by the children directly
but by proxies. The working age population used in the analysis of the census data
was also those 6 years old and above. This is one of the differences between the
population census and the EICV surveys which could impact on the statistics obtained
from both sources.

In the population census, the economically active population was defined as all those
6 years old and above who were either employed or unemployed. A person was
categorized as employed if for that person the response was ‘employed or temporarily
unemployed’ to the question “During the period from 15/07 to 15/08/2002 did …..
have any gainful employment”. The person was unemployed, presumably, if the
response to the question for the person was ‘first time job seeker or jobless’. This
definition differs from the two used in EICV, and so the results are likely to be affected.
It is also not the same as the then international standard definition of employment.

The main variables for which data were collected in the population census were
economic activity status (active employed, active unemployed and not active) as well
as the following characteristics of the main job: occupation, status in employment,
industry and institutional sector. Used along with other variables in the census, these
statistics relating to labour supply could be disaggregated by many factors such as
educational attainment, sex, age, province, residence etc.

As acknowledged in the thematic report, ‘Analysis of results, Characteristics of the

economically active population’, the population census data is not expected to be of
the same quality as the data from EICV for various reasons. One of these is that the

31
form of questioning used to categorize the population into the activity status groups
(employed, unemployed, inactive) is such that it would have provoked many
misclassifications. Nevertheless, the data is still useful in the following context:

 When the economic activity status is used as a cross-classificatory variable

explaining the observed phenomena form the other census variables;
 When detailed disaggregation statistics are required for geographic variables
(e.g. at sector or village level) as well as for occupation and industry ( to 3 and
4 digit levels);
 When good quality statistics on the economically active population are not
available from other sources.

As highlighted above and in particular for computing many of the related labour market
indicators, the population census is a source of statistics on labour supply. However,
the analysis of these statistics should take note of the above limitations in the data. In
particular, any comparison or joint use of values of these statistics or indicators with
those from EICV should be done with caution.

The thematic report was produced in 2005 for data collected in 2002, a gap of about 3
years. Although this is not unusual for the analysis of economic characteristics from
population census data, it means that the results for those statistics that are volatile,
like the unemployment rate, are largely historic. The results for structural statistics like
the distribution of employed population by industry/occupation would however still
have some relevance. The thematic report and many additional tables can be viewed
and downloaded from the NISR portal and public-use micro data can be obtained from
NADA. Although the population census results are usually released with great fanfare,
this does not always translate to the thematic reports. So awareness of these valuable
statistics by researchers and the population at large may be limited.

Establishment Census

The establishment census took place in 2011 with data collection done in 2 months.
The analysis was based on 91 tables and 10 diagrams. It is worthy of note that the
report was published by August 2011, a remarkable achievement, all the more so as
this was the first establishment census in Rwanda. The establishment census
collected data on the sex and nationality of the owner, working status, place of work,
institutional sector, industry, legal status, registration, type of taxes, keeping of regular
accounts, capital and number of employees according to education, sex, nationality
and type of contract. The objectives were to provide (i) a comprehensive profile of
economic activities of establishments in Rwanda towards establishing a business
register; (ii) statistics based on the above characteristics of the establishment and their
employees; (iii) data for use in classifying establishments by size and identifying them
as being in the formal or informal sectors.

32
A clear definition of an establishment based on international standards (ISIC) was
adopted. However certain establishments were excluded a priori from the compilation.
These were those operating from within households and so not visibly identifiable,
those that were itinerant in nature such as transport drivers, those that were temporary
and roads and buildings construction sites. An important point to note is that the
census was of establishments and not enterprises. Several establishments may
belong to the same enterprise and an establishment may be an enterprise in itself. As
presently constructed, the questionnaire used for the census cannot be used to return
statistics on enterprises, not even the number of enterprises.

The challenge is that since 2011 no effort has been made to maintain the register of
establishments from the census exercise. Some of the identified establishments in the
census register may no longer exist and quite a few new ones may have been
created. So the value of the listing of establishments in the census is decreasing over
time.

Manpower survey

This first manpower survey was implemented in 2011 within the context of a common
EAC project. It was a multi-faceted survey aimed at collecting data from
establishments as well as from employees within these establishments.
Establishments were sectored into (a) formal sector establishments in the private
sector and those in the public sector; (b) informal sector establishments; (c)
education/training institutions. In each of the first two of these sectors, a sample of
establishments was selected and data collected using an employer module. A
separate module was used for the heads of the education/training institutions. In each
of the selected establishments in sector (a), a sample of employees was selected for
direct data collection using an employee module. In the other two sectors, a census of
employees in the selected establishments was enumerated using the same employee
module.

The employer’s module collected data on the establishment’s characteristics, the

characteristics of the employees as reported by the employer, the total number of
posts and number of filled posts, total gross remuneration, number and type of vacant
posts, future manpower projections and staff development. The employee module
collected information on the characteristics of employees as reported by the
employees. The module for heads of education/training institutions collected
information on the labour, current and future enrolment and training output.

The design is the standard design for manpower surveys but the breadth of topics
covered is quite wide. The published reports, in 2 volumes, contain a total of over 750
tables and 212 diagrams. These provide a rich source of statistics on both labour
demand and the employment component of labour supply. As indicated in the table
above, this survey is often the best source for labour demand statistics and for some
groups of labour supply statistics.
33
However the survey did not cover all establishments or all employees. Establishments
with only one worker were excluded, which means most own-account establishments
were excluded. Also the statement “Politicians, military and police forces were not
considered as employees.” in the reports suggest these persons were also not
included. It is not clear whether their establishments were also excluded.

The sampling frames for the establishments were (i) Civil servant census for the public
sector establishments; (ii) 2011 Establishment census for the private formal and
informal establishments; (iii) MINIEDUC Schools database for primary and secondary
schools; (iv) High Education Council database of higher type of learning institutions;
and (v) TVET database from WDA for TVET schools. Different sample designs were
used to select establishments from these frames. Details on weights and estimation
methods are provided as well as precision estimates for key indicators.

The reports are available in print form as well as e-documents in the NISR portal.
Micro data files can be obtained from NADA.

National Child Labour Survey, 2008

The purpose of this survey was to provide estimates on the extent of child labour for
use in developing appropriate strategies for its elimination. It was directed at collecting
data on the economic and non-economic activities of children (aged 5 to 17 years) and
the impact these may be having on their education, health and development. It was a
household based sample survey using the standard two-stage design of enumeration
areas selected with probability proportional to size and households within each
enumeration area selected with simple random sampling. Note children not living in
households were excluded by virtue of this approach. As usual for such surveys,
questionnaires were addressed to adults as well as to children. They collected data on
the characteristics of the household, the head of the household as well as the children.
It was an ad hoc survey to properly measure child labour using the international
standards for collecting statistics on child labour.

Informal Sector Survey, 2006

This first survey of informal sector enterprises was essentially an enterprise survey
directed at generating data mainly about the enterprise itself to serve national account
purposes and decision making about enabling the activities of the enterprises. In this
context data on the size and characteristics of the persons engaged in the sector were
also collected. These included age, sex, nationality, status in employment (owners,
permanent employee, and temporary employee), geographic location, industry and
income. The survey had challenges relating to the definition of the informal sector and
to the lack of any previous experience of such a survey. It excluded ambulant
operators and those operating from within their households, away from any specified
location.

34
4.2.2 Source institution: Rwanda Social Security Board (RSSB)

Data source: Administrative records

Enterprises employing staff are obligated by law to pay social security contributions for
their staff to the RSSB. This process entails submitting data on both the employer
(employer members) and the employees (contributors) at registration and
subsequently every quarter. The data covers the socio-demographic and geographic
location characteristics of the employer members and the contributors, the socio-
security characteristics of the establishments, their size, industry, institutional sector
and occupational injuries, the monthly income of contributors and their nationality.
RSSB itself has data on beneficiaries, amount paid as benefits, and on pension
beneficiaries.

The data is used by RSSB to produce statistics on employer members and

contributors which are submitted to LMIS regularly as quarterly statistics (most recent,
November 2013) and published in their Annual Statistical Bulletin as annual statistics
(most recent, 2009-2010). As noted in the above table, these statistics fit into different
areas of the framework on both the labour supply and labour demand sides. They are
particularly useful with respect to their wages statistics. The major challenge in using
them is that the coverage of establishments and of employees seems to be low when
compared to other sources such as the Rwanda Revenue Authority, the Rwanda
Development Board, the 2011 Establishment Census (although this last covered
establishments and not enterprises) and estimates from EICV. The under-coverage is
estimated by the LMIS as .... for the enterprises and ... for the employees. There are
also issues relating to inaccurate reporting of age and unique identification of
enterprises. The usefulness of the statistics is thus limited by these constraints.
Nevertheless, it can be assumed that trends in their values over time follow the same
pattern as for all enterprises and so reveal good information on growth rates relative to
the latter.

The dissemination of the RSSB statistics is through their printed publication and the
LMIS website.

4.2.3 Source institution: Rwanda Revenue Authority

Data source: Administrative records

All enterprises are obliged by law to pay various forms of taxes relating to the business
and to the income paid to employees of the business, if any. The latter is done through
a Pay As You Earn (PAYE) system in which employers submit details of their
employees who are taxable and the amount of tax deducted at source from their pay.
Employees who earn below RWf 20,000 per month are not taxable and so their details
are not included in the information submitted by the employers. Although individual
records are submitted by employers for each of their taxable employees, these are not

35
captured into the RRA database and are used only for periodic audit checks on
employers. Only aggregate information on numbers of employees and numbers of
enterprises is available. These can be disaggregated by the geographical location of
the enterprises, their industry, institutional sector and the status in employment of the
employees (regular/temporary). However the categorization by industry is not reliable
due to serious and frequent errors in the way enterprises self-classify themselves.
Employees with multiple jobs are also identified.

For own-account enterprises without employees, their tax is based on their profits and
processed through another system, PIT. Details are therefore available on some of the
owner’s characteristics and their profits, which can be taken as their income.

The RRA generates statistics relative to the enterprises and the employees from their
database which are submitted to LMIS for dissemination on their websites. The
statistics that are disseminated are however LMIS-adjusted statistics in order to
remove outliers in the reports submitted by RRA. The usefulness of these statistics
would be enhanced if individual employee records were available for use in the
analysis as their individual characteristics could then have been used to disaggregate
the statistics further. This is particularly important with regard to the income statistics
that could be derived from such the RRA records. Also there is some under-coverage
due to enterprises not registering as required by law and some under-reporting of
employee numbers and income by enterprises wishing to minimize the amount paid as
tax. The extent to which these happen is shown in the LMIS analysis which shows ....

4.2.4 Source institution: Ministry of public service and labour (MIFOTRA)

Data source: Reports of labour inspectors

Labour inspectors, located in every district, send in data collected during inspections
of establishments in their district. The data sent enable the production of statistics on
the characteristics of employees such as wages, education, occupation, age, sex,
nationality, geographical location, type of establishment, industry and access to social
protection as well as on the characteristics of the enterprises, including compliance to
labour laws. One limitation with respect to these statistics relates to coverage of
establishments and their employees. Although the aim is to cover all establishments
listed during the 2011 establishment census, difficulties are experienced tracking
some of the establishments and the workload of the inspectors is heavy. Another
limitation comes from the use of very different reporting periods by the inspectors,
which makes aggregation a challenge. These periods can vary by over 12 months.
Further, the completeness and accuracy of the data collected during inspections may
be affected by deliberate actions on the part of the employers and/or the employees
as well as the amount of time available for the visit.

Data source: IPPS

36
This information system located in MIFOTRA has information on all public servants,
both current and former, relating to their age, sex, field of education, level of
education, contract details, wages, location of their institution, performance and
disciplinary records. It would seem the data is however not used for any statistical
purposes.

4.2.5 Data source: Department of Migration and Immigration

The Department submits statistics on migrant workers on a monthly basis to

MIFOTRA. The tables include classifications by age, sex, nationality, occupation,
education level, income and province. There is some issue about the classification
used for industry between the Ministry and the Department.

The statistics from all the above data sources are submitted for dissemination on the
LMIS website.

4.2.6 Source institution: Rwanda Development Board (RDB)

Data source: records from registration of enterprises

The Department is responsible for registration of new businesses during which

information is collected about persons to be employed in these enterprises. Based on
these records, the Department produces statistics on registered enterprises and on
the characteristics of persons to be employed, including on their nationality. These are
submitted to LMIS for dissemination.

4.2.7 Source institution: Ministry of Education (MINEDUC)

Data source: annual census of institutions

Each year the Ministry collects data at all levels of education. These levels are pre-
primary, primary, secondary, vocational training centres and tertiary and include both
private and public institutions. The data collected includes enrolment at all levels, field
of study at senior secondary level, exits from the secondary system, type of
qualification enrolled for at the tertiary level, staff and key services. Wherever
applicable, the data is disaggregated by sex, type of institution (private, public) and
disability status. Data is also collected on participation in adult literacy. Questionnaires
are used for data collection from institutions at all levels below tertiary whilst data is
collected directly from tertiary institutions.

The annual statistics are disseminated in a yearbook, the most recent of which was in
2013 relating to data for 2012. It contains 42 main tables, 29 annex tables and 20
diagrams. Historic data is used in the yearbook to analyse trends. The statistics are
also submitted to LMIS for use in their website.

37
4.2.8 Source institution: Workforce Development Agency (WDA)

Data source: Administrative records of institutions

Data is collected jointly with MINEDUC covering vocational training centres, technical
secondary schools and post-secondary technical polytechnics. The data covers
enrolment, teaching staff, type of school, infrastructure and equipment. The statistics
produced are disseminated through MINEDUC.

4.2.9 Other institutions

Labour statistics are produced by some other institutions such as the LMIS of RDB
and the Institute of Policy Analysis and Research (IPAR) through targeted topic-
specific surveys. Examples are the surveys on women economic empowerment,
governance frameworks and performance contracts done by IPAR and their school-to-
work transition survey which is now being implemented. LMIS is conducting an
establishment survey on new jobs created over the last 5 years. The statistics
generated from these surveys are used primarily for analysis and are not always part
of the publically available network of labour statistics.

38
Chapter 5: Bridging the Gap

From the previous sections, it is evident there are a plethora of data sources in
Rwanda for the production of labour statistics for all of the thematic groups in the
structure of the labour statistics framework. The challenges in using them relate to

 the incomplete coverage,

 inconsistent concepts, definitions and classifications,
 different data collection methodologies,
 poor quality of some of the sources,
 variability in frequency of collection and dissemination,
 lack of coordination, especially between the producers of administrative
statistics, and
 full exploitation of the potential of administrative sources.

Many of the producing institutions are already taking steps to further improve their
activities for producing quality labour statistics.

5.1 NISR

The institution plans to institute an annual labour force survey from 2016, an annual
integrated business survey from 2014/15, a 3-yearly establishment census from 2017
and to continue its 3-yearly EICV, the current one, EICV4, being conducted now in
2014. These will be in addition to its other survey and census activities that are not
directly related to the labour force. Taken together these surveys and census would go
a long way to bridge the gaps with respect to labour supply statistics and labour
demand statistics.

5.1.1 The Annual Labour Force Survey

The institution plans to start annual labour force surveys as from 2016. As noted
earlier, this is the best source for labour statistics relating to persons in the labour
force, especially labour supply statistics. Properly designed, it has the potential to
provide good quality data for statistics in all groups of the labour supply categories in
the labour statistics framework as well as some of the labour demand groups. In terms
of completeness in coverage, accuracy of measurement and scope of topics, the
labour force survey will be preferable to the EICVs for generating these statistics.
Given the wide topic scope of the latter,

 the range of labour market variables that can be included is limited;

 the concepts, definitions and classifications used to measure the labour market
variables may not be consistent with international standards;
 the training and supervision of enumerators cannot be as rigorous; and

39
  the smaller sample size required to keep interviewing time for all the many
survey topics within reason and resources reduces the precision of some of the
labour market estimates, especially at lower levels of disaggregation.

Although NISR has developed a good experience in conducting household surveys, it

may be useful to initially limit the ambitions of the labour force survey. For example,
the ideal in order to measure stock as well as flow statistics is to use some form of
panel design. Initially, it may be advisable to implement a one-off sample survey but
with the sample subdivided into 2 nationally representative sub-samples implemented
in 2 semesters to capture seasonality. Then, over the latter few years, the periodicity
can increase to quarterly or monthly nationally representative sub-samples. The
ultimate objective of a continuous survey using a panel design should be deferred for
now.

The survey should as much as possible use concepts, definitions and classifications
consistent with international standards, particularly with respect to the key and basic
variable of labour force status (employed, unemployed and out of the labour force). In
this regard, it will be necessary to take into account the recently adopted international
standards for the measurement of all forms of work. However, in order not to overload
the questionnaire on an annual basis, it may be useful to use a rotating modular
approach to cover some of the forms of work. A core module for employment-work,
and possibly work for own-use production of goods, should be implemented each
year. The definition of employment as an unspecified ‘any’ work over a 12-month
period used in EICV should not be adopted for the labour force survey, nor should that
of the 2002 population census.

In addition to the core labour force status variable, data should be collected to
measure labour underutilization and its components of time-related underemployment
and unemployment and the potential labour force. Measuring other forms of
inadequate employment would also be useful as a proxy for low individual labour
productivity. The survey should collect information on the characteristics of the
employed including the standard ones of industry, occupation, status in employment
and institutional sector. It should extend this to cover benefits and entitlements so as
to measure informal employment and conditions of work of the employed population.

Two important but challenging topics should be measured as best as possible. These
are employment in the informal sector and income from employment. For the former,
the questionnaire and survey design should provide for adequate information on the
characteristics of persons engaged in such enterprises over and above on the
characteristics of the enterprises themselves. This is on the assumption that the policy
issue of prominence is employment in informal sector enterprises. In EICV and the
2006 Informal sector survey, this was not the case. Greater emphasis was given to the
characteristics of the enterprises such as output, financing mechanisms, operating
characteristics and support mechanisms. It should be noted, however, that when
approached from the position of persons working in informal sector enterprises as
40
proposed earlier, there is a risk of double counting of enterprises. The design used
should therefore depend on the purpose of the statistics required. If interest is in the
informal sector enterprises, such as estimating the number of such enterprises, the
design used in EICV and the 2006 Informal sector survey is preferable. Accurate data
to measure income from employment, especially from self-employment, can be difficult
to collect in a labour force survey. It is however necessary to do so, conscious of the
limitations of the data collected. A useful approach is to use different methods for
income from paid employment and for income from self-employment. The former can
be analysed with much greater confidence than the latter. In some instances, a
household enterprise module is used to obtain the income of the self-employed as the
mixed income of their enterprises.

Other topics that can be covered rotationally are data on occupational injuries and
diseases, membership of trade union and employer’s organization, pension receipt,
access to employment services and experiences in industrial relations.

5.1.2 The annual integrated business survey

Establishment surveys have been acknowledged above as the best source for
statistics in all the thematic groups of the labour statistics framework on the demand
side. This is demonstrated by the wealth of labour market data from the 2012
manpower survey that can be used as indicated above for these groups. The
proposed integrated business survey could therefore provide the necessary data for
these groups on an annual basis from the best source for such data. Although it will
not be at the same level of detail and scope as the manpower survey, it is expected
that sufficient information on persons employed in the selected businesses will be
available. This is one of the major challenges in integrated surveys dealing with many
topics, as discussed in the previous section for the EICV.

The establishment can only provide aggregated data on its employees such as their
distribution by sex, nationality, occupational groups (possibly), training provided,
qualification levels, income groups (possibly levels) and so on. In addition, information
can be provided on variables common to the business and the employed persons
such as industry, institutional sector, formal/informal categorization, geographic
location. The establishments may also be able to give information on vacancies and
newly created jobs, but the disaggregation of these may be limited. Information on
training provided and missing and mismatched skills would be useful but may be
challenging. Statistics on the establishment itself such as size, type of ownership,
characteristics of owners, output and plans for future expansion should be available.

It may be useful to consider a modular approach in which modules on specific topics

are attached to the core rotationally. Given there are no plans to repeat the manpower
survey, it may be possible once every 3 to 5 years to attach a manpower-like module
along the lines of the employer’s module of the 2012 manpower survey to the core
establishment questionnaire. In addition a sub-sample of establishments can be

41
selected once every 5 years so that a shortened form of the employee module can be
administered to a sample of employees.

For the annual integrated business survey to be successful, careful attention should
be paid to maintaining a good establishment register. This would require close
collaboration between NISR and other institutions interested in using such a register
and possible use of institutions operating at district and sector level. It may be useful
to consider supplementing or complementing these activities with some form of area
listing of establishments.

The 2012 manpower survey excluded establishments with only 1 worker. The reason
for doing so was however not presented in the reports. It may be useful to review the
decision for the integrated business survey, as otherwise many own-account
establishments will not be included in the survey. The survey also excludes ambulant
businesses and those operating from within the household with no fixed identifiable
location elsewhere. Whilst this is understandable in terms of the limitation of the frame
of establishments, it may be useful to use other ways to get similar information about
the omitted enterprises; for example a module on household enterprises attached to a
labour force survey or to EICV.

5.1.3 3-yearly establishment census

The 2011 establishment census was well designed and successfully implemented,
although it was quite a hectic exercise over a period of about 6 months. Repeating this
activity every 3 years may however be quite demanding in terms of human resources,
given the other old and new activities being planned by NISR. A 5-year periodicity may
be more attainable. The data collected in 2012 could not be used to link up
establishments with their parent enterprises. Having information on the number of
enterprises in the census could serve as check for under-coverage of data from
institutions collecting data at the enterprise level, e.g. RSSB and RRA.

5.1.4 3-yearly EICV

With the introduction of the annual labour force survey, the employment module in
EICV should be much reduced but not completely eliminated. It will still be necessary
to have data to categorize the population by their labour force status. This will enable
the cross-classification of labour force status against other variables from the survey
such as the poverty status of households. This is in fact the only source for producing
good quality statistics on the working poor, an MDG indicator. In addition, it would be
necessary to also continue using the ‘at least 1 hour in 12 months’ measurement
method for employment to link up the results with the previous EICVs until a sufficient
number of overlapping labour force surveys have been conducted.

The analysis of the EICV4 should give pre-eminence to the 7-day reference period
over the 12-month period to produce statistics according to international standards.
Nevertheless, the relationship between them should be examined by analysing a table
42
of employment statuses classified by the 12-month framework versus the 7-day
framework. This would give the basis to better understand the implications of using
one or the other. An issue that would need to be resolved is the extent to which the
analysis of characteristics such as occupation, industry etc. should reference the job
done in relation to the 12-month period or to the 7-day period. EICV3, as the other
EICVs before this, lays emphasis on the former. The analysis of the data should also
clearly distinguish between statistics based on jobs and statistics based on employed
persons, both of which can be done from the data collected.

5.1.5 Dissemination

NISR uses printed publications and websites to disseminate its results. Micro data
from the surveys and censuses can also be accessed through NADA. Yet, from the
user satisfaction survey conducted during the preparation of NSDS2, many users
expressed dissatisfaction in particular with access and availability of NISR
employment statistics. This disjuncture between the actual situation and the one
perceived by users may be due to lack of awareness by the users of the existence of
the results, despite the release timetable on NISR’s internet homepage. The existing
methods are passive and depend on users going out to seek the results by obtaining
publications or consulting the website. Also, although there are news flashes of NISR
statistics on its internet home page, it seems none are on employment statistics. A
different strategy may be needed to push out employment statistics to users. One
possibility is to actively and continually engage the media on the results that are
available and on the implications of these results for the way of life of Rwandese. Not
all the results have to be pushed out at the same time. Emphasis should be on
continuing presence in the media by slowly releasing the results from the same source
over time.

The use of results published on the NISR website and micro data from the surveys
and censuses should be monitored in order to assess the extent of usage of the
statistics. In particular, it may be useful for NISR to engage directly with users,
including the media, policy-makers, researchers and politicians,

 to promote a good understanding of the results;

 to improve awareness of those that are not being consulted; and
 to encourage further analysis of the available data.

5.2 RSSB

Normally, social security institutions are a good source of statistics to monitor trends in
wages. Although they have limited coverage vis-a-vis the employed population as a
whole, the trends in the wages of their contributors may mimic that in the wider
population. They are also a good source of the number and characteristics of
beneficiaries and benefits, especially of pensions and claims resulting from
occupational injuries. The limitations in their use come from the under-coverage of

43
enterprises, the under-reporting of employees to minimize contributions. These are in
fact precisely the limitations in the statistical production activities of the RSSB. This
affects the usefulness of the statistics produced for general analysis of the labour
market. An improvement in coverage and reporting would benefit both RSSB and the
labour statistics system.

5.3 RRA

The data collected have the same limitations as those of RSSB with respect to under-
coverage and under-reporting, but probably less so. The data processed is however
limited to those aspects of direct interest to the tax authorities, which is surprising
since reportedly the data are submitted electronically. It would seem that individual
data records of employees submitted by employers under the PAYE system are not
used to generate statistical tables, even though they are collected. They are used
principally for audit checks. This severely limits the possibilities for statistical analysis.
Another challenge in using the RRA data, as well as those of RSSB, is in the
classification of enterprises by industry. The quality is poor. The existence of a
national coding index for industries should greatly improve this classification. Other
challenges are that not every employee is taxable, some enterprises may be
exempted from national business tax (they only pay local tax at district level, for
example) and there is a tendency to under-report income.

Despite the above challenges, RRA administrative sources could provide good data
on the number of paid employees, the number of the self-employed and their income
(levels and distribution) classified by various person and enterprise characteristics.
These are useful for producing indicators such as low-income incidence, inequality of
employment income, average income of persons at very low levels of disaggregation
by geographic location, occupation and industry. The availability of these data at the
required level would go a long way to filling the gap between the data required for
some labour demand statistics and the data available to do so. Means should be
found to persuade RRA to produce statistical tables using individual records of
employees in order to profit from their data to the maximum. One possibility is for
NISR to work with RRA to develop an appropriate software application for generating
the required tables from the individual records. Then the demands on RRA to actually
produce the tables will be routine and much reduced.

The introduction of a unique PIN for businesses across various data producing
institutions should positively impact on coverage both for RRA and RSSB.

5.4 MIFOTRA

To contribute to bridging the gap between what is available and what is desirable the
labour inspector reporting system needs to be upgraded. Steps that could be taken
include increasing their number, planning their operations to have reports sent in at

44
regular intervals, streamlining their operations, developing common understanding of
the variables on which they are reporting to ensure consistency across reporters, etc.

As the proposed secretariat of the coordinating unit for NEP, the Monitoring &
Evaluation unit of MIFOTRA should develop simple project and programme monitoring
indicators that can be used at the most disaggregated geographical level. The
country’s administrative structure has a presence at village level that could be used to
quickly collect local level data for these indicators. A mobile telephone based system
could be developed for fast transmission of data on key variables. For this purpose,
the indicators selected should be simple with a limited number of response options.
NISR, as the coordinator of the national statistical system, should work with the
MIFOTRA and the Ministry of local government to develop these indicators and the
reporting system. A pilot study will be necessary.

The rich data set on public servants, IPPS, is not exploited for statistical purposes.
Even though coverage is limited to public servants, the analysis of the data in this
system could still provide useful insights on the employment situation in Rwanda. The
Ministry should therefore make an effort to analyse the data.

The Ministry uses LMIS as its dissemination mechanism for the statistics produced
from the records of labour inspectors and from the submissions of the Migration
department. It would be useful to have some publication, like a yearbook, in which
some analyses of the statistics are presented, along with the proposed analysis of
IPPS data.

5.5 RDB

Enterprise registration data are useful for maintaining the establishment register and
aiding coverage by RSSB and RRA. NISR, RSSB and RRA should examine the
means of doing so. The data provided on employees are not of mush use beyond that
of the institution itself, if at all.

5.6 MINEDUC

It is important to continue with the development of the computerized student and

teacher system as this will facilitate the production of statistics on enrolment, exits and
qualifications. The Ministry should collect and process data on field of study at VCTs
and other TVET institutions to provide statistics on the nature of skills availability for
these levels. Output data at all levels, both successful exits and drop-outs, are
important and should be provided.

5.7 WDA

The planned tracer study should provide important information on skills mismatches,
provided it is properly designed and managed. The Authority should endeavour to
repeat it every 3 to 5 years.
45
5.8 Other Institutions/sources

Data from the targeted surveys and other data collection activities should be treated
as public and made available to all who wish access to it. NISR, as the agency giving
visas for these activities, could make this a condition for issuing these visas.

LMIS and IPAR should work closely with major producers to improve primary data
collection. Other partners in the labour statistics framework should be kept informed of
plans for targeted surveys either by NISR as the visa issuing authority or by the
implementing institution.

5.9 National Statistics System (NSS)

The NISR, as the coordinating institution for the NSS, should put in place adequate
coordinating mechanism for labour statistics across the major producers. This could
build on the recent joint workshops and meetings that have been organized in the
context of the development of NSDS2, introduction to the new international standard
for measuring employment and the development of the labour statistics framework.

The mechanism should be at informal, technical and policy-making levels. Informal e-

contacts could be encouraged through an exchange of email addresses and other
electronic contact information between focal points (statisticians) from the major
producers. This would enable quick exchange of information on activities, mutual help
in the implementation of these activities and quick responses to demand for statistics.
This informal system should be supported by regular formal meetings of this technical
group, possibly every 6 months. Then, depending on needs, this technical group
should report to a meeting of senior policy-makers of their respective institutions.

There are many person, household, posts and enterprise characteristics that are
common not just across the labour statistics system but the entire NSS. These relate
to concepts, definitions and classifications used across the NSS. It may therefore be
useful to develop national equivalents for use by all producers in the form of a
dictionary of concepts, definitions and classifications of labour market-related
statistical variables in Rwanda. In particular with respect to the classification of
industry and occupation, the development of coding indexes could have a positive
impact on accuracy, consistency and ease of coding of these variables.

46
Chapter 6: Programme for the Production of labour market
statistics for the RLSF (2014/15 – 2019/20)

Part I: Situation Analysis

The EDPRS2 has identified job creation as crucial, especially for the growing youth
share of the population. Further, policies and programmes that have evolved out of
EDPRS2, especially the National Employment Programme (NEP) awaiting final
approval, have objectives which depend on the availability of good labour market
statistics. This importance of labour market statistics is consequently highlighted in
the National Strategy for the Development of Statistics (NSDS2). It however notes
that there are still some lacunae in the statistics required to make evidence-based
decisions, even though a wide range of these statistics are being produced in the
National Statistical System (NSS). In response to this and as a follow up to the
proposal in NSDS2, NISR as the coordinator of NSS embarked on the development
of a “Labour Statistics Framework for Rwanda (RLSF)”.

The proposed RLSF, whilst recognizing the particular needs of EDPRS2 and NEP, is
based on an all-encompassing approach to organizing labour statistics in Rwanda. In
this light, labour market indicators have been identified, prioritized and described in a
Metadata Handbook within the context of a structure for the statistics required to
compute them. The choice of indicators also takes into account the reporting
obligations of Rwanda to regional and international systems such as the MDGs, East
African Community and the International Labour Organization. The RLSF examines
the ideal data sources for these statistics, identifies and assesses the existing data
sources and recommends ways of moving from the current towards the ideal through
the achievable. Amongst the challenges noted in the RLSF as faced by the existing
sources are:

 incomplete coverage,
 inconsistent concepts, definitions and classifications,
 different data collection methodologies,
 poor quality of some of the sources,
 variability in frequency of collection and dissemination,
 lack of coordination, especially between the producers of administrative
statistics, and
 poor exploitation of the potential of administrative sources.
The programme for the production of labour market statistics is the first step in
implementing these recommendations. The programme responds to the identified
lacunae in the existing labour statistics system along the lines of Strategy 1.2 and 1.3
of NSDS2. It recommends (i) the use of a judicious mix of surveys, censuses and
administrative sources to meet the above data needs; (ii) aggressive dissemination of

47
the results; and (iii) organization of the labour statistics system within the NSS, as
proposed in NSDS2.

The SWOT analysis done in NSDS2 also applies to this programme. During the
recent years, NISR and partner institutions have successfully carried out a number of
important statistical exercises, which has built up considerable confidence and
experience of their specialists in the implementation of household surveys and use of
administrative sources. However, the staffing and other resources required to take on
additional surveys and other activities are a challenge. The 2012 user-satisfaction
survey identified the dissemination of labour statistics as the weakest amongst all the
statistics systems. At the same time, the approval of NSDS2, the successful
implementation of NSDS1, the existence of national policies and programmes such
as EDPRS2 and NEP as well as the commitment and enthusiasm of their staff have
created windows of opportunities for the successful implementation of the
programme. Provided the resources are available and the institutions do not over-
extend themselves, any threats to implementation of the programme will be
containable.

Part II: Sub-Programme A – Production of labour statistics from Surveys

and Censuses (2014/15 – 2019/2020)

II.1 The annual labour force survey (Responsible agency: NISR.

Collaborating agencies: MIFOTRA, MINICOM)

As already indicated in NSDS2, NISR has decided to institute an annual labour force
survey (LFS) as from 2016.

The LFS will be wide-ranging covering topics through both an annual core module for
employment work and own-use production of goods and specialized 2 to 3-yearly
rotating modules for other forms of work and some other topics, as follows

(iv) Questionnaire (v) Topics

(vi) Core (i) population; labour force status; education and training;
(ii) employment and labour underutilization, including unemployment,
underemployment and the potential labour force; Other inadequate
employment situations;
(iii) Characteristics of the employed such as industry, occupation, status
in employment, institutional sector, hours worked, employment in the
informal sector, informal employment; income from employment
(main and secondary jobs)
(iv) Participation in own production of goods
(vii) Module 1 (iv) other forms of work such as unpaid trainee/apprentice work, volunteer
work and own-use production of services (domestic work)
(viii) Module 2 (v) work experience of the employed such as method used to find a job;
(vi) access to employment services and characteristics of previous job of
the unemployed and those out of the labour force.
(ix) Module 3 (vii) safe work (occupational injuries and diseases)

48
The LFS will as much as possible use concepts, definitions and classifications that
are consistent with international standards, especially for the basic labour force
status variable. When designing the survey, account will be taken of the new
international standards for measuring employment.

Preparatory activities for the survey will start in 2015 with field operations
commencing in first half of 2016.

For the first 3 years of implementation, the sample will be divided into 2 nationally
representative sub-samples with the sub-samples being administered 6 months apart
to account for seasonality. Depending on experience gained, the frequency will then
be increased to 4 (quarterly sub-samples) to produce quarterly national level
estimates of key indicators. This increase may require a reduction in the above scope
of the survey so that results can be processed in time and reports produced. One
possibility could be to have the smaller surveys during 3 quarters and to include the
omitted topics once a year in the 4th quarter. There are several country examples that
could be studied in making these decisions. Sometime much later, when NISR has
gained sufficient experience in using panel designs from the EICV, the LFS design
could upgrade to a continuous panel survey.

II.2 The 3-yearly EICV (Responsible agency: NISR. Collaborating agencies:

MIFOTRA, MINICON, PSF)

NISR intends continuing the EICV on a 3-yearly basis starting from EICV4 in 2014.
With the introduction of the annual labour force survey, the employment module used
in these surveys (up to the current EICV4) will be reduced, but not eliminated. Some
information on the basic labour force status of individuals; the employed and
unemployed population; and the characteristics of the employed population such as
industry, occupation, status in employment, institutional sector; income from
employment (for all jobs) will still be collected. This is particularly important to be able
to cross-classify labour force status with other survey variables, such as poverty
status of households. The EICV is in fact the only survey source for producing good
quality statistics on the working poor, an MDG indicator. Its usefulness is also in
terms of generating the above data for all jobs, which will not be done in the LFS.
This permits the production and analysis of statistics both in terms of employed
persons and jobs.

The EICV will continue using both the 7-day reference period and the 12-month
reference period in measuring employment. This will ensure continuity with the
previous surveys so as to carry out comparative analysis of the variables of interest
over time, whenever necessary. The definition of employment will be in relation to
employment work only, following the recent International Conference of Labour
Statisticians’ standard for measuring employment as only income-generating work.
Consequently the same goes for jobs. It will be expected that this new definition will
impact on the estimates of employment produced and so raise questions of
49
comparability with previous EICVs. However, data on other forms of SNA work
activities should still be collected, at least for the first 2 or 3 EICVs, in order to overlap
with the old definition of employment. The definitions of the working age population
and other labour market variables, including classifications, will be consistent with
international standards. As much as possible, the analysis will be on jobs, as this is
the main strength of this source.

Both the EICV and the LFS will inevitably be organized in the same year on some
occasions, i.e. 2017 and 2020. Since all the topics in the EICV are also in the LFS
there will be two sets of estimates whenever this happens. Almost surely, there will
be differences in the estimates produced from these two sources, even though they
will be using the same definition of the working age population, employment and
unemployment. The analysis and dissemination processes will need to be carefully
managed to avoid confusing users and raising questions about one set of estimates
or the other. In so far as it concerns persons, the LFS estimates are the best and will
be the only ones recognized as ‘official estimates’; for example the official
unemployment rate will be that produced from the LFS. In instances when the
variables used in cross-classification or other forms of analysis are only available
from the EICV, then the EICV results relating to these variables will become ‘official
estimates’. For example, when analysing working poverty, the resulting EICV results
as they relate to the poverty of working persons will be official. In addition, if both
estimates are disseminated NISR will explain the reasons for any differences.

II.3 The yearly Integrated Business Survey (Responsible agency: NISR.

Collaborating agencies: MINICON, MIFOTRA)

NISR has already designed this survey and it is in the pilot stage. The first full survey
will be implemented in 2015 and thereafter annually. It is expected to yield data on
jobs (persons) in the sampled establishments, the characteristics of these jobs such
as industry, location, hours worked, earnings as well as the characteristics of the
establishment. As noted in the RLSF, it is the best source of these and other labour
demand statistics.

In the absence of any follow-up plans for another manpower survey, in addition to the
survey as planned, a module will be attached to the IBS in 2017 to collect data on
training provided, vacancies and their characteristics, labour costs etc. from
employers (similar to some of the questions in the employers’ module in the 2011
manpower survey). It could also collect data directly from a sample of employees in
sampled establishments, on training and skills utilization. In the competing demand
for space and time, these factors will not be part of the regular survey.

The success of this survey depends crucially on the quality of the sampling frame
and so is linked to projects II.4 and II.7.

50
II.4 The 3-yearly Establishment Census (Responsible agency: NISR.
Collaborating agencies: RSSB, MIFOTRA, RRA, RDB)

The establishment census is being implemented in 2014 and thereafter it will be

repeated every 3 years. The periodicity of 3 years will be useful in terms of
minimizing the work required to maintain the register of establishments. It is however
quite short, considering the many other survey activities planned over the next 5
years. Depending on the experience gained, it may be necessary to consider using a
periodicity of every 5 years with intense efforts put into keeping the register up to
date through other means as described in project II.7.

II.5 The 5-yearly Sector surveys (Responsible agency: RDB. Collaborating

Agency: NISR, MIFOTRA, WDA, Relevant sector ministry, NCBS)

The sector surveys produce useful data on skills availability, skills shortages and
mismatches. Another round will be implemented in 2017, 5 years after the last series
in 2012.

II.6 The 5-yearly Tracer Studies (Responsible agency: WDA. Collaborating

agencies: NISR, MINEDUC, RDB, NCBS)

These studies provide very good information for training institutions, if properly
designed. The current study is taking place in 2014. Another one will be done in
2019. The second one will build on the experience gained from the first survey in
2014. Its success would depend on lessons learnt from the survey and design
methodology used in 2014. It will be extended to include university graduates.

II.7 The National Statistical System (Coordinating agency: NISR.

Collaborating agencies: All major producers)

II.7.1 Establishment register maintained.

Either during or immediately following the establishment census in 2014, NISR will
organize an inter-agency team including RDB, RSSB, RRA, MIFOTRA, MINALOC to
draw up plans on how to jointly exploit their various sources to maintain the business
register. District statisticians and labour inspectors will be invited to collaborate with
the Team with respect to their own possible inputs into the exercise. The Team will
also explore the possibility of using local tax collectors to provide information on new
establishments at local level and on the ones that no longer exist. The working group
will use all relevant material on how best to do this from partner countries that already
have such registers as well as from regional and international agencies. In particular
the manual on business registers developed by the African Development Bank
(AfDB) will be studied. Support for this activity will be sought from AfDB.
Implementation may involve some study visits to countries which could be identified
in consultation with the AfDB. This activity will start in 2015.
51
II.7.2 National classification of industry

Work that was started to develop this classification will be reignited in 2014 to
develop a Rwanda classification of industry consistent with ISIC, rev. 4. This work
transcends labour statistics and must bring in partners from economic statistics and
relevant sectoral ministries dealing with industry data. If necessary some external
technical support will be sought from the UN Statistics Division. Once the
classification is established, work will start on the development of a coding index to
facilitate the coding of industry both in surveys and in administrative systems.

II.7.3 National classification of occupation

Like the industry classification, the development of a national classification of

occupation that is consistent with ISCO-08 will be started in 2015. The work will be
undertaken by a technical group including all parties involved in using this kind of
classification, especially education and training institutions. Technical support will be
sought from ILO, if necessary. A coding index will be developed as soon as the
classification has been created.

II.7.4 Rwanda Dictionary of Labour Statistics

The labour statistics sources have many variables in common amongst themselves
as well as with other statistics system. In order to promote some level of
harmonization of statistics across these sources, as a first step, an exercise will be
undertaken to develop and agree on national definitions of some key labour market
variables. These may include definition of household; headship of household;
standard age groups; classification of establishments as micro, small, medium and
large; definition of employment and unemployment; definition of time-related
underemployment and potential labour force; and so on. Nothing precludes the
national standards being the same as the international standards for some variables.
In any case, as much as possible, all of the definitions should be consistent with the
international standards, even when they are not exactly the international standards.

The group of statisticians from different agencies who have been participating in the
NISR workshops relating to employment statistics will be constituted as the Technical
Group on Labour Statistics to implement this activity. Their method of work will
initially be informal networking through use of electronic means such as exchange of
emails, mobile telephones, etc. NISR will take the lead in promoting these
exchanges, starting in 2014.

II.7.5 Rwanda Labour Statistics Council

The possibility of creating a formal body at policy level of all major producers and
users of labour statistics will be examined. This body will be the sectoral Sub Working
Group referred to in NSDS2. It will be part of the NSS, reporting to the Steering
52
Committee of NSDS2. It will receive technical support from the Technical Group
created in II.7.4, which will itself then take on a formal role in this capacity. It is
expected that this system will greatly improve coordination across the RLSF, avoiding
overlaps, duplications and gaps. The study of this possibility will start early in 2015
and a decision taken before the end of 2015.

II.7.6 Labelling of official labour statistics

In its capacity as coordinator of the NSS and the institution responsible for giving the
go ahead to non-NISR survey activities through the visa system, NISR will consider
the need to go one step further by accrediting certain statistics as ‘official Rwandan
statistics’ provided they meet certain criteria. This will not be an attempt to prevent
other statistics being produced and disseminated, but one directed at assuring the
public at large, especially those outside Rwanda, of the quality of the statistics being
disseminated to them. Accreditation will be at the request of the producer of the
statistics. The system, currently used in the UK statistics system will be studied in
2015 and a decision made by mid 2016.

Part III: Sub-Programme B – Production of labour statistics from

Administrative Sources (2014/15 – 2016/17)

III.1 Statistical tables based on RRA employee records (Responsible agency:

NISR. Collaborating agency: RRA, MIFOTRA/LMIS)

Data on individual employees are submitted electronically to RRA by employers in

connection with PAYE tax payments. These data are however not used to produce
statistical tables based on the characteristics of these tax payers. Thus only a few
basic tables of taxpayers by industry, location and institutional sector are
disseminated.

In the project, NISR will work with RRA and other collaborating agencies to develop a
software application for constructing tables based on RRA individual taxpayer
records. RRA can then routinely use this application to generate a wide range of
statistical tables. This project will start in the second half of 2014.

III.2 Unique identification number for enterprises (Responsible agency: NISR.

Collaborating agencies: RRA, RSSB, MIFOTRA, RDB)

Activities are already being implemented to develop a system for assigning unique
identification numbers to enterprises. The project will give impetus to this procedure
so that it can achieve results soon. It is expected that RSSB and RRA will then be
able to take necessary steps to improve their coverage of enterprises and
employees. The project will end by mid-2015.

53
III.3 Labour inspector census module (Responsible agency: MIFOTRA/LMIS.
Collaborating agencies: NISR, MINALOC)

The project will take steps to use standard reporting period (quarterly) for all data
submissions from the labour inspectors. A training workshop will be held with labour
inspectors to (a) improve their understanding of the concepts and definitions that will
be used when recording information in establishments; (b) agree on suitable methods
for extracting the information from the enterprises. This will start in 2015.

III.4 IPPS data (Responsible agency: MIFOTRA. Collaborating agency:)

The IPPS data will be used to regularly, possibly quarterly, produce statistics on
employment in public services disaggregated by socio-demographic characteristics,
occupation and income. The start date for this project will be second half of 2014.

III.5 Local level data for M&E of NEP projects (Responsible agencies:
MIFOTRA, MINALOC. Collaborating agency: NISR)

As the host of the Secretariat for NEP, MIFOTRA working with MINALOC will develop
a system for quickly accessing local level monitoring and evaluation data. With
guidance and assistance from NISR and LMIS, suitable indicators will be identified
depending on the nature of the NEP project. These will be such that data responses
can be easily transmitted through a mobile reporting system.

The mobile reporting system will be developed with assistance from LMIS. The
reporters at local level will be local tax collectors or some other local staff of
MINALOC. The information will be obtained through a mix of observation and use of
key informants. Training sessions will be held for these persons to have a common
understanding of what will be reported and how. Given the possibility of the phones
attracting unwanted attention or used for other purposes, the mechanism will be
developed in collaboration with a selected service provider so that the phones are
dedicated phones useful for only this purpose. Alternatively, only very cheap mobile
phones will be used. The system will have a computerised system for direct receipt of
transmitted messages and immediate processing of the information across reporters.

The project will start in the latter part of 2014 with the objective of going operational
by the second half of 2015.

III.6 MIFOTRA & LMIS dissemination system improved (Responsible agency:

MIFOTRA/LMIS. Collaborating agency: NISR )

The LMIS is the vehicle for disseminating statistics produced by a wide variety of
institutions, including MIFOTRA. The development of the web-pages of the LMIS
website that are currently showing they are still under development will be completed
within 6 months or taken down from the site. As much as possible, more elaborate
54
metadata will be made available for the indicators that are being disseminated. These
activities will be completed by end 2014.

MIFOTRA will take steps to have a regular printed publication with some analysis of
the indicator values being disseminated. This will be either an annual publication or
quarterly write-ups on specific issues that are current. The publication will be made
available on the Ministry’s website in pdf format. The first publication will be
disseminated by mid-2015.

III.7 Increased data on skills availability from the education and training
system (Responsible agency: MINEDUC & WDA. Collaborating agency:
NCBS)

Through the newly developed computerized reporting system for all education
institutions, data on fields of study will be made available for enrolled students and
successful exits at all post-junior secondary levels including VTCs and other TVETs.
Exit and transmission data at each level of the pre-primary, primary, secondary and
tertiary levels will be disseminated as a guide to the future availability of skilled
persons from the system. This will be an annual activity as from 2015.

Part IV: Conclusion

If successfully implemented, the Programme will yield the following outcomes:

(a) Annual sets of comprehensive labour statistics available for the monitoring of
national employment policy and programmes, including in particular the
assessment of trends in the creation of jobs, working poverty, shifts in industry
sectors, etc.;

(b) All results publically available in printed material and the web as well as through
regular media events and statistical literacy seminars;

(c) National and international users, investors, entrepreneurs and individuals with
access to reliable, consistent and up-to-date annual labour statistics based on
international standards which will strengthen the overall data credibility and
facilitate international comparisons and analysis.
Measuring the programme’s implementation progress will be accomplished
using standard monitoring procedures. The project will be subject to evaluation
in accordance with the policies and procedures established for this purpose

55
Annex 1: Work Plan and budget
Activities Responsib Collaboratin 2014 2015 2016 2017 2018 2019 2020 Total cost
le Agency g Agencies (RWF
thousands)
No. Description
Part II: Sub-Programme A – Production of labour statistics from Surveys and Censuses
II.1 The annual labour NISR MIFOTRA,
force survey MINICON

II.1.1 Preparatory
activities
II.1.2 Implementation

Annual budget estimates (RWF thousands) 65,000 730,800 750,000 770,000 790,000 810,000 3,915,800
II.2 The 3-yearly EICV NISR MIFOTRA,
MINICON,
PSF
Annual budget estimates (RWF thousands) 1,477,300 3,649,900 3,709,900 8,837,100
II.3 The yearly Integrated NISR MINICON,
Business Survey MIFOTRA,
PSF, RDB

Annual budget estimates (RWF thousands) 230,000 241,500 253,600 266,300 279,600 293,600 1,564,600
II.4 The 3-yearly NISR RSSB,
Establishment Census MIFOTRA,
RRA, RDB,
PSF

Annual budget estimates (RWF thousands) 534,200 550,900 567,600 1,652,700

56
 Activities Responsib Collaboratin 2014 2015 2016 2017 2018 2019 2020 Total cost
le Agency g Agencies (RWF
thousands)
No. Description
II.5 The 5-yearly Sector RDB NISR,
surveys MIFOTRA,
WDA,
Relevant
sector
ministry,
NCBS
Annual budget estimates (RWF thousands) 100,000 100,000
II.6 The 5-yearly Tracer WDA NISR,
Studies MINEDUC,
RDB,
NCBS
Annual budget estimates (RWF thousands) 180,000 180,000 360,000
II.7 The National NISR All major
Statistical System producers

II.7.1 Establishment register

maintained

Annual budget estimates (RWF thousands) 6,500 5,000 5,000 5,000 21,500
II.7.2 National classification
of industry:
Alphabetical
classification index
Annual budget estimates (RWF thousands) 2000 2,500 60 20 4,580
II.7.3 National classification
of occupation:
Alphabetical
classification index
Annual budget estimates (RWF thousands) 2000 3,000 60 20 5,080

57
 Activities Responsib Collaboratin 2014 2015 2016 2017 2018 2019 2020 Total cost
le Agency g Agencies (RWF
thousands)
No. Description
II.7.4 Rwanda Dictionary of
Labour Statistics
Annual budget estimates (RWF thousands) 1000 1,000 2,000
II.7.5 Rwanda Labour
Statistics Council

Annual budget estimates (RWF thousands) 200 200

II.7.6 Labelling of official
labour statistics
Annual budget estimates (RWF thousands) 2,000 2,000 4,000

Part III: Sub-Programme B – Production of labour statistics from Administrative Sources

III.1 Statistical tables NISR RRA,
based on RRA MIFOTRA
employee records / LMIS

Annual budget estimates (RWF thousands) 60 300 360

III.2 Unique identification NISR RRA,
number for RSSB,
enterprises MIFOTRA,
RDB
Annual budget estimates (RWF thousands) 120 40 160
III.3 Labour inspector MIFOTR NISR,
census module A / LMIS MINALOC

Annual budget estimates (RWF thousands) 4,500 4,500

III.4 IPPS data MIFOTR
A
Annual budget estimates (RWF thousands) 300 300 600
58
 Activities Responsib Collaboratin 2014 2015 2016 2017 2018 2019 2020 Total cost
le Agency g Agencies (RWF
thousands)
No. Description
III.5 Local level data for MIFOTR NISR
M&E of NEP A,
projects MINALO
C
Annual budget estimates (RWF thousands) 300 35,000 35,300
III.6 MIFOTRA & LMIS MIFOTR None
dissemination system A / LMIS
improved

Annual budget estimates (RWF thousands) 3000 40,000 43,000

III.7 Increased data on MINEDU NCBS
skills availability C&
from the education WDA
and training system

Annual budget estimates (RWF thousands) 100 50 20 20 20 20 230

Total cost (RWF thousands) 2,200,280 385,940 979,470 5,304,460 1,041,320 1,254,620 5,381,120 16,536,920

59
Annex 2: Labour Market Indicators*

Category 1: Labour Supply (Labour Force)

Group No. Title Priority

Distribution of Working age population by Age, Sex
1 II
and Labour force status

Population 2 Dependency ratio II

3 Age structure of population III

4 Labour force participation rate I

Labour Distribution of persons outside the labour force by

5 III
Force Reason

6 Labour force status by Nationality IV

7 Employment-to-population ratio I

8 Distribution of Employed population by Industry I

Distribution of Employed population by Status in

employment

9  Vulnerable employment rate I

 Proportion of employees in precarious

work
10 Informal sector employment rate I
Employment
11 Informal employment rate I

12 Distribution of employed population by Occupation II

Distribution of employed population by Working time

bands
13 II
 Excessive hours

14 Average weekly hours worked II

Share of paid Employment in non-agricultural

15 III
employment

60
 Group No. Title Priority
Average weekly hours worked by Employed
16 III
population by Status in employment
Average weekly hours worked by Employed persons
17 III
in selected 2-digit level Industries
Average weekly hours worked by Employed persons
18 III
in selected 2-digit level Occupations
Distribution of employed population by Industry and
19 III
Occupation

20 Employees by Industry IV

21 Employees by Institutional sector IV

22 Employees by selected 2-digit level Industries IV

23 Employees by Occupation IV

24 Employees by selected 2-digit level Occupations IV

25 Employees by Industry and Occupation IV

26 Employees by weekly hours worked by Industry IV

Employees by weekly hours worked by selected 2-

27 IV
digit level Industries

Employment in the informal sector by main &

28 IV
secondary activity

29 Distribution of employed persons by nationality IV

30 Unemployment rate I

31 Youth unemployment rate I

Labour 32 Youth not in education and not in employment rate I

Underutilizati
on 33 Time-related underemployment rate I

34 Rate of Labour underutilization I

35 Long-term unemployment rate II

61
 Group No. Title Priority

36 (Former) Relaxed unemployment rate II

Unemployed population by Category (1st time,

37 IV
Previously employed)
Unemployed persons previously employed by former
38 IV
Industry
Unemployed persons previously employed by former
39 IV
Occupation

40 Registered job-seekers rate IV

Average hourly earnings of employees by

Industry, Occupation
41 I
 Average hourly earnings of
employees in selected 2-digit or lower occupations

42 Average hourly earnings of employees by deciles II

43 Wage index III

Wages
Average income-from-employment for self-employed
44 III
by Industry, Occupation

45 Minimum wage rate III

Mean weekly earnings of employees by selected 2-

46 IV
digit level Industries
Mean weekly earnings of employees by selected 2-
47 IV
digit level Occupations
Distribution of Labour force by Educational
48 I
attainment
Distribution of unemployed persons by Educational
49 I
attainment
Distribution of employed persons by Occupation and
50 I
Educational attainment
Distribution of Employed persons with certification by
type of certification ( Vocational certificate, Bachelor’s
Skills 51 I
degree, Masters degree, Ph. Ds, Professional
qualification) by Industry
52 Distribution of employed foreigners by Industry I

Distribution of trained persons by areas of training

53 I
and level of training

54 Literacy rates of the labour force II

62
 Group No. Title Priority

55 Distribution of required skills by type of establishment III

Distribution of number of outputs (trained persons)
56 accessed professional internship from public & III
private institutions
Distribution of Working age population by
57 III
Educational attainment

58 Share of women in non-agricultural paid employment II

59 Share of youth in non-agricultural paid employment II

Female share of employment in senior and middle

60 II
management

61 Wage gap II

62 Occupational segregation III

Trade union density rate of employed persons for

63 III
each Status in employment category by Industry
% of Employed persons covered by collective
64 III
bargaining for each Status in employment category
Employment
Equity & % of Employed persons covered by collective
65 III
Industrial bargaining by Industry
relations Distribution of Number of strikes and lockouts by
66 III
Industry
Distribution of Number of workers involved in strikes
67 III
and lockouts by Industry
Distribution of Number of registered complaints by
68 III
type and Industry

69 Trade union membership by type of member IV

Trade union density rate of employed persons for
70 each status in employment category by Institutional IV
sector
Employed persons covered by collective bargaining
71 IV
by Institutional sector
Days not worked due to strikes and lockouts by
72 IV
Industry

Safe Work & 73 Fatal occupational injury rate II

Social
protection 74 Labour inspection rate II

63
 Group No. Title Priority

75 Non-fatal occupational injury rate II

76 Rate of occupational diseases II

Ratio of labour inspectors to number of workplaces

77 II
coverable
Share of persons in labour force with pension
78 II
coverage
Distribution of Social security beneficiaries by
79 III
Industry
Number of self-employed persons contributing to
80 III
social security
Proportion of persons over 64 years entitled to a
81 III
pension

82 Cases of non-fatal occupational injury by Occupation IV

Cases of non-fatal occupational injury by Type of

83 IV
incapacity and Industry
Days lost due to cases of occupational injury with
84 IV
temporary incapacity for work by Industry

85 Cases of fatal occupational injury by Industry IV

86 Cases of fatal occupational injury by Occupation IV

87 Workers in reference group by Industry IV

88 Workers in reference group by Occupation IV

Registered workplaces that could be selected for

89 IV
labour inspection

90 Working poverty rate I

Poverty of 91 Earnings inequality II

employed
persons 92 Low pay rate III

Working-age population living below the national

93 IV
poverty line by labour force status

64
Category 2: Labour supply (Other work activities)

Group No Title Priority

Other labour 1 Distribution of persons in (own-use production of I
input into SNA goods, unpaid trainee/apprentice work, volunteer
work in market enterprises and volunteer work for
own-use production of goods for other households)
by Industry
2 Rate of subsistence foodstuff producers II
3 Rates of own-use producers of goods by activity II
cluster
4 Unpaid trainees/apprentices in establishments by III
Industry

Category 3: Labour demand

Group No. Title Priority

Vacancies & 1 Vacancy rate I
Jobs
2 Distribution of Jobs created in Formal/informal I
sectors by Industry,

3 Registered Vacancies IV
4 Evolution of professional service providers last 5 IV
years
5 Evolution of employment agencies last 5 years IV
Costs to 6 Labour costs II
employer
Establishments 7 Distribution of Establishments by Formal/Informal I
sectors
8 Growth rate of Micro, small & medium enterprises II
(MSMEs)
9 Distribution of establishments by Ownership II
10 Distribution of establishments by Size (Capital or II
employment)
11 Distribution of work permits by Industry and IV
Nationality

65
Category 4: Context

Group No. Title Priority

Context 1 Labour productivity (Growth rate of the ratio I
GDP/E)

2 Labour income share of Gross value added I

3 Employment elasticities (Growth of employment as I
ratio of growth of GDP)
4 Inflation rate I
5 Children’s economic activities / Child labour II
6 Proportion of working age population (labour force II
or employed) living with HIV/AIDS
7 Public social security expenditure as proportion of II
GDP
8 GDP growth rate II
9 Private investment as proportion of GDP II
10 Public expenditure on health as proportion of GDP III
11 Public expenditure on education as proportion of III
GDP

*: Core Rwanda Labour Market Indicators are in bold

66