Botswana

Title of the survey

Survey of Employment and Employees

Organization responsible

Central Statistics Office

Periodicity of the survey

Bi-annual, in March and September.

Objectives of the survey

To provide up-to-date estimates of employment and monthly earnings. In certain years, the survey also provides information on occupations, training needs, conditions of work and labour turnover, which are used for manpower planning.

Main labour topics covered by the survey

Employment, earnings and job vacancies.

Reference period

The whole months of March and September.

Coverage of the survey

Geographical

The whole country.

Industrial

All divisions of economic activity in the formal sector. Excluded are the armed forces, household services, and informal sector activities such as domestic workers, barbers, cobblers, street vendors, etc.

Establishments

All types and sizes of establishments, except those which do not have employees, such as associations, thrift loan societies, borehole syndicates, syndicates, foreign missions and donor organizations, as well as ownership of dwellings.

Persons

All persons engaged in the above-mentioned establishments.

Occupations

All occupations are covered in principle. Data are collected on employment and earnings by occupation certain years only (see also History of the survey).

Concepts and definitions

Employment

All persons engaged include working proprietors, unpaid family members who worked for at least one hour during the reference month, and employees. Employees are defined as persons on the payroll on the payday nearest to the end of the survey month, who worked for at least one hour during the reference period, regardless of whether they were paid on a monthly, bi-weekly, weekly, daily or hourly basis. They include paid family members and company directors, employees who were temporarily absent (e.g. on paid leave) and who will return to work after their absence, persons leaving or joining the establishment during the month, permanent and temporary workers and part-time workers, as well as all categories of employees (seasonal, casual, apprentices and trainees, piece workers and commission agents, etc.). Data are collected separately on:

Earnings

Data are collected on the total remuneration (i.e. total gross earnings) paid to employees, as defined above, during the survey month, by sex and citizenship. Total gross earnings include wages, allowances, overtime payments, bonuses and other cash payments and cash benefits paid during the month, including payments made to employees who left or joined the establishment during the month. They exclude non-cash benefits. In March 1993, data were collected separately on (a) gross wages and salaries paid during the month, and (b) bonuses, gratuities and payments in kind paid during the past 12 months, by sex and citizenship.

Wage/salary rates

Not relevant.

Hours of work

Not relevant.

International recommendations

The definition of total cash remuneration used in this survey conforms to the international guidelines on regular cash earnings. It excludes earnings in kind.

Classifications

Industrial

Data on employment and earnings are classified according to the International Standard Industrial Classification - Botswana adaptation (BISIC), which is linked to the International Standard Industrial Classification of all economic activities (ISIC), Rev.2, 1968 at the three-digit level.

Occupational

Where relevant, the International Standard Classification of Occupations (ISCO-68) is used (see "History of the survey").

Others

The survey data are classified by sex and citizenship.

Sample size and design

Statistical unit

The reporting unit is the establishment in the formal sector.

Survey universe / sample frame

This consists of the Central Statistics Office's Enterprise and Establishment Register (EER) which covers all establishments in the formal sector. The register is maintained using a variety of sources, such as the Register of Companies maintained by the Ministry of Commerce and Industry, the approval of trading and transport licenses, telephone and fax directories, newspaper advertisements. The EER covers some 18,000 active establishments.

Sample design

The survey design combines fully enumerated strata and sampled strata. The following establishments are fully enumerated: A sample is drawn from all private companies employing fewer than 50 employees and those whose employment size is unknown: the EER is first stratified by economic activity (ISIC divisions), then a sample of establishments in drawn from each division. The sampling fraction varies with the number of establishments in each ISIC division, as follows: The overall sampling fraction, including the fully enumerated stratum, is normally around 25 per cent of the total active establishments in the EER.

Field work

Data collection

The survey is conducted by post. Questionnaires must be returned within 30 days. Postal reminders are sent to non-respondents, and follow-up visits are made by members of the Labour Statistics Unit.

Survey questionnaire

Various forms are used, for private and parastatal organizations, central Government and local Government, which follow the same format, with some slight adaptations. The questionnaires comprise questions on: Instructions are provided in the questionnaire with regard to the main inclusions and exclusions of categories of employees and components of remuneration.

Substitution of sampling units

There is no substitution of sampling units in the case of total non-response. Mailed reminders and personal visits are used to follow-up on non-respondents.

Data processing and editing

Data are processed by computer, using PCs. They are edited by machine and field work whenever possible. Consistency checks are made, particularly for average earnings, based on the previous survey sub-sectors average earnings.

Types of estimates

Number of employees: totals, percentages and annual changes. Earnings: average cash monthly earnings and indices.

Construction of indices

Indices of average earnings are computed by sector, economic activity and citizenship.

Weighting of sample results

The following formula is used to gross-up the survey results, at the three-digit ISIC level:

Ti=(Ni/ni)xi+TLi

where,

Ti=
the estimate of total employment in ISIC subindustry i,
Ni=
the number of establishments in ISIC subindustry i, with 50 or less employees or employment size unknown,
ni=
the number of establishments in ISIC subindustry i, with 50 or less employees or employment size unknown, which positively responded to the survey,
xi=
employment in the ni in ISIC subindustry i,
TLi=
employment in the establishments in ISIC subindustry i, classified as either being parastatal, multi-establishments or private company with 51 or more employees.

The above formula is used for producing national estimates of number of employees by citizenship, sex and occupation (where relevant), and estimates of wages.

Adjustments

Non-response

In the final estimates, data for non-responding establishments from the fully-enumerated strata are imputed using the previous survey's returns. Establishments which do not respond and establishments for which CSO communications are returned unopened are assumed to be similar in employment aspects to the establishments which responded. The stratum weights are therefore adjusted to take account of non-response.

Other bias

None.

Use of benchmark data

Not relevant.

Seasonal variations

None.

Indicators of reliability of the estimates

Coverage of the sampling frame

Every effort is made to cover all establishments in the formal sector.

Sampling error / sampling variance

Not computed.

Non-response rate

The non-response rate averages 30 per cent in terms of number of units.

Non-sampling errors

Inaccuracies may occur because of coding and data entry errors. The sample stratification is based on employment size and ISIC group; hence, estimates for other variables, particularly those for location, are not of the same standard of reliability as those for economic activity.

Conformity with other sources

No other current source of data on employment and earnings exists. The survey results are compared with the previous same survey.

Available series

Published results include cross-tabulations of:

History of the survey

An Employment Survey was first introduced in 1972, and conducted by the Central Statistics Office on an annual basis, in August of each year, until 1984. It was used to collect data on employment, monthly earnings, occupation, work experience, hours of work, labour turnover, etc. In 1984, the Employment Survey was replaced by an employee-based survey, the Survey of Training Needs and Conditions of Work (STNCW), later renamed "Survey of Employment and Employees". Starting from 1985, two surveys were conducted: a comprehensive and employee-based survey in March, and a more simple survey in September, which collected information on employment and earnings by sex and citizenship, and unfilled vacancies. The March survey collected detailed information on a sub-sample of employees' characteristics such as occupation, sex, age, citizenship, work experience, hours of work, monthly earnings, etc. Due to financial and manpower constraints, this survey could not be conducted annually as initially planned and, since 1987, data have been collected on employment and earnings by sex and citizenship only, during the months of March and September of each year. In March 1993, an occupational employment and wage survey was re-introduced. For each occupation reported by the establishment, the survey collects data on the job description, title or occupation, minimum education and training required, unfilled vacancies, and normal hours of work, number of employees and earnings by sex and citizenship, as well as labour turnover.

Documentation

Central Statistics Office: Labour Statistics 1986-1990 (annual; Gaborone). idem: Employment Survey 1972-1985 (annual; ibid.). Both publications contain some methodological information. The results of the most recent surveys may also be made available on diskette, upon request.

Confidentiality / Reliability criteria

The publication and release of data are subject to confidentiality rules, whereby individual units should not be identified.

Other information

Data supplied to the ILO for publication

The following data are published in the Yearbook of Labour Statistics: The corresponding half-yearly series of paid employment are published in the Bulletin of Labour Statistics.