Statistical Methods, Standards and Guidelines
55 4.8 INTERNATIONAL COMPARISON PROGRAM (ICP) The International Com parison Programs is a global project, managed centrally by the World Bank, with an organised hierarchy of regional management structure. The responsibility of managing its African component ICP -Africa lies with the African Development Bank (AfDB). The main objective of the ICP is to compare the economic aggregate and the volumes of gross domestic product (GDP) between the countries.
The targeted population is the set of all goods and services that are consumed by household from outlets during the benchmar k year, covering the whole country (including rural and urban areas), except expenditure of housing rent on residential (building), education and health services which are subsidized in most of the countries.
The ICP price survey is integrated with the ex isting system of price collection for the CPI except to some items which are not available in CPI but are needed in ICP according to the agreement of country members.
The ICP price collections are carried out in urban and rural areas (weekly market) of th e sampled centres. Stratification allocation and purpose are the methods used to select the region for ICP price collection survey. In Tanzania, there are seven zones including Zanzibar which constitute Unguja and Pemba.
The country members capture and va lidate ICP price collection using the same method as used in CPI but further analysis is done by African Development Bank (AFDB).
4.9 HARMONISED CONSUMER PRICE INDEX (HCPI) The Harmonised Consumer Price Index (HCPI) Compilation Model is an Excel add -in based Visual Basic for Application (VBA) program developed for providing technical assistance to Fund member countries of SADC for consumer price statistics compilation. The package is a prototype model that could also be used for teaching, training, and research purposes. It can likewise serve as a framework from which price collection formats and product classification systems can be integrated to fit country-specific practices.
Its main function is to compile the HCPI for multiple areas at multiple levels of product items to obtain aggregate national level price statistics.
Statistical Methods, Standards and Guidelines
56
It’s Methodology.
-
The Modified Laspeyres Formula is used to compute HCPI statistics based on monthly price quotations (or monthly average price quotations) and annual expenditure information using a combination of manual and computer tabulation procedures.
It/0 = i pi tqb i / i pi 0qb i which can be writtenIt/0 =
i wi . Ii
t / Ii
0
where
wi is the weight used for product i,
It/0 is the price index for product i between the price reference period 0 and period t;
pi
t is the price of product i in period t;
pi 0 is the price of the same product i in period 0; qb i is the base-period quantity of product i expressed as a proportion of the expenditure on product i to total expenditures covered in the HCPI.
-
HCPI is calculated using the equivalent of a recursive procedure, in which previous period‟s base-weighted long -term price relatives, pt-1, q0 are updated by the current period‟s price relatives.
-
The geometric mean method is used in computing the price level at aggregate level in view of its multiple advantages.
-
The program adopts the Matching Price Observation me thod in imputing the areas‟ average prices, in which the price averages are calculated on the basis of “matched observations”. Whenever a particular price observation is missing from either the previous month or the current month, the corresponding price o bservations will be dropped from the other period. This is to ensure consistent sample of price quotations in each period.
Statistical Methods, Standards and Guidelines
57 5. The program calculates missing variety prices based on Short Term Price Relative STPR (actual or imputed) and previous period price, and stores them in the database with a flag. These calculated price data can be retrieved into spreadsheets for the next period imputation process.
4.10. INTEGRATED LABOUR FORCE SURVEY. Objective of Labour Force The Integrated Labour Force Survey intends to obtain comprehensive data on the current status of National Labour Market and to provide up to date data needed by the government and other stakeholders on human economic activities particularly those related to the informal sector and its magnitude, unemployment, underemployment child labour and time use.
Use of Product The findings of the survey are used in planning, policy implementation, monitoring and evaluation of government programmes aimed to determine the magnitude of the Labour force in the c ountry and to collect information on employment status so as to introduce necessary changes in the country‟s employment policies where needed.
Methodology Design of the Sample The Integrated Labour Force Survey used the existing National Master Sample (NMS). The NMS is a generalized set of area units that can be used as Primary Sampling Units (PSUs) for conducting various household surveys. It is a fixed sample of rural and urban clusters, which, among other things, makes it possible for the performance o f a continuous Survey Programme as well as ad hoc sample surveys.
Frame of the Sample
The sampling frame for the current NMS is based on the preliminary results of the 2002 Population
and Housing Census . The Primary Sampling Unit (PSU) is the village for the rural and EA for
urban areas respectively. A probability proportional to size without replacement (ppswor) –
systematic sampling procedure is used for the selection of PSU. About two months before the
commencement of the field work a household listing exercise is done. All households within each
cluster are listed. The household listings give the sampling frame of households for each cluster.
Statistical Methods, Standards and Guidelines
58 Sample Size Determination The selections of EAs follow the Probability Proportion to Size (PPS) sampling whil e the selection of households and individuals follow a simple random sampling procedure.
Estimation Procedure The sampling procedure for both the urban and rural samples suggests good estimates at national, and cluster levels. Regional estimates can also be worked out. For urban sample it is possible to get estimates for the three domains of study, i.e. Dar Es Salaam city, nine municipalities and other towns. Estimation of individual towns and households‟ sizes can also be obtained by some imputation methods.
Data Collections
Labour force survey uses five types of questionnaires. First, questionnaire (LFS1) is administered
to the head of the household or his/her representative intending to collect household particulars.
The second questionnaire (LFS2) aims to collect the information of labour force details for
individuals. The third questionnaire (CLS1) is administered to parents or guardians of all child aged
5 to 17 years. The fourth questionnaire (CLS2) aims to collect information of children age 5 to 17
years. The fifth questionnaire on the time use (TUS) is designed to collect the information on the
routine activities of the respondents and administered on seven consecutive days to each member
aged 5 years and above of the selected households.
Field work The field work includes listing exercise, the supervisors and enumerators must be trained on map reading and listing. The regular field visits for ensuring close supervisors is made by national and regional supervisor. Supervisors are also responsible for ensuring the quality control of the data at all stages of data production.
Data Collection Data collection is conducted in teams, each team consist of supervisor, enumerators and a driver. Supervisors are responsible for the overall administrative wo rk in the field including checking the quality of the questionnaires before departing from the cluster. Data Processing Data processing starts soon after receiving questionnaires from the filed. The data processing personnel includes supervisors and a ques tionnaire administrator, who are responsible for checking
Statistical Methods, Standards and Guidelines
59 the number of clusters (EAs) in a region and number of each household in the cluster. Followed by manual editing, coding questionnaires, data entry using CSPro as a package to capture data and data cleaning and validation is done by experienced data processing personnel.
4.11 EMPLOYMENT AND EARNINGS SURVEY. The Employment and Earnings Survey, is an annual survey conducted by the National Bureau of Statistics. The enumeration covers three main categ ories of employing establishments in both private and public sectors. The categories involved are: All establishments of public sector; all registered private establishments employing at least 50 persons; and a sample of all registered private establishmen ts whose employment capacity is between 5 to 49 persons in Tanzania Mainland. The survey does not include domestic servants in Private households, non -salaried working proprietors and non-salaried family workers.
Objective of the Survey The main objective of employment and earning s is to obtain a comprehensive data on the annual status of employment and earnings as well as data on socio -economic characteristics of the labour market.
Use of Product The findings of employment and earnings survey are used fo r estimating the labour market indicators that could be used for planning, policy formulation and examining the contribution to Gross Domestic Product (GDP) of different categories of employment.
Methodology (a) The Selection of Establishments The Employment and Earnings Survey used the existing Central Register of Establishments (CRE) frame. The selection of establishments from the CRE frame falls under the following groups: - i). All establishments of public sectors found in the current CRE frame are taken; ii). All establishments of private sector with at least 50 employees found in t he current CRE frame are taken; iii). The list of surveyed establishments of private sector employing persons in the range of 5 to 49 is based on a sample.
Statistical Methods, Standards and Guidelines
60 (b) The Sample Design i). A sample of 10 percent of establishments is selected in the employment size group of 5 to 9 employees;
ii).
A sample of 33 percent of establishments is selected in the employment size
group of 10 to 49 employees.
(c) Sample Selection A random sampling method is used to select the number of establishments to be enumerated according to the sample size in each employment group.
4.12 NATIONAL PANEL SURVEY 1 Background The Technical Committee of the MKUKUTA Monitoring System has requested the National Bureau of Statistics (NBS) in late 2006 to establish a technical team to clarify the way forward in conducting a potential series of regular annual National Panel Surveys. The team is formed, and is inclusive of Government, research organizations, and development partners. The team is chaired by NBS and links to the Survey and Routine Technical Working Group of MKUKUTA Monitoring.
2 Purpose of Tanzania’s Panel Survey The National Panel Survey (NPS) is intended to achieve multiple objectives. These objectives are:
To provide information required for MKUKUTA monitoring, monitoring of other development objectives (MDG, PAF) and monitoring of specific programs;
To provide high quality nationally representative information on income dynamics at the household level, to provide annual consumption estimates to monitor poverty in years between HBSs and to provide reliable agricultural statistics to feed into the National Accounts;
To provide a flexible survey instrument able to accommodate ad hoc data requests and to assess impact of (new) policy interventions
Statistical Methods, Standards and Guidelines
61 3 What is a Panel Survey? Panel surveys collect data about individuals, households and communities over time, in order to assess change. Because panel surveys usually collect data from the same sources each year the y are well suited to assess change, to provide information about the causes of change and poverty dynamics and to assess impact of interventions.
4 Why is it important for monitoring MKUKUTA and other development initiatives? For MKUKUTA (but also PA F), information from Panel Surveys will help to provide indications of the degree of chronic and transitory poverty, and allows to measure some MKUKUTA outcomes annually (see annexes 1 and 2). This could complement the 5 -year HBS cycle, e.g. poverty incidence, inequality or access to clean water.
It may also be a useful instrument/platform to reduce the numerous surveys planned for specific projects, by consolidating them into a single operation. Therefore for MKUKUTA monitoring it potentially links to output reporting via the Annual Implementation Report.
The Panel Survey will also regularly provide information on the state of the agricultural sector and rural livelihoods, and supply information for the monitoring and evaluation of the ASDP and other sectoral interventions.
5 Timeframe The timeframe reviewed here for the NPS consists of the first 3 years of the panel. This includes 4 months of planning (described below) and 3 annual rounds (12 months each), each including a preparation phase for the subsequent data collection.
When implemented successfully, the panel may be maintained for a much longer period. Some panels have been maintained for as long as 20 years. For outer years, budget information as presented for year 3 is indicative.
6 Institutional Arrangements The National Bureau of Statistics takes overall responsibility for the National Panel Survey. In doing so they will report progress and seek technical advice and financial approvals through the Survey and Routine Data Technical Working Group, and its on -ward link to the MKUKUTA Technical Committee.
Statistical Methods, Standards and Guidelines
62 The NBS may choose to form partnerships with other institutions or may contract out elements of the preparation, implementation, data processing or analysis of the National Panel Survey.
Public procurement procedures will be followed in the implementation of this work.
7 Survey Sample The overall sample design will provide for annual poverty estimates for three strata in mainland Tanzania (rural, Dar es Salaam, other urban areas) plus Zanzibar, and will enable the study of poverty dynamics. The sample will supply annual production estimates for main crop and livestock at the national level and, if feasible and deemed of policy relevance, for major agro-ecological zones. Given the proposed sample size, agricultural data can also be analyzed by ex-post farm typologies e.g. by farm-size, degree of market integration, etc. Information on average cost of production for selected crops at the same level of geographical disaggregation will also be supplied, though less frequently.
The sampling will emphasize randomization to ensure full representative ness at the selected domains of inference. The NPS sample will be drawn from the National Master Sample (NMS). The draft sample design is described in Table 1 below and was based on inputs from sampling/survey experts. It entails interviewing 3,456 households (2,240 rural and 1216 urban) in mainland Tanzania and 576 in Zanzibar (288 rural and 288 urban) for a total annual sample of 4,032 households. The total rural sample will be 2,528 households.
The precise and final sample design will be finalized in the coming months with the support of a sampling expert. At that time, the exact PSUs and sequencing of fieldwork will be identified.
Table 1: Sample Design
Households Primary Sampling Units (PSUs) Mainland 3,456 216 Rural 2,240 140 Dar es Salaam 608 38 Other urban 608 38
Zanzibar 576 36