The Annual Establishment & Institution Survey (AEIS) has been developed as part of an integrated programme of the Bangladesh Bureau of Statistics (BBS) in order to provide timely information for various economic sectors needed for national accounting and other policy, management purposes. The Industry and Labour Wing has been formed in order to conduct economic surveys on the sectors that are not related to agricultural activities. Those sectors are: (i) Manufacturing (ii) Trade (iii) Service and (iv) Household based economic activities. These sectors have been divided into two groups, namely, small establishments and large establishments. In the manufacturing sector the small establishments are those that employ 19 people, and those that employ 10 or more people have been classed as large establishments. In the trade and service sectors the establishments that employ 1-19 people are called small establishments and those employing 20 or more people are large establishments. The household sector has been placed only in the small group irrespective of the number of persons engaged. It has to be specifically mentioned that the large manufacturing sector has been separated from the above mentioned sectors and placed in a separate survey called the Census of Manufacturing Industries (CMI) for which it is outside the scope of the AEIS. The small manufacturing sector and all other sectors, both small and large along with the Household sector, are covered by the AEIS. In order to capture the seasonal activities of the sectors, the surveys are based on monthly national sub-sample of EAs, and collect data through out the twelve months of the year using separate questionnaire for each sector.
The main objective of the AEIS is to provide annual estimates for the key variables in each sector. The variables are:
1. value of products sold or manufactured and services rendered (gross output)
2. value of raw materials and other inputs (industrial cost)
3. operating cost
4. input cost
5. employment by sex and category
6. employment cost
7. value of inventories (change in stock)
8. fixed assets
9. fixed assets information (capital expenditure)
10. gross value added
11. value added at factor cost
The information obtained for each non-agricultural economic sector is defined by an industry group following the Bangladesh Standard Industrial Classification (BSIC).
Kind of Data
Sample survey data [ssd]
The Annual Establishment & Institution Survey (AEIS) sample design is based on two independent frames, a list frame and an area frame: The list frame covers the large establishment of th6 trade sector and the service sector. As has been explained in the background at the beginning the large manufacturing sector has been separately placed in the Census of Manufacturing Industries, and the household sector has also not been included in this frame. The area frathe covers all the small establishments that have been excluded from the list frame. A total of 1500 EA's were selected from the area frame consisting of 1,96,720 EA's to cover manufacturing establishments with 1-9 persons engaged, other sector establishments with 1-19 persons engaged and various household economic activities. For household economic activities, there were no cut-off relating to person engaged. Therefore, all households with relevant economic activities were included in the area frame.
Producers and sponsors
Authoring entity/Primary investigators
BANGLADESH BUREAU OF STATISTICS
Statistics Division, Ministry of Planning
The sampling strategy for each sector was determined on the basis of the survey objectives and the nature of the sectors. The total persons engaged (TPE) for the survey sectors (Wholesale Trade, Retail Trade and Services) are heavily concentrated in the area frame, since most of the stablishments are small. The sampling for each of these sectors was designed to provide reliable result at a more aggregate level, as well as for the larger BSIC groups. The universe covered by this component of the AEIS consisted of all permanent establishments with 20 or more TPE in the wholesale Trade. Retail Trade and Services sectors. The sampling frame consisted of a directory list of permanent establishments with names and addresses. The directory was compiled from data obtained .from the economic census of 1986.
The overall sample size was determined on the basis of the precision required for the individual BSIC domains with an upper limit set by resource constraints. At the first step, the overall sampling strategy for each sector was determined on the basis of distribution of the establishments by size and optimum allocation criteria. Table-i represents the distribution of establishments in the directory list frame by economic sectors and TPE size groups, and Table-2 shows the sampling rates for each sector by TPE size group determined for the sampling approach.
A one-stage systematic sample selection from a list frame with specified sampling rates is fairly straight-forward. For each stratum this sampling rate was determined by using Table-2. The sampling interval (Ih)Will be the inverse of the sampling rate, obtained an integer random start (Rh)between I and Ih. The sample establishments were selected for stratum "h" beginning with the random start, and multiples of the sampling interval were added until the end of list for stratum "h" was reached. The selected establishments carried the following order numbers: Rh, Rh+Ih, Rh+2Ih, Rh+31h, etc.-----
Deviations from the Sample Design
Sectors TPE size groups
20-29 I 30-49 I 50+
Wholesale Trade 25 % 50 % 100 %
Retail Trade 50 % 100 % 100 %
Services 10 % 20 % 100 %
This sampling strategy in an overall sample of approximately 1513 permanent establishments.
The sample for the Survey was based on a stratified multistage design. The primary sampling units (PSUs) were the enumeration areas (EAs) which were selected with probability proportional to size (PPS). In order to expand the AEIS data to the national or domain levels, it was necessary to apply a weight (expansion factor) to the data from each survey record. The basic weight would be equal to the inverse of the probability or selectiori .(ciculated by multiplying the probabilities at each stage of selection). Since the EAs were selected with PPS a separate weight was calculated for each EA. In the case of BSIC groups which were sub-sampled within a sample EA, it was necessary to take into account this second stage probability. The overall probability of selection for each sample establishment or household was as follows:
PhIJ=[nh*MhI)/SUTVIh(MhI)] *eh../Eh..1 where: PhIJ=sampling probability for economic units in the j-th BSIC group with the i-ht sample EA of stratum h nh=number of sample EAs selected in stratum h MhI=measure of size for the i-th sample EA in stratum h, SUMh(Mh)=cumulative measure of size for stratum h, that is, the sum of the measures of size across all the EAs in the frame for stratum h ehij number of sample economic units selected for the j-th BASIC group within the i-th sample EA of stratum h Ehij = total number of economic units listed for the j-th BSIC group within the i-th sample EA of stratum h.Note : In the case of BSIC groups for which no sub-sampling was carried out within the sample EA, ehij = EhI, and the second' component of probability dropped out (since it was equal to 1). The basic weight (wh )to be attached to the record for each unit (establishment or household) was, therefore, the inverse of this probability, or:
Whi = [SUIVIh (MhI )/(flh * Mhj)]* Eh1 /ehij ],
Whij = weight or expansion factor for each sample economic unit in the j-th BSIC group in the i-th sample EA of stratum h
These basic weights were adjusted for non-interviews and for any other problems which affected the probabilities of selection.
Some of the most important survey estimates were in the form of totals, such as total production or employment. The survey estimates of the total obtained are as follow:
X = SUIVIhIi (WhI * XhiJk), where
X = estimate of total for variable x
Whij = final adjusted weight for sample economic units in the j-th BSIC group within the i-th EA
of stratum h
Xhk = value of variable x for the k-th sample economic unit in the j-th BSIC group within the i-th
EA of stratum h
The expression " SUMhij ,, represents the summation within these levels. For particular domains, the summation would only be across the observations in that domain. In the case of the survey estimate of the total number of economic units with a certain characteristic, the variable X was defined as follows:
X = 1 if the economic unit has that characteristic, or
X = 0 if otherwise
Dates of Data Collection (YYYY/MM/DD)
Mode of data collection
Ministry of Planning
After completion of the enumeration, the enumerators submitted the filled-in questionnaires to the headquarters. After receiving the questionnaires the respective officers
checked the work of the enumerators. If any gross mistake or omission or duplication was observed the concerned officers, as far as possible, corrected it accordingly, and advised the concerned enumerators not to make such errors in future. Then the filled-in questionnaires were manually edited as per instruction provided to the editors. The edited schedules were checked on sample basis. If the error rate was found beyond acceptable limit for any editors, then 100% check was made.
After completion of editing and coding work, the filled-in schedules were sent for computer processing of the data. The data were keyed into diskettes and subsequently transferred into magnetic tapes, and were edited by the computer thoroughly to identify the suspected relationships among the data items and possible incorrect markings by the enumerators. Complete listings of the potentially erroneous questionnaires were prepared. These were reviewed for inconsistencies and corrected whenever found necessary. The data were then prepared for processing. Necessary computer programmes were written to produce statistical tables as per tabulation plan. A total of 86 statistical tables were prepared comprising all sectors/sub-sectors of the economy covered by the AEIS both through area sample and the list sample.