The project on "Promoting Competitiveness in Micro and Small Enterprises" (MSE) was initiated in 2000 by the Economic Research Forum, with the main objective has been to expand the knowledge on this sector in the Middle East and North Africa region, with the ultimate aim of designing relevant policies and specific programs to help this sector fulfill its enormous growth potentials. Constituting an average of 95% of the number of enterprises in the region, it is presumed that promoting this sector will have a positive spill-over effect on the economies of the region.
Discussions on the results of the project have pointed to an emerging consensus that it will be filling a knowledge gap related to the micro and small enterprises sector in the MENA region. Policies and strategies designed to promote this sector have not been adequately targeting their needs, and thus this project is considered to be of great relevance to the policy making process.
Specifically, the main contributions may be summarized as follows:
1) The database gathered through the project based on field surveys is considered unique, as to the number of enterprises covered (18,000), and the information produced, including information on the enterprise, the entrepreneur and the household. A special focus on women entrepreneurs have been made throughout the survey. This mine of data will undoubtedly provide background information that enables policy makers to design relevant policies.
2) The "Policy Briefs" gives a concise summary of the outcome of each country study and highlights the recommendations reached based on the analysis.
3) The current Country reports series is prepared based on the findings of the surveys, detailed information about the performance of the enterprises, determinants of success and prospects for the future are given. Special focus on the status of women entrepreneurs is also made.
4) The Synthesis report will have a comparative analytical approach of the case studies of the four countries. This report will asses the MSE sector in the four countries and will draw relevant policy recommendations for the region.
It has been evidently shown that promoting this sector could contribute to the solution of the increasing unemployment problem in the region, and a means to alleviate poverty through income generation. The spillover effects that this sector if properly developed will positively affect the development of the countries concerned. However, the real level of knowledge about the MSEs is
The Micro and Small Enterprises survey (MSEs) study in Morocco attempts to identify the following fundamental research questions:
1) What are the factors that determine the competitiveness and dynamism of the Micro and Small Enterprises?
2) What are the necessary reforms and policies for the development of the Micro and Small Enterprises?
3) How to involve the private sector and civil society organizations in the development of Micro and Small Enterprises sector?
4) What is the role of female entrepreneurs or workers in this sector?
Kind of Data
Sample survey data [ssd]
Unit of Analysis
Small and Micro Enterprises
Geographic Levels and areas study adopted are:
· National Urban, National Rural
· Grand Casablanca
· Rabat_S_Z_Z; Gharb_C_BH, Tangier-Tetouan
· Méknès_T; Fez _ B; Taza _H_T; Oriental
· Doukala_Abda; Marrakech_ T_H; Souss_M_D
The survey covered a national sample of Micro and Small Enterprises.
Producers and sponsors
Economic Research Forum
European Commission (through the FEMISE 2 project)
Arab Fund for Economic and Social Development
International Development Research Centre
University of Mohamed V, Morocco
Project’s Advisory Committee
Michigan University, USA
Project’s Advisory Committee
Project’s Advisory Committee
Preparation of frames
On instruction of the Minister of Economic Forecasting and Planning, the Department of Statistics provided the survey statistician project a copy of the file containing the master produced by the DS sample. This file was used to select, probability proportional to size, a sub-sample of primary units of the master sample. Then one by Direcet Census of Primary Unite was chosen to meet the needs of this study.
Industrial zones and clusters of economic units have been identified and a probability sample in these units was drawn in each of the provinces / prefectures affected by this survey.
Two types of scanning were conducted at each of DC / clusters / selected industrial areas: (1) a scanning preparing a comprehensive list of economic units of interest of local, (2) a scanning for households to complete, possibly, the list of economic units and establish a list of economic units of interest for which the activities are carried out at home. However, to stay within the budget allocated for the study, it was necessary to limit, sometimes, the door to door (2nd type of scanning) to a sample of houses in this case, the weights were adjusted by the number of households raked over the total number of households in the PU / DC.
The preparatory work of the sampling frames and the list of economic units to be investigated preceded the actual collection of data that was carried out by a small team of team leaders and their best elements. It is necessary to note that before drawing of samples, training in three independent areas of the study was completed in close collaboration with the person designated by the Department of Statistics. In this sense, the three frames are mutually exclusive and exhaustive.
Overall sample size and distribution
The overall sample size is based on firstly, the degree of reliability / accuracy set by the main parameters of the study and, secondly, material and human resources available and the time allocated to completion of the entire project.
The overall sample size was set at 5000 economic units. To achieve this figure, it was necessary to perform the probabilistic draw a total of 268 primary units (PU) .
The overall sample size fixed at around 5000 economic units. However, the probabilistic procedures adopted led to a total sample of 5210 economic units, an average of 19.44 economic units by PU [master sample (19,35), industrial zones (18,59) clusters (20,30)].
A more detailed description of the different sampling stages, allocation of sample across areas and drawing probabilities are provided in the methodology available among external resources.
The total non-response rate is about 21% on average, with some variations, sometimes significant, by study area, by the urban/rural residence and type of sample.
The rate of partial non-response of the study variables considered to be sensitive is negligible: less than 0.5% for all variables for all enterprises in the baseline survey sample (5210) up to a maximum 1.2% for the basic variables considered in field of study (urban, rural), and follow-up survey. However, it has been made a number of adjustments and calibrations for some answers but deemed inconsistent. The adjustments made by some variables are based on data from the Department of Statistics, the Economic Census for 2001/2002 and the National Informal Sector Survey 1999/2000.
A more detailed description of the distribution of the non-response rates across areas are provided in the methodology available among external resources.
The three sampling frames used for drawing samples does not allow direct extrapolation of the results of the survey. However, their combined use survey results with some specific economic study units obtained objects (1) of the National Survey and the Informal Sector (2) of the National Economic Census were used to extrapolate, straighten and achieve Timing results found.
Thus, a variable "weight" was created and integrated into the overall survey file. This variable has the weighting of sample data and hence lead to the results for the study be on the population level.
The extrapolation of the sample factors were developed using the sampling methodology described in paragraph 1 above and calculating the probability of inclusion of economic units in the sample in each stratum / sub considered strata.Inclusion probabilities are the product of three probabilities:
1. inclusion probabilities of primary units in the sample in each stratum / sub stratum ;
2.the selection probabilities of the selected DC;
3. the selection probabilities of economic units.
The final weights were calculated taking into account the adjustment factors and timing prepared based on the relevant data of some variables of interest obtained from the Department of Statistics in particular the variable size of the economic unit classes
( "1", "2", "3", "4-5", "6-9", "10-19", "20-49"):
1. the ENSI_1999-2000 data on economic home units not covered by the National Economic Census (2001-2002);
2. National Economic Census 2001-2002: statistical tables showing the distribution by size (size-classes specified above) economic units by region, medium, industrial districts and primary units / clusters.
The data needed for the various quantitative and qualitative analyzes performed (statistical tables, indices, different measures of central tendency and dispersion, ...) have been prepared by the data processing specialist closely with the survey statistician team.
Dates of Data Collection
Data Collection Mode
Data Collection Notes
A) Pilot Survey
To achieve the objectives of the study, questionnaires and instructions and coding manual were developed in accordance with the rules and standards established for the four countries involved in this study on micro and small enterprises.
Given the complexity of the study and the importance of obtaining reliable data, it was necessary to conduct a pilot study to test all documents and methodological approaches in the field prior to the implementation of the main survey.
The main objectives below were assigned to carry out the pilot survey:
·assess administrative difficulties and location of PU / DC;
·test the validity of the methods used for the preparation of frames and data collection, including those relating to: (1) scanning the economic units and households accompanying sample selection and data collection and (2 ) scanning economic units and households drawing samples at the office and conducting the survey the following day.
·test the adequacy and the degree of applicability of the sampling methodology;
·test the validity of survey documents (questionnaires, instructions, ...);
·various evaluations: organization, team training, contacts difficulty, motivation, difficulties in collecting, sensitive issues, completion time, cost ...
To meet the objectives listed above, the pilot survey was conducted for 10 days in January 2001, according to the methodology described in the preliminary report of this study (Preliminary sampling methodology, December 31, 2000) and concerned the sites below:
1) Casablanca 4 PU / DC
- Stratum habitat "Luxury and Modern": 1 PU / DC
- Stratum habitat "New Medina": 1 PU / DC
- Stratum habitat "Old Medina": 1 PU / DC
- Stratum habitat "Precarious & Clandestine" 1 PU / DC
2) Big City 3 UP / DR: Meknes
- Stratum habitat "New Medina": 1 PU / DC
- Stratum habitat "Old Medina": 1 PU / DC
- Stratum habitat "Precarious & Underground": 1 PU / DC
3) Small / Medium City: Tiflet
- 1 PU / DC
1) Province: Khémisset (a Douar a dependent rural town Tiflet)
- 1 PU / DC
A7-day training was provided to the entire survey staff of 16 interviewers, 3 team leaders supervisors and 3 back-up interviewers .The staff has been selected from a total of 30 people who have received training.
The analysis of quantitative and qualitative data of the preliminary investigation has corrected anomalies methodological documents prepared and realize the documents used in the main survey. The results of this work have been documented in a report entitled " Projet Micro et Petites Entreprises : Résultats de l'Enquête Pilote/Maroc ; 2001 "
B) Phase 1 /baseline survey
The success of a field interviewing depends largely on the manner in which it is planned and executed. Thus, during the phase of data collection in the field, a suitable device has been set up to ensure the smooth running of the operation. This device includes the preparation of a proper investigation logistics, recruitment and training of survey personnel and organization of data collection in the field.
Taking into account the relative importance of the sample size and its distribution by province / prefecture objects of study, the recruitment of a survey of 55 people staff was needed. To ensure the quality of information collected and the efficiency of work in the field, each of the supervisors and interviewers was recruited based on his qualifications and past experiences in the field of surveys by direct interview. In addition, each of the interviewers / supervisors achieved a grade at least equal to university degree. A team of experts mobilized for this study supervised the staff survey.
To ensure effective implementation of the data collection in the field, the entire interviewing staff received a 13-day training. of preparation for the execution of the survey. This training was provided by the survey experts and has processed all its aspects, focusing on the concepts and basic definitions. To confirm the full assimilation of interviewers training, training sessions at the office were supplemented by simulations of field investigation.
Also, a set of instructions including explanations of all the steps to follow in the actual interviews and to be used as a reference in the field, was made available to auditors and interviewers.
Additional training devoted to sampling methods, methods of control and organization of field monitoring was provided to the supervisors.
A rigorous selection process involved all trained survey staff, and only 42 of the 55 people trained were selected, 34 interviewers from which 4 are back-up and 8 supervisors with one back-up.
The launch of collection of field data was carried out in close collaboration with 1) the local authorities and 2) regional offices of the Ministry of Economic Forecasting and Planning.
In addition to the control system, a system of supervision by the experts of the study and coordinated management of the fieldwork was developed.
The first field day allowed supervisors and controllers to install teams at each site investigated and make contact with local authorities and the Regional Delegation of the Ministry of Planning, which were previously informed of the survey implementation. These contacts aimed to ensure the support and assistance of the local authorities and facilitate the work of interviewers.
The monitoring team conducted subsequently and permanently, tours of monitoring and control to ensure the proper conduct of the field survey. She followed with the team leaders the progress of the data collection state and replace, if necessary, sample units by reserve units.
Data collection involved a sample of 5210 economic units representing almost all economic units of the Kingdom. The collection operation is divided into two periods: May-August 2002 and October 2002, which completed the sample.
Main problems encountered during the data collection in the field
1) Households screening
The total household screening needed a lot of time and caused a lot of problems without achieving significant results in terms of location of new economic units at home or inside the PU. A partial screening would have been more appropriate (optimal) for this kind of surveys.
The duration of interview ab was very long, an average of 2 hours and 8 minutes (109 minutes for the questionnaire "Enterprise" and 19 minutes to the questionnaire "household").This time the interview is at least twice as long as we deem best practice in data collection of acceptable reliability (Suggestion made by men land-interviewers and supervisors with extensive experience in collection data. These results were indicated in reports of work completion of data collection in the field or given verbally in meetings to discuss the problems of collection.)
Indeed, experience in this field shows that even if the collection is performed by a qualified, reliable data still function key elements for both the interviewer and the respondent, namely the psychological effect (large form), fatigue and tiredness leading to the formulation of rapid, and not accurate answers, and formulating insufficient explanations... It will be necessary in the future to review and better target aims to reduce by at least 1/3 the amount of information of the current questionnaire "Enterprise".
Information on variables, including those related to the performance of the company are highly reliable and sometimes questionable. Indeed, a significant proportion of entrepreneurs / managers are illiterate and few respondents hold regular accounting records. It is therefore necessary to have auxiliary information (auxiliary variables highly correlated with the study variables) sufficiently reliable to make the necessary adjustments and calibrations of survey data.
Rapid assessment of difficulties filling two questionnaires:
· Questions 1 to 104 do not generally pose problems the respondent answers easily.
· Questions 105-111: some respondents begin to ask questions about the true purpose of the investigation and any related taxes.
· Questions 112-123: some entrepreneurs find it difficult to sufficiently precise percentages.
· Questions 124-160: usually little or no problems with regard to these issues.
· Questions 161-179: the estimated working capital for entrepreneurs is difficult and requires a lot of time and thought
· Questions 180-248: little or no problems with these issues.
· 249-265 questions are the most difficult questions, require a lot of time and effort both on the part of the respondent that the interviewer arise at a time when the respondent is tired by the effort on answering the previous 248 questions.
The problem is worse with illiterate entrepreneurs / managers and / or not holding regular accounting records.
· 266-275 Questions: little or no problems here.
· Questions 276 to 322: the only difficulty lays in the estimation of in kind benefits.
Some respondents may have questions about the usefulness of the questionnaire and its relationship with the Enterprise questionnaire.
The household survey is relatively easy except in matters related to household income (questions 18 to 29) for which it is necessary to make adjustments based on other reliable sources.
The basic problem is to find the entrepreneur / manager available. Indeed, for a significant proportion of respondents, several contacts and visits were needed, especially in industrial areas, before completing the questionnaire. However, former migrant entrepreneurs easily accept to be interviewed.
Respondents belong to different social and educational backgrounds levels, so they do not respond to the survey in the same way. Some therefore agree with little or no problems answer questions, while others are suspicious and sometimes aggressive but all formulate the same point: the questionnaire is too heavy.
To enter data for the survey of micro and small enterprises, two computer applications were developed using SPSS DATA ENTRY software that has features well suited to this kind of operation. It is developing two models :
1) Enterprise questionnaire 2) Household questionnaire.
The designed applications have the quality of being both simple in scope of data entry, a reliable and rigorous information input. In fact, the software has several advantages:
· It allows you to make an input interface similar to the questionnaire that allows the agent to find the same entry configuration point of view as between the input interface and the questionnaire and thus facilitate its work and minimize the size of errors;
· It offers the opportunity to provide the implementation of all the possible validity checks only codes varying in a predetermined interval are accepted at the time of entry. entry errors of this type is automatically accompanied by a beep and a text message displayed on the screen, indicating the type of error;
· It also helps establish consistency checks which are very useful and extremely important in this type of operation. Thus, the main logical relationships between different variables of the questionnaire were considered. This type of control is not only put to ensure the quality of information at the time of entry, but also to correct any errors made by interviewers at the time of filling the questionnaire.
Before starting the data entry of the survey, appropriate training has been provided by the IT team for the data entry staff mobilized to accomplish this task. The training sessions were an opportunity to clarify the whole process of handling two computer applications and explain the various commands provided in this context. Data entry operators were none other than engineers new award-winners of the National Institute of Statistics and Applied Economics.
Furthermore, the data entry was carried out in the control room office, which has created a synergy between agents and supervisors before consolidating and controlling questionnaires. This way of working has helped to rehabilitate, in the early days of the operation, computer applications to the needs of the data entry, including some open questions whose coding is done in the office.
In order to ensure the smooth running of the operation, the IT team ensured ongoing supervision and intervened whenever the need was felt. It should be noted that the entry concerned the two questionnaires, household and enterprise, and lasted for thirty full days, this resulted in the complete entry of 5,210 questionnaires each type.
Once the Data entry is completed, the files have been prepared for the important phase auditing and cleaning. This phase is to ensure greater data quality and identify some unavoidable and difficult to control at the time of entry errors. Therefore, and given that the data entry of the two questionnaires was conducted separately, it was initially necessary for each data entry of “enterprise” questionnaire to enter the corresponding “Household” questionnaire.
Given the relatively high number of questionnaires (5210), this has been done in different provinces or in the selected primary units level. This need to split the parent file for each survey in 10 files so that each corresponds to a given survey area. This approach was of great importance since it easily identified not entered or entered incorrectly questionnaires. Two large files were made (household and enterprise file).
Moreover, and for purposes of analysis, the two files were merged into a single file so that each observation is composed of variables related to the enterprise and the corresponding household.
Estimates of Sampling Error
The estimate of the accuracy of the results depends on several factors including:
· The level of analysis: national, area of study, urban/rural residence... It is about the size of the sample and thus the accuracy of the results;
· The sampling plan adopted: a complex design was prepared to draw probability samples. Therefore, it is necessary to consider the effect of sampling on one hand, the positive effect of multiple stratification used (gain accuracy) and on the other hand, the negative effect of clusters (loss of accuracy: Er = 1 + (m - 1) ? where m denotes the average cluster size (19,44) and c is the coefficient of correlation within clusters).
The detailed statistical analysis, using the SPSS software for all study variables showed that:
· At national level, the relative error does not exceed 4.5%, with a confidence level equal to 95%, for any study variable whose relative dispersion is less than 120% and the effect of plan that survey does not exceed 1.8.These conditions are satisfied by the basic variables of the study;
· At area of the study level, urban-rural, follow-up level: under the same terms as national level and depending on the actual size of the sample, the margin of error for the estimation of the basic variables of the study, is between 3.5% and 14.5% with a confidence level equal to 95%.
Economic Research Forum
Economic Research Forum (ERF) - 21 Al-Sad Al-Aaly St., Dokki, Giza, Egypt
To access the micro data, researchers are required to register on the ERF website and comply with the data access agreement.
The data will be used only for scholarly, research, or educational purposes. Users are prohibited from using data acquired from the Economic Research Forum in the pursuit of any commercial or private ventures.
Licensed datasets, accessible under conditions.
The users should cite the Economic Research Forum as follows:
OAMDI, 2013. Micro and Small Enterprises Survey (MSEs), http://www.erf.org.eg/cms.php?id=erfdataportal. Version 1.0 of Licensed Data Files; Morocco MSEs 2002. Egypt: Economic Research Forum (ERF).
Disclaimer and copyrights
The Economic Research Forum has granted the researcher access to relevant data following exhaustive efforts to protect the confidentiality of individual data. The researcher is solely responsible for any analysis or conclusions drawn from available data.