Transcript Uganda - United Nations Statistics Division

SYSTEM OF ECONOMIC
SURVEYS in UGANDA
Addis Ababa, 17th October 2007
Musana Atai Imelda
Principal Statistician, Uganda Bureau of Statistics.
Mobile. 256-772680087
[email protected]; [email protected]
www.ubos.com
What are Economic Surveys?
A Business Inquiry is an Economic Survey
aimed at collecting data from all Industrial
sectors necessary for the computation of
the main economic indicators such as:
Gross Output
Value Added,
Intermediate Consumption
Operating surplus
Productivity among others.
Types of Surveys needed and
why?




Censuses/Large Surveys conducted every 5 or
10 years (depending on availability of funds).
Inter-Censual Businesses Surveys or Annual
Surveys, conducted every 2 or 3 years
(depending on the availability of funds). This will
provide an update of the census data
Other Adhoc Surveys, to be conducted when
there is no Inter censual Survey or otherwise.
..\History.doc
Objectives of Business
Inquiries (BI)

Compute data on main Economic
Indicators such as Value Added by Sector, Gross Output
by sector, and Intermediate Costs among others
Value Added by sector is the basis of
weight development in GDP estimation
thus improvement of estimates of GDP
 Determine sector productivity and
contribution of each sector to economic
growth

Objectives of BI

Development of Input –Output Tables
(IOT), Supply and Use Tables (SUT) which
finally feed into the Social Accounting
Matrix (SAM).

Provide data required for Macro Economic
Modeling as required by the Ministry
responsible for planning and policy
formulation

Provide a framework for monitoring
economic development over time
Objectives of BI

Development of weights for use in index number
computation such as Index of Industrial
production (IIP), Producer Price Indices (PPI),
Construction Sector Indices (CSI) etc.

Sampling Frame for other detailed economic
surveys for instance:



The Foreign Private Capital Flows survey undertaken by Central Bank;
The Investment Surveys by Investment Authority and other such
surveys
Facilitate in-depth analysis by Researchers and
their policy implication
What are the good practices in BI survey
design and organization?
K
E
Y
S
T
E
P
S
1
Questionnaire Design
2
Sample Design and Selection
3
Training of Staff
4
The Pretest and Pilot
5
Organization of Field Work
6
Organization of Office Work
7
Data Editing
Questionnaire Design

Separate questionnaires should be designed for the
different sectors to take in account the peculiarities in
each sector and collect data that will be useful.

The large questionnaire which is more detailed issued to Business
Establishments with Final Accounts, containing questions relating to:
 The small questionnaire contained less detailed questions but relating to
data required as mentioned and administered to those establishments
that did not keep final accounts

The questionnaires should be able to provide data
relating to:


Activities and location of the business;
General information on employment and ownership and statement of Fixed
Assets;
 Capital expenditure at the end of each year; and
 Total sales, income and costs.
Types of Questionnaires
Description
1
Agriculture and Fishing
2
Mining & Quarrying
3
Manufacturing
4
Construction
5
Trade
6
Hotels & Restaurants
7
Business & Personal Services
8
Finance
9
Insurance
10
Education & Health
11
Non-Governmental Organization
12
Small Trading & Services (All Service Sectors
13
Small Mining & Manufacturing (All production sectors)
New Questionnaire to
be administered in the
next Survey due to the
need for Institutional
Satellite Accounts.
Sample Design and Selection

The sample design for an economic survey is
complex as compared to the Household Survey
however the sample should be at least 5% of the
population. Procedure could be:
 Stratify
the population by industrial classification
 Within each Industry, further stratify by number of
employees based on predetermined employment size
bands for each industrial group
 e.g 1 – 4; 5 – 9; 10 – 19; 20 – 49; 50 – 99; and 100
Plus.
Sample Selection

To minimise the expected variance of total
turnover (for the large establishments) over all
strata, within each stratum CUT -OFF sampling
method should be applied for some business
establishments whose employment is equal or
above a certain employment size band, are
selected in totality.
 (in
Uganda all businesses employing 20 or more
persons were selected, about 30% of the population)

For the remaining businesses Proportional
Allocation was used to determine sample sizes
for each stratum and thereafter businesses were
selected using Systematic Sampling.
Training of Staff

Need for in-depth training of staff aimed at:
 Acquainting
staff with the basic accounting ideas and
terminologies,
 Understanding the different coding system sucgh as
the ISIC, CPC etc
 Understanding GDP and System of National
Accounts, SNA 1993
 Understanding the structures of Final Accounts
 Understanding the proposed UBI questionnaires to be
used
 Extracting information from the Final Accounts and an
example of a questionnaire
Training of Staff


Examinations and Role Plays
After each training sessions, participants should be
subjected to examinations to test their ability to extract
data from specific final accounts.

Further, Role-Plays should be emphasised because it is
argued that the presentation of an enumerator before a
General Manager will in most cases determine the kind
of response obtained.

Good public relations and building of rap pours between
the enumerator and the respondent increases the
response rate

Finally the Issue of confidentiality of the information
should be more than emphasized during the training. All
staff must take the Oath of Secrecy.
Pretest & Pilot Survey

The Pilot Survey should be undertaken
with a few selected Business
Establishments. This will help:
 polish
up the instruments to be used,
 remove any redundancies and
 further train the enumerators in the
assignment among others things
Organization of the Field work

Field Teams should be properly designed.
Teams are organized in such a way that they are
centralized and administer questionnaires in a
specific geographical location this has the
advantage of reducing the costs of the Survey
 (Ideally it would be good to allocate questionnaires of a specific sector to a team
thus ensuring consistency).

The Field Team should have a Supervisor/Team
Leader, Enumerators and Drivers, each with
clearly spelt out assignments:
Organisation of Office work

Proper organisation of
Office work with TEAM
SPIRIT is a key to a
successful Survey.

The office activities
involve among other
things:







preparation of fieldwork
activities,
providing technical support,
staff payments,
receiving of returns from the
field,
editing of returns,
data capture,
managing and maintaining of
vehicles; and
other administrative work etc
The
different roles should however be distributed among those
involved in the office activities.
Data Editing

Data editing should be taken seriously and it should be
undertaken both at the field and in the office.
Field Editing

In the field the interviewer/enumerator should be the first
Editor who would then pass on the completed
questionnaire to his/her supervisor for further editing
including:




Allocation of appropriate ISIC, CPC codes
Ensuring the return is completed
Ensuring the questionnaire is balancing etc
When an error is discovered, the supervisor re-visits the
Establishment to rectify the problems with the concerned
authorities. This should be done before submission of
the questionnaire to the office
Office Editing

Office Editing would include:

Receipt and assigning of different return
codes depending on the type of
questionnaire.


The return codes could be: 1. Good Return; 2. Fixed Assets
Missing, 3. Refusal 4. Partial data, no employment etc.
The return type codes assigned help in
analysing the status of each establishment
and the performance of each interviewer
at a given time.
Office Editing

Ensuring that the right
ISIC & CPC codes were
allocated and all the
administrative units are
coded.

Undertaking Intelligent
Edits.

E.g issues like ‘Cost of
Staff’ being greater than
the ‘Total Income of the
business’ should be
queried
Challenges in Data Processing




Selection of the right personnel
Training and Clear allocation of duties
Developing a proper system of Data Entry, what
application should be used?
Does it provide for
 self
editing/checking?
 Error Levels?
 Record workloads for each Data Entrant?
 Access levels for all staff with passwords?
 What are the security checks?
Challenges in Data Analysis





What Statistical Package will be used, SAS or
STATA or otherwise?
Do you have the Tabulation Scheme, was is
discussed and agreed?
How have you provided for weights and
consequently grossing up?
The fact that businesses have different
accounting periods, this is a challenge;
Different currencies used. During analysis all
currencies should be converted into one.
Uniform exchange rates should be used for all
businesses providing data in say dollars
Challenges in Data Analysis



Imputation for missing data is a challenge what
approach should be adopted?
Provision of District estimates is not possible
since the sampling was done using the Industry
(ISIC), and employment size bands as the main
sampling variables. Challenge is to provide
Gross Output and Value Added by District, what
should be done?ugandamap.jpg
The formulae for computation of Value Added
and Gross Output vary from sector to sector and
this should be provided for during analysis.
Challenges of Conducting a BI

How to ensure that BIs are conducted at least every 5 years
although this will depend on the availability of funds. In some
countries every 10 years. Undertake inter censual surveys every 2
or 3 years to update the census information? To what extent can this
be achieved?

Induce Governments to put more emphasis on the Business
Inquiries and therefore fund them. In Uganda most of the Inquiries
have been mainly funded by Donors

Apathy by the Business Community due to the time taken to fill a
specific questionnaire, how can we increase their interest?

Most of the data required is obtained from the Final Accounts and
access to these accounts normally requires prior clearance from the
Board, Top Management, and this takes a lot of time.
Challenges of Conducting a BI

Poor training may lead to poor filling of questionnaires.
Need for thorough understanding of Final Accounts and
the different Audit formats

Undertaking intelligent edits to determine whether the
data has been falsified, Need to put in checking
questions.

Proper coding of activity codes according to the UN ISIC
Rev 3. When should we adopt ISIC Rev 4?

How to handle secondary activities in the business, Can
we apportion value Added to the secondary activities?
Challenges of Conducting a BI

Businesses which are partly private and partly
government, difficult to determine their
contribution.

The different accounting periods adopted by the
businesses is a great challenge, how should the
be handled?

Is the computer system for data entry and
analysis integrated? To what extent are security
systems provided in the system?

Penalties for Non Response, to what extent can
the Statistics Office invoke the law and ensure a
response is obtained?
Challenges of Conducting a BI

How do you achieve response rates above 60 percent?

How big should the sample be?

How can we determine value addition in specific
sectors?

The sampling frame for the survey, how often should it
be updated?
Challenges in surveying the
Informal
sector
 The informal sector is one of the difficult sectors to
measure in the country. This sector cuts across all the
Industrial sectors but especially Manufacturing, Trade,
Agriculture, Mining and Quarrying, Transport. It has
been difficult to determine the actual contribution of the
informal sector to the overall growth.

The sampling frame for the informal sector is difficult to
determine. Propose that we use the Businesses
Register and population figures to determine the location
of the informal sector activities and then administer
questionnaires using simple random sampling, is this
feasible?
Challenges in the Informal
Sector

The definition of the Informal Sector is a challenge. Businesses employing
less than 5 persons excluding the Service Sector? Businesses without Final
Accounts, Businesses without a fixed location?

The nature of the informal sector businesses, which have a life span of 3
months. After developing the sampling frame one goes to administer the
questionnaire and the business has closed.

Border-to-Border businesses, how to measure their contribution to growth?

Household Enterprises, how to measure their value added especially when
the purchases are both for the household and for the business.

Most of the businesses are not registered, do not have a name, and most
times the respondents are not the owners and therefore cannot provide
accurate information.
Conclusion

Undertaking a Business Inquiry is a
REAL challenge that we should
address and try to see solutions
around it.

The presence of a highly productive
Informal Sector, which is not
measured is however a bigger
challenge.
THANK_YOU