Transcript Evaluating national statistical systems in the CARICOM

Patrick Kent Watson
Sir Arthur Lewis Institute of Social and
Economic Studies
The University of the West Indies
St. Augustine
1.
Introduction

Some leading questions in designing the study

Conducting the study

Specific concerns about the data categories
1.
Conclusion
 National Statistical Offices (NSO) in CARICOM region have
relatively long history of collecting and disseminating data.
 Anecdotal evidence suggests that there is a lot of room for
improvement in the quality, coverage, timeliness and
methods of dissemination.
 Data inadequacy, and response to it, is going to become
and even more pressing issue as these countries seek even
further social and economic development:
 “Good decision-making, in both the public and private sectors,
depends on the timely availability of information that is
accurate, current and appropriate for its intended use.”
(GORTT 2007).
 Sufficient concern expressed about data
inadequacy to warrant a comprehensive review of
the data collected and disseminated by the NSO of
the CARICOM region.
 Imperative to determine where we are now,
where we wish to go, and the steps needed to
accomplish that objective.
1. Given where we are now, what are the data needs and
what are the gaps?
2. Given where we want to go, what are the needs, what
are the gaps and what is the roadmap to get there?
3. Are there flaws in the measurement of data currently
collected?
4. Are there data currently collected that we really do not
need?
5. How do the NSO score on data sharing and data
dissemination? What are the methods used to share and
disseminate data?
6. How well do the CARICOM NSO compare with each other
NSO in quality, coverage, timeliness and dissemination
especially give the existence for some time now of a data
harmonisation exercise?
7. How well do the CARICOM NSO compare internationally
in quality, coverage, timeliness and dissemination, both
with countries of similar development, with NIC and with
developed countries?
 For each of the generic data types
mentioned (macroeconomic data etc.) and
for each of the CARICOM NSO, it will be
necessary to identify
 the actual data series currently constructed
 the instrument(s) used to obtain the
component parts (Census, Household
Budgetary Survey, Survey of Business
establishments etc.).
 Some pertinent questions to be asked include
1. (1) what sampling technique?
2. (2) what coverage?
3. (3) what margin of error?
This will involve a careful study of the documentation
used and a detailed study of the procedures followed
to obtain the data.
Also likely to involve the cooperation of the Heads of the
NSO as it is very likely to involve interviews with the
key personnel of the NSO (Head, Chief Statistician,
Heads of Division, Statisticians).
 It will almost inevitably involve interviews with
respondents to questionnaires and with final users
of the data and appropriate instruments will have
to be designed to do these things.
 Final users shall be asked to express an opinion on
what is currently available and what they would
like to see available.
 Users are particularly concerned about quality,
ease of access, coverage and timeliness of data.

National Income and Components
 Perhaps the most widely used indicator of development is
the level and growth of real gross domestic product (or
some other income measure) per capita.
 This requires data on the real gross domestic product
(GDP) and on population.
 GDP is based on the ‘value added’ of sectors so that all
activities that add value should be included for coverage
to be complete.
 Questions about quantification and valuation of inputs and
outputs inevitable arise as well as about the completeness
of coverage.

National Income and Components (cont’d)
 The answers will differ according to the sector of activity.
 A matter of some interest shall be compliance with SNA
1993 now that SNA 2008 is about to be implemented.
 I am particularly concerned about the non publication
(annually) of Supply and Use Tables (SUT) as is required by
SNA 1993.
 A basic SUT has at least one major advantage: it allows for
cross-validation of the data.
 A more detailed SUT allows for the construction of InputOutput and Social Accounting Matrices which have
widespread use in modern policy analysis and
implementation.
Savings and Investment
 Given the importance of investment in the process of
growth, and the words of Sir Arthur Lewis about the
fundamental role of savings in the developmental process
(Lewis 1954), it is extremely important to have confidence
in the data on these two entities.
 There is a lot of uncertainty about the nature and quality
of this data.
 Private savings calculated as a residual
 Private investment largely absent
 What about Flow of Funds Accounting?
Demographic data
 The CARICOM regions are perhaps most sorely lacking in
‘social’ data.
 It is important to have confidence in data that go into
indices of ‘well-being’ like nutritional intake, life
expectancy and illiteracy rates.
 These are used (along with economic data) to construct
the much touted Human development Index and the
World Development Indicators.
 Yet many of these may be based on approximate (non
existent?) data.
 The resource implications must be worked out
 Data may be ordered by prioity items, based on
developmental priorities.
 Enough lip service has been paid to the
development of the national data bases
 Time to put our money where our mouth is