This is the html version of the file http://unstats.un.org/unsd/accsub/2002docs/sa-02-4.pdf.
Google automatically generates html versions of documents as we crawl the web.
Inter-agency Meeting on Coordination of Statistical Activities SA/2002/4 New York, 17-19 September 2002 11 September 2002 Item
Page 1
Inter-agency Meeting on Coordination of Statistical Activities
SA/2002/4
New York, 17-19 September 2002
11 September 2002
Item 5 of the provisional agenda
IMPROVING THE QUALITY OF DATA USED FOR INDICATORS FOR THE
MILLENNIUM DEVELOPMENT GOALS AND TARGETS
Note by UNSD
1.
Introduction
The Statistics Division reported to the ACC Subcommittee at its 35
th
session, held in
Vienna, 18-20 September 2001, on arrangements for the use of statistical indicators to
monitor progress in achieving the Millennium development goals set out in the
Millennium Declaration. This report described the action taken by the General Assembly,
the Assembly's resolution on follow-up (A/RES/55/162), the preparation of the "Road
Map" report of the Secretary-General (A/56/326) and agreed arrangements for the
preparation of indicators for the Secretary-General's annual reports and for coordination
with agencies providing data and with UNDP, which is responsible for managing the
preparation of country reports on achievement of the goals and targets.1
The Sub-committee commented on the work and plans of the United Nations Statistics
Division relating to the review of global progress towards the goals of the United Nations
Millennium Declaration and reiterated the need to approach all the appropriate source
agencies relevant to the targets listed in the document.
2
Two inter-agency expert group meetings organized in 2001 by the Executive Office of the
Secretary General, including representatives of the Department of Economic and Social
Affairs and the Statistics Division, United Nations Development Group, Bretton Woods
institutions and OECD, confirmed the respective responsibilities for monitoring and
reporting at the global and national levels.
2.
Review of the process for the preparation of the 2002 Report
Based on the inputs provided by each of the designated agencies on the relevant
target/indicators and on the outcome of the discussion at the two interagency and expert
group meetings held in New York in March and April 2002, the Statistics Division
prepared the “Report of the Inter-agency Expert Group on MDG Indicators”, for the
EOSG’s use in preparing the Secretary-General’s report to the General Assembly.
3
The final report to the EOSG was prepared according to the following schedule:
1 The relevant General Assembly resolutions and reports are available through the Statistics Division site
millenniumindicators.un.org.
2 Administrative Committee on Coordination, Report of the Subcommittee on Statistical Activities on its
thirty-fifth session Vienna, 18-20 September 2001, ACC/2001/2.
3 United Nations Millennium Development Goals, Data and Trends, 2002. Report of the Inter-agency
Expert Group on MDG Indicators. ST/ESA/STAT/120, in http://millenniumindicators.un.org; and General
Assembly document “Implementation of the United Nations Millennium Declaration Report of the
Secretary-General”, A/57/270, also available through http://millenniumindicators.un.org

Page 2
2
6 May – agencies submitted drafts of their inputs to UNSD
22 May – the draft was circulated for comments to all agencies participating in the
effort and other relevant offices within the UN system.
31 May – the final version, incorporating all comments was submitted to the EOSG
Parallel to the preparation of the report, work has been ongoing since mid-2001 on the
MDG Database of UNSD.
4
Data series on the 48 MDG indicators were provided by the
agreed source agencies at different stages. Following the December 2001 meeting, where
agencies agreed on a near-final list of indicators, agencies began work to update the
needed series and sent them to UNSD. In some instances, country data series were
submitted to UNSD for incorporation in the database prior to the submission of the
inputs–regional/global figures and text–for the report. In other cases, while regional and
global estimates were sent for inclusion in the report to the EOSG, country data were not
at all available or were not submitted.
Checking data quality
One key element in preparing the quantitative assessments on developmental goals is
ensuring that the data used as a basis for the indicators are the best available. The 2002
document was the first progress report in a fifteen-year process. Agencies will work
together to better understand and reduce the uncertainty in the data in the coming-year
reports.
The time allowed for the preparation of this year report was hardly adequate for
completing the checking of data and for further exchange of information and discussion
with the agencies, after the expert group meetings.
MDG database and submission of data series
The MDG database at UNSD is the focal point for interagency work on data compilation
and for the improvement of data used in the reporting. It contains data series related to
the 48 millennium indicators, continuously updated as new data become available, and
technical notes on the indicators and the sources used.
To the extent possible, there should be consistency between the country series contained
in the database and the global and regional figures presented in the annual and
quinquennial reports and which will also be included in the database in the next few
weeks. This is difficult for several reasons. First, country data change over time and
while changes can only be reflected in regional and global values once a year in the
report, the database can be updated more frequently or with a different schedule. Also, in
some cases, regional estimates are not based on a specific country data series even when
country data exist, due to poor quality of existing data. Finally, in this year’s report, given
the tight schedule for the submission of the document to the EOSG, some regional/global
figures were last minute calculations, not fully consistent with the data series that had
been provided earlier for the MDG database.
Some issues for discussion on the improvement of the process are the following:
4 In http://millenniumindicators.un.org

Page 3
3
Inter-agency expert discussion on data issues. Now that the reporting process and the list
of indicators have been fixed, time can be devoted to more in-depth discussion and
consultation on the improvement of the data used. Also, more work can be devoted
specifically to assess the quality and availability of data for some of the more problematic
indicators. To this end, UNSD in close collaboration with the agencies can identify
indicators for special attention. (See also section below on “Conceptual and
methodological issues”.)
Consistency between regional and country data. One first step to increase consistency
between regional and country data could be improving the mechanism of submission of
data series and regional figures/estimates. There was a disconnection between the
regional data sent for the preparation of the main report to the EOSG and the country data
submitted earlier. To the extent possible the MDG database should contain only the data
series on which the regional/global averages or estimates are based, and a description of
the methodology used for estimates. For the next reports, it would be helpful to have the
regional and country data sent to UNSD together, with possibly some explanation of how
they relate to each other, and on the methodology of regional estimations.
Inter-agency and expert group meetings. The main forum for further discussion and
decisions on future steps for the improvement of data quality could be, as in the past,
interagency and expert group meetings that bring together the various partner agencies
and the experts on selected statistical fields of interest. The possibility of having one or
two thematic meetings on the more problematic indicators/data series, rather than
meetings covering all 48 indicators, can also be considered, for instance, grouping
together all the health indicators for one meeting and the economic indicators for the
second meeting.
3.
Coordination with UNDP work at the country level
Country reports on progress towards the MDGs are intended to raise public awareness
and help focus the national debate on specific development priorities and on the
necessary action in terms of policy reforms, institutional change and resource allocation.
In its capacity as United Nations coordinator at the country level and chair of the United
Nations Development Group, UNDP is supporting United Nations country teams in the
preparation of national reports. Country teams generally include full representation of the
partner agencies.
A necessary condition to obtain better data for the MDG indicators is to improve the
capacity of countries to produce data. The preparation of the country reports represents an
opportunity to help build national capacity for monitoring and reporting. In this context,
coordination between global and national reporting would help identify priorities for
building national capacity for the production, analysis and dissemination of data.
Finally, the country-based information compiled and analysed by the UN country teams
can be used as additional inputs –beside international sources – for the report to the
General Assembly, after data quality and sources are carefully scrutinized. Analysis of
the data and data sources used in the country reports will provide for a further review of
existing problems and data gaps in countries.

Page 4
4
4.
Harmonization of regional groupings
At the last three meetings (in 2001 and March and April 2002) partner agencies agreed to
present the regional data and estimates according to the UN grouping for geographical
regions and sub-regions and for developing, developed and least developed countries, as
given in the Standard Country or Area Codes for Statistics Use.
5
However, for some
agencies it was impossible to recalculate estimates or averages with a different grouping
of countries in the short time available.
In order for the report to have a consistent format and to avoid confusion in the
presentation of the findings, it is essential that data be presented with a homogeneous
format in all sections. Also, data contained in the MDG database should be presented, to
the extent possible, by region and countries. To this end, the latest regional averages and
estimates for the M-49 groupings should be sent to UNSD as soon as possible in order to
be posted in the website.
5.
Coordination of global surveys
Data for the compilation of the MDG indicators are provided by specialized UN-system
agencies, Bretton Woods institutions, OECD and the World Trade Organization. They are
derived from data submitted to relevant agencies by national statistical offices or directly
collected in countries through surveys sponsored and/or coordinated and implemented by
the agencies. Agencies are responsible for ensuring international coordination of the
content and geographical coverage of surveys and for analyzing the data for MDG
reporting to the extent possible, as long as they are consistent with international standards
and with the agreed indicators for the MDGs.
Some of the survey programmes pertinent to the data for the MDG indicators are the
following:
Multiple Indicator Cluster Survey (UNICEF)
HIV/AIDS, various epidemiological surveys (UNAIDS)
LSMS (World Bank)
TB/DOTS (notification programme); Roll Back Malaria (WHO)
Pilot surveys in selected countries to test/improve methodologies of data collection on
labour force (ILO)
Pilot surveys in selected countries to test/improve methodologies of data collection on
nutrition (FAO)
Pilot small-scale studies on education/literacy (UNESCO)
Access to PCs and internet (ITU)
Secure tenure and slum improvement (Habitat)
Some issues to be considered for discussion:
Public access to the original agency databases and technical notes drawn on for
submission to the Millennium Indicators Database at UNSD
5 Standard Country or Area Codes for Statistics Use, ST/ESA/STAT/SER.M/49/Rev.4 (United Nations
publication), and in http://millenniumindicators.un.org/unsd/methods/m49/m49regin.htm

Page 5
5
Schedule of updates
Scope and geographical coverage of surveys
Other surveys planned or ongoing not included in the list above
Other issues related to the use of global surveys data for MDGs monitoring
6.
Conceptual and methodological issues
As part of the common effort to better understand and reduce the uncertainty in the data,
partner agencies and UNSD could focus on some specific areas/indicators where, for
different reasons, obtaining data has appeared to be more difficult.
Inconsistent internationally published series
There are indicators where a variety of sources and published data series exist and
possible confusion in their use may arise. They are generally data obtained through
differences in estimate methods, basic data sources, recency or frequency. .
Examples:
Poverty estimates: a review is needed of the different methods used by World
Bank and UNCTAD, household-surveys-based estimates versus SNA-consistent
estimates
Infant and child mortality estimates: there are differences between the UN
Population Division, UNICEF and World Bank data and various methodology
issues that need to be clarified.
Other agencies are also publishing reports on the MDGs (for instance the UNIFEM
“Progress of the World’s Women” and HDR). We should try to ensure some consistency
between MDGs data series/global figures presented in the GA report and in the Web site,
and those presented in other reports/Web sites.
Insufficient data for trend analysis
Other estimates are produced by a single agency and represent the only international
source available. They are however still in the process of being refined and have not been
used consistently over time, so that trend data are not available.
Examples:
Maternal mortality
Secure tenure: regional and global estimates of the secure tenure index were
provided for 1993 only. The methodology used to calculate the index needs to be
further reviewed and discussed.
Insufficient coverage
In other instances, country coverage for the underlying data is limited and therefore
trends and regional analysis are difficult.
Examples:
Women’s paid employment in non-agricultural sector

Page 6
6
Youth unemployment
Ratio of children orphaned by AIDS to non-orphaned children who are attending
school
Households using solid fuels
Problematic estimates
Finally, there are data sets that because of poor quality of existing data in countries
present a high degree of uncertainty.
Examples:
Survival rates from grade 1 to grade 5. For some countries, data indicate drastic
changes from one year to another.
Points for discussion
In conclusion, the meeting may wish to:
--Comment on coordination and harmonization issues which agencies may have
experienced in the course of preparation of the 2002 inter-agency report on indicators for
the Secretary-General's report to the General Assembly;
--Suggest indicators or topics which may need special efforts, such as a technical
meeting, in order to reach a more acceptable level of reliability, especially for trend
analysis, and coverage;
--Comment on complementary country programmes which can be of greatest
value in meeting national and international requirements for monitoring achievement of
the Millennium development goals;
--Call attention to agency programmes which are expected to yield new datasets
and insights on development progress in the next year;
--Take note of the need for updating and harmonizing the Millennium indicators
in the Statistics Division's Internet database according to agency schedules for
completing new datasets, and in any case no later than March of 2003 for the next report
of the Secretary-General.