39652
Volume 3, No. 1, February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
What is ICP? Notes from the Global Manager
The International Comparison Program (ICP) Dear Readers,
is a global statistical initiative involving nation-
al, regional and international agencies. It pro- The conclusion of this round of the International Comparison Program is in sight
duces internationally comparable price levels, as all of the regions plan to publish preliminary results for Actual Final Consump-
expenditure values, and Purchasing Power Par- tion by the end of this year to be followed shortly by preliminary estimates at the
ity (PPP) estimates informing users such as na- full Gross Domestic Product level. These preliminary estimates will be based on
tional policy makers, economists, international the most recent expenditure shares calibrated to the 2005 GDP. Final Purchasing
organizations and academics about the relative Power Parities (PPPs) and volume indices will be prepared after mid-year 2007
purchasing power of citizens of countries. The when most countries will have final estimates for the 2005 GDP. The respective
PPP between two countries is the rate at which global publications will follow the regional schedule.
the currency of one country needs to be con-
verted into that of a second country to ensure Having said this, there is much more to do and many challenges remain. Data
that a given amount of the first country's cur- collection remains to be done for capital formation, housing, government and
rency will purchase the same volume of goods health in addition for the Ring Comparison. These are "difficult" items to mea-
and services in the second country as it does sure. However, some of the major hurdles regarding measurement concepts were
in the first. resolved during a recent meeting with the regional coordinators.
In This Issue Efforts to date have concentrated on the collection of price data. However, the
final results will be based upon expenditure shares for 155 basic headings com-
Notes from the Global Manager ...................... 1 ing from a breakdown of each country's GDP to compute PPPs and per capita
measures. Therefore, a recent meeting was held with national account experts to
News in Brief ................................................. 2
� New Working Group Shaping Data agree upon methods to evaluate expenditure shares and determine where data are
Access Policy inconsistent between countries.
� Regional Coordinators Reviewed Program in
Washington The Global Office and the regional coordinators, for the second time as a group,
� ICP e-Newsletter Published in 5 Languages
� ICP Tool Pack Simplified for Regional Use reviewed regional preliminary PPPs and country average prices. Much was learned
� Executive Board to Meet in New York from these exercises about the quality of the data being collected. The primary
lesson learned, however, was the importance for the data review and sharing to
Feature ......................................................... 3 actually take place.
� ICP - Enhancing the Quality of Global
Economic Statistics The ICP has a unique feature that sets it apart from other international statistical
Methodology ................................................. 4 endeavors - a country cannot compute a PPP by itself. A closely related feature
� Harmonizing ICP and HES Data to Calcu- is that a region or the Global Office cannot produce satisfactory results unless all
late PPPs for the Poor countries and regions adhere to the same methodology and procedures.
� Compilation, Validation and Documenta-
tion of Expenditure Weights
� SPD Software Improves Product List The agreed-upon principle to ensure a satisfactory outcome is based upon trans-
Development parency. Transparency requires that countries within a region see the average
prices submitted by each other to make sure the same products were priced, and
Regional Spotlight .......................................... 6 that the coverage of outlets is comparable. In addition, each regional coordina-
� Lessons Learned: The ICP Latin
America Experience tor needs to see and understand the average prices estimated by other regions to
� Strategic Framework for Capacity ensure they are following the same methodology and collection principles to not
Building in Africa bias the global comparison.
� Regional ICP Board Met in Asia
These principles are essential and will be the basis to determine what is published
Regional Reports in Brief .............................. 8 in the global report. This does raise questions about data confidentiality and data
Country Spotlight .......................................... 9 access. As a result, a Data Access Working Group has been formed to consider
� Lebanon: ICP & Tool Pack Offer New these many issues, with more discussion in another part of this newsletter.
Vision of Data Collection
�� ICP Experience Will Refine Egypt's CPI
Matching Indonesia's Construction In closing, I want to thank those who contributed material for this newsletter,
Materials Survey with the ICP which focuses on the theme of capacity building. As you can tell from this issue,
� Nepal Finds ICP Useful for Building there are many developments. Please visit the ICP website (www.worldbank.org/
National PPI data/icp) to keep regularly posted on new events and material.
Regards
Calendar of Events ..................................... 12 Fred Vogel
Published in Arabic, English, French, Russian and Spanish. www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
News in Brief
New Working Group Shaping Data Access Policy ICP e-Newsletter Published in 5 Languages
Last December, a Data Access Working Group was established and Notably, the e-Newsletter has gone through considerable changes
itsmembersselectedbytheGlobalOffice,incloseconsultationwith in format, presentation and substance, beginning with the last is-
the Chairman of the ICP Executive Board. The Working Group sue in November 2005. A new section was introduced in the No-
was mandated to develop a policy to protect the integrity of ICP vember issue to focus on "Country Spotlights." The section pro-
micro data holdings, and to formulate data access guidelines for re- vides National Coordinators and other country experts with space
searchers and the general public. to contribute succinct articles on their experiences. That issue also
marked another important milestone: The e-Newsletter was pub-
The management of the ICP is shared by the Global Office and lished and distributed in five languages for the first time, including
five regions headed by Regional Coordinators (RC). The RCs co- Arabic, English, French, Russian and Spanish.
ordinate the statistical process in their respective regions and man-
age the horizontal review of data provided by the national authori- The enhancements continue in the current February 2006 issue. A
ties of participating countries. For the latter, they need access to new Methodology section highlights procedural guidelines, meth-
micro data from their participating countries, which is the rationale odological developments, and summary reports of research papers
for developing a clear policy for data submission and data access and case studies. The e-Newsletter is posted on the ICP website
to micro data holdings. for the wider readership.
At the same time, many researchers seek access to unpublished data ICP Tool Pack Simplified for Regional Use
to expand knowledge boundaries on issues like measuring poverty The ICP Tool Pack (TP) software, in use by Regional Coordina-
and improving methodology. If we want to promote the use of tors, is now distributed in a new and simplified design. This was
PPPs, the guiding principle for a balanced data-access policy should done in response to requests made by the CIS and Africa Regions
encompass appropriate use of data by facilitating access, while re- to remove the Web-based interface, which required Microsoft's
specting the confidentiality and integrity of micro data. IIS software component to be installed on the computer. The IIS
requirement posed a major problem due to stringent IT policies
The Working Group met on 6 February 2006 in Washington to con- against installing IIS in the Africa and CIS Regions. The simpli-
sider these multiple issues. The group has prepared preliminary rec- fied TP brings all user interaction into the Windows interface of
ommendations to be considered by the RCs and subsequently by the the popular Data Processing Module (DPM). It also comes in a
ICP Executive Board. "client" version of the installation, which is necessary in offices
where several users work on the same data in a networked envi-
Regional Coordinators Reviewed Program in Washington ronment. The new TP allows the main DPM installation (includes
The Regional Coordinators Workshop (February 7-10) included the Database) to be installed on one central computer, while the
an indepth review of Quaranta Tables and supplementary analysis "client" installations (without the DB) reside on other computers
provided via the Dikhanov Tables for four of the five ICP regions. on the network. Both the client and main installations can update
The Regional Coordinators and the Global Office also reviewed data and generate reports.
technical issues and the timetable for health, housing, equipment,
construction, and government. The work plan for the Ring Com- The TP developers have taken utmost care in this simplification
parison was also reviewed and tentatively agreed upon. The work process to ensure that any new patches (changes) to the software,
plan will be reviewed by the regions and participating countries including the present one, will NEITHER affect the data (product
with the final version to be prepared for the next meeting of the lists, specifications, prices � individual or average) NOR the func-
ICP Executive Board. tionalities for those using the earlier Web-based version.
Varioussessionsof theworkshopsalsobenefitedfromvaluablein- While simplifying the TP installation, the Global Office has in-
put from TAG members Alan Heston (chair) and Kim Zieschang, cluded a few important enhancements based on expressed and
and Global Office consultant Derek Blades. Actions flowing from anticipated requests. These include the Country and Dikhanov
this workshop include the finalization of guidelines for the review diagnostic reports, customization in Index number and Quaranta
of national account data, and data collection for equipment, con- processing, and flexibilities in viewing index number reports.
struction, health and housing (The full set of materials and min- For more information, please e-mail Vilas Mandlekar (vmandlekar@worldbank.org).
utes is available on the ICP website.)
Executive Board to Meet in New York
Determining if a country/region meets the requirements to be The ICP Executive Board will meet in New York City on March 6,
included in the comparison was another key theme in the work- 2007, ahead of theannual meeting of theUnited Nations Statistical
shop. The policy guidelines from this workshop --Requirements Commission. The Board will review the status of data collection,
for Participation in the International Comparison Program -- will publication plans, issues raised by the Data Access Working Group,
be presented to the ICP Executive Board. and the outcomes of the regional coordinator workshop such as
policies for inclusion in the ICP.
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Feature
ICP -- Enhancing the Quality of Global Economic Statistics tions be closely checked for variations over time or from place
Kimberly Zieschang to place to guard against direct price comparisons of dissimilar
The ICP is widely known as the premier source products. When differences in specification arise, compilers must
of information on international price compari- take measures to assess the importance of the difference and ap-
sons as well as, on the volume side, international ply appropriate statistical methodology to correct relative price
comparisons of gross domestic product (GDP) comparisons for the difference. The current ICP round's SPD
and its consumption, capital formation, and net methodology is an internationally consistent approach to prod-
export expenditure components. Accordingly, uct specification. It allows more effective detection of significant
the ICP has an active user community critically variations in product specifications from country to country. The
evaluating the results of each round, and some data quality is- same methodology also offers rigorous options for correcting
sues have surfaced in reviews of earlier ICP rounds. The current relative price comparisons for differences in product specifica-
round addresses these issues with two new initiatives: the Struc- tions. This may involve further price collection on a revised speci-
tured Product Description (SPD) method for systematically de- fication found in the areas compared, or it may involve methods
veloping product specifications, and a renewed effort to obtain of statistically correcting for the differences using, for example,
improved GDP expenditure weights for the current ICP round's regression models explaining country-to-country price variations
155 basic headings. A valuable by-product of these new ICP data by variations in the coded product characteristics information in-
quality initiatives is to support and, in some instances, set in mo- herent in the SPD approach. In many ICP participating countries,
tion national and international efforts to improve the quality of application of the SPD approach has led to new insights into the
national accounts and price statistics of the countries participat- adequacy of national product specification practices for headline
ing in the ICP. price indicators such as the consumer price index (CPI).
Identifying data quality problems. The ICP does this, first, by Encouraging continuous improvement in the quality of national
identifying inconsistencies among national statistics on prod- statistics. While the first step to quality improvement is knowl-
uct-related breakdowns of GDP as well as national price sur- edge of quality problems, the second step is to change survey
vey practices in need of management and/or methodology and/or
improvement. Taking GDP compilation practices and/or the staff, equip-
product detail first, because "regional
The ICP has become an important force for building ment, and software complement to address
the ICP comprises an ag- statistical capacity in Africa. .... It has helped those problems. The lack of published prod-
gregate of purchasing power to identify areas where national statistical compilers need uct detail in GDP by expenditure in many ICP
parity (PPP) calculations for technical assistance to address data quality issues in both participating countries has spurred a new focus
155 detailed basic headings price statistics and national accounts .... it has spurred
technical assistance providers to sharpen their focus on on the quality of national accounts statistics in
weighted by their shares in the most pressing needs. selected regions of the world, by both national
GDP, it opens detailed GDP " compilers and providers of statistical technical
data to scrutiny by regional assistance. In many instances, issues the ICP
teams of international experts. Countries compiling GDP us- surfaces in the form of suspiciously high or
ing detailed, product-by-product supply and use balancing are low national shares for certain basic headings can have implica-
able to head off quality problems even at fairly high levels of tions for GDP level and growth estimates. In a number of ICP
detail. Yet in many parts of the world, compilers have not yet participating countries, errors in ICP basic heading shares may
been able to undertake a robust supply and use methodology, indicate insufficient coverage of the nonobserved economy and
and detailed GDP statistics usually are not subject to this level of other coverage issues that significantly affect expenditure in cer-
review. Whether countries use advanced and data-intensive GDP tain product groups.
compilation methods such as supply and use, or rely on less reli-
able techniques, detailed GDP breakdowns are rarely subject to a On the quality of national price data, the SPD methodology is
review further comparing one country's GDP product composi- finding acceptance among national price statistics compilers in
tion with another's. Such comparisons are, of course, inherent in countries participating in the ICP. Countries adopting this prod-
ICP calculations. uct specification framework in producing national price indica-
tors can expect to see improvements in their ability to monitor
Turning to price survey methods, price indices must compare and control for variations in product quality (i.e., product specifi-
similar items when assessing relative price differences, whether ... continued on page 11
over time in the usual national consumer or producer price index
context, or from country to country in the context of the ICP. To * The author is Chief of Data Dissemination Standards Division, Statistics
assess the "similarity" of items in price indices, products must be Department, IMF, and member of the Technical Advisory Group of ICP.
described or specified systematically, and the descriptions must The views in this article are those of the author and should not be interpreted as
include the most important price-determining characteristics of the views of the International Monetary Fund.
the products. Good survey practice also requires that specifica-
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Methodology
Harmonizing ICP and HES Data to Calculate PPPs for the Poor ICP BH expenditure in the total household consumption, for dif-
Quantifying the purchasing power differences across countries ferent population categories (e.g. by income level, urban/rural).
and income groups is critical for measuring global poverty and
monitoring the effectiveness of poverty alleviation policies. Since the HES questionnaires are not designed according to the
Crucial steps in compiling such poverty incidences are the con- ICPBHclassifications,someproductsandserviceswillneedtobe
version of the US$1/day international poverty line into the re- regrouped. For example, the household survey questionnaire may
spective national currency units of countries using PPPs and de- collect data on "carrots," "tomatoes," and many other vegetables,
termining the number of people who fall below that threshold. which need to be aggregated into a single ICP BH -- "Fresh Veg-
This measure is a key indicator for monitoring the Millennium etables." In other cases, when the HES questionnaire lacks detail,
Development Goals. data may be split under separate BHs. Obviously, such efforts are
governed more by what is feasible than what is desirable. This
In the absence of poverty-specific PPPs, the common practice is first step of the work will result in data files providing the house-
to use PPPs for aggregate consumption. This has two limitations. hold annual consumption by BH, for as many BH expenditures
First, the PPPs are based on prices of consumption items for as possible.
all countries in the comparison. Consequently, the PPP estimates
for developing countries are unduly influenced by the consump- The work will ultimately cover about 80 developing countries, de-
tion baskets and spending habits of their developed counterparts. pending on the timeliness, quality and scope of national HES. So
Second, the PPPs are derived using national average expenditure far,datafor35countrieshavebeenfinalized.Itisexpectedthatthe
weights. Therefore, goods that are important to the poor and work for the remaining countries should be done by the end of
comprise a large part of their expenditure carry proportionally July. For an additional 40 or so countries in the 2005 ICP Round,
less weight. the required household expenditure data are either unavailable, old
or of relatively poor quality. This group will be treated differently:
A Poverty Advisory Group established to address the limitations the required expenditure shares by income group will be generated
of current PPPs has recommended that poverty-specific PPPs be utilizing an econometric model. This involves conducting econo-
computed for countries where poverty is prevalent, using price metric analysis of unit record data from a number of benchmark
relatives from ICP sources together with weights representing countries, to identify the determinants of household expenditure
expenditure patterns of the poor. The data to construct such patterns for predicting expenditure patterns for those countries
weights can only come from nationally representative household (and time periods) where expenditure share data are not available.
expenditure surveys.
The global poverty-PPP project is being managed separately from
A characteristic feature of the ICP survey framework is that indi- the ICP and is funded by the World Bank, to avoid undue bur-
vidual consumption expenditure by households is broken down den on the ICP program, now going through significant changes
into 90 classes according to the Classification of Individual Con- and improvements. The Asian Development Bank and the Afri-
sumption According to Purpose (COICOP). Some classes are can Development Bank are funding independent regional pilot
further divided into 110 Basic Heading (BH) expenditures. The research initiatives. The global and regional initiatives are coordi-
ICP provides detailed price data for specifications of individual nated to avoid duplication of efforts. The work from this round
items that are classified into the 110 BHs. Though the ICP pro- will be leveraged to further refine the Poverty-PPP methodology
vides a detailed account of regional variations in prices, its expen- in future ICP rounds. The ultimate objectives are to integrate pov-
diture side of the matrix leaves much to be desired. On the expen- erty PPPs into the mainstream work of the ICP and to allow in-
diture side, Household Expenditure Surveys (HES), in contrast, ter-regional cost-of-living comparison within a country.
typically provide a detailed breakdown of expenditures by regions
and income groups. But the price side of their matrix is mostly Compilation and Validation of Expenditure Weights
sparse, if not virtually blank. One very important off-shoot of The new ICP strategic framework is designed to meet the short-
the 2005 ICP Round is the work to integrate the ICP and HES term goal of producing reliable, timely and relevant PPP data and
expenditure classifications. to address the long-term objective of building capacity in price
and national accounts statistics upon which the quality of PPP
For this purpose, a two-step expenditure data compilation work data depends. As part of this effort, an expert group on National
has been initiated. The first step involves computing the weight- Accounts met for three days in February to discuss issues related
ing coefficients representing consumption patterns of the poor, to the compilation, validation and documentation of the expen-
for countries that have reliable and recent data on household diture weights needed to calculate PPPs. In context, Basic Head-
expenditure. Where possible, the computation of these sets of ing (BH) PPPs are averaged to higher levels of aggregation using
weights takes into account regional and rural-urban variations in detailed expenditure weights. Accordingly, participating countries
consumption patterns. The work involves: (i) recoding and ag- are required to compile BH expenditure weights for 155 break-
gregatingdataonexpenditurebycommodityaccordingtothe110 downs of GDP.
BH classification of the ICP and (ii) deriving the share of each
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Methodology
The experts convened with a two-fold objective. The first goal SPD Software Improves Product List Development
was to agree on the key adjustments needed to ensure consistency The task of building a list of product specifications (PSs) has
with SNA 93 or, where this is deemed implausible, to propose been simplified and significantly improved in this ICP round us-
the kind of metadata that should be gathered and maintained by ing software tools developed by the Global Office (GO), thus
Regional Coordinator (RCs). The second target was to review overcoming the problems of comparability, representativity and
the status of national efforts in compiling BH weights, review the inadequate documentation from earlier rounds. The software is
national structure of consumption for consistency and plausibil- based on the conceptual framework in the Structured Product
ity, and create a working document that will help RCs validate BH Description (SPD) approach, adopted from the US Bureau of
shares submitted by their constituents. The meeting focused on Labor Statistics' checklist methodology.
householdfinalconsumptionexpenditure(HFCE). Sinceregions
are publishing consumption PPPs beginning this summer, it was The GO has invested in developing the SPD software tools nec-
thought that priority should be accorded to HFCE. essary to support the process of SPDs and PS generation, which
result in PSs that are comparable and representative. The soft-
The experts agreed that the first objective could be met using the ware supports this through an efficient iterative process between
IMF GDDS metadata. They noted that the use of a standard plat- national offices, regional implementing agencies and the GO. Be-
form, such as the GDDS metadata, would improve transparency sides facilitating the development of a global standard for prod-
of data compilation practices, identify weak areas in a country's uct specifications, the software also introduces an international
statistical infrastructure and allow participating countries to ben- product coding structure.
efit from best practices through a regional approach. The GDDS
metadata will shed light on important questions such as: Is HFCE The software is being augmented to support the generation of
estimated independently or is it obtained as a residual? Are the PSs in multiple languages simultaneously, and will include a glos-
weights for basic headings within HFCE based on household ex- sary of terminologies in various languages. The software has a
penditure surveys or on supply-use table? common database of products and characteristics. In the future,
different regions will be able to describe the same product differ-
The expert group recommended that the ICP Global Office ently, for example, the description "minced beef" may be pre-
(GO) should collaborate with IMF to identify areas of immediate ferred in one region while "ground beef" is used in another. The
relevance to the ICP and prepare a summary form to be filled by software will assign the same code for similar descriptions in its
national statistical offices. For some countries, this may involve glossary, thereby permitting regions to use their preferred descrip-
simply updating the relevant GDDS/SDDS records. For others, tion while still allowing the development of a standard global list.
it may mean filling the forms afresh. It was noted that this has Regardless of the language or region where the PS is defined, in
an added benefit of providing information to update the IMF's future rounds it will be possible to identify matching products
metadata records. both within and across regions, based on the associated codes
for products and characteristics. This will have a far-reaching im-
The meeting discussed the status of the ongoing expenditure plication in the global comparison. It will allow a robust regional
share compilation work and reviewed data from Latin America linking because it will be possible to identify countries that have
and the Asia-Pacific. The process involved both horizontal and maximum PS overlap with different countries across regions.
vertical checks. Vertical reviews were performed within countries
to check for consistency of the structure of consumption expen-
diture. It was also proposed that where possible, the horizontal
checking should include consistency of structure over time, com-
paring data from two or more years. Horizontal or cross-section
validation involved inter-country checks. A macro-based Excel
spreadsheet prepared by the GO was used for this purpose. The
tool follows a standard procedure to calculate the coefficient of
variation (CVs) of the country weights for each basic heading and
highlights those that appear to be anomalous. It combines the CV
test with an examination of standard deviations in which coun-
tries are color-coded if their weights fall outside one, two or more
standard deviations from the regional mean. This was done both
for nominal and real shares expressed in PPP terms. The tool
also allows validation at various levels of aggregation, including
at summary category, groups, classes, or basic heading levels. This
would allow regions to first focus on getting the higher categories
right and then moving further down to groups and subsequently
to classes and ultimately to basic headings.
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Regional Spotlight
Lessons Learned: The ICP Latin America Experience the only way for a country to realize that its numbers do not make
Louis Marc Ducharme sense is when its data is confronted by a neighboring country.
In November 2003, Statistics Canada (STC) This was the case not only for the price specialists, but also for
agreed to participate in and coordinate the In- the national accountants. There is nothing that can replace face-
ternational Comparison Program (ICP) for 10 to-face discussion.
Latin American countries over a period of three
years. STC is cooperating with the Economic Quality control: There is no substitute for very intensive quality
Commission for Latin America and the Carib- control. In Latin America, the quality control was done: a) before
bean (ECLAC) on the project; the Agency's role the price collection, when defining specifications and choosing
is funded by the Canadian International Development Agency. outlets b) during price collection in the field and c) at the process-
ing and aggregation stage at the STC. The regional coordinators
In early 2003, the ECLAC started harmonizing the basket of con- have established a schedule for visiting and assisting countries
sumer goods and services with the help of participating countries. over the four quarters, ensuring the widest geographical coverage
In 2004, the STC and ECLAC devoted time and resources to re- within limited resources.
view product specifications and assess the degree of preparation
of participating countries in the field. The region staged a "price- III. Learning by Building Capacity - An exercise of such magni-
collection dress rehearsal" in all 10 countries, in preparation for tude has tremendous spillover effects on the National Statistical
actual price collection in 2005. System of the participating countries. In our experience, a num-
ber of areas have benefited from the ICP. Let me mention two:
Thanks to the efforts of the 10 countries and two organizations
coordinating the process, the Latin American region has acquired Price Statistics: As mentioned, the success of this exercise could
experience and best practices in executing a complex program. not be achieved without a great deal of effort in harmonizing con-
The first data on consumer spending will be published in June cepts, definitions and practices in the collection of price data. The
2006, and the project will be completed in the first quarter of confrontation of the results collected by the 10 countries showed
2007 with the publication of the GDP. the importance of carefully defining the specification of the prod-
uct and service to be priced. The use of the SPD concept proved
Many lessons were learned in the course of the first two years of useful in harmonizing specifications among different countries,
this project. Amongst them, it is worth mentioning the following: and ensuring proper control of the item to be priced. Also, the
many field visits and instructions given to interviewers in the pro-
I. Learning by Doing - Dress rehearsal: Before conducting the cess of this exercise showed the importance of collecting charac-
first real price collection, a region should plan a dress rehearsal teristics of product, outlet and neighborhood to explain discrepan-
ahead of time. It is the only way to discover problems and put in cies in the results and improving the survey frame. Although most
place solutions to correct them. countries viewed this exercise independently from their current
operations, they have integrated some of the lessons learned.
Frequency of collection and revisions: It is important to have as
many as four rather than a single collection during the reference National Accounts: The data confrontation of national accounts
year, as this will allow time to improve operations in the field. It expenditures was a revealing experience for all participating coun-
is also essential to have many revisions to the weighting patterns. tries, as it showed the weakness of some key headings and high-
For this region, it is clear that without a third and fourth price lighted the need for rapid improvement of the national structure
collection, the data for some of our participating countries would of consumer expenditures. A few countries are already imple-
have been of poor quality. On the other side of the equation, it menting some changes to the next revision of their national ac-
took three attempts to harmonize the classification and weights counts. It is only in the context of this kind of exercise that
for the GDP and even after doing so, there are still some serious national accountants have a real opportunity to confront and im-
problems to resolve. prove their national statistics.
Establishment and monitoring of a realistic timetable: The estab- For all the reasons mentioned above, all participants of the re-
lishment of a strict timetable and its monitoring are imperative. gional meetings in Latin America agreed that results from the ICP
Given the size and complexity of the ICP and the very tight dead- are the fruit of substantial investments and that it would be a pity
lines for the world comparison, the regional and national coordi- if a new round is not launched immediately after the current one
nators cannot tolerate any deviation from the agreed timetable. to ensure that knowledge is not lost and that improvements to
national statistics are consolidated.
II. Learning by Confronting Results - Post-mortem meeting: One
important aspect of learning lies in the good practice of holding The author is Director General, Economics Statistics, STC and Chair of the
post-mortem meetings immediately after each price collection, ICP Data Access Working Group
where data of all participating countries are confronted. Often,
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Regional Spotlight
Strategic Framework for Capacity Building in Africa Regional ICP Board Met in Asia
Michel Mouyelo-Katoula, ICP-Africa Coordinator The Regional Advisory Board (RAB) for the ICP Asia-Pacific re-
The second Forum on African Statistical Devel- gion met in Manila, 8-9 December 2005. The Board is respon-
opment (FASDEV II) was held in Addis Ababa, sible for providing leadership and determining strategic priorities
Ethiopia, from 9-10 February 2006. It brought for Asia and Pacific region within the framework of the ICP. The
together some 150 participants - from African Boards also provide a forum for member countries to be involved
National Statistics Offices (NSOs) and their line in the regional project. This was the fifth meeting of the RAB.
ministries, PRSP units, sub-regional organiza-
tions, the UN Economic Commission for Africa Dr. Ifzal Ali, Chief Economist of the Asian Development Bank
(UNECA), the African Development Bank (AfDB), the World (ADB) and member of the global ICP Executive Board welcomed
Bank, the OECD Development Co-operation Directorate (DAC), the RAB members. Dr. Ali acknowledged the hard work and fi-
and Partnership for Statistical Development in the 21st Century nancial resources that participating countries have put into the
(PARIS21), and representatives of the donor community. The program. He stressed that for the first time, the ICP has placed
participants approved a Reference Regional Strategic Framework ownership of the program where it belongs, with National Statis-
(RRSF) for Statistical Capacity Building in Africa. The Frame- tical Offices and regional implementing agencies. He supported
work, which is in line with the Marrakech Action Plan on Statis- this welcome departure from past practices. He concluded with
tics, is the result of inter-institutional agreement reached at the the observation that irrespective of the PPP data that will result
first FASDEV meeting in May 2004 in Addis Ababa, Ethiopia. from this round, capacity building will be the most important
legacy of the ICP in the Asia-Pacific region.
The RRSF is meant to contribute to improved development out-
comes and good governance in Africa by guiding and accelerat- Ms. Carmelita N. Ericta, Executive Director of the Philippine
ing sustainable statistical capacity building activities. It reflects the NSO, was elected as the new Chair. Ms C. Ericta succeeds two
desire of African countries, with the support of the international former RAB Chairs before her, Mr. Frederick Ho, Commissioner
community, to meet the data challenges of the results agenda by of the Hong Kong Census and Statistics Department who retired
2015. It helps in particular to focus on strategic directions and ap- from the government service, and Mr. Paul Cheung, formerly
propriate implementation instruments for improving the planning, Chief Statistician of the Singapore Department of Statistics who
financing, management, and coordination of statistical develop- had relinquished the
ment activities to meet demands for statistics of good quality. Chair after assuming .. irrespective of the PPP data that will
new responsibilities "result from this round, capacity building
It is expected that national statistical institutions would refer as head of the UN will be the most important legacy of the
to the framework for guidance on statistical capacity building; Statistical Depart- ICP in the Asia-Pacific region. "
that international organizations would be able to use it as an in- ment in New York. Ifzal Ali,
put in their selection of programs and projects for supporting Chief Economist
statistical development; and that all stakeholders would be able The Board reviewed The Asian Development Bank
to use it to promote effective partnerships and coordination price data from the
of capacity building efforts on the part of both national and second-quarter data collection cycle and approved steps taken by
international institutions. the ADB and National Statistical Offices to improve data quality.
The RAB supported a meeting of NSO heads to discuss the level
NSOs would be responsible for implementing the Framework at of detail that can be published. The Board suggested drafting a
the country level. Sub-regional and regional organizations have communication plan and planning pre-launch activities prior to
play important roles as well, especially in the area of promoting thereleaseof finalPPPdata. Membersof theBoardstressedthat
statistical development, monitoring implementation and provid- the next round should be immediately organized at the end of the
ing forums for sharing knowledge, experience and expertise. If current round to capitalize on the present momentum and experi-
the Framework is to be effective, it will also be essential for inter- ence gained. The sixth RAB meeting is tentatively scheduled for
national development agencies and donors to be fully involved in June/July 2006.
its implementation.
AfDB and UNECA, jointly, will coordinate the implementation,
monitoring and reporting of the RRSF through appropriate co-
ordination mechanisms yet to be agreed between the two insti-
tutions. AfDB's contribution to the implementation and coordi-
nation of the RRSF will be provided in the framework of the
ICP-Africa.
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Regional Reports in Brief
Africa participated to the meeting. The objectives of the meeting were to
Forty-nine countries are conducting data collection, with varying review: i) progress in the region and at the global level; ii) the pre-
submission rates of edited and validated data. Since data col- liminary data of the third quarter of 2005 for consumer expen-
lection started in mid-2005 and will continue through mid-2006, ditures; iii) the revised data of the first two quarters of consumer
methods are being developed to calibrate the data to a 2005 annual expenditures; iv) the purchasing power parities of consumption
reference period. A quarterly price collection for pharmaceutical by major groups and basic headings; v) the consumer weights pro-
products began in November-December 2005, with data received vided by the national accounts of the participating countries; vi)
from 10 countries to date. Twenty-six government occupations the preliminary data for housing; vii) the preparatory work for the
have been selected as the most common jobs in participating other areas of the GDP (construction, machinery and equipment
countries. Looking ahead, surveys for equipment goods will be and government expenditures).
conducted in April-May 2006, using the Ring List core products as
theprimarybasis with somesupplements toreflectthelocalsitua- Compared to the two first quarter, the third one showed substan-
tion. Collection for construction components will also take place tial progress in the quality of data the number of products priced
in April-May 2006. Meanwhile, the partnership formed with the and the number of observation made. Expenditure weights were
ONS(UnitedKingdom)andtheINSEE(France)isprovingtobe received from all participating countries and reviewed. For the
productive from the technical viewpoint. first time all individual parities and the global consumption ex-
penditures per capita were examined and analyzed together by the
Asia price specialist and national accountants of the ten participating
Three series of Data Review Workshops in 2005 have been con- countries. Thiswasamajorstepforwardasitisthefirsttimethat
ducted. Actions have been taken to improve the comparability in this region, both price specialist and national accounts could
of price data on the basis of decisions taken in the Data Review examine the result of price collection and weight estimations. In
Workshops in July 2005 and the Regional Coordinators meet- December 2005, the South American Region completed its fourth
ing in Washington in September 2005. These actions include: collection of the consumer prices. The preparation for capital
revisions on the product list; preparation of the "What to Price" formation and government is underway.
guide; and preparation of the "List of Products Needing Special
Attention". Asia plans to conduct two more Data Review Work- Western Asia
shops to improve the reliability/quality of data for the household During the last regional workshop for Western Asia, the second-
consumption items. GDP expenditure weights have been received quarter data of all 11 countries was revised using the Dikhanov
from 22 countries, with varying reference periods from 2002 to analysis and Quaranta tables. Through this revision, the regional
2004. The expenditure data from the countries were consolidated office and the member countries were able to detect problem-
into regional and sub-regional tables and then evaluated on the atic products and to identify means to tackle them. The product
average expenditures and statistical diagnostics. list for consumption of this region was quite large (1,285 prod-
ucts), which allowed for deletion of repetitive or non-represen-
Three workshops on Poverty PPP, construction survey and equip- tative products. In some areas, such as clothing and furniture,
ment goods surveys were organized the last week of November some participants expressed the view that some countries may be
and the first week of December, 2005. Data collection for con- pricing lower-quality products while others may have gone for a
struction and equipment is scheduled for the second quarter of higher-end product. It was agreed that a number of national ICP
2006. Surveys for education, dwellings and government occupa- price experts should travel to their neighboring countries to check
tions will be carried out during Q1 2006. for quality differences, and to develop adjustment factors where
quality differences were noted.
CIS
The region has collected data for both the 2004 and 2005 refer- Two parallel workshops are organized for March 2006. A work-
ence periods and will be providing PPP estimates and associated shop is organized for data validation to finalize data from the sec-
indices for both years. The prices for the 2004 reference year ond quarter and validate third quarter data. National Accounts
have been validated. The 2004 results will be published at the experts are invited to take part in this workshop, to discuss basic
end of 2006. The region will receive average prices for the 2005 heading expenditure weights and also to take part in the data vali-
reference year at the end of Q1 2006. GDP expenditure weights dation process. A second workshop will bring together national
for 2005, which will used for comparison results calculation, will construction experts to go over construction SPDs and to agree
be available next year. on implementation plans and the schedule.
Latin America
The third regional meeting of the South America Region was held
inMontevideo(Uruguay)14-19November2005.Pricespecialists
and national accountants from all ten South American countries
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Country Spotlight
Lebanon: ICP & Tool Pack Offer New Vision of Data Collection ICP Experience Will Refine Egypt's CPI
Maral Tutelian Mohamed F. Abulata
The ICP offers Lebanon a great opportunity to The present round of the ICP is the third in the
extend price collection of goods and services all ICP series in which Egypt has participated. This
over the country, since the CPI price collection round, however, is characterized by its wider geo-
is currently limited to the capital Beirut and its graphic coverage so that average prices may be
nearby suburbs. At the same time, the ICP also more expressive of the real price levels prevailing
provides a platform for Lebanon to refine its
method of defining specifications of goods and in Egypt. One other major feature of the present
services in the CPI. round is the larger number of items to be priced,
with due consideration given to the detailed Structured Product
The Central Administration for Statistics (CAS) is responsible for Description (SPD) associated with each product.
the CPI and National Accounts in Lebanon. In the early days of
the ICP in the ESCWA region, Lebanon fully participated and co- Egypt is taking part in the present ICP round in two different
operated with the regional committee in specifying and translating regions, Africa and West Asia. Egypt is also among the 19 coun-
SPDs into Arabic. Once the specifications were ready, the CAS tries selected from the six regions to take part in the Ring Com-
selected a special team to implement the ICP under the super- parison that is designed to link regional purchasing power parity
vision of the CPI Team. The team is composed of one senior estimates.
statistician, three Assistant Statisticians, and 25 surveyors and con-
trollers. To ensure national coverage, Lebanon was divided into The geographic coverage of the ICP price survey has been ex-
five sub-regions, with each one including at least one major city to tended to cover eleven governorates representing nearly 70 per-
ensure the capture of large-scale markets. Each sub-region was cent of Egypt's population. Both urban and rural areas have been
divided into sectors to ensure consumption diversity in goods and included whenever the selected governorate consists of both
services. Outlets with big turnover in each consumption category types of areas (nine governorates). The remaining two governor-
were selected from these sectors. A pre-survey was conducted in ates are basically metropolitan cities (Cairo and Alexandria).
early January to confirm SPDs and compare, for each product, the
different specifications encountered in the different sectors. The The sample outlets of different types was as large as 5,171 outlets
survey was postponed selected from all governorates included in the survey. The size of
to the second quarter .. ICP and the Tool Pack offered Lebanon outlet sample varies from one governorate to another, depending
due to the country's "a new vision of organizing data collection on the population size and the availability of all types of outlets.
political situation. and management of regional databases
and an experience in harmonizing product The sample size ranges from 348 outlets in Behera governorate
A total of 1,261 items specifications and price collection practices to 618 in Cairo governorate. The frequency of price collection
were priced in more across regions." has been monthly for food and beverages, and quarterly for the
than 2,000 outlets. remaining items with the exception of certain services, for ex-
Prices of fruit and vegetables were collected daily, while prices ample, education services, which were priced biannually due to
for the rest of food and non-alcoholic beverages were collected therelativestabilityintheirpricesovertime.Thefieldworkstarted
weekly. Other prices were collected monthly. This decision was in January 2005.
taken to ensure a great amount of data to analyze seasonality and
comparability between the regions. The ICP Tool Pack software A total of 108 price collectors, 26 team leaders and four senior
was a great help for managing data strorage. It also offered op- supervisors have implemented the field operations. Prices were
tions for data comparison at several levels and provided a speedy collected simultaneously for both the African and West Asian
tool to manage the field and to correct outliers. lists using two different questionnaires. The local employees of
the Central Agency of Public Mobilization and Statistics (CAP-
The experience that the CAS has gained from the implementation MAS) have been recruited on a part-time basis to carry out the
of the ICP will provide a solid background of comparable data fieldwork. Besides their own staff, few local statistical offices have
between the Lebanese regions for the first time in the history of temporarily appointed additional price collectors.
Lebanon. This will allow regional comparison and analysis of
the socio-economic situation and the level of poverty. The ICP Office editing and data processing are carried out centrally at the
is now an important input to the smooth implementation of CPI CAPMAS premises in Cairo. Tool Pack and SEMPER software
extension, to stretch data collection to the whole country. were used for data entry and validation for West Asian and Afri-
can ICP prices, respectively.
In conclusion, the ICP and Tool Pack has offered to the adminis-
tration a new vision of organizing data collection and managing
regional databases, as well as an experience in harmonizing speci- Undoubtedly, the present ICP has given substantial attention to
fications and price collection practices across regions. capacity building in the national statistical offices in the fields of
price statistics and national accounts. Egypt will build upon the
The author is Director General, CAS, Lebanon .
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Country Spotlight
experiences gained from the ICP to improve its CPI in several as- indices that will result from the proposed BOCC survey. The
pects (1) Adopting detailed SPD to ensure that the reported price new comparable indices will give a clearer picture of the construc-
is exactly related to the same product, which is priced in various tion sector's role in the country's economic performance. It will
data collection centers; (2) Increasing numbers and types of outlets clarify how strong the relationship is between construction and
so that a better price average may be attained; (3) Devising soft- economic growth.
ware with a built-in validation capacity (Tool Pack or SEMPER)
in compiling CPI; and (4) Upon satisfactory testing of the CPI The next step to integrate the BOCC approach with national sta-
module of the Tool Pack and assurance from the World Bank for tistical work is to assess the existing questionnaires and to review
its continued support, Egypt would like to use the ICP Tool Pack the method of calculating the indices. A pilot study should be
forCPIcompilation.ThiswouldbeagreatbenefittoboththeCPI planned to apply the new questionnaire in selected areas and pos-
and ICP as it allows greater integration of the two programs. sibly through the national statistical capacity building initiatives.
Clearly, an ICP "construction community" composed of the
The author is Senior Advisor, Central Agency for Public Mobilization and workshop participants, guest lecturers, and ADB and World Bank
Statistics, Arab Republic of Egypt experts responsible for the ICP Construction Survey, will be a
strong force for sharing experiences of the pilot studies.
Matching Indonesia's Construction Materials Survey with the ICP The author is Division Chief of Mining, Energy and Construction Statistics,
Dedi Walujadi BPS-Statistics Indonesia
In Indonesia, data users eagerly await the Whole-
sale Price Indices of Construction Materials by
Type of Construction. Both the central gov- Nepal Finds ICP Useful for Building National PPI
ernment and local authorities use these indices, Sher Bahadur Budha
each fiscal year, in their budget proposals for in- Gross Fixed Capital Formation (GFCF) is a very
0 frastructure projects for development planning important component among the five main ICP
purposes. The indices are also important in the expenditure aggregates. Within this expenditure
construction business, where economic actors estimate project aggregate, Machinery and Equipment and Con-
costs and forecast business revenue and profit. struction are the main two pillars which support
the national economy for sustainable develop-
The Asian Development Bank and the World Bank initiated a ment. In the Nepalese context, GFCF holds an
very interesting occasion, the ICP Construction Workshop, in approximately 20 % share of total GNP. The change in this share
December 2005. This Manila workshop aimed to compare and is considered an indicator of the country's development pace.
build a standardized survey questionnaire by using the Basket of
Construction Components (BOCC) approach, introduced by Nepal is among 23 Asia-Pacific countries taking part in the 2005
the Global Office, so that the indices will be comparable across ICP Round. As part of the regional preparation for Machinery
countries. Questions raised on the floor included: What types of and Equipment Goods surveys, a regional workshop was orga-
commodities should be accommodated? What should be the ap- nized by the Asian Development Bank in Manila from December
propriate size for each commodity, for it to be well accepted by 1-3, 2005. National ICP Coordinators from the NSOs and their
the respondent as an economic actor in the construction sector? national equipment experts attended the workshop. An interna-
What are the weightings for different types of constructions ac- tional equipment expert and ICP coordinators from the ADB and
tivities, and how big is the sample size of the survey? The fruitful the ICP Global Office (GO) led the workshop.
discussion, which followed an explanation of the materials and
results of some preliminary studies in Africa and Central Asia, led After a short presentation of the hybrid Structured Product De-
to the conclusion that the ICP Construction Survey may be ap- scription (SPD) and Product Specification (PS) approach pro-
plied in every country, with some modification. The participants posed for the 2005 ICP Round, the discussions focused on de-
also agreed on the importance of socializing the new method veloping regional SPDs and narrowing down the specifications
through national statistical capacity building. on each item to reflect market realities in the region. The starting
point was the global SPD list prepared by the GO. The following
To make the Wholesale Price Indices of Construction Materials questions were considered over three days: i) Product Selection:
by Type of Construction using the BOCC comparable in the near Is that particular item representative for most of the participat-
future, some work needs to be done. Like other countries in Asia, ing countries? ii) Product Addition: Is there any need to expand
Indonesia needs appropriate infrastructure to support its econo- the list to cover areas of prominent commercial activities in the
my. It is believed that the proposed Government infrastructure region? iii) Product Specification: What specification should be
budget will give an accurate measure, if it can be backed-up by a fixed for each selected item (including manufacturer and model)
theoretically sound deflator. The performance of the construc- so that most countries can price it?
tion sector can also be measured appropriately by using the new
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Continued from previous pages
The national equipment experts represented various specialties, ... continued from page 3 (ICP: Enhancing quality of statistics.)
including IC Engine, construction equipment, medical equip-
ment, fabrication, etc. The experts contributed a great deal to cations) through time. In the-mid 1990s, the Boskin commission
the discussion, and this was very vital to the creation of a draft attributed about 60 percent of its estimate of upward bias in the
regional list. To finalize the list, an electronic expert group was US CPI to insufficient accounting for quality change. For coun-
created to help address some pending issues. Nepal was involved tries not yet using SPD-type methods, their implementation can
in the electronic discussion. The e-group members undertook only lead to significant first-stage improvements in the countries'
pre-surveys and research in their own markets and finalized the ability to control for product quality change in national price in-
list through electronic discussion. The Asian list is used by the dices. They do this by providing better information (1) on ex-
GO as the core equipment list for the Ring Comparison. actly how product specifications are changing and (2) in a form
prospectively allowing compilers to use statistical modeling to
The ICP has proven useful in developing new methodologies for develop product quality adjustment methods..
comparing costs of capital goods, including construction proj-
ects and equipment goods. Besides their intended aim of com- The ICP has become an important force for building regional
paring costs of capital goods across countries, the methodologies statistical capacity in Africa. First, it has helped to identify areas
will help countries develop and/or improve national indexes. In where national statistical compilers need technical assistance to
Nepal's case, the ICP has contributed to capacity building in the address data quality issues in both price statistics and national
Central Bureau of Statistics (CBS) in these ways. accounts. Second it has spurred technical assistance providers to
sharpen their focus on the most pressing needs for methodologi-
A. As a part of the comprehensive system of price statistics, the cal and organizational capacity improvement in national statisti-
ICP is contributing to further improvement in the country's cal offices.
existing system of price measurement.
B. It is hoped that after the current ICP Round is completed, Continuous improvement in the ICP. Ultimately, the quality of
Nepal's CBS will be able to estimate GDP from the expendi- the ICP rests on the quality of its national source information.
ture approach as well. As a result of a general improvement in national statistical ca-
C. Since the CBS has initiated price data collection to compile pacity, future rounds of the ICP can expect quality benefits to
and publish PPI in the near future, the ICP has made it pos- flow from the improvement in the quality of the price and na-
sible to identify and understand Structured Product Descrip- tional accounts source data in the participating countries. This
tions and Product Specifications, which will be very useful for improvement should be helped significantly by the quality issues
the precise measurement of PPI. previous rounds of the ICP have identified as national authori-
D. It is expected that the CBS will be able to strengthen and ties and multilateral and bilateral technical assistance providers
build sustainable capabilities to develop various kinds of price build the statistical capacity to remedy them.
statistics through the development of expertise during ICP
activities and regional workshops.
Looking ahead, an increased number of workshop days with
field trips will provide even greater opportunity to examine and
discuss SPDs and PSs more closely. The Electronic Working
Group also needs to be supported by a separate website to fol-
low up on the progress of the work and to share country experi-
ences with all ICP participating countries.
The author is Equipment Expert, Central Bureau of Statistics, Nepal
www.worldbank.org/data/ICP
Volume 3, No. 1 February 2006
ICP News A quarterly e-newsletter for the International Comparison Program
Calendar of Events
Event Western Asia 6th Regional workshop Event Meeting of the Balkan Group on the first
Date March 13-17, 2006 survey 2006 "Food, drinks, tobacco"
Venue ESCWA, Beirut, Lebanon Date April 20-21, 2006
Participants National Coordinators, Price Experts and Venue Ljubljana
Construction Experts from 11 countries Participants 3 countries of the Balkan Group, Statistics
Slovenia, Eurostat, OECD
Host ESCWA Host Slovenia as group leader
Contact Farah Nsouly
person nsouly@un.org Contact Marjanca Gasic
person Marjanca Gasic@gov.si
Event Overlap Meeting for the first survey 2006 Event Planning Meeting for the second survey 2006
"Food, drinks, tobacco" "Personal Appearance"
Date March 15-17, 2006 Date April 27-28, 2006
Venue OECD, Paris, France Venue Eurostat, Luxembourg
Participants 4 group-leader countries, Eurostat, OECD Participants 4 group-leader countries, Eurostat, OECD
Host OECD Host Eurostat
Contact Francette Koechlin Contact Jarko Pasanen
person Francette.KOECHLIN@oecd.org person Jarko.Pasanen@cec.eu.int
Event LAC Regional meeting
Date April 3-12, 2006
Venue Santiago, Chile
Participants National Accountants and Price Specialists of the
10 participating countries of South America
Host ECLAC, Santiago
Contact Louis Marc Ducharme
person louismarc.ducharme@statcan.ca
How to reach us:
For more details about the International Comparison Program please visit our website
http://www.worldbank.org/DATA/ICP
or contact
Fred Vogel, Global Manager fvogel@worldbank.org
Yonas Biru, Deputy Global Manager, ybiru@worldbank.org
Comments and contributions to ICP News should be sent to Amy Lee, slee9@worldbank.org.
International Comparison Program (ICP)
The World Bank
1818 H Street NW, MC2-209
Washington D.C. 20433
USA
Email: ICP@worldbank.org
www.worldbank.org/data/ICP