In 2013 the African Development Bank

Transcription

1 BANQUE AFRICAINE DE DEVELOPPEMENT AFRICAN DEVELOPMENT FUND FONDS AFRICAIN DE DEVELOPPEMENT Statistics Department December Table of Contents Chief Economist Complex Preface 1 Acknowledgments Profiles of peer reviewers 2 2 Abbreviations Executive summary Background 6 2. Resources and infrastructure 8 3. Statistical classifications and best practices recommended by SNA Compilation of national accounts statistics Revisions to national accounts compilation Price indexes and deflators Dissemination practice Improvements in progress Communication strategy Recommendations Main conclusions 18 The case of Kenya Preface In 2013 the African Development Bank () established a program of peer reviews for economic statistics, initially focusing on national accounts. This supports the long-standing objective of the Bank to enhance statistical capabilities in the region. To initiate the peer reviews, the Statistical Capacity Building Division within the Statistics Department of the invites its regional member countries to discuss the sources and methods of their national accounts with qualified experts from up to three other African countries. The discussions take place in the country under review, with the aim of identifying good practices by the country being reviewed, as well as areas where improvements should be made. Suggestions for improvements are made, taking due account of the practical challenges faced by the country under review. All those participating in the review process both reviewers and reviewees benefit from the free exchange of information and ideas. As the reviews are published by the, the national and international community has access to these detailed and transparent descriptions of the sources and methods underlying the compilation of national account statistics in the African region. To date, peer reviews have been carried out in seven countries Ghana, Zambia, The Gambia, Botswana, Burundi, Egypt and Kenya. Reports on these reviews are published on the website The is grateful to all those who have helped to bring this peer review of Kenya to a successful conclusion. This includes the Kenya National Bureau of Statistics and its national accounts experts, who agreed to be reviewed, as well as the national statistical offices of Ghana, The Gambia, and Zambia and their national accounts experts who volunteered to conduct the review. Statistics Department African Development Bank Group Compiled by Bernice Serwah Ofosu-Baadu Head of National Accounts and Economic Indicators Section of Ghana Statistical. Litia Simbangala Head of the National Accounts Branch of the Central Statistical Office of Zambia Aminata Deen Senior Statistician and Acting Head of the National Accounts Unit of The Gambia This report was prepared by the national accounts experts from the national statistical offices of Kenya with the support of Statistical Capacity Building Division of the Statistics Department, in the Office of the Chief Economist at the African Development Bank. The findings reflect the opinions of the authors and not necessarily those of the African Development Bank or its Board of Directors. Every effort has been made to present reliable information obtained through the peer review process. A f r i c a n D e v e l o p m e n t B a n k

2 African Development Bank Acknowledgments We would like to commend the Statistical Capacity Building Division within the Statistics Department of the African Development Bank () for introducing peer reviews to assess the reliability of national accounts statistics in the Africa region. This peer review of the national accounts of Kenya has provided us with the opportunity to learn from each other s experiences with regard to the compilation, revision, and reliability of national accounts through the constructive exchange of views and ideas. We also wish to thank Mr. Zachary Mwangi the Acting Director General of the Kenya National Bureau of Statistics (KNBS), Mr. Collins Mathew Omodi the Acting Director of the Macroeconomics Statistics Directorate, and Mr. Benjamin Mucharia Muchiri the Acting Senior Manager of National Accounts for their most valuable co-operation, support and professional assistance. We extend our profound gratitude to all others who contributed in various and diverse ways to the success of the mission, in particular the support staff of the KNBS and the staff of East Africa Regional Resource Center in Nairobi. We especially wish to thank Mr. Oliver Chinganya, Manager of the Statistical Capacity Building Division of the, for his contribution to and oversight of the process of peer review. His support of statistical capacity building in Africa in highly appreciated. Profile of peer reviewers Bernice Serwah Ofosu-Baadu Head of the National Accounts and Economic Indicators Section of the Ghana Statistical Service since Ms. Ofosu-Baadu has eight years experience in compiling GDP data through the production and expenditure approaches on a quarterly and annual basis. Litia Simbangala - Head of the National Accounts Branch of the Central Statistical Office of Zambia. Has eight years experience in compiling national accounts statistics (annual GDP; planning, development and coordination of business surveys; satellite accounts; GDP forecasting; and analysis of economic performance indicators). Aminata Eyatunde Deen Senior Statistician and Acting Head of the National Accounts Unit of The Gambia Bureau of Statistics since Ms. Deen has twelve years experience in compiling annual GDP data through the production and expenditure approaches. From , she worked with the Tourism Unit of the then Central Statistics Department compiling and analysing tourism statistics. 2

3 African Development Bank Abbreviations BPM 6 CIP COFOG COICOP CPA CPI FISIM GDP GFCF GNI GVO GVA HIES ICP IGAD ILRI ISIC KPHC KIHBS KNBS MFSM MRDS NAS NPISH NSDS NSS PFCE PPI QNA SNA SUT UNSC VAT African Development Bank Balance of Payment Manual -6th edition Census of Industrial Production Classification of Functions of Government Classification of Individual Consumption by Purpose Classification of Products by Activity Consumer Price Index Financial Intermediation Services Indirectly Measured Gross Domestic Product Gross Fixed Capital Formation Gross National Income Gross Value of Output Gross Value Added Household Income and Expenditure Survey International Comparison Program Intergovernmental Authority on Development International Livestock Research Institute International Standard Industrial Classification Kenya Population and Housing Census Kenya Integrated Household Budget Survey Kenya National Bureau of Statistics Monetary and Financial Statistics Manual Requirement Data Set National Accounts Section Non-profit Institution Serving Households National Strategy for the Development of Statistics National Statistical System Private Final Consumption Expenditure Producer Price Index Quarterly National Accounts System of National Accounts Supply and Use Tables United Nations Statistical Commission Value Added Tax 3 A f r i c a n D e v e l o p m e n t B a n k

4 African Development Bank Executive summary The African Development Bank (), through its statistical capacity-building program, has embarked on an initiative that supports peer reviews of the reliability of the national accounts produced by national statistical offices in the Africa region. The peer review process provides an opportunity for a small team of experts from countries within the region to review the robustness of national accounts compilation in another African country that has agreed to be reviewed. The reviewers engage their peers in frank discussions on technical issues with respect to the methods and data sources they employ to compile their national accounts. The outcome of this process is to serve three main purposes. These are: (i) To improve methodology and international comparability through adherence to the internationally agreed 1993 System of National Accounts (SNA 93) and the various international standard classifications of the United Nations with the eventual aim of adherence to the SNA 2008 for those countries that are at a commensurate level of statistical development; (ii) To provide credence to the compilation process and instil confidence among the users of national accounts statistics; and (iii) To up the advocacy on the credibility and robustness of the national accounts estimates within the Africa region at the international level. National accounts statistics in Kenya are compiled by the National Accounts and Fiscal Statistics Section in the Macroeconomic Statistics Directorate of the Kenya National Bureau of Statistics (KNBS). The Section comprises seven professional staff. It compiles annual GDP by the three approaches: the production approach, the expenditure approach and the income approach. The production and expenditure approaches are produced at both current and constant prices. For quarterly GDP, the estimates are at constant prices only and are derived solely by the production approach. The constant price estimates are currently based on the prices of Prior to the recent rebasing, the guidelines and recommendations of the SNA 93 were generally followed in the compilation of national accounts in Kenya. Revision 3.1 of the International Standard Industrial Classification of All Economic Activities (ISIC Rev. 3.1) was used to classify the structure of the economy into three main sectors: agriculture, industry and services. However, to improve the compilation of national accounts, the KNBS has adopted the SNA 2008 and its accompanying statistical infrastructure. The Classification of Products by Activity 2008 is used to classify products by the activity that produces them. The Classification of Individual Consumption by Purpose is used to classify household expenditure. It is also used in the compilation of the consumer price index (CPI).) The Classification of the Functions of Government is used to classify Government expenditures by function. The KNBS has just completed revising its national accounts estimates from the base year 2001 to the new base year A supply and use table (SUT) and a set of integrated economic accounts up to the generation of income account have been compiled and are ready to be published by the end September Findings The KNBS has made great efforts to improve the compilation of its national accounts statistics. It has rebased the accounts on 2009 and at the same time it has adopted SNA 2008 and ISIC Rev. 4. The previous national accounts series had 2001 as a base year and were compiled in line with the SNA 93 and ISIC Rev The need to update the base year from 2001 to 2009 became necessary in order to capture the new economic developments and to better reflect the performance of the economy as well as to meet the needs of users. 1. The rebasing has allowed a wider coverage of activities and the inclusion of new methods such as the estimation of Financial Intermediation Services Indirectly Measured. 2. The GDP is compiled by economic activity at the four-digit industry level and disseminated at the two-digit industry level according to ISIC Rev In terms of coverage, the estimates of gross value added (GVA) of informal sector activities is based mainly on the Kenya Integrated Household Survey conducted in 2005/06 as no standalone survey of the informal sector has been conducted. 4. The KNBS uses NADABAS (Excel and Access) based software to adjust the 4

5 African Development Bank data obtained from different sources (censuses, surveys, and administrative) from their specific format to the required national accounts format. All subsequent adjustments to the data are made and recorded by the software thus capturing a complete compilation. 5. A SUT has been developed for the first time since 1976 with 2009 as the reference year. It covers 81 Industries and 150 products. 6. Information from the balance of payments is used to generate other national accounts statistics such as the gross national income (GNI). Estimates of GNI and gross national disposable income are estimated and disseminated. 7. Annual GDP estimates are compiled using all the three approaches. However, quarterly GDP estimates are compiled only at constant prices using only the production approach. 8. Integrated economic accounts are compiled up to the generation of income account, but institutional sector accounts are only being compiled for general government and rest of the world. However, the KNBS plans to compile the whole set of institutional sector accounts in the revised national accounts statistics. 9. Besides the CPI, other price deflators are used including the producer price index for manufacturing, the building cost index for construction, and the unit value indexes for exports and imports. 10. The informal sector is not explicitly shown in Kenya s GDP although it is implicitly included in the estimates. Collection of firewood and water, which are defined in the SNA as economic activities are monetized and included. 11. Production data obtained for subsectors from the Census of Industrial Production need to be complemented with a dedicated informal sector survey to improve the estimates. 12. The inclusion of VAT data and data from the 2010 Integrated Survey of Services has considerably improved the compilation and estimation of GDP for the service sector. 13. To obtain the trade and transport margins for each economic activity in the compilation of the SUT, the Survey of Trade Margins was conducted in Exports and imports are being deflated by the recommended unit value indexes. 15. Appropriate components of the CPI are used as deflators for specific sectors/industries. 16. VAT data are widely used to complement other sources of data for national accounts compilation. 17. The KNBS maintains a business register which serves as the frame for all its enterprise/establishment surveys. 18. The KNBS adequately disseminates its statistical releases and publications: quarterly and annual reports through the website, hard copy reports, media releases and CDs. 19. Any major increase of the estimates arising from the revision is the result of the inclusion of new data sources (censuses, surveys, and administrative). 20. Given the size of Kenya, and the demands for data set out in the Requirement Data Set (MRDS) by the United Nations Statistical Commission, the staff strength is inadequate, particularly as the MRDS has been broadened to include (1) quarterly accounts for the nominal and real measure of GDP by industry or by expenditure components and (2) the quarterly compilation of the integrated accounts up to net lending for the total economy and the rest of the world. See Annexes 3 and 4 for more details. 21. Back casting of the previous years of the old base year to yield consistent time series is in progress. 22. Purification and distribution of water is excluded from the estimation of GVA for collection, purification and distribution of water 5 A f r i c a n D e v e l o p m e n t B a n k

6 African Development Bank 1. Background The history of organized statistical activities in Kenya goes back to the 1920s when the Colonial Government appointed its first Official Statistician in The enactment of the Statistics Act on 4th July 1961 fully integrated the Kenya Statistical Unit within the government machinery as the government statistical office. The Statistical Unit was formally established as the Economics and Statistics Division of the Treasury. In 1962, the Division undertook the second Population Census in Kenya prior to attaining Independence. In 1963, the Ministry of Planning and Development was established and the Division was transferred from the Treasury to the new ministry. Subsequently, the Division was split into two units that were elevated to departments namely the Planning Department headed by a Chief Economist and the Statistics Department headed by a Chief Statistician. However, the Agricultural Statistics Section was physically located in the Ministry of Agriculture, an arrangement which continued up to During this year, the Statistics Department of the Ministry of Planning and Development was renamed the Central Bureau of Statistics (CBS) and the head of the Department was designated Director General instead of Chief Statistician. The CBS expanded its operations in the 1960s and 1970s both at the headquarters in Nairobi and at the field level. During this period, it undertook a wide range of data collection activities and kept a healthy publication programme. This impetus continued up to the mid 1980s when the CBS greatly expanded its field survey programme to respond to the need for district-level data following the adoption of the District Focus for Rural Development strategy in However, the period that began in the mid-1980s and continued into the 1990s was characterized by low level data collection efforts, minimal processing and analysis of collected data and discontinuation of issuance of publications which, in the past, used to be regular features of CBS activities. The factors responsible for the downward trend included inadequate budgetary allocation as a result of reduction in government expenditure and an inadequate number of professional staff, particularly at senior levels, in the 1990s to prepare the CBS for the challenge of the 21st century. The Kenya National Bureau of Statistics (KNBS) was established by the Statistics Act of 2006 to replace the CBS. The Act establishes the KNBS as a semiautonomous government agency under the Ministry of Devolution and Planning, with a vision statement to be a centre of excellence in statistics production and management, and a mission to effectively manage and coordinate the entire National Statistical System (NSS) to enhance statistical production and utilization. The Statistics Act 2006 specifically mandates the KNBS to: 1. Act as the principal agency of the government for collecting, analysing and disseminating statistical data in Kenya; 2. Act as the custodian of official statistics; 3. Conduct the Population and Housing Census every ten years, and other censuses and surveys as the Board of Directors may determine; 4. Maintain a comprehensive and reliable national socio-economic database; 5. Establish standards and promote the use of best practices and methods in the production and dissemination of statistical information across the NSS; and 6. Plan, authorise, coordinate and supervise all official statistical programmes undertaken within the NSS. Its core mandate is the collection, compilation, analysis, publication and dissemination of statistical information for public use, with an additional role of coordinating, monitoring and supervising the NSS. The KNBS aims at providing quality statistical and sex disaggregated data which is key to achieving excellence in performance levels based on the fundamental principles of official statistics. Under the new framework, the KNBS is headed by the Director General, who is assisted by six directors. These are the directors of Strategy and Development; Population and Social Statistics; Macroeconomic Statistics; Production Statistics; Information, Communication and Technology; and Finance and Administration. The Board of Directors of the KNBS provides overall supervision and policy guidance. The aspirations of the autonomous KNBS are to promote awareness of the important role and functions of the institution, to improve the timeliness and integrity of the statistics being produced, as well as playing an advocacy role to increase resource mobilization since it has the statutory mandate to produce and provide government, the private sector, 6

7 African Development Bank parastatal organisations, international organisations, civil society and the general public with official statistics. National accounts statistics in Kenya are compiled by the National Accounts Section, which is a unit within the Macroeconomic Statistics Directorate of the KNBS. The Section consists of seven professionals and has the capacity both in terms of human and material resources to undertake the compilation of national income statistics on a timely basis. The KNBS has a fixed release calendar for the annual GDP as well as the quarterly GDP. The preliminary annual GDP of the latest year is released in April the following year and it takes 16 months for the final estimates to be released (see Annex 2). The KNBS has just rebased the national accounts from 2001 to As part of the rebasing exercise, the KNBS adopted the SNA 2008 and this is reflected particularly in the implementation of the new method of estimating Financial Intermediation Services Indirectly Measured (FISIM); the work-in-progress technique for crops, livestock, and construction; the development of a supply and use table (SUT) as an integral part of national accounts compilations; the construction of Input-Output Tables and a Social Accounting Matrix; and the adoption of ISIC rev 4. This is an improvement over the SNA 93 recommendations that were previously followed. In addition, the quality of the revised national accounts is much better than the former estimates due to the use of new and improved data sources.. The Classification of Products by Activity 2008 was used to classify the products and activities for the SUT. The Classification of Individual Consumption by Purpose was used to classify household consumption and in the compilation of the consumer price index (CPI). The Classification of the Functions of Government is used to classify the functions of Government. This compilation is done by the KNBS using the data from the Office of the Controller of Budget, Ministry of Finance and County Governments reports while the Government Finance Statistics Manual 2001 (GFS 2001) is used to classify government receipts and expenditure. The KNBS has a comprehensive business register that is used to conduct annual economic surveys and which needs to be regularly updated. The questionnaires are posted to the respective respondents and follow-ups are then conducted to improve the response rate. Response rate of at least 60 percent are attained before analysis is done. For large companies employing 50 and above, a response rate of 70 percent is preferred. Data from other sources such as annual reports from nonmarket government corporations, the National Social Security Fund, the National Health Insurance Fund, the utility companies and the telecommunications networks operators are used to complement the surveys. Changing to the SNA 2008 has much improved the scope, detail and quality of the national accounts and their supporting economic statistics. The adoption of the sixth edition of the Balance of Payments Manual (BPM 6) and the 2010 Monetary and Financial Statistics Manual (MFSM) are still in progress but the recommendations in these manuals were followed to the extent the data would allow during the revision exercise. This report describes the methods, concepts, and definitions employed by the KNBS to compile GDP and other national accounts aggregates for Kenya both before and after the revision and the rebasing. It identifies the challenges with respect to the compilation methods, data sources, and adherence to best practices. The report also gives recommendations on how to improve the estimation of GDP and the revision of the national accounts of the country. Specifically, the tasks undertaken by the Peer Review were: Review the current national accounts compilation methodology; Review the compilation methods, the data sources, and harmonized international classifications used in the revision estimates; Review the dissemination practices; Identify possible areas of improvements, if any; and Provide recommendations on ways to improve the process. 7 A f r i c a n D e v e l o p m e n t B a n k

8 African Development Bank 2. Ressources and infrastructure The KNBS head office is housed in three different locations within the vicinity of Nairobi. The Macroeconomic Directorate, of which the National Accounts Section is a part, is housed on the eleventh floor of Bima house in Nairobi. The Section is made up of seven staff, all university graduates and professionals by status; four are well trained and experienced in national accounts compilation while three have just two years experience. Though the new staffs have some training in national accounting they still need further training in national accounts statistics both in-house and externally to produce the statistics demanded by the adoption of the SNA One such demand is the production of annual SUTs. This requires a lot of data collection on all the sectors, integrated economic accounts as well as institutional sector accounts. The office space in Bima house is very congested in order to accommodate the staff and the equipment necessary for the efficient production of national accounts statistics on a timely basis. There is a functioning IT network with internet facilities that links with the line ministries and the Central Bank for the timely exchange of information, including data used in the compilation of national accounts. This reduces the need for the physical circulation of staff between offices, speeds up the transmission of information, and contributes to staff effectiveness. Kenya has a history of strategic planning for statistical development. The process of designing the third plan, called the National Strategy for the Development of Statistics (NSDS), has begun. The first five-year Statistical Plan for the CBS, a government department in the Ministry of Planning and Development, covered the periods This plan was instrumental in the transformation of the CBS into the semi-autonomous government agency the KNBS. The second plan covered the period It was based on a comprehensive review of the previous strategic plan and the lessons learnt. While the first two plans focused on the Bureau, the third plan will cover sectors: government ministries, departments and agencies the MDAs - with the KNBS and counties (local levels of administration) as special sectors. International best practice requires that the NSDS design process should follow a sectoral (or bottom-up) approach whereby a Sector Statistics Plan (SSP) is designed for each sector and the SSPs are used as building blocks for the overall NSDS. This approach leads to enhanced coordination of the NSS. It is fortuitous that the design of the NSDS has coincided with the design of the strategic plans for MDAs and counties. This coincidence creates an opportunity for inter-linkage between the SSPs and the strategic plans for MDAs. The design of the NSDS also takes place at a time when a new Constitution has come into force. The Constitution places a lot of emphasis on accountability which calls for evidence-based policy, planning, decision-making, monitoring and evaluation at the centre and in the counties, and hence the need for comprehensive and quality data. There is currently a document outlining the NSDS which will serve as a roadmap for statistical activities in the country over the next five years. The KNBS has a Stakeholder Unit which handles issues pertaining to the functioning of the NSS. However, this cannot substitute for a comprehensive NSDS document, setting out the statistical priorities of the NSS for a specific period of time. Indeed, the NSDS could usefully incorporate an action plan with monitoring benchmarks to ensure a steady progression toward achievement of the national statistical strategy. The directorates collaborate with each other by providing the data required for national accounts compilation. These data are used by the KNBS to estimate the GDP. For instance, the Agricultural Statistics Section, a unit under the Production Statistics Directorate, provides data on crops, livestock and other products such as horticulture, forestry and logging, and cut flowers. The Foreign Trade Section and the Financial Statistics Section, units under the Macroeconomic Directorate, collect data on foreign trade and financial statistics, which are used as inputs into the compilation of imports and exports of goods and services. The KNBS conducted the Kenya Population and Housing Census (KPHC) in 2009 throughout the country. This 8

9 African Development Bank provided some key information to update the indicators currently used in the compilation of national accounts revision. In addition to the KPHC, other comprehensive surveys were also conducted (as shown in Annex 1) and these have provided vital information for use in the revision. 3. Statistical classifications and best practices recommended by the SNA Statistical classifications are a key element in the compilation of statistical indicators in general and the compilation of national accounts statistics in particular. Since classifications 1 are an exhaustive and structured set of mutually exclusive and well-described categories, often presented in a hierarchy that is reflected by the numeric or alphabetical codes assigned to them, they are used to standardise concepts and compile data. To this end, the statistical infrastructures compliant to the SNA 2008 are: Standard Industrial Classification of All Economic Activities Revision 4 (ISIC Rev.4), Classification of Products by Activity 2008 (CPA 2008), Standard International Trade Classification Revision 4 (SITC Rev.4), The Classification by Broad Economic Categories Revision 3 (BEC Rev.3), Classification of Individual Consumption According to Purpose (COICOP), Classification of the Functions of Government (COFOG), Classification of the Purposes of Non- Profit Institutions Serving Households (COPNI), Classification of the Outlays of Producers According to Purpose (COPP). Best practices according to the SNA 93 include the following: Output should be valued at basic prices, while all transactions on the uses of goods and services (final consumption, intermediate consumption, and capital formation) should be valued at purchasers prices. Thus, in principle, GVA at basic prices is output valued at basic prices less intermediate consumption valued at purchasers prices. When the data available are not sufficiently reliable and robust to permit the use of double deflation, estimates of value added at constant prices should be obtained by deflating current value added by an output price index or extrapolating base year value added by an output volume index. These are acceptable secondbest solutions. The base year should be changed every five years if chain-linked volume indices (which constitute the preferred option) are not compiled. (The SNA 2008 recommends the use of previous year s prices and chain linking for volume measures.) Non-profit institutions serving households (NPISHs) should be included. Adjustments for residents expenditures abroad and non-residents expenditures in the domestic economy may need to be included in private final consumption expenditure. A complete set of integrated economic accounts should be compiled for the economy. SUTs should be compiled on an annual basis. FISIM should be allocated to users. Exports and imports of goods should be deflated with unit value indexes. 1 Source: Standard Statistical Classifications: Basic Principles, E Hoffmann and M.Chamie. Paper presented to the 30th UN Statistical Commission, A f r i c a n D e v e l o p m e n t B a n k

10 African Development Bank 4. Compilation of national accounts statistics The KNBS compiles and disseminates quarterly and annual GDP estimates. While the annual GDP estimates are compiled and published using the production, expenditure and income approaches, the quarterly GDP estimates are compiled using the production approach and published only at constant prices. Current and constant price estimates are compiled for both the production and the expenditure based annual GDP. Thus, meeting the minimum required data set for the implementation of the SNA 2008 for the annual national accounts. Currently, the KNBS compiles the national accounts within the scope of the SNA 93. Revision 3.1 of ISIC is used to classify the structure of the economy into three main sectors: primary, secondary, and tertiary. In using ISIC Rev. 3.1, the compilation is done at the group (four-digit) level, although the final figures are mostly published at the section (two-digit) level. The CPA 2008 is used to classify products, COICOP is used to classify household final consumption expenditure and COFOG is used to classify government final consumption expenditure. An Excel based IT tool, SNAPC, is used for the compilation. The base year for the constant price series is currently 2001 (to be replaced by 2009 in September 2014) Approaches and methods to compute annual national accounts before the revision The KNBS compiles annual national accounts using the production, expenditure, and income approaches. i. Production Approach The economy is grouped into three sectors: primary, secondary, and tertiary. Estimates for these sectors are compiled routinely and disseminated at both constant and current prices. The growth rates as well as the contribution of the sectors to GDP are also compiled and disseminated. GVA by industry is estimated as follows: Primary sector: The primary sector, which relates to agriculture, is disaggregated into crops, livestock, forestry, hunting, and fishing. Crops: Data for the compilation of GVA for crops is mostly collected from the Ministry of Agriculture and the Livestock and Horticulture Development Authority. The data covers crops such as maize, beans, wheat, potatoes, coffee, tea, cut flowers, fruits, vegetables, rice, etc. The data collected are used to derive value added as the difference between total output and intermediate consumption for the each of the crops covered. The value added estimates also include horticulture activities. Constant price GVA is obtained by deflating current price value added using appropriate CPI components. In some cases indirect methods are used. Livestock estimates are derived through the commodity flow method. The information obtained to estimate the GVA for livestock includes the total stock of animals in the reference period by type, the opening and closing stock and the numbers born, sold and slaughtered including imports and exports. Constant price GVA is obtained by deflating current price value added using appropriate CPI components. Forestry and logging GVA is estimated using information from government accounts, the 1998 Labour Force Survey (LFS) and the 1999 KPHC. Fishing GVA is estimated based on the data on quantities and values of fish landed provided by the Department of Fisheries. Includes inland, marine and aquaculture fishing activities. To derive the constant price estimates, volumes are used as an extrapolator on 10

11 African Development Bank the previous year s GVA. Output of own-production in agriculture is estimated through benchmarks derived from the 1997 Welfare Monitoring Survey Secondary sector: Mining and quarrying GVA is estimated using the information obtained from data on quantities and values of minerals obtained from the Department of Mines and Geology. Volume indices are used to derive constant price estimates. Manufacturing GVA is estimated using information from the Monthly Survey of Industrial Production. In addition, the 1998 LFS provides benchmark information on micro and small scale enterprises. All subdivisions of the manufacturing subsector are covered. Electricity GVA is estimated separately for production and for distribution using the annual audited financial statements of the Kenya Power Authority. The output of the distribution component is measured by the total value of the trade margins 2 realized on the goods they purchase for resale. GVA at constant prices is calculated by extrapolating the quantities produced by different categories of users. Water GVA is estimated using the data collected from the 1999 KPHC. In addition, the time spent in minutes in collecting water by the households is based on the 1998 LFS. The time spent is then monetized using the rate of pay for agricultural unskilled labour under 18 years. Purification of water in Kenya mainly takes place at the county levels of government, and as such it is captured in GFS. Construction GVA, which includes buildings and roads, is estimated using the input approach for the modern sector. Data on stocks of traditional dwellings are obtained from the 1989 KPHC and the 1999 KHPC as well as estimates of own account (informal) construction. Government accounts are used to derive estimates for general government. The building cost index is used to inflate GVA at constant prices to obtain GVA at current prices Tertiary sector: Wholesale and Retail Trade output is measured by the total value of the trade margins realized on the goods they purchase for resale on all commodities. The trade margins are also based on the values of similar economies and expert opinion. Intermediate consumption is derived from the input-output ratio and subtracted from the trade margins to obtain GVA at current prices. Constant price GVA is derived by deflating the current price GVA by the overall CPI. Hotels and restaurants. The GVA at constant prices for hotels are obtained through extrapolation using the hotel bed occupancy rate as extrapolator. To obtain current price GVA, the CPI component is used to inflate constant price GVA. Informal activities are estimated using information from the 1998 LFS. For restaurants, information from the 1998 LFS is used. Transport, storage and communication GVA estimates are derived using information obtained from the Annual Survey of Transport. Survey information is complemented with the financial statements from the providers of railway, water and air transport. Auxiliary activities such as clearing and forwarding, storage and warehousing, and tour operators are also included. Estimates of informal road passenger transport are derived using total number of registered communal taxis (matatu), adjusted for non-compliance, daily average revenues and expenditure (repair and maintenance). Annual financial statements of the telecommunication providers as well as that of the post and couriers are used to produce the GVA for communication component of the subsector. Constant price GVA is obtained by deflating current price GVA by relevant CPI components. Financial and insurance are mainly derive from Central Bank's statistical 2 A trade margin is defined as the difference between the actual or imputed price realized on a good purchased for resale and the price that would have to be paid by the distributor to replace the good at the time it is sold or otherwise disposed of. 11 A f r i c a n D e v e l o p m e n t B a n k

12 African Development Bank bulletin which contains the requisite data on commercial banks and auxiliary banks. For insurance, the value of claims and premiums are obtained from data collected from the Insurance Regulatory Authority. Real estate GVA GVA estimates are based on the number of households in different types of houses according to the 1989 and 1999 KPHCs. Actual rents and imputed values are obtained from the 2005/06 Kenya Integrated Household Budget Survey (KIHBS). GVA at constant prices are calculated by extrapolation using the population growth rate as an extrapolator on the benchmark values. Intermediate consumption is estimated at 10 percent of output for modern houses and 5 percent for other houses. No intermediate consumption is calculated for traditional houses Business services activities GVA is derived using the 1998 LFS benchmark estimates. The annual LFS provides the volume indicator used for extrapolation. Public administration and defence GVA is estimated directly from government accounts of both the central government and local authorities. The number of government employees is used as an indicator for extrapolating GVA at constant prices. Social security activities are estimated using the financial statement. Education and health GVA for these activities is obtained from the government accounts for public sector institutions. Private sector components are generated from the 1998 LFS and from enrolment data. Community, social and personal activities GVA is estimated from the 1998 LFS and the annual LFS for the number of employees in the sector and their earnings. Input-out ratios are used as derived the intermediate consumption. Taxes on products are estimated through the commodity flow method using an adjusted VAT rate to cater for non-compliance. The taxes are computed only for the goods and services to which they apply. Comments In terms of coverage, the estimates of GVA by economic activities do not adequately capture the contribution of the informal sector, since there has not been a stand-alone survey for the informal sector. The only available information on the informal sector was captured through a module in the 2005/06 KIHBS and does not wholly reflect the situation on the ground. According to the SNA 93, the valuation of output should be at basic prices with intermediate consumption being valued at purchasers prices. GVA at current prices is then derived as output less intermediate consumption. The KNBS estimates the GVA at current prices for all economic activities in line with the SNA 93. The SNA 93 recommends double deflation as the way of estimating constant price GVA. However, the KNBS does not use - double deflation. A mix of methods is used instead. Some economic activities are deflated, others are extrapolated, while for some the inputoutput share is used. There is an overreliance on the CPI as a deflator in the current national accounts. However, this concern has largely been addressed in the revised estimates as additional deflators have been employed. These include the use of the producer price index (PPI), the unit value indexes for exports and imports and implicit deflators derived from the SUT. ii. Expenditure Approach All the components of the expenditure account are covered, compiled and disseminated at current and constant prices on an annual basis. Private Final Consumption Expenditure (PFCE) is estimated annually using the commodity flow method and the COI- COP classification. Data sources for some expenditure estimates, such as non-resident expenditure in the domestic market and resident expenditure abroad, have still to be fully explored and captured. Consumption expenditure of NPISHs is estimated as part of PFCE and not separated. Government final consumption expenditure is estimated using data from central and local government and nonmarket government corporations. Consumption of government fixed capital is estimated using the perpetual inventory method. Gross capital formation consists of two main components: Gross Fixed Capital Formation (GFCF) and changes in inventories. (No estimate is made for acquisition less disposal of valuables.) GFCF is estimated using the commodity 12

13 African Development Bank flow method. The commodity flow approach utilizes the total supply of capital goods: that is, domestically produced capital goods plus imported capital goods. GFCF is estimated for machinery and equipment, buildings and structures and civil engineering projects such as roads. Changes in inventories are estimated indirectly as a percentage of GDP of the previous year. Exports and imports of goods and services are estimated from data obtained from the External Trade Section and the Balance of Payment Section of the KNBS. iii. Income Approach The income approach is estimated for the total economy as the sum of: compensation of employees, gross operating surplus, mixed income, taxes on production and imports less subsidies. The data sources used for the estimation includes the 1998 LFS, the financial statements of industries, the government accounts, and the 2005/06 KIHBS. iv. Other annual accounts produced 1) Annual production account by industry 2) Gross National Income 3) Gross National Disposable Income and Saving up to net lending Items 2 and 3 are estimated using balance of payments data in addition to GDP data Methods to compute Quartely GDP (QGDP) before the revisions The KNBS only compiles QGDP at constant prices and only from the production side. The output of agriculture (crops, forestry, livestock, hunting and fishing) is complicated by the fact that the process of production may extend over many months or even years. Thus, according to the SNA 93, it should be treated as workin-progress. This is further complicated by the fact that the output is not sufficiently processed in a form that is ready to be marketed: output 3 according to the SNA 93 should be in a form of finished goods or services ready to be supplied or provided. The SNA 93 recommends that the value of agricultural output should be recorded as being produced continuously and distributed over the quarters in proportion to the total cost incurred in each period. This is because production cycles stretch over several quarters for some produce. When quarterly and annual series are derived from the same data sources, the annual figures are calculated by summing the estimates for the four relevant quarters. But, when quarterly and annual series are derived from different data sources, then the quarterly estimates need to be aligned to the annual estimate in a way that preserves the quarterly distribution over time and provides quarterly estimates that sum to the annual estimates. The latter is what applies to the KNBS. This requires quarterly compilation techniques that meet international best practice. Such a benchmarking technique is the IMF s XLPBM. The XLPBM is an Excel function implementing the Denton proportional benchmarking method and its enhancement for extrapolation. Annual estimates are benchmarked with the quarterly indicators. The quarterly compilation should derive seasonally adjusted quarterly estimates as well as unadjusted estimates. Quarterly national accounts are all about seasonal patterns. Hence, there should be a process of estimating and removing the seasonal effects from time series to reveal non-seasonal features. For this, X-12 ARIMA (Auto-Regressive Integrated Moving Average) or other software using similar techniques is recommended. Seasonally adjusted quarterly national accounts estimates are an international. As a pilot country of the IMF s Enhanced Data Dissemination Initiative 4 project on quarterly national accounts, the KNBS follows these procedures in compiling and publishing QGDP at constant prices using 3 Reference SNA 93, paragraphs and A f r i c a n D e v e l o p m e n t B a n k

14 African Development Bank the production approach. The data sources for the QGDP compilation include the VAT turnover data (communication, wholesale and retail trade, business service activities and community, social and personal service activities), Monthly Survey of Industrial Production, Monthly Survey of Hotels, and the Kenya Power and Lighting Company. Where data are not readily available on a quarterly basis, trend extrapolation is used. The QGDP series are available from 2001q1 to 2014q1. 5. Revisions to national accounts compilation National accounts reflect the performance of an economy. Rebasing enables the national accounts to capture the real picture of the economy by taking account of factors, such as movements in relative price, and structural changes in production and consumption patterns, which, if ignored over time, may contribute to an under or over estimation of GDP. According to the SNA 93, the base year for national accounts should be changed every five years if chain-linked volume indexes (which constitute the preferred option) are not compiled. However, the SNA 2008 recommends the use of previous year s prices and chain linking for volume measures. It is on this basis that the KNBS, wishing to improve the reliability of its national accounts statistics, took the decision in 2010 to rebase and revise them and to change the base year from 2001 to Main Sectors Revised The rebasing started with the development and compilation of a SUT for The SUT comprises 81 industries and 150 products. This was compiled using a new IT tool, NADABAS (Excel and Access based). To compile the 2009 SUT, comprehensive data collection took place as a SUT is data hungry, it requires a lot of data. The SUT is the main framework upon which the revisions of the national accounts compilation were linked. VAT turnover data, although used in the quarterly compilation, were included in the annual compilation for the first time. The VAT data were used to complement the data from Integrated Services Survey (ISS) so as to improve the services sector estimates in particular. For list of the sources of data used for the SUT compilation, see Annex 1. The main subsectors that were extensively revised include crops, livestock, manufacturing, and almost all the subsectors under the services sector for which recent data were available. 1. Crops and livestock subsectors including poultry: Information from the 2011 Cost of Agricultural Production Survey and from livestock reports by IGAD and ILRI 5 were used to improve subsector estimates. This led to an increase in the levels for growing crops and for horticulture from 399,474 (KSh Million) to 454,709 (KSh Million), and the farming of animals from 127,704 (KSh Million) to 132,968 (KSh Million) as shown in Table Manufacturing: the 2010 Census of Industrial Production (CIP) was complemented with VAT turnover data. 3. Construction: The ICP 2011 Construction Survey on input costs was incorporated into the estimation of the construction sector. 4. Services sector: VAT turnover data complemented with data from the 2010 ISS and the 2010 Survey of Trade Margins contributed to the improvements in the sector s estimates. 5. Education: The National Education Economic Accounts compiled by Ministry of Education and UNESCO in collaboration with the KNBS in 2012 helped to improve the estimation of the sector. 4 Established in 2010 by the IMF. 5 GAD and ILRI reports were not prepared by the KNBS. 14