2. Outline
• Introduction and analytical content (Sam)
o History, data sources & compilation, description
o Global trends in public expenditures
o Findings from third-party analysis
• Access and interactive tool (Nilam)
o Archiving and how to download data
o Data visualization tool
• Reaching clients (Yifei)
o Webpage visits and data download statistics
o Promotion
3. Why SPEED?
• Public expenditure (PE) is a powerful instrument for
achieving sustainable growth, poverty reduction,
and transformation
o Need to understand linkages between different types of
PE and development for better allocation of resources
o Calls for reliable data on use of public resources
• SPEED database developed to provide
policymakers, researchers, etc. with the most
comprehensive PE information
o Officially launched in 2010
o Two updates: 2013 and 2015
4. Brief history and funding
• Database work
o Shenggen Fan and Anuja Saurkar
o Bing Yu and Sangeetha Malaiyandi; and Eduardo Magalhaes
• Support
o CKM (Luz Marina Alvare and Nilam Prasai)
o Several colleagues (Xinshen Diao, Tewodaj Mogues, Frances
Cossar, Dario Debowicz, Inigo Verduzco Gallo, Athur Mabiso,
Edina Metili Mwangi, Alejandro Nin-Pratt, Gissele Gajate
Garrido, Yifei Liu, Alvina Erman)
• Funding
o Initial stages funded by USAID via its support to the Regional
Strategic Analysis and Knowledge Support System (ReSAKSS)
programs in Africa and Asia
o More recently, by the CGIAR Research Program on Policies,
Institutions, and Markets (PIM)
5. SPEED data sources
• Primary source: IMF Government Financial
Statistics (GFS) Yearbooks
• Supplemental sources
o IMF Statistical Appendices and Selected Issues
o World Bank Public Expenditure Reviews
o International and multilateral organizations (e.g. West
Africa Central Bank, Eurostat)
o National sources—e.g. budget documents and statistical
abstracts
Downloaded from websites (ministry of finance, statistics
bureau, accountant general’s office, central bank)
• Based on UN Classification of the Functions of
Government (general government expenditure)
6. Data compilation (3 main issues)
1. Missing data points within series—interpolation
o Single point: simple average of non-missing points
o Multiple consecutive points: linear trend between non-
missing points
2. Missing data points at the beginning or end of the
series—extrapolation
o Five-year average growth rate of non-missing points to
extend series backward or forward
3. Competing data points from multiple sources and
consistency of filled data points—compare growth
rates and shares
o Missing data from any one source are sometimes replaced by
corresponding non-missing data from other sources—
complicated with non-missing data from multiples sources
o Checking that filled data for subcomponents of total
expenditure add up to the actual or non-missing total value
7. Overall consistency checks
1. Visual inspection using graphs to identify
unrealistic deviations from trend or outliers
2. Sum of parts to check that the sum of sector
expenditures is smaller than the total
expenditure
3. Exclusion of negative expenditure values as a
result of certain accounting procedures (e.g. sale
of a nonfinancial asset particularly for countries
in Europe)
4. Detailed source matching by documenting the
source or method used to obtain each data point
8. What’s in the latest (2015) database?
• 147 countries from 1980 to 2012
o Organized into 8 regions: East Asia and Pacific (EAP), Europe
and Central Asia (ECA), Latin America and the Caribbean
(LAC), Middle East and North America (MENA), South Asia
(SA), Africa south of the Sahara (SSA), Other Europe, and High
Income (HI)
• 9 sectors
o Agriculture, defense, education, fuel and energy, health,
mining, social protection, and transport & communication
(separately and combined)
• Several indicators
o Expenditure in constant 2005 purchasing power parity (PPP)
dollars (PPP$) and constant 2005 US$
o Percentage of sector expenditure in total expenditure
o For agriculture, percentage of expenditure in agGDP
o Percentage of total expenditure to total GDP
10. Summary of data—global trends I
0
20
40
60
80
100
EAP ECA HI
Europe
Other HI LAC MENA SA SSA
Agriculture Education
Health Defense
Transport & communication Social protection
Other
Annual average agricultural expenditure
(% total expenditure, 1980-2012)
• Social protection attracted the largest share in the developed regions (HI
Europe and Other Europe) and in ECA and LAC
• Education was top (10-17%) in other developing regions, except in MENA where
defense was top spender (17%).
• Infrastructure and agriculture attracted the least shares (3-7%)
11. Summary of data—global trends II
1980-1989 1990-1999 2000-2012
Rank Share (%) Rank Share (%) Rank Share (%)
EAP 4 8.9 5 6.2 6 4.6
ECA - - 5 4.1 6 3.9
HI Europe 6 4.1 6 3.1 6 1.8
Other HI 5 3.3 6 2.6 6 2.0
LAC 6 4.1 6 3.7 6 2.2
MENA 6 4.1 5 3.2 6 2.5
SA 3 9.0 4 6.6 5 5.1
SSA 4 7.4 6 3.5 6 3.0
Rank is from 1 to 6, with 1 being the first or top rank.
Rank and share of agricultural expenditure share in total
expenditure, 1980-2012
Rank and share of agriculture expenditure got worse over time
12. Summary of data—global trends III
-10
0
10
20
Growth rate (%)
0
50
100
150
200
per capita (2005 PPP$)
0
5
10
% of total expenditure
0
10
20
30
% of agricultural value added
Annual average agricultural expenditure, 1980-2012
14. Third-party analysis (32 publications)—I
Type # Institutions
Journal
articles
11 American Political Science ReviewA*, Agricultural
EconomicsA, World DevelopmentA, Food and Nutrition
BulletinB, Journal of Technology Management and
Technopreneurship, Corporate Ownership and Control,
Journal of Economics and Sustainable Development,
Philosophical Transactions of the Royal Society, The BE
Journal of Macroeconomics, Journal of International Finance
Studies, Nepalese Journal of Agricultural Economics
Reports 11 World Bank, European Union, African Development Bank,
Climate Policy Initiative, Inter-American Development Bank,
FAO, UNDESA, OECD
DPs/BPs 7
Dissert.
/ Thesis
3 Vilnius, ISM - University of Management and Economics,
Lund University, Addis Ababa University
15. Third-party analysis—II
• Country/regional coverage
o Country specific (2—Malawi and Ethiopia)
o Regional, cross-country comparisons: Latin America (4),
Africa (8), South Asia (2), low and middle income
countries (13)
• Thematic focus
o Agriculture (17), social protection (3), cross-sectoral (5)
• Analytical issues
o Relationship between government spending and growth
and productivity (3)
o Effect on or controlling for government spending (3)
o Composition of spending (1)
16. Third-party analysis—findings
• Relationship between spending and growth and productivity
o Thapa et al (2015): increase agricultural expenditure or investment accelerate
agricultural and total GDP growth and, consequently, achieving MDG1—
Bangladesh, Nepal, Cambodia, and Lao PDR
o Shittu et al. (2014): strong/positive influence of public agricultural expenditure
(PAE) on productivity in SSA; increase in annual PAE by US$3,000/worker (25x
present level) needed to raise agTFP by 1%
o Musaba et al (2013): significant positive effects of agricultural and defense
expenditure on economic growth. Expenditures on other sectors were negatively
related to economic growth
• Effects on government spending
o Doyle (2015): effect of remittances on government spending on social protection
is negative (used three different measurements of government social spending)
o Sánchez, et al (2015): spending on social protection (unlike social sector)
increased steadily in low-income countries irrespective of the recent global
financial crisis
o Ceriani et al (2011): social protection spending is negatively related to income
distribution (gini)
• Composition of spending:
o Lowder et al (2012): on-farm investments in agricultural capital are more than 4x
the size of government spending in agriculture in low and middle income
countries
17. New directions of SPEED to consider
• Disaggregation by economic use
o Compensation of employees, use of goods and services,
consumption of fixed capital, interest, subsidies, grants,
social benefit, other
• Disaggregation of agriculture expenditure by sub-
function
o Research (reconcile with ASTI), extension, irrigation,
marketing infrastructure, farm support, etc.
• Other sectors, identify key sub-functions for
poverty and nutrition (other?), e.g.:
o Expenditure in rural areas for all sectors
o Primary education and health care
o School feeding in education sector
18. Functions (COFOG)
701. Gen. Pub. Services
702. Defense
703. Pub. Order & Safety
704. Economic Affairs
705. Environ Protection
706. Housing &
Community Amenities
707. Health
708. Rec., Culture, & Rel.
709. Education
710. Social Protection
Economic classification
21. Compensation of
employees
22. Use of goods and
services
23. Consumption of
fixed capital
24. Interest
25. Subsidies
26. Grants
27. Social benefits
28. Other
Agricultural
sub-sector
Crops
Livestock
Forestry
Fishery
Agricultural
sub-functions
Extension
Research
Irrigation
Land management
Market infrastructure
Farm support
Policy, planning, M&E
Regulation, licensing
Statistics
…
Level
Central
State/Region
Local/District
…
Policy objective
Food security
Productivity/Growth
…
Target
Demography
Geography
…
Planning/Execution
Budget
Expenditure
Source of financing
Domestic (loan, tax)
External (loan, grant)
Different levels of disaggregation
Editor's Notes
Also implied in the CAADP NAIPs as the different budgets are allocated differently according to the different classifications