Quality%20of%20public%20finances:%20some%20illustrations

1
Quality of public finances some illustrations

• António Afonso
• (European Central Bank ISEG/UTL-Technical
  University of Lisbon UECE-Research Unit on
  Complexity and Economics )
• Política fiscal y coordinación de políticas
• San Sebastian, 24 July 2009
• These slides reflect the views of the author and
  do not necessarily reflect those of the ECB or
  the Eurosystem.

2

• Introduction and motivation
• Measuring performance and efficiency
• Methodology
• Illustrative examples
• Overall public sector
• Education
• Health
• Social spending
• Conclusions

A. Afonso
3
Public expenditure ratios have steadily
increased in the euro area since the 1960s before
peaking and, in some cases, declining in more
recent years. Public expenditure is nevertheless
much higher than in most other industrialised
countries. According to many observers, it
exceeds the levels required for the efficient
provision of essential public services. (ECB,
Monthly Bulletin, April 2006, p. 73).
Introduction and motivation
The need to improve competitiveness, concerns
about fiscal sustainability and growing demands
by taxpayers to get more value for public money
as well as the need to reconsider the scope for
state intervention in the economy has prompted
efforts to increase the focus of budgets on more
growth-enhancing activities and gear the tax mix
and the allocation of resources within the public
sector towards better efficiency and
effectiveness. (EC, 2007, p. 9)
The question we ask today is not whether our
government is too big or too small, but whether
it works (). Where the answer is yes, we intend
to move forward. Where the answer is no, programs
will end. And those of us who manage the public's
dollars will be held to account to spend
wisely, reform bad habits, and do our business in
the light of day because only then can we
restore the vital trust between a people and
their government. (Barack Obama inaugural
speech, 20 January 2009)
A. Afonso
4
Public finances efficiency and economic growth
Introduction and motivation
A. Afonso
5

• Public finance developments, notably the growth
  in the size of the government, have increasingly
  been in the focus of policy debates.
• The existing fiscal framework in the EU has
  increased the awareness of the relevance of
  fiscal sound behaviour.
• The EC, the Lisbon Reform Agenda and the
  Stability and Growth Pact argue for assessing
  fiscal policy developments also by taking into
  account the quality of public finances,
  especially the efficiency and effectiveness of
  public spending.
• At the EU level the Working Group for the Quality
  of Public Finances was created in the Economic
  Policy Committee (in 2004).

Introduction and motivation
A. Afonso
6
Total General Government spending ( of GDP)
  1980 1995 2000 2007 Change 80 to 07
Austria 50.2 56.0 51.3 48.0 -2.1
Belgium 54.7 51.9 49.0 48.8 -5.9
Germany 46.6 48.3 45.1 43.9 -2.7
Denmark 52.7 59.3 53.5 50.6 -2.1
Spain 30.8 44.4 39.1 38.8 7.9
Finland 40.1 61.6 48.3 47.4 7.3
France 45.7 54.4 51.6 52.6 6.9
Greece 29.0 46.6 46.6 43.1 14.1
Ireland 45.6 41.1 31.6 36.4 -9.2
Italy 40.8 52.5 46.2 48.5 7.8
Luxembourg 42.3 39.7 37.6 37.5 -4.8
Netherlands 55.8 51.6 44.2 45.9 -9.9
Portugal 33.5 43.4 43.1 45.7 12.2
Sweeden 59.6 65.2 55.6 52.5 -7.1
United Kingdom 47.2 43.9 36.8 43.6 -3.6
Euro Area 44.5 50.6 46.2 46.3 1.8
EU 15 44.9 50.3 44.9 46.1 1.2
Japan 33.5 36.9 40.6 38.2 4.8
United States 33.8 35.4 32.5 35.6 1.8

• General government sector increased in the euro
  area and in the EU15 from 1980 to 1995
• In 2007, higher than US and Japan

Introduction and motivation
Source EC Ameco database and EC spring 2008
economic forecasts. DE, ES, GR, IE, SE, EA, EU15
values for 1980 are from the Ameco autumn 2006
database (old definition). For 1995, values
reflect the euro area 13 whereas for 1980 values
reflect the euro area 12 and West Germany
respectively.
A. Afonso
7
Measuring public sector performance and efficiency

• Main questions
• Are public services satisfactory considering
  the amount of resources allocated to its
  activity?
• Could one have better results using the same
  resources?
• Could we have the same results with lower
  expenses?
• Can we measure cross-country efficiency and
  determine benchmark countries?
• Can we explain measured inefficiency?
• systemic component,
• environmental or non-discretionary component.

Performance and efficiency
A. Afonso
8
Measuring performance and efficiency

• Public sector performance can be measured via
  output/outcome indicators
• Health, education, infrastructure, income
  distribution
• gt need for good indicators

Performance and efficiency

• Public sector efficiency relates outcomes to
  the resources used/inputs
• gt need for homogenous and matching data
  (heterogeneity is a limit)

Key issues methods and (homogeneous and right)
data to assess performance and efficiency.
A. Afonso
9

• The common production function relates inputs
  (xi) to output (y)
• y F (x1,x2)
• Alternatively F (x1, x2) is a production
  possibilities frontier
• Note that
• typically there are several outputs, (y1, y2,
  ...)F (x1,x2, ...)
• their joint production depends on several inputs
• and on other environment variables.

Methodology (1)

• Non-parametric methods commonly used in the
  literature
• FDH, DEA, both
• Non-discretionary inputs should be considered
• There are some examples of two-step
  (tobit/bootstrap) analysis.
• Parametric methods stochastic frontier
  analysis.

A. Afonso
10
Examples of possible methods
Cost efficiency Technical efficiency
Productivity
Methodology (2)
Total Factor Productivity
Partial Indicators
Frontier Analysis
Malmquist Indices
Parametric
Non-parametric
Deterministic (COLS)
Stochastic (SFA)
DEA
FDH
Two-step analysis
Tobit
Extensions for Panel Data
Bootstrap
Fixed Effects
GLS
Random Effects
A. Afonso
11

• One should be able to
• i) estimate output efficiency scores for EU/OECD
  countries, taking into account the resources
  employed
• ii) explain efficiency scores, controlling for
  environment factors (non-discretionary inputs).
• Most used methodologies
• raw efficiency scores DEA (data envelopment
  analysis)
• stochastic frontier
• explaining inefficiency
• tobit regression,
• bootstrap technique

Methodology (3)
A. Afonso
12
DEA

• y - column vector of outputs, x - column vector
  of inputs, X - input matrix,
• Y - output matrix.
• q - efficiency score (qlt1).

Methodology DEA
q lt 1, inefficiency q 1, efficiency
Note ? is the measure of efficiency, given by
the ratio between the weighted average of the
outputs (y) produced and the weighted average of
the inputs (x) used. See Coelli et al. (1998) for
more details.
A. Afonso
13
DEA and FDH illustration
Ds output inefficiency
Methodology DEA
A, C efficient B, D less efficient.
Ds input inefficiency
A. Afonso
14
Non-discretionary inputs and two-step procedure
(1)
Ds environment corrected output
score d1c/(d1cd2c)
Methodology exogenous factors
Ds output score d1/(d1d2)
1 gt d1c/(d1c d2c) gt d1/(d1d2), the environment
corrected score is closer to the frontier.
A. Afonso
15
Non-discretionary inputs and tobit two-step
procedure (2)

• Non-discretionary inputs
• Socio-economic differences play a role in
  determining heterogeneity and influence outcomes
  (for either schools, hospitals, local governments
  or countries achievements in an international
  comparison).

Methodology exogenous factors
Two-step approach Efficiency scores (d) are
regressed on non-discretionary factor (z)
The efficiency scores are not higher than 1 (or
always lower than one according to the setup),
which allows using a tobit regression approach.
A. Afonso
16
Malmquist Productivity Index MPI (constant
returns to scale)
output

• The DMU produces less than feasible under each
  periods production frontier.
• The MPI indicates the potential rise in
  productivity as the frontier shifts from period t
  to t1.
• The DMU at time t could produce output yp for
  input xt
• With the same input xt it could produce output
  yq at period t1.

Methodology MPI
Efficiency change index
Technology change index
input
A. Afonso
17
Coelli et al. (2005).
Stochastic frontier analysis

• i country, t time period
• yit output, GDP per worker
• Xit vector of inputs, private and public
  capital per worker and human capital
• ß set of production function parameters to be
  estimated
• eit normally distributed random error
• hit non-negative efficiency effect, assumed to
  have a truncated normal distribution
• zit non-discretionary factors (the governance
  indicators) that explain inefficiency
• set of efficiency parameters to be estimated
• A translog functional form for F() seems a
  sensible option.

It is possible to produce a likelihood ratio
statistic to test g0 If g0, there are no random
inefficiency effects.
18
SFA production possibility frontier
Stochastic frontier
A. Afonso
19

• Van den Eeckhaut, Tulkens and Jamar (1993),
  efficiency in Belgian municipalities.
• De Borger and Kerstens (1996), efficiency of
  Belgian local governments.
• Evans, Tandon, Murray and Lauer (2000),
  efficiency of national health systems.
• Gupta and Verhoeven (2001), education and health
  in Africa.
• Clements (2002), education in Europe.
• St. Aubyn (2003), education in the OECD.
• Afonso, Schuknecht and Tanzi (2005, 2006), public
  sector in the OECD and in emerging markets.
• Afonso and St. Aubyn (2005a, b), health and
  education in OECD.
• Afonso and St. Aubyn (2006, 2007), health and
  education in OECD using bootstrap methods.
• Afonso and Fernandes (2006, 2008), Portuguese
  municipalities.
• Afonso and Scaglioni (2007), Italian regions.
• Sutherland et al. (2007), education in OECD.
• Eugene (2007), health, education, public order
  and safety and general public services in EU15.
• Afonso, Schuknecht and Tanzi (2008), social
  spending and income distribution in the OECD.
• St. Aubyn (2008), law and order efficiency
  measurement.
• Geys, Heinemann, and Kalb (2008), German
  municipalities.
• Afonso and St. Aubyn (2009), public and private
  inputs in aggregate production in OECD.
• Another strand for instance, study of the
  determinants of (education) quality using
  cross-country regressions, Barro and Lee (2001),
  Hanushek and Luque (2003).

Some literature
A. Afonso
20
Illustrative examples of public sector
cross-country efficiency analysis
Overall public sector Education Health Social
spending
A. Afonso
21

• Public spending policies are more useful when
  they
• are limited to core/productive spending
  (including basic safety nets)
• provide services in an efficient manner
• Cross-country, sector level analysis is important
  to highlight best practices
• Social protection, health, and education
  accounted for 64-65 of total spending in the
  euro area/EU in 2006 (focus on these items)
• DEA/tobit/bootstrap/stochastic frontier
  procedures have been recently used in the context
  of cross-country efficiency analysis
• Non-parametric analysis has the advantage that a
  priori conceptions about the shape of the
  production frontier are kept to a minimum
• Parametric analysis has the advantage of allowing
  for hypothesis testing
• Care is needed in selecting as homogeneous as
  possible data as well as the right data
  (physical vs. financial resources, etc.)
• Countries far from the efficiency frontier not
  necessarily inefficient (non-discretionary
  factors)
• QPF indicators and efficiency assessments can
  help EU fiscal surveillance. SPs/CPs include a
  section on the quality of public finances
• An indirect cost of public sector provision
  inefficiency is the increase in the excess burden
  of taxation, (Afonso and Gaspar, 2007).

Conclusions
A. Afonso
22
PSP
Source Afonso, Schuknecht and Tanzi (2005).
A. Afonso
23
Public sector overall efficiency, 2000
Source Afonso, Schuknecht and Tanzi (2005).
A. Afonso
24
Illustrative evidence on public sector
performance and efficiency (considering general
government spending)
Good performance (two right-hand side quadrants),
include lower efficiency/higher spending
(Finland, Sweden, and Denmark) and higher
efficiency/lower spending (Austria, Japan,
Ireland, US).
Source Adapted from Afonso, Schuknecht and Tanzi
(2005).
A. Afonso
25
General Government functional spending ( of GDP)
Source OECD.
A. Afonso
26
Source OECD.
A. Afonso
27
DEA results
Note in this example inefficient values are
higher than unity.
With the same inputs, it would be possible to
increase the output.
A. Afonso
Source Afonso and St. Aubyn (2006).
28
Results from education tobit
Note in this example inefficient scores (d) are
higher than unity.
A. Afonso
Source Afonso and St. Aubyn (2006).
29
Health expenditure
OECD, 2003 8.7 of GDP, of which 72.5 is
public spending.
Source Afonso and St. Aubyn (2007).
A. Afonso
30
Health inputs and outputs summary
Outputs
Inputs
Source OECD.
Infant survival rate (ISR) 1000-infant
mortality rate/infant mortality rate
A. Afonso
31
Principal component analysis (PCA) for health
analysis

• PCA reduces the dimensionality of multivariate
  data
• Afonso and St. Aubyn (2007) in the case of health
  in OECD
• apply PCA to the 4 input variables
• use the first 3 principal components as the 3
  input measures (they explain around 88 of the
  variation)
• applied PCA to the three output variables
• selected the 1st principal component (it accounts
  for around 84 of the variation)
• This reduces the problem to 1 output 3 inputs
  (helpful since, as as a general rule of thumb,
  there should be at least 3 units for each input
  and output)

A. Afonso
32
Health output efficiency results DEA
Note in this example inefficient values are
higher than unity.
With the same inputs, on average, output could
increase.
Source Afonso and St. Aubyn (2007).
A. Afonso
33
Results from 2nd step health Tobit
Source Afonso and St. Aubyn (2007).
Note in this example inefficient scores (d) are
higher than unity.
A. Afonso
34
Income distribution efficiencyProduction
possibility frontier (1 input, 1 output)
Source Afonso et al. (2008).
A. Afonso
35
DEA income distribution efficiency (1 input,
public social expenditure 2 outputs, Gini
coefficient, income share of poorest 40)
Source Afonso et al. (2008).
A. Afonso
36
Production possibility frontier, CRS, 1 input
(social spending-to-GDP), 2 outputs (income share
of poorest 40, Gini)
Source Afonso et al. (2008).
A. Afonso
37

• Afonso, A. and Fernandes, S. (2006).Measuring
  local government spending efficiency Evidence
  for the Lisbon Region, Regional Studies, 2006,
  40 (1), 39-53.
• Afonso, A. and Fernandes, S. (2008). Assessing
  and Explaining the Relative efficiency of Local
  Government, Journal of Socio-Economics, 37 (5),
  1946-1979.
• Afonso, A. and Gaspar, V. (2007). Dupuit, Pigou
  and cost of inefficiency in public services
  provision, Public Choice, 132 (3-4), 485-502.
• Afonso, A. and St. Aubyn, M. (2005a).
  Non-parametric Approaches to Education and
  Health Efficiency in OECD Countries, Journal of
  Applied Economics, 8 (2), 227-246.
• Afonso, A. and St. Aubyn, M. (2005b). Assessing
  Education and Health Efficiency in OECD Countries
  using alternative Input Measures, in Public
  Expenditure, 361-388. Banca d Itália.
• Afonso, A. and St. Aubyn, M. (2006).
  "Cross-country Efficiency of Secondary Education
  Provision a Semi-parametric Analysis with
  Non-discretionary Inputs", Economic Modelling,
  23 (3), 476-491. ECB WP 494, 2005
• Afonso, A. and St. Aubyn, M. (2007). Assessing
  health efficiency across countries with a
  two-step and bootstrap analysis, ISEG/UTL WP
  33/2006/DE/UECE.
• Afonso, A. and St. Aubyn, M. (2009). Public and
  Private Inputs in Aggregate Production and
  Growth A Cross-country Efficiency Approach,
  mimeo.
• Afonso, A. and Scaglioni, C. (2007). Efficiency
  in italian regional public utilities provision,
  in Servizi Publici Nuove tendenze nella
  regolamentazione, nella produzione e nel
  finanziamento, pp. 397-418, eds. M. Marrelli, F.
  Padovano and I. Rizzo, 2007, FrancoAngeli,
  Milano, Italy. ISBN 978-88-464-8786-5.
• Afonso, A., Schuknecht. L. and Tanzi, V. (2005).
  "Public sector efficiency An international
  comparison," Public Choice, 123 (3), 321-347.
  ECB WP 242, 2003

References (1)
A. Afonso
38

• Afonso, A. Schuknecht, L. and Tanzi, V. (2006).
  Public Sector Efficiency Evidence for New EU
  Member States and Emerging Markets, ECB Working
  Paper n. 581, Applied Economics, forthcoming.
• Afonso, A., Schuknecht, L. and Tanzi, V. (2008).
  Income Distribution Determinants and Public
  Spending Efficiency , ECB Working Paper n. 861.
• Barrios, S., Pench, L. and Schaechter, A. (2009,
  eds.). The quality of public finances and
  economic growth Proceedings to the annual
  Workshop on public finance, European Economy -
  Occasional Papers n. 45.
• Barro, R. and Lee, J-W. (2001). Schooling
  Quality in a Cross-Section of Countries.
  Economica, 68, 465-488.
• De Borger, B. and Kerstens, K. (1996). Cost
  efficiency of Belgian local governments A
  comparative analysis of FDH, DEA, and econometric
  approaches. Regional Science and Urban Economics
  26, 145-170.
• Clements, B. (2002). How Efficient is Education
  Spending in Europe? European Review of Economics
  and Finance, 1 (1), 326.
• Coelli, T. Rao, P. and Battese, G. (2005). An
  Introduction to Efficiency and Productivity
  Analysis. 2nd ed., Kluwer, Boston.
• EC (2007). The EU economy 2007 review, Moving
  Europe's productivity frontier. November
• EC (2008a). Public Finances in EMU 2008.

References (2)
A. Afonso
39

• ECB (2006). The importance of public expenditure
  reform for economic growth and stability, ECB
  Monthly Bulletin, April, pp. 61-73.
• Eugène, B. (2007). The efficiency of the Belgian
  general government in an international
  perspective, mimeo, National Bank of Belgium.
• Evans, D. Tandon, A. Murray, C. and Lauer, J.
  (2000). The Comparative Efficiency of National
  Health Systems in Producing Health an Analysis
  of 191 Countries, GPE Discussion Paper Series
  29, Geneva, World Health Organisation.
• Geys, B., Heinemann, F. and Kalb, A. (2008).
  Voter Involvement, Fiscal Autonomy and Public
  Sector Efficiency Evidence from German
  Municipalities, ZEW Discussion Paper 08-024.
• Hanushek, E. and Luque, J. (2003). Efficiency
  and equity in schools around the world,
  Economics of Education Review, 22, 481-502.
• Simar, L. and Wilson, P. (2007). Estimation and
  Inference in Two-Stage, Semi-Parametric Models of
  Production Processes, Journal of Econometrics,
  136 (1), 31-64.
• St. Aubyn, M. (2003). Evaluating Efficiency in
  the Portuguese Education Sector, Economia, 26,
  25-51.
• St. Aubyn, M. (2008). Law and Order Efficiency
  Measurement A Literature Review, ISEG/UTL WP
  19/2008/DE/UECE.
• Sutherland, D. Price, R. Joumard, I. and Nicq,
  C. (2007). Performance indicators for public
  spending efficiency in primary and secondary
  education, OECD Economics Department WP 546.
• Van den Eeckhaut, P., Tulkens, H., and Jamar,
  M.-A. (1993). Cost-efficiency in Belgian
  municipalities, in Fried, H. Lovell, C. and
  Schmidt, S. (eds.), The Measurement of Productive
  Efficiency Techniques and Applications. New
  York Oxford Univ. Press.

References (3)
A. Afonso
40

• Corrective arm of the SGP
• Mentions that the Commission and the Council,
  when assessing and deciding upon the existence of
  an excessive deficit, shall take into account
  developments in the medium-term budgetary
  position (in particular, fiscal consolidation
  efforts in good times, debt sustainability,
  public investment and the overall quality of
  public finances). (see also EC, 2008a)

Quality and efficiency in the SGP
Regulation of the European Council, N.º 1467/97
of 7 July 1997, modified by Regulation N.º
1056/2005 of 27 June 2005, on speeding up and
clarifying the implementation of the excessive
deficit procedure.
A. Afonso
41
Non-discretionary inputs and tobit two-step
procedure (3)
Problems with tobit traditional procedure
- Each efficiency score estimate depends on all
observed inputs and outputs ei is serially
correlated. - The environmental variables are
correlated with both inputs and outputs ei is
not independent from zi.
Methodology exogenous factors
Simar and Wilson (2007) propose alternative
estimation and inference procedures based on
bootstrap methods. They assume
where ei is a left truncated normal random
variable.
A. Afonso
42
Data for education analysis
Source OECD.
43
Health analysis 2nd step (bootstrap)
Source Afonso and St. Aubyn (2007).
A. Afonso
44
PSP
Overall public sector
Source Afonso, Schuknecht and Tanzi (2005).
A. Afonso