Professional Documents
Culture Documents
Different Countries
gove_1465
7..36
MATT ANDREWS*
Work on good governance implies a one-best-way model of effective government. This has isomorphic inuences on development, whereby governments are inuenced to adopt a one-size-ts-all approach to get things done.
This article challenges whether such an approach exists, proposing that
models actually do not hold even for the so-called effective governments.
Governments look different, even if they are similarly called models of good
government. This proposition is examined through a study of public
nancial management practices in a set of Organisation for Economic
Co-operation and Development (OECD) and non-OECD countries. The
study shows that effective governments are not more likely to exhibit better
practice characteristics implied in one-best-way models. Good public nancial management means different things in different countries. The article
concludes by suggesting that good governance models give way to menus
and the development community invest more time in examining why different countries select different menu items.
Introduction
Variation is one of the worlds core characteristics, manifest in our abilities
to categorize things on the basis of uncountable variables and in the many
manifestations of global inequality. The international need for social and
economic development is also a broadly felt twenty-rst-century issue. In
pursuit of this second issue, however, many observers have forgotten
the rst, applying routine development solutions to different countries,
regardless of variation. The good governance movement is an example. It
manifests in inuential indicators composed of multiple dimensions that
seemingly constitute a one-best-way model of good or effective government around which all countries should convergeespecially those
needing to develop. Political and administrative reforms in many countries are directly shaped by good governance indicator scores and their
underlying best practice dimensions, with countries apparently buying
into the implied story that this is what good government looks like.
I challenge such a story in this article, arguing that the good governance
version of good or effective government is a hollow one imposing a false
*Harvard University
Governance: An International Journal of Policy, Administration, and Institutions, Vol. 23, No. 1,
January 2010 (pp. 735).
2010 Wiley Periodicals, Inc., 350 Main St., Malden, MA 02148, USA, and 9600 Garsington
Road, Oxford, OX4 2DQ, UK. ISSN 0952-1895
MATT ANDREWS
FIGURE 1
Government Effectiveness Score, 2006 (-2.5 to +2.5)
2.5
2
1.5
1
0.5
0
-0.5
-1
-1.5
-2
FSU SSA
SA
LA
EA
MENA EaU
OECD
FSU, Former Soviet Union; SSA, Sub-Saharan Africa; SA, South Asia; LA,
Latin America; EA, East Asia; MENA, Middle East; EaU, Eastern Europe;
OECD, Organisation for Economic Co-operation and Development.
10
MATT ANDREWS
11
12
MATT ANDREWS
Research Approach
With these apparent characteristics of good PFM identied, I adopt a
research approach to test the proposition implicit in the governance indicators that good government means the same thing in different countries.
I ask primarily if data show that governments considered good or more
effective have higher or lower convergence around these characteristics
than others and then examine differences in structure among governments considered more effective (the nine identied as scoring the
highest points in the WGI government effectiveness sample referenced
in Figure 1Denmark, Canada, Sweden, Australia, the Netherlands,
United Kingdom, United States, Belgium, and Germany).
I look at four specic areas related to the above list: scal rules, lump
sum appropriations and performance management, internal audit, and
legislative authority. I use data from the 2007 OECD Budget Practices and
Procedures Database, which provides quantitative data with regard to 89
detailed questions about PFM systems in 38 countries (OECD 2007a).8 The
data were collected through the OECDs Senior Budget Ofcers network
and represent the answers of these ofcers. It is the broadest one-stop
information source in the world currently and is referenced fairly widely
(e.g., see Curristine 2007; Lienert 2005). There are, however, some concerns about data quality, particularly related to the lack of stringent quality
13
control by the OECD. This necessitates careful quality control when using
the data, which I conducted (where possible) through triangulation of the
OECD data with other primary or secondary sourcesincluding government reports and legislation, and academic articles. Even with such
quality control, it is possible that the information may yield debate, partly
because some of the answers are subjective or require respondents to have
information they may not be privy to. While noted, I do not believe this
should undermine the value of my analysis. Ultimately, data on how things
are done in any administrative process are often open to subjectivity and
debate.
PFM in Good Governments: Commonly Good Outcomes but
Different Structures
What I nd in the study largely conrms the earlier discussion and provides interesting information from which to start inductively building a
framework that explains why good governments actually look different in
different countries.
I begin by looking at the degree to which good governments have
similar outcomes (something I argued above in saying that these governments are similarly successful). Decits are the only PFM outcome
measure commonly available across countries, but thankfully, these are
also a vital PFM outcome measure given current struggles with economic
downturns and scal management.9 Figure 2 shows average decits across
a section of countries for 19902006. Governments considered more effective on the WGI are at the right with decit averages below 2.5% on GDP
over this period, compared with higher averages in many of the other
governments. The gure thus suggests a convergence around lower
average decits in the more effective governments.
FIGURE 2
Average Decits for 19902006
2
Deficit as % GDP
0
-2
-1.5
-1
-0.5
-2
0.5
1.5
2.5
-4
-6
-8
-10
-12
Government effectiveness score
14
MATT ANDREWS
FIGURE 3
Decit Records for the Nine Good Governments, 19872006
0.2
0.1
USA
Deficit/surplus
20
05
20
03
20
01
19
99
19
97
19
95
19
93
19
91
-0.1
19
89
19
87
United Kingdom
Sweden
The Netherlands
-0.2
Germany
Denmark
-0.3
Canada
Belgium
-0.4
Australia
-0.5
-0.6
Year
15
FIGURE 4
Public Expenditure Patterns for the Nine Good Governments,
19872006
0.8
0.7
Australia
Spending as % of GDP
0.6
Belgium
Canada
0.5
Denmark
0.4
Germany
The Netherlands
0.3
Sweden
United Kingdom
0.2
USA
0.1
05
20
03
20
01
20
99
19
97
95
19
19
19
93
91
19
19
19
87
89
Year
0
-1.5
-1
-0.5
0.5
1.5
2.5
16
MATT ANDREWS
No rules
Budget balance rule
Debt rule
Expenditure, revenue, budget
balance rules
Australia
Belgium
Canada
Denmark
Germany
The Netherlands
Sweden
United Kingdom
United States
Fiscal Rule
Country
TABLE 1
Fiscal Rules in the More Effective Governments
Expenditure Rule
Other
No
For some types of expenditure at a
chapter level
For all expenditure at chapter level
18
MATT ANDREWS
in recent years (at least at the national level and when looking at expenditures in a strict sense).12 Germany, in contrast, found rigid rules in the
European Stability and Growth Pact impossible to enforce because of
economic challenges associated in part with unication and also with
recession.13 The United States actually had scal rules from the late 1980s
and formally still has some on the books, but these are not reected in the
OECD database partly because of their perceived lack of presence and
inuence.14 In both the United States and Germany, social and economic
challenges (the Iraq war, unication, and economic pressures) were partly
to blame for undermining the inuence of rules. Other governments in the
effective government sample are also experiencing pressure in this regard
as they face the challenges of other special costs associated with aging
populations and, more recently, the 2008 global economic downturn.15
These costs contribute uncertainty to the PFM agenda and make rigid
rules less appropriate devices for scal management. One could also argue
that they redene the role of scal decit measures as PFM outcome
indicators; in the face of spending challenges or economic downturns,
some governments might nd it less appropriate to control decits rigidly
in some years, rather allowing some slack to accommodate new policies or
demands.16
Identifying how uncertainty inuences the appropriateness of scal
rules assists one in understanding why different governments have different PFM processes in place. Hallerberg et al. underscore the importance
of this kind of understanding in their direct reference to unanswered
questions about adopting scal rules themselves: While rules seem
attractive and straightforward to contain the spending and borrowing bias
of proigate governments, it is by no means clear what institutional design
they need and how they should be embedded into the government budgeting process to be effective (Hallerberg et al. 2007, 335). They suggest
other political process inuences on scal rule adoption, identifying two
institutional approaches in countries attempting to enhance top-down
budgetary inuencedelegation and contracts (Hallerberg et al. 2007).
Delegation involves a minister of nance using rules to enforce his/her
inuence, and contracts involve actual contractual agreements about scal
behavior within the executive and between the executive and legislature.
The authors argue that delegation is appropriate for single-party governments where ideological distance and political competition is small in the
party, while contracts are appropriate for coalition governments and for
single-party governments where ruling party ideological distance and
political competition is signicant. The authors emphasize the materiality
of these differences: The European framework [of rigid rules] may be less
effective in countries whose budget process is shaped by the delegation
approach . . . [and] . . . the two are not easily interchangeable for a given
country (Hallerberg et al. 2007, 339). The story of PFM systems, related to
scal rules, thus emphasizes context and the need to embrace contextual
variation.
19
FIGURE 6
Input Control Relaxation and Government Effectiveness Scores
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
-1.5
-1
-0.5
0.5
1.5
2.5
No, expenditure
specied below
agency level
No, expenditure
specied below
agency level
Australia
Belgium
Canada
Denmark
Germany
Country
Lump Sum
Appropriations?
B. Often
C. Sometimes
D. Almost never
A. Rarely
A. Almost never
B. Often
C. Rarely
D. Almost always or often
A. Almost never
B. Rarely
C. Almost never
D. Almost never
A. Rarely
B. Rarely
C. Almost never
D. Rarely or almost never
A. Sometimes
B. Almost always
C. Sometimes
D. Almost never
TABLE 2
Relaxed Input Controls and Performance Measures
Continued
Other
No
Performance against
Objectives Routinely
Presented to Legislature
20
MATT ANDREWS
Sweden
United Kingdom
United States
A. Almost never
B. Almost always
C. Almost never
D. Almost never
A. Almost never
B. Often
C. Almost never
D. Almost never
A. Rarely
B. Sometimes
C. Rarely
D. Almost never
A. Almost never
B. Rarely
C. Rarely
D. Almost never
Lump Sum
Appropriations?
The Netherlands
Country
TABLE 2
Continued
Yes
Performance against
Objectives Routinely
Presented to Legislature
22
MATT ANDREWS
23
These observations once again suggest that there are appropriate reasons
for variation in intergovernment and intragovernment structures related to
tasks being undertaken, politics, and culture. The lack of performancebased contracts or pay may not denote ineffective government, just different government context. Context matters, as Curristine notes in discussing
performance budgeting information derived from the 2003 OECD Budget
Survey: These reforms have been introduced into an existing institutional
context and budget process (Curristine 2005, 124). She notes, In most
cases, they have not completely transformed or shifted systems away from
inputs, suggesting that, in the short run at least, the new processes actually
run in parallel to the old.17 Brinkerhoff and Goldsmith (2005, 199200) call
this institutional dualism, which further complicates the nature of
systems in countries observed: Because there are different combinations of
new and old systems likely in all domains, it is very difcult to understand
the true institutional qualities in place.18 Curristine (140) recognizes this
important issue by stating, Reformers do not begin with a blank sheet;
performance indicators and targets are introduced into existing and established systems of accountability and control, which have both informal and
formal components. Again referring to the 2003 OECD survey, she notes
further that the question reformers and observers of government effectiveness should concern themselves with centers on establishing the appropriate mix of systems for specic countries:
Traditional accountability mechanisms designed around input controls have not
been extensively relaxed in some countries. Accountability for performance will
co-exist alongside traditional mechanisms. The issue is not about completely
replacing input controls with outputs/outcomes, it is more a question of how to
nd the desired mix of mechanisms within the [individual country] system
(Curristine 2005, 140, italics added).
24
MATT ANDREWS
FIGURE 7
Internal Audit in Line Ministries and Government Effectiveness
Scores
% Budgetary entities with internal
audit
100
90
80
70
60
50
40
30
20
10
0
-1.5
-1
-0.5
0.5
1.5
2.5
25
States): Some governments legislate the need for internal audit, while
others do not; while most governments have internal audit directly in
budgetary entities, there are important exceptions that provide internal
audit through central entities; some governments have central standardsetting entities for internal audit, while others do not; internal audit entities in all governments produce similar reports (reviews of internal control
systems, nancial audits, legislative compliance audits, and performance
audits), but the time spent on the various types is very different (as is time
spent on assurance and consulting activities); and the ratio of civil servants
to internal auditor varies signicantly, from 247 in the United States to 752
in the Netherlands and 979 in Canada.
Instead of emphasizing the supposed similarities, the article could have
examined reasons for what are really notable differences, some of which it
even alludes to. Consider, for example, the differences in institutionalization of the internal audit profession (in the formation of professional
institutes in the United States in the 1940s20 vs. in many European countries like the Netherlands and Belgium in the 1970s21). This explanation is
reected partially in the articles observation that human resource organizations in governments must be able to attract and retain talent but speaks
to a larger social and cultural context in which such talent is produced.
And consider the fact that internal audit was legally recognized and
mandated as a public sector function at very different times in the
governments1978 in the United States and 2001 in the Netherlands. The
article emphasizes the importance of establishing a legal mechanism but
does not discuss how time since passing legislation might affect internal
audit coverage, activity, and inuence. The article also mentions the
importance of managerial acceptance of internal auditing as a function
(especially the modern version) but does not allude to what this acceptance hinges oncultural awareness (and the professionalization issue
already referenced), social and economic pressures to manage risks, risk,
and uncertainty avoidance, perhaps?
The Story with Accountability StructuresLegislative Authority. The
budgetary authority of the legislature is also an issue raised in recent good
governance work, especially pertaining to PFM. The PEFA instrument
used to assess the performance of PFM systems in developing countries
devotes 2 of 28 indicators to this issue, noting that government accountability is undermined where the legislature does not effectively exercise its
power. PEFA notes that the legislature should have authority over budget
review, in-year adjustments to the budget, and review of nancial reports
and audits. The OECD database asks whether legislatures do indeed have
such authority in these areas, as well as whether they have staffs to assist
and whether the time allowed for review is sufcient for the exercise of
authority.22 Figure 8 shows a legislative authority index calculated on
the basis of answers to these questions in the 2007 survey. Scores on the
exact same index are provided for 2002 in Lienert (2005).
26
MATT ANDREWS
FIGURE 8
Legislative Budgetary Authority and Government Effectiveness
Scores
12
0
-1.5
-1
-0.5
0.5
1.5
2.5
27
TABLE 3
Different Kinds of Legislative Budgetary Authority
Country
Australia
Belgium
Canada
Denmark
Germany
The Netherlands
Sweden
United kingdom
United States
Type of Government
Legislatures
Budgetary Authority
Westminster
Parliamentary monarchy
Westminster
Parliamentary monarchy
Parliamentary republic
Parliamentary monarchy
Parliamentary monarchy
Westminster
Presidential
1
4
1
5
4
6
9
1
10
Sources: 2007 OECD Budget Practices and Procedures Database; Lienert (2005).
spending bias, which was argued particularly in Swedens case in the late
1980s (Wehner 2007). This counterargument focused mostly on the inuence legislatures can have in adjusting budgetary allocations, however,
which would only make up three points in the authority measure in
Table 3, leaving eight points focused on the accountability-enhancing
aspects of legislative engagements. The fact remains that some more effective governments score substantially less than 8 and that the variation in
scores is still evident.
The observation further weakens any one-best-way argument about
what good or effective government looks like and suggests the importance
of examining context to understand what kind of system a country needs.
Table 3 alludes to one important contextual aspect that seems to dene the
level of legislative authority that may be possible or appropriate to a given
governmentits type. The table suggests that presidential governments
like the United States tend to have higher levels of legislative authority,
while parliamentary governments (and Westminster governments in particular) have lower levels. Lienert (2005) argues that this may not in fact be
as strong a pattern as the table suggests, partly because most government
types are now hybrids. Other important inuences relate to political
legacy (with French-inuenced systems looking quite different from
others, for example),24 the broader political engagement of citizens and
civil society, demand for information, and demand for legislative activism.
Where these factors vary, one can expect variation in government structures, which seems appropriate and necessary for contextually signicant
effectiveness.
One can also expect differences in acceptability of certain practices that
dene dimensions of good or effective government, like corruption and
responsiveness. Various observers explain pork barrel spending as a function of the peculiar government type and legislative strength in the United
States, for example. Most decry the practice, as referenced by Brinkerhoff
and Goldsmith (2005, 213):
28
MATT ANDREWS
Concluding Thoughts
The data and discussion around Figures 58 constitute a challenge to the
good governance proposition that good government looks the same in
different countries. Countries that exhibit good outcomes can have very
different governance structures. This evidence should challenge the
current predilection for one-best-way models of PFM systems and government structures in general. These models are being foisted on developing countries with the implied promise of development but without
evidence that the developed countries themselves uniformly adopt the
model elements. Countries that come out reecting good government
according to the inuential good governance indicators actually look very
different, varying on the very dimensions that indicators imply are central
to good government. I show this point here with regard to basic public
nancial constructs where many would probably believe there is one way:
Imagine the variation with more controversial aspects of governance.
I believe that the development community should pay more attention
to this variation. Such attention will require closer focus on the importance
of context in shaping governments. Instead of building an ever-growing
list of good governance characteristics we would like developing countries to adopt, researchers should focus on better understanding what
structures government actually do adopt and why.
It is important to note that elements of the good government picture
painted by governance indicators are already xtures in global public
sector reform programs, and context appropriateness is not a strong point
in most development engagements. A recent study of 31 African countries PFM reforms nds governments commonly pursuing multiple best
practice constructs in the form of a model: like multiyear budgets (29 of
the 31); program, activity, or performance budgets (25 of the 31); external
audit and legislative reform (28 of the 31); and internal audit (25 of the 31)
29
30
MATT ANDREWS
31
Countries with low capacity may accommodate lower levels of structural innovation in governments, necessitating more basic systems.
Government entities providing different services should have different structures, reecting task sensitivity, the degree to which tasks
are performance oriented, and the degree to which users inuence
task production.
Future research should aim to develop these ideas into real answers to
the questions posed earlier. This would be very valuable for developing
countries, not only showing an honest respect for the fact that good government looks different in different contexts but also actually trying to
explain why.
Notes
1.
2.
3.
4.
5.
6.
7.
8.
These six countries that scored above 2 and make up the top 5% in the
81-country sample.
These countries scored above 1.5 and rounded out the top 10% in the sample.
The WGI regulatory burden element has, as one of its core sources, scores
on the Heritage Foundation/Wall Street Journal Index for government
intervention in the economy, which is measured in terms of the following:
government consumption as a percentage of the economy, government ownership of businesses and industries, share of government revenues from
state-owned enterprises and government ownership of property, and economic output produced by the government.
The numbers draw from my own assessment of question 4, ak, in the 2007
OECD Budget Practices and Procedures Database, which asks about the
legal basis of the following: the form and structure of the annual budget and
related legislation, the timing of the annual budget process, roles and
responsibilities of different parts of the executive in budget formulation and
execution, roles and responsibilities of the legislature and the executive in
the budget process, provisions on what happens when the budget is not
approved by the beginning of the scal year, requirement for legislative
authorization of spending, requirement for legislative authorization of taxes,
rules for the use of contingency or reserve funds, requirement for audit of
government accounts by the supreme audit institution, requirements for
internal audit structures in line ministries, and management and reporting
relating to off-budget expenditures.
The entire group of governments was in scal trouble in the early to mid1990s, the tail of a scal expansion period that led to some signicant
adjustments in the past 15 years.
See also British Department for International Development (DFID) (2001)
and World Bank (1998), or the Asian Development Banks online toolkit:
http://www.adb.org/Governance/gov_toolkits.asp#3.
Anderson (2006) is a good reference point for this, making the prescriptive
comment: In sum, bottom up: disaster; top down: success.
Argentina, Australia, Austria, Belgium, Brazil, Canada, Chile, Costa Rica,
Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary,
Iceland, Ireland, Israel, Italy, Japan, Luxembourg, Mexico, the Netherlands,
New Zealand, Norway, Peru, Poland, Portugal, Slovak Republic, Slovenia,
Korea, Spain, Sweden, Switzerland, Turkey, United Kingdom, United States,
and Venezuela.
32
MATT ANDREWS
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Anderson (2006, 2) noted that decits were reentering the agenda in 2006,
stating: Fiscal decits have reclaimed their place as a pressing public policy
issue around the world.
Not just different degrees of adoption as suggested, for example, by Blndal
(2003).
See OECD (2002) for more detailed analysis of scal rules that conrm the
information in the table. One (Australia) has had surpluses the past few
years, while the other (United States) has recorded decits.
Anderson (2006, 31) argues that the rigid rules have not stemmed expenditures at the local level (not covered by rules) and have led to increased use
of tax expenditures to introduce new policies without breaching the ceiling
or requiring balancing measures.
Posen (2005, 5) writes: Germany, along with other Eurozone members
France, Portugal, and more recently Italy, has been in repeated violation of
these rules. Posen argues that this is partly because Germany has of course
suffered from a prolonged recession and historically high unemployment
since 2001, which has put signicant pressure on scal policy (5).
The U.S. experience is well discussed in Anderson (2006) and in Schick
(2005), who discusses the situation as such: The Gramm-Rudman-Hollings
laws (GRH) enacted in 1985 and 1987 purported to limit annual budget
decits; the 1990 Budget Enforcement Act (BEA) capped annual appropriations and required that any legislation increasing the decitor decreasing
the surplusbe offset. BEA expired at the end of scal 2002, but some of its
rules have been reimposed in congressional budget resolutions. These have
not been effective.
Interestingly, a country like Sweden may face less pressure from such costs
because of the historical role government has played in providing social
welfare (something that was criticized in the decit years of the early 1990s).
This established role decreases uncertainty about future demands.
In some instances, this will be reected in structural decit measures, which
should account for economic cycles, but these measures do not reect potential social challenges that may be demographically induced, or other challenges that governments may face. Countercyclical budget management
approaches are increasingly being introduced to facilitate policy continuity
and guide spending.
Andrews and Hill (2003, 126) note this in a study of performance budgeting
in the U.S. states: In essence, PBB [performance-based budgeting] reforms
involve the introduction of new rules and norms to drive budgeting behavior, which have to overcome the inuence of pre-existing rules and norms
(usually associated with incremental and program budgeting systems) in
order to inuence behavior.
This is a problem for clean, one-best-way-type theory and for data heads
gathering information about such. How do you develop clean theory or
gather clean data about the extent of a practices implementation when such
is implemented in the presence of an existing practice, is often interrupted
by the introduction of new practices, and has formal and informal dimensions? The common approach is to look at legislation and see whether
innovation A or B is in place, which is obviously problematic.
Some countries interpret this function as the equivalent of inspectorates, for
example, and in some Eastern European settings, the internal audit ofce is
literally the remnants of the KGB (the National Security Agency of the
USSR). In the countries identied here, one can expect very signicant
differences in who the internal auditors are, what they do, etc.
The U.S. Institute of Internal Auditors (IIA) was founded in 1941, although
the internal audit function was entrenched in industry in the 1930s.
21.
22.
23.
24.
25.
33
References
Anderson, Barry. 2006. Top Down Budgeting and Its Relationship to Fiscal Rules
and Delegation of Authority. Presentation at the Meeting of Senior Budget
Ofcials from Latin American Countries, Santiago, Chile, January 25, 2006.
http://www.eclac.org/ilpes/noticias/paginas/5/22145/TOP%20DOWN%
20BUDGETING%20&%20ITS%20RELATIONSHIP%20TO%20FISCAL.pdf.
(March 2009).
Andrews, Matthew. 2004. Authority, Acceptance, Ability and Performance-Based
Budgeting Reforms. International Journal of Public Sector Management 17 (4):
332344.
. 2008a. The Good Governance Agenda: Beyond Indicators without
Theory. Oxford Development Studies 36 (4): 379407.
. 2008b. Managing Money for (a) Change. Working Paper, World Bank.
Andrews, Matthew, and Herb Hill. 2003. The Impact of Traditional Budgeting
Systems on the Effectiveness of Performance-Based Budgeting: A Different
Viewpoint on Recent Findings. International Journal of Public Administration 26
(2): 135156.
Arndt, Christiane, and Charles Oman. 2006. The Uses and Abuses of Governance
Indicators. Paris: OECD.
Bizony, Piers. 2006. The Man Who Ran the Moon. New York: Thunders Mouth Press.
Blndal, Jon. 2003. Budget Reform in OECD Countries: Common Trends. OECD
Journal of Budgeting 2 (4): 133.
Boivard, Tony, and Elke Lfer. 2003. Evaluating the Quality of Public Governance: Indicators, Models and Methodologies. International Review of Administrative Sciences 69: 313328.
Brinkerhoff, Derick W., and Arthur A. Goldsmith. 2005. Institutional Dualism and
International Development: A Revisionist Interpretation of Good Governance.
Administration & Society 37: 199224.
34
MATT ANDREWS
35