Workshop by the Federal Ministry of the Interior and the Forum of Federations on
Benchmarking in Public Administrations in Federal States held in Berlin on 15
On 15 February 2008 the Federal Ministry of the Interior and the Forum of
Federations hosted an international workshop on benchmarking in public
administrations in federal states. The workshop was based on proposals made by the
German government in the framework of the second stage of the reform of
Germanyâ€™s federal structure to institutionalize benchmarking as a tool to compare
performance in order to improve the quality of administrative action (effectiveness,
efficiency, practicability, responsiveness to citizensâ€™ needs, staff satisfaction).
The following conclusions were drawn from the workshop:
1. Benchmarking in public administrations is a valuable tool to compare
performance in a federal state. Administrations should be in a position to
compare themselves with each another and to learn from best practices.
Benchmarking requires permanent evaluation and adaptation to fulfil its
functions. This also includes the definition of the relevant data basis.
2. Benchmarking studies can be conducted in public administrations and
possibly in businesses performing the same/similar tasks. Not only the
cooperation of experts from the administration, but also of external experts
(methodological expertise) and if possible of partners from industry has
3. The introduction of benchmarking requires the political will of the highest
level of the federation and the constituent units. This political will was
expressed by the decision taken by the Conference of Minister-Presidents on
15/16 December 2004 and by the statements of the Federal Government and
the coalition parties on benchmarking as part of the second stage of the reform
of Germanyâ€™s federal structure.
4. If the constitution provides for close administrative and financial relations
between the Federation, the constituent units and local authorities, all three
levels must be included in the benchmarking procedure.
5. It is desirable to ensure that the benchmarking of public tasks not only focuses
on costs, but also takes the effects of state action into account.
6. Organizational solutions for the federation and the constituent units are
necessary to implement benchmarking.
7. Benchmarking results should be made public. A comprehensive discussion of
benchmarking results and of possible conclusions conducted among the
benchmarking partners should be an integral part of the benchmarking
8. The speakers reported that benchmarking so far has not been enshrined in
their countriesâ€™ constitutions. However, they felt that any support for the
development of a sustainable benchmarking culture in federal systems was
After State Secretary Beus and the President of the Forum of Federations, Mr. George
Anderson, opened the workshop, the following speakers presented their experience:
– Dr. Gottfried Konzendorf, desk officer, Project Group on the Second Stage of the
Federalism Reform, Federal Ministry of the Interior, on the German discussion on
benchmarking within the framework of the second stage of Germanyâ€™s
– Jeff Tryens, consultant, former Executive Director of the Oregon Progress Board,
on federal-state benchmarking in the United States with three examples from
– John Wright, former Deputy Minister, Saskatchewan, Canada, on benchmarking
healthcare in Canada,
– Dr. Reto Steiner, University of Berne, Competence Centre for Public Management,
on benchmarking in the public administration in Switzerland, and
– Roger Wilkins, Head of Government & Public Sector Group, Citigroup (Australia
and New Zealand), former Director-General of the Cabinet Office of New South
Wales, Australia, on federalism and benchmarking in Australia.
The presentations specifically highlighted the following:
Dr. Gottfried Konzendorf (Germany, Federal Ministry of the Interior) :
Benchmarking can be regarded as a method of learning
Â· to improve the quality of workflows, organizational structures and work
results (input, output, outcome)
Â· based on a systematic comparison and exchange of experience of best
Â· between partners (organizations, local and regional authorities) performing
the same or similar tasks or trying to solve similar problems.
Germany, like other countries, participates in international benchmarking studies,
such as PISA or studies on better regulation. Benchmarking is also used for public
tasks in Germany. In healthcare, for example, there are ten comparative studies
entitled "Benchmarking healthcareâ€. The tool is also used in other policy areas. The
German consumer association "verbraucherschutz bundesverbandâ€ (vzbv) has
already conducted a number of federal benchmarking studies on consumer
protection policy. Other successful applications include the benchmarking studies of
tax offices, local benchmarking studies of the IKO network and the benchmarking of
When analyzed, these successful benchmarking processes produce the following
– Benchmarking in federal systems is possible and useful.
– Benchmarking helps improve the quality of administrative action.
– The methodological approach depends on the issue at stake. It is not possible to
develop a uniform method that can be applied to all issues.
– Innovation partnerships between politicians, administrative staff and external
consultants are recommended for the implementation of benchmarking to be
Despite these positive examples there is still a huge demand for the use of
benchmarking in the federal system of Germany. The decision taken by the
Conference of Minister-Presidents on 15/16 December 2004 on comparisons of
quality and performance has not been consistently implemented. Furthermore there
is no sufficient data basis for comparative studies beyond the constituent states.
In order to extend benchmarking in our federal system a new section on
"Cooperationâ€ with the following article should be included in the Basic Law (the
(y) The Federation and the LÃ¤nder shall conduct benchmarking studies to identify
and promote the performance of their administrations and publish the results.
The Federation and the LÃ¤nder should also take the decision to set up a
benchmarking agency aimed at improving the implementation of benchmarking.
This may be done by the Deutsches Forschungsinstitut fÃ¼r Ã¶ffentliche Verwaltung
Speyer (FÃ–V, German Research Institute for Public Administration), in which both
the Federal Government and the Federal LÃ¤nder cooperate and which is renowned for
its methodological expertise. The following points must be taken into account, when
assigning this responsibility to the FÃ–V:
Â· When conducting or assigning benchmarking studies, the FÃ–V will be
supported by an advisory board of national and international administrative
Â· The FÃ–V management board (members: Federation and the LÃ¤nder) will
decide on benchmarking topics upon the proposal of the members of Working
Group VI of the Conference of Interior Ministers, the Federal Ministry of the
Interior or other Ministers Conferences.
Â· The methodological basis and the results of benchmarking studies will be
published by the FÃ–V in an appropriate form.
Â· The Federal Government and the Federal LÃ¤nder are free to decide whether to
participate in a benchmarking study and whether to draw any political
Â· In December each year, a report will be published on the results of the
benchmarking studies of the past year together with a plan for the coming
year, which will be discussed by the Federal Chancellor and the Minister-
Presidents of the LÃ¤nder. This report will be forwarded to the Conference of
Interior Ministers (Working Group VI) and possibly to other Ministers
Jeff Tryens ( USA, Oregon):
Most programs in the U.S. involve two, and often three, levels of government.
Typically, the federal government provides significant funding and rules. State
governments add additional funding and assume management responsibility for the
program. Local governments, when involved, deliver services for the state
government or oversee the work of those who do. State and local governments are
fiercely protective of their policy-making prerogatives, balking at undue oversight
from above. Benchmarking initiatives are further complicated by the sheer number of
state and local governments involved. Over 80,000 local government units operate
within the 50 states.
Perhaps because of this vast proliferation of sub-federal governments, federal
initiatives that force performance-based competition between states are basically
non-existent. The speaker was not aware of any federal program that pits states,
head-to-head, in competition based on their respective measurable results.
This presentation will examine three types of results-focused federal programs from
one stateâ€™s perspective. The three types could be called: cooperative, soft prescriptive
and hard prescriptive.
The cooperative example, known as the Oregon Option, was aimed at creating better
results for Oregon by reworking intergovernmental relations. The soft prescriptive
example, Healthy People, is based on a federal executive initiative designed to spur
action at all levels of society by creating a well documented, target-based vision for
the future. The hard prescriptive approach, No Child Left Behind, is based on a federal
law requiring schools receiving certain types of federal funding to improve the
academic performance of all their students.
This paper is based on the speakerâ€™s experience as former executive director of the
Oregon Progress Board. As the group responsible for administering the stateâ€™s
results-based strategic vision, Oregon Shines, the Board has had some involvement in
each of these approaches over its 19 year history.
Cooperative Approach: Oregon Option
In 1994 a united front of Oregon state and local elected leaders convinced the Clinton
Administration that Oregon was an ideal proving ground for trying out new
approaches to government devolution based on shared objectives. Oregon was a
known policy innovator. It was far from the glare of East Coast media. And, perhaps
most importantly, it had the Oregon Benchmarks: a statutorily-required set of
measures focused broadly on the stateâ€™s well being.
Over the next four years, Oregon and the federal government worked to improve
results in three "clustersâ€: child health, family stability, and workforce development.
The initiative was considered a key cog in Vice President Al Goreâ€™s reinventing
government machinery. Thousands of hours at all levels of government were
invested in working out new ways to do business.
In preparation for this presentation, I polled ten key players involved in the Oregon
Option. According to this admittedly small sample, their greatest hopes for the
Oregon Option were: 1) achieve better outcomes in the clusters; 2) change the culture
of federal-state or local relationships; 3) change the internal functioning of
federal/state/local agencies. While the respondents felt that better outcomes had been
achieved in the clusters, they felt that culture change and improved bureaucratic
operation had been only somewhat or not at all successful. In short, the big picture
results were disappointing.
Soft Prescriptive Approach: Healthy People
In existence for nearly three decades, this U.S. Department of Health and Human
Services (DHHS) program is described as enabling "diverse groups to combine their
efforts and work as a team.â€ It provides data and national targets for 28 health topics
with 467 specific objectives ranging from HIV to health education. Some preventionrelated
grants and no penalties are associated with the program. State coordinators
are encouraged to share information with one another. The federal government also
supplies standardized analysis and promotional materials. Comparative data on
state progress toward the national targets is not a feature of the program.
Metrics included in programâ€™s first iteration, Healthy People 2000, were Oregonâ€™s
measures of first choice when establishing the health-related Oregon Benchmarks.
And the national targets set by DHHS guided the Progress Boardâ€™s discussion
regarding state targets. As federally-generated measures, the Healthy People metrics
have had a certain cache with public health professionals. Oregon has participated in
a national consortium that was formed to advance the goals of Healthy People 2010.
While still in existence, Healthy People 2010 appears to have lost steam at the state
level over the years. A former state "Healthy Peopleâ€ coordinator told me that
Healthy People 2010 drives no decisions at the state health department. Nevertheless,
the federal government is now preparing Healthy People 2020.
Hard Prescriptive: No Child Left Behind (NCLB)
This federally mandated program, begun in 2002, requires all states to "have
assessment systems, report disaggregated data and target federal resources to serve
their neediest studentsâ€ according to the federal secretary of education. Each state
sets its own targets, without regard to national benchmarks, which schools receiving
certain types of federal funding are required to meet for all students. Data is broken
down into specific subsets of students for every participating school. Individual
schools are subject to state and federal interventions, both positive and negative, as
part of the initiative.
An interventionist, performance-based federal education policy is revolutionary in
the U.S. where local elected school boards reign. Oregon state government, which
has little control over what schools do, has welcomed the initiative. According to one
state official, it provides "more and better opportunities to force change.â€ On the
other hand, the federal government appears to treat NCLB as a one-size-fits-all tool
with little collaboration among states or between states and the federal government.
Again, state-to-state comparisons are not relied upon as part of act implementation.
While many problems remain, NCLB has established the federal governmentâ€™s role
in setting standards for educational performance. Perhaps its greatest achievement is
that children are no longer passed from grade to grade without the requisite learning
needed. The legislation is currently up for renewal in the U.S. Congress.
1. Desired outcomes must truly be shared. Despite powerful federal political support
for the Oregon Option, Oregon participants felt that the federal bureaucrats
involved never really bought into the process. Sometimes mid-level federal
bureaucrats acted as though the Oregon Option was a politically motivated freepass
that they had an obligation to thwart.
2. Accountability for results must be carefully defined. The Oregon Benchmarks, while
important, were generally too high level to ascribe cause and effect relationships
between government intervention and changes in benchmarks. A much more
sophisticated set of metrics based on strategy implementation is needed.
3. True systems change can only happen over the long haul. Somehow, cooperative
commitments must transcend terms of office. The Oregon Option was the product
of a connection made between a state governor and the vice president of the
United States. The departure of that governor in 1995 marked the beginning of
the decline of the Oregon Option long before the end of the Clinton
Administration, which sounded its death knell.
1. Data alone doesnâ€™t do it. Comparative data alone will generally motivate high
performers to even higher levels but do little for others. Even in a field like public
health, where most players are comfortable with data, the mere existence of
regular, reliable, comparable data has done little to change the direction of health
trends in Oregon in my opinion.
2. Focus is critical. A data-rich environment, allowing comparisons across a broad
range of related issues, is some peopleâ€™s idea of heaven. Itâ€™s my idea of hell.
Imagine for a moment a yearly matrix of Healthy People data with 267 objectives
across 50 (or even Germanyâ€™s 16) states. Layering and priority-setting must be
employed when tackling big issues.
3. Soft incentives, like recognition for performance, are worthwhile. Oregon uses sub-state
level data to motivate counties to focus on benchmark-related issues by
recognizing high performing counties and high improvement counties through
an awards program. Free analysis is also provided upon request. More than once
Iâ€™ve visited with a county leader whose office displayed the Progress Boardâ€™s
certificate of achievement.
1. Even mandates should be soft around the edges. Benchmarking programs should be
carefully constructed to allow states to measure progress in ways that work. For
instance, Oregon is changing its measurement system to track progress over the
course of a classroom year rather than comparing achievement of last yearâ€™s fifth
graders to this yearâ€™s.
2. Mandates alone wonâ€™t create a performance culture. A school principle in Oregon
recently told a state official that if he just had two fewer of a certain kind of
student he could get out from under the No Child Left Behind requirements. What
kind of a learning atmosphere is that for a disadvantaged child?
3. The theory of change must be apparent. Mandates are hard to swallow. Without a
clear understanding of what the performance model is, even nascent supporters
will balk. Is the model based on front line empowerment or top down authority,
Dr. Reto Steiner (Switzerland, Canton Berne) :
In Switzerland, the federal administration, the cantons and the local authorities use
benchmarking in very different ways:
Â· Federal level: Switzerland increasingly participates in benchmarking projects
conducted by international institutions, such as the OECDâ€™s PISA studies.
Theses comparisons are intended to assess the effectiveness of individual
policy areas. Furthermore, the Federal Government promotes benchmarking
projects between the cantons, e.g. in education and healthcare. This is
implicitly required by legal acts laying down the principle of economical and
efficient performance (e.g. Art. 43a of the Federal Constitution), which must be
evaluated on a regular basis (e.g. Art. 170 of the Federal Constitution and Art.
36 of the Act on the Organization of the Government and the Administration).
The Federal Government often acts as a facilitator, e.g. by tasking the Federal
Statistics Office with the development of indicators, with organizing basic
data or even with the interpretation of data. Federal offices are still hardly
ever compared with one another.
Â· Cantonal level: At cantonal level, benchmarking projects are mainly initiated by
the Federation or by conferences of directors (i.e. cantonal ministersâ€™
conferences). They usually focus on comparisons of outputs and outcomes
between the cantons, rather than on comparisons of entire organizational units
within cantons. Furthermore, cantons promote benchmarking at local level,
because they want to make sure that the tasks assigned to local authorities are
fulfilled in line with the requirements established by the cantons and that
financial transfers are used efficiently.
Â· Local level: At local level there is a large number of benchmarking projects.
They are initiated by the local authorities, by cantons or by other
organizations, such as associations or consultancies. Since local authorities are
usually small organizational units, there is a greater readiness to conduct
comprehensive comparisons of organizations. In addition to systematically
conducted projects, local authorities regularly carry out informal comparisons,
as has been shown by a national survey of local authorities. The effects of such
informal benchmarking should not be underestimated.
The following factors influence the success of benchmarking projects:
Success factor 1: Commitment of political leaders
Â· There must be a clear commitment to benchmarking, i.e. the support of
political leaders and the heads of administrations.
Â· Staff members must be involved in the process ("onlookers must become
participantsâ€). Comprehensive information and training are a possibility to
reduce existing barriers (barriers based on a lack of knowledge, abilities,
wishes and legal norms). Fears of staff members as to whether a
benchmarking process will eventually make staff redundant or involves
individual sanctions, must be taken seriously.
Success factor 2: Integration and form of controlling
Â· Benchmarking is part of a comprehensive controlling and quality control
strategy. This tool particularly makes sense if it is integrated in a new
management and control model of the institution concerned. Quality
management (e.g. the Common Assessment Framework) should be part of the
Â· For a systematic benchmarking the administration should know its products
and design accounting in such a way as to enable absorption costing.
Success factor 3: Partners and objects of comparison
Â· The institutions chosen for a comparison in the first benchmarking study
should be partners with similar tasks, i.e. with only minor differences. In a
first stage, this would increase acceptance and facilitate work.
Â· Benchmarking partners from the private sector are an added benefit.
Â· Benchmarking can be initiated by institutions themselves, by the superior
state level or by independent institutions. These three possibilities have their
specific advantages and disadvantages. Cooperation with professional
methodological experts and/or the creation of centres of expertise can be
useful to benefit from know-how synergies.
Success factor 4: Applied procedures and tools
Â· The cost-benefit ratio should always be reasonable. Bureaucratic and
oversized benchmarking procedures should be avoided.
Â· State benchmarking should not exclusively focus on costs (input orientation),
but instead centre on processes and output/outcome.
Success factor 5: Evaluation, communication and improvement
Â· The internal discussion of results is essential.
Â· Results should be communicated in public and used as a basis for real
improvements. Communication at the start of the project and after the
evaluation stage creates a certain pressure to act. Again, it is necessary to have
the support of superiors and political leaders. This makes change more
probable. It is important to pursue benchmarking continuously.
John Wright (Canada, Saskatchewan):
The use of comparable health care indicators for clinical, policy, program and
research purposes is reasonably extensive in Canada. However, the use of
benchmarking, based on best practice or medical evidence, is relatively new. In
Canada, provinces provide health care services to 97 percent of the population. The
federal government contributes less than 25 percent of the cost of provincial services
through annual cash transfer payments to the provinces. Provinces are protective of
their constitutionally assigned jurisdictions and do not welcome federal intrusions
into their areas of responsibility.
Health care data has been traditionally collected by the provinces, Statistics Canada
and the Canadian Institute for Health Information (CIHI). CIHI is a federalprovincial
agency that is jointly funded and administered with a mandate to be a
"source of unbiased, credible and comparable health information.â€
At the beginning of this decade, a rising sense of urgency to improve the quality and
timeliness of health care led the federal government to inject additional funding into
the system and in exchange the provinces were to provide increased transparency
through comparable indicator reporting.
In 2000, 2003 and 2004, the First Ministers agreed to a series of health care accords
that, among other initiatives, greatly increased the number of comparable indicators
that were to be reported annually to the Canadian public. In addition, the 2004
accord required the provinces to expand comparable indicator reporting to four
surgical areas and to produce evidence based benchmarks for each of these areas by
December 2005. There is no legal basis for the accords and the benchmarking
In the development of the comparable indicators and benchmarks, no rigorous
methodology was employed. A functional or collaborative approach was used with
the following elements:
Â· Organize: A steering group of Deputy Ministers of Health was formed along with
a working group that included Statistics Canada and CIHI officials.
Â· Plan: Clear definitions for comparable indicators were established and the
necessary data infrastructure was defined. Similarly, a process for determining
benchmark definitions for the four surgical areas was created.
Â· Collect Data: Best practices for data collection infrastructure across the country
were examined and shared and implementation began in most provinces.
Numerous challenges were encountered. A federal funding agency for health
research was contracted to determine the evidenced based surgical benchmarks.
Â· Report Progress: Reports on up to 70 comparable indicators by province were
made public in 2002, 2004 and 2006. Additionally, the evidence based
benchmarks (8 in total) were published in late 2005.
Â· Analyze/Refine: Provinces were also required to produce by December 2007 a
public plan on how they were going achieve the evidence based surgical
benchmarks. No province met the deadline.
Â· Adopt Best Practices: Provinces are currently implementing best surgical practice
methodologies and programs. Progress to achieving the benchmarks is being
made. Each province has implemented one wait time guarantee relative to one of
There was some early resistance to the development and publishing of the
comparable indicators and benchmarks. Some provinces did not want to be
compared to others. Also the cost and difficulty of collecting consistent data was
deemed a barrier by some provinces. This resistance was largely overcome due to
peer pressure, the public commitments of First Ministers, public pressure and
pressure from various health care provider organizations.
Experience has shown that better upfront planning is needed. Work will continue
with CIHI to ensure consistency in data definitions and collection and the sharing of
A benchmark research agenda is needed. A process to establish the next set of
benchmarks including who decides, the order of priority and how the research will
be undertaken needs to be established. Provinces need to look outside the health care
system to establish certain benchmarks. The involvement of business schools,
industrial engineering faculties and other partners is critical to providing robust
benchmarks based on clinical evidence or best practice.
Public transparency and collaboration among the provinces have improved. The
patient is on the road to recovery.
Roger Wilkins (Australia, New South Wales):
Since 1993, Australia has had a system of benchmarking government services
delivered by State governments. There is no particular legal basis, only a decision of
the Prime Minister and the Premiers.
There is an annual report on performance published by an independent Federal body
called the Productivity Commission. The report covers services such as housing,
health, education, policing, courts and justice, community services, and indigenous
affairs. All of these services are delivered by the States. The Productivity Commission
has been integrated into the following structure:
Progress has been made over the last 15 years in refining the system, but much still
remains to be done in agreeing on and defining outcome measures.
State politicians sometimes find the report politically embarrassing, and the level of
State support has been mixed. It has, however, enabled a degree of comparability
and competitive federalism to drive efficiencies and innovation.
The new Australian Government (Labor elected in November 2007) has an agenda of
wide ranging reform in government service delivery. It also works to get the federal
system to work more efficiently. A focus on outcome measures and benchmarking is
likely to see the reinvigoration/reform of the current benchmarking system.
A major feature of Australiaâ€™s federal system is the level of vertical fiscal imbalance
the Federal Government raises most of the revenue; and there are large transfers to
the States. This is an important reason why Federal governments have been able to
encourage or coerce the States into benchmarking and reporting performance. But
State Treasuries/Finance Departments and Premier and Treasurers have also found
benchmarking a useful discipline in driving efficiencies.