MEASURING CORRUPTION

1
MEASURING CORRUPTION AT THE COUNTRY LEVEL
Training programme
Photo by Adam Rogers/UNCDF
Photo by Catherine Jaimeson/ Flickr
United Nations Development Programme
2
What we will cover this afternoon

• Contrast how different global composite
  indicators define and measure corruption
• Examine the strengths and limitations of
  composite and original indicators
• Identify data that is actionable
• Identify at least 4 ways in which indicators can
  be made pro-poor and gender sensitive

3
What we will cover this afternoon

• Weigh up the advantages and disadvantages of
  using perception and experience-based data
• Explain the rationale for using complementary
  indicators (input vs. outcome de jure vs. de
  facto)
• Distinguish the distinct purposes of various
  assessment tools
• Introduce a web-based toolbox of ready-made
  corruption assessment methodologies

4
Users Guide to Measuring Corruption (2008) By
UNDP Global Integrity

• Literature review of best practices for
  corruption measurement, and classification of
  existing tools
• Based on the views of professionals
• tasked with measuring corruption
• Identifies good practices
• Case studies

5
Key Findings

• Practitioners want actionable data, and existing
  global indicators are inadequate.
• Perception-based data viewed as the least useful
  data.
• Indicators customized to country specificities
  are most useful and seen as more credible by
  policymakers
• Numbers need context qualitative data and
  narrative, political economy studies

6
Part One International sources of corruption
Indicators
Photo by Konomiho/flickr.
7
What are international corruption indicators used
for?

• 1. Ranking countries tracking changes over time
  is being used in order to
• Name and shame governments and actors seen to be
  the worst corruption offenders
• Praise good performers and promote a virtuous
  cycle of competition
• Inform decisions about aid allocations
• Inform investment decisions

Photo by takeabreak/flickr
8
Ranking countries with composite indicators
Composite indicators of corruption combine
existing indicators into a single, aggregate
measure. International efforts to measure
governance / corruption began to take off in the
late 1990s with the use of composite indices.
9
The mushrooming industry of indicators
10
Discrepancies in rankings
Corruption in China in 2007 3 indices 3 stories

11
Discrepancies in rankings
Corruption in China over time (2004-06) Two
indices, two stories

According to the World Bank, China is doing
better in 2006 than it was in 2004
According to TI, China is doing worse in 2006
than it was in 2004
12
How to explain these discrepancies in rankings?

• What is each index measuring?
• What types of indicators make up the composite
  score for each index?
• What data sources are used?
• Are component indicators and data sources the
  same every year?

13
What is each index measuring, and what data
sources are used?

Corruption in public sector as perceived by
experts
Existence, effectiveness and citizen access to
anti-corruption mechanisms, assessed by national
experts
Corruption in public private sector ( some
indicators at household level) as perceived by
experts opinion polls (incl. NGO experts)
14
The CPI Digging deeper...

• The CPI is the most famous source of information
  about the level of corruption.
• To what extent is this ranking reliable?
• To what extent is it useful?
• Three sets of challenges
• 1. Ambiguities in definition
• 2. Methodology
• 3. Perception data

15
1. Ambiguities in definition

• What is measured by the CPI?
• Surveys ask different questions related to
  corruption they do not cover the same issue!
• Political (grand corruption) vs. lower-level
  bureaucratic corruption
• Frequency of corruption acts
• Size of bribes
• Legal or illegal activities (What is called
  corruption in one country differs from another
  country)
• Cost to society

16
1. Ambiguities in definitions

• Composite indicators have no explicit
  definition, but instead are defined implicitly by
  what goes into them.

Selected sources (out of 14 in 2007) Subject asked Respondents Coverage
World Economic Forum (WEF) Undocumented extra payments or bribes connected with various government functions Senior business leaders domestic and international companies 125 countries
Bertelsmann Transformation Index (BTI) The governments capacity to punish and contain corruption Network of local experts inside and outside the organization 125 countries
Economist Intelligence Unit (EIU) The misuse of public office for private (or political party) gain Expert staff assessment 166 countries
17
2. Annual variations in methodology
Rank 71th
Out of 145
9 surveys out of 12
16 surveys out of 18
Rank 70th
Out of 163
14 surveys out of 16
Rank 78th
Out of 158
18
3. Perception data

• Lack of diversity in sources providing data for
  CPI
• Private-sector biased 12 out of 14 institutions
  have a private sector focus
• But preferences of firms dont always match those
  of society (e.g. negative ratings to strong
  labour / environmental regulations, high taxes)
• Typical respondent Male economically
  well-off
• Missed perspectives / interests of vulnerable
  groups, women unofficial businesses which
  employ majority of population in developing
  countries

19
3. Perception data (cont.)

• CPI often used as a true (instead of perceived)
  corruption level estimate,
• BUT a gap between perceptions and actual level
  of corruption is to be expected
• Experts read same reports and are influenced by
  other experts
• CPI and WB Control of Corruption scores for
  China have 7 sources in common
• Experts influenced by the same (biased /
  misleading) sources, e.g. media, court cases
• THEREFORE A country score showing high
  accuracy (low margin of error) is in fact
  derived from sources influenced by the same
  mainstream information!

20
Tip 1 Ignore an indexs label and dig underneath

• Different scores on 2 different indices does not
  mean that one index is right and the other is
  wrong.
• Only that the 2 indices are assessing different
  concepts
• Quick tip
• 1) Know what want to measure
• 2) Ignore an indexs label
• 3) Unpack component questions / indicators
• 4) Which index really measures what you want to
  measure?
• 5) Or do you need to generate your own research?

21

• Part two
• National corruption indicators
• Complementarity in the use of indicators
• input vs. outcome-based indicators
• de jure vs. de facto indicators

Photo by Konomiho/flickr.
22
Second generation of measurement tools focused
on single countries Original indicators

• Original indicators of corruption based on
  originally (and locally) generated rather than
  third party data
• Can be single-country in focus, but many repeat
  methods across countries for comparison (e.g.
  Open Budget Index, Global Integrity Index, TIs
  Global Corruption Barometer)
• Major strength produce highly detailed and
  contextualized information that can be
  disaggregated to reveal local nuances (e.g.
  patterns between corruption incidence and income
  levels)
• Weaknesses
• A high degree of contextualization makes
  cross-country comparisons difficult

23
What are national corruption assessments used for?

• Country-specific corruption assessments may be
  used for
• Understanding the drivers of corruption and the
  blockages to reform (diagnostic)
• Assessing the effectiveness of anti-corruption
  mechanisms (monitoring)
• Identifying capacity gaps (capacity development)
• Understanding the impact of corruption, e.g. on
  marginalised groups, on business, on social
  service delivery (costs of corruption)
• Monitoring national AC strategies (and/or UNCAC
  other international AC conventions)

24
Activity Identifying actionable indicators

• Which, if any, of these indicators are
  actionable? (drawn from CPI and WBs Control of
  Corruption indicator)
• 1. How often do firms make extra payment to
  influence the content of new legislation?
  (Business Enterprise Environment Survey)
• 2. How many judges and magistrates do you think
  are involved in corruption? (Regional barometers)
  
• How would you make them, more actionable?

25
Activity two examples
Non-actionable Actionable
1. How often do firms make extra payment to influence the content of new legislation? (Business Enterprise Environment Survey) 2. How many judges and magistrates do you think are involved in corruption? (Regional barometers) In practice, are the regulations restricting post-government private sector employment for national legislators effective? In practice, are the requirements for the independent auditing of the asset disclosure forms of members of the national legislature fulfilled? In practice, when necessary, does the judicial disciplinary agency initiates investigations? In practice, when necessary, does the judicial disciplinary agency impose penalties on offenders?
26
Actionable indicators a word of caution.
Not everything that counts can be counted, and
not everything that can be counted counts.
Albert Einstein
27
Actionable or action-worthy?

• What are the pitfalls of actionable indicators?
• Simply because something can be measured does
  not mean that it is an important constraint on
  good governance.
• Actionable does not mean action-worthy
• E.g. Measuring whether a country has an
  independent anti-corruption commission when there
  is no guarantee it will reduce corruption
• E.g. Measuring the speed of judicial proceedings
  when not clear that increasing the speed will
  ensure that justice is done
• Risk of measuring things because they are easily
  measurable, leading to teaching to the test and
  reform illusion

28
Determining what is an action-worthy indicator

• Is it measuring interventions that are truly
  beneficial?
• Assessing performance of interventions only
  makes sense if these are first deemed appropriate
  for the particular country context
• Prior research can identify what reforms to
  prioritize, and ones that risk further
  entrenching corrupt systems e.g. political
  economy studies such as National Integrity
  Systems assessments by TI
• 2) Can the performance measured be attributed to
  specific policy inputs? (actionable indicator)

29

• Complementarity in the use of indicators
• Input/de jure indicators vs. outcome/de facto
  indicators

• Input/de jure indicator In law, is there an
  agency with a legal mandate to address
  corruption? Does it receive regular funding, a
  professional full-time staff?
• Outcome indicator In practice, is the
  anti-corruption agency effective?
• What are some factors which may affect the
  effectiveness of the agency?

30
What are some factors which might affect the
effectiveness of the agency?

• Head of ACA protected from removal without
  relevant justification through a formal
  process?
• Difficulties accessing politically sensitive info
  necessary for carrying out investigation?
• Subject to favourable or unfavourable criticism
  by govt?
• Political appointments vs. based on professional
  criteria?
• Conflicting family relationships / personal
  loyalties?
• Threats / harassments?

31
Complementarity in the use of indicators 1.
Input vs. outcome indicators

• Combine input and outcome-based indicators to
  show discrepancies between
• change in law resources (, infrastructure,
  procedures, staff)
• and
• change in practice (improved governance in
  practice)

32
Complementarity in the use of indicators 1.
Input vs. outcome indicators - example

• Input indicator In law, is there an agency
  with a legal mandate to address corruption?
• Outcome indicators
• In practice, is the anti-corruption agency
  effective?
• Does the ACA make regular public reports (e.g. to
  legislature)?
• When necessary, is the ACA able to independently
  initiate investigations?
• b) Can citizens access the anti-corruption
  agency?
• Does the ACA act on complaints within a
  reasonable time period?
• Can citizens complain to the ACA without fear of
  recrimination?

33
What is the value of monitoring the input and
outcome sides of reform? 4 groups
34
Discrepancies in law and practice are revealing

• Are laws and organisational reforms translating
  into impact on the ground? If not, why not?
• Inappropriate reforms?
• Economically perverse incentives?
• Vested political interests?
• Lack of capacity (e.g. human resources,
  information retrieval systems)?
• High tolerance of corrupt conduct by the
  public?
• Different understandings of what is corrupt
  conduct?

Photo by S.P. Schuman/ State University New York
35
Part three Experience-based vs.
perception-based indicators
Photo by Konomiho/flickr.
36
Experiences vs. perceptions of corruption
Example of a fact-based / experience-based
question In some areas there is a problem of
corruption among public officials. During 2009,
has any government official, for instance a
customs officer, police officer or inspector in
your own country, asked you or expected you to
pay a bribe for his service? yes no dont know
37
Experiences vs. perceptions of corruption
Example of a perception-based question It is
known that in some countries the problem of
corruption among public officials is perceived to
be high by citizens. Imagine a person who needs
something that is entitled to him/her by law. Is
it likely or not likely that this person would
have to offer money, a present or a favour (e.g.
more than the official charge) to get help from
Member of Parliament? Elected municipal
councillors? Customs officials? Hospital
staff? Police officers ?
Tax/revenues officials?
Teachers/Professors? Courts officials?
Inspectors? Private sector?

38
Vote

• Raise your hand every time you agree with a
  statement
• Perceptions are better indicators of progress
  than facts
• Facts are better indicators of progress than
  perceptions
• Both perceptions and fact-based indicators are
  useful
• Neither is helpful, it is not possible to
  measure corruption

39
Module 7 Complementarity in the use of indicators

40
Experiences vs. perceptions of corruption

• Experiences of victimisation are only a small
  part of what affects peoples perceptions about
  corruption.
• Respondents more likely to perceive corruption
  tend to
• feel unsafe have been victims of other offences
  (e.g. theft, assault)
• live in neighbourhoods with low levels of
  solidarity and social cohesion
• Perception is not a proxy for actual levels of
  corruption. There are many social factors
  affecting perceptions, other than victimisation.

41

• How can perception data
• help or hinder
• corruption monitoring?

42
Drawbacks of perception data

• Seen as biased, leading some to minimize its use,
  or rule it out entirely
• Politically sensitive, especially in polarized
  political climates, where the media is seen to
  have undue influence in magnifying public
  perceptions of corruption
• Lag effect perceptions often lag behind reforms

43
But perception data matters!

• High perceptions of corruption are correlated
  with low perceptions of state legitimacy
• High perceptions can fuel corrupt practices by
• encouraging people to believe they must pay
  bribes
• reducing the likelihood of citizens reporting
• leading those with power to believe that there is
  nothing wrong with accepting bribes.
• Perceptions and opinion reveal important
  information about how corruption works in a
  specific context

44
Afghan Civil Servant Corruption Survey

• A) Perceptions of corruption in the civil
  service 
• Perceptions of the extent of corruption (now
  over time)
• Perceptions regarding whether exposing acts of
  corruption would hurt the civil servants career.
  
• If perceptions of corruption are high
• lowers the barriers to corruption (such as
  guilt, fear of punishment, reputational harms)
• increases corrupt practices

45
Perceptions of corruption in the civil service

• Imagine the total number of people in all 3
  branches of government in Afghanistan was 100 and
  you had to say how many of these 100 are corrupt.
  How many would you say?

46
Afghan Civil Servant Corruption Survey

• B) Conceptions of corruption in the civil
  service
• Do they align with definition of corruption in
  AC reforms? (e.g. can civil servants recognize
  acts of corruption from a given list?)
• Personal attitudes and beliefs may dominate
  legal definitions of corruption...
• If discrepancy between civil servant conceptions
  of corruption and the law
• promotes transgressions and will undermine
  reform efforts.

47
B) Conceptions of corruption in the civil service

• Referring to the public sector, which of the
  following acts do you consider to be acts of
  corruption
• Stealing money from government (embezzlement)
• Threatening individuals with harm if they do not
  give money (Extortion)
• Influencing government contracts in return for
  financial or other gains
• Bribery
• Promotion of unqualified individuals for
  political reasons or other gains
• Favouring relatives and friends e.g. appointing
  them to good positions (nepotism)

48
Afghan Civil Servant Corruption Survey

• C) Tolerance of corruption in the civil service
• Extent to which certain corrupt practices are
  considered acceptable within the civil servants
  values system or justified as a display of
  hospitality or gratitude for efforts of the civil
  servants.
• If tolerance of corruption is high
• Raises the probability that a civil servant will
  participate in corrupt acts
• Reduces the probability that corrupt acts will
  be exposed.

49
Tolerance of corruption in the civil service

• Please indicate how much do you agree with the
  following statements
• I would feel comfortable helping my family member
  get a job in my govt office, provided they were
  qualified.
• even if they were not as qualified as another
  candidate.
• I would feel obligated to use my influence as a
  civil servant to help my friend / relative with a
  problem if I could.
• Corruption by low-level employees is more
  acceptable than corruption by high-level
  officials.

50

• How would you customize this survey to your own
  country context?
• Which sections are most relevant?
• Which sections would need to be revised?

51
Part four Pro-poor and gender sensitive
indicators
Photo by Konomiho/flickr.
52
Poverty gender sensitive indicators

• Corruption exacts a higher price on the poor
  as an additional tax, and by lowering the
  quality of services
• Gendered forms and dimensions of corruption
  men more often tend to be in positions of power,
  and therefore to be the beneficiaries of
  corruption
• Need for assessments to be sensitive to the
  impact of corruption on marginalised groups, and
  to the effectiveness of anti-corruption
  mechanisms for these particular groups
• Indicators in this area too often tend to be
  gender and poverty blind

53
What makes indicators pro-poor or
gender sensitive?

• UNDP proposes 4 ways to make indicators sensitive
  to vulnerable groups
• Disaggregating by poverty/gender
• Specific to the poor/women
• Implicitly poverty/gender sensitive
• Chosen by the poor/women

54
Pro-poor and gender sensitive indicators

• Disaggregating by poverty/gender
• The proportion of poor households using public
  services who experienced corruption directly in
  the last 12 months in comparison to non-poor
  households. (Measuring experience of corruption)
  
• The proportion of lower to higher income citizens
  who could afford to bring a legal suit ( of
  lower income citizens may be measured as those
  earning less than the median yearly income)
  (Proxy indicator measuring resistance to
  corruption)

55
Pro-poor and gender sensitive indicators

• 2. Specific to the poor/women
• Size of funds allocated to legal aid in
  provincial budgets (per capita)
• Amount of money spent on servicing one
  beneficiary of legal aid (legal aid budget /
  number of beneficiaries of legal aid)
• Number of attorneys as of citizens in need of
  one
• Average duration of cases processed by the
  public defenders office compared to average
  duration of the type of case in question (as )

56
Pro-poor and gender sensitive indicators

• 3. Implicitly poverty/gender sensitive
• The percentage of small retail business that
  can afford to bring a legal suit.
• Backlog of small cases of little financial
  value
• The proportion of public agencies for which
  public expenditure tracking surveys (PETS) are
  regularly conducted.

57
Pro-poor and gender sensitive indicators

• 4. Chosen by the poor/women
• Level of satisfaction with public services
  expressed by women in poor households in
  comparison to men.
• Agencies perceived to be corrupt by low-income
  households in contrast to higher income
  households.
• Womens trust in the police and its ability to
  provide women with redress if they file a
  complaint.

58
Scenario Assessing the performance of local
complaint offices

• Decreasing number of public complaints filed
  each month
• Is it evidence that the national AC strategy is
  working in the province? Is it evidence that the
  incidence of corruption has gone down?
• Meanwhile, corruption cases are still making
  the news in the provincial papers...
• How to explain the such a contradiction?

59

• Through interviews with the staff of the office,
  their clients
• and a survey of the public, the following
  findings emerge
• The office is not easy to access for ethnic
  minorities living in the periphery of the
  province
• Fewer respondents from low-income areas know the
  process for filing reports, and many assume there
  is a fee for reporting cases

• Many respondents do not expect their cases
  would receive a fair investigation and have
  little faith in the institution
• Different opinions on what is considered
  corruption amongst staff the public

60
?
?
What indicators could be used to measure the
severity of each one of these challenges,
especially for vulnerable groups, and progress
in tackling these challenges?
?
?
?
?
?
?
?
?
?
?
61
Proposed framework of indicators for assessing
the performance of the office, especially in
servicing vulnerable populations
Possible challenges Possible pro-poor, gender sensitive indicators
Political will (Yearly) change in expression of public confidence that the provincial mechanism is willing and able to act upon complaints (disaggregated by income, gender, ethnicity, etc.)
Independence Change in public perception of independence of the provincial mechanism (e.g. do you think that members staffing the complaint mechanism are independent from political and private powers in your community?) (disaggregated by income, gender, ethnicity, etc.) Change in perception of equal treatment (disag.)
62
Possible challenges Possible pro-poor, gender sensitive indicators
Whistle-blower protection Change in the spectrum of small claims poor complainants file Change in ratio of perception of problems solved to problems exacerbated among litigants (disag.)
Accessibility citizens who have access to a phone connection (if complaint mechanism is a hotline) (disag.) citizens who believe that complaint boxes are placed in strategic and sufficient locations in the province (disag.) citizens who know the process for filing a complaint (disag.) responses to a request for information provided for a fee (e.g. photocopying fee)
63
Possible challenges Possible pro-poor, gender sensitive indicators
Complaints handling Is the staff representative of the local population (disaggregated by gender, ethnicity, etc.)? Change in time between filing and first hearing (disag.) Change in time from filing to disposition in cases of small financial value (disag.)
Enforcement Change in proportion of complainants who are satisfied with outcomes of filing a complaint (disag.) Change in proportion of cases that are finalized in 12 months (disag.) Change in rates of non-enforcement for poor/vulnerable vs. different groups of (wealthy or otherwise privileged) defendants
64
Part five Integrity indicators (to measure
the opposite of corruption, i.e. resistance to
corruption)
Photo by Konomiho/flickr.
65
Corruption indicators A quick recap

• Number of reported cases of bribery within the
  police
• Police ranked as second most corrupt institution
  in public national perception surveys
• Bribe payments as of income among individuals
  in low-income neighborhoods
• Existence of accessible and safe reporting
  mechanisms
• Proportion of sanctions to registered complaints

66
Corruption indicators can measure
The incidence of corrupt transactions
experience (victimisation) vs. perception data
The impact of corruption Data sensitive to
vulnerable groups gender The existence and
effectiveness of anti-corruption
mechanisms Integrity indicators
67
Integrity indicators

• Also called transparency or accountability
  indicators
• Focus on the effectiveness of anti-corruption
  mechanisms which are within the control of policy
  makers
• Therefore tend to be actionable
• How can we design such integrity indicators?
• Can they capture the particular experiences of
  women and vulnerable groups in using
  anti-corruption mechanisms?

68
Requests for new connections
Hiring touts/ middlemen to fill in formsBribes
to get filled forms depositedPayment to expedite
new connectionsBribes to secure a new
connectionBribes to speed up an
applicationBribes to ensure forms are accepted
Authorization of new connections
Offered extra payment to resolve problemAsked
for payment to resolve problemBribes to stop
reporting illegal connectionsPayments to
expedite repair workPayments to make illegal
connectionsUnofficial private repairsChoice of
contractors personal or political gain
Service delivery
Meter arrangement
Install meters/replacement of meters
Extort payment to reduce meter readingBribes to
reduce meter readingPayment to overlook an
illegal connectionFalse meter readings
Meter reading
Enter meter reading into billing system
Billing
Outputs billing/estimate billing for unmetered
accounts
Payment to reduce bill
Computerized billing
Account summaries prepared and billing prepared
for distribution
Step 1 corruption risk mapping
Payment
.
WEDC, Loughborough University
69
Designing integrity indicators
The method
1. Trace the steps in the delivery of a service
2. Identify corruption hot spots
3. Identify corresponding anti-corruption
mechanisms
4. Design indicators to match
5. Quantify indicators
70
Example procurement (Macedonia RTA Index)
Hot spot Publication of and invitation for
bids done in a way that does not inform as many
bidders as possible AC mechanism Public
procurements, particularly those of a larger
value, should be published in the daily newspaper
with the largest circulation, instead of in
newspapers with limited circulation Indicator
(quantitative) Number (as of total number of
procurements) and value (as of total value of
procurements) of procurementrs advertised in
papers of large circulation
71
Example of a qualitative indicators of integrity
in procurement Developing a scale to quantify
qualitative indicators
In practice, are major public procurements
effectively advertised? Score 100 There is a
formal process of advertising public
procurements. This may include a government
website, newspaper advertising, or other official
announcements. All major procurements are
advertised in this way. Sufficient time is
allowed for bidders to respond to
advertisements. Score 50 There is a formal
process of advertising but it is flawed. Some
major procurements may not be advertised, or the
advertising process may not be effective. The
time between advertisements and bidding may be
too short to allow full participation. Score 0
There is no formal process of advertising major
public procurements or the process is superficial
and ineffective.

72
How to quantify qualitative indicators? Example 2
Example In practice, are national level judges
protected from political interference? Scale
100 points, 50 points, 0 point (Source Global
Integrity Scorecard)
100 points - National level judges operate
independently of the political process, without
incentive or pressure to render favourable
judgments in politically sensitive cases. Judges
never comment on political debates. Individual
judgments are rarely praised or criticized by
political figures.
73
How to quantify?
50 points - National level judges are typically
independent, yet are sometimes influenced in
their judgments by negative or positive political
incentives. This may include favourable or
unfavourable treatment by the government or
public criticism. Some judges may be demoted or
relocated in retaliation for unfavourable
decisions. 0 point - National level judges are
commonly influenced by politics and personal
biases or incentives. This may include
conflicting family relationships, professional
partnerships, or other personal loyalties.
Negative incentives may include demotion, pay
cuts, relocation, threats or harassment.
74

Part six Corruption Assessment Tools

Photos by Adam Rogers/UNCDF
75
Existing instruments for assessing corruption
assessing AC mechanisms

Theres no point in reinventing the wheel!
Many existing corruption assessment
methodologies can provide templates for a
research focus, research design and even
indicator sources.
76
Country-level corruption and Anti-corruption
assessment tools
Transparency/Accountability/Integrity
Corruption
Diagnostic Assessments
Compliance monitoring
Perception
Experience/ victimisation
Public opinion
General population / vulnerable groups
Institutions
Processes
Experts
Sectors
Public sector
Public sector
Mapped by Transparency International
Local level
Private sector
77
www.gaportal.org Profile for each tool

• Purpose
• Type of data used
• Methodology
• Pro-poor / gender sensitive aspects
• Actionability
• Complementarity (de jure / de facto inputs /
  outcomes)
• Example indicators
• URL

78
www.gaportal.org
www.gaportal.org
79
Selecting an appropriate assessment tool

• Start with a clear objective
• Map out relevant tools, and select a
  methodology that is suited to your purpose
  and focus
• Dont just replicate a generic tool developed
  by an international organization, or a tool used
  in another country adapt it to your setting
  (method data sources, country-specific
  indicators, scales to quantify qualitative
  indicators)
• Make sure to use indicators that are actionable
  action-worthy indicators, and complementary
  (input vs. outcome de jure vs. de facto)
• Invest in proper design, testing and initial
  consultation of stakeholders