How the Provincial Information & Analysis Unit (PIAU) pitched for the Performance Management System mapping KPIs of over 300,000 government officials across Punjab, project was deployed in 2010.
2. Problem Statement
Service Delivery Gap
1947 1995 2009
Service Delivery
Expectations
Actual Service
Delivery
Source: WDI, World Bank and Structured Interviews with Ministers, Secretaries and Field Officers2
3. Service
Delivery Gap
Resource Gap
Federal Govt.
Operational
Efficiency Gap
Capacity
Willingness to
Work
Institutional
Design
Policy
Framework
Institutional
Framework
Causes of Service Delivery Deficit
Source: Basic Theory of Informational Economics, Personnel
Economics: Past Lessons and Future Directions, Edward P Lazear
3
4. Which one is the Binding Constraint?
ResourceShortage
HRCapacity
Willingnessto
Work
InstitutionalDesign
Service-delivery
deficit
Binding
constraint?
If constraint is
non-binding,
improving it
won’t decrease
deficit
Maximumservice-deliverycapacityin
currentenvironment
4
5. So is Capacity a Binding Constraint?
• Government sector teachers are paid 5 times
more than private sector teachers
Source: Leaps Project
5
6. • Government sector teachers are more educated
and better trained than private sector teachers
Source: Leaps Project
6
7. • Yet children in private schools are 1.5-2.5
years ahead of children in government schools
Source: Leaps Project
7
8. • If we look at students that switch schooling
regime (private to public or vice versa)
• The same child performs worse when in a
government school than when in a private school
Source: Leaps Project
8
9. Service
Delivery Gap
Resource Gap
Federal Govt.
Operational
Efficiency Gap
Capacity
Willingness to
Work
Institutional
Design
Policy
Framework
Institutional
Framework
Causes of Service Delivery Deficit
Source: Basic Theory of Informational Economics, Personnel
Economics: Past Lessons and Future Directions, Edward P Lazear
9
10. Probability of being caught
(increases with information)
𝑃 F = 𝑝𝜏 ∙ 𝑝 𝛼
Probability of being punished
(decreases due to collusion)
Utility Function of a Public Sector Manager
𝑈 M = 𝑤𝑎𝑔𝑒 − 𝑒𝑓𝑓𝑜𝑟𝑡 + 𝑎𝑙𝑡𝑟𝑢𝑖𝑠𝑚 + 𝑝𝑒𝑟𝑠𝑜𝑛𝑎𝑙 𝑔𝑎𝑖𝑛
𝑤𝑎𝑔𝑒 = 𝑤𝑠 + 𝑤𝑝
salary perks &
privileges
𝑒𝑓𝑓𝑜𝑟𝑡 = 𝑒 𝑞 𝑝
‘units’ of public goods delivered
𝑝𝑒𝑟𝑠𝑜𝑛𝑎𝑙 𝑔𝑎𝑖𝑛 = 𝐸(𝜋)
𝑄
𝑞 𝑝
sum of units of public goods converted into private goods
Q = total units in control
expected return on
each unit
𝑃 S = 1 − 𝑃(F)
Probability of success
Probability of failure
Expected Return of Allocative Distortion
Reward of succeeding
𝐸 𝜋 = 𝛽 ∙ 𝑃 S − δ ∙ P F
Cost of failure
(via allocative distortion)
10
11. Service
Delivery Gap
Resource Gap
Federal Govt.
Operational
Efficiency Gap
Capacity
Willingness to
Work
Institutional
Design
Policy
Framework
Institutional
Framework
Service
Delivery Gap
Resource Gap
Federal Govt.
Operational
Efficiency Gap
Capacity
Willingness to
Work
Collusion
Information
Failure
Institutional
Design
Policy
Framework
Institutional
Framework
Causes of Service-delivery Deficit
11
12. 𝑈 M = 𝑤𝑠 + 𝑤𝑝 − 𝑒 𝑞 𝑝 + 𝜇
+ 𝛽 ∙ (1 − 𝑝𝜏 𝑝 𝛼) − δ ∙ 𝑝𝜏 𝑝 𝛼
𝑄
𝑞 𝑝
Utility Function of a Public Sector Manager
Allocative Distortion
Can be increased with improved monitoring
12
13. Where We See Allocative Distortion
• Irrigation & Power
– Unequal distribution of water
• Education
– Teacher absence
• Health
– Non-provision of medicines
– Medical staff absence
17
14. How to Differentiate Employee
Performance in Presence of Other Gaps
Average
performance
Theoretical limits of service delivery outcomes
Performance spectrum of employees working in similar environment
Compare top-
performing and
lowest-performing
employees with
reference to
average
performance
18
15. WHERE will allocative
distortion show up?
1. Amount of unaccounted water
2. Dry tails
3. Reporting discrepancies
4. Missing gauges
WHOis doing it?
1. Superintendent Engineer (SE)
2. Divisional Canal Officer (XEN)
3. Sub-divisional Canal Officer (SDO)
WHAT are
departments doing?
HOW are they doing it?
1. Making Changes in Water Distribution
2. Ensuring Channel Safety Against Damage
3. Regulation of Irrigation Channels
4. Monitoring of Outlet Performance
Ensuring Equitable
Water Distribution
How to Define the Best Indicators?Example Irrigation
19
16. Developing an Evaluation System
Work
Breakdown into
Activities
Process
Mapping
Responsibility
Assignment
Key
Performance
Indicators (KPIs)
System
Development
Secretary’s
Dashboard
20
18. Current Progress
• WBS nodes: 7,312
• Process maps: 3,427
• Process steps: 34,448
• Managers identified: 2,520
• KPIs developed for 6 departments
– Health
– Higher Education
– Schools Education
– Irrigation & Power
– Board of Revenue
– Home (Punjab Prisons)
• Data collection started for
– Higher Education
– Irrigation & Power
22
19. Way Forward?
Role of Key Performance Indicators in
Wider Punjab Good Governance
Initiative
23
20. Appendices
• KPI progress
• Other interventions
– Citizen Feedback Model
– Litigation Monitoring System
– CM Directives
24
21. Citizen Feedback Model
• Aim: Decrease Petty corruption
• Model: Gather feedback through phone calls
from the citizens who received a service from
the government office
• To be implemented in 7 districts for 5 services.
– Registration of Property
– Provision of Free Medicine in Emergency
– Surgical Procedures
– Issuance of MLC
– Pension cases
25
22. Proposed Interventions
• Service provider is made bound to provide Cell# of the citizen
availing the service
• Manager/ Senior Office of the Provincial government calls the
citizens in a systematic fashion to know about their experiences.
• Over time the telephonic feedback will generate patterns of service
provider’s behavior.
• Such patterns once analyzed will throw up worst service delivery
nodes.
• It is believed that once the worst service delivery nodes are
identified, the service managers can be sorted out through
administrative actions like transfers and/ or targeted investigations.
• The intervention explained so far captures non-collusive corruption
only. Therefore, supplementary intervention was introduced.
26
23. Role of PIAU
Step1: Development of Paper Based Data Capture at Service
Delivery Node
Step2: Development of Online Data Entry System
Step3: Development of Feed back Capturing Forms
• DCO: Forms will be generated to enter the feedback
• Call Center: A representative random sample of transactions are
selected for gathering feedback.
Step 4: Development of Online Data Entry system of the
Collected Feedback
Step 5: Development of System to detect Collusive Corruption
Step6: Analysis of Patterns (to be put up to concerned
departments, CS and CM)
27
24. Intervention Cycle for Services
Development of
Data Entry
Forms at
Service Delivery
Node
System for
Online Data
Entry System
Development
of Feed back
Call Forms
System for
Online Data
Entry of the
Collected
feedback
Development of
System to
Detect Collusive
Corruption & In-
efficiency
Revenue
Registration of
Property
Completed Completed Completed Completed Completed
Health Services
Provision of Free
Medicine in
Emergency
Completed Completed Completed Completed
Minor Surgical
Procedures
Completed Completed Completed Completed
Issuance of MLC Completed Completed Completed Completed
School Education
Pension & Leave
Cases
In Progress In Progress
28
28. Litigation Monitoring System
• Current System
– Court summons are forwarded to relevant departments.
– Summons sometimes incorrectly sent to wrong department
• Need Gap
– The Law department is unable to effectively monitor litigation activity because of
delayed intimation/responses from relevant departments
– Delayed response to court summons (due to lack of coordination) results in
adverse judgments and expiration of appeal windows
• Solution
– Litigation Monitoring System - track status of cases to help streamline the
monitoring process for the Law and concerned departments
– Value Addition
• Availability of quick snapshots will help Secretaries identify gaps, timely escalation of
overdue cases will improve response time and assigning of responsibility will become
easier in case of lapses
• Status
– Development work for the initial release of the software is underway
32
29. Entering Cases
For priority cases
Has case been sanctioned
by Law department?
May be different from
concerned
department
Post-admittance,
evidence, arguments, etc.
Once the case is entered, it will appear on the Solicitor/AG’s dashboard and a law counsel will be assigned to the case
33
30. Updating Cases
Select existing case from
database
Y/N (In case date was
postponed without hearing)
Did gazetted officer represent
the department?
Immediately flags to
Solicitor/AG office
Option to attach scanned
documents
Person representing
department
•When next hearing date passes and case is not updated, case is flagged automatically
•The case proceedings will be updated by law counsel. The Litigation officer will be able to view and comment on updates
34
31. CM Directives Monitoring System
• Current System
– Computerized Diary number issued for CM Directives but not used to
track implementation progress of Directives
• Need Gap
– Increase in volume of directives has rendered it difficult to track
progress on directives
– No pre-emptive information on delays, response always reactionary
• Solution
– Online CM Directives Monitoring System - tracks the entire activity
that takes place on a particular Directive once it is issued
– Value Addition
• Shorten the response time
• Improve coordination between departments
• Ensure relevant departments receive directive
• Status
– Preliminary software version under testing
35
35. Health Expenditure Per Capita
0
5
10
15
20
25
30
35
2002 2003 2004 2005 2006
HealthExpenditurepercapita(inUSD)
Bangladesh India Pakistan
Source: World Development
Indicators, World Bank 39
36. The Problem of Collusion
• Collusion is an exogenous variable
– Society
– Patronage-based politics
• Organizations are a subset of society
– Behavior dictated by society instead of vice-versa
– Minimum level of collusion given in society
– Change requires external intervention
• Pakistan
– High societal dependence on beradri
– High nepotism
– Social network guarantees insurance against shocks
– Khwaja, Mian (2005) study of loans data on 90,000 firms
from 1996-2002 find politically-linked firms borrow 45
percent more and have 50% higher default rates
40
37. Political-Administrative Interaction
• Government operating at two levels – De facto
and De jure
– De jure
• National leadership, CM, CS, Administrative Secretaries,
Commissioners, DCOs
– De facto
• Service managers – SHOs, Patwaris, Headmasters, Bulldozer
operators, etc
• Each tier faces different incentives
– Policy for improvement of one tier will not necessarily
affect the other if they have competing incentives
41
39. Incentives of Politicians
• National party leadership desires national/ provincial
seat
– Needs popular support for success
– Bargains for as many votes as possible
– Two types of votes available
• Patronage-based votes via local politicians
• Party-based votes via public popularity
– Negotiates with local politicians for patronage-based
votes using party-ticket
• Stronger party-based votes, more bargaining power
• Party-based votes can only be increased by provision of public
goods
43
40. Incentives of Politicians
• Local leader desires role in government
– Bargains using patronage-based vote bank that
can be used with any party
– Desires to get seat without making extra effort to
gain more votes (by providing public goods)
– Only do as much work as is required to maintain
patronage-based vote bank
• Provide public goods only to patrons (i.e. convert them
into private goods)
– Land tenure security, coercive force (police), health services,
etc
44
41. Incentives of Politicians
Maintenance of patronage-based
vote bank
• Patronage-based distribution of
public goods
• Long-term capture of position in
government (without permanent
alignment to political party)
Maximization of party-based
votes (i.e. gain support of
popular masses)
• Equitable distribution of public
goods
• Long-term sustainability of
party appeal
Locally elected
representatives
MNAs
MPAs
DE FACTO
(LOCAL)
POLITICIANS
DE JURE
(NATIONAL)
POLITICIANS
PM
CM
Party leaders
45
42. Interaction of Politicians with
Administration
• Incentives create a market for exchanging
political favours for public goods
– Administrator uses political favours to lower
probability of being held accountable
– Politician uses public goods as private goods for
community leaders of patronage-based voters
46
43. Trade of Public Goods for Political
Protection
Local
leader
s
National/
Provincial
leadership
Patron of local leader Political protection
Public goods traded as private goods
Service
Delivery
tier
Senior
public
managers
47
44. Institutional Design
• 18th Amendment
• Medium Term Development Framework
• Valuation of rural land
• Review of PLGO
• LRMIS
• Punjab Health Sector Reform Programme
• Punjab Education Sector Reform Programme
• Punjab Resource Management Programme
48
45. Learning and Educational Achievements
in Punjab Schools (LEAPS)
• Rigorous academic study carried out by Tahir Andrabi
(Pomona), Jishnu Das (World Bank), Asim Khwaja (Harvard),
Tara Vishwanath (World Bank) and Tristan Zajonc (Harvard)
• 112 villages (having at least 1 private school) studied for 4
years (2003 to 2007) in districts
– Attock
– Faislabad
– Rahim Yar Khan
• Sample
– 823 schools
– 5,000 teachers
– 12,000 children tested in Urdu, Math and English
– 2,000 households
49
46. Information Failure in Observing
Employee Performance
UNOBSERVABLE
PORTION
OBSERVABLE
PORTION
OBSERVED
PORTION
COMPLETE
INFORMATION
SET (only known
to employee
him/herself)
50
47. Empirical Support of Low Service-delivery
as a result of Information Failure
• Health service in Delhi, India
– Public-sector doctors are more competent, but exert
less effort
• Civil works (roads) in Indonesia
– Missing funds decreased from 27.7% to 19.2% by
increasing to monitoring.
• Driving licensing in India
– 2/3rd people obtaining licenses are not qualified to
drive
51
48. The Dire Straits of Medical Practice in
Delhi, India – Das and Hammer (2005)
• 205 doctors in Delhi assessed by interviewers acting as
patients
• Same doctors observed a month later in clinical
practice
• Observation
– Public sector doctors more competent than private sector
ones, but exert lesser effort than their counterparts
• Policy Insight
– To improve medical services for the poor greater emphasis
needs to be laid on changing the incentives of public
providers rather than increasing competence via training
52
49. Monitoring Corruption: Evidence from a
Field Experiment in Indonesia
• To study the impact of increasing top down monitoring on
corruption.
• Setting:
– 608 villages in Indonesia were selected where roads were to be built. Out
of these, some villages were selected randomly and told that their funds
usage will be audited by central authority. The missing funds ( reported
usage- actual usage) were taken as an indicator of corruption. The change
in percentage of missing funds before and after the intervention was
studied.
• Finding:
– The missing funds decreased from 27.7% to 19.2% .
• Policy Insight:
– The knowledge that a tighter level of monitoring is being exercised by the
superiors reduces corruption. However, this must be complimented with
higher chances of being punished.
53
50. Corruption in Obtaining Driving Licenses in India
(Bertrand, Djankov, Hanna and Mullainathan, 2006)
• Study undertaken in India to observe the process
of obtaining driving licenses by following 822
individuals through the application process
• Observations
– 1/3rd of those that obtained a License did not take the
driving test; all on average paid 2.5 times more that the
requisite fee and 2/3rd were unqualified to drive
– Extra legal payments were unavoidable and not paid
directly to corrupt bureaucrats, but to “agents”
– Process fails to implement the social goal it was
implemented for 54
51. Information Services
• KPI should not be looked into in isolation
• It will help Business Process Improvement for all
depts.
• Enterprise Service Bus:
– Get data from 36 Districts & 38 Departments
– Give services to public/employees
• Resultantly, the existing systems of departments
improve, e.g. Health:
– DHIS + PHSRP MIS
– Clerks bring Flash Drives from Districts to Lahore
– Amalgam of 5 systems with duplications
52. 56
CIVIL SECRETARIAT
LAHORE
CM SECRETARIAT
ITD, AIWAN-E-IQBAL
COMPLEX
MOST ISLAMABAD
NTC RACE COURSE
LAHORE
Comm. Tower
Comm. Tower
Comm. Tower
W
ireless
Link
W
ireless
Link
Satellite dish
Digital Cross Connect
(Dxx) (1 Mbps)
Digital Cross Connect
(Dxx)(1 Mbps)
Digital Cross Connect
(Dxx)(128Kbps)
Digital Cross Connect
(Dxx)(1 Mbps)
BRI ISDN
BRI ISDN
INTER - LAN CONNECTIVITY (THREE SITES)
These Links will be used for internet Access, inter
departmental communication, Email, Messaging
53. Linkages
• GATEWAY Portal: $3 million with Microsoft
– Dashboard for CM is only front-end
– What is the right information for the indicators? Where
is it going to come from?
– PITB has requested PIAU for Enterprise Service Bus
• The Urban Unit
– Established wireless networking* with PIAU for their GIS
data needs (e.g., Health data for PHSRP)
– CS + Secretaries would have GIS Decision Support
System
* All Wireless Network Equipment has been procured & installed but not yet operational since TUU’s Vendor defaulted
55. HR-MIS / PIFRA
• HR-MIS for ALL Depts. (ITD & PIFRA)
– PIFRA data currently being obtained in real-time
from AG office and ported to ITD system
– Transfer/Enquiry/Promotion orders for employees
NOT to be issued until first updated in system
– Pilot: Higher Education & Sialkot CDG
59
56. ... Linkages: Result Based Mgmt. (RBM)
• RBM is a public sector management philosophy
and approach that focuses on achievement of
goals and objectives for improved result delivery:
• Strategic planning
• Systematic implementation
• Effective resource usage
• Performance monitoring, measurement & reporting
• Evaluation to improve result delivery
– Next tranche of $150 million from PRMP
– Being dove-tailed formally with PIAU
•Thomas, Kosha, Malaysia: “Integrated Results Based Management,” 2008 Community of Practice on Managing for Development
Results,Annual Publication
•The World Bank/IBRD, “Designing and Building a Results-Based Monitoring and Evaluation System, A Tool for Public Sector
Management,” A Workshop for Government Officials and Their Development Partners, 2000
57. Case Study: Mohatir’s ICU
• New Remuneration System 1993
• Public Sector Agencies given freedom on
policies/procedures, akin to businesses
• Five Central Agencies for reform:
– Implementation Coordination Unit (ICU JPM) within
PM’s Office
– Malaysian Administrative Modernization And
Management Planning Unit (MAMPU), Economic
Planning Unit(ECU), Public Sector Department (PSD)
• Annual Contract System with Employees
– Goals set at beginning of year
– Remuneration dependent upon performance/goals
• Program Based Budgeting
Notas del editor
RBM slides
Riots, law and order, transparency international. Wapda installations attacked?
Multiple constraints on performance, but not all constraints are equally binding. Most binding constraint has to be removed first.
Principal-agent problem, hidden information, hidden action
Job market slide
Principal cannot observe all information, action. Employee full info set- theoretically only known to him. Government wants to increase information. ACRs are not best record of performance. ACR: is a measure of the time spent with the person. Correlation direct. Nicely the time spent, not looking at work he did. Not necessarily related to work he does. Optimal contract. High risk, high pay job – weaker monitoring, public service requires greater monitoring. Contract mode not the best mode. Organizational memory, expertise, unique skills.
Political patronage -> if caught
Information economics -> build step-wise
Graph of each evidence paper
No of distorted slots.
Unaccounted water – cusecs
Department wise – wbs on this many pages
Process maps – pages
RAM – pages, # of managers identified, activities identified.
Work done so far / work to be done
Citizen feedback model
RBM
How collusion operates in society. Political + social. Nepotism data. Developed societies, insurance cover given by state – fair trial, disease and sickness. Our society: use support network to get fair trial, ensure services. Asim Khwaja, firms in KSE