Avoiding the Morning After Blues: Building State Capability while Times

are Still Good Lant Pritchett HKS and CGD

India Policy Forum Lecture July 12, 2017

Outline of the Presentation

• Times are good, but will almost certainly get worse…and how bad the ‘worse’ will be depends on what India does now to build capability

• What is state capability for policy implementation (SCPI)? Clarification of concepts

• The types of capability needed • How not to get state capability—techniques of

successful failure • How you might get SCPI (what we propose): PDIA

(Problem Driven Iterative Adaptation)

Growth is good now…but is unlikely to last

Growth in developing countries is strongly episodic and has strong “regression to the

mean”

• Generating episodes of rapid growth in developing countries is not

actually rare…it is common • …but episodes of rapid growth all end...

– Sometimes in moderate slowing (e.g. Korea, Taiwan) – Sometimes to sudden stops (e.g. Brazil, Mexico in the 1980s) – Sometimes to extended downward spirals (e.g. Venezuela) – Sometimes to big crises and then bounce back (e.g. Indonesia) – Sometimes to extended crisis (e.g. Iran post Shah)

• The long-run prosperity depends on keeping the downturns short and shallow---but that depends on what happens when times are good

An overall international index of “state capability” (average of rule of law, bureaucratic quality, service delivery, government effectiveness and control of corruption) suggests India is above the developing

country average…but not getting better over recent decades

India is above average …but trend 1996-2012 is

negative

And you cannot run the “institutions and growth” logic in reverse…there is zero correlation of growth and improvement

in governance

Strong correlation of growth and income levels on levels

Zero correlation of growth on level or changes in quality of institutions or governance

While times are good India needs to

act to build the capabilities to

• Keep the good times rolling…as sustained growth requires increasingly sophisticated capabilities

• Cope with the slow downs that will come

• Prevent a slowdown from becoming an extended growth trap or catastrophe

What is state capability for policy implementation and what does it

matter?

100

90

90

90

60

43

30

9.2

21.2

0

0 10 20 30 40 50 60 70 80 90 100

Czech Republic

Finland

Uruguay

Colombia

Top quartile by income

Second quartile by income

Third quartile by income

Lowest quartile

Bottom half of countries by…

Lowest 25 countries

Percent of 10 misaddressed letters coming back to USA within 90 days (all countries agree to

return within 30 days)

Includes not just Somalia and Myanmar but Tanzania, Ghana, Nigeria, Egypt, Russia, Mongolia, Cambodia, Honduras, Fiji, etc.

Source: Chong, A., La Porta, R., Lopez‐de‐Silanes, F. and Shleifer, A. 2014. Letter grading government efficiency. Journal of the European Economic Association 12(2), pp. 277-299

Does Policy matter for outcomes? All of these countries have a “policy” that all children should be

able to read by the time they finish grade 6—yet outcomes vary from essentially zero to 100

Source: Pritchett and Sandefur 2017, ‘Girl’s Schooling and Women’s Literacy: Schooling Targets Alone won’t reach literacy goals.”

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

0 1 2 3 4 5 6

Percent of women 25-34

who can read a single sentence

(latest DHS data)

Rwanda

Indonesia

Average 51countries

India

Bangladesh

Nigeria

What is state capability for policy implementation?

State Capability for Policy Implementation

• The ability of organizations to induce their agents to undertake the fact-contingent actions that further the public purpose objectives of the organization

– Resources

– Technical Capacity

– Motivation of agents to do what they know

In practice the worst and best medical care in rural Madhya Pradesh came

from the same people

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0

5

10

15

20

25

30

Private Sector (ALL) Public Sector (ALL) Public Sector Doctors in theirPublic Clinics

Public Sector Doctors in theirPrivate Clinics

Stan

dar

diz

ed C

hec

klis

t Sc

ore

(St

and

ard

Dev

aiat

ion

s)

%C

hec

klis

t C

om

ple

tio

n

The same provider has the lowest checklist adherence in their public sector clinic....and the highest in their private sector clinic

% Checklist Completion Standardized Checklist Score

Difference in performance for the same provider in their public and private clinics

Source: Das, Jishnu, Alaka Holla, Veena Das, Manoj Mohanan, Diana Tabak, and Brian Chan. "Standardized Patient Study Reveals Severe Quality Deficits in Rural and Urban India." Health Affairs, Forthcoming.

Two conceptually distinct steps in policy implementation

• Determination of the “administrative facts”

• Acting in accordance with those facts

Truth-telling was low under traditional third-party auditing system status quo

Source: Duflo, Esther, Michael Greenstone, Rohini Pande, and Nicholas Ryan. 2013. “Truth-Telling by Third-party Auditors and the Response of Polluting Firms: Experimental Evidence from India.” Quarterly Journal of Economics 128 (4): 1-48.

How not to get state capability: Techniques of successful failure

Isomorphic Mimicry is gaining survival value by looking like something else

(Remember: Red and black, friend of Jack, Red and Yellow, Kill a Fellow)

Ecosystem for organizations

Organization

Agents

How Open is the System?

How is Novelty Evaluated?

Strategies for Organizational Legitimation

within the Ecosystem

Leadership Strategies

Front-line Worker

Strategies

Open Closed

Enhanced Functionality

Agenda Conformity

Isomorphic Mimicry

Demonstrated Success

Value Creation

Performance Oriented Self-interest

Organizational Perpetuation

How do you destroy organizational capability? Premature load bearing

Import tax collection agency

Available rewards to Non-compliance for individual agents

Rent collectors

Existing tax code

Maximally Feasible tax code

Trainin

g

No amount of training is going to increase organizational capability in practice when the stress of existing policy implementation already far exceeds capacity

With heavy regulation and weak capability the “rule” and the actual outcome diverge

A typical firm says “30 days” to get a permit in countries where the Doing Business says 50 and 500—and everything in between

Firms in the same country report widely different times Fastest 10 percent (circles) Slowest 10 percent (triangles) Slovenia: Fast say 14 days, slow say 365 days

When regulations exceed enforcement already, even large reductions in “official” regulations leave typical behavior unchanged (or more)

Compliance times often increased even though “reform” reduced required time

Suppose you were designing a physical training program to build athletic

capability

• Agility

• Quickness

• Leaping

• Stamina

•Power •Strength •Maximum one lift peaks

Four analytical questions about policy implementation for your activity that you wrote down

Is your activity…

Does producing successful outcomes from your policy….

Transaction intensive?

Require many agents to act or few

Locally discretionary?

Require that the implementing agents make finely based distinctions about the “state of the world”? Are these distinctions difficult for a third party to assess?

Service or imposition of obligation

Do the people in direct contact with your agents want or not want the agent to succeed?

Based on a known technology

Is there an accepted handbook or body of knowledge for doing what you are trying to do or will this require innovation (not just context)

Typology of tasks by capability intensity needed for implementation

Health Finance

Policy making Or elite services

Iodization of Salt/Tertiary care

Monetary policy

Logistics

Vaccinations Payment systems

Implementation intensive service delivery

Curative care Loans

Implementation intensive imposition of obligations

Regulation of private providers

Regulation of private providers

Wicked hard

Promotive health

Equity financing of start-ups

or

This is an activity oriented classification not a “sector” or even overall task

Primary School

Transaction Intensive

Locally Discretionary

High Stakes

Known Technology

Type?

Setting Standards

Building Schools

Delivering Inputs

Classroom teaching

Student Assessment

21st Century is about state capability implementation intensive challenges

Health Finance

Policy making

Iodization of Salt

Monetary policy

Logistics

Vaccinations Payment systems

Implementation intensive service delivery

Curative care Loans

Implementation intensive imposition of obligations

Regulation of private providers

Regulation of private providers

Wicked hard

Preventative health

Equity financing of start-ups

or

Response to Low Capability: Make it logistics

Get rid of “discretionary” part of driver’s license? Doctors to give patients all the same advice? Teachers to follow a day to day script?

Procrustes and his bed

Once systems decide that logistics of input compliance is all they do….

Enrolments Inputs

Enrolment

in

government

schools

Enrolled in private

Percent

in

government

Percenta

ge with

drinking water

Percentage

with girl's toilet

Pupil teacher ratio

2004/05 5,487,221 4,297,171 56.1 79.8 25.4 55

2011/12 4,226,225 5,229,293 44.7 100.0 75.3 29

Gain/loss -1,260,996 932,122 -11.4 20.2 49.9 -26

A million children abandon public schools, learning is getting worse and worse and the Ministry declares success—with data.

India needs three capabilities for the future

• Capabilities in “implementation intensive service delivery”—e.g. education, health, policing, livelihoods, sanitation and these cannot be reduced to logistics

• Capabilities in “implementation intensive imposition of obligations”—e.g. environmental regulation, collection of taxation, dispute resolution, land titling/land use restrictions

• Capabilities in the “wicked hard” space of sustaining dynamic innovative growth—e.g. skills formation, job matching, responsive provision of needed government services for nascent and emerging industries

Let me be clear…ICT will not, cannot, be the key to the capabilities India

needs

• Moore’s law and the vast improvement (10 orders of magnitude in computing power) have made ‘Moore’s law amenable’ products and services massively cheaper and better.

• But much of governmental capability is not Moore’s law amenable

• “Transparency” is a limited notion…the world is not “transparent”

Problems cannot be solved by the same level of thinking that created them.

Albert Einstein

The feasible path has to start from where you are….

“You cannot cross a chasm in two jumps”

…because of the first fails you are at the bottom, with broken bones

Experiment with improving police in Rajasthan

Doesn’t matter what results are if they cannot be implemented

The “treatment” (of giving policemen days off) just wasn’t done

• …because these reforms were executed within the framework of existing conditions, incentives for the implementation of reforms was not incorporated into the design. … Therefore, while the senior police leadership consistently supported the reforms and gave orders for their implementation, the long term, system-wide benefits perceived by the leadership were not internalized by the police station chiefs. As a result, police station staff gradually ceased to carry out the program elements, perhaps even going so far as to falsify the community observer records, and the project stopped functioning over 18 months.

0

5

10

15

20

25

30

35

Control

Days sincelast day off

Percentwith dayoff in last 8days

Source: “Can Institutions be Reformed from Within? Evidence from a Randomized Experiment with the Rajasthan Police,” (with Raghabendra Chattopadhyay, Esther Duflo, Daniel Keniston and Nina Singh), NBER Working Paper No. 17912, March 2012.

Trying to change the facts about attendance merely created an administrative fiction

0

10

20

30

40

50

60

Present (full day) Absent or casual leave Exempt

Feb. 2006

Apr-07

“Exemptions” went up—fiction

replaced fact

During the course of the field experiment to motivate nurses to attend their clinics in Rajasthan…

“Absence” went down

But “presence” went down too

Source: “Putting a Band-Aid on a Corpse: Incentives for Nurses in the Indian Public Health Care System” (with Esther Duflo and Rachel Glennerster), Journal of the European Economic Association 6(2-3), pp. 487-500, April 2008.

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

Doctor Attendance Nurse, labtechnician,Pharmacist

Treatment Effect on Attendance

Impact of typical introduction of ICT “solution” to low attendance

-0.2

-0.15

-0.1

-0.05

0

Staff availability Staff Quality Knowledge ofentitlements

Treatment impact on patient perceptions

-0.3

-0.2

-0.1

0

0.1

0.2

PHC Large Hospital (Public orPrivate)

Percentage change in place of delivery

0

0.1

0.2

0.3

Birthweight Reduction in low birthweight

Percentage improvement in birth outcomes

PDIA (Problem Driven Iterative Adaptation) as an approach to

building capability of state organizations while producing

results

Four Principles of PDIA (Problem-Driven Iterative Adaptation)

1. Local Solutions for Local Problems

2. Authorizing Problem Driven Positive Deviance

3. Try, Learn, Iterate, Adapt

4. Scale Learning through Diffusion

This section is based on Andrews, Pritchett and Woolcock 2013 (forthcoming)

Local Solutions for Local Problems

o Agenda for action focused on a locally nominated (through some process) concrete problem

o Not “solution” driven that defines the problem as the lack of a particular input (e.g. “teacher qualifications”) or process (e.g. “EMIS”)

o Rigorous about measurable goals in the output/outcome space (e.g. cleaner streets, numbers of new exports, growth of exports)—can we know if the problem is being solved?

• Good problems

Examples of “problem driven”

Compliance Driven

• Enforce existing regulations on workplace safety

• Hire teachers with required qualifications

• Comply with procurement regulations

Problem Driven

• Reduce fatalities/accidents at work places

• Attract and retain teachers who help student progress

• Buy things effectively

Pushing Problem Driven Positive Deviation

o Authorize some agents (not all) to move from process to

flexible and autonomous control to seek better results

o An “autonomy” for “performance accountability” swap (versus “process accountability”)

o Only works if the authorization is problem driven and measured and measurable… increase the ratio of “gale of creative destruction” to “idiot wind”

Authorizing positive deviation

o Allow flexibility in methods against specified and agreed to problems

o “Fence breaking” activities that allow deviations from process controls for designated activities

o Rapid feedback loops to search over design space

Iteration and adaptation

• Feedback loops on performance that allow practices to change (rather than stop gap addressing individual cases)

• Use evidence in management time (not ex post impact evaluation)

• Have sequenced steps: “what did you do?” “what happened?” “what did you learn?” “what will you do next?”

Scaling through diffusion

• Since the basic problem with dysfunctional organizations is a collapse of internalized norms of performance…this has to be reversed

You cannot beat a turtle into moving

The head has to come out for the body to move

Organizations can survive external attack…by not moving

Making Hay while the Sunshines

• India is in the middle of a historically transformative episode of rapid growth.

• India needs to build the capabilities for the future while times are good.

• But…this (was not) happening…and while logistics can be made much better with ICT that is a very limited set of what governments do

• Capability for implementation intensive tasks requires a different method—particularly when starting from low performing organizations.