mung“. awe ‘

s. ,
f

,2 .2493 2. .

1. K
3:1"
.1:

.Lnflfiu .
x: um“ r:

 

 

 

 

 

 

THE'fng

. LEBRARY
Michigan State

5 > University i

 

 

This is to certify that the

dissertation entitled

CHANGE AGENTS AND POLICY ENTREPRENEURS
AT THE LOCAL LEVEL

presented by
Gregory Allen Cline
has been accepted towards fulfillment

of the requirements for

PhD degree in Political Science

warm 5 LCM secek‘

 

Major professor

Date Aprll 20, 2001

MSU is an Affirmative Anion/Equal Opportunily lnxlimlion 042771

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDuep65p. 1 5

CHANGE AGENTS AND POLICY ENTREPRENEURS AT THE LOCAL LEVEL
By

Gregory Allen Cline

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Political Science

2001

UMI Number: 3009096

Copyright 2001 by
Cline, Gregory Allen

All rights reserved.

®

UMI

UMI Microform 3009096
Copyright 2001 by Bell & Howell Information and Learning Company.

All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.

 

 

Bell & Howell Information and Learning Company
300 North Zeeb Road
PO. Box 1346
Ann Arbor, MI 48106-1346

 

ABSTRACT
CHANGE AGENTS AND POLICY ENTREPRENEURS AT THE LOCAL LEVEL
By

Gregory Allen Cline

This research study is of four county-level efforts (three funded by an external
foundation change agent, the fourth indigenous) to develop locally designed innovations
to extend health care coverage to the uninsured. This research posits, tests and supports a
relationship where external change agents will stimulate (and indeed rely on) the
appearance of policy entrepreneurs when funding innovative change efforts at the local
level. The findings, supported by the development of two formal decision models, show
that external change agents can play an important role in encouraging policy innovation at
the local level, that the pre-existing local market context is an important predictor of the
success of innovative efforts, and that the appearance of policy entrepreneurs to lead
externally-funded innovations increases the probability of the successful development and
launch of desired innovations. Another key finding is that collective entrepreneurship,
posited to occur when innovation is desired in a technically complex issue area, is key to
the successful development and launch of innovations at the local level. Collective
entrepreneur teams are groups of several persons who possess the differing components
of a policy entrepreneur skills and assets, and appear in issue areas where complexity is
so high that it is unlikely that one person possesses all of the necessary skills and assets to

single-handedly act as a policy entrepreneur.

Perhaps the most useful finding of this research is that earlier theories of policy
entrepreneurship are tOO simplistic, and fail to capture, define and explain the complexity
that policy entrepreneurs and their collective teams face when pursuing innovation. By
building on existing theories of policy subsystems and theories of punctuated policy
change, this study posits and supports the idea that innovative policy changes will
inevitably have to overcome a policy subsystem in order to develop and launch a desired
innovation. The power of policy subsystems was found to differ across communities at
the same point in time, thus leading to the conclusion that innovations that can succeed in
some communities may not succeed in others at the same point in time.

Two formal decision models of innovative change at the local level were
developed and assessed; one of externally initiated change, the other of internally initiated
change. In both models, it was found, and supported by the qualitative case studies, that
two keys to the decision to pursue change were the ability to accurately assess the
readiness of the community to support change, and the innovator‘s (external or
indigenous) valuation Of the best and worst outcomes. For a policy entrepreneur, it was
also found that this person’s ability to self-assess his or her own abilities, and match them

to the innovative challenge, were additional components of the decision to pursue change.

Copyright by
Gregory Allen Cline

2001

DEDICATION

This dissertation is dedicated to my darling wife, Kelly, who
persevered with me in its development, and to my mother, Marilyn, who always believed

this would happen.

ACKNOWLEDGEMENTS

This research project was funded through a generous grant by the W. K. Kellogg
Foundation (Battle Creek, MI) to Applied Research (East Lansing, MI). This research
would not have been possible without the support of the principal investigators of the
cluster evaluation of the Comprehensive Community Health Models of Michigan
initiative, James Dearing, Ph.D., and R. Sam Larson, PhD. Their financial support,
however, was dwarfed by their moral support, unflagging encouragement, and regular
insightful feedback on some of the concepts contained herein, as those concepts were
developed and applied throughout the course of the cluster evaluation. I remain in their
debt for offering me the incalculable, career-shaping opportunity to participate as a
member of their interdisciplinary evaluation team. Such freely given friendship and faith
earns high rewards in this life, and in others.

Carol Weissert, Ph.D., at the Michigan State University, is a mentor nonpareil.
Dr. Weissert provided the perfect blend of academic criticism, moral support, and
sensitivity that a student employed in a full time career requires when also engaging in the
arduous process of conducting dissertation research. Her willingness to spend numerous
hours working with me on all of the concepts contained herein and reviewing multiple
(long) versions of the narrative made the process more enriching and enjoyable than I
could have ever imagined. Words alone do not suffice to provide due respect to the
largely anonymous and time-consuming task shouldered by dedicated graduate student

mentors such as Dr. Weissert. It must be said that it is a unique feeling of

vi

accomplishment to know that she deemed me worthy of the opportunity to meet her high
standards of academic research and scholarship.

My committee members, Cynthia Jackson-Elmoore, Ph.D., Michael Mintrom,
Ph.D., and, Ken Williams, Ph.D., provided valuable feedback to me during the
dissertation proposal writing process and at my defense that immeasurably increased the
quality of this work. Dr. Jackson-Elmoore, while serving for three years alongside me as a
mentor and collegial member of the cluster evaluation team of the Comprehensive
Community Health Models of Michigan initiative, continuously provided additional
insights into many of the concepts contained herein as these were developed. Dr.
Mintrom provided a great deal of inspiration and insight into my efforts to transport
theories of entrepreneurship from the federal and state levels to the local level. Dr.
Williams patiently worked with me to iron out the kinks of the decision models through
many early iterations, and provided appropriate praise when these models met his
exacting standards.

Chris Butler, Ph.D., proved a true friend and colleague as, over a period of several
months, he questioned every assumption in my decision models throughout these models’
development. He politely, but firmly, forced me to defend every aspect of the design of
these two models, and guided me through a literature that I had only partially remembered
prior to attempting to resurrect these rusty skills for application in this research study.

Margaret Woolgrove, a colleague with wonderful editorial skills, patiently read

through the manuscript after its completion and immeasurably increased the grammatical

quality and internal consistency of this document. I must extend personal thanks to her for
reading about topics of which she had little personal interest in order to add this value.

My colleagues at the Michigan Public Health Institute were another source of
unending support and politely listened to me as I (endlessly) discussed the concepts
contained herein. As important, my work colleagues were willing to remind me, in my
most despairing moments, that they believed I would accomplish this first goal in my
career as a social science researcher. With their help, I have.

Although listed last, my wife, Kelly, my daughters, Eva and Elli, and my family
and friends proved to me that, without a strong network of personal and familial support,

success, although perhaps obtainable. is largely meaningless. Their love, support, and

 

occasional sharp reminders that I “should be working on my dissertation,“ were the spurs
that kept me on track for many months. My wife, Kelly, deserves most of the credit for
the completion of this dissertation. Early on she saw in me more positives than I saw in
myself, and her loving faith coupled with timely firmness have brought out the best of
what I have to offer the world. Most of all, her loving dedication to her step children
permitted me the time needed to complete this research project after work hours.

And last, but by no means least, my mother, who trained me from earliest
childhood to believe that “Dr. Cline” was a title that in her eyes I was somehow destined
to achieve. It took years to believe that her dreams for me could also become my own.
Yet she always knew this day would come. Bless her for waiting 36 years and 8 months

for that day to arrive.

viii

 

The aforementioned persons have done all that they could do, the shortcomings of

this dissertation are the sole responsibility of the author.

ix

 

 

TABLE OF CONTENTS

Chapter One: Overview and Organization .....................................
Overview ........................................................
Organization ......................................................

Chapter Two: The Federal, State and Local Context for Health Care Coverage ........
Introduction to the Federal, State and Local Policy Environment .............

The Provision of Health Care Coverage/ Services in the United States .........

Mixed Methods of Purchasing Health Care ............................. 1

The Market for Health Care in the US ................................. l

The Intergovernmental Nature of Health Policy in the United States ......... l
Possibilities for a Political (Policy) Solution ............................ 1

The Uninsured ................................................... 2
Summary ....................................................... 2
Chapter Three: Change Agents, Policy Entrepreneurs and Policy Change ........... 2
Change Agents ................................................... 2
Foundations as Change Agents ................................ 2

Government Agencies as Change Agents ........................ 3

Summary of Foundation and Government Agencies as Change Agents

................................................... 3

Entrepreneurs and Policy Change .................................... 39

Characteristics of Policy Entrepreneurs and their Activities ................ 45
Collective Entrepreneurship ................................... 48

The Role of Information ..................................... 50

The Importance of Context ................................... 52

The Market Entrepreneur ..................................... 55

The Policy Entrepreneur ..................................... 56
Combining the Roles, and the Use of Information ................. 57

Costs and Benefits .......................................... 60

From Whence do Policy Entrepreneurs Appear? ................... 61
Summary ....................................................... 63
Chapter Four: Research Design and Methodology ............................. 66
Introduction to the Research Design .................................. 66
Case Study Design ..... ' ..................................... 67

Study Questions .................................................. 70
Propositions and Units of Analysis ................................... 71
Change Agents ............................................. 71

Policy Entrepreneur ......................................... 74
Exogenous Factors .......................................... 78
Generalizability ............................................ 81

Criteria for Interpreting Findings: Links Between Interview Data and Propositions

......................................................... 81

xi

Links between interview data and propositions ...................
Variables ................................................
Variable Coding and Scoring Process ..........................
Criteria for Interpreting Findings: Two Decision Models .................
Model Structure ...........................................
Outside-Inward Model of Change .............................
Information Constraints ...............................

Payoffs in the Outside-Inward Model of Change ............

Probabilities within the Outside-Inward Model of Change . . . .

Propositions of Relationships Among Probabilities in the OUlSlt
Inward Decision Model .........................

Operationalizing the Outside-Inward Decision Model .......
Conclusions from the Outside-Inward Model Of Change .....
Model of Inside-Outward Change .............................
Information Constraints ...............................
Payoffs for the Inside-Outward Decision Model ............
Probabilities within the Inside-Outward Model of Change . . . .

Propositions of Relationships Among Probabilities in the Inside
Outward Decision Model ........................

Operationalizing the Inside-Outward Decision Model .......
Conclusions from the Inside-Outward Model of Change .....
How Iteration Affects the Two Models .........................

Comparing the Models ......................................

xii

Chapter Five: Case Studies, Assessment of Propositions and Decision Tree Models . .
Introduction ....................................................

Case Study Of a Foundation Change Agent: William K. Kellogg Foundation
(WKKF) .................................................

WKKF’s Concept for CCHMs ................................
WKKF’s Goal for CCHMS ..................................
WKKF’S Conceptualization of its own role as a Change Agent ......
WKKF’S Conceptualization of Project Leadership ................
WKKF’S Selection of Demonstration Sites ......................
Summary of WKKF as a Change Agent ........................
Case Study of the State as a Change Agent ............................
Summary Of MDCH as a Governmental Change Agent ............
Calhoun County Case Study .......................................
Summary of Calhoun County Case Study .......................
Muskegon County Case Study ......................................
Summary Of Muskegon County Case Study .....................
St. Clair County Case Study .......................................
Summary of St. Clair County Case Study .......................
Wayne County Case Study .........................................
Summary of Wayne County Case Study ........................
Assessment of the Propositions across all Cases ........................

Change Agents ............................................

xiii

Policy Entrepreneur ........................................ 21 l

Exogenous Factors ......................................... 219
Generalizibility ............................................ 233
Assessment of the Decision Models ................................. 235
Constraints on the Assessment of the Decision Models ............ 235
Assessing the Outside-Inward Decision Model ................... 236

Discussion of the Outside-Inward Decision Model’s Propositions . . . . 239

Final Conclusions from the Outside-Inward Model Of Change ....... 241

Assessing the Inside-Outward Decision Model ................... 242

Final Conclusions on Both Decision Models ..................... 243

Chapter Six: Conclusion and Implications for Future Research .................. 245
Answering the Research Questions .................................. 246
Contributions of this Research ...................................... 250
Implications for Future Research .................................... 253
APPENDIX A ........................................................ 259
APPENDIX B ........................................................ 267
APPENDIX C ........................................................ 276
APPENDIX D ........................................................ 286

xiv

BIBLIOGRAPHY ..................................................... 298

XV

 

LIST OF TABLES

Table I, National Spending on Health Care, 1965, 1980, 1994 and 1998.

Table 2, Percent of National Spending on Health Care, by Sector, 1965, 1980, 1994 and
1998.

Table 3, Differential Information Use Across Milieus.
Table 4, Comparison of Change Agent Roles by Type
Table 5, Count of Interviews by Case.

Table 6, Timing of Interviews by Case.

Table 7, Mathematical Calculations of Simulated Values for the Outside-Inward Decision
Model, Example 1.

Table 8, Mathematical Calculations of Simulated Values for the Outside-Inward Decision
Model, Example 2.

Table 9, Mathematical Calculations of Simulated Values for the Outside-Inward Decision
Model, Example 3.

Table 10, Proposition #1, Variables by Case.
Table l 1, Proposition #2, Variables by Case.
Table 12, Proposition #3, Variables by Case.
Table 13, Proposition #4, Variables by Case.
Table 14, Proposition #5, Variables by Case.

Table 15, Decision Model Outcome Paths for the CCHMS Cases.

 

LIST OF FIGURES

Figure 1, A Nested Market View of the Health Care System.

Figure 2, Uninsured Families by Work Status Of Family Head, 1999.

Figure 3, Workers Aged 18-64 without Health Insurance, by Firm Size, 1999.
Figure 4, Model of Outside - Inward Change.

Figure 5, Model of Inside - Outward Change.

xvii

 

 

INTRODUCTION

Chapter One: Overview and Organization

Overview

Supported by a foundation change agent, policy entrepreneurs in some
communities have sought to build partnerships that seek to extend health care coverage to
the working uninsured through the development and launch of a specific program or
product. These partnerships have involved a myriad of actors whose participation in the
partnership ranges from passive to active, and whose role in the partnership varies with
both the political and market context within which the innovation (either a program or
product) exists. Because these innovations require local and state policy activity as well
as the development of a health care coverage product or mechanism, the policy
entrepreneurs must engage in activities in both political and market contexts.

This research consists Of six case studies and two formal decision models of
change efforts and policy innovation at the local level. Two of the case studies are of two
different entities, a private foundation (the W. K. Kellogg Foundation, based in Battle
Creek, Michigan) and a state agency within Michigan (the Michigan Department of
Community Health) assuming the role of an external change agent that pursues a change
effort at the local (in these cases, county) level. The literature review presented in Chapter
Three and the propositions presented in Chapter Four elucidate and define the expected

roles of both entities acting as change agents at the local level.

The four remaining case Studies (all within Michigan) are of county-level change
efforts attempting to develop and launch local innovations to extend health care coverage
to the uninsured. Three of these change efforts were funded by the Kellogg Foundation,
the other was undertaken by an indigenous local-level policy entrepreneur in the absence
of an external change agent. All four county change efforts were expected to have
interactions with the state agency of interest, which may or may not have chosen to act as
a change agent.

The four local-level case studies conducted as a part of the research presented here
use theories of both change agents and policy entrepreneurship (theories developed
largely, but not entirely, for entrepreneurs acting at the federal and state level) to explain
how these local change efforts did or did not develop and launch an innovation designed
to extend health care coverage to the working uninsured. The literature review focuses on
the skills and assets that entrepreneurs are expected to possess, and the ways in which
political and market contexts in technically complex issue areas affect the probability for
success of a change effort within such an issue area.

As mentioned, this study also develops two decision models that illustrate the
ways in which policy entrepreneurs are constrained by the local political and market
context in their efforts to build winning coalitions that support the development and
launch of proposed innovations. Although the models demonstrate the constraints on
policy entrepreneurs, analysis of qualitative data collected for the case studies
demonstrates how policy entrepreneurs may use their skills and assets to convince

potential members of a supportive coalition that the benefits to these potential members

are positively changed through collective action, or also alter the members’ perceptions of
the benefits of collective action.

The models developed and assessed in this study present a two step process for
entrepreneurial change efforts; the first step is the building of a coalition of critical local
actors within the issue area that will support a change effort, the second step is the
development and launch of an innovation within the technically complex issue area.

This research presents a blended approach to the study of policy change that also
presents a theory for the role for both external change agents and local policy
entrepreneurs or collections of policy entrepreneurs in the policy change process. This
approach is not intended to negate previous theoretical work on policy change, but rather
to serve as an expansion of the literature. Mintrom (2000) presents a theory of policy
entrepreneurship that, rather than seeking a monocausal explanation that focuses on the
activities of policy entrepreneurs in the introduction of policy innovations, instead
explains the political context (expanded in this work by drawing on Schattschneider.
1975, and Baumgartner and Jones, 1991) within which policy entrepreneurs operate.
Mintrom also refined Wilson’s typologies of policy making (1973), the role of the policy
entrepreneur in coordination across issue networks (Heclo, 1978), target groups
(Schneider and Ingram, 1997), and within Terry Moe’s model of structural choice (1984,
1990). Roberts and King (1996) present a theory of collective entrepreneurship that
extends existing theories of policy entrepreneurs by recognizing the importance of

building teams in complex, technical issue areas. This research builds on this and related

research by adding in the role of change agents, and applying the notion of market

entrepreneurship intertwining with political entrepreneurship to the local level.

This research achieves two goals. The first is to develop a model of collective
policy entrepreneurship at the local level and to assess the probability that policy
entrepreneurs can succeed in building a supportive coalition given a complex and
technical issue. These probabilities may be affected by the presence or absence of change
agents, the local political and market contexts, and the personal characteristics of the
persons filling the role of local policy entrepreneur.

The second goal is the presentation of a theory of how private philanthropic
foundations (and to a lesser extent government agencies) may act as agents of change at
the local or community level to encourage the search for innovations that improve social
and/or economic well being within targeted communities.

The study questions for this research are:

1. How [did] a foundation play the role of change agent to encourage a local
partnership that attempts, through the development and launch of some program
or product, to extend health care coverage to the uninsured?

2. How [did] state government play the role of change agent to encourage the
development of a local partnership that attempts, through the development and
launch of some program or product, to extend health care coverage to the
uninsured?

3. What persons filled the role of local policy entrepreneur, and how did their

placement within the local policy and market contexts (i.e., their social position

and employment role) affect their ability to pressure for the development and
launch of some program or product to extend health care coverage to the
uninsured?

4. How [did] local policy entrepreneurs go about building collective
entrepreneurship team to permit the development and launch of some program or
product to extend health care coverage to the uninsured?

5. How [did] the local policy entrepreneur alter the structure of costs and benefits for
those members of the partnership? Did the local policy entrepreneur alter the
members’ perceptions of cost?

6. How [did] the political and market contexts affect the ability of local policy
entrepreneurs to build a partnership around the issue of the development and
launch of some program or product to extend health care coverage to the
uninsured?

7. Could models developed at the local level in Michigan apply elsewhere?

The contributions of this research to the literature are:

1. An exploration and beginning explanation of the role of foundations and, to a
lesser extent, state agencies, as change agents that may spark the innovative
activity of a policy entrepreneur or collective entrepreneurship team.

2. An exploration of how or if a local entrepreneur or collective entrepreneur team
may act as both a political and market entrepreneur in order to achieve his/her

goals.

 

3. Two explorative models of policy entrepreneurship that define the different major
contextual constraints faced by local policy entrepreneurs, assess and assign
different probabilities associated with these constraints, and define, assess and
assign probabilities associated with those constraints that an entrepreneur will
engage in the described entrepreneurial activity based on the presence or absence

of change agents and the personal characteristics of the policy entrepreneur.

Organization

The format of this dissertation is as follows:

Chapter I: Overview and Organization

Chapter 2: The Federal, State and Local Context for Health Care Services and
Coverage

Chapter 3: Literature Review

Chapter 4: Research Design, Proposed Methodology and Analysis

Chapter 5: Case Studies, Assessment of Propositions and Decision Models
Chapter 6: Conclusion and Implications for Future Research
6

 

 

 

 

Chapter Two: The Federal, State and Local Context for Health Care Coverage

Introduction to the Federal, State and Local Policy Environment

The difficulties associated with achieving national health care reform to improve
access to health care were once again demonstrated in 1994 when the Clinton health plan
failed passage in Congress. Weissert and Weissert (1996) point out that the political
feasibility of achieving health care reform at the national level is a function of several
important factors, any one of which may make a national reform plan impossible to pass,
let alone implement. At the state and local level, governments may avoid some of these
factors, or the factors may be weaker and less likely to sink a reform plan. Yet policy
activities around access to health care at these levels face common problems: limits on
the availability of resources necessary to implement expansions in health care coverage,
the complexity of the issue area and the few actors outside of those who are suppliers that
understand the complexities, the market and legal power of the suppliers, and the way in
which supply and demand for health care services is skewed by the market and legal
power of the suppliers. At the local level the issue of resources is particularly acute.

With the demise of the Clinton health care reform plan, the provision of health
care coverage to uninsured Americans defaulted to private markets and state and local
governments. The latter two entities are best conceptualized not as structures or
organizations but rather as arenas where resources are assigned, where “winning and
losing” occurs. This conceptualization allows us to see private markets and state and local

governments as performing similar activities, although using different mechanisms.

—_—__

 

Across the country local markets for health care coverage and services are
changing, although the change in local markets varies considerably from one locale to
another. The changes are anointing new winners among both new and more established
actors as well as creating losers of some established actors. At the same time, state
governments have begun to take steps into the governmental void left behind after the
effectual ending of debate on health care reform at the national level (Oliver and Shaheen,
1997).1 With established interests weakening, new interests rising and the rapid change
within the market, we have the textbook recipe for creating successful policy initiatives
intended to redistribute or share resources, thus potentially creating new winners and
losers (Wilson, 1973). At this time, a fuller discussion ofthe provision of health care

coverage/services in the United States is necessary.

The Provision of Health Care Coverage/Services in the United States

The United States spends more on health care, both in terms of Gross Domestic
Product and per capita, than any other country in the world. Expenditures for health care
have been rising steadily (with a recent drop in the rate of this increase) for decades (see

Table 1). Despite these huge and increasing public and private expenditures for health

 

‘ The Health Insurance Portability and Accountability Act (HIPAA) and
State Children’s Health Insurance Program (SCHIP) both nibble at the
edges of the problem of the uninsured, although SCHIP has the potential
to cover a great many currently uninsured children. This doesn’t alter the
fact that the average uninsured person is over 18, and thus not eligible for
coverage under SCHIP. It would also seem likely to a prudent observer
that the national level government may well become too consumed with
the issue of maintaining a viable Medicare system to pay close attention to
issues of expanding coverage.

8

_—

care, the proportion of the population that is uninsured remains in the double digits
(18.4% of persons under age 65 in 1998 [Fronstein, 2000]2). The proportion of the
population that is uninsured has been growing steadily for over a decade, despite a vibrant
economy and multiple federal and state eligibility expansions for the Medicaid program,
as well as a variety of local-level initiatives throughout the country (Quinn, 1997;
McBride, 1994). This stubborn rise in the uninsured is perhaps more surprising if we
consider that the US population has also been aging over the same time period, increasing
the proportion of the population over age 65 and eligible for Medicare.

There are a variety of reasons cited in the literature for the disparity between
expenditures and number of persons with some form of health care coverage. Four
culprits appear consistently throughout the literature:
> The growth in the number of small businesses, which have historically not

provided health insurance (McLaughlin, 1994; Chollet 1994).
> A change in the economy with growth in job sectors where insurance has not been

prevalent (Acs, 1995).
> The rising cost of health insurance (Acs, 1995; Kronick, 1991).
> The rapid growth of health care expenditures as compared to personal income.

(Kronick & Gilmer, 1999).

 

Uninsured statistics are calculated for persons under the age of 65 because
Medicare provides some form of coverage to all persons aged 65 and
older.

———

 

Mixed Methods of Purchasing Health Care3

The United States is alone among the richest industrialized nations in the mixed
method by which health care services are purchased for its citizens. Unlike most other
industrialized countries, over half ofthe health care services in the United States (54.5%)
are purchased by private sector entities (HCFA, 2000). The remaining health care services
are purchased by the public sector, with the federal government purchasing 32.8%, and
state and local governments purchasing the remaining 12.7%. The private sector’s share
of these expenditures has dropped sharply since the mid-19605 (see Table 2). The bulk of
this decrease in the proportion of private sector spending has been the result of the
increase in federal spending on health care services through the Medicare and Medicaid
programs. (The Medicaid program expenditures include formulaic state funds.) However,
the remainder of the decrease has been the result of a net loss in the provision of health
care coverage by businesses (primarily small businesses, Donelan, et al., 1996) and a

consequent increase in the number of persons who are uninsured.

 

The terms “purchase” and “purchaser” have specific meanings in health
care finance, as do the terms “pay” and “payer”. “Purchasing” normally
refers to some organization/individual that is buying a defined package of
health care services from a specific provider or set of providers or from a
payer, often prepaid, with some decision having been made among
available providers of health care services based on notions of cost and
quality of care. “Paying” refers to a more narrow and passive role where
some organization/individual is reimbursing a provider or set of providers
for the delivery of health care services.

10

—_

The Market for Health Care in the US

It is useful when thinking of the mixed methods by which health care is purchased
(and delivered) to imagine that we are talking about a nested market system (see Figure
I). The term “nested” reflects the variability in health care markets across the country at
both the state and local level, while allowing for the role of the federal and state
purchasing that cuts across local market boundaries. We can determine the geographic
boundaries of these different markets, defined either by service delivery areas or
politically, along state, county or municipal lines.4 Each market has a permeable
membrane, because each market is able to some extent to control the way in which the
market operates within its own boundaries. However, each market is also increasingly
affected by what is occurring in nearby markets, and the larger state and national markets
within which they are nested (through Medicaid and Medicare expenditures which are
largely beyond these local market actors’ control). Also, although some health care
coverage/services tend to still be primarily local (e.g., primary care or outpatient surgery),
other services and products such as pharmaceuticals and laboratory tests, generally '
operate within markets that are regional (multi-county, multi-state and national) and thus
increase the permeability of the smaller markets for health care services.

It is also useful to keep in mind that health care markets are far from free markets.

These markets have long been dominated by two guild-like organizations (the American

 

4 See the Dartmouth Atlas of Health Care (Wennberg, 1999) for a
description and examples of defining health care boundaries by delivery of
service. The two most commonly used boundaries are hospital referral
regions and hospital services areas.

11

—_ ‘4

Medical Association and the American Hospital Association) that have controlled the
market entry of new suppliers, as well as the way in which supply is structured and who
has access to all forms of market information (Starr, I982). The AMA and the AHA have
maintained, until very recent times, a stranglehold on the supply side of the market, the
effects of which have been argued to materially affect most of the demand within the
market (Wennberg, 1997). These two groups (each of which have well-organized sub-
organizations at the state and regional level, and for the AMA at the county- and
sometimes even municipal-level) have the ability to manipulate legislation that affects
their professions and the markets within which they operate. They also affect the ability
of purchasers to understand their options for health care, retarding demands for
competition amongst providers, retarding access to quality and cost information, and
slowing purchaser demands for changes in legislation and regulation. Only with the fast-
paced increases in health care costs of the 19705 and 19805, as well as the introduction of
prospective payment by Medicare and Medicaid, have we witnessed these markets
becoming more “free.” With efforts by various purchasers and payers to shift the costs of
providing health care to their populations onto others. purchasers demanding and in some
cases creating competition, an increase in the availability of quality and cost information
made available to very large purchasers, and the increased permeability of the membranes

of the many smaller local markets these markets have begun to open (Rice, 1998)}:6

 

The rise in health care costs in the 705 and 805 was largely the result of
rapid technological advances that provided effective but expensive care
that prolonged life, especially in the last six months of life. The recent
slower rise in health care costs at the end of the 19905, although carrying a
similar high technology component, is driven mostly by increases in the

12

I . , .

Physicians, hospitals/health systems and health plans have all felt pressures from
the change in purchaser/payer behaviors, however, it is arguable that physicians have
perceived themselves as having lost the most power (Marmor, 2000). Physicians perceive
themselves to have lost both market power and decision making power over how best to
treat their patients. The loss of market power has been felt as they have faced pressures to
reduce reimbursement levels for services from both traditional fee-for-service plans and
from managed care plans, as well as from public payers/purchasers represented by the
federal Medicare system and state Medicaid agencies. The loss of decision making power
has come as a result of the increase in managed care organizations that have placed limits
around the possibilities and pathways of treatment for patients enrolled in their plans.

At the same time, physicians have seen the slow but steady erosion of their legal
power. From the mid-18005 until the recent present, physicians have successfully used the
argument that only they with their acquired expertise were capable of exercising the
power to prescribe and provide a broad array of services to their patients (Starr, 1982).
This legal power gave physicians a strong grip on the supply side of the market. This
power was institutionalized by both federal and state legislatures and rigorously enforced

by the American judicial system (Starr, 1982). With relatively recent increased pressures

 

costs of new and more effective, but expensive, pharmaceuticals (Health
Care Trends Report, 1999).

6 “Prospective payment” is a method established by the federal government
where the costs of health care services are collected into diagnosis related
groups (DRGs). Services within each DRG are paid at preset rates, with
these rates changing annually at the recommendation of the Prospective
Payment Assessment Commission.

13

 

from payers and purchasers for reductions in the costs of providing health care, we have
seen the slow erosion of this power through changes in legislation granting other health
care professionals (e.g., nurse practitioners, physician assistants and certified nurse
midwives) the legal power to both provide and prescribe services that in the past were the
sole purview of physicians.

One can still argue that in many communities all health care is local and that there
is no dominant larger market for health care beyond the community-level, which captures
the fact that health care providers often live in or near the communities in which they
practice or operate hospitals and health systems, and for the most part all of their patients
are local (Health Care Trends Report, 1999). However, hospital/health systems are also
contracting for access to patients with large and growing health plans (payers) whose
cachement areas extend beyond the scope of any single community, even large cities, as
well as sometimes crossing state borders.7 As well, the federal government and the states.
acting sometimes as payers and sometimes as purchasers, cut across local markets in their
supply of services through Medicare and Medicaid.

These same health plans are then contracted by a multitude of firms large and
small (purchasers), and are increasingly contracting with state agencies (purchasers) for
the right to provide care to the Medicaid population. Some firms are small, with
employees that reside only locally, and these employees tend to represent a small

proportion of a health plan’s enrollees. Other firms may be large, and also may represent

 

7 A cachement area is the geographic area within which a health plan
operates. It is defined by both the location of providers and enrollees.

l4

 

 

a significant proportion of a health plan’s enrollees. Lastly the “newest” set of market
players on the scene are states, who by entering this market with Medicaid enrollees being
bid out to health plans at the lowest price (instead of acting merely as payers of Medicaid
claims), are actors with enormous market share that is translating itself into the power to
affect substantial change in the structure of local markets for health care. The effects of
these recent changes in the methods by which states are beginning to purchase coverage
for their Medicaid populations are still unknown, and the predictions of how the presence
of these “new” market actors will affect change are many and sometimes contradictory.

At the local level efforts at providing health care services to the uninsured have
centered around both public and private actors. In many communities, local health
departments have been charged or have taken on the responsibility of providing primary
health care services to persons without access to the normal health care system, as a result
of being uninsured or underinsured. These organizations have long provided a partial
health care safety net in communities, funded by local and state public dollars. It is also
not unusual for there to exist nonprofit, community-based organizations (CBOs) that
began as providers of human services, which also provide varying levels of primary
health care services. Not surprisingly, communities vary widely as to the extent of
services available from these kinds of public agencies and private organizations.

In almost all communities, hospitals and health systems provide free or reduced
cost health care services to persons with limited ability to pay and/or have insufficient or
no health care coverage. The American system of health care has long held a premise that

no individual may be denied emergency care regardless of their ability to pay (Starr,

15

 

1982). This partial safety net is rarely coordinated, and the provision of these services
comes through either hospital emergency rooms or through hospital- or local health
department-operated clinics.8 Interspersed throughout the many local markets for health
care are individual physicians and group practices that voluntarily provide varying levels
of free or reduced health care. These private actors (nonprofit and for profit as well as
public9) have historically recouped the costs of these services through the shifting of these
costs onto those persons with health care coverage through increases in how much these

organizations charge the insured for services.

The Intergovernmental Nature of Health Policy in the United States

The development, enactment and implementation of health policy in the United
States occurs at the federal, state and local level. The federal government for the most
part originates policy and sets guidelines for implementation, while state governments
implement federal policy (Medicaid & SCHIP) but also engage in innovation within
existing federal programs as well as introducing new policy initiatives, e. g. risk pooling,
health care coverage for the self-employed or for high risk individuals (Blumberg and

Nichols, 1995). Local governments have traditionally been seen as implementers of

 

An example is Wayne County, Michigan, where two related programs
provide additional health care coverage to both indigent persons as well as
the working uninsured using a mixture of federal, state and local funds, as
well as systems of coordinated care for the uninsured that rely solely on
the sharing of existing resources by a mixture of private and public health
care organizations.

9 Some hospitals are publicly-founded and governed, either by state or local
(county or city) governments.

l6

 

 

public policy as directed by the federal and state level. All of these activities may be both
coordinated and uncoordinated. Activities may be coordinated such that either the federal
or state level government may coordinate the activities of implementation of policies by
levels below themselves (i.e., Medicaid policy enacted at the federal level and
implemented by states, and the public health code enacted by the state and implemented
at both the state and local level.) Policy activities may also be uncoordinated as states and
local communities attempt to grapple with policy problems that actors at these levels of
government perceive to be unaddressed by levels of government above themselves.

The efforts of local governments to address health care problems of the uninsured
have tended to be limited, primarily because of limited resources. Local level policy
efforts have tended to rely on the resources directly controlled by public health officials,
mostly the marginal revenues derived from full or partial payment for health care services
delivered to both insured and uninsured patients who present at public health department
clinics. It also the case that few county or city officials propose reallocating existing
resources or extracting additional resources to allocate to the problem of the uninsured.
Despite these low levels of policy activity, local policy makers have frequent contact with
actors within the local health care system, as well as regional (health plans and larger
health systems) and state actors (the Medicaid agency).

This contact consists of both market and political interactions. Market interactions
at the local and regional level involve working with local providers in their role as

providers of human services and health care services to the poor and uninsured, or

negotiating contracts with regionally-based health plans to provide services to the
Medicaid enrollees of these plans.

The political interactions include implementing state policies such as the
establishment of clinics to serve particular populations (e.g. in Michigan the
implementation of Maternal and Infant Support Services clinics) with state agency
bureaucrats and their state representatives in the legislature. Directors of local health
departments interact politically through professional organizations and personal contacts.

These interactions place local governments and local government agencies solidly
within both the market and political context at the local, regional and state level.
However, it is important to remember that local officials are not strong policy actors.
These interactions are characterized by their low level of locally-controlled resources,
their role of health and human service provider and their role as implementers of state-
level policy. However, local government officials are the only actors that regularly and

directly interact in both market and political contexts at both the local and the state levels.

Possibilities for a Political (Policy) Solution

In their book, “Governing Health” (1996), Carol Weissert and William Weissert
detail the relatively few opportunities for the US. Congress to pass legislation for
national health insurance in the last several decades similar to what exists in most other
advanced industrialized nations. The most recent attempt at the creation of national health
care insurance failed in 1994, and most health policy experts agree that another

opportunity for addressing the issue of national coverage is several years into the future.

18

 

 

There exist several reasons for this. One is that “socialized medicine” (also known as
“single payer”) solutions in other industrialized countries are experiencing real problems
related to both increasing costs and long waiting lists for services, as well as problems
related to the “rationing” of high cost or experimental services. The other is that the
mixed method of purchasing health care through both private and public mechanisms is
so firmly rooted in the American way of delivering health care and providing health care
coverage that a wholesale conversion to an entirely different system seems politically and
financially impossible.

The problems associated with a political solution at all govermnental levels for
reducing the number of uninsured are reminiscent of the ‘problems chasing a solution’
model posited by John Kingdom (1995). In the case of finding a political solution of
providing coverage to the uninsured we find that there are a multiplicity of models at
different levels that focus on different aspects of the problem, but because of the
complexity of the problem, and the ultimate costs of any solution, these solutions are
relegated to discussions amongst academics and policy experts within and without of
government.

What is most interesting for this study, and as a topic for future research, is that as
we move from the federal to the state to local levels of government and governance, we
appear to witness increases in the amounts of activity undertaken to address this problem.
As mentioned earlier in this chapter, observation suggests that this is because of the
decrease in the number of actors that need to be involved in a solution (which is

exacerbated by the mixed methods of providing health care coverage in the United

19

 

States), the decrease in the complexity of the issues faced by focusing on a single local
health care market, and the fact that the need for either coverage or at a minimum
coordination of resources is felt on a day-to-day basis by all actors.

As a result of the defeat of the Clinton national health care coverage plan, the
near—term opportunities for increasing publicly—financed coverage at the federal level
have been restricted to coverage that targets those who are least able to alter their
insurance status, such as the State Children’s Health Insurance Program (SCHIP).
Therefore, the assessment by Weissert and Weissert that the likelihood of a near-term
political solution at the federal level to the uninsured seems to be holding. The only other
“major” activity that has occurred at the federal level has been the Health Insurance
Portability and Assurance Act of 1996 (HIPAA), which did not provide funding for
increased coverage, but did allow enrollees to maintain their employer-based coverage in
a health plan after departing their employer. However, the costs of maintaining coverage
are borne solely by the consumer.

The private sector, led by large firms, has begun to make aggressive attempts to
control health care spending. These attempts have lowered the costs of these large
purchasers through the introduction of managed care arrangements and reduction of both
administrative costs and cost shifting. These actions have exacerbated the problem of

covering the costs of providing care to the uninsured.10 Despite an economy that has been

 

'0 In an effort to reduce the rates of large purchasers (both private and public)
health plans are not only reducing the funds available for providing
uncompensated care, but are also shifting the “savings” provided to large
purchasers to smaller purchasers who face information and collective
action constraints in forestalling this shifting (Blumberg and Nichols,

20

growing, in the face of numerous federal and state expansions of the Medicaid program
and an increase in the proportion of the population eligible for Medicare the number of

uninsured continues to grow. This begs the question, who are the uninsured?

The Uninsured

The uninsured tend not to be the poorest members of our society, because the
poorest members are eligible for Medicaid, which is publicly-supported health care
coverage. Medicaid eligibility is based on means-testing, that varies from state to state,
within broad federal guidelines. (Because the Medicaid program is funded from both
federal and state coffers, states have some discretion in setting eligibility standards.)
Means-testing is a method used by state governments to determine eligibility for the
Medicaid program. Means-testing methods consider family income, family size, and
family assets to determine the eligibility of a family and specific members of a family for
Medicaid benefits, which are designed to provide health care coverage to the poorest
segment of the population. The uninsured tend to be the working poor, or the family
members of the working poor (see Figure 2). Data from the March 1999 Current
Population Survey estimate that 61.2% of the uninsured live in a family where the head of
the household is employed full time, for a full year (this percentage includes one-person

families). This group is ineligible for the means-tested Medicaid program, due to either

 

1995). This results in smaller purchasers dropping more expensive
coverage, increasing the number of uninsured.

21

 

 

income level, total assets, or both. Only 16.8% of the uninsured live in a family where the
head of household is unemployed and no one in the household works.

The uninsured are concentrated in families with lower incomes. The families most
likely to be uninsured are also single parent families, and/or where the family head is
working in industries that are less likely to offer health insurance (Fronstein, 2000).
Perhaps the most striking statistic, other than the one that shows that most of the
uninsured are in families where the head of household works full time all year long is the
likelihood that the uninsured are employed by smaller firms (see Figure 3).

The last aspect about the uninsured that must be clarified is the length of time
people spend without health care coverage, and the costs these long periods without
coverage impose on both uninsured people and the health care system. Among the
uninsured three-fourths of these persons are without health care coverage for more than a
year, and over half are without coverage for over two years (McBride, 1994). The people
who experience long spells without coverage have negative cost effects on the health care
system as a whole, as well as the individual-level adverse health effects for these people
(IOM, 1993). The absence of health care coverage reduces a person’s ability to access
health care services, thereby reducing the likelihood that a person will obtain health care
services when these services are needed.

This raises cost issues for the health care system as a whole because although
persons without health care coverage consume far fewer health care services than those
with either private or public health care coverage, these persons do still become ill or

injured, and once seriously ill or injured, they receive care by the system, regardless of

22

 

their ability to pay (IOM, 1993). However, researchers have long argued that the cost of
treating uninsured persons if and when they become seriously ill is higher than the cost of
providing them with lower levels of health care. The recent field of study of ambulatory
care sensitive (ACS) conditionsll has shown that these conditions are primarily occurring
among the uninsured population (Billings, et al., 1991, 1996). John Billings has also
compared the prevalence of ACS admissions in several major US cities to ACS
admissions occurring in Canadian cities, and found very large differences in the number
of persons presenting at hospital emergency rooms with ACS conditions. These
differences seem attributable to nothing other than the fact that Canada has no uninsured
citizens (1991).

The interaction between the market-driven lowering of costs to private, and now
increasingly public, purchasers and payors, and the effects of charity care for the
uninsured are as yet not well understood, although reductions in the amount of charity
care being provided have been documented (HCFO, 1997). Partially this is because the
market has not squeezed all of the shifted costs out of the costs of privately and publicly
purchased health care coverage; it is also due to the recentness of the changes in the
behavior of large purchasers from being very passive to aggressively searching for lower
costs and forcing reductions or freezes on the costs onto the providers providing health

care to their employees.

 

” Ambulatory care sensitive (ACS) conditions are conditions for which
persons are hospitalized that, if the person had received adequate
preventive or primary care, the condition would most likely not have led to
a hospitalization. Examples are hospitalizations for ear infections or
complications arising from poorly managed diabetes.

23

 

In the Detroit area we are beginning to see the long-term effects of the squeezing
out of cost shifting as well as the reduced revenues obtained from Medicare as a result of
the 1997 Balanced Budget Agreement, which made deep cuts in Medicare payments to
academic teaching hospitals. The evidence is in the fiscal crisis at the Detroit Medical
Center (DMC) and its decisions to stem financial losses. A long time provider of
uncompensated care to area residents, the DMC has recently begun issuing “bills” to
patients who are uninsured, which has had the reported effect of reducing the number of
uninsured patients presenting at DMC emergency rooms and outpatient clinics. It is not
clear yet if these uninsured patients are going elsewhere for care or if these persons are

simply forgoing care.

Summary

This chapter has been included as a part of this research study for three reasons:
(1) because the general reader is expected to be unfamiliar with the issue area of health
care, (2) to enhance reader understanding of the discussion surrounding the development,
maintenance and weakening of policy subsystems in technically complex issue areas (to
be described in Chapter Three), and (3) to prepare the reader for understanding the
purpose of the innovations studied, which seek to address the issue of reducing the
uninsured population within their geographic areas. Preparing the reader to understand
the broader issues surrounding policy change in this issue area, why so little has occurred

at the national level, the power of the various industry groups within the issue area at the

24

 

 

 

 

 

local and state levels, and then the characteristics and problems associated with uninsured

populations has set the stage for the remainder of this research effort.

25

 

Chapter Three: Change Agents, Policy Entrepreneurs and Policy Change

This chapter will present a literature review of change agents, especially
foundation change agents, and of policy entrepreneurs and their role in policy innovation

and change.

Change Agents

The literature on change agents is sparse, and the use of the term “change agent”
is extremely inconsistent throughout the various social science literatures that use the
term (e.g., Schneider, Teske and Mintrom, 1995; O’Gorman, 1978; Havelock and
Havelock, 1973, Perlstadt, et al., 1999). The term change agent is applied to both
individuals (as in Schneider, Teske and Mintrom, 1995; O’Gorman, 1978) and to
organizations (as in Cunningham, 1972; Havelock and Havelock, 1973, Perlstadt, et al.,
1999). Fortunately, the behavior ascribed to both the individual and the organizational
change agent is similar, and it is useful as a starting point to state that it is individual
decision makers within change agent organizations that engage in the described
behaviors, which accounts for some of the similarities in the focus on both individuals
and organizations as change agents.

This portion of the literature review will look at two types of change agents, the
first, and most important for this study, will be of private philanthropic foundations acting
as change agents. The second portion will briefly discuss the literature on government

agencies and their role as change agents.

26

 

 

Foundations as Change Agents

At the end of the 19‘h Century, large foundations in the United States discovered
their ability to influence public policy at all levels of government by funding specific
research projects and through the funding of educational programs (Rabinowitz, 1990).
However, these contributions were relatively insignificant until the period after World
War II, when American foundations began to take on a more activist role, and the
proliferation of foundations as well as the growing endowments of existing large
foundations permitted this sector to reach a critical mass that could be seen to have an
effect as compared to the comparatively much larger amounts of funding produced by
federal and state governments.

A 1995 study of foundation efforts to influence health policy decision making
through conscious funding strategies identified a typology of four types of foundation
strategies for influencing public policy (Knott and Weissert). These strategies were
identified across two dimensions; when a foundation enters an issue area with funding
(early or late), and how long a foundation continues funding within an issue area
(continuous or sporadic). Knott and Weissert contended that these two dimensions
provided a useful typology of foundation behavior, and defined the types of behaviors
consistent with the four types of strategies exhibited across these two strategies. Pioneers
were identified as foundations that enter issue areas early and maintain consistent funding
patterns. Explorers were those foundations that enter issue areas early but were less
consistent in their funding, perhaps because of lower resource levels. Foundations that

enter issue areas late but are consistent funders were termed ranchers. Lastly, foundations

27

that were late entrants into an issue area and had less consistent funding were termed
itinerants, they routinely entered issue areas late, funded for just a few years and then
moved onto another issue area.

Knott and Weissert’s research is helpful in understanding the foundation under
study here, the W. K. Kellogg Foundation, as this foundation was one of the cases
assessed in their study. They found Kellogg to be hard to categorize as a result of some
inconsistent fimding behaviors across the health issue areas they studied. They tentatively
assigned Kellogg the classification of rancher, a late entrant that continuously funds once
it enters an issue area. However, Knott and Weissert also noted that Kellogg appears to be
less of a follower of either other foundations’ funding behavior or of what issues are the
most salient in Congress. Rather, Kellogg’s funding behaviors appeared to be more
driven by assessments of its own priorities, irrespective of others.

In a report to the WK. Kellogg Foundation Perlstadt, et al. (1997) identified seven
broad themes addressed in the literature concerning change agents: systems change,
knowledge transfer, civic engagement, inclusion, decision making, project organization,
and project leadership.

Brian O’Connell defined a set of specific goals of foundations based on his
observations of foundation activity throughout American history. These goals include: (1)
discovering new knowledge, (2) encouraging excellence, (3) enabling of individual and
group potential, (4) relieving misery, (5) preserving and enhancing democracy, (6)
building better communities, (7) nourishing the human spirit, (8) creating tolerance,

understanding and peace, and (9) remembering the dead (1987). O’Connell and other

28

—_;

 

researchers of foundations agree that foundations often combine a number of these roles
when engaging in grant-making decisions (Rabinowitz, I990; Cunningham, 1972).

Shaheen and Perlstadt (1982) made a start on defining the ways in which social
change occurs by noting that two methods appear to be common, an inside-outward effort
and an outside-inward effort. The first is a situation where persons and organizations
within a community empower themselves to bring about change, thus not requiring the
presence of a change agent. The second situation is where social change begins with an
external entity such as a government agency, foundation or philanthropic-minded firm
supplying resources (external or “outside” empowerment) to persons and organizations so
that they may bring about change.12

Perlstadt, et a1. (1997) argued, based on previous research on citizen participation
in regional health planning funded by the federal government (Wandersman, 1984;
Langton, 1978) and evaluation data concerning the planning phase of CCHMs “ that
many federal agency and program administrators are ambivalent about citizen
participation, many citizens who participate in such activities are ambivalent and
confused about their purpose on committees and boards. Hence the general conclusion
that an outside change agent of some sort is needed” (Perlstadt, et al., 1997, 74).

In their article based on their evaluation of the planning phase of the

Comprehensive Community Health Models of Michigan (“CCHMs”) Initiative, which

 

'2 This study proposes looking at three cases of outside-inward social change
and innovation, and one case of inside-outward social change and
innovation.

29

 

supplies three of the four proposed cases for study in this effort) Perlstadt, et al., noted
that there were five antecedents to successful citizen participation in health planning;
First, the external change agent must be willing to play multiple supportive roles.
Second, there must be some attempt to balance power between the dominant
influential stakeholders and the usually voiceless groups coming to the table.
Third, one of the sub-goals is to create a collaborative process for policy
formulation.
Fourth, once the playing field is level, there is a greater likelihood that the
management of the change process will be shared by those involved.
Fifth, the community entity working toward change eventually seeks to sustain
itself through institutionalization. (Perlstadt, et a1. 1999, 78-79; see also Havelock
and Havelock, 1973; Checkoway, 1981; Tye, 1973.)
In defining the role of the change agent, Havelock and Havelock , who studied
federal agencies as change agents found that change agents perform a combination of four

separate roles (1973). These roles are:

1. A catalyst, that gets a system moving and encourages people to develop solutions;
2. A solution giver, that provides an idea of how to approach and resolve a problem;
3. A process helper, that facilitates the learning process for people involved in the

initiative; and
4. A resource linker, that ensures participants to some change process are capable of

using the variety of potential resources at their disposal.

3O

   

The role of catalyst is familiar to anyone studying grant making activities.
Perlstadt, et al., used Havelock and Havelock’s roles in arguing that foundations have as
their main purpose the provision of grants and other assistance to communities to tackle
problems that the foundation perceives exist within the community (1997, also WKKF,
1998). In this role foundations identify a social problem, and either use a competitive
proposal process or purposively select communities that exhibit the social problem and
are determined by foundations to be in a position, with external funding, to address the
problem (1998). Thus the roles of change agents as defined by Havelock and Havelock
(1973) can be seen to apply equally well to both federal agencies and private foundations.

The role of solution giver can be seen as an extension of the role of catalyst. If a
foundation or other change agent as catalyst identifies some social problem and provides
funds to support a process that will address the problem, then it is natural that the change
agent believes its has some notion of what the set of potential solutions are, and may even
tie these solutions into the competitive bid process (Havelock and Havelock, 1973).

The process helper role is one where the change agent engages in knowledge
transfer to the participating individuals/organizations attempts to attack some social
problem (Doty, 1980). Because community capacity to understand the complexities of a
problem is often low, either due to a shortage of local experts or to the concentration of
these experts in one sector, the change agent must play the role of teacher to ensure that
all participants in a community change process have an equal and adequate level of
understanding of the problem being addressed (Doty, 1980). This role is especially

critical when dealing with complex social problems, such as those faced by consumers in

31

 

 

the American health care system (Perlstadt, et al., 1999). In dealing with problems in the
issue area of health care, “... participants in a health planning processes must be given
concrete information on the how-to’s of health care reform. This includes information on
the financing, operation and mechanisms of the delivery system.” (Perlstadt, et al., 1999,
79.) This information must be provided equally to all participants to ensure a comparable
knowledge base, and as a check against elitism (Langton, 1981).

The resource linker role is one where the change agent makes the participants to a
change process aware of the existence of potential resources available both external and
internal to their own community to address the problem. This role may also include some
effort on the part of the change agent to formally introduce local experts within a specific
sector (such as the local health care system) to the change process (Tye, 1973). Thus, the
resources to be linked may be in the form of individual expertise as well as financial.
Individual expertise may be voluntary (a hospital CEO) or paid (health care consultant).
Thus the linkage may come about through encouragement to participate such as a
political linkage agent role (Tye, 1973) or through the funding of consultants (Perlstadt, et
al., 1999).

An aspect of change agent activities that directly affects this proposed research is
the identification of leadership for the community project. According to Perlstadt, et al., a
“... traditional procedure for finding community leadership [for a project] is for a small
group of influentials to approach an individual with prior volunteer activities or service
who would be able to call upon his or her business or organization for staff support, and

encourage him or her to become the project champion (1999, 83).” However, “...

32

_ I

coalitions or partnerships created by granting agencies face the dual challenge of
identifying leaders and creating a local organization since the individuals who might
spontaneously organize in response to specific community problems have not voluntarily
done so” (83). This proposed research will seek to prove that this search for leadership is
in reality a search for a policy entrepreneur who may be supported by the change agent in
the collaborative development and launch of some innovative solution(s) to the social
problem(s) identified by the change agent.

The development of philanthropy in America is deeply rooted in American
political philosophy, specifically American beliefs concerning the importance of
pluralism in society and the maintenance and improvement of its communities, and the
wealthy citizen’s role in supporting, nurturing, and enhancing both (O’Connell, 1987;
Rabinowitz, 1990; Cunningham, 1972; Dickinson, 1962).

Regardless of the ideological bent of the foundation, the goals of foundations tend
to support pluralism in society and the belief that the increased citizen participation that
comes with pluralism would benefit American communities. Thus even those foundations
whose goal is to influence national public policy focus much of their funding efforts
within communities and at the “grass roots” level (Covington, 1997). These long-held and
mutually reinforcing goals have led to a common philosophy (if not common ideas as to
what was “best” for communities) across all types of foundations concerning the
importance of positive social change achieved through projects/programs aimed at
encouraging, enhancing and supporting citizen participation and planning at the

community level (O’Connell, 1987; Rabinowitz, 1990).

33

 

 

O’Connell developed a useful profile of foundation characteristics that distinguish

them from other public and private actors within society.'3 These characteristics are:

1.

Independence. The principal purpose of the foundation community is to represent
a different view of the world, a view that is not hampered by a specific set of
constituents. This allows for more freedom of thought and encourages
experimentation with concepts that would otherwise not be given an opportunity.
Leadership. Foundations may lead with both funding and with the leverage that
their board members have within the communities in which they interact.
Intelligence Function. Foundations tend to specialize within specific issue areas,
and this specialization leads to networking within issue communities, as well as
the development of specialized knowledge concerning the inner workings of
communities and how these issues affect them.

Flexibility. Unlike government actors, foundations may spend their funds in any
manner and on any topic they choose and have the dexterity to augment, reduce or
shift funds as is deemed necessary. Foundations may tenaciously stick with an
issue over time which is more difficult for an actor who must answer to some
constituency.

Innovation and Daring. Foundations may gamble their funding on daring (others
may label them crazy or unworkable) innovations because of their independence

from any particular constituency.

 

'3 Unfortunately as the reader will see, O’Connell’s characteristics overlap,
and are thus not mutually exclusive. However, of the profiles reviewed, his
was the most specific and least prone to overlap or pontification.

34

 

 

 

6. Enthusiasm. Foundations don’t just gamble on daring innovations, they
enthusiastically pursue opportunities to fund creativity, whether scientific, social

or artistic. (1987, pp. 271-274.)

As will become apparent to readers after reading the section on policy entrepreneurs and
policy change, these characteristics of foundations presented by O’Connell demonstrate a
fair amount of congruence with the characteristics of policy entrepreneurs, and we should
not be surprised that we would find these two actors acting in concert. Perhaps we should
only be surprised that research linking the two types of actors has not yet occurred.

The sparse literature that specifically assesses the role of private foundations as
change agents focuses primarily on citizen participation.14 This intertwining of citizen
planning and the literature on change agents, especially foundations, is expected, as
foundations are often interested in addressing problems at a community level (even if the
initiative is open to communities across the nation, because the social problem is believed
to be widespread). It is consistent because the historical chroniclers agree on the notion
that since the earliest stages of private foundation-based philanthropy in America, a core
goal of foundations has included the pursuit of community change through the

democratization of decision making for change (Cunningham, 1972; O’Connell, 1987;

 

The literature on foundations contains little theory, but is rather filled with
thick descriptions of foundation history, funding streams, the effect of
changes in federal tax laws specific to foundation giving, and foundation
activities and only focuses on the proto-theoretic development of the
characteristics of foundations and their behavior. The rigor of these
observed characteristics has not always been high, and several suffer from
the use of categories that are not mutually exclusive, and thus add little to
the development of even proto-theory.

35

—_—_—___4

 

Rabinowitz, 1990). Foundations encourage citizen involvement in planning to improve
existing programs, increase program responsiveness, and gain community acceptance for
programs designed to address social problems (Perlstadt, et al., 1999). This involvement
may also be undertaken in order to restrain experts and policy makers from designing
programs that do not take local opinion and context into account when in the program

formulation stage (Selznick, 1979, Weinstein, 1997).

Government Agencies as Change Agents

Private foundations are not the only organizations that act in the role of change
agent. Government agencies at all levels also sometimes assume this role, and some
government agencies have as a key part of their mission the pursuit of goals that are
typical of change agents.

As discussed above, it is not just resources that are important to outside-inward
change agents, but the flow of relevant information and the encouragement of the growth
of information networks is perhaps just as important. Mossberger and Hale (1999) tie
theories of information diffusion with the role that federal agencies play in
intergovernmental networks that promote federally-funded innovation. Mossberger and
Hale argued that “[F]ederal agencies promote the diffusion of information to subnational
governments to encourage the adoption of certain policies and administrative processes.
[in this activity] the federal government acts as a policy entrepreneur or change agent”
(1999, 2). Mossberger and Hale not only demonstrated that federal agencies were

important primary agents of change with a central role in the formation of

36

 

 

 

 

intergovernmental information networks, but also noted in their mixed use of terms the
overlap between the behaviors of policy entrepreneurs and change agents.

Federal agencies have been directed to act as change agents in a manner somewhat
similar to that taken by foundations with federal agencies’ focus on citizen participation
in regional health planning from the 19605 to the 19805 (Wandersman, 1984; Langton,
1978). In these efforts their activities moved beyond the scope of the deveIOpment of the
intergovernmental information networks described by Mossberger and Hale and took on
more of the roles described by Havelock and Havelock (see previous section).

If we examine the literature on government agencies’ role as change agents, we
immediately note that much of the literature is dominated by the study of federal agencies
acting as change agents abroad. This is not surprising since the US. government funds
agencies whose explicit purpose is the pursuit of social, economic and/or administrative
change and development in countries perceived to lack the domestic resources necessary
to pursue change through an inside-outward approach (Kamel, 1999). Although the scale
of these change agents’ efforts differ from domestic change agents, there exist close
similarities between the conclusions of this research and the research presented earlier.

The US. agency change agent enters into a social or economic change process in
a donee country because domestic resources to support change processes are low,
expertise is limited, and the political resistance to change is high (Kamel, 1999; Lin,
1989; Duncan, 1986; Dabasi-Schweng, 1965). Thus these change agents fulfill the roles
defined by Havelock and Havelock (1973). Especially noteworthy is the similarity

between the efforts of foundation and federal change agents to increase citizen

37

 

____——;— A

 

 

 

 

participation in planning, and foundations’ historical efforts at representing unvoiced or
unempowered citizen interests against established political interests and the efforts of
overseas change agents to overcome political resistance to reform and change amongst
established bureaucracies.

Regrettably, the contribution of this literature is the study of change agent failures.
One key reason cited by Rondinelli (1993) for failure is that the problems being addressed
are intractable and plans for addressing them are not sufficiently long term. This
correlates well with the perceptions of the chroniclers of the history of American
foundations that foundations are better positioned than government agencies to begin,
support and achieve change because of their freedom and independence from external
constituencies, and their ability to plan, implement and stick to long term change efforts.
Change strategies of government agency change agents have also been accused of
focusing on interests that are transient and serve short term political interests (Kamel,
1999). Finally, a key reason for failure often cited in this literature is the change agent’s
inability to aggressively access and develop native expertise and intellectual capacity, '
instead relying heavily on international experts with little knowledge of local contexts
and their associated problems (World Bank, 1992; Dow, 1985). This last distinction in
overseas change agent behavior seems counter to the traditional inclusive and citizen-
directed social planning approach ingrained in the American foundation community and
also exhibited by federal change agents acting within the United States. It is beyond the

scope of this study to assess why the difference exists.

38

 

 

Summary of Foundation and Government Agencies as Change Agents

Foundations in America have a long history deeply rooted in American political
philosophy, specifically American beliefs concerning the importance of pluralism in
society and the maintenance and improvement of its communities, and the wealthy
citizen’s role in supporting, nurturing, and enhancing both. Both foundations and
government agencies may act as change agents, and their roles have been defined as
similar, with the chief difference in the filling of this role being the freedom and
independence of foundations from controlling or constraining constituencies. Research
has demonstrated a set of defined change agent roles, and the assessment of change agent
failures in the literature on overseas change agents presents a rigorous assessment of these
change agents’ violation or failure to fulfill one or more of these roles. Change agents
share many characteristics of policy entrepreneurs, and the literature on change agents
provides an inkling of the link that this proposed research intends to make between these
two types of actors as change agents often seek to locate a policy entrepreneur to lead

innovative local change efforts.

Entrepreneurs and Policy Change

This portion of the literature review is centered around the theory of policy
entrepreneurs, market entrepreneurs, collective teams of entrepreneurs and their role in
policy change. Although there are several theoretical components that explain the

development and implementation of these local level policy innovations to provide

39

 

 

 

coverage to the working uninsured, policy entrepreneurs are central to understanding the
development and implementation of some policy innovations at the local level.

A complication exists in my review of the literature on policy entrepreneurs.
Although political science theories address the inseparability of markets from public
policy (Mintrom, 2000), there has been little recent effort to incorporate ideas concerning
market context and its role in change with political explanations of change.15 Oliver and
Shaheen actually mix these terms and refer to political markets within which policy
entrepreneurs develop innovative products (1997). Oliver and Shaheen are referring in the
first term to state level political networks a la Heclo, and in the second term to policy
innovations and change within state legislatures. Oliver and Shaheen also mention that a
state’s economy can help create a favorable context, but this role is presented as
background and does not include the possibility of market actors’ participation in any
policy change and product development processes, nor does it include any sense that the
market context (local, state or national) may affect change efforts.

In the context of the local level initiatives being studied here, both market and
policy entrepreneurship are required for the successful implementation of the innovation,
because these innovations do not exist solely in either a market or political context. This
is not inconsistent with the literature on policy entrepreneurs, as Schneider and Teske
have called for a synthesis of political and economic theory in order to develop better

theories around the emergence of policy entrepreneurs and the nature of their activities

 

‘5 One notable exception to this is Baumgartner and Jones (1991), who
integrate the effect of market venues on the process of policy change.

40

 

 

 

 

(1992, 93). Similarly, Mintrom draws heavily on economic theories of market
entrepreneurs to develop a milieu of the market entrepreneur (2000). However,
Mintrom’s study was based on policy entrepreneurs in education below the post
secondary level, where we would not expect to find developed private markets, and thus
he did not need to develop a theory that blended the two roles.

The output of a policy entrepreneur is a policy innovation that represents a change
from existing policy. The output of a market entrepreneur is a new product and/or service
(Mintrom, 2000; Schneider and Teske, 1992, 1993). Because the entrepreneurs under
study here must produce two kinds of outputs, we must expect their activities to occur in
the two contexts. Thus either or both the local political and local market context would
constrain or facilitate the ability of the policy entrepreneur to affect change.

Although policy entrepreneurs are incidentally referenced by Baumgartner and
Jones (1991) in their study of policy subsystems and policy change, their theory is
relevant to the issue of how local policy entrepreneurs may go about interacting in both
political and market venues (components of Mintrom’s milieus, introduced below) in
order to achieve change. Baumgartner and Jones presented a theory explaining both
policy stability and punctuated changes in policy within the United States. They argued
that both stability and punctuated change were manifestations of a process that is “the
interaction of beliefs and values concerning a particular policy [the policy image]
with the existing set of political institutions [venues of policy action]” (1045).

Baumgartner and Jones argued that when political images (created by elites and

broadcast to the public through mass media) are positive, government permits and even

41

 

 

 

facilitates industry experts in the development of a policy subsystem that controls policy
related to that industry (e.g., the design and enactment of industry regulations). They posit
that the development of the policy subsystem is an effort to insulate the industry from the
influence of democratic forces. Also noteworthy is their argument that the development
of a policy subsystem is more likely when the issues at hand appear to be technical and
specialized in nature, requiring the turnover of control to persons trained to deal with
these issues. The policy subsystem that is created controls policy activity through the
limitation of the number of venues available for change, and also attempts to maintain a
positive policy image.

Paul Starr’s (1982) seminal (and Pulitzer Prize-winning) socio-historical study of
the practice of health care in America offers hundreds of pages of supporting evidence
related to the issue area of health care to Baumgartner and Jones’ theory on how policy
subsystems develop and persist within technically complex issue areas. Starr recounts in
detail how the complexities of health care service delivery, the complexities related to the
specific delivery of medical care for disease elimination or amelioration, and the
development of America’s mixed financial methods of funding health care service
delivery came in to being.

Baumgartner and Jones argue that policy subsystems may be broken, however, by
the activity of those actors outside of the policy subsystem who disagree with its outputs.
The avenues for breaking this subsystem are the myriad other more receptive venues
available to these actors (sometimes referred to by Baumgartner and Jones as policy

entrepreneurs). These other venues include, but are not limited to, elected officials,

42

 

 

legislative bodies, government agencies, the mass media, other industries, the public at
large, and the market. The process of breaking up a policy subsystem is marked by what
they term a “succession of ‘venue-access events’” (1052). These events result in the
expansion of discussion of issues related to those in the hands of the policy subsystem to
venues beyond the subsystem’s control. They state that “ the job of a policy
entrepreneur is often that of identifying the most receptive alternative venues for the
policy” (1052).

Coincident and interactive with the venue-access events is a change in policy
image. The policy image becomes less positive, and the exposure of images related to the
policy (positive and negative) increases. These two types of events are interactive and
mutually supporting as the increase in the number of venues looking at a problem is
increasing, and journalists are drawn to the issue because of increasing elite attention.

Changes that may weaken a policy subsystem occur largely outside the public eye
in the early stages, with little media attention at the beginning stages of venue expansion
and image change. As mentioned above, the possible venues for weakening the policy
subsystem include state involvement, advances in research and technology and market
activity.

In health care a time bomb for venue expansion was put in place with the
enactment of the Medicaid program, the costs of which are shared by the federal
government and the 50 states. This program was described in the previous chapter, and
the venue expanding time bomb waiting to explode was simply the occurrence of a

sustained and rapid increase in the costs of providing these services. When the Medicaid

43

 

 

 

program was enacted in 1965 health care costs were low and so was the rate of growth of

costs. However, with expansions in Medicaid eligibility by the federal government and
also by many of the states, and the introduction of new, costlier technology, as well as the
increasing cost Shifting from the steady growth of the uninsured (all described in the
previous chapter) led to an increase in program costs beyond what was originally
anticipated. It should come as no surprise, then, that as states, many of which have
balanced budget requirements and fewer resources than the federal government, began to
search for alternative methods of purchasing these services that ultimately served to
weaken the policy subsystem.

Baumgartner and Jones specifically argue that the private market is “a primary
venue for political action” (1051). Decisions made in the private market may influence
the course of public policy, and may play a role in weakening policy subsystems through
the actions of individual investors in an industry. The perceptions of these investors are
shaped by the perceptions and activities of elites, and based on these changing
perceptions, investors alter their investing behavior. For Baumgartner and Jones, market
actors thus follow and act based on the initial outcomes of venue—access events and elites’
changing political images.

The entrepreneurs of this proposed study must then develop arguments that
address the market-driven benefits and costs of the innovation in order to obtain
agreement and the signing of the private sector agreements necessary for the launching of
the innovation. At the same time they must develop arguments for the public policy

benefits of the innovation in order to obtain the changes in public policy necessary for the

44

—_¥

launching of these innovations. They will then carry these arguments to alternative venues
in their search for cracks in the existing policy subsystem. Although this complicates the
explication of my theory, it is an exciting extension of the theory of policy

entrepreneurship.

Characteristics of Policy Entrepreneurs and their Activities

The characteristics of policy entrepreneurs have been discussed by many authors
on the subject, but I will use Mintrom’s detailed description to frame this discussion.
Mintrom’s characterization is consistent with Oliver and Shaheen’s (1997) description of
policy entrepreneurs at the state level, with the behavior of policy politicians at the
national level as described by Heclo (1978), and with the description of local
entrepreneurs in Schneider and Teske (1992, 1993). According to Mintrom, a policy
entrepreneur must have social perceptiveness, social connectedness, be able to creatively
frame social problems and to develop widely acceptable policy innovations.

Mintrom describes the characteristic of social perceptiveness as the ability to '
listen to others and to understand their needs. Mintrom argues that this characteristic is
the starting point for entrepreneurial creativity. Entrepreneurs gain insight on a policy
problem by listening to others and using this information to interpret and understand the
problem in the broadest possible terms. After listening to the concerns of a variety of
actors, the policy entrepreneur thinks creatively about policy in the issue area, and also

thinks about what motivates these actors in their pursuit of policy goals.

45

 

 

Social connectedness corresponds to Oliver and Shaheen’s criterion of
institutional access, although is somewhat broader (1997). Social connectedness
(institutional access) is the extent to which a policy entrepreneur is already connected or
can become connected through social or political ties to the key players in the issue area.
Mintrom argues two benefits arise from high levels of social connectedness: one is the
ability to detach oneself from a particular view on an issue and view it in an objective
fashion, the other is entrepreneurs have opportunities to present arguments for policy
change to all or most of the relevant actors within a policy issue area.

The ability to detach from a particular viewpoint on an issue is critical to the
development of new thinking concerning the social problem. Mintrom specifically argues
that “... [T]he likelihood that an entrepreneur will perceive opportunities for gain where
others do not will be enhanced by his or her movement across various social and
professional communities (2000, 126).” Oliver and Shaheen (1997), support this idea by
arguing that entrepreneurs, unlike others, are not afraid of change, but rather see change
as natural and positive.

The ability to creatively frame social problems in a compelling manner and then
offer a solution is Mintrom’s third characteristic of the policy entrepreneur. Mintrom
characterizes this ability as problem solving and the search for solutions. Policy
entrepreneurs are unwilling to accept business as usual solutions, but are rather
adventurous, holding their skepticism in abeyance while they incorporate information
from the variety of social realms within which they operate (2000). After arriving at a

policy innovation designed to address some social problem, the policy entrepreneur

46

¥——_—‘

 

creatively frames the arguments supporting the development and introduction of the
innovation to the variety of interests within the scope of conflict.

Mintrom observes that the perception of problems and solutions is indicative of a
wider phenomenon, that both problems and solutions are socially constructed (also Cobb
and Elder, 1983). He notes that this line of reasoning holds implications for the way in
which a policy entrepreneur must go about framing problems and solutions. Socially
constructed problems and solutions may be reframed through argumentation and the re-
presentation of problems (2000). Thus refraining an argument requires the entrepreneur to
be able to understand the frames of her/his different audiences, and explain a problem and
offer a solution that is consistent with their audiences’ frameworks. Therefore an
entrepreneur must have the ability to develop a solution to a problem that is transportable
across these different frameworks, yet remains consistent as it is presented across flames.

For the cases within this study, the entrepreneur must be able to frame a solution
in terms of risk pools, reimbursement and organizational mission to health care providers,
and then transport the same solution to local and state public officials as a solution
understandable in terms of the redirection of public funds and the possibilities of using
those funds to leverage other funds to provide coverage to their Lurinsured constituents.
The entrepreneur must then be able to develop a marketing plan for the product that
convinces firms that purchasing this low cost product will produce healthier, more
productive workers, will increase workforce stability through retention (thus reducing the

costs associated with staff turnover) while minimizing the cost of providing the benefit to

47

 

 

 

their employees. Lastly, the product must be viewed by the employees of the target firms

as worth the amount of their contribution to the product’s premium.

Collective Entrepreneurship

If we reflect on the set of entrepreneurial skills and assets outlined above, and
recall the complexity of the issue area and the groups of industry and policy actors
involved as described in Chapter Two, the potential that a single entrepreneur would
possess all of the necessary skills and assets described across the issue area and the
involved actors would seem unlikely. Roberts and King argue that, “As issues grow more
complex, constituencies more diverse, and change more discontinuous and radical, we
expect groups to supplant individuals as the primary unit of analysis [of entrepreneurial
activity]” (1996, 162). This would certainly be the case for an entrepreneur seeking to
develop an innovative program or product that may (or may not) exist in both the political
and market realms. Developing and launching an innovation that will extend health care
coverage to the uninsured requires some level of knowledge in the following areas:
health, health care, service delivery, financing, risk pools, reimbursement schedules,
formularies, the characteristics of the uninsured, and the characteristics of firms that do
not regularly provide health care coverage. Finding one individual with all of this
knowledge who also has the characteristics of an entrepreneur would be an arduous, if not
impossible task. Thus we would expect that in place of an individual, we may instead see

a team of individuals, perhaps led by one who is primus inter pares, attacking the

48

 

 

 

 

problem and pooling their knowledge, expertise, argumentation skills and social
connections to achieve some common end.

Katzenbach and Smith (1994) defined the look of such team as follows, “... a
small number of individuals with complementary skills who are committed to a common
purpose, performance goals, and approach for which they hold themselves mutually
accountable” (45). A collective entrepreneurship team that is complementary and diverse
avoids “groupthink”, promotes creativity, and is better able to build a winning coalition
from diverse constituencies (Roberts and King, 1996). Roberts and King further argued
that such a team, “... needs support from the community to survive battles with
entrenched interests” (45). It is useful when thinking of linking external change agents to
local-level policy entrepreneurs to also take into account that Roberts and King warn that,
“It is not clear who answers the phone when entrepreneurs call in less resource-rich
communities” (45).

This is because innovation requires resources as coalitions will need to be built to
support the drive for change (King, 1990; Kanter 1988). Communities with low capacity
will likely have few people around with the necessary skills and characteristics of policy
entrepreneurs (Roberts and King, 1996). These innovators require support in order to
devote significant portions of their time to entrepreneurial tasks. Interestingly, Roberts
and King believe that when pursuing radical changes, the policy entrepreneurs are likely
not politicians, because elected officials cannot be single-minded enough to pursue
change. Thus, someone outside of government is better placed to focus on developing and

launching radical change efforts than elected officials. Again, the literature on policy

49

 

 

 

entrepreneurs unsuspectingly dovetails with the literature on philanthropic foundations’

activities.

The Role of Information

Schneider and Teske discuss the role of information for both market and policy
entrepreneurs. For market entrepreneurs, the ability to keep information private is
paramount, because this aids the restriction of competition (1992). They note that,
“Entrepreneurs glean some new piece of information from the environment and then bet
on the future based on their insight. However, once this information becomes public, the
entrepreneur loses the competitive edge and entrepreneurial profits are bid away” (1992,
740)

However, Schneider and Teske struggle with converting the role of information
for market entrepreneurs into the role it plays for policy entrepreneurs. They argue that
“Most public sector entrepreneurs need to garner votes or support in public opinion to
succeed” (740). Citing Doig and Hargrove (1987), they further note that “building and
maintaining a political coalition is critical to the success of a political entrepreneur, and
revealing information and ideas is central to this coalition process” (740).

In this struggle to understand the role of information for policy entrepreneurs
Schneider and Teske seem to back away from the natural but radical conclusion that the
role of information is reversed for policy entrepreneurs as compared to market
entrepreneurs. However, based on their conceptualization, no other conclusion seems

reachable. Schattschneider (1975) argued that there is always a struggle between the

50

“privatization” (narrowing of the scope) and “socialization” (widening of the scope) of
conflict, and this was in part a function of visibility (16). Thus policy entrepreneurs must
publicize some problem and then offer an innovative policy solution that they believe will
receive broad support, broad enough to permit the adoption of the innovation in the face
of opposition. Bachrach and Baratz explicitly supported this argument when they argued
that the second face of power is the ability of a community’s power elite to restrict the
kinds of policy changes that may be brought up for discussion (1962). Mintrom’s
emphasis on the communication skills and social connectedness (detailed above) is yet
another argument that it is the dissemination of information, not the guarding of
information, that is one of the keys to the success of a policy entrepreneur.

When considering the issue of health care, information is a difficult resource to
obtain but critical (Perlstadt, et al., 1999). Rice (1998) argued that information on all
aspects of health care markets is difficult to access, and once accessed, is difficult for
anyone who is not an expert to understand. Compounding the problems of accessing and
using information is the fact that for almost all consumers of health care services, services
are actually purchased by an intermediary (either by an employer or by the state or federal
government), and then in most cases, some other organization mediates between the
providers of care and the consumers, e.g, an HMO, a third party administrator, or some
similar type of insurer/health plan. Thus the purchasers of health care are not the
consumers, and a number of organizations and persons must access and use information,

and then act collectively based on their understanding of the information in order to effect

51

 

change. Rice, probably without knowing it, highlighted the different roles of information
for market and policy entrepreneurs as applied to the issue of health care.

Thus we would expect that a policy entrepreneur would look to the health issue
area and incorporate ideas concerning the benefits of managed care versus fee for service
care, an understanding of the likely makeup of the target population of both firms and
their uninsured workers, as well as the impact of changes in Medicare and Medicaid in
order to demonstrate how his or her innovation fits into and complements the existing
policy and market environment. These arguments would be tailored to the audience,
stressing the aspects of the innovation that are salient to each. It should also be expected
that aspects of the innovative idea would change as the policy entrepreneur goes about
making arguments to different audiences. These changes would be made to satisfy the
specific concerns of an audience, and must be made in a manner that may be reconciled
with the concerns and needs of the other audiences.

The Importance of Context

Mintrom (2000) and Oliver and Shaheen (1997) (and to a lesser extent Schneider
and Teske, 1992, 93) highlight the importance of context to the potential entry of a policy
entrepreneur into a policy arena. Mintrom argues that context matters and develops his
typologies to demonstrate the possible rnilieus (i.e., contexts) within which we could
expect to see policy entrepreneurs engaging in the development of policy innovations.
Mintrom argues that no policy entrepreneur is a free agent, because there are limits placed
on what s/he is able to do (specifically his matrix on the possibilities of introducing

policy innovations) and these limits are based on institutional structures for policy

52

 

making, policy settings, and the behaviors of the groups and individuals around them.
This milieu shapes the possibilities of action by a policy entrepreneur and allows us to
develop expectations about the possibility of generating support for policy innovation.

Oliver and Shaheen argue that one alternative conclusion of their study of state
level policy entrepreneurs is that the external determinants of policy innovation are more
powerful than the internal determinants. However, Oliver and Shaheen argue that the
leadership role in policy innovation is understated.l6 While accepting their argument that
leadership is important, because the entrepreneurs under study for this research must act
in both policy and market realms, it is preferable to note at this time that context is
exceedingly important. As well, Oliver and Shaheen mention the role of a state’s
economy in setting favorable conditions for policy change that requires resources,
relegating the role of the economy essentially to that of shaping a state’s tax base (1997,
738).

For Mintrom, the political context depends on the openness of political venues
and the pace of change within the issue area, whereas for Wilson context is dependent on
the distribution of costs and benefits of the policy under consideration. Oliver and
Shaheen expanded on Wilson’s typology and argued that policy entrepreneurs operating
within Wilson’s typology attempt to dilute concentrated benefits and create new benefits

or concentrate existing diffuse benefits to build winning coalitions (1997). Schneider and

 

‘6 Oliver and Shaheen’s use of the term “leadership” is not entirely consistent
with the use of that term in political science. Their use of the term appears
to reflect leadership behavior on the part of a policy entrepreneur, not
leadership as defined in the work of Kaufman (1981) or Behn (1991).

53

 

Teske’s organizational milieu was posited within a theory developed to understand local
policy entrepreneurs (1992). Baumgartner and Jones similarly argued that the outcomes
of political debates vary with the arena or venue (1991). Wilson focused primarily on
federal policy change, while Mintrom examines policy change at the federal, state and
local levels. This proposed study will, however, examine policy change primarily at the
local level, with a lesser emphasis on policy change at the state level. As mentioned
earlier, the innovation will consist of both policy change and the creation of a health care
coverage program or product. Thus the context which must be understood to begin to
predict the possibilities of the successful introduction of both policy change and a new
health care coverage product is broader than either Wilson’s context, or Mintrom’s
context for the policy entrepreneur, but must also include the contextual factors in
Mintrom’s milieu of the market entrepreneur; i.e., the pace of change within the industry,
and industry concentration (2000).

This combining of the market and political contexts to understand the possibilities
for the introduction of policy innovation and a product for market are not new.
Schattschneider argued that both the political and market contexts determine the scope of
any conflict, and the extent to which the scope of conflict may be expanded (1975). Thus
a policy entrepreneur seeking to introduce a policy innovation and a product for market
must consider a whole host of issues, including; the structure of the market, the extent to
which the market is undergoing change, the openness of political venues, and the pace of

change in both the issue area and the affected industrial sector. All of these factors affect

54

 

the ability of the local level entrepreneur to succeed in widening the scope of conflict and

building a winning coalition.

The Market Entrepreneur

Mintrom developed two typologies of the entrepreneur, one derived from his
observations of the behavior of market entrepreneurs, the other from his observations of
the behavior of policy entrepreneurs (2000). As with Wilson’s earlier typology (1973),
explicit in both of Mintrom’s typologies is that context affects the ability of both market
and policy entrepreneurs to introduce innovations. Mintrom’s emphasis on milieus
(contexts) is echoed by Schneider and Teske, who emphasized the importance of a
community’s organizational milieu within which the activities of potential policy
entrepreneur occur (1992). Like Mintrom, Schneider and Teske emphasize collective
action restraints as the key feature of the organizational milieu.

In the milieu of the market entrepreneur, Mintrom observes that two contextual
factors affect the opportunities available for profitable market entry, the pace of change
(innovation) within the market, and the concentration of firms within the market. In
Schneider and Teske’s article on antigrowth entrepreneurs, they note that the local growth
machine holds a privileged position vis-a—vis local politicians in part because the market
for businesses in the grth machine is not primarily local (1993). Within health care
markets the reverse is less true. The market for health care services for health care
providers (hospitals and physicians) tends to either be entirely local, or primarily local,

depending on the size of the local hospital and the availability of similar services in

55

surrounding communities, as well as the geographic proximity of the majority of persons
using services to providers of the services. This privileged position should be expected to
vary considerably across local markets, such that in highly urbanized areas where tertiary
care hospitals exist as well as the very specialized high end procedures are more
prevalent, the privileged position of health care firms will approach (but not reach) that
described in literature. In local health care markets where these functions are absent, the
privileged role will be considerably weakened. Thus this aspect of the external privileged
position is variably reduced within health care markets.

However, one wrinkle to this understanding of the market players with a
privileged position is raised with respect to the need to have large firms represented at the
table when a local-level entrepreneur is attempting to build a winning coalition. If a
substantial minority of the community’s privately-insured population is employed by
firms that are controlled outside of the community, then any effort to engage these market
actors represents an extreme version of the privileged position described by Schneider
and Teske, because the growth machine is not entirely local, but parts of the local growth

machine may in fact be controlled from outside the community.

The Policy Entrepreneur

Mintrom (2000) also developed a milieu for the policy entrepreneur based on his
observations of the milieu of the market entrepreneur and on Wilson’s typology of the
politics of policy issues (1973). This is consistent with Schneider and Teske’s

‘organizational milieu’ which is based on the idea that there exist varying possibilities for

56

 

successful collective action (1992). Mintrom’s typology of the constraints faced by a
policy entrepreneur is based on two contextual factors that affect the ability of policy
entrepreneurs to enter the policy arena and introduce policy innovations. The first
contextual factor is the pace of change within the policy milieu and could be measured
“... in terms of citizens’ and interest groups’ preferences, and the introduction of public
sector management reforms in and around a given jurisdiction” (2000, 118). The second
contextual factor is the number of venues for political participation. Mintrom considers
the venues where participation may occur to consist at all levels, from town halls to state
and federal legislatures. Mintrom described the openness of these venues as a continuum
stretching from direct democracy with full and equal participation by all voters to a milieu

where only a few powerful interest groups interact across the venues.

Combining the Roles, and the Use of Information

By drawing on the policy entrepreneur literature and the literature on policy
change presented so far, it becomes possible to consider how a policy entrepreneur may
combine these roles. Mintrom (2000) has argued that the roles are similar, although as
Schneider and Teske tentatively point out, the role of information is dissimilar. However,
Schattschneider (1973) and Baumgartner and Jones (1991) have argued that along with
traditional political venues for achieving policy change, the private market must also be
considered a venue for action. Thus postulating the combination of these roles for this

research is a natural extension of previous research.

57

For both market and policy entrepreneurs, the issue of the Openness of venue
(represented by policy networks or industry concentration) constitute one of two
constraints on their activities. The other constraint relates to the pace of change within the
issue area (whether market or political). Thus it becomes relatively simple to postulate a
set of expectations for entrepreneurs who must act in both realms, since the postulated
constraints for both milieus are the same. We would expect that if venues and pace of
change in the policy milieu present opportunities for a policy entrepreneur, but the market
venues and pace of change present no opportunities, then the entrepreneurs under study
here would be expected to fail. However, the discussion of the interaction of different
policy venues by Baumgartner and Jones (1991) suggests that such a disjunct between a
policy issue area’s milieu and its corresponding market issue area’s milieu is unlikely. In
fact, we would expect that these two milieus would manifest similar constraints and/or
opportunities for entrepreneurs at coincident moments in time, based on their argument
that these venues are synergistic and mutually reinforcing. Adding Baumgartner and
Jones’ terminology to Mintrom’s milieus, it can be argued that since these two sets of
venues are interactive and affect each other, the openness of each venue in the milieus as
well as the milieus’ pace of change should be expected to covary.

Thus the only concern is the difference in the use of information, tightly
controlled by market entrepreneurs, but disseminated widely by policy entrepreneurs.
Fortunately, the literature on market and policy entrepreneurs does not suggest that the
use of infOrmation by entrepreneurs should be expected to change across the

combinations of constraints and/or opportunities within either milieu.

58

——___4__4

The matrix presented in Table 3 illustrates the information use problem. The
entrepreneurs to be studied operate in the mixed milieu, so I postulate that they will
guardedly disseminate information as they go about building a winning coalition that may
then develop and launch an innovative program or product. Guarded dissemination would
represent the careful dissemination of information in the coalition building stage, perhaps
by using causal arguments that do not stipulate the specific design aspects of some
program or product to address the problem, but rather discuss the aspects of the problem,
and suggest a number of possible approaches to solving or ameliorating the problem.
Once the policy entrepreneur has built what is believed by her/him to be a winning
coalition, the entrepreneur would then engage in more closed discussions with the
minimum number of partners determined to be necessary to the successful development
and launch of the program or product.

This careful information-use tightrope suggests that the policy entrepreneur must
possess very high levels of the communication skills described above. Guarded
dissemination could also be expected to draw heavily on the entrepreneur’s existing '
social connections, such that the entrepreneur may discuss ideas, problems and solutions
with a wide number of influentials, believing these discussions will be held in confidence.
It also suggests that the entrepreneur may have to engage in more “politicking” than is
described in the literature, as more of the discussions that are a part of the coalition
building process may have to occur in private, one-on-one settings. Thus the task of
combining the roles should be expected to be difficult, more difficult than would be

expected when separately performing either role.

59

 

Costs and Benefits

In addition to the manner in which context affects the ability of a policy
entrepreneur to engage in policy innovation, the notion of the personal costs and benefits
to entrepreneurial activity is also addressed in the literature. Ricketts (1987) and
Schneider and Teske (1992) argue that benefits and costs affect the rate at which
entrepreneurs are attracted to such activities, as they have skills and talents that could be
employed elsewhere. Ricketts argued that the rate at which entrepreneurs are attracted to
the local political environment is a function of the costs they face in entering the political
arena and the benefits they garner if they succeed as political entrepreneurs (1987).
Schneider and Teske address this issue by stating that “... costs are a function of the
collective action problems [context] and the ease with which problems can be solved.”
(1992). “... benefits (or “profits”) entrepreneurs reap are a function, among other things,
of the budgetary slack of their local community, which affects the entrepreneur’s ability
to reallocate resources to achieve the policy goals held by the entrepreneur” (1992, 73 7).

Unfortunately, Schneider and Teske’s description muddied the understanding of
benefits by describing a part of the function of benefits without adding to our
understanding of what the benefits are. Oliver and Shaheen briefly address this issue by
stating that“... leaders [policy entrepreneurs] likely seek a different combination of
“profits” from innovation: substantive experts may be more interested in the “purposive
benefits” of solving problems and advancing an ideological agenda, while strategic
experts may be more attracted to the “material benefits” of winning elections or

advancing their prestige and power in the political arena” (1997, 745).

60

—¥

So by extrapolating from Ricketts and Oliver and Shaheen that a successful policy
entrepreneur will reap benefits from success, then we can hypothesize these benefits
would be based on the chosen career path of the individual. For politicians the benefit of
successful entrepreneurship is an increased probability of re-election, and increased
influence with other local governing bodies and among the local political/power elite
(Kingdon, 1995). For a civil servant, such as a city manager, the benefits of success will
be increased prestige and beliefs of his/her abilities that could result in opportunities to
obtain a similar position in a larger jurisdiction or continuance in office. For an
entrepreneur in the business sector, the benefits could be higher profits, as well as
increased influence among business peers within the same sector and/or within their
community. For other entrepreneurs, such as directors of nonprofit organizations, we
could expect that the benefits would include increased influence in the local policy
community, increased ability to attract grant funding from foundations and government
agencies, as well as the potential for better job opportunities within and outside of their

community.

From Whence do Policy Entrepreneurs Appear?

The preceding discussion of the costs and benefits of acting as a policy
entrepreneur begs the question, “Where do policy entrepreneurs come from?” Most
discussions in the literature begin with Kingdon who argues that policy entrepreneurs can
come from anywhere, although many of the examples he provides are from either the

legislative or executive branch of government (1995). Kingdon argues that policy

61

entrepreneurs exist essentially everywhere, and are constantly attempting to tie policy
innovations to existing policy problems, often without success. Policy entrepreneurs are
most successfiil in tying solutions to problems in moments of crisis, which are what
Kingdon calls “windows of opportunity” for policy innovation and change. Citing one
respondent, Kingdon wrote: “You [the policy entrepreneur] keep your gun loaded and you
look for an opportunity to come along. Have idea, will shoot (183). ” (Emphasis in
original.)

Kingdon’s research was targeted at policy change at the national level, so to
understand where local level policy entrepreneurs come from we can turn to Schneider
and Teske. Schneider and Teske argued that at the local level, “... there is a population of
potential political entrepreneurs distributed across local governments. The size of the
local population with entrepreneurial skills and ambitions is a function of the
characteristics of the community, such as its income and education level” (1992, 737).
This reasoning is appealing, although it appears that Schneider and Teske thereby restrict
local level policy entrepreneurs to elected officials and government officials. This is too
restrictive and inconsistent with broader notions of who may play the role of entrepreneur
at the local level (Mintrom, 2000). Regardless, Schneider and Teske agree with Kingdon
that policy entrepreneurs exist essentially everywhere, and these entrepreneurs stand
ready to take action, based on their perceptions of the costs and benefits of taking action
at any given time on any given problem issue. Most interesting about their reasoning is

the contextual issue they raise concerning a community’s income and educational level.

62

This raises questions about the possibilities of finding a policy entrepreneur in smaller

communities, as well as in poorer communities (be they urban or rural).

Summary

The preceding discussion has highlighted the relevant research and provided a
conceptual framework for this research study of change agents and their support of policy
entrepreneurs at the local level who, in the presence or absence of a foundation change
agent, develop and launch a policy innovation that attempts to extend coverage to the
working uninsured. What should be pointed out is that there are similarities in the
characteristics and activities of both change agents and policy entrepreneurs, which
probably explains the application of the term “agent of change” to policy entrepreneurs by
some researchers (e.g., Schneider, Teske and Mintrom, 1995). In retrospect it is perhaps a
little surprising that the linking of the potential interactions between philanthropic
foundations acting as change agents and policy entrepreneurs has not been previously
studied, given this congruence. Regardless, these two types of actors appear to have some
congruence of characteristics and activities, including the perception of possibilities or
innovations not previously given air time, separation from elected constituencies, and a
willingness to attempt to bring these innovative possibilities into the real world.
Foundation change agents support community-level change through citizen participation
and planning, and policy entrepreneurs go about change through the building of coalitions

broad enough to support innovation.

63

Based on the preceding literature review, we would expect that a change agent
will provide the resources necessary for (1) a policy entrepreneur to devote a significant
proportion of her/his time to the development of a program or product and (2) to the
provision of the staff and information structure to support what will become a full time
information gathering, innovation search and development, and intensive political and
market lobbying effort culminating in the launch of the innovation. These resources are a
necessary component because only a well-supported policy entrepreneur who may
dedicate significant amounts of his or her time can succeed in cracking the closed
political venues that dominate the health care policy subsystem at all levels. The change
agent will attempt to ensure the flow of pertinent information to the policy entrepreneur
in terms of information sharing on possible policy innovations, networking opportunities
with like-minded policy entrepreneurs and health policy experts, and the potential
facilitation of entry into state-level policy networks. The change agent will also provide
the policy entrepreneur with sufficient discretionary spending to permit the hiring of
expert consultants in the relevant areas of health care policy, financing and marketing.

Thus this study is of policy entrepreneurs who have in some cases emerged from
the activities of a foundation change agent that acts as a catalyst, solution giver, process
helper, and resource and information provider, transferring its legitimacy to the policy
entrepreneur as s/he goes about building a team of collective entrepreneurs, developing an
innovation, building a coalition (often a public private-partnership with broad citizen
participation) for the support of the innovation, developing a product, and marketing the

product to the local employer market.

64

 

These policy entrepreneurs have had to operate within two different but
overlapping milieus or contexts, one market and one political. These two contexts
constrain the use of information by the policy entrepreneur in different and sometimes
contradictory ways. In the policy context, the use of information is marked by spreading
the innovative concept as far as possible. In the market context, information on the
innovation is tightly held. It has been proposed here that entrepreneurs Operating across
both contexts must guardedly disseminate information on their innovation.

These entrepreneurs must have existing social connections across the potential
partners, strong social skills to build new relationships, strong communications skills to
support their coalition-building efforts, flexible, and open-minded thinking to enable
them to conceptualize an innovation that can fit into complex political and market

environments.

65

 

 

Chapter Four: Research Design and Methodology

Introduction to the Research Design

The research design uses a case study approach. The case studies will be used to
inform the development of heuristic decision models to define and assess the strength and
role of constraints on the change agent and local policy entrepreneurs in the study. This
study examines four county-level policy initiatives within Michigan to develop a better
understanding of the applicability of these theories to the sub-state level, and to develop
suggestions for a general theory of change agents and policy entrepreneurship at the sub-
state level. This research study examines the role of two change agents. The role of the
Michigan Department of Community Health (MDCH) as a state agency change agent in
all four sites, based on the role elicited in the literature review, is assessed. Senior
Officials within MDCH were approached by members of the collective team of
entrepreneurs in three of the four study sites to act as a state agency change agent. The
role of the W. K. Kellogg Foundation (WKKF) acting as a foundation change agent, as
elicited in the literature review, is also assessed. WKKF supported activities related to a
broad change initiative in three of the four study sites.

Two decision models are developed to enhance interpretation both within cases
and in the cross-case analysis through the illustration of the constraints imposed by the
local political and market context. These contextual constraints are modeled in the ex
p031; facto probabilities of the paths within the model, since the probabilities are

hYpothesized to change as a result of changes in context.

66

Case Study Design

This research was conducted using a case study design based on the methods
developed by Yin (1994). Yin differentiated the case study approach from other
nonstatistical approaches, arguing that “... you would use the case study method because
you deliberately wanted to cover contextual conditions-believing they might be highly
pertinent to your phenomenon of study” (13). The literature review has highlighted the
importance of market and political contexts on policy entrepreneurship at the local level.
Since a goal of this research is to disentangle the relationship between the economic and
political context constraints on policy entrepreneurship, the case study is the appropriate
research method.

This research is organized using Yin’s five components: 1) study questions,
2) propositions, 3) unit(s) of analysis, 4) links between data and the propositions, and
5) criteria for interpreting the findings.

Four case studies were conducted of situations where policy entrepreneurs were

provided the opportunity (or seized the opportunity) to develop and launch an innovative

67

program or product through the formation of a coalition of organizations necessary to

make the program or product a reality.

17,18

Three of the case studies were originally promised funding over a seven-year

period by WKKF as part of a single multi-site initiative. The initiative, known as the

Comprehensive Community Health Models of Michigan Initiative (CCHMs, pronounced

as “chums”), funded separate projects in three different counties in Michigan. These

projects were mandated to begin a community dialogue on health and health care within

their communities, a dialogue that would incorporate representatives from the purchaser,

payer and consumer communities. The forum for this dialogue was to be a governing

board made up of representatives from each of the three groups in equal proportion.

Without explicitly stating so, the Foundation was attempting to force the creation of an

alternate venue for action outside of the existing local health policy subsystem. Among

the specific charges of WKKF to the projects and their respective governing boards was

 

The innovative program or product expected is health care coverage for
the working uninsured. Because the bulk of health care coverage is
employer-based, i.e., firms most often originate the purchase of coverage
(even though employees are often expected to contribute either through
copays, premium contributions or both), the coverage would be designed
and marketed to employers that do not traditionally offer such coverage.
This health care coverage may be developed and launched as a separate
program offering a product, or a new product may be developed and
launched within an existing program.

The terms “partnership” and “coalition” have been used interchangeably
because the change agent literature concerning foundations uses the first
term most often, while the political science literature on policy
entrepreneurs tends to use the second term.

68

the need to explore and develop ways to expand health care coverage to the uninsured
within their communities.

The CCHMS cases in the study include Calhoun, Muskegon and St. Clair
counties. The start dates of the Calhoun, Muskegon and St. Clair processes were slightly
staggered, with Calhoun beginning in 1993, between eight and twelve months prior to
Muskegon and St. Clair, both of which began in 1994. Each was funded for the same
length of time, although late in 1998 St. Clair’s funding was cut short due to major
shortfalls in Kellogg Foundation revenue from its invested endowment.

The fourth case study is of Wayne County, Michigan, where a program
(“HealthChoice”) was developed and launched that provides an employer-based health
care coverage product targeted at firms that do not traditionally offer such coverage to
their employees. This innovative program is similar to the type innovation that was one of
the goals of the CCHMS initiative — moving toward community-wide coverage, but was
not directly funded by a foundation or other outside change agent. (The Muskegon
CCHMS Project adopted many aspects of the program and product developed in Wayne
County.) The Wayne County case is included as an example of inside-outward change to
contrast with the other three cases where a foundation change agent initiated an outside-
inward change effort. The Wayne County innovation did use state and federal funding
mechanisms to aid in the launch of its health care coverage program and product.
HealthChoice was launched in January 1994 following a number of changes in state

funding for indigent health care services.

69

 

 

 

 

Study Questions

1.

How [or did] a foundation play the role of change agent to encourage a local
partnership that will, through the development and launch of some program or
product, extend health care coverage to the uninsured?

How [or did] state government play the role of change agent to encourage the
development of a local partnership that will, through the development and launch
of some program or product, extend health care coverage to the uninsured?
What persons filled the role of local policy entrepreneur, and how did their
placement within the local policy and market contexts (i. e., their social position
and employment role) affect their ability to pressure for the development and
launch of some program or product to extend health care coverage to the
uninsured?

How [or did] local policy entrepreneurs go about building a collective
entrepreneurship team to permit the development and launch of some program or
product to extend health care coverage to the uninsured?

How [or did] the local policy entrepreneur alter the structure of costs and
benefits for those members of the partnership? Did the local policy entrepreneur
alter the members ’ perceptions of cost?

How [or did] the political and market contexts affect the ability of local policy
entrepreneurs to build a partnership around the issue of the development and
launch of some program or product to extend health care coverage to the

uninsured?

70

 

7. Could programs or products developed at the local level in Michigan (and models

of how these came about) apply beyond Michigan?

Propositions and Units of Analysis

Change Agents
Proposition #1: A change agent is a necessary (but not sufficient) precursor to the
development and launch of the county-level program or product in counties where
there exists no “slack” in resources, or where the local health policy subsystem is
weak enough to be opened with outside assistance. In this situation, a foundation
and/or state government may both become a factor in the development and launch
of the program or product.
Units of Analysis: Change Agent(s), County

As stated in Chapter Two, although it has been argued that health care is a local
matter, the resources (both private and public) available to extend coverage within a
given locality to currently uninsured workers are scarce. It is the charge of hospitals to
serve persons in need of emergency care regardless of their ability to pay. However, the
recent strains on the health care system documented in Chapter Two, (changes in federal
funding primarily related to the 1997 BBA and changes within the market brought on by
the increased use of managed care arrangements), increased purchaser demands for cost
reduction or reductions in the rate of cost increases, and in Michigan, the effects of state

purchasing of Medicaid services through capitated contracts with managed care

71

 

 

 

 

 

organizations have all dramatically reduced the flow of financial resources to hospitals
and health systems, in many cases driving these organizations into annual budget
shortfalls. This situation means that any local effort to develop and launch a program or
product intended to extend health care coverage to the working uninsured will require
accessing funds from outside the geographical boundaries of the community.

In addition to the need for external resources to support a program or product, the
resources necessary to support a process of coalition building, and the technical aspects of
the development and launch of the innovative program or product suggests that a change
agent often must exist. A change agent should anticipate that it may be able to provide
targeted resource inputs to encourage specific innovations, provided the local policy
subsystem has been weakened enough to not inhibit a change effort. In communities
where there exist no resources for the pursuit of innovation, but where the policy
subsystem is weakened, a change agent can be the critical element that ‘jump starts’ the
innovative activities of a policy entrepreneur to develop and launch an innovation. In this
role the change agent may act as a catalyst, provide solutions, act as a process helper, and
resource linker. However, in communities where the policy subsystem has not been
weakened, it should be anticipated that no amount of externally-provided resources will
stimulate innovative activity that will not be stymied by the members of the policy
subsystem.

Lastly, the change agent plays a role in the selection of the site for the initiative.
The 5 election process should depend heavily on developing and applying a set of decision

rules and criteria intended to identify counties with political and market contexts that are

72

 

 

 

 

favorable to the creation of alternate venues for action outside of the local health policy

subsystem. A county with a favorable context would be one where the contexts have
weakened the local health policy subsystem to the point where it can not inhibit the
development and launch of the desired innovation.

Based on the literature review, we would expect a foundation change agent to take
on the four roles defined there. In the case of the role of state government as a change
agent, we would expect it to take on the role of information networker and resource linker
described in the literature review, especially as defined by Mossberger and Hale (1999).
See Table 4 for a visual presentation of the change agent roles assumed by foundations
and state agencies

According to the literature review, the change agent should play a role in the
development of the criteria used to identify a project leader, which this research intends to
prove is a person who should be able to combine the roles of policy and market
entrepreneur if the chosen leader is to succeed in the development and launch of the
desired innovation. The process for locating the entrepreneur may be expected in some
way to alter the pool of potential policy entrepreneurs vying for the position. Thus we
must examine how the change agent involved itself in the choice of project leader in order

to understand to what extent the pool of entrepreneurs is constrained by this process.

73

 

 

 

 

 

Policy Entrepreneur
Proposition #2: A local policy entrepreneur team is a necessary component to the
development and implementation of the policy/market innovation.

Unit of Analysis: Policy Entrepreneur, County

In order for the local partnership to develop and launch a health care coverage
program or product that will extend health care coverage to the working uninsured within
a community, a policy entrepreneur must exist. (In all cases we would expect there to be a
small number of persons working together who individually fit this definition, although
we would expect one at least to be primus inter pares, if not clearly identified as the
leader of this group.) As stated in the discussion around Proposition #1, the selection of
the project leader is the focal point for the selection of the policy entrepreneur. The
supply of policy entrepreneurs is expected to be steady across cases, constrained only by
the income and education levels extant within the county (as discussed in Chapter Three,
see Schneider and Teske, 1992).

Ideally this policy entrepreneur should already possess extensive knowledge of the
nature of the health care system within his or her community as well as an understanding
of the growing problem of the uninsured. However, this person would be expected to
engage in an information gathering effort that attempts to define the problem within the
community, preferably in collaboration with the partners determined to be necessary to
the development and launch of the desired innovation. Indeed, this community—specific
information gathering and problem definition process would be expected to take place

with the participation of potential coalition partners in a manner that would elicit these

74

 

 

 

"_ Cu- 2..

partners’ views on the scope of the problem, and, broadly, what kinds of innovations
these partners perceive would best ameliorate the problem.

A coalition of issue area actors that will support the development and launch of
the innovation is proposed as a critical first step in a two step process. The first step is to
build the supportive coalition, the second step is to engage in the technical aspects of the
development and launch of the innovation. Consider the building of the supportive
coalition as ‘getting the ducks in a row,’ a process where the policy entrepreneur engages
in information gathering, as described in the previous chapter, thinking critically about
the issues and problems, and listening to the thoughts of the actors within the issue area
as to the problem to be addressed, and these actors’ notions of how best to ameliorate the
problem. The policy entrepreneur then uses their own information and the information
gleaned from prospective coalition partners, and proposes an innovation that s/he thinks
best meets the needs of the potential coalition partners and also could ameliorate the
problem.

This person should also have an established network of contacts that extend both
into local policy networks and into the local market for health care services, or have the
ability to develop these contacts. The policy entrepreneur should have strong social and
political skills, and have the ability to think creatively about the problems of covering the
working uninsured within the broad structures of funding health care coverage that exist
at the federal, state and local level. At least as important, the policy entrepreneur must be

able to listen to how potential coalition partners perceive the problem and its potential

75

 

 

 

solutions, and then frame their arguments for an innovation using these partners’, and
his/her own innovative ideas for a solution.

The entrepreneur should understand how to frame an issue, which will be revealed
by his/her ability to craft causal arguments that support his/her proposed innovation, and
can mold these causal arguments to suit the multiple organizations/individuals that must
be convinced to participate in the partnership that will develop and launch the innovation.
Lastly, the entrepreneur must understand how to place an issue on the local political
agenda, and how to keep an issue high on local policy agendas long enough to implement
the innovation.

Because of the complexities of the issue area, we would expect that the policy
entrepreneur or project director, will build a collective entrepreneurship team to
maximize the availability of the skill sets and characteristics mentioned above. The more
diverse and complementary the skills of this team, the more likely they are to succeed.

Because a coalition must be developed in order for the innovation to succeed, the
entrepreneur must also possess coalition-building skills. These skills will include the
ability to both reframe problems so that actors will see benefits in participation where
before they saw only costs, as well as the ability to find ways to apply pressure to some
necessary actors in order to achieve cooperation. It is important to note here with respect
to the CCHMS project sites that the governing boards Should not be mistaken for the
coalitions that must be built. Rather, the entrepreneur must develop a coalition of like-
minded organizations/individuals, many or all of whom are also on the governing board

in order to succeed in the development of the innovation.

76

 

 

 

 

 

 

Finally, the policy entrepreneur (or some members of the collective
entrepreneurship team) will also be acting as a market entrepreneur, and will face the set
of dual constraints described in the literature review. The entrepreneur will need to
navigate both sets of constraints in order to form a winning coalition as well as develop

and launch an innovative program or product.

Proposition #3: Local policy entrepreneurs either alter the actual structure of costs
and benefits of the members of the coalition vis-a-vis participation in the
development and launch of a new program or product, or they alter the members’
perceptions of the costs and benefits of the new program or product.
Unit of Analysis: Policy Entrepreneur

The policy entrepreneur may either alter the actual costs and benefits for the
potential coalition partners necessary to the development or launch of the program or
product designed to extend health care coverage to the working uninsured or they may
use causal arguments to alter the potential partners’ perceptions of the costs and benefits
of participation. They may also do both. The policy entrepreneur may alter the actual
costs and benefits by accessing external sources of funding to support the launch of the
innovation, and they may be supported by a change agent so that they may foot a portion
of the costs of the development of the program or product. In this manner the policy
entrepreneur changes the actual costs and benefits to potential partners in their coalition.

However, the policy entrepreneur may also make causal arguments to change

potential coalition partners’ perceptions of costs and benefits. These causal arguments

77

 

 

 

would be expected to center around the unknowns surrounding the expected success (
the program or product, especially the financial risks that the potential coalition partnr
perceive in committing to the partnership and the launching of the program or product
The characteristics, skills and ability of the entrepreneur to alter real costs or benefits,
just members’ perceptions of costs and benefits, will be represented in the decision trc
under both the coalition-building phase, where their importance will be secondary, an
the program or product development and launch phase, where their importance will be
primary.

Regardless, the policy entrepreneur must build a coalition of the key entities
within the issue area in order to obtain support for the development and launch of the
desired innovation. The task of convincing these potential coalition partners of the wo
of supporting the entrepreneur’s effort is a critical precursor to the development and

launch of the desired innovation.

Exogenous Factors
Proposition #4: The local political and market contexts constrain the ability of th
policy entrepreneur to build a winning coalition that can develop and launch an
innovative program or product.
Unit of Analysis: County
As discussed in the literature review, both the political and market contexts wi
place constraints on the ability of the entrepreneur to build a winning coalition that car

then develop and launch an innovative program or product. For the political context t1:

78

limits are based on institutional structures for policy making, policy settings, and the
behaviors of the groups and individuals around them. This milieu shapes the possibilitie:
of action by a policy entrepreneur and allows us to develop expectations about the
possibility of generating support for policy innovation. For the market context these limi
stem from two contextual factors, the pace of change (innovation) within the market, anc
the concentration of firms within the market.

Based on the discussion presented in Chapter Three, we should expect that the
bulk of these constraints will be revealed in the coalition-building stage, prior to any
effort to develop and launch an innovative program or product. These constraints are
represented by the availability of alternative venues for presenting the issue, as Opposed
to the closed venues of the health care policy subsystem. Although the literature on polic
entrepreneur behavior suggests that these two processes occur at the same time, we woul
expect the highly technical aspects of program or product development would not occur
until after a winning coalition had been formed to support and participate in this part of
the process. Also, the concept of guarded dissemination suggests that policy
entrepreneurs would try to make the two processes more sequential than simultaneous, tc
protect information during the coalition-building stage that is necessary and requires

protection for later use in the program or product development stage.

79

 

Proposition #5: The local political and market contexts must be undergoing change
in order to increase the probability of successfully building a coalition that will
develop and launch some program or product to extend health care coverage to the
uninsured.

Unit of Analysis: County

The political venues of the health policy issue area are largely closed and
dominated by the suppliers of health care services. These actors have a great deal of
political power which has led to significant legal power over the supply-side of the
market. However, the political venues in some states and communities, despite
remaining largely closed, have been cracking open. Especially noteworthy is that
physicians have been losing their grip on the authority to make diagnoses and prescribe
medications to patients to nurse practitioners and physician assistants through changes in
state law as well as the spread of managed care arrangements. Physicians have also begun
to feel financial pressures related to changes in reimbursement and reimbursement
structures for both Medicaid and Medicare, and have reacted negatively to the constraints
of private sector managed care arrangements.

The market context has undergone significant change. The reduction of federal
reimbursements for Medicare claims, the state purchase of Medicaid services through
managed care organizations, the broad introduction of managed care arrangements in the
private sector of the market, and the increase in purchaser demands for lower costs or at

least lower rates of increase in costs have all weakened these actors’ power within the

market.

80

 

 

 

 

The weakening of these actors is perhaps critical to the development of political
and market contexts that permit the introduction of a policy innovation that rests on a

broad public-private coalition in order to succeed.

Generalizability
Proposition #6: Although the cases to be proposed all exist in Michigan, the lessons
learned from these cases will yield generalizable concepts that may be applied
throughout the U.S.
Unit of Analysis: County

The health care context in Michigan is unique in two ways, e.g., all hospitals are
nonprofits, and it is one of only 12 states that has not initiated market reforms that permit
wider development of individual or small group health care coverage products. However,
the main thrust of this study is on change agents and policy entrepreneurs, and the
insights gained from this research will be widely applicable to other communities in the
U.S., as well as to theories of local-level policy entrepreneurship in other issue areas (e.g.,

antigrowth or education policy entrepreneurs).

Criteria for Interpreting Findings: Links Between Interview Data and Propositions
Links between interview data and propositions

Interview data was collected from the key stakeholders (both local and state)
involved in the CCHMS Initiative within each of three separate counties, as well as within

Wayne County. Using a common interview instrument developed for the evaluation of

81

 

the CCHMS Initiative, interviews were conducted with all key stakeholders in the four

counties (see Table 5.) These data are used to provide answers to propositions one

through seven in the next chapter.
Interviewees within each county included:

1. Local policy entrepreneurs (Project Directors at each CCHMS project site, and
County Executive officials for Wayne County);

2. Members of the potential or realized partnership (hospitals/health system CEOs
and/or their designated representatives); physician participants (or their
representatives); local health department participants (if any); and county-level
public officials participants (where applicable).

3. Foundation change agent representatives (interview data from a previous
evaluation of the Planning Phase of the CCHMS Initiative);

4. A state agency representative from MDCH; and

5. Relevant consultants used by each partnership in the development and launch of
the innovation (where applicable).

The specific persons who were interviewed varied. For example, a county

commissioner was principally involved in only three of the four counties.
Variables

The variable list used in coding these interviews of demonstration site key

informants is provided below.

82

 

 

Proposition #1 Variables"

FCA_CAT:

FCA_SOL:

FCA_HLP:

FCA_LNK:

SCA_LNK:

SCA_NET:

Foundation change agent acting/not acting as a catalyst for innovation.
Foundation change agent acting/not acting as a solution giver to aid
innovation.

Foundation change agent acting/not acting as a process helper for change.
Foundation change agent acting/not acting as a resource linker to support
innovation.

State agency change agent acting/not acting as a resource linker to support
innovation.

State agency change agent providing/not providing a networking

opportunity to encourage innovation.

Proposition #2 Variables

LPE_SUP:

LPE_KNO:

LPE_CON:

 

Reference (+/-) to the supply of policy entrepreneurs.

Presence/absence of issue-specific knowledge on the part of the policy
entrepreneur (or a member of the team assembled by the policy
entrepreneur).

Presence/absence of social/business connections on the part of the policy
entrepreneur (or a member of the team assembled by the policy

entrepreneur).

Assessing the ability of the change agent to select communities that are
‘ready’ will be performed using the context variables presented for
propositions #4 and #5.

83

 

 

LPE_CAU:

LPE_COA:

LPE_LME:

LPE_ING:

LPE_COL:

LPE_MED:

LPE_PRO:

LPE__COM:

 

Presence/absence of the use of causal argumentation on the part of the
policy entrepreneur (or a member of the team assembled by the policy
entrepreneur).

Successful/unsuccessful coalition-building activity on the part of the
policy entrepreneur (or a member of the team assembled by the policy
entrepreneur).

Successful/unsuccessful market entrepreneur activity on the part of the
policy entrepreneur (or a member of the team assembled by the policy
entrepreneur).

Presence/absence of information gathering activity on the part of the
policy entrepreneur (or a member of the team assembled by the policy
entrepreneur).

Successful/unsuccessful use of collective entrepreneurship team.
Successful/unsuccessful use of local media on the part of the policy
entrepreneur (or a member of the team assembled by the policy
entrepreneur).

Successful/unsuccessful problem definition activity on the part of the
policy entrepreneur (or a member of the team assembled by the policy
entrepreneur).

Successfiil/unsuccessful use of a community venue on the part of the

policy entrepreneur (or a member of the team assembled by the policy

entrepreneur).

84

 

 

 

Proposition #3 Variables

PER_COS:

PER_APC:

PER_ACO:

Perceptions of cost (+/-) on the part of potential/actual collaborating
participants in the development of any innovation to extend health care
coverage to the uninsured. This variable is intended to capture potential
partners’ general beliefs of the cost of extending health care coverage to
the uninsured.

Successful/unsuccessful attempt on the part of the policy entrepreneur (or
a member of the team assembled by the policy entrepreneur) to alter the
perceptions of the cost to potential/actual collaborating participants during
the development of a specific innovation to extend health care coverage to
the uninsured.

Successful/unsuccessful attempt on the part of the policy entrepreneur (or
a member of the team assembled by the policy entrepreneur) to alter the
actual costs to potential/actual collaborating participants during the
development of a specific innovation to extend health care coverage to the

uninsured.

Proposition #4 Variables

LCO_LPC:

Effect (+/-) of the local political context at the start of the process of
innovation on the development/launch of the innovation. Reference to the
openness of venues/structures is positive, reference to closed

venues/structures is negative.

85

LCO_LMC: Effect (+/-) of the local health care market context at the start of
innovation on the development/launch of the innovation. Reference to
conditions that have been described as weakening health care subsystem
actors are positive, reference to conditions that would not weaken these
actors is negative.

Proposition #5 Variables

LCC_PCC: Effect (+/-) of the current changes in the local political context during the
process of innovation on the development/launch of the innovation.
Reference to opening of venues/structures is positive, reference to closing
of venues/structures is negative.

LCC_MCC: Effect (+/-) of the current changes in the local health care market context
during the process of innovation on the development/launch of the
innovation. Reference to conditions that have been described as weakening
health care subsystem actors are positive, reference to conditions that

would not weaken these actors is negative.

Variable Coding and Scoring Process

The interviews were each read several times to ensure consistency in the coding
and scoring of all variables. Scoring was assigned to each appearance of each variable
based on whether it was a positive or negative appearance of the variable. Due to the
difficulties associated with determining the relative effect of any appearance (whether

positive or negative) of a variable on the development/launch of an innovation, positive

86

 

 

and negative effects were simply scored either 1 (for a positive effect) or -1 (for a
negative effect).

The appearance of a variable was not defined by the use of specific words, but
rather descriptions of behaviors, actions, social connections or environmental
descriptions. Thus an appearance of a variable almost always occurred in units that were
no smaller than a sentence, and often were more than one sentence in length.

An appearance of a variable related to a specific event was counted only once
within each response to a question on the interview instrument. However, if the event was
discussed again in a response to a different question on the interview instrument, it was
counted as another appearance of the variable.

Some one sentence responses were counted as appearances of two variables. For
example, the statement, “Obviously I considered the media campaign and the placement
of the county executive’s name on the product as absolutely crucial,” was scored as a
positive 1 for both use of the media and for the local connections of the policy
entrepreneur. The rationale for assigning a positive score on the first variable is self- '
evident. The rationale behind assigning a positive score on the second variable is that the
use of the name of the lead member of the collective policy entrepreneurship team was
perceived by the respondent as lending credibility to the innovation. Thus this response
was interpreted as revealing the strong positive name recognition of the individual among
a group of stakeholders important to the development and launch of the innovation.

Each set of coded and scored key informant interviews for each county were then

entered into an Excel spreadsheet, one row for each informant interview. These sheets

87

 

 

were then linked to an interim aggregated spreadsheet, and then to a finalized aggregated

spreadsheet. The finalized aggregated spreadsheet presented the aggregate score on each

variable, the total number of appearances of each variable, and the total number of

interviews with appearances of each variable. The total number of interviews with

appearances of each variable was used to provide a measure of the validity of the score in

the following manner:

1. For variables where the percentage of respondents was 50% or greater, these
scores were considered to possess high validity.

2. For variables where the percentage of respondents was equal to 25% and less than

50%, these scores were considered to possess medium validity.

3. For variables where the percentage of respondents was less than 25%, these scores
were considered to possess low validity, and are not reported.

Table 5 shows the number of interviews completed for each case study, who
conducted them, and when they were conducted. Table 6 shows the timing of interviews
by case. The count of interviews for the foundation change agent and the state agency
change agent does not include all respondents who remarked on these two organizations’
roles, only those organizations’ representatives who were interviewed. Most of the
interviews were conducted in person, captured on tape, and transcribed to facilitate the
coding and analysis process. The state agency respondent declined to be taped. Only one
of the Wayne County respondents was conducted in this manner, the remainder were

conducted via telephone and were not taped. This was the result of the difficulty in

88

 

 

 

obtaining time to speak with these respondents because of their relatively higher level
positions within government or large health systems.

The reason the number of respondents by case is different for the foundation and
state agency case studies was related to the few people involved within these
organizations, however data were also provided by each of the county case study
respondents.20 The number of St. Clair County respondents was low because few people
were ever involved in the effort to develop and launch an innovation. For the Wayne
County case study the number is low because the pool of potential respondents were
individuals whose status was relatively much higher than those in the other counties,
making demands on their time for interviewing a low priority. A number of refusals for
interviews occurred as a result. The relative differences in the number of interviews
conducted, and then necessarily differences in the weight of the scores of the variables for
each case will be accounted for when the propositions are assessed across cases.

For the Wayne County case, the few persons interviewed nonetheless captures the
necessary information from the pool of actors involved. The four interviewees include the
primary consultant for health care financing and benefit plan design. This interviewee
also was employed at the state’s Medicaid agency and was intimately familiar with the
events that occurred around indigent funding of health care services at the state and
county level that eventually led to the development of HealthChoice in Wayne County.

Also interviewed was a Detroit Medical Center senior official, who had participated the

 

20 One foundation change agent interview was conducted as a focus group

with three respondents.

89

 

 

most in the benefit plan design for HealthChoice. The other two interviewees were the
past and current directors of HealthChoice, the former having participated in the
development of the general model for HealthChoice, and the specifics of the model and
its benefits plan after the two health systems agreed to participate.

The case studies will be presented one at a time, with first a review of the
historical activities that occurred around the effort to develop and launch an innovation,
then the scores on all variables that reach medium or high validity will be reported.
Wherever appropriate and usefully illustrative, quotations from the taped and transcribed
interviews will be included for each variable.21 As a result of the human subjects
anonymity protections provided to each of the key informants prior to their agreement to
be interviewed, and the small number of interviewees in each case, quotations and the
complete transcripts are not provided in this dissertation. Each case will then be

concluded with a summary of the findings from that case.

Criteria for Interpreting Findings: Two Decision Models

Two decision models have been developed to aid in the interpretation of the
findings and further the development of a theory of a change agent and a theory of an
independent local level policy entrepreneur engaged in the pursuit of innovations at the
local level. The first model is of outside-inward change (see Figure 4) and presents the

shifting probabilities of successful entrepreneurial innovation based on the local (county)

 

2‘ For the Wayne County case, because only one interview was taped and

transcribed, no quotations will be included, to avoid providing too much
interpretative weight to the quotations of one interviewee.

9O

political and market contexts where an external change agent initiates an outside-inward
change effort. Using the first model and data collected for answering the propositions,
probabilities proposed here will be assessed in the next chapter for accuracy for the
segments of the model to represent the constraints faced by a project director or a policy
entrepreneur in each county. These probabilities will not be assessed in the sense that a
mathematical analysis of outcomes will yield specific probabilities, but will rather be
assessed based on an analysis of the interview data, which are based on the propositions.
The second model is of inside-outward change (see Figure 5) where a policy entrepreneur
indigenous to the geographic area targeted for the desired innovation decides to either
pursue change, or avoid change.

Both models are forms of a lottery (von Neumann and Morgenstem, 1944). A
lottery is a decision model where one actor or decision maker is presented with a single
choice opportunity between alternatives. The outcomes stemming from each alternative
are not assumed to be within the complete control of the decision maker. One alternative
of the lottery usually has a known outcome, such as “no bet.” The alternatives (assuming
only two exist) could result in any of several other outcomes. The decision maker
associates probabilities with these outcomes occurring and values (payoffs) for all
outcomes, such as winning $4 million (minus one dollar) and losing (minus one dollar).
Thus the solution to the model is derived by calculating the expected value (EV) of each
alternative and assuming the decision maker chooses the alternative with the highest EV.

The key choice is the determination by either the external change agent (in the

three CCHMS case studies this is the WK. Kellogg Foundation) or a policy entrepreneur

91

 

 

 

(for the Wayne County case study) of the readiness of the political and market contexts to
support change, as well as an assessment of how these two contexts will affect the
probabilities of forming a strong/weak coalition that will then support the development
and launch of an innovation, or not. Thus this is an assessment by either an external
change agent or an indigenous policy entrepreneur of the presence or absence of the
classic “window of opportunity” postulated by Kingdon (1995).

In the model of outside-inward change (Figure 4), the actor is the external change
agent, thus the payoffs represent the value of the outcomes for the change agent, not the
project directors, or the counties where the initiative was funded. In the model of inside-
outward change (Figure Five), the actor is the local policy entrepreneur, and the payoffs
represent the value of the outcomes for the local policy entrepreneur, not the county

within which the local policy entrepreneur operated.

Model Structure

Both models have similar structures, that being a decision to pursue some change
effort with an end goal (payoff/outcome) of developing and launching some innovation.
In this study, the innovation is expected to be a program and/or product that exists in both
the public and private sectors, and extends health care coverage to the uninsured, likely
the working uninsured. The key difference between the models is that the decision to
pursue (fimd) change in the outside-inward model is made by an external change agent (a
foundation), and the decision to pursue change in the inside-outward model is made by an

individual policy entrepreneur.

92

 

 

 

In both models, the decision to pursue change is based on an assessment of the
presence or absence of a window of opportunity for change in the issue area. This
assessment is based on the decision maker’s (whether it is a foundation or an individual
policy entrepreneur) perceptions regarding the readiness of the context to support the two
phases that will result in the desired goal of developing and launching an innovation:
building a coalition of the key entities within the issue area necessary for the development
and launch of the innovation, and then the actual process of developing and launching the
innovation.

Thus, there exist three separate sets of probabilities to be assessed by the decision
maker prior to pursuing or not pursuing change. These are associated with one decision,
and two sets of activities:

1. Are E political and market contexts amenable to change? As presented in the
literature review, this is a function of the pace of change and the concentration of
interests within both issue areas.

2. Given that the contexts are amenable to change, does some individual exist with
the necessgy set of entrepreneurial skills and assets to build a strong coalition to
support the development and launch of the innovation? As presented in the
literature review, this would be an individual with strong social connections,
issue-specific knowledge, and the ability to creatively frame causal arguments that
can be transported across the frames of reference of the entities necessary to the

formation of the supportive coalition.

93

 

 

 

 

 

3. Given that the contexts are amenable to change, and a strong coalition is built,
does some individual have the necessary set of skills to develop and launch the
desired innovation? As presented in the literature review, this is an individual
with the ability and enough issue-specific knowledge to design and launch an
innovation. This person must also be expected to encounter situations where they
must either alter the specific structure of the innovation, as necessary, to meet the
demands of the different entities within the coalition, or perhaps the ability to alter
these entities’ perceptions of the benefits of the innovation. Lastly, in complex
issue areas, the ability of this person to form a collective team of entrepreneurs in
order to develop and launch a technically complex innovation would also be a part
of the probabilities.

These three steps are sequential. The sole decision to be made is the decision to
pursue change, which is the combination of the assessment of the amenability of the local
contexts for change, and the perceived ability of some individual (a project director or an
indigenous policy entrepreneur) to form a strong coalition that will then support the '
development and launch of the desired innovation (payoff/outcome). Then an effort is
made by either the project director or the indigenous policy entrepreneur to form a strong
coalition to support the third step, the development and launch of the desired innovation.

Assessing these probabilities across these two models highlights the key
difference between the models: holding issue area and desired innovation
(payoff/outcome) constant, although the assessment of the amenability of the local

context is the same, the assessment of the ability to form a strong coalition, and then

94

 

 

 

 

 

develop and launch an innovation are different. After assessing the local contexts, in the

outside-inward model of change, the change agent must assess the ability of some
individual (the project director) within the targeted community to carry out the formation
of a strong coalition and then the development and launch of the desired innovation. In
the inside-outward model of change, the local policy entrepreneur must self assess his

own abilities to build a strong coalition and develop and launch the desired innovation.

Outside-Inward Model of Change

The first model provided here illustrates the choices available to change agents

determining whether or not to begin some change effort at the local level (see Figure

Four). The first node of this model is an assignment by nature of the readiness of the

county to support change within a given issue area. The probability of assignment of high

readiness to a county is p0, while the probability of assignment by nature of a low
readiness level is l-p0 (see the description of these probabilities provided below).

The next node is the sole decision point within this model, the decision by the

change agent to fund/not fund an externally-initiated change effort. This decision point is

one where complete and perfect information is absent, the change agent is incompletely

and imperfectly aware of whether or not the county in question has a high or low state of

readiness for change within the given issue area. However, the change agent may assess

the county in order to better understand whether or not it possesses high or low readiness

to support change within the given issue area. Once the change agent makes a choice to

95

__K, ,

 

fund, it has to rely on the grantee, specifically the project director, to accomplish the goal
of the change agent.

If the change agent decides not to fund, the model reaches an endpoint. The
payoffs associated with this ‘not fund’ decision are a function of whether or not the
decision to ‘not fund’ is a potential missed opportunity (not funding a community with a
high readiness level) or the avoidance of likely failure to reach desired goals (not funding
a community with a low readiness level).

The next set of nodes is where the project director begins the two step process of
achieving the funded goal(s) of the change agent. The two steps are: (1) to build a
supporting coalition (which may be strong or weak) consisting of most or all of the actors
who will be involved in or necessary to support the (2) development and launch of the
desired innovation (which may or may not occur). The node where the project director
goes about building the supportive coalition is constrained by the readiness of the
community to support change within the issue area, and the entrepreneurial skills and
assets possessed by the project director. This decision model does not permit the change
agent the opportunity to defilnd a project before it reaches its endpoint.

Where the choice by the change agent was to fund a high readiness community,
the probability of building a strong coalition is p,, and the probability of forming a weak
coalition is l-p,. Where the choice by the change agent was to fund a low readiness
community, the probability of building a strong coalition is p2, and the probability of

forming a weak coalition is l-p2 (see the description of the probabilities presented below).

96

Once a coalition (weak or strong) has been formed, the proj ect director must then
engage in the second step of the process, developing and launching the desired
innovation. This step is constrained by the readiness of the community to support change,
the entrepreneurial skills and assets possessed by the project director, and the strength of
the supportive coalition. Where the change agent has funded a community with a high
readiness level, and a strong coalition has been built, the probability of developing and
launching the desired innovation is p3. The probability of not developing and launching
an innovation with these precursor events is 1-p3.Where the change agent has funded a
community with a high readiness level, and a weak coalition has been built, the
probability of developing and launching the desired innovation is p4. The probability of
not developing and launching an innovation with these precursor events is 1-p4. Where
the change agent has funded a community with a low readiness level, and a strong
coalition has nevertheless been built, the probability of deveIOping and launching the
desired. innovation is p,. The probability of not developing and launching an innovation
with these precursor events is 1-p5. Lastly, where the change agent has funded a
community with a low readiness level, and a weak coalition has been built, the
probability of developing and launching the desired innovation is p6. The probability of
not developing and launching an innovation with these precursor events is 1-p6.

There are four payoffs to the change agent for reaching the eight possible
endpoints where the change agent choice to fund has been made to fund an innovative
change effort at the community level. These payoffs vary based on the strength of the

coalition that was formed, and whether or not an innovation was developed and launched.

97

‘n—n‘ rm: way __. _ .—

 

 

It is proposed here that a strong coalition will support the development and launch of an
‘innovative’ innovation, whereas a weak coalition will support the development and
launch a ‘less innovative’ innovation. The other payoffs are where either a strong or weak
coalition does not lead to the development and launch of an innovation. The payoffs vary
here, too, because it is proposed that the change agent may still be able to point to the
formation of a strong coalition as a partial success, whereas the formation of a weak

coalition is of little or no benefit to the change agent.

Information Constraints

The sole decision node is a node where information is imperfect. This decision
node is modeled as a situation where the change agent is imperfectly aware of the
accuracy of its decision rules/criteria for choosing counties with appropriate contexts, as
well as the ability of the change agent to measure the components of community
readiness that it has identified as important. Thus, this also places a constraint on the
project director such that s/he may be incompletely aware of the strength of the local
health policy subsystem, and the likely success of using alternate venues for the purpose
of building a strong coalition to support the development and launch of an innovative
program or product. Regardless of whether the project director knows where she is at
(high or low community readiness), she still must attempt to build a strong coalition and

develop and launch an innovation.

98

 

 

 

Payoffs in the Outside-Inward Model of Change

There are six hypothesized payoffs for the external change agent in this model of

outside-inward change. The payoff ordering is:

a>a-1>5+3>6+2>6+1>6”
This payoff is associated with the outcome where the project director has,
regardless of context, successfully formed a strong coalition and then
developed and launched an innovative program of product. This is the best
payoff, and thus outcome, for the change agent.
This payoff is associated with the outcome of the project director building
a weak coalition that fails to launch an innovative program or product.
This is the lowest valued payoff for the change agent because the resources
expended did not result in the development of a strong coalition or an
innovative program or product, and also incorporates the missed
Opportunities that using these funds for other uses would have offered.
This payoff is associated with the outcome where the project director has,
regardless of context, formed a weak coalition and then developed and
launched an innovative program or product. This outcome is ordered
below the best outcome because it is presumed that a weak coalition
would not be as capable of supporting as innovative a program or product

as would a strong coalition.

 

22

Note that 06-1, 5+3, 5+2, and 5+1 are ordinal in nature for simplicity.

99

 

 

 

 

5+3:

5+2:

5+1:

This payoff is associated with the outcome where a change agent decides
not to fund an initiative in a community with low levels of readiness. This
payoff is ordered third because it represents the potential use of change
agent resources in another productive endeavor(s), as well as the
avoidance of outcomes where resources are expended and the change
agent’s desired outcomes are not achieved.

This payoff is associated with the outcome where the project director has,
regardless of context, succeeded in developing a strong coalition but has
been unable to develop and launch an innovative program or product. This
is ranked above the worst because the newly formed strong coalition yields
other benefits to the community, despite the failure to launch an innovative
program or product. Thus, the change agent will derive some positive
benefit from its expended resources, despite the failure to launch an
innovative program or product. However, it is ordered fourth because the
change agent could have perhaps expended the resources on other projects
that would have fully realized their goals.

This payoff is associated with the outcome of the change agent choosing
not to fund a community with high levels of readiness. This payoff is order
fifth because although the change agent avoids expending resources on an
initiative that may fail to achieve its goals, it has missed a high probability

opportunity to achieve success with a high profile initiative.

100

Probabilities within the Outside-Inward Model of Change

p03

pl & p2:

This probability is nature determining the readiness of the county for
implementation of some change initiative.“Readiness” as used here refers
to the amenability of the local political and market contexts to the
development of alternate venues of action beyond the control of the local
health policy subsystem. The amenability of these two contexts is expected
to covary as hypothesized in Chapter Three. The value of p0 is
unrestricted, thus, p0 may assume any value between 0 and l. Permitting
this probability to be unrestricted is important because it allows nature to
assign any probability between 0 and 1 that a community will have a high
readiness level.

These are the probabilities of successfully building a strong coalition,
given high or low levels of community readiness. These two probabilities
are affected by the context within which the project director takes action,
and the entrepreneurial qualities of the project director. These probabilities
are therefore dependent on the accuracy of the selection criteria used by
the change agent, and how well the change agent applied the selection
criteria. These two probabilities are also affected by the selection of the
person to fill the project director position, a position hypothesized as
requiring a policy entrepreneur. It is expected that p1 > p2, because the
alternate venues available for action are proposed to be more plentiful in

the county where the local health policy subsystem is weaker as a result of

101

 

P3 ' P63

a more favorable context. The value for pl is unrestricted, thus, pl may
assume any value between 0 and l. Permitting this probability to be
unrestricted is important because it allows nature to assign any probability
between 0 and 1 for the building of a strong coalition. Then, regardless of
what value is assigned to pl, p2 must be less than p,.

These probabilities represent the development and launch of an
innovation, given high/low levels of community readiness, and the
formation of strong/weak coalitions. These four probabilities are affected
by the context within which the project director takes action, the ability of
the project director (acting as a policy entrepreneur) to be aware of the
need to combine the roles of policy and market entrepreneurs and then use
guarded dissemination techniques to build a strong coalition within the
local health policy subsystem to support the development and launch of an
innovative program/product. We would expect that at these nodes, context
affects these probabilities less than the success of the project director in
building a strong coalition using guarded dissemination techniques. It is
expected that p3 > p5 > p4 > p6. The ordering places p3 higher than p5
because the context is more favorable, although the other components of
the probability are equal. The probability p5 is ordered second because
although the context is less favorable than for p4 the project director has
succeeded in building a strong coalition despite the unfavorable contextual

constraints. The probability p4 is ordered third because the context for p4 is

102

more favorable than the context for p6. The value for p3 is unrestricted,
thus, p3 may assume any value between 0 and 1. Permitting this probability
to be unrestricted is important because it allows nature to assign any
probability between 0 and l for the development and launch of an
innovation. Then, regardless of what value is assigned to p,, p5 must be

less than p,.

Propositions of Relationships Among Probabilities in the Outside-Inward Decision Model
The probabilities within the model are assumed to be related in the following

ways:

Outside-Inward Decision Model Proposition #1

Increases/decreases in pO are positively related to increases/decreases in the value of p,.

This is because pl is based in part of the readiness of the community, as well as the

entrepreneurial skills/assets of the project director.

Outside-Inward Decision Model Proposition #2

The probability pl is monotonically and positively related to p3. This is because the

building of a strong coalition directly and positively affects the project director’s ability to

then develop and launch an innovation. Thus, increases in pl will result in increases in p3.

Outside-Inward Decision Model Proposition #3

The probability l-pl is monotonically and negatively related to p4. This is because the

building of a weak coalition directly and negatively affects the ability of the project

director to develop and launch an innovation.

103

 

Outside-Inward Decision Model Proposition #4

Increases/decreases in p2 are monotonically and positively related to p,. This is because
the building of a strong coalition directly and positively affects the project director’s
ability to then develop and launch an innovation. Thus, increases in p2 will result in
increases in p5.

Outside-Inward Decision Model Proposition #5

The probability l-p2 is monotonically and negatively related to p6. This is because the
building of a weak coalition directly and negatively affects the ability of the project

director to develop and launch an innovation.

Operationalizing the Outside-Inward Decision Model

The expected values of the decisions for the external change agent for funding
[EV(F)] and not funding [EV(NF)] are provided below.

EV(F) = po(p.(p30€ + (1 - p3)(5 + 2)) + (l-p.)(p4(06-1 + (1 - p.05» +
(1 - po)(pz(psa + (1 - ps)(5 + 2)) + (1 - p2)(p6(a - 1) + (1 - p.05».
EV(NF) = p0(5 + 1) + (l - p0)(5 + 3).

Thus the change agent chooses to fund only when EV(F) 3 EV(N F ).

The key questions to be answered using these equations are:
1. How do changes in each probability affect this decision?
2. How does the value of or affect the evaluation of the decision?
The first question is based on the determination of the existence of a window of

opportunity for an external change agent to fund a community to pursue innovation. The

104

 

޴

second question is the determination of the value of the best outcome for the change
agent, if the choice is made to fund.

Answering both questions entails a discussion of the probabilities associated with
nature’s assignment of readiness and each of the two activities, building a strong/weak
coalition and developing and launching an innovation or not, as well as a discussion of
the value of the outcomes for the external change agent. This will be illustrated through
the presentation and comparison of examples of choices to fund/not fund based on
possible values of oz and 5, and possible values for the various probabilities.

For example, when the values of the payoffs are a = 5 and 5 = 0, and the values
ofthe probabilities are p0 = 0.47, p1 = 0.53, p2 =0.46 , p3 = .70, p4 = .24, p5 = .47, and p6 =
.21, then EV(F) = 3.3, and EV(N F) = 2.1 (see Table 7 for the mathematical calculations
of the expected values). Thus the choice of the change agent faced with these values for
both the given payoffs and probabilities should be to fund. Note that in this example, the
change agent (assuming their information on community readiness is complete) perceives
that the probability is a little less than even that the funded community is one of high
readiness. Yet the probability that a strong coalition could be built is 0.53, and the
probability that the desired innovation will be developed and launched is 0.70. This
example would be a situation where some uncertainty as to the readiness of the
community is mitigated by the higher probabilities associated with building a strong
coalition and developing and launching the desired innovation. These latter two
probabilities are affected by the set of skills and assets of the project director. It is

reasonable to expect that in this situation, the choice to fund was based on beliefs of the

105

 

ability of the project director (provided with resources by the change agent), to build a
strong coalition and develop and launch the innovation, despite some uncertainty as to
whether or not the community is indeed one possessing high readiness.

Another possible set of values for the probabilities results in a different choice.
For example, when the values of the payoffs are the same as in the first example, and the
values ofthe probabilities are p0 = 0.31, p1 = 0.47, p2 =0.28 , p3 = 0.31, p4 = 0.15, p5 =
0.30, and p6 = 0.02, then EV(F) = 2.2, and EV(NF) = 2.4 (see Table 8 for the
mathematical calculations of the expected values). Thus the choice of the change agent
faced with these values for both the given payoffs and probabilities should be to not fund.
In this example, with the payoffs held constant from the first example, but with all of the
probabilities being lower than in the first example, EV(F) is now slightly less than
EV(NF).

What happens when the gap between at and 5 increases? Using the same
probability set from the second example, and holding 5 = 0, but increasing or = 10,
EV(F) = 6.6, which results in a choice to fund (see Table 9 for a mathematical calculation
of the expected value). Thus, if the gap between on and 5 widens, assuming probabilities
remain constant, then the choice to fund becomes more likely. This results in the
conclusion that not only does the change agent’s assessment of the probabilities matter in
the choice to fund/not fund, but also the values assigned by the change agent to 06 and 5

also affect this choice.

106

 

 

 

Conclusions from the Outside-Inward Model of Change

Perhaps the most important conclusion to be derived from the preceding
discussion of the probabilities and payoffs is that the change agent should perhaps already
know who the project director will be prior to determining whether or not to fund. This is
because the ability of the project director to act as a policy entrepreneur is directly linked
to (although not wholly responsible for) that individual’s ability to build a strong coalition
and develop and launch the desired innovation. In both phases, the project director’s
connections, ability to make causal arguments, and issue specific knowledge all
contribute to their ability, within the existing political and market contexts, to build a
strong coalition, and then to develop and launch an innovation, and thus to the probability
that the desired outcome of 05 is achieved by the choice to fund. Therefore, after
accounting for the readiness of a county to support change, the change agent can best
maximize its probability of achieving or by assuring itself that the project director
possesses the necessary qualities to achieve the desired outcome through the selection of
the project director prior to funding the change initiative.

This conclusion should not come as surprise, since both foundations (although not
the foundation under study here) and federal agencies often employ two-stage grant
making processes that on the surface appear designed to reduce the risk of failure in the
funding of communities to develop and launch innovations. These two-stage processes
begin with funding for planning grants that encourage communities to prepare for the
development and launch of an innovation through the planning of the process by which

an innovation will be developed and launched. The second stage is where the funder

107

provides resources to those communities that it deems to be best able to implement the
innovation they have planned. Thus, change agents that use such a two-staged process
appear to be funding the coalition building phase with the planning grant, and then fund
the development and launch of innovations in communities that have presumably
identified an entrepreneur to lead, and have formed a strong coalition.

The reduction in the risks for change agents in employing two-stage grant making
processes is obvious. By using some smaller portion of their overall basket of grant funds
to fund planning grants, they gain some ability to better assess those communities that can
locate a policy entrepreneur to lead the planning grant process, and these persons have the
opportunity to build a strong coalition. If no entrepreneur is located, or if the entrepreneur
fails to build a strong coalition, the employment of grant funds in a losing effort is
reduced, as fewer grant funds are employed for the planning processes. Thus, at the end
of the planning grant process, the change agent has more information from its targeted
(funded) communities as to these communities’ abilities to develop and launch an
innovation. Only those communities that have identified an entrepreneur for the role of
leadership, and have built a strong coalition should be selected for implementation
funding.

This concern over the importance of project leadership on the probabilities
associated with the building of a strong coalition and the development and launch of the
innovation should not obscure the fact that it is the readiness of the community in terms
of its context and the relative strength of any existing policy subsystem that primarily

affects the probability for achieving or. The amenability of context is modeled as

108

 

 

‘i———_T

important, which is consistent with the policy entrepreneur literature and the emphasis
this literature places on windows of opportunity and what constitutes a window of
opportunity. However, in the outside-inward model of change, the change agent only
assesses context, it does not actually go about the process of coalition-building and the
development and launch of the innovation. These activities are in the hands of the project
director. Thus, in a model where the decision to fund change is not broken up into a
planning stage and an implementation stage, should a foundation preselect the project
director prior to funding, it reduces its own uncertainty related to the probabilities for the
building of a strong coalition and the development and launch of the desired innovation.
The choice to fund/not fund is also very dependent on the values that the change
agent places on or (the development and launch of the desired innovation), and 5 (the
failure to develop and launch any innovation). The larger the difference between the
perceived values of or and 5 (holding all probabilities constant), the more likely is the
change agent to choose to fund. Conversely, as the perceived valuation of at and 5
become nearer to each other (holding all probabilities constant), it becomes less likely

that the change agent will choose to fund.

Model of Inside-Outward Change

The second model provided here illustrates the choices available to policy
entrepreneurs determining whether or not to begin some change effort in the absence of
an external change agent (see Figure 5). This model is similar to the first model except

that it is the local level policy entrepreneur who determines whether the context (the

109

readiness of the community) is favorable for the building of a strong supportive coalition
and the development and launch of an innovation. The policy entrepreneur in the inside-
outward model has the freedom to either begin or not to begin a change effort, because
there is no external change agent present that requires this activity to occur once it has
initiated grant funding for a specific change-oriented project. However, the model does
not present the policy entrepreneur with the opportunity to change his mind once he
decides to pursue a change effort. Rather, once an effort has begun, the model does not
conclude until the policy entrepreneur has both attempted to build a strong coalition, and
then proceeds to attempt to develop and launch the desired innovation.

As in the first model, nature assigns the community either a high or low readiness
to support a change effort within a given issue area. The probability that nature has
assigned a high level of readiness to the community is p0, and the probability that nature
has assigned a low level of readiness to the community isl -p0.

The next node is the sole decision point within this model, the decision by the
policy entrepreneur to pursue/not pursue an indigenous change effort. This decision point
is one where complete and perfect information is absent, the policy entrepreneur is
incompletely and imperfectly aware of whether or not the community within which s/he
resides has a high or low state of readiness for change within the given issue area.
However, the policy entrepreneur may assess the county in order to better understand
whether or not it possesses high or low readiness to support change within the given issue
area. Once the policy entrepreneur makes a choice to pursue change, s/he has to rely on

his/her own skills and assets to accomplish his/her goals.

110

 

 

 

If the policy entrepreneur decides not to pursue a change effort, the model reaches
an endpoint. The payoffs associated with this ‘not pursue’ decision are a function of
whether or not the decision to ‘not pursue’ is a potential missed opportunity (not pursuing
change in a community with a high readiness level) or the avoidance of likely failure to
reach desired goals (not pursuing change in a community with a low readiness level).

The next set of nodes is where the policy entrepreneur begins the two step process
of achieving his/her desired goal(s). The two steps are: (1) to build a supporting coalition
(which may be strong or weak) consisting of most or all of the actors who will be
involved in or necessary to support the (2) development and launch of the desired
innovation (which may or may not occur). The node where the policy entrepreneur goes
about building the supportive coalition is constrained by the readiness of the community
to support change within the issue area, and the entrepreneurial skills and assets s/he
possesses. This decision model does not permit the policy entrepreneur the opportunity to
halt a decision to pursue change before it reaches its endpoint.

Where the choice by the policy entrepreneur was to pursue change in a high
readiness community, the probability of building a strong coalition is p,, and the
probability of forming a weak coalition is l-pl. Where the choice by the policy
entrepreneur was to pursue change in a low readiness community, the probability of
building a strong coalition is p2, and the probability of forming a weak coalition is l-p2

Once a coalition (weak or strong) has been formed, the policy entrepreneur must
then engage in the second step of the process, developing and launching the desired

innovation. This step is constrained by the readiness of the community to support change,

111

 

the entrepreneurial skills and assets possessed by the policy entrepreneur, and the strength
of the supportive coalition. Where the policy entrepreneur has pursued change in a
community with a high readiness level, and a strong coalition has been built, the
probability of developing and launching the desired innovation is p3. The probability of
not developing and launching an innovation with these precursor events is 1-p3.Where the
policy entrepreneur has pursued change in a community with a high readiness level, and a
weak coalition has been built, the probability of developing and launching the desired
innovation is p4. The probability of not developing and launching an innovation with
these precursor events is 1-p4. Where the policy entrepreneur has pursued change in a
community with a low readiness level, and a strong coalition has nevertheless been built,
the probability of developing and launching the desired innovation is p,. The probability
of not developing and launching an innovation with these precursor events is l-p5. Lastly,
where the policy entrepreneur has pursued change in a community with a low readiness
level, and a weak coalition has been built, the probability of developing and launching the
desired innovation is p6. The probability of not developing and launching an innovation
with these precursor events is l-p6.

There are four payoffs to the policy entrepreneur for reaching the eight possible
endpoints where the policy entrepreneur choice to pursue change has been made. These
payoffs vary based on the strength of the coalition that was formed, and whether or not an
innovation was developed and launched. It is proposed here that a strong coalition will
support the development and launch of an ‘innovative’ innovation, whereas a weak

coalition will support the development and launch a ‘less innovative’ innovation. The

112

other payoffs are where either a strong or weak coalition does not lead to the development
and launch of an innovation. The payoffs vary here, too, because it is proposed that the
policy entrepreneur may still be able to point to the formation of a strong coalition as a
partial success and may be used in a future iteration of the model, whereas the formation

of a weak coalition is of little or no benefit to the policy entrepreneur.

Information Constraints

The sole decision node is modeled as a situation where the policy entrepreneur is
imperfectly aware of the accuracy of her decision rules/criteria for understanding her
community’s contexts, as well as the ability of the policy entrepreneur to measure the
components of community readiness that it has identified as important. The constraint
this places on the policy entrepreneur is that she may be incompletely aware of the
strength of the local health policy subsystem (the extent to which the pace of change and
the concentration of interests exist within both the political and market contexts), and the
likely success of using alternate venues for the purpose of building a strong coalition to

support the development and launch of an innovative program or product.

113

Payoffs for the Inside-Outward Decision Model

There are six hypothesized payoffs to the policy entrepreneur in this model. The

payoff ordering is:

a:

5:
(X-l
5+3:

0t>O£-l>5+3>5+2>5+l>5.23
This payoff is associated with the outcome where the policy entrepreneur
has, regardless of context, successfully formed a strong coalition and then
developed and launched an innovative program of product. This is the best
payoff and outcome for the policy entrepreneur.
This payoff is associated with the outcome of the policy entrepreneur
building a weak coalition that fails to launch an innovative program or
product. This is ordered as the least favorable outcome because it results in
a loss of face, prestige, and perhaps damages the future career
opporttmities of the policy entrepreneur.
This payoff is associated with the outcome where the policy entrepreneur
has, regardless of context, formed a weak coalition and then developed
and launched an innovative program or product. This outcome is ordered
below the best outcome because it is presumed that a weak coalition
would not be capable of supporting as innovative a program or product as
would a strong coalition.
This payoff is associated with the outcome where the policy entrepreneur

determined that it was prudent not to pursue a change effort because

 

23

Note that 06-], 5+3, 5+2, and 5+1 are ordinal in nature for simplicity.

114

 

5+2:

5+1:

community readiness was too low. This payoff is ranked third because the
policy entrepreneur avoids the loss of face, prestige and influence that
would have occurred had s/he engaged in a failed attempt to pursue
innovation.

This payoff is associated with the outcome where the policy entrepreneur
has, regardless of context, succeeded in developing a strong coalition but
has been unable to develop and launch an innovative program or product.
This is ranked above the worst payoff because the newly formed strong
coalition yields other future benefits to the policy entrepreneur, despite the
failure to launch an innovative program or product. Thus the policy
entrepreneur will derive some positive future benefit from his/her efforts,
despite the failure to launch an innovative program or product. However, it
is ordered fourth because the policy entrepreneur did not fully realize
his/her goals.

This payoff is associated with the outcome where the policy entrepreneur
failed to pursue innovation in a context favorable to such action. This
payoff is ordered fifth because the policy entrepreneur missed a favorable
opportunity to pursue innovation that would have increased his/her

prestige and future career opportunities.

115

Probabilities within the Inside-Outward Model of Change

p0: This probability is nature determining the readiness of the county for
supporting some change initiative. “Readiness” as used here refers to the
amenability of the local political and market contexts to the development
of alternate venues of action beyond the control of the local health policy
subsystem. The amenability of these two contexts is expected to covary as
hypothesized in Chapter Three. The value of pO is unrestricted, thus, p0
may assume any value between 0 and 1. Permitting this probability to be
unrestricted is important because it allows nature to assign any probability
between 0 and 1 that a community will have a high readiness level.

pl & p2: These are the probabilities of successfully building a strong coalition,
given high or low levels of community readiness. These two probabilities
are affected by the context within which the policy entrepreneur takes
action. These probabilities are therefore dependent on nature’s
determination of community readiness. These two probabilities are also
affected by the capabilities of the person cast in the policy entrepreneur
position. It is expected that p l > p2, because the alternate venues available
for action are hypothesized to be more plentiful in the county where the
local health policy subsystem is weaker as a result of a more favorable
context. The value for pl is unrestricted, thus, p1 may assume any value
between 0 and 1. Permitting this probability to be unrestricted is important

because it allows nature to assign any probability between 0 and l for the

116

———-—_4

 

 

P3 ' P63

building of a strong coalition. Then, regardless of what value is assigned to
p,, p2 must be less than pl.

These probabilities represent the development and launch of an
innovation, given high/low levels of community readiness, and the
formation of strong/weak coalitions. These four probabilities are affected
by the context within which the policy entrepreneur takes action, the
ability of the policy entrepreneur to be aware of the need to combine the
roles of policy and market entrepreneurs and then use guarded
dissemination techniques to build a strong coalition within the local health
policy subsystem to support the development and launch of an innovative
program/product. We would expect that at this node context affects these
probabilities less than the success of the policy entrepreneur in building a
strong coalition using guarded dissemination techniques. It is expected that
p3 > p5 > p4 > p6. The ordering places p3 higher than p5 because the context
is more favorable, although the other components of the probability are
equal. The probability p5 is ordered second because although the context is
less favorable than for p4, the policy entrepreneur has succeeded in
building a winning coalition despite the unfavorable contextual
constraints. The probability p4 is ordered third because the context for p4 is
more favorable than the context for p6. The value for p3 is unrestricted,
thus, p, may assume any value between 0 and l. Permitting this probability

to be unrestricted is important because it allows nature to assign any

117

 

probability between 0 and 1 for the development and launch of an
innovation. Then, regardless of what value is assigned to p,, pS must be

less than p,.

Propositions of Relationships Among Probabilities in the Inside-Outward Decision Model
The probabilities within the model also are assumed to be related in the following

ways:

Inside-Outward Decision Model Proposition #1

Increases/decreases in p0 are positively related to increases/decreases in the value of p].

This is because pl is based in part of the readiness of the community, as well as the

entrepreneurial skills/assets of the policy entrepreneur.

Inside-Outward Decision Model Proposition #2

The probability pl is monotonically and positively related to p3. This is because the

building of a strong coalition directly and positively affects the policy entrepreneur’s

ability to then develop and launch an innovation. Thus, increases in pl will result in

increases in p3.

Inside-Outward Decision Model Proposition #3

The probability l-pl is monotonically and negatively related to p4. This is because the

building of a weak coalition directly and negatively affects the ability of the policy

entrepreneur to develop and launch an innovation.

118

 

Inside-Outward Decision Model Proposition #4

Increases/decreases in p2 are monotonically and positively related to p5. This is because
the building of a strong coalition directly and positively affects the policy entrepreneur’s
ability to then develop and launch an innovation. Thus, increases in p2 will result in
increases in p5.

Inside-Outward Decision Model Proposition #5

The probability l-p2 is monotonically and negatively related to p6. This is because the
building of a weak coalition directly and negatively affects the ability of the policy

entrepreneur to develop and launch an innovation.

Operationalizing the Inside-Outward Decision Model
The expected values of the decisions for the policy entrepreneur to pursue change
[EV(PC)] and avoid change [EV(AC)] are provided below.
EV(PC) = P0(P1(P3 a + (1 - ps)(5 + 2)) + (1-p1)(p4(0€-1 + (1 - 1305)) +
(1 - Po)(P2(P5 01 + (1 - p5)(5 + 2)) + (1 - P2)(P6(a - 1) + (1 - 1305))-
EV(AC) = p0(5 + 1) + (1 - p0)(5 + 3).
Thus the policy entrepreneur chooses to pursue change only when EV(PC) 2 EV(AC).
The key questions to be answered using these equations are:
1. How do changes in each probability affect this decision?
2. How does the value of or affect the evaluation Of the decision?
The first question is based on the determination of the existence of a window of

Opportunity for a policy entrepreneur to pursue innovation. The second question is the

119

 

determination of the value of the best outcome for the policy entrepreneur, if the choice is
made to pursue change.

Answering both questions entails a discussion of the probabilities associated with
nature’s assignment of readiness and each Of the two activities, building a strong/weak
coalition and developing and launching an innovation or not, as well as a discussion of
the value of all of the outcomes for the policy entrepreneur.

The examples for assessing these questions can be the same as those used for the
previous model, as the equations for determining EV(F) and EV(NF) are the same as for
EV(PC) and EV(AC). Thus, the probabilities directly affect the decision to pursue/avoid

change, but so also do the perceived values of or and 5.

Conclusions from the Inside-Outward Model of Change

The most interesting conclusion to be drawn from the preceding discussion of the
probabilities is that a rational policy entrepreneur, after assessing the readiness of the
local context to support change (the presence/absence of a window of Opportunity), must
also assess his/her own ability to build a strong coalition and develOp and launch some
innovation. This would seem to place a new demand on the policy entrepreneur, one not
extensively studied or discussed in the literature, which is the ability to assess the whether
one’s own skills/assets are sufficient to pursue the desired goals within the existing
environment. In other words, a policy entrepreneur’s assessment of his/her own abilities
to some extent determines the parameters of that person’s definition of a window of

Opportunity. A person who believes s/he possesses policy entrepreneur skills/assets in

120

 

abundance will define a window of opportunity more broadly than a person who believes
s/he is less well supplied with these skills/assets.

Another component Of the window for Opportunity calculation that comes out
from the discussion of the probabilities is that, if the policy entrepreneur perceives the
need to form a collective team of entrepreneurs in order to achieve the desired outcome,
the policy entrepreneur’s perceived ability to form such a team would also then be
incorporated into his/her definition of what constitutes a window of Opportunity. It should
be expected that this assessment of the ability to form a collective entrepreneurship team
would be related to the financial resources available to the policy entrepreneur, the
availability of persons to flesh out such a team, and the ability of the policy entrepreneur
to lead such a team.

As in the model for outside-inward change, the choice to pursue/avoid change is
also very dependent on the values that the policy entrepreneur places on at (the
development and launch of the desired innovation), and 5 (the failure to deveIOp and
launch any innovation). The larger the difference between the perceived values of at and
5 (holding all probabilities constant), the more likely is the policy entrepreneur to choose
to pursue change. Conversely, as the perceived valuation of or and 5 become nearer to
each other (holding all probabilities constant), it becomes less likely that the policy

entrepreneur will choose to pursue change.

121

 

How Iteration Affects the Two Models

An interesting question across and within the models is whether or not iteratit
plays a role in the decision to pursue/avoid change made by the policy entrepreneur. '.
policy entrepreneur is assumed to be indigenous to the county within which they won
pursue/avoid change. It is possible that the policy entrepreneur has experienced previ-
iterations of this model, thus becoming familiar with the contextual constraints presei
by the issue area within which s/he wish tO pursue change, and perhaps have already
attempted to build a strong coalition and develop and launch an innovation. This wou
increase the ability of the policy entrepreneur to assess the contextual readiness to suj
change, making him/her more accurate at assessing the probability of achieving or. S/
would also become better at assessing his/her own abilities, and the match of her abil
tO the situation she believes she is facing. Thus if the policy entrepreneur has experier
previous iterations of the modeled situation, this would also affect his/her definition c
window of Opportunity. Additionally, if the need for a collective team of policy
entrepreneurs is perceived, previous change activities within this issue area would ha‘
provided the entrepreneur with a sense Of the availability of suitable members for suc
team, the cost of obtaining these members, and the policy entrepreneur’s ability to lee
such a team.

Conversely, the policy entrepreneur could decide to risk an outcome less than
the policy entrepreneur believed the probability Of at least building a strong coalition
sufficiently high, and that the resulting coalition could be used for similar purposes if

future iteration. In other words, if the policy entrepreneur believed s/he could a avoid

122

 

very negative outcome, and build a strong coalition (that would persist into a future
iteration), this would increase the probability of the appearance of future windows of
Opportunity for future iterations, because the definition Of a window of Opportunity would
have been broadened.

Speculating on the possibility for iteration of the modeled process for an external
change agent within the same county is more difficult, because the sparse literature on
change agents does not discuss returning to the site of a failed effort to try again.
However, the same calculations of the value of funding a change process would hold true
if the change agent believed that future iterations were a possibility.

Another aspect of iteration as applied to change agents is that their search for a
project director (which is an attempt to reduce their uncertainty over reaching their
desired outcome, and the risk of expending funds in an effort that does not produce their
desired outcome) could incorporate the idea that the person best suited to the role had
previous experience building coalitions and launching innovations, preferably within the

issue area targeted for change.

Comparing the Models

For both models, increases in p0 and pl also increase EV(F) and EV(PC).
However, the relationship is not one to one. For increases in p3, p4, p5 and p6, the
relationship between how increases in these probabilities lead to increases in EV(F) and

EV(PC) is less ambiguous.

123

 

The key question raised by comparing these two models is whether or not a
naturally occurring policy entrepreneur is better able to identify the situation being faced
and to decide to initiate change as opposed to a project director cast in the role of policy
entrepreneur with no choice but to pursue change. The literature makes clear that
although naturally occurring policy entrepreneurs may decide to act and then fail, by
definition, these persons have some innate and/or developed ability to determine the
readiness of the community to accept a change effort, and thus would seem less likely to
appear in situations where poor identification of community readiness by the change
agent forces a project director to engage in a change effort that may have a low
probability of success.

An important difference between the two models is that the payoffs, despite the
use of the same variables and ordering, are different for the change agent than they are for
the policy entrepreneur. This is most obvious when comparing the worst payoffs (5) for
each. The worst payoff for the change agent (a foundation) is the loss Of face and the
Opportunity cost of the potential use of the same funds in more productive endeavors.
However, foundations are funded from endowments, and although the funding, once
spent, cannot be recouped, the failure of the effort does not affect the flow Of future
dollars from the endowment.24 For the policy entrepreneur, though, the worst payoff could

result in damage to his career prospects. Thus, failure could affect the future career

 

2" For philanthropic foundations that hope to expand their endowments
through future giving (whether by individuals or organizations) towards
endowment expansion, a negative payoff would constitute a more
damaging loss of face.

124

 

prospects and income of the policy entrepreneur, who may lose elected Office, or face
stagnation (or worse) within his career field (whether within a bureaucracy or the private
sector).

Regardless Of the differences of the values of the payoffs between the change
agent and the policy entrepreneur, we can say that as Ct increases, so will the value of the
payoff of choosing to either fund or pursue change. Thus, as differences in the perceived
value of 0t and 5 increase, the likelihood that a choice to either fund or pursue change
also increases.25 Conversely, as the perceived valuation of Ct comes closer to 5, the

likelihood that a decision to either fund or pursue change decreases.

25

In these two models, the worst payoff, 5, is different than the worst payoff
when playing the lottery. When playing the lottery, the worst outcome is
the lost dollar. However, in both of these models, the negative value of 5
is likely to be much larger than -$1.

125

 

Chapter Five: Case Studies, Assessment of Propositions and Decision Tree Models

Introduction

The study of WKKF as a change agent was hampered by the fact that new
interviews of foundation staff for this research project were discouraged, and that many of
the most important internal WKKF documents concerning the genesis Of the
Comprehensive Community Health Models (CCHMS) Initiative were not released.26
However, this research was informed by interviews of the key foundation staff involved
in the genesis and implementation of the CCHMs’ concept. The interviews were
unstructured, thus containing a great deal Of information irrelevant to this study. Also, the
respondents were often reticent in their responses, for reasons that will become clear in
the discussion below. Nevertheless, these interviews provide partial or complete answers
to several of the propositions.

Prior to analyzing the data and conducting the case studies, background
information on the CCHMS initiative must be provided. This background information

includes WKKF’s concept for CCHMS, WKKF’s goals for CCHMS, WKKF’s

 

2" The interviews conducted for this research were conducted as a part of a
larger external evaluation of the CCHMS initiative, an evaluation funded
by the W. K. Kellogg Foundation. There were two teams contracted to
perform this external evaluation, the first team evaluated the ‘Planning
Phase,’ (1994 to 1997) so called because the three demonstration sites
were engaged in planning their sites’ CCHMS efforts during this period.
This first team had its contract voided by the foundation at the end Of the
Planning Phase, and a second team was hired to complete the external
evaluation throughout the ‘Implementation Phase’ (1998 to 2001). This
researcher was a member of the second evaluation team.

126

conceptualization of its role as an external change agent, WKKF’s awareness of the need
for a policy entrepreneur to fill the role of project director, and WKKF’s selection of its
demonstration sites.

The interviews with respondents from the demonstration sites about WKKF as a
change agent provided incomplete information to supplement this study. This may be a
result Of the study focus on the development of an innovation that was embedded in a
larger process, thus the interview instrument focused on the development and launch Of
the innovation and less on the role of change agent. It may also be that, because the
innovation was embedded in a larger initiative, it was difficult for respondents tO separate
their perceptions of the role Of WKKF in terms Of the overall initiative from its role
specific to the innovations. Lastly, WKKF is a major philanthropic institution both in
Michigan and across the country. Thus, responses that cast a negative light on the
foundation may have been avoided by many respondents for fear that, despite the human
subjects’ protections, their responses could in some way be traced back to the speaker and

make the receipt of future funding for other projects more difficult.

Case Study of a Foundation Change Agent: William K. Kellogg Foundation (WKKF)
Prior to launching into the case study, a characteristic of the CCHMS Initiative
requires explanation. Each project site was to divide the initiative into two phases, a
Planning Phase, where the governing board (described below) developed a “Community
Health Investment Plan” (CHIP), which, upon approval from WKKF, would lead them

into the Implementation Phase, implementing the CHIP.

127

WKKF’s Concept for CCHMS

WKKF’s concept for CCHMS rested on the notion that if communities could
harness the variety of health care dollars flowing into their communities, they could
reassign these dollars in a way that would broaden access (coverage) to health care, and
improve community health. This belief rested on the notion that because the dollars that
flow into communities for health care come from different sources, and the assignment of
the dollars to capital resource and new health care service development and maintenance
is held by only a few suppliers within the local market (primarily hospitals), that these
dollars were probably used in an inefficient manner. Coupled with this notion of
inefficient assignment of dollars coming from different streams (federal, state, and private
sources), was the notion that market competition amongst the primary assigners Of
resources — hospitals — exacerbated this inefficiency through the unnecessary duplication
of services and capital resources. One WKKF respondent put it this way, “...if you could
just start with a clean slate, what would a rational health care system look like and what
would it cost? you could just basically engineer the whole system, redirect it.”

Carrying this concept further and incorporating some notion of the resistance to
the CCHMs’ concept that was encountered, the respondent noted, “...how do you begin to
really start something Of this magnitude outside [referring to the role of change agent]
because folks are looking only at what they see as a realm of possibilities in their current
reality and to say, ‘well, why don’t you restructure the reality’ is more than most folks are
prepared to think about.” The respondent later defined exactly what the overarching

outcome should be as a result of the CCHMs Initiative, “A comprehensive health care

128

system where the incentives are in keeping people well, that is inclusive of the entire
community, a system that works together. It’s not fragmented, you know, and ideally I
think that if you could capture the money that comes in a community all the funding
stream[s] that keep everything apart and administer those funds in a different way.”

Interestingly, the notion of restructuring reality was echoed in the interviews
conducted within each county. Many respondents, including some hospital respondents,
noted that competition amongst hospitals within all three CCHMS communities had
resulted in the unnecessary duplication of services, sometimes in areas where there had
previously been collaboration and sharing of resources between hospitals with
overlapping cachement areas. In one county, respondents felt powerless to prevent this
duplication of effort arising from market competition.

Unfortunately this concept ran into several Obvious barriers in real life. First,
harnessing dollars flowing into a community from sources controlled by governmental
entities well above the community level (federal and state) is seriously hampered by the
reimbursement methods employed by these governmental entities. Second, WKKF
appeared to underestimate the complexities of market dynamics as represented by
hospital competition, the extent Of managed care penetration within the communities, and
nature of the private market for health care within each community. Finally, WKKF
appeared unaware of the strong health care policy subsystem that exists at the community,
state and federal levels, and how this subsystem would perceive its concept as a serious

threat to its own future.

129

WKKF’s Goal for CCHMS

WKKF’s goal when it developed CCHMS was an explicit attempt to create a new
venue for action through the development Of its C-P-P27 governing board for each project.
The literature suggests that the success of this venue-creation activity in each county
would be directly related to a pre-existing weakening of the local health policy subsystem
within each county, the nature of each community’s market for health care services, and
the person who filled the role of project leader (preferably a policy entrepreneur).

Nowhere in existing WKKF interviews or documents concerning the planning Of
the CCHMS initiative do we find reference to health policy subsystems at any level of
governance. However, we do see statements about including the voices of persons who
have in the past had no voice in health care decision making at the local level and the
inclusion of these voices into a shared model of community governance Of health care
resource assignment and decision making. This demonstrates a partial but incomplete
understanding of the nature Of policy subsystems and the need for alternate venues for
action, and would also explain their incomplete assessment Of the components of
community readiness (described below), as WKKF planners did not include many ideas

of how context affects the potential for creating alternate venues for action.

 

27 C-P-P refers to the three groups intended to be present in equal numbers
on the governing board, consumers, providers and payers.

130

 

WKKF’s conceptualization of its own role as a change agent
As mentioned in the literature review, the first of five prerequisites for successful
citizen participation in health planning is that an external change agent must be willing to

play multiple supportive roles, including the remaining four prerequisites (roles):

1. A catalyst, that gets a system moving and encourages people tO develop solutions;
2. A solution giver, that provides an idea of how to approach and resolve a problem;
3. A process helper, that facilitates the learning process for people involved in the

initiative; and
4. A resource linker, that ensures participants to some change process are capable of

using the variety of potential resources at their disposal.

Using the interviews of WKKF staff and the interviews conducted of persons involved in
developing innovations within each county, it is possible to assess how or if WKKF
performed each of these roles.

In general, WKKF staff struggled with each Of these roles, and the struggle was
directly related to WKKF’s historical grant making role, which was at odds with the roles
of a change agent. Throughout WKKF’s history it had responded to communities that
sent proposals to them that fit within one of their programming areas, one of which is
health. The foundation had “prided itself ’ on the way that it supported communities that
had identified their own problems and had then approached the foundation for help
supporting the funding of activities that arrived at a solution. A WKKF staff person
directly associated with CCHMS had this to say, “[It was an internal WKKF]

philosophical dilemma ideally you would want [the] community to be coming to you

131

and telling you what their needs are rather than [WKKF] coming from outside it’s a
topdown approach, but you really want it to go in the Opposite direction.” Carrying along
these thoughts about the role WKKF took on with the development Of the CCHMS
Initiative, the respondent continued, “...hopefully the process would be one of opening up
an opportunity and then the dynamism coming from the community and then how do you
feed the information, how do we provide the support systems, the technical assistance to
help this is a very different approach than anything going in the foundation .” The
philosophical dilemma facing WKKF did not bode well for its success in transforming
itself into an outside-inward change agent. It is useful to note, though, that WKKF’s
dedication to the support of community generated ideas is completely in line with the
ideas raised in the literature Of American foundations’ long history of support for
pluralism in communities, and that the increased citizen participation that comes with
pluralism benefits communities

The historical method of grant making had provided WKKF with few clues as to
how to behave as a change agent. One WKKF respondent noted, the old way of doing
business hadn’t allowed the foundation to amass enough information in particular areas
on the strength of communities and health care decision making or reform that were
important to put out as policy lessons, so part of this was a shift to a different model in
the hopes that, in fact, more could be gleaned from this kind of process.” This statement
suggests that the foundation was aware that it had much to learn about being a change

agent, and was approaching the CCHMS Initiative perhaps as a way of learning more

132

about how to act as a change agent in helping communities alter the way they engaged in
health care decision making and reform.

There appeared to be two different internal pressures within the foundation that
forced the development of the CCHMS Initiative onto a fast track. By federal law,
foundations must disburse through grant making a certain percentage of their gross
revenue each year. Specifically, unanticipated increases in the gross revenue fiom
WKKF’s endowment (as a result of a buoyant stock market) presented the foundation
with a need to expend more funds through grant making than they had prepared to
expend. Thus there was internal pressure to locate worthy projects for funding as
described by one WKKF respondent, “... if you’re an organization that has to give away
money for a living, you have to meet certain targets and we don’t have the luxury of not
doing something that’s going to be a payoff for federal requirements.”

The other internal pressure was the idea that foundation financial support of
internal planning processes was finite, and that the staff involved in the internal planning
process were running out of time to produce a plan for the CCHMS Initiative. Noted one
WKKF respondent, “... there’s a limited amount of time in any organization that you can
plan, and there’s a limited amount of time in any organization of policy support for
developing an idea, and we were starting to run into that .” Thus it is possible that one
of the reasons that some WKKF staff perceived the Initiative as a lab from which the
foundation could learn the ropes of being a change agent was because there was
insufficient time to complete planning for such a role within a time frame that would be

internally supported with continued resources.

133

Catalyst Role

The role of catalyst is fulfilled by an entity that gets a system moving and
encourages people to develop solutions. As mentioned above, this part Of the change
agent role ran counter to the historical role assumed by WKKF in its grant making.
WKKF was accustomed to communities coming to them having already deveIOped
solutions, only needing external funding to support implementation. A WKKF respondent
said, “... the idea was really to help the community understand to demystify the system
in such a way that they really could exert control and transform the system. but the real
trick was in getting them, the communities, to become the driving force we didn’t
have the communities going to energize the process.” Later the respondent noted that, “...
the dilemma seems to be, how do you activate the community? How do you select the
communities and how do you expect the communities to be activated? This wasn’t the
normal way that a community sort Of activated

Unfortunately, this research study does not have access to all of the internal
WKKF documents used to select communities that were perceived as ready. However, the
statements in the preceding paragraph, made after the CCHMS Initiative was already
underway in all three demonstration sites, suggest that the respondent believed that the
demonstration communities were insufficiently energized through their own efforts for
this Initiative (i.e., prior to being funded for this Initiative), and that WKKF felt unable to
and unsure Of how to fully assume the role of catalyst to provide outside energy to get the

system moving and encourage people to develop solutions.

134

The foundation was so wary of this role that it placed a neutral financial
intermediary between itself and the implementation of the Initiative in each Of the three
project sites through the granting of demonstration funds to each county’s community
foundation. This notion Of a local, neutral intermediary was tried because the belief
(revealed by subsequent events to be correct) that conflict would occur in each
demonstration county as a result Of the funding Of this Initiative. Conflict over an
initiative of the foundation that directly involved the foundation would apparently not be
tolerated by its Board. One respondent noted, “... [O]ne of the biggest Obstacles was
around this area of potential conflict, one letter fiom somebody who threatened
comflakes can be more powerful than almost anything else.” The respondent clarified
what this meant, “SO it was always the deal that if you could do this and don’t stir up
conflict, go to it.” Thus the foundation felt unable to directly fulfil the role of catalyst.

Interview responses from the key informants within each demonstration site
brought forth only nine coded examples Of WKKF acting as a catalyst, within interviews
from eight respondents (35% of those interviewed, medium validity).28 The aggregate
score was a -3, which means that of the nine responses, six of those portrayed instances
where WKKF did not fulfil the role of catalyst, while three portrayed instances where
WKKF did firlfil the role of catalyst. As mentioned above, this negative view of WKKF
acting as a catalyst by participants in the Initiative in the demonstration counties is
restricted to its activities surrounding innovations designed to expand health care

coverage to the uninsured, not to its larger role in the CCHMS Initiative.

 

2” All variable scores are presented in Tables 10 through 14.

135

Solution Giver Role

The role Of solution giver is fulfilled by an entity that provides an idea of how to
approach and resolve a problem. The foundation had a solution, the one provided by
forming the C-P—P Governing Board which would then attempt to harness the different
streams of health care funding entering the county and redirect those funds in a more
efficient, equitable fashion. However, the foundation was reluctant to impose this
solution, but rather wanted to see each county adopt the solution with little effort.
Although no foundation staff interview evidence discusses this, it is possible that this was
the reason WKKF imposed a two phase process on each demonstration site: the first
phase of planning appears to be intended to permit each demonstration site’s C-P-P
Governing Board the Opportunity to gel while discussing how best to plan for the
achievement of the Initiative’s goals. The process also permitted the foundation input into
the plans before funding the second phase, thus permitting the foundation to redirect what
it perceived as ‘errant’ planning.

Here the key informant interviews from the demonstration sites are helpful.
Although there were still relatively few responses (less than half of respondents), the
answers given were uniform across the sites. Nine respondents (3 9%, medium validity)
brought forth only ten coded examples of WKKF acting in the solution giver role, and the
aggregate score was -2. The persons that did manifest this variable all were pleased with
the initial idea of a C-P-P Governing Board, although the seeds of conflict were already
being planted. In all three counties at least one respondent said something similar to what

this respondent reported, “... everybody was curious and Kellogg was talking about

136

money. SO that’s initially why people come to the table, curiosity, they’re afraid someone
else is going to be there first and say you don’t care... .” Regardless, all respondents on
this variable reported that in the earliest stages of the planning phase there was a great
deal of interest, energy and excitement among the participants. However, this excitement
evaporated when each county in turn produced a CHIP that was turned down by WKKF
and sent back for revision. It is not clear if the participants in each county misunderstood
the process, or if the foundation was too ambiguous in their directions, but some
respondents in each site felt that they had been given carte blanche to produce whatever
plan their C-P-P Governing Board came up with. However, in each demonstration site,
WKKF staff determined that the CHIPS were not innovative enough, or were too stuck in
the old ways of providing health care, and gave each demonstration site suggestions on
how to revise their CHIP, including planning for an innovation to extend health care
coverage to the uninsured. Whether because the planning phase was farther away in time
than the requirement for CHIP revision or because the participants that responded
concerning CHIP revision felt more strongly (one used the word “betrayed”), the negative
scores, all related to the return of the CHIPS for revision, outweighed the positive
responses, which all referred to WKKF’s help in setting up the C-P-P Governing Board at
the beginning of the planning phase.

The foundation played little role in the provision of solutions other than described
here, although they did set up a conduit of information and access to consultants external
to the foundation through the funding of an Operations Office, which is described in more

detail in the process helper and resource linker sections.

137

Process Helper Role

The role Of process helper is fulfilled by an entity that facilitates the learning
process for people involved in the initiative. Kellogg in some ways mingled the process
helper role with the solution giver role, in that their solution was to set up a process of
shared governance over health care planning within each demonstration site. However, as
stated above, one WKKF respondent noted that the foundation struggled with how to
perform this role, and said that they were unsure how to “feed” the communities.
Ultimately WKKF settled on the notion of setting up an intermediary organization
(commonly termed an IMO) to handle the role of process helper for the foundation. This
entity was titled the CCHMS Operations Office. As in the catalyst role, the foundation
was uncomfortable with filling this role directly, and also underestimated the need for this
role. One WKKF respondent noted that, “... I didn’t envision it as a full time job but
clearly it is.”

The Operations office served two conflicting roles, one was as a monitor of
activities occurring in the project sites, the other was as an advocate for the project sites
to the foundation.29 A third role was that Of a provider of technical assistance, discussed
in the resource linker section, below. There are no WKKF interview data that discuss the
Obvious conflict between the two roles, and the interview data from the demonstration
sites provides only six responses from four different informants (17%), thus conferring

low validity to their responses. The coded responses’ aggregate score was -2, which

 

2’ The Operations Office is the focus of a separate study funded by WKKF,
funded by the same grant that supported this research.

138

resulted in a four negative responses and two positive responses on whether or not this

role was fulfilled. This relative absence of interview data related to the process helper role

suggests that it was incompletely filled by either WKKF or the Operations Office.30
Resource Linker Role

The role Of resource linker is fulfilled by an entity that ensures participants to
some change process are capable of using the variety of potential resources at their
disposal. Like the solution giver role, the foundation appeared to be more willing to fill
this role than either the catalyst or process helper roles. However, the foundation was
unwilling to directly fill this role, and instead handed this role to the Operations Office,
where it fell under the rubric of technical assistance.

For the most part, resource linking consisted of the funding of consultants and
other information provision activities made available to the demonstration sites by the
Operations Office. One of the three project directors became so dissatisfied with these
forms of technical assistance that she began to discard without reading all informational

mailings from the Operations Office, and also hired her own consultants to supplant those

 

3" The usefulness of role of the Operations Office was questioned by one

WKKF respondent who stated, “... it’s clear we need somebody out there
working with the communities, but whether the Operations Office was the
way to do it, because again, the ideal situation would be for the
communities to be generating what it is they need .” This suggests that
some WKKF staff were still struggling inside the boundaries of their old
way of doing business, and having trouble grappling with the fact that the
demonstration sites required continual help with process beyond the C-P-P
Governing Board approach outlined by the foundation at the beginning of
the Initiative.

139

 

provided by the Operations Office.3| Another project director was so dissatisfied with the
deliverables of one particular consultant funded by the Operations Office that she hired
her own consultant to redo the work. The third project director expressed no Opinions of
the Operations Office resource linking efforts.

The responses were mixed, with 18 coded responses related to this activity, from
only five key informant interviews, yielding low levels of validity to their responses. The
aggregate score was 2, suggesting that respondents believed that with respect to the
efforts to develop and launch a health care coverage innovation, the foundation and the
Operations Office weakly linked the CCHMS communities to the resources they needed

to complete their tasks.

WKKF’s Conceptualization of Project Leadership

In neither the interviews conducted by the Planning Phase Evaluation Team nor in
WKKF documents is there found a belief by WKKF staff that project leadership would be
an issue beyond the establishment of the C-P-P Governing Board. Thus there was no
search for an individual with the characteristics of an entrepreneur to fill this position. In
fact, it was such a non-issue that the hiring of a project director was left up to the C-P-P

Governing Board. This had grave consequences in St. Clair County, as described below.

 

3‘ The use of the feminine pronoun in no way identifies the respondent, as all

three project directors were women.

140

WKKF’s Selection of Demonstration Sites

The methods used by the foundation in assessing the readiness Of communities to
successfully carry out the CCHMS Initiative were murky and difficult to assess because
WKKF staff were reticent on this subject, and did not release information concerning the
decision rules used to assess whether a county was “ready” for this Initiative. The
decision on county readiness was apparently based on two reports from consultants; the
first report was conducted by a Lansing-based private consulting firm that met with
predefined key stakeholders within a subset often Michigan counties to assess these
stakeholders’ interest in and support for the change proposed in the CCHMS Initiative.
This report was not released by WKKF, thus it is unknown how foundation staff
narrowed their set Of possible sites to ten counties within Michigan.32

The second report was a study often Michigan counties across a range Of social,
economic and health indicators. This second document, titled. “Comprehensive
Community Health Models (CCHM) Project: Community Profiles,” was produced to help
WKKF staff determine counties that were most ready for this Initiative based on the
following criteria; racial mix, ratio of urban population, ratio of age groups within
population, education levels, unemployment rates, income levels, crime rates, number and
type of large employers, employment by industry, distribution of firms by number of
employees, an index of medical underservice, extent of prenatal care, infant mortality,

leading causes of death, immunization rates, number Of providers (physicians and mid-

 

” The ten counties profiled in the second report are: Berrien, Calhoun,
Grand Traverse, Jackson, Lenawee, Marquette, Midland, Monroe,
Muskegon and St. Clair.

141

level) per 100,000, number and characteristics of hospitals (number of beds, admissions,
occupancy rates and births), number and characteristics of long term care facilities and
services, service area of HMOs active within each county, average health and hospital
expenditures, Medicaid payments by vendor, welfare payments, 881 payments, payments
from the State of Michigan’s Office Of Services to the Aging, funding from the Michigan
Department Of Public Health, funding from the Michigan Department of Substance
Abuse, and funding from the Michigan Department of Social Services (Spitz, et al.,
1993)

This is a fairly complete list of available county-level data concerning social,
economic and health indicators. It is unknown how foundation staff used these indicators
to narrow their search for “ready” counties. However, these criteria miss two critical
aspects of a determination of the readiness of a community to engage in the change
proposed within the CCHMS Initiative as noted in Chapter Two, specifically, the
congruence between a county’s geographic borders and its market for health care
services, and the extent of the penetration of managed care arrangements within the
county.

The first missing indicator, congruence between a county’s geographic boundaries
and its local market for health care services, is important because it should be expected
that as the local market for health care services becomes more congruent with a county’s
geographic boundaries, the stakeholders for change are more likely to exist within the
county funded to pursue change. Thus, as a county’s geographic boundaries become less

congruent with its market for health care services, we should expect that some key

142

stakeholders (e.g., hospitals and employers) are external to the county and will see no
benefit to participating in a change process. Congruence between a county’s geographic
boundaries and the market for health care services within which it is embedded is a
critical indicator for assessing readiness as this congruence affects the ability of the
project director to build a coalition of key stakeholders within the county to bring about
change.

The second missing indicator is the extent to which managed care arrangements
had penetrated the local market for health care. This was noted as important in Chapter
Two because managed care arrangements are one of the elements that weaken some of
the actors within the local health care policy subsystem. Because the initiative intended to
target the way that health care resources are assigned within a county, the strength of the
local health care policy subsystem was of paramount importance. It should have been
anticipated (and to some extent the recognition that conflict would arise suggests that it
was) that the members of the local health care policy subsystem would resist changes to
the assignment of health care resources. Thus any indicator that could shed light on a 1
potential weakening of this subsystem should have been included. As described in
Chapter Two, the introduction Of managed arrangements differentially weakens both
hospitals and physicians, by curtailing reimbursements and requiring preapproval of
services, thus lessening the market power Of both sets of actors.

Two indicators that were not mentioned in the literature, but which I proved
important, were the number of public sector employees and the private ownership Of

firms. The assessment Of the number of public employees within the county, whether

143

 

federal, state, or local, was important because it predicted the ability to engage the
employers of these residents in a discussion Of the reallocation of health care resources is
very difficult, if not impossible. Calhoun County had a substantial federal presence in the
form of active duty military personnel assigned to an Air Force Reserve base and an
Army National Guard Base. This took these persons’ health care dollars and their
employers out of the pool Of prospective health care stakeholders and funds that could be
considered Open for reallocation.

The nature of the ownership of the firms within each county was an important
piece of information that could affect the assessment of the readiness of key private firm
stakeholders within a county to engage in a discussion of the reallocation of local health
care resources. If a number of firms, or just several large firms, are owned by firms
outside Of the county, it is very possible that decisions concerning how the firm pays for
employee health care coverage are made outside of the county. This would, as in the case
of public employers, remove these stakeholders from discussions on the reallocation of

health care dollars.

Summary of WKKF as a Change Agent

The Kellogg Foundation approached the role of outside-inward change agent with
a great deal of caution, trepidation and some awareness that they did not fully understand
the role. The foundation staff members planning the initiative also encountered the barrier

of time in terms of the willingness of WKKF’s board to support a detailed but time

144

consuming planning process for the implementation of an initiative that was very
different than previous WKKF grant making activities.

The foundation’s historical method of grant making, reacting to already mobilized
community groups, was diametrically Opposed to the role Of an outside-inward change
agent. Foundation staff planning the implementation of the initiative appeared somewhat
aware of the role, and that this type of grant making would probably lead to conflict
within the demonstration communities. This expectation of conflict, and WKKF’s desire
to avoid it at all costs, led the foundation to use two separate intermediaries to buffer
WKKF from the expected conflict. Thus the foundation did not attempt to directly fill the
shoes Of a change agent. The key IMO was the Operations Office, which was originally
conceptualized as filling roles that were not exactly the same as those predicted of a
change agent, although the IMO developed awareness over time of the need for these
roles, and that the need for some roles was greater than originally anticipated by the
foundation staff responsible for designing the initiative. These problems all led the
foundation to score poorly on the change agent variables, which suggests that the
predicted need was there, but that the predicted need was unfulfilled.

The foundation did not anticipate the need for a person with the characteristics of
a policy entrepreneur to fill the position of project director, and thus played no role in the
initial selection of the person who filled these positions in the three counties.

The foundation had some sense that both the political and market contexts within

which their initiative was implemented might matter, but their assessment of the

145

 

readiness of communities to engage in a change initiative appears to have been

incomplete, leading to the potential selection of sub-optimal counties.”3

Case Study Of the State as a Change Agent

The study of the role of the State of Michigan as a change agent is limited to the
Michigan Department Of Community Health (MDCH), the largest of Michigan’s
agencies, that encompasses Medicaid, public health, mental health and substance abuse
services, services for the aging, long term care, and crime victim’s services.34 This agency
is the primary recipient of both federal and state funding for health care services to the
state Of Michigan (not including Medicare, which is administered directly by the federal
government). Thus any access to public dollars beyond the limited resources available at
the county level required approaching high-level Officials within MDCH.

The likelihood that MDCH was going to play a change agent role was clouded
from the moment it was invited to the table at the beginning of the CCHMS Initiative. The
MDCH respondent expressed incredulousness at the goals of the CCHMS Initiative as

presented by WKKF project staff, and believed from the beginning that its goals were so

 

’3 On of the findings of the overall evaluation of the CCHMS initiative was

that the foundation just included the ‘usual suspects’ when it set up
governing boards in each county, rather than thinking through how the
issue area, and its associated contexts should alter who best should have
been involved in the initiative within each county.

3" None of the three CCHMS counties seriously engaged elected officials

beyond the county level during the time frame covered by this study.
However, the project director in Muskegon successfully engaged her
representatives to the Michigan legislature to secure a portion of the state’s
share of tobacco funding for the Access Health innovation.

146

 

broad and far-reaching as to be unattainable. The MDCH respondent also remarked that
the absence of a competitive bid process extended the process of coalition-building,
information gathering and problem definition within each community more than would
have been the case had competitive bids been sought.

Several state-level activities limited the attention span of these higher-level
Officials throughout the time period Of study. During the first two-thirds of this time
period, MDCH went through a process of shifting 750,000 Medicaid recipients from
either fee-for-service or a limited managed care reimbursement system to a fully capitated
managed care system. Throughout the entire time period (stretching to the present)
MDCH was also attempting to shift all mental health services (those funded by Medicaid
funds or by the state) from a fee-for-service reimbursement system to managed care
arrangements. In the latter third Of the time period covered, MDCH was planning for and
implementing MIChild, Michigan’s programmatic operationalization of the State
Children’s Health Insurance Program which passed the U.S. Congress and was signed
into law by President Clinton in 1997. Thus, there were several major health care-related
policy changes occurring throughout the time period of the CCHMs Initiative under study
here.

Despite the distractions occupying the attention of these high-level MDCH
Officials, each CCHMS demonstration county approached the state as a potential source of
innovative ideas, potential links to other local-level health care coverage innovators in
Michigan and for resources to help fund an innovation. Only one of the three CCHMS

sites developed and launched an innovative program/product for covering the uninsured,

147

 

(Muskegon County). The reasons why Muskegon County succeeded where the other two
failed will come out in more detail in each county’s case study, however, the respondents’
perceptions Of MDCH’s role are presented below. It is important to note now that the
only social/business connections between a high-level staff member of the Medicaid
agency (within MDCH) and a member Of a demonstration site work group attempting to
develop a health care coverage innovation were in Muskegon County. The Director of the
Medicaid agency was from west Michigan, near Muskegon County, and a key member of
the project director’s team had worked directly for the Director of the Medicaid agency
previously when he had run a West Michigan health management organization.
Resource Linker Role

In terms of the variable concerning the state’s role as a resource linker, seven
interviews (29%, medium validity) manifested an appearance of this variable, with an
aggregate score of 7. However, four of the seven respondents were from Muskegon
County, which succeeded in linking to public dollars through MDCH. The other two
counties each had only one respondent manifest this variable, and in both of these two
respondents’ interviews the variable yielded a negative score. The seventh respondent
was from MDCH. Thus we can assume that the state was willing to act as a resource
linker, but it appeared to require the existence of strong, preexisting social/business
connections in order for the state agency Official tO assume this role. This is entirely

consistent with the theory of policy entrepreneurship presented in the literature review

and the propositions.

148

 

Provider of Networking Opportunities
On the variable concerning the state’s role as a provider of networking
Opportunities, only three respondents (9%) manifested the variable in their interviews. It
would be inappropriate to draw any conclusions from the aggregate score on this variable
(3), as so few respondents manifested the variable. However, the absence of data points
suggests that MDCH did not fulfill this role, perhaps due to the large number of

distractions occupying high-level Officials during the course Of the initiative.

Summary of MDCH as a Governmental Change Agent

The high-level staff of MDCH were quite skeptical of the CCHMS Initiative’s
goals as described to them by senior WKKF staff at the beginning of the initiative. Their
skepticism was based on their beliefs that the goals Of the initiative were too broad and
far reaching, and thus doomed to failure. These staff also believed that the absence of a
competitive bid process forced the communities to engage in more work than would have
been the case in the areas of problem definition, information gathering and coalition
building.

The high-level staff of MDCH were at the time of the initiative preoccupied with
a number Of large-scale statewide policy changes, and thus did not have a great deal of
attention available to aid the staff and volunteers of county-level demonstration projects.
State agency staff acted in the resource linker role for only one of the three demonstration
counties, the one where the Director of the Medicaid agency had close personal and

business ties with two members of the collective entrepreneur team, and from the same

149

geographic area of Michigan that the Director was from. The state also did not perform
the predicted role of provider of networking opportunities, perhaps also because Of beliefs
that the Initiative would fail, and the distractions related to co-occurring large, statewide

policy change activities.

Calhoun County Case Study35

Calhoun County was the first of the CCHMS sites funded by the foundation, but
its claim to first place is clouded by the manner in which this was achieved. As the
internal WKKF staff planning the CCHMS Initiative were beginning to perceive that they
were running out of time for continued internal support of a planning process, Battle
Creek (a city in Calhoun County and the home of WKKF) experienced the merger Of two
Of its hospitals. This was a major event for the local health care market, and stimulated
the WKKF planners to propose that Calhoun County was ripe for the implementation of a
C-P-P Governing Board-directed health planning initiative. Not all WKKF staff were in
agreement on this proposal, in part because of their concern that the anticipated conflict
that could arise as a part of the CCHMS Initiative should be kept farther away from the
foundation than its own backyard. One WKKF respondent said, “Now, Calhoun County
would not have been our first choice for a number of reasons. One, if you want to keep
conflict outside your doors, this is not the place, you know, that wouldn’t be your choice
to try to do this.” However, the foundation had a long history of grant making in Calhoun

County, especially in Battle Creek. The respondent noted this and thus the inevitability of

 

3’ A total of ten interviews were conducted for this case study.

150

 

Calhoun as a CCHMS demonstration site, “... the amount of investment we have in this
community, there wasn’t any way of avoiding that [the decision to make Calhoun County
a CCHMS Initiative site] and I think that Calhoun County did not start in the ideal way.”
The MDCH respondent characterized the choice of Calhoun as political, and noted that
the federal government is a major employer in the county, and thus would deprive the C-
P-P governing board of a key stakeholder that provides employer-based health care
coverage to a significant proportion Of the county’s insured population.

Other WKKF respondents struggled to explain the choice Of Calhoun County in
their retrospective 1995 interview36 with a previous evaluation team, stating, “In some
respects Calhoun County up until a year ago was considered a related, but not a CCHMS,
site.” This implies that for its first year of funding, Calhoun was more of a pilot site for
CCHMS concepts than the first implementer of the initiative.

Regardless of its unusual beginnings, the Calhoun County demonstration, titled
the “Calhoun County Health Improvement Project” (CCHIP), got under way, formed a C-
P-P governing board and proceeded to develop a plan for implementing the CCHMS
initiative goals within the county, including expanding health care coverage to the
uninsured. The formation of the C-P-P in Calhoun presaged the sorts of problems and
conflicts that would arise in both sister sites. Three key problems run across all three sites

at this point in time. The first one is of the complexity of the health care system and the

 

3" As mentioned in Chapter Four, two separate teams of evaluators were

contracted to evaluate the initiative. The first team had its contract ended
at the end of the ‘Planning Phase.’ This quotation was drawn from
interviews conducted by that team.

151

level of understanding of this complexity by persons not directly working within the
system. This problem is an echo of the very same problem faced by the federally-funded
citizen based health planning initiatives of the late 1960s and early 19705 studied by
Havelock and Havelock (1973). The second is that selecting a “consumer” representative
is problematic, as all members Of any community are consumers of health care services.
This second problem, coupled with the problem Of low levels of understanding of the
health care system lead to the appointment of a local large private firm CEO to the C-P-P
governing board as a consumer representative.” The third barrier was the conflict
anticipated by several WKKF staff and board members over any decisions that reassigned
health care dollars. Such reassignment was bound to be viewed as zero-sum by at least
some of the actors in the local health care market, especially in a market environment
squeezed by private and public payers, and differentially experiencing the stresses
brought on by the replacement of fee-for-service reimbursement with managed care
arrangements. Not surprisingly physicians were the most concerned over the reassignment
of resources, given the description of their weakened position described in Chapter Two.
Based on the interview responses from Calhoun County (as in the two other
counties, the processes were staged in similar manner at the beginning, as well as being
consistent with the literature on policy entrepreneur behavior), the first step taken in their
development of a health care coverage innovation was information gathering on

innovations elsewhere, both within and outside of Michigan. Six of the ten respondents

 

37 This far from the only such labeled appointment that was a bone of

contention in Calhoun or the other two counties.

152

(60%, high validity) manifested this variable, and the overall score was 13, with only one
negative manifestation of the variable. Thus the information gathering phase was
certainly considered effective and fruitful by the majority of the respondents.

Where is the policy entrepreneur? This is where events diverge significantly from
what was predicted in the propositions related to the policy entrepreneur. At some point
in time during the earliest stages of information gathering, the expected policy
entrepreneur, the CCHIP project director, was supplanted by a coalition of local business
interests, and then for the most part shut out of the development of the innovation that
was launched in Calhoun County, although with the cooperation and eventual formal
approval of the C-P-P governing board . How did this happen?

Early on in the information gathering process, it was discovered that in the county
that borders Calhoun to the west (Kalamazoo), a private-sector group was forming a
purchasing alliance.38 This group, the Southwest Michigan Healthplan Purchasing
Alliance (SMHPA), was interested in expanding their pool of employers in order to
increase their purchasing clout with health plans. Someone, it is unclear who, struck upon
the notion that the purchasing alliance could use some of the savings realized by the
participating firms tO fund health care coverage for the uninsured within both counties.
How this was to occur was never fleshed out, but a clause in SMHPA’s charter was

eventually inserted that this would occur and be discussed by the Board of SMHPA once

 

’8 A purchasing alliance is a coalition of private firms that band together in
order to pool their purchasing power and covered lives to lower the cost of
the health care coverage they provide to their employees.

153

it began Operating in the black.39 It was also suggested by local participants that if the
newly formed Medicaid Qualified Health Plans (the capitated entities providing benefits
to the Medicaid population) could purchase coverage through SMHPA at reduced cost,
these savings for the state Medicaid agency would also remain in the two counties and be
used for this purpose. A number of discussions were held with high-level Medicaid
Officials to make this a reality, but ultimately the state Medicaid agency provided virtually
no incentive for these QHPs to participate in SMHPA, and refused to permit any savings
realized by the QHPs to remain in the two counties for the charitable extension of
coverage to the uninsured.

At this point a discussion of the interview data becomes paramount. First, the
variables that are manifested in the interviews reveal little about who actually acted as the
entrepreneur, or a part of a collective entrepreneurial team. Nor are there many
revelations concerning the benefits of lower health care costs that would be realized by
the Calhoun County employers pushing for this innovation, indeed who took over and led
the development and launch of the innovation. This would appear to be the result of the
innovation existing entirely within the private sector, only the charitable purpose clause

(not to be discussed, let alone activated, until SMHPA had become profitable), marks it

 

3’ This use of a purchasing alliance that would eventually discuss a method

of returning savings to the community through the funding a health care
coverage for the uninsured should strike readers as not particularly
innovative, and wonder why WKKF did not request a revision in the
Calhoun CHIP prior to approving it. That this did not happen is apparently
because the Kellogg Company, also in Battle Creek, and which shares a
number of board members with WKKF’s board, Obtained agreement from
the foundation to accept this innovation in Calhoun’s CHIP.

154

as an innovation intended to extend health care coverage to the working uninsured. Thus
one trademark behavior of market entrepreneurs, closely guarding information, appears to
have been in full force. Only the efforts to gain formal approval from the C-P-P
Governing Board, which was the entity that controlled the CCHMS Initiative funding that
helped support expanding SMHPA into Calhoun County, are discussed.40

However, revealed in the interviews is prototypical entrepreneurial behavior by a
team of Calhoun County businessmen who must convince, through public interactions,
the C-P-P governing board that their innovation is the best route to take in achieving the
CCHMS initiative goal of expanding health care coverage to the uninsured. The strongest
manifestation of the policy entrepreneur variable in the Calhoun County interviews is that
Of coalition building. All ten respondents (100%, high validity) manifested this variable,
with a total Of 62 appearances of the variable in the transcripts, and an aggregate score Of
-44. One respondent stated, “I think it was quite Obvious that the payers [private firms]
wanted the [SM]HPA, the purchasing alliance, very badly and the providers did not. It
turned into win-lose, who got the most votes [referring to the C-P-P Governing Board
formal vote]. I think that hurt the initiative.” This begs the question, how could this
innovation have been developed and launched with such a negative score on coalition

building?

 

4° The charitable purpose clause was important to WKKF because it could

not give funds to a for-profit entity, regardless of the intermediaries in
between. Internal Revenue Service regulations prohibit foundations from
funding the development of for-profit enterprises. Thus the inclusion of
the charitable purpose clause was crucial to the use of WKKF funding to
support the development and launch of SMHPA.

155

Compounding this question is the score revealed on the problem definition
variable. Problem definition is a process whereby the entrepreneur or entrepreneurial
team work with their coalition Of supporters to define the nature and scope of the problem
to be tackled. Nine respondents (90%, high validity) manifested 18 appearances of this
variable in the interviews, with an aggregate score Of -12. This negative score also causes
one to wonder how the innovation managed to be developed and launched. If problem
definition was unsuccessfully carried out, how did the entrepreneurial team succeed in
gaining resources for the development and launch of the innovation?

The answer is found in the interview transcripts. There was a critical C-P-P
governing board vote for the formal approval Of the SMHPA innovation. Prior to the
meeting, the entrepreneurial team supporting the SMHPA innovation relentlessly lobbied
each member of the C-P-P governing board until the entrepreneurial team was certain that
they would carry the vote. The vote was unanimous in support of SMHPA as the
innovation to be used to extend health care coverage to the uninsured. The transcripts
provide no window on these one-on-one lobbying efforts, which one would expect given
that market entrepreneurs as a matter of course guard this information.

This lobbying effort is the root of most of the appearances Of the use of causal
arguments variable in the Calhoun interviews. Seven respondents (70%) manifested this
variable in their interviews, with an aggregate score of 6. The entrepreneurial team of

Calhoun County businessmen appeared to be viewed as largely succeeding in their

156

 

arguments in support of the SMHPA as the innovation that would serve to meet the
CCHMS initiative goal Of extending health care coverage to the uninsured.“

Another reason the development and launch of this innovation occurred appears to
be related to the number of appearances and the score of the market entrepreneur variable
in the transcripts. Six respondents (60%, high validity) manifested 16 appearances of this
variable, with an overall score of 6. Some of this activity went down avenues that
eventually did not pan out, thus the aggregate score is lower than the number of
appearances.

Other predicted entrepreneurial activity appeared in the transcripts. Variables
related to issue—specific knowledge, local social connections, and the development/use of
a community venue as an alternate venue for action all appear in the transcripts at
medium validity. The entrepreneurial variables Of supply Of entrepreneurial leadership,
collective entrepreneurship, and use Of the media as an alternate venue for action are all at
low levels of validity, and hence not appropriate for analysis or the drawing of
conclusions.

For the variable of issue-specific knowledge, four key informants (40%, medium

validity) manifested eight appearances Of the variable, with an aggregate score of 6. This

 

4‘ However, not all were convinced, especially powerful representatives from

the county chapter of the Michigan State Medical Society. Although not a
part of this study, the dissatisfaction of this player in the health policy
subsystem would ultimately prove fatal to the innovation. About a year
after launch, the SMHPA closed its doors. Anecdotal evidence suggests
that this was the result of the Calhoun County chapter of MSMS and local
managed care organizations exercising their health policy subsystem
connections to convince private firms not to join. The SMHPA was folded
when it became clear that it would not achieve profitability.

157

is perhaps not a surprise, since businessmen from large firms were pushing for an
innovation that would enable them to pool their purchasing power and lower their own
firrns’ health care costs. Given that the innovation existed solely within a market context,
and these businessmen run the firms that would benefit from lower health care costs, it is
safe to state that they understood the component of firm cost that results from the
provision of health care coverage to their own employees, and the potential for cost
savings that would arise as a result Of pooled purchasing. Thus the entrepreneurial team
did possess issue-specific knowledge, as predicted by the propositions.

For the variable of social/business connections, four key informants (40%,
medium validity) manifested nine appearances of the variable, with an aggregate score of
-5. This seems unusual given that the innovation succeeded in receiving support in a
formal vote of the C-P—P governing board, but is likely the result of unwillingness to
reveal guarded information on the part of a team of pure market entrepreneurs, as well as
their lack of connections to high-level staff within MDCH. In other words, the social
connections used to intensively lobby prior to the formal vote of the C-P-P governing
board are not manifested as appearances of the variable, yet we should expect that, based
on the literature review, personal, arm-twisting use of these connections probably
occurred. Thus the negative value is probably an artifact of hidden information, not the
absence of the activity.

On the variable of the development/use of a community venue as an alternate
venue for action, three key informants (3 0%, medium validity) manifested five

appearances with an aggregate score of -5. The negative score is not difficult to interpret,

158

as the only effort that was ultimately necessary to secure CCHIP support for the
innovation was the one formal C-P-P governing board vote. Absent this event, there was
considerable mistrust on the part of the community (in Calhoun this included a formal
CCHIP-sponsored member organization of approximately 900 persons) in the actual goal
of the innovation. One informant noted, “SO I think the [SM]HPA would have happened
either way [referring to the formal vote of the C-P-P governing board], but I just don’t
think it should have been given the stamp Of, that it was a community idea. Everybody
understands profit margins.” Thus we can state that the team of market entrepreneurs that
pushed for their innovation was ineffective in developing and using a community venue
as an alternate venue for action. It should be noted that, absent the one formal vote, the
use of a community venue was unnecessary to the development of a private-sector
innovation.

When assessing the market entrepreneurs’ ability to either alter costs or alter
perceptions of cost, we need to begin with the key informants’ perceptions of the costs of
the innovation at the outset. In Calhoun County, as in every other CCHMS demonstration
county, the initial perceptions of the costs Of any innovation to extend health care
coverage to the uninsured were negative. Six (60%, high validity) key informants
manifested 15 appearances of this variable, with an aggregate score Of-15.NO respondent
had a positive view of the costs of the innovation at the outset. However, three key
informants (30%, medium validity), manifested seven appearances of this variable, with
an overall score of 5. Thus, the market entrepreneurial team succeeded in changing

perceptions Of the cost of the innovation to critical partners through argumentation. What

159

 

is interesting is that the innovation itself was not altered to widen the coalition. Not a
single key informant manifested one appearance of the variable denoting an activity
where actual costs were altered for a key project partner.

With respect to the effects of the political and market contexts (existing contexts
and whether these contexts were undergoing change), the data tell different stories for
each context. There were no appearances of the effect of the local political context
(existing or undergoing change) variable, again not surprising as the innovation existed
solely in the market realm. However, the market context variables both appeared in the
interview data, with six key informants (60%, high validity) manifesting 24 appearances
of the existing market context variable, with an aggregate score Of -14. The imbalance
between the negative and positive appearances is likely the result, again, of information
being withheld by market entrepreneurs of their private assessments of the ripeness of the
market to support SMHPA. Some of the negative codes are from providers noting the
resistance of physicians to the fact that the nature of the health care market in Calhoun is
fragmented, with the western portion of the county’s market being actually an extension
of the larger market for health care services in the more populous county to the west
(where SMHPA was begun), Kalamazoo. Thus Calhoun County does not represent a self-
contained health care market for most health care services, but rather this market crosses
over the western border, and is to some extent dominated by the larger demand for and
supply of services in Kalamazoo County. A few negatively-coded appearances of the
local market undergoing change variable (3 0% of respondents, medium validity, four

appearances, score -4) relate to the absorption Of two locally-owned firms by out-Of-state

160

corporations, thus removing these firms from being members of SMHPA as the decisions
over how and where to purchase health care coverage for these firms’ employees was

suddenly moved out of state.

Summary of Calhoun County Case Study

Calhoun County was not an ideal choice for a demonstration site, because the
conflict expected to arise would occur within the same county (much Of the conflict in the
same city) where WKKF was located. Regardless, the march to developing and launching
an innovation diverged from what was predicted by this research early in the information
gathering stage, and the expected policy entrepreneur, the CCHIP project director, was
supplanted by a team of CEOs and high level businessmen from several large firms
within the county, including the Kellogg Company, which shares many board members
with the board of WKKF. Although the team of market entrepreneurs hid many of their
activities throughout the interviews, and from other key informants, a picture of weak
policy entrepreneur characteristics and activities emerges. Though scoring poorly on the
variables of coalition building, problem definition, use Of a community venue, and social
connections (this last one likely skewed by hiding information from others of their
activities), the team of market entrepreneurs scored well on the market entrepreneurship
variable, as well as the variable for altering key partners’ perceptions of the cost of the
innovation (if not altering the innovation’s costs). Thus they were able to garner enough

support for one critical formal vote of approval to Obtain CCHIP support Of SMHPA as

161

the Official CCHMS innovation that would some day provide health care coverage to the
uninsured, albeit without any formal mechanism or timetable for doing so.

The political context had little impact on this market-based innovation, and thus
no scores were recorded on either of these two variables. However, the market context
was viewed as largely negative because the western portion of Calhoun County is a subset
Of the larger Kalamazoo County health care market, which was perceived as threatening
to Calhoun County-based providers. Small negative changes to the local market were
manifested in terms Of the absorption of two local firms by larger corporations that were
based out of state, but it was not enough to prevent the SMHPA from moving through
development to launch. These two mergers did, however, remove these firms from the
pool of private firm stakeholders available for membership on the C-P-P Governing
Board and for discussions around the reallocation Of health care resources envisioned by

WKKF staff at the beginning of the Initiative.

Muskegon County Case Study42

The Muskegon Community Health Project (MCHP) was the organization that
formed to implement the CCHMS Initiative within Muskegon County. MCHP
successfully developed and launched an innovation addressing the community-wide
coverage goal of the CCHMS Initiative, ‘Access Health,’ that Offered an employer-based
health care coverage product targeted toward employers that do not traditionally provide

health care coverage to their employees. Within its first three years of Operation Access

 

‘2 A total Of nine interviews were completed for this case study.

162

Health was expected to provide coverage to 3,000 working uninsured, roughly one-third
of the uninsured within the county. Employer targeting occurred through two selection
criteria, (1) the median wage of employees is equal to or less than $lO/hour, and (2) the
employer has not purchased coverage for its employees in at least 12 months. The product
is known as a one-third share plan, where the employer pays for one-third Of the
premium, the employee pays for one-third of the premium, and the last third is paid for
through a mix Of state funds and federal matching funds. The use of state and federal
matching funds places the innovation within both the political and market realms.
Because the innovation is the Offering of a low-cost, employer-based health care coverage
product, MCHP chose to take a step to achieving the CCHMS initiative goal of
community-wide coverage using a voluntary mechanism that fit with the most prevalent
existing source of health care coverage. It should also be noted that the innovation is
loosely based on a similar product developed in Wayne County, the subject of this
research project’s fourth case study. Because the funding mechanism is the same as that
used by the Wayne County model, descriptions of the specific state law and federal
funding mechanism that permit it to operate are not repeated here, but can be found in
their entirety in the Wayne County case study.

An aspect of the Wayne County model must also be discussed here. In order to
Obtain state funds to be used with the employee and employer premium share payments as
a match for federal funds, and thus to firnd Access Health, the team Of collective
entrepreneurs also had to assume responsibility for the funds for the administration of the

State Medical Program (SMP) as administered in Muskegon County. These funds provide

163

reimbursement for outpatient services provided to qualifying indigent persons who are
not otherwise covered by some other form of public or private insurance. These funds
would need to be used to set-up a health care coverage product for the indigent population
within Muskegon County that was similar to PlusCare (discussed in more detail in the
Wayne County case study, below). The program and product that was ultimately
developed to Obtain these SMP funds for the federal match was entitled ‘Muskegon
Care.’43
There are two hospitals/health systems in the county, and unlike the other two
counties, managed care arrangements had been introduced by private sector health plans
approximately ten years prior to the beginning of the Initiative. One hospital is a member
of the larger Sisters of Mercy system, while the other, Hackley, was not affiliated with
any other hospital or health system. Muskegon was also unusual in that nearly all of the
primary care physicians are under contract to one or both of the physician health networks
within the county, each being a wholly-owned subsidiary of one Of the two hospitals. This

fact centralizes a great deal of supply-driven market power in the hands of the two health

systems.

 

‘3 One third of the funds used to pay for Access Health premiums come from

federal dollars that are matched to state and local dollars. All of the state
dollars used in the matching process come from existing SMP
expenditures for Muskegon County. In order to use the SMP dollars to
obtain a federal match, these SMP funds were transferred to the county,
which then had to assume the responsibility for providing care to the
population eligible for SMP assistance. The managed care program
developed for this purpose was ‘Muskegon Care.’

164

 

Residents of Muskegon County obtain nearly all of their health care services
within the county, with the exception Of tertiary care, which is Obtained in either Grand
Rapids or southeast Michigan. Thus the bulk of the local market for health care services
fits neatly within the county’s geographic and political boundaries.

The Muskegon County version Of the CCHMS Initiative began, as did the others,
with the formation of a C-P-P governing board, the hiring Of a project director, and the
formation of work groups to begin the detailed work of addressing the Initiative’s major
goals. The workgroup formed to address the CCHMS goal Of community wide coverage
was named the Uninsured Workgroup. Shortly after the formal set up of MCHP as the
organization that would act as the entity that pursued CCHMS goals, a project director
was hired to run MCHP.

Throughout the first two years of Operation, the C—P-P governing board was
deadlocked, and unable to take an active role in guiding the activities of the various
workgroups formed to pursue specific CCHMS goals. The board was deadlocked by the
resistance Of the members Of the local health care system, and the inability of the
representatives of the two hospitals to put aside their intense competition in the local
health care market long enough to consider how to pool their resources to achieve
common goals."4 This situation was seized by the project director as an opportunity to

start a broad spectrum of workgroups to address a variety of local public health concerns,

 

4" The two hospital systems had been fierce competitors for many years, and

the competition only increased when, prior to the funding of the CCHMS
initiative the Mercy hospital system absorbed a small osteopathic hospital
in Muskegon, leaving only two hospitals in the county.

165

as the attention of the board was not focused on these workgroup activities, but rather on
political infighting amongst its health system members.

Another key element Of the larger CCHMS process as it occurred in Muskegon
was the engagement of an experienced local media consultant by the project director to
aid her in information gathering and information use. This consultant had extensive
experience in both marketing and in politics, having sustained a private-sector media
consulting firm for many years in Muskegon, and prior to that conducting polling work
on two presidential election campaigns. The political expertise and local knowledge
offered by this consultant would prove invaluable to the project director.

MCHP’s project director early on realized that the complexity associated with the
issue of community-wide coverage would require more specialized knowledge and
expertise than was possessed by any current persons on staff at MCHP, including herself.
Although the project director was broadly familiar with the issue area of health care and
the problems of the uninsured, having previously served as a member on a former
Michigan Senator’s staff, she felt unable to become deeply involved in the day-tO-day
work required to successfully develop and launch an innovation while at the same time
running the rest of MCHP’s Operations. Thus the project director hired a Muskegon
county resident who had extensive prior experience in the design and administration of
health plans. The new hire was given the title of “Deputy Director,” highlighting both the
experience and status of the person hired and the importance placed on the goal of

community-wide coverage by the project director.

166

The hiring Of the deputy director, whose primary task was to guide the Uninsured
Workgroup to its goal Of developing and launching an innovation that would expand
access to health care coverage, is the first step taken by the project director to develop a
collective entrepreneurship team. This step towards developing a collective
entrepreneurship team is the first signal that the MCHP project director possessed enough
of the qualities of a policy entrepreneur to understand the context within she was
working, the complexity of the issue faced, and the skills that would be required to
achieve the goal.

As in the Calhoun County, the members of the Muskegon Uninsured Workgroup
who were not employed in the health care industry were daunted by what they did not
know about how health care is purchased, supplied and delivered. In the same way that a
policy subsystem develops, the persons who were not employed in the health care
industry tended to defer to the judgement of those who were. However, one key member
Of the Uninsured Workgroup was a respected retired physician who was willing to think
broadly about health care and the ways in which a community could consider expanding
coverage to uninsured persons. This key member, along with the deputy director,
provided the workgroup with considerable issue area expertise from outside of the local
health policy subsystem.

The Uninsured Workgroup began a nationwide search of local-level initiatives to
expand access to coverage for the uninsured. The search, as in the other two counties, was
broad, and yielded several different models, including the one-third share plan model

eventually settled on. This process consumed several months.

167

Recognizing that the members of the workgroup who were from outside the health
care industry were operating with lower levels of knowledge, the deputy director
undertook to educate these members on the complexities of building a health care
coverage product. This approach was reported by respondents to have worked, but shortly
after this process was complete the deputy director left MCHP for another career
opportunity.

At this point the project director reentered the meetings of the workgroup on a
more regular basis, and once again realized that the workgroup required extensive expert
guidance if it were to move forward towards its goal. At this point she hired a consultant
with a background similar to the deputy director’s, but with even more experience within
the field, including participating in the development and launch of an innovation targeting
the uninsured in a community in Florida, with prior business and social connections to the
Director of the Medical Services Administration (the Medicaid agency) at the Michigan
Department of Community Health. The consultant reconstituted the workgroup (which
had a several month hiatus from meetings) and placed several new members on the
workgroup who were from the local health care system, but who she knew were
interested in pursuing innovation."5 It was with this membership that the workgroup chose
a model for innovation and went through the detailed work Of tailoring the model to the
needs of the stakeholders in Muskegon who would have to agree (and sign binding

contracts) to support the launch of the innovation.

 

4’ These new physician members were described as “young and enthusiastic

forward-thinkers.”

168

The development of the details Of what services would be covered and how were
delegated to a new group formed by the consultant, the Physicians Subcommittee. This
group met regularly to assess what services would be covered and to set up fee schedules
and the mechanism by which persons with the coverage would access services. Although
the membership was largely physicians the directors of the two physician networks
owned by the two hospitals, and two pharmacists, one from each hospital, were also
members. The chair of the subcommittee (chosen by the project director) described the
reasoning behind its membership composition, “Providers have for a long time been burnt
by HMOs in terms of losing dollars, etc. One Of the most interesting things I saw was
when I had the Opportunity to lead up the [subcommittee] and when I came to them and
said ‘Look, what I have been told is we get to build it, we tell them what the benefits need
to be, and we tell them what we want to be reimbursed for it. And they’ll try and work it
out. it was kind of a revelation [that] you are going to include us in this product, you
are going to allow us to make it ours.” Thus the choice of forming a subcommittee made
up mostly of physicians (specially chosen for their support of innovation) gave the
ultimate product designed a moral ‘leg up’ when it was eventually marketed to physicians
and the two hospitals for contracting.

The product designed was a “Cadillac” model, meaning it had a remarkably rich
benefits package with low OOpays and broad access to services. However, risk was
significantly reduced because the subcommittee decided to limit access to services
provided in Muskegon by Muskegon area providers. Thus no services were covered that

were used outside of Muskegon. Additionally, the program limited the scope of coverage

169

for catastrophic (re: very expensive) services. Given the likely income levels of the
persons covered by Access Health, it was decided that these persons could Obtain
coverage for these low probability high cost services through Medicaid’s Spend-down
program, which provides reimbursement for the services used by persons who experience
catastrophic health events.46 This decision dramatically reduced the risk to providers who
signed contracts to provide services to persons enrolled in Access Health.

At roughly the time that the consultant replaced the deputy director and had
reconstituted the Uninsured Workgroup, C-P-P governing board disagreements between
the representatives from the two hospital systems reached a head, and one Board member
complained to a member of the local media. The result was a newspaper article critical of
the C-P-P governing board and its inability to reach agreement on what it was, and what it
was supposed to be, doing. This news event apparently embarrassed all Of the members of
the governing board, and they agreed to hold a retreat with an outside facilitator to work
through their disagreements. The outcome of the retreat was an agreement as reflective of
the fears of both hospitals concerning competition and loss of control which outlined that
the board would not:

1. Become a certificate Of need agency (an organization that would determine how

many of what kind of services should be permitted within a community);

 

4" Although they are normally not eligible for Medicaid, persons who are in

this income category can become eligible for the Medicaid spend-down
program if, in any given month, their health care costs, subtracted from
their income, would lower their household income to a level where they
would be eligible for Medicaid benefits. Thus Medicaid spend-down
eligibility is a month-by-month catastrophic health care safety net for the
lower income working uninsured.

170

2. Take control of large streams of funding and redirect them; or
3. Become a “super board” that would control the two health systems.

After reaching these collective decisions, which, incidentally, ran counter to the
larger goals of the CCHMS initiative, the hospital representatives felt able to cooperate
and cease bickering about what the board could do. However, once the board came out of
its multi-year period Of self-imposed gridlock, the project director reported, “... [the
board] suddenly realized that the community had moved ahead of it, and not just on
Access Health.” Essentially the board had been faced with a fait accompli across several
projects being developed by workgroups that had been working for several years on
various CCHMS goals. In the case of Access Health, they were faced with a project that
had already developed broad community support, had support from within the physician
community, and was backed by extensive causal arguments presented by a variety of high
status members Of the community, who also had access to the supportive results of
telephone surveys.

Because the chair of the Physicians Subcommittee was also the director of one
hospital system’s physician network, that hospital CEO was easily convinced of the value
of Access Health, once the guarantees against high risk health events had been explained.
The same was not true of the other hospital CEO, and the process by which that hospital
system agreed to the signing of the contract necessary to become a provider of services
for persons enrolled in Access Health is murky, as should probably be expected Of market
entrepreneur activities. However, one respondent reported that the reason that this

hospital eventually signed was the belief that Access Health would launch without them,

171

and that the negative media fallout from not being in at the beginning of a program and a
product designed to provide health care coverage to the working uninsured would
seriously damage the hospital’s reputation.

At this point we can assess the entrepreneurial qualities Of the project director and
her collective entrepreneurship team. The lead member of the team was the project
director, the other members of the team included the retired physician, the consultant
coordinating the workgroup, and several key members of the workgroup placed there by
the consultant. An occasional member of the team was the local media consultant, who
aided the workgroup in information gathering within Muskegon through telephone survey
interviews, and in the provision of information to local media to stimulate interest in
workgroup activities, and support for innovation. Lastly, once the workgroup decided on
a model, a Lansing-based consultant was hired who had set-up the one-third share plan in
Detroit used as a model for Muskegon. This consultant had previously worked in the
state’s Medicaid agency, and due to her ongoing work in health care financing at a
national level health care consulting firm, had expanded her connections from that time
within the state bureaucracy.47

For this case only, scores on the twelve policy entrepreneur variables will be
presented for both the collective team and for the project director. This is done because
this case is one of two with a clear team of collective entrepreneurs, and because, unlike

the other clear case Of collective entrepreneurship, Wayne County, enough data exists to

 

47 This consultant was also interviewed for the Wayne County case study.

This was the only person interviewed twice for two separate cases.

172

 

perform this extra analysis. This analysis is based on appearances of these ten variables
where it is clear in the appearance that the project director is named as a part of the
appearance, or, if no individual is referred to with respect to an event or situation, it is
known that the project director was directly involved. The analysis of the project
director’s involvement in these variables is probably still understated, as many
appearances of these variables were difficult to trace back to discrete individuals. Likely
this is because this collective team was rather self-effacing in interviews, and often used
the term “we” when referring to entrepreneurial activities, rather than to name a
themselves, or other specific individuals or subset of individuals. This also likely
accounts for the attainment of no more than medium validity on all but one policy
entrepreneur variable by the project director.

It is appropriate now to present the scores on those local policy entrepreneur
variables that achieved medium or high validity. I will begin with the collective
entrepreneurship variable, as Muskegon represents the strongest manifestation Of this
variable, and the preceding discussion has highlighted the extent to which a team Of
entrepreneurs was involved in the development and launch of Access Health. The
collective score for this variable was 35 (89%, high validity), from a total of 37
appearances of the variable. This score represents the high levels of activity by each of the
members of the collective entrepreneurship team, and is a score far higher than that
manifested in either of the other two counties. The project director’s score was 3 (44%,

medium validity) from a total 5 appearances traceable to her.

173

On the variable of issue-specific knowledge, the score was 19, (78%, high
validity) from a total of 27 appearances Of the variable. The project director’s score was 1
(33%, medium validity) from a total of 3 appearances traceable to her. This strong
positive score represents the high levels Of issue-specific knowledge possessed by the
team.

On the variable of pre-existing local and state connections, the score was 24
(56%, high validity) from a total of 25 appearances of the variable. The project director’s
score was 6 (44%, medium validity) from a total of 6 appearances of the variable. This
high score represents the remarkably high level of connections that the project director
and members of the team had both locally (all but the Lansing-based consultant were
from Muskegon, and resided in Muskegon), and at the state level in both the bureaucracy
and the state legislature.

On the variable concerning the ability to make persuasive causal arguments, the
score was 29 (78%, high validity) from a total of 31 appearances of the variable. The
project director’s score was 5 (44%, medium validity) from a total Of 5 appearances of the
variable. This high score represents a team with remarkably high argumentation skills that
were transferable across different frames. Many of the appearances of this variable were
tied to two individuals, the project director and the consultant hired to replace the deputy
director. Both individuals appear in the transcripts as people who spent a great deal of
time convincing key stakeholders that the positive benefit Of extending coverage through
the chosen innovation would be worth the risk associated with signing the necessary

contracts. The consultant spent a great deal of time asking opposing members of the

174

 

workgroup to offer ways to alter aspects Of the model to make it work, and then worked
these changes into the program. This refinement process was facilitated by her creating a
subcommittee to the workgroup made up of local physicians and their representatives
(whom she selected) who spent a great deal of time developing the specifics Of what
services would be covered, the restrictions on these services, and how persons with
coverage would access services.

On the variable of coalition building, the score was 28 (89%, high validity) from
a total of 48 appearances of the variable. The project director’s score was 8 (44%,
medium validity), from a total of 8 appearances of the variable. The few negative
appearances of this variable are almost all related to remarks on the abrasive personality
of the consultant hired to coordinate the workgroup after the departure Of the deputy
director. Some respondents described her personality as “pushy” and this view was
reported by even those respondents who were always supportive of innovation. However,
it is worthwhile to note that the respondents who were members Of the collective
entrepreneurship team believed that it was her tenaciousness that caused many to view
her as sometimes frustrating the building of the supportive coalition. This tenacious
aspect of her personality, and her unwillingness to end the pursuit of an innovative
solution, that contributed to the high score on the causal argumentation variable, noted
above. It is also worthwhile to note that the project director’s score, likely to be
understated, accounts for nearly a third Of the positive score.

On information gathering, the score was 26 (89%, high validity) of a total of 26

responses. The project director’s score was 10 (44%, medium validity) from a total of 10

175

appearances of the variable. The scope of information gathering efforts was far broader in
Muskegon than in either Of the other two counties. The project director’s understated
score demonstrates that she was a main driving force behind this effort. In addition to the
search for existing innovations that was also conducted in Calhoun and St. Clair, and the
collection of information concerning health care costs within the county as was conducted
in Calhoun, MCHP also commissioned several telephone surveys throughout the process
that (among other things) assessed the extent of the uninsured within the county, the types
of firms that routinely did not offer health care coverage to their employees (and why),
and the extent Of public and employer support within the county for an innovative health
care coverage product targeting firms that did not Offer health care coverage. This
information was used to shape and support the team’s causal arguments, and was
provided to the local media to help shape public opinion in favor of innovation. The
information was also routinely shared with the county Board of Commissioners at
regularly scheduled public hearings, which also received media coverage.

On the variable of problem definition, the score was 35 (78%, high validity) of a
total of 35 appearances of the variable. The project director’s score had low validity, and
thus, is not reported. Again, the high score represents a remarkable level of agreement on
the problem faced. Although all three counties had positive scores on this variable, what
is interesting about Muskegon’s score is that there were no negative appearances Of the
variable. This means that even Opponents Of the innovation were in agreement concerning
the problem faced. Of course, agreement on the scope of a problem does not necessarily

lead to agreement on the methods of alleviating or resolving the problem. However, what

176

is clear from the transcripts is that the collective entrepreneurship team did not have to
expend much effort during problem definition bringing all stakeholders to a consensus on
the scope Of the problem to be addressed.

On the variable Of use of the media as an alternate venue, the score was 7 (44%,
medium validity), from a total of seven appearances of this variable. The project
director’s score was 5 (3 3%, medium validity) from a total of 5 appearances Of the
variable. This suggests that she was the primary driving force behind the effort to use the
media as a venue. Only Muskegon registered any appearance of this variable, and this is
likely due to Muskegon being the only funded county to retain a media consultant.
Although the score is low, and possessing medium validity, the fact that the score is
positive (with no negative appearances), demonstrates that the project director knew that
using the media as an alternate venue for action was an Option, and pursued it through the
hiring of a media consultant, and the targeted release Of information to local media to
create a favorable public image.

The thought and effort that went into MCHP’s use of the media is captured from
the transcript of the media consultant. “Your product better be able to deliver, and you’re
going to need to know that your physician community stands behind what you’re going to
market. You’re going to have to do it in an environment where the average person is
bombarded with 50,000 words everyday. How do you get your message through? SO,
we’ve had to develop a media plan that is going to clearly penetrate that, it’s going to use

each piece Of media for what it does best and not make it do what it doesn’t do.”

177

On the variable of use of the community venue, the score was 19 (67%, high
validity) Of a total of 19 appearances of the variable. The project director’s score was 6
(44%, medium validity) from a total of 6 appearances of the variable. The project director
can thus be traced to almost one third of the total positive score. Again, Muskegon scored
well above the other two cases, demonstrating the understanding of the project director
that this venue was available, and also how to use it. As previously mentioned, the project
director seeded workgroups across all of the CCHMS goals in the absence of the attention
of the C-P-P governing board. Unlike the other two project directors, she apparently did
not believe she required the prior approval of the C-P—P governing board to form
workgroups to pursue CCHMS goals. Instead, she perceived this activity as within her
authority as the project director, and merely reported her activities at their regularly
scheduled meetings, where no challenges concerning her activities were raised.48 More
importantly, these workgroups were perceived to have, on average, more balanced
representation from the different constituencies within the community. The former deputy
director Of MCHP compared her work there with a current community project in another
community, “... there was certainly support in some of the broad groups of people who
participated in health project activity. Actually, very good community representation. I
say that, in part, because the project I am currently working on in Grand Rapids has

limited community input, compared to what I saw in Muskegon. I think from the

 

‘8 It should be added that MCHP’s project director, unlike her two
counterparts, also perceived her spending authority over discretionary
funds more broadly, which is perhaps why she was able to hire more
consultants than did the others, and also spent far more on information
gathering efforts.

178

community perspective that she’s [the project director] done a wonderful job at reaching
out to people who previously had no experience with health care delivery issues.”

The score on the market entrepreneur variable was 6 (22%, low validity), from a
total of six responses. The project director registered no appearances Of this variable. This
low score (albeit, positive) with low validity makes it difficult to draw conclusions
concerning how well the collective team performed in this role. However, the score was
positive, and the following quote from a transcript is illustrative of how key physician
members of the collective entrepreneurship team convinced physician and hospital
stakeholders to support a market-based innovation of which they had originally been very
skeptical, “[U]ltimately, it comes down to will they [the hospitals and physician groups]
sign the contracts we’ve put out there and take our fee schedule, and they said yes.
[W]e have several physicians on the Board of Access Health [the innovation], I think that
they are just enthusiastic. The younger physicians are spinning it in a very positive
way. And when I came to this project everything was about how we were going to come
in and tell them how to do their business. There was great suspicion around it. This has
been a major flip-flop.”

Despite the low validity of the score on the market entrepreneur variable, we get a
sense of these activities by examining the scores Of the perceptions of the cost of an
innovation, altering the perceptions of cost of the innovation, and the altering of the costs
of the innovation. The initial perceptions Of the cost of any innovation to extend health
care coverage to the uninsured, as in all three cases was negative, with a score of -20

(78%, high validity) of a total of 34 responses. This is consistent across communities

179

probably because if this issue had been deemed easy to tackle, it would have been
resolved long ago.

However, the scores for altering perceptions of cost and altering the actual costs
of the innovation were both positive, with no negative manifestations of either variable.
The variables’ appearances are tied to all members of the collective entrepreneurship
team. For the variable of altering stakeholders’ perceptions of the costs of the innovation,
the score was 14 (67%, high validity), from a total Of 14 appearances of the variable. The
project director’s score was 3 (33%, medium validity), from a total of 3 appearances. This
means that the members of the team, including the project director, were adept at
changing stakeholders’ negative perceptions of the cost of an innovation through
argumentation, supported by the high score reported above for that variable.

For the variable of altering the actual costs of the innovation to stakeholders, the
score was still higher, with a 21 (67%, high validity) from a total of 21 responses. The
project director’s score was 6 (56%, high validity), from a total of 6 appearances. This
means that the members of the team were willing to alter the nature of the innovation to
accommodate the cost concerns of some key stakeholders, while maintaining a cohesive
coalition of supporters. It should be pointed out that not all stakeholders were brought on
in this way. It is noticeable that this willingness to alter the actual costs of the innovation
to the stakeholders has ahnost one third of its appearances traceable to the project
director. This suggests that she may have set the tone for compromise amongst the
collective team of entrepreneurs. One of the most important stakeholders, a hospital,

appeared to eventually sign on only after becoming convinced that Access Health would

180

launch with or without them, and it was the perception that this would result in negative
impacts on their image that led them to sign their contract.

Of the four context variables, only one manifested any appearances. The
favorableness of the local market context had a score of 11 (67%, high validity), from a
total Of 24 appearances. This is one Of the most mixed scores from the Muskegon case,
which is a positive score, nonetheless. This means that the local market context was
perceived tO be, in general, favorable to an effort to develop and launch an innovation
targeting the working uninsured. What is interesting, and confirmatory, was that it was
the weakness of some of the actors in the local health policy subsystem that was cited as
making the context favorable. The physicians, empowered to participate in the design of
the benefits plan, were eager to participate, given that their experiences of control over
benefits and reimbursement had been usurped by HMOs for years. The fact that no other
context variables appeared in the transcripts will be discussed in the following section on

assessing the propositions.

Summary of Muskegon County Case Study

The Muskegon County CCHMS project successfully developed and launched an
innovation that was planned to extend health care coverage to approximately 3,000
working uninsured in the county, roughly one-third of the uninsured within the county. At
the same time, although not a subject of this study, the team also assumed control of
county SMP fimds and developed a health care coverage program and product to cover

the indigent in the county, thus permitting the use of state funds for drawing down further

181

matching federal funds to support Access Health. Although this fell short of the CCHMs’
goal of providing community wide coverage, this innovation went much farther towards
that goal than was achieved in the other two funded counties. The project director and the
members of the collective entrepreneurship team that she formed manifested high
positive scores on virtually all of the entrepreneurship variables, and many of the scores
incorporated either no or very few negative appearances Of the variables. The project
director successfully used alternate venues for action which were left unused by the other
two project directors, including the local media and the community. The local political
context variable, the changing political context variable, and the changing market context
variable all had zero appearances in the transcripts. However, the local market context
variable had a positive score, albeit with some negative appearances Of the variable. Thus
the only local contextual impact on the development and launch of the innovation was a
generally favorable market context. In this case “favorable’ refers to the weakening of
some of the actors in the local health policy subsystem, making them interested in
receiving the Opportunity, once Offered, in participating in the design Of the innovation.
This weakening in the local health policy subsystem also meant that some of the most
weakened actors, in the case of Muskegon, physicians, would have two thoughts
concerning innovation. First, that if they could be re—empowered in some way, they would
accept the opportunity, and second, that their allegiances to other members of the health
policy subsystem were weak enough that they would be willing to support an innovation
that they helped design in the face of strong Opposition from one of the other members of

the subsystem.

182

St. Clair County Case Study49

The CCHMS initiative in St. Clair County, known as CCHMS of St. Clair County,
was the only site of the three that was defunded prior to its originally scheduled end date.
As a result of shortfalls in projected revenue from its endowment, which were driven by
huge losses in the stock market, WKKF staff were forced to look across all multi-year
funding commitments and make hard decisions to make funding cuts. WKKF staff
continue to be reticent about how they made decisions on which grantees to cut, but it
seems likely that CCHMS of St. Clair was out due to the inability of the project director
(the third person in that job) to form a strong coalition and make progress towards any of
the CCHMS Initiative’s goals. Thus CCHMS Of St. Clair County was defunded two years
ahead of its planned end date.

As mentioned, CCHMS Of St. Clair had a total Of three project directors
throughout the course of its four-year life span. Throughout its first year and a half, the
head of the community foundation acted as its project director, then a permanent full time
director was hired, and held the post for about a year and a half. The second director was
apparently hired by members of the C-P-P governing board as a favor, and because some
on the C-P-P governing board deemed him to be a person who would not vigorously
pursue innovation. A year and a half later, WKKF staff dissatisfaction with the second

project director’s performance led to this person’s removal, and the third and final project

 

‘9 A total of five interviews were conducted for this case study, one Of which

was not coded because the respondent was unwilling to answer most of the
questions. With so few interviews, only scores with high validity (2 or
more respondents with at least one manifestation of a variable) are
reported.

183

director was hired. The third director was hired from outside of the county; the only of the
three project directors who did not have existing deep roots within the county. The lack of
consistent, strong project leadership appears to be part of the reason why CCHMS of St.
Clair County never moved beyond the information gathering stage.

Unlike in Calhoun, where the development and launch of an innovation in health
care coverage was taken over by business leaders, and in Muskegon where the project
director Operated with a high degree of independence in the first several years without an
attentive C-P-P governing board, the governing board in St. Clair paid close attention to
the activities of the project director and project staff, and also kept a tight rein on all
project expenditures. This situation was not one that was perceived to be amenable to
change by the third and final project director.

The St. Clair market for health care services is a subset of the largest market for
health care services in Michigan, that of greater southeast Michigan or Detroit. Thus, few
high level services provided in the county attract the bulk Of county residents who require
those services; many residents leave the county to obtain these services within the greater
Detroit area. Only primary and secondary care services attract more than half of those
who consume services within the county. This is a sharp contrast with Muskegon.

The health care providers (hospitals and physicians) were largely insulated from
the changes affecting health care markets across the nation in the decade prior to funding.
Not a single managed care organization Of any kind Operated in the county until several

years after the start of funding, and even today the penetration of managed care

184

arrangements is below ten percent of the covered population (Michigan Insurance Bureau,
2000)

There was little competition amongst the three hospitals in the county until shortly
before the arrival of WKKF funding. One hospital, River District, is very small and Offers
few services. The other two hospitals, only a couple Of miles apart in the county’s largest
city, did not treat each other as competitors until after a 1993 Federal Trade Commission
ruling prevented the merger of the two hospitals. Thus hospital competition did not begin
in earnest until about the time that CCHMS of St. Clair was funded. However, all
respondents reported that competition, once begun, was fierce, and seriously hindered C-
P-P governing board efforts at identifying common interests and working on projects that
both hospitals could agree to support.

The locus of health care coverage purchasing activity was also the least favorable
of all Of the counties. Most Of the employed who work in the county are employed by
entities at one of the three levels Of government, and all Of its large employers (with the
exception of its hospitals) are owned by parent companies external to the county. More
importantly, many employed residents work for firms outside of the county.

The stunted effort in St. Clair County resulted in few interviews (there were few
people involved in addressing the problem of extending health care coverage) and many
variables never manifested a single appearance. The score on the information gathering
variable was 7 (50%, high validity) from a total of 11 appearances of the variable. This
suggests that the effort to gather information was largely successful, and in fact, the

information gathered on innovations elsewhere was as extensive as gathered by the other

185

two CCHMS projects. There was also an effort to collect and analyze information to
understand the extent of uninsured within St. Clair County, although the report produced
appears to have been ignored, or at least not used by anyone either among the project staff
or the C-P-P governing board.

Because literally no other efforts outside of gathering information on innovations
elsewhere and the report on the uninsured within St. Clair County occurred, we see few
appearances of the other variables, but it is appropriate to report on the few that did
register appearances. The variable for knowledge Of the issue area registered a score of l
(50%, high validity) from a total of 6 appearances of the variable. This suggests that the
knowledge possessed by the different project directors was at best mediocre, and probably
was one of many reasons for the failure to move beyond information gathering.

The variable for local connections Of the policy entrepreneur scored a 0 (50%,
high validity) from a total of 2 appearances of the variable. The low number of responses
makes interpreting this score unwise, but since we know that the final project director
was new to the county, we should be surprised that it was not negative. This suggests that
perhaps the project director had some facility in establishing contacts once she arrived.

On the variable Of coalition building the score was —1 8 (100%, high validity), from
a total of 18 appearances of the variable. This suggests that there were no successful
efforts at coalition building throughout the project. This is a very large negative score
given the number of respondents, and the fact that all respondents registered several
negative appearances of the variable bears further discussion. One CCHMS of St. Clair

staff respondent, discussing the effort to develop support for a Muskegon-like innovation,

186

reported, “... no one in the county, no one institution, hospitals, the health department,
F IA, none Of them was willing to say we’ll take a lead on this and take the risk here.”
Another respondent, referring to the last project director’s effort to develop and launch an
innovation targeting the uninsured said, “... the second thing is that it ran into a buzz saw,
no matter how good you worked, where we had two hospital that are in major
competition and throwing a lot of dollars at it [act of competing].” Another staff member
reported concerning members of the C-P-P governing board, “I think people wanted to
support her [the third project director], but they wouldn’t go as far as supporting her
based on what their organization could bring to the table, more supported conceptually,
maybe her altruism.” This suggests that the members of the C-P-P governing board
remained first and foremost representatives of the organizations for which they worked,
and were members of a board that was supposed to pursue collective change a distant
second. It also suggests that these members perceived the entire CCHMS effort as
unrealistic and unachievable.

On the variable of collective entrepreneurship, the score was -5 (75%, high
validity) from a total of 9 appearances Of the variable. This score suggests that an effort to
build a collective team of entrepreneurs failed, and indeed, there appear to have been at
best three members of the team, one of whom was a consultant who became frustrated
after a year Of effort and left the project for a job outside of the health issue area. Staff
turnover at the project seems to be partially to blame, not only at the project director

position but also in all positions, including secretarial. No clues are found in the

187

transcripts as to why this was the case. The only positive scores refer to the one
consultant and to a local physician who had long been an advocate for the uninsured.

The variable of problem definition yielded a score of 1 (75%, high validity) from
a total of 11 appearances Of the variable. This very low positive score is likely indicative
Of the difficulties noted earlier of defining the problem in a way that satisfied the key
members of the governing board, the two largest hospitals. This score is low despite the
commissioning of the sole effort to gather local-level information on the uninsured within
St. Clair County, a report that nonetheless could not tell the governing board or the
Medicaid/Uninsured Workgroup (the small group formed to study the issue) how many
uninsured existed in the county.50 This is not surprising since the only estimates of the
uninsured available come from state-level surveys conducted by the Bureau of the
Census. Thus only an investment in a survey of St. Clair County residents, such as was
conducted by a private firm for MCHP in Muskegon, would have supplied the necessary
information. One respondent, when asked about problem definition, said, “It [the
definition] was a moving target. There wasn’t a strategy [a deliberate approach to
problem definition].”

What is not surprising in St. Clair County is that it registered high validity scores
on two of the variables related to the political and market context. For the variable of the

amenability of the existing local political context for innovation, the score was -13 (50%,

 

’0 The very title of the workgroup formed suggests the problems that

CCHMS of St. Clair had in defining the nature of the problem it needed to
address. The concern over the Medicaid population was one of access to
health care services, while for the uninsured it is access to health care
coverage.

188

high validity) from a total of 13 appearances of the variable. This high and uniformly
negative score is related to statements that refer to the two different issues cited by
respondents, an insular, conservative populace, and an “Old boys’ network” Of influentials
within the county who controlled the C-P-P governing board and felt unable to act in an
innovative manner and remain influential." One respondent, speaking to this issue and
what was perceived as the stonewalling of the Medicaid/Uninsured Workgroup’s
activities, noted “I’m sure [a Board member] knew she was going to run for County
Commissioner, and she didn’t want to do anything to mess that up.”

For the variable of the amenability of the local market context at the start of the
project, the score was -16 (75%, high validity) from a total of 16 appearances of the
variable. These negative appearances mostly relate to the failure of the two hospitals to
merge, and the fierce competition that broke out between them in the aftermath of the
merger. The remainder refer to the absence of HMOs within the county, and to the

absence of a constituency of large, local employers fi'ustrated with high health care costs.

Summary of St. Clair County Case Study

The summary of the events that occurred with the funding Of the CCHMS
initiative in St. Clair County is one of a group of local influentials placed on the C-P-P
governing board who appear by all accounts uninterested in pursuing the kinds of

innovation for which they were funded. The governing board moved slowly to hire a full

 

5' The gender used is inaccurate, several women, one a hospital CEO, were

considered members of this network.

189

time project director, and then finally hired a local who was deemed unfit by those
respondents who referred to this person. After being removed at the request of a
foundation that had otherwise taken great pains to avoid any conflict-raising issues, a
project director who reportedly was more competent was hired, but hired too late to save
the project’s prospects according to three respondents. Then, when the foundation ran
into budget shortfalls, funding was discontinued before either problem definition or
information gathering was ever considered complete.

The effort to develop and launch an innovation addressing the uninsured never
moved past the early stages of information gathering, and the members of the C-P-P
governing board were never able to agree on commitment to any model, agree to develop
a new model, nor appear to ever have used the information contained in the one report on
the state of the uninsured within St. Clair County. The only positive effect on the issue
area of the uninsured within the county appears to be that some members of the
Medicaid/Uninsured Workgroup agreed to continue to meet to discuss the issue after the

termination of WKKF funding.

190

Wayne County Case Study52

The Wayne County case study, unlike the CCHMS county case studies, is an
example of inside-outward change. NO external change agent selected the county and
offered to fund a change effort. Thus, this is a case of the appearance of a local-level
policy entrepreneur who saw an opportunity to pursue an innovation and seized it. The
roots Of the innovation that was developed and launched in Wayne County differ in many
ways from the other three county case studies.

Prior to discussing these roots, it is necessary to detail the differences between
Wayne County and the CCHMS counties. Wayne County is by far the most populous
county in Michigan. Its large population results in a large tax base, and a high demand for
county govemment-provided services. As a result, Wayne County also has the largest and
most articulated county government structure in Michigan, in order to respond to the
demand for services. Wayne County also has the largest population of Medicaid eligibles,
the largest population Of persons below poverty, and the largest population of uninsured.

The roots of the innovation that was eventually developed and launched began in
the 1980's when a system known as County Care replaced two previous systems of

providing health care services to the indigent and uninsured. These two systems were

 

’2 Four interviews were conducted for this case study. A high number of
refusals were encountered, resulting in fewer opportunities for interviews.
Only one of the interviews was taped and lasted an hour, the other three
were conducted over the phone and were no more than thirty minutes each.
Lastly, the shortness of the three interviews resulted in less Opportunity for
the appearance of variables. Due to the small number of interviews, only
variables with high validity (2 or more respondents manifesting the
variable) are reported.

191

general assistance (GA) and resident county hospitalization (RCH). State law prior to the
enactment Of County Care in Wayne County required that counties share in the cost of
providing fee for service outpatient and inpatient care to the poor and uninsured.
Reimbursements for services supplied under both programs were well below existing
Medicaid rates, which were at that time 90% of usual and customary rates.53 In 1979, state
law was changed (after strong lobbying by hospitals) to require the low reimbursement
levels for the RCH program to be raised to the equivalent of existing Medicaid rates for
reimbursements. This change in reimbursement levels, along with the onset Of steep rises
in health care costs in the 19703 and early 19803, found counties unable to provide
enough funding to support their share of the costs, and the state was unwilling to assume
more Of the cost burden. SO state law was changed and these two systems of care ended.
The senior elected officials within the Wayne County Executive, (lobbied hard by
the CEOs of inner city Detroit hospitals), determined that the cost Of providing services to
the indigent population (who would inevitably appear in the emergency rooms of its inner
city hospitals) would result in huge cost overruns that would dramatically affect these
hospitals’ ability to remain in operation. Thus County Care, funded by Wayne County,
was enacted. County Care immediately ran into debt, and Wayne County appealed to both
the state legislative and executive branches of government for help. The state covered the
existing Wayne County debt, and altered state law again in 1987 to permit the setting up

of PlusCare, a managed care program that provided health care coverage to the most

 

’3 At that time “usual and customary” reimbursement for services was set by

the Michigan Insurance Bureau based on an assessment of costs for
various services within predefined geographic regions.

192

indigent residents of the county. This law permitted Wayne county to set up municipal
health facility corporations which would be responsible for establishing, modifying,
Operating and managing health services (RA. 1987, NO. 230). In Wayne County the
nonprofit corporation set up to receive the funding and manage PlusCare was entitled
Urban Hospital Care Plus (U HCP).

This PlusCare program was funded by Wayne County, state funding, and an
innovative matching of federal funds available to hospitals that provided a
disproportionate share of their services to persons unable to pay for services (GAO,
1994).54 This system was investigated by the General Accounting Office, which lobbied
Congress to alter federal law to prevent this activity from occurring in Michigan,
Tennessee and Texas. Of the three states investigated, only Michigan used this
mechanism to fund a county-level indigent care program. The GAO report provides a
detailed assessment of the funding mechanism that permitted the state and Wayne County
to supplement their combined funds to attract federal dollars to help support the PlusCare
program. The process began with Wayne County contributing $15.5 million and the state
of Michigan contributing $7 million to seven urban Detroit hospitals. The federal
government then provided matching DSH funds totaling $28.5 million to the seven
hospitals at a match of 79 cents on the dollar to the seven hospitals, who then forwarded

the entire $51 million to UHCP to administer and run PlusCare (GAO, 1994).

 

’4 A hospital is eligible for DSH payments when it provides at least six
percent of its services to the indigent patients in the state (GAO, 1994).

193

After several years of Operation, senior Wayne County Executive officials noted
that UHCP was running in the black, because the costs Of running PlusCare were below
projections and less than the total of county, state and federal funds flowing into UHCP.
At that time the county engaged a health care consulting firm and a legal firm to
determine if these funds could be used to fund an additional program that would provide
health care coverage to the working uninsured. This activity was in effect the formal
gathering of a collective entrepreneurship team. The members Of the team included three
Wayne County Executive officials (one the senior elected Official within the county), and
the consultants from both the consulting firm and the legal firm.

The program that was developed was based on a model that members Of the health
care consulting firm had developed when these persons had been employees of the state’s
Medicaid agency, and the agency had funded through a Robert Wood Johnson Foundation
grant.55 This prior program set up a one-third share plan, where premiums for a health
care coverage product were funded through contributions of one-third from the employer,
one-third from the employee, and one-third from grant funds. The model developed by
the health care consulting firm used the same methodology, with UHCP contributing the
one-third share provided by grant funding in the earlier program. The reimbursement
levels for the services Offered by HealthChoice were slightly below Medicaid

reimbursement levels for similar services.

 

’5 This program, known as the Health Care Access Project, ended when

foundation funding ended.

194

After deciding on a general model for a new health care coverage product, senior
Wayne County Executive officials sought to sell the proposed program, known as
HealthChoice, to the UHCP Board and the two inner city Detroit health systems that
would be asked to sign contracts to provide the services offered in the health care
coverage product. These officials presented the model to the UHCP Board and senior
executive staff at the two health systems, arguing that the losses incurred by the health
systems accepting lower reimbursement levels under the proposed HealthChoice product
would be more than made up by the fact that ultimately the health systems provided
services to this population anyway, when they became ill enough to seek care through
emergency rooms. Because the emergency room care was uncompensated, the hospitals
ultimately lost money on the proposed HealthChoice population, and thus by receiving
some reimbursement for services the hospitals would, in the end, realize an increase in
income for services rendered. As well, the Wayne County Executive officials and their
health care consultants argued that HealthChoice brought in new resources from the
private sector in the form of the premium shares paid by employers and employees.

The selling Of HealthChoice to the UHCP Board was easier than casual
observation may anticipate, because the Wayne County Executive controlled the
appointments of all members of the Board. Thus the approval of the UHCP Board was
essentially a fait accompli. However, the selling of HealthChoice to the two health

systems was more arduous. Although the causal argument for increased revenue was

195

sound, the hospitals were concerned about the risks they would face should a person
covered by HealthChoice experience a serious and costly health event.56

Wayne County Executive officials used two ‘carrot’ arguments and a ‘stick’ threat
to Obtain agreement from the two health systems to sign HealthChoice contracts. The first
carrot argument was that the population targeted by HealthChoice are the working
uninsured. Studies of the working uninsured have shown that this population tends to be
healthier than the rest of the population (privately and publicly insured), primarily due to
their youth. Also, the Wayne County Executive Offered the health systems a lead role in
designing HealthChoice’s benefit plan along with the health care consultants.

The ‘stick’ threat used by county officials was their reminder to these two health
systems that the Wayne County Executive held contracts with both health systems for the
provision of health care coverage to Wayne County employees and inmates of Wayne
County prison facilities. It was suggested to these two health systems that these contracts
could go to other Detroit area health systems should these two health systems not agree to
sign contracts for providing coverage under HealthChoice. Not surprisingly, the two
health systems agreed to work with the health care and legal consultants hired by the
Wayne County Executive to develop a benefit plan along with the health care consultants.
The health systems then signed contracts for the provision of services for persons enrolled

in HealthChoice. It is not clear whether the ‘stick’ threat was more important than the

‘carrot’ causal arguments in convincing the health systems to cooperate. However, the use

 

56

HealthChoice, unlike Muskegon’s Access Health, does not rely on
members’ access to Medicaid Spend-down eligibility to cover this risk.

196

Of the ‘stick’ threat would suggest that Wayne County Executive officials felt that the
health systems would not have agreed had it not been made.

One last hurdle for the development of HealthChoice remained, obtaining
agreement from the Michigan Insurance Bureau (MIB) to set up a quasi-govemmental
health care coverage product that would be exempt from the state’s insurance regulations.
Meeting state regulations for reporting and the provision Of funds for a risk pool were
beyond the resource capabilities of HealthChoice as proposed, and thus required a
decision by MIB to let HealthChoice avoid these requirements. At this point, lobbying by
the Director of the Michigan Department of Management and Budget (MDMB) and the
Director Of the then Michigan Department of Public Health (MDPH) appear to have been
key. The Directors of both MDMB and MDPH had previously been employed by the
Wayne County Executive, and were long time colleagues of the members of the
entrepreneurial team from the county. Both of these directors, upon requests from their
colleagues in Wayne County, approached the Director of MIB and lobbied him to
interpret state regulations in a manner that would favor HealthChoice. This resulted in a
decision by the Director of MIB to refer to HealthChoice as health care ‘coverage’ and
not ‘insurance’, which exempted HealthChoice from MIB regulations on reporting and
risk pools.

Using the data derived from the interviews of Wayne County respondents we may
now assess the scores registered on the proposed variables. A few caveats should be
provided prior to reporting scores on these variables. The few interviews conducted and

their brevity reduced the Opportunity for the appearance of variables, thus fewer variables

197

made appearances than would have occurred had more and longer interviews been
possible.

On the variable of the state acting as a resource linking change agent, the score
was 4 (100%, high validity) from a total of 4 appearances of the variable. The state’s
Medicaid agency was an obvious supporter of PlusCare, whose funds ultimately made the
development of HealthChoice possible. However, the state’s role in the deveIOpment of
HealthChoice was largely passive, other than the enabling Of the funding mechanism for
HealthChoice and the permission Obtained from the Michigan Insurance Bureau tO set up
a health care coverage product that would be exempt from state insurance regulations.

On the variable of social connections of the entrepreneur, the score was 11 (75%,
high validity) from a total of 11 appearances of the variable. The senior Wayne County
executive and his deputy were long time fixtures of Wayne County’s political landscape.
They had worked with the health care systems throughout their careers, had worked very
hard throughout the years to ensure that funding for indigent care services would be
available to the health systems, and as noted, held contracts with these health systems to
provide services to both Wayne County Executive employees and to inmates Of Wayne
County prisons. They also, as noted above, had close contacts with members of both the
executive and legislative branches of state government. As it turns out, on the key issue of
avoiding MIB regulation, these state-level connections proved critical. It should also be
noted that the health care consultants hired to help design the benefit plan also had close
connections within state government, as these persons had all been previous employees of

the Medicaid agency and MDMB.

198

On the variable Of the use of causal arguments, the score was 11 (75%, high
validity) from a total of 11 appearances Of the variable. All of the members of the
collective entrepreneurship team appear to have developed and used the same set of
arguments to the health systems and their colleagues at the state level who lobbied on
behalf of HealthChoice to MIB. It should be noted that the ‘stick’ threat was not the kind
of causal argument one would expect to hear based on the literature review of
entrepreneurial behavior. However, it was a ‘cause and effect’ argument insofar as it was
pointed out to the health systems that the refusal to sign contracts to provide services to
HealthChoice enrollees could have certain negative financial outcomes for the health
systems. One might rather label the ‘stick’ threat plain coercion on the part Of the Wayne
County Executive.

On the variable Of coalition building, the score was 4 (100%, high validity) from a
total of 12 appearances of the variable. This score, although positive, contains negative
appearances of the variable related to the use of the ‘stick’ threat on the health systems.
We should perhaps expect this mixture of scoring from a coalition built on a combination
of long standing connections, causal argumentation and coercion.

On the variable of collective entrepreneurship, the score was 4 (50%, high
validity), from a total Of 4 appearances of the variable. This low score is at first a little
puzzling, as of the four counties studied, this collective team of entrepreneurs seems more
cohesive than any other formed. However, this seems due to the fact that the three
members of the team from the Wayne County Executive took their longstanding

relationships and ways of working together for granted, and thus felt little need to refer to

199

it. Regardless of the low score, it is clear that a team was formed of persons whose skills
were essential to the development and launch of HealthChoice. A question that can be
asked but can not be answered here is, once the health systems agreed to participate,
should they be considered members of this team, given that they participated in the
development of HealthChoice’s benefit plan? If they are included as members Of the team
at that point, then our ideas concerning the formation of collective teams of entrepreneurs
based in part on coercion should be considered.

The only other variables that manifested valid scores are related to the existing
political and market contexts within the county. For the variable on the existing political
context within Wayne County, the score was 4 (75%, high validity), from a total of 4
appearances of the variable. These are all related to the fact that the innovation originated
from the Wayne County Executive, and thus the strong support the innovation received
from that quarter. On the variable of local market context, the score was 7 ( 100%, high
validity), from a total of 7 appearances of the variable. These appearances are all related
to references to the long standing problems faced by health systems in Wayne County
(especially within Detroit) of providing uncompensated care to the uninsured when they
presented in hospital emergency rooms. The automatic provision Of care in hospital
emergency rooms to persons unable to pay was noted in Chapter Two as a fixture of the
American health care system, and a cost of doing business that is unavoidable by
hospitals. The high levels of such care provided by health systems within Wayne County
made them acutely aware Of its costs, although this awareness by itself was apparently not

quite enough to convince them of the need for a program like HealthChoice, at least

200

without the coercion that accompanied the causal arguments of the Wayne County

Executive.

Summary of Wayne County Case Study

The innovation under study in the Wayne County case is very different from those
studied in the three CCHMS case studies. The innovation had deep roots in the struggle to
provide reimbursement for services provided to Wayne County indigents over more than
a decade. Officials of the Wayne County Executive, and its partnering health systems had
for years developed and implemented solutions to their local problems within a shifting
political and market environment that was in many ways beyond their control. Thus soon
after officials from the Wayne County Executive became convinced that they had a
surplus of funds left over from PlusCare, they immediately sought an opportunity to use
these funds in an innovative way.

Although the score for acting as a market entrepreneur was not deemed valid due
to the low number of total interviews, the variable appeared twice in reference to the way
in which Wayne County Executive officials quietly formed a collective entrepreneurship
team, developed a general model for an innovation, and developed their causal arguments
prior to approaching the two health systems targeted for the signing of contracts.

The collective team of entrepreneurs used both carrots and a stick to Obtain
collaboration from the two health systems, and upon securing their collaboration, used
their strong connections at the state level to secure the critical decision from MIB to

exempt HealthChoice from regulation. Then the team spent a mere nine months

201

developing and finalizing the details of the benefits plan prior to enrolling eligible
employers and their employees into the plan.

Although a case study based on only four interviews, three of which were
foreshortened due to respondents’ time constraints, should be expected to result in fewer
appearances of variables, the absence of valid scores on some variables should be
assessed. The variables Of defining the problem, issue-specific knowledge, and efforts to
either alter the costs of the innovation to partners, or alter partners’ perceptions of cost
never appeared in the interviews.

The absence of valid scores on these variables would appear to be related to the
long struggle over covering some of the costs of the uninsured, primarily the indigent,
that county Officials and inner city Officials had engaged in for over a decade prior to the
appearance of the opportunity to develop the innovation that became known as
HealthChoice. For example, on the variable Of problem definition, the reason this variable
failed to appear may be that unlike in the other three cases, the problem was already
known and a consensus on what the problem was had already been arrived at by the
necessary set of actors. They had been struggling with issues of reimbursement for
services for the uninsured for over a decade, and so referring to a process of defining the
scope of the problem faced were intimately bound up in this struggle.

On the variable of issue-specific knowledge, the reason it did not appear is likely
also related to the same issue of the struggle to maintain reimbursements. These actors

were already so aware of the issues that they never referred to the need to Obtain issue-

202

specific knowledge, since they already possessed it at the point in time where develOping
the HealthChoice innovation occurred.

The absence of valid scores on the issue of perceptions of cost, altering
perceptions of cost, or altering the actual costs of the innovation would also seem to be
partially explained by this history of the key actors working in this issue area. It appears
that all participants to the development and launch of the innovation were intimately
aware of the costs of the innovation, and only the one health system respondent
commented on its cost to providers. It was considered inappropriate to code the ‘stick’
threat as an altering of either perceptions of the cost of the innovation, or its actual cost to
the health system partners, as this cost, though apparently perceived as real, was not
actually a component of the innovation that was developed and launched.

What this case study highlights are the real differences between counties with
large pOpulations and highly articulated county government and much smaller counties.
It also demonstrates how one well-connected and experienced politician at the local level
may be able to Operate exceedingly well as an entrepreneur The primus inter pares
member of the collective team of entrepreneurs was a long standing fixture of county
government and Wayne County politics who was intimately familiar with health care
issues concerning the uninsured, and their financial impact on his county and its health
systems. This person used his accumulated experience and the experience of the other
members of his team of entrepreneurs to persuade and coerce cooperation from two
health systems for the development and launch of an innovation targeting the working

uninsured. Elected county Officials in smaller counties have fewer similar Opportunities

203

for such experiences, and certainly have no levers over health systems similar to the

levers available to the Wayne County Executive.

Assessment of the Propositions across all Cases

With the separate case studies complete, it is now time to assess how well the
propositions were supported by the data across all Of the cases. This analysis will be
conducted across cases, meaning that interview data will be aggregated across the cases to
assess each proposition. This will result in changes in validity of many variables, as the
denominator for validity is the sum of persons interviewed across cases. The permutations

among cases will not be ignored, but accounted for in the assessment of each proposition.

204

 

Change Agents

Proposition #1: A change agent is a necessary (but not sufficient) precursor to the
development and launch of the county-level program or product in counties where
there exists no “slack” in resources, or where the local health policy subsystem is
weak enough to be opened with outside assistance. In this situation, a foundation
and/or state government may both become a factor in the development and launch
of the program or product.

This proposition will be assessed across cases in two parts, first for the role of
WKKF as a change agent, across the three CCHMS counties, second of the state as a
change agent, across all four county—level case studies.

Units of Analysis: WKKF as a Change Agent, County’7

Of the three CCHMS case studies, one developed and launched a mixed public
and private innovation that extended health care coverage to a portion of its county’s
uninsured. A second county saw private sector interests develop and launch a purely
private sector innovation that was sanctioned as the CCHIP effort to extend health care
coverage to the uninsured, although the mechanism for so doing was never articulated,
and its start date left open. The third county never developed and launched an innovation.

The question becomes one of establishing a counterfactual in the two cases where
an innovation was developed and launched. Would these innovations have been

developed and launched in the absence of the change agent? The answer for Muskegon is

 

’7 The four interviews from the Wayne County case study are not included in

the calculation of validity for the aggregate scores on change agent
variables related to WKKF.

205

a resounding no. For Calhoun it is a tentative maybe, with the caveat that the future goal
of using savings to extend health care coverage to the uninsured would probably have
never been included in the SMHPA charter without the presence Of the WKKF-funded
CCHMs Initiative.

In the Muskegon case it seems clear that the activities of the project director
would never have taken place in the absence of the WKKF—funded intervention. The
project director spent over three years, and extensive financial resources provided by
WKKF, to go about the process of developing and launching the Access Health
innovation. She hired staff and several consultants, commissioned several telephone
surveys, and had the freedom to engage in a variety of other entrepreneurial activities to
develop an innovation to extend health care coverage community-wide. No public or
private organization in Muskegon, and no collection of like-minded organizations, had
the desire and financial wherewithal to fund such activities. The resistance to change
within this issue area as evidenced by the infighting within the MCHP’s C-P-P Governing
Board suggest that the innovation would never have been developed and launched
without the presence of WKKF acting as an external change agent.

In the case of Calhoun County, the process of problem definition and information
gathering led to the creation of an Opportunity for private sector interests to realize a
reduction in their own health care costs. Thus these interests seized the process of
developing and launching an innovation from the WKKF-funded project director and
pursued innovation largely on their own, but using some CCHIP funds for the process.

Because these private sector interests were also representatives on CCHIP’s C-P-P

206

Governing Board, they were able to Obtain release Of these resources through the Official
sanctioning by the Governing Board of this innovation as the innovation that would
extend health care coverage to the uninsured. It is entirely possible that the staff of
SMHPA in Kalamazoo County would have eventually come into contact with these
private sector interests in Calhoun County, and that the Calhoun—based employers would
have joined SMHPA in the absence of the WKKF-funded CCHIP. What would likely not
have happened would have been the inclusion, however fuzzy, of a charitable purpose
clause in SMHPA’s charter stating that once savings from pooled purchasing were
realized, these savings would somehow be translated into an effort to extend health care
coverage to the working uninsured.

Despite the fact that WKKF can point to one clear success and one partial success
does not mean that the foundation fully assumed the role of an external change agent.
Rather, these successes occurred despite the unwillingness of the foundation to assume
the role, from the earliest point of internal planning for the initiative, to the selection of
counties for funding, to the support of these counties as they went through their separate
CCIrIMs-firnded processes. The fact that the foundation permitted the takeover of
innovation development and launch in Calhoun County, and their reluctance to intervene
in a slow and ultimately flawed process for the selection of project leadership in St. Clair
County are both manifestations of the foundation’s fear of the conflict that would seem to
be inherent in the funding of change initiatives, and the role of a change agent.

The imperfect filling of the role of change agent comes through in the two

variables with valid scores across the CCHMS cases, that of acting as a catalyst and of

207

acting as a solution giver. Both scores are negative (catalyst = -3, 33%, medium validity
from a total of 9 appearances of the variable; solution giver = -2, 35%, medium validity,
from a total of 10 appearances of the variable), but not strongly negative. Thus we can
interpret these two variables’ scores as the foundation insufficiently filling these two
aspects of the role.

Regardless Of the low negative scores on these two variables, the proposition
appears valid: these under-resourced counties would not have addressed the issue of
extending health care coverage to the uninsured without the presence of WKKF acting,
however imperfectly, as an external change agent. The innovation in Muskegon and the
not very innovative innovation in Calhoun would not have been developed and launched
as they were in the absence of WKKF funding and the stated goals Of the CCHMS
initiative. Thus WKKF, acting however imperfectly as a change agent, was a necessary, if
not sufficient, condition for the development and launch of these two innovations.

In the defense of WKKF staff, this was the first time they had considered taking
on the role Of an external change agent, and they did so with much trepidation and real
confusion as to how to fill a role that they did not completely understand. Early on they
understood that this effort was a test Of their ability to assume this role. Only time will tell

if WKKF staff have learned from their missteps on the CCHMS initiative.

Units of Analysis: State Agency as a Change Agent, County
In the four county-level case studies, two Of the four counties developed and

launched mixed public and private innovations that would extend health care coverage to

208

the working uninsured, Muskegon and Wayne Counties. In a third county, Calhoun, an
innovation based solely in the private sector was developed. and launched that was
intended to ultimately use savings from pooled private sector purchasing to somehow
extend health care coverage to the uninsured. In the fourth case, no innovation was ever
developed and launched.

In the first two cases, the state agency played a relatively passive change agent
role as a resource linker but not as a provider of networking Opportunities. The cross case
score for the state agency on the variable of acting as a resource linker was 8 (39%,
medium validity) from a total of 19 appearances of the variable, from 11 respondents.
The somewhat low score (considering this is an aggregate across all cases) and medium
validity are directly related to the fact that the state performed this role for only two of the
teams of entrepreneurs from the counties, Muskegon and Wayne. In both cases, the teams
Of entrepreneurs had strong pre-existing social and business connections to the state-level
actors necessary to establishing the resource linking role. These pre-existing connections
appear to have been necessary because the key state agency Officials, especially within the
Medicaid agency, did not seem to have the time to proactively step into the role of
provider of networking Opportunities to local level officials engaged in innovation. This
variable had too few appearances from too few respondents to register above a low
validity score.

The persons pursuing the goals of the CCHMS initiative in St. Clair County were
unable to establish contact with state agency Officials. The market entrepreneurs who

attempted to Obtain cooperation from state agency officials to include Medicaid

209

purchasing in Calhoun County also failed in their effort, apparently also because they had
no close ties to the necessary state-level officials.

This problem was anticipated by Roberts and King (1996) in their study of policy
entrepreneurs in the issue area of education in Minnesota, where they argued that, “... the
constraints on elected and appointed office make it difficult for an incumbent to be a
single-minded change agent” (1996, 178). The constraints referred to by Roberts and
King include both the demands Of these officials’ day-to-day jobs, and the diverse
constituencies to whom they are answerable.

The results from this study do not preclude state-level officials from acting as
providers of networking Opportunities, and thus initiating contact for the purpose Of
innovation with local level Officials. The results do suggest that Roberts and King’s
argument has merit, and that these Officials may be too busy with current business, as
appears to have been the case for the three CCHMS counties, or too concerned about
other constituencies to assume this role.

Regardless, it is clear that in the two cases where an innovation was developed
and launched, the passive resource linker role of the state was essential in the
development and launch of these innovations. This holds true despite the fact that one of
the counties had substantial resources of its own to dedicate to the innovation. Thus we
can state that in these two cases, this role Of the state was a necessary, although not
sufficient, condition for both the low and high resource counties in their development and

launch of these innovations.

210

Another unanticipated role of the state was in the favorable interpretation of
MIB’s rules governing the regulation of health care insurance, and what constitutes
‘insurance’ and is thus subject to regulation. This favorable interpretation of existing
administrative regulations, and the favorable exercise of regulatory authority by another
state agency constitutes at a minimum an unanticipated facilitating role for state agencies
in the support of change efforts at the local level. This favorable exercise of regulatory
authority aided the development and launch of HealthChoice in Wayne County, and the

precedent thus set also aided the development and launch of Access Health in Muskegon.

Policy Entrepreneur
Proposition #2: A local policy entrepreneur team is a necessary component to the
development and implementation of the policy/market innovation.
Unit of Analysis: Policy Entrepreneur, County

It was proposed that in order for a local-level group to develop and launch a health
care coverage program or product that will extend health care coverage to the working
uninsured within a community, a policy entrepreneur must exist. As stated in the
discussion around Proposition #1, the selection of the project director is the focal point
for the selection Of the policy entrepreneur. The supply of policy entrepreneurs was
expected to be steady across cases, constrained only by the income and education levels
extant within the county (as discussed in Chapter Three). NO interview evidence appeared
to support this idea that the supply of policy entrepreneurs was constrained in this

manner. Thus no basis exists for accepting or rejecting this concept.

211

What was not anticipated as a part of this research study was that the supply of
policy entrepreneurs might be artificially constrained by the selection process for the
position Of project director in any of the three CCHMS counties. However, this appears to
have occurred in St. Clair. Future research of extemally-initiated change projects should
consider the real possibility that selection processes may hinder the search for a policy
entrepreneur to fill such a role.

Perhaps as important as the artificial constraints on the selection process is the
fact that the conceptualization of the role of the project director by WKKF staff did not
include the idea that the persons filling these roles would need to possess any of the
qualities Of an entrepreneur. This underspecification Of the project director role had
negative consequences in St. Clair County, where foundation staff ultimately had to
request the firing of one project director and the hiring of a replacement.

It was proposed that the policy entrepreneur (in the three CCHMS cases this was
expected to be the project director) should already possess extensive knowledge of the
nature of the health care system within his or her community as well as an understanding
of the growing problem of the uninsured. Also, this person should have an established
network of contacts that extend both into local policy networks and into the local market
for health care services, or have the ability to develop these contacts. The policy
entrepreneur should have strong social and political skills, and have the ability to think
creatively about the problems Of covering the working uninsured within the broad

structures of funding health care coverage that exist at the federal, state and local level.

212

On the variable Of issue specific knowledge, the cross case score was 30 (50%,
high validity) from a total of 45 appearances of this variable. That this variable registered
positive after the negative scores from the Calhoun and St. Clair cases is indicative of the
extensive knowledge possessed and referenced concerning the members Of the collective
teams of entrepreneurs in Muskegon and Wayne. This variable, as mentioned in these two
cases, appeared critical to the success Of the development and launch of the innovations
in Muskegon and Wayne.

The variable Of social connections had a cross case score of 32, (54%, high
validity) from a total of 49 appearances of the variable. This relatively strong positive
score is driven almost entirely by scores from the Muskegon and Wayne cases. This is no
surprise, as these were the cases where innovative programs were developed and
launched that succeeded in extending health care coverage to the working uninsured. As
mentioned in the case studies, these connections were both within the counties and with
key state-level Officials. Only in Wayne did the members Of the policy entrepreneurship
team have pre-existing contacts with the two health systems necessary as partners in the
innovation.

It was proposed that the entrepreneur should understand how to frame an issue,
which would be revealed by his/her ability to craft causal arguments that support his/her
proposed innovation, and can mold these causal arguments to suit the multiple
organizations/individuals that must be convinced to participate in the partnership that will
develop and launch the innovation. Lastly, the entrepreneur must understand how to place

an issue on the local political agenda, and how to keep an issue high on local policy

213

agendas long enough to implement the innovation. The assessment of the use of causal
arguments is presented in the assessment Of Proposition #3, below.

Because of the complexities of the issue area, it was proposed that the lead policy
entrepreneur would build a collective entrepreneurship team to maximize the availability
of the skill sets and characteristics mentioned above. The more diverse and
complementary the skills of this team, the more likely the policy entrepreneur would be to
succeed.

The aggregate score of collective entrepreneurship across all cases was 37 (50%,
high validity), from a total of 61 appearances of the variable. This score highlights the
fact that in three of the four cases studied, a collective team of entrepreneurs was formed,
although the characteristics of the team of market entrepreneurs in Calhoun County are
difficult to assess. In both Wayne and Muskegon cases, teams were formed from among
existing staff of the organization undertaking innovation, along with consultants, and in
the case of Muskegon, two other persons volunteering their time as a part of the CCHMS
initiative’s activities. The teams appear to have been formed for the reasons proposed, the
issue area’s complexity was such that no single individual had the set of all necessary
knowledge, skills and connections to develop and launch an innovation alone. Thus, in
Muskegon and Wayne, both of the members of the entrepreneurial team who were
acknowledged as the leaders of those teams, the project director and the County
Commissioner respectively, hired or contracted with individuals who could provide the

knowledge, skills, and connections that they were missing. In St. Clair, where no

214

innovation was developed or launched, no team was ever really formed that possessed all
of the proposed set of qualities deemed necessary for success.

Because a coalition must be developed in order for the innovation to succeed, it
was proposed that the entrepreneur must also possess coalition-building skills. These
skills would include the ability to both reframe problems so that actors will see benefits in
participation where before they saw only costs, as well as the ability to find ways to apply
pressure to some necessary actors in order to achieve cooperation. It is important to note
here with respect to the CCHMS project sites that the Governing Boards should not be
mistaken for the coalitions that needed to be built. Rather, the entrepreneur needed to
develop a coalition Of like-minded organizations/individuals, many or all of whom may
also have been on the Governing Board (for the CCHMS cases), in order to succeed in the
development of the innovation.

On the variable of coalition building, the aggregate score across all cases was -30
(93%, high validity) from a total of 126 appearances Of the variable. This relatively high
negative score is heavily influenced by the very negative score achieved on this variable
in the Calhoun case study. Within two cases, the Muskegon and Wayne cases exhibited
positive scores. As mentioned in the Calhoun case study, the negative score was hard to
interpret because so much activity surrounding the development and launch of a strictly
private sector innovation appears to have been hidden during the course of the interviews.
If we rely only on the successful examples of the development and launch of mixed
public-private innovations in Muskegon and Wayne, the score is 32 (92%, high validity)

from a total of 52 appearances of the variable.

215

Finally, it was proposed that the policy entrepreneur (or some members Of the
collective entrepreneurship team) would also be acting as market entrepreneurs, and
would face the set of dual constraints described in the literature review. The entrepreneur
will need to navigate both sets of constraints in order to form a winning coalition as well
as develop and launch an innovative program or product.

This variable did not achieve medium validity across cases, despite the fact that
the Calhoun case was a purely private-sector innovation. This appears to be related to the
guarding of information related to market entrepreneur behavior as proposed. However,
what was not anticipated was that this behavior extended to the guarding Of this
information during the key informant interview process. This leads to implications for
future research on market entrepreneurial behavior, that the gathering of information from
these kinds of key informants of their activities should be anticipated to be difficult.

Another possibility is that, given the nature of the mixed public-private
innovations, market entrepreneurial behavior was not a major component in the
successful development and launch Of these innovations. It is difficult to assess whether
or not this may be true, since clues to the hiding of information exist in the Calhoun
transcripts. However, the Muskegon and Wayne transcripts exhibit few clues in this
direction, and in neither case was a medium or high validity score manifested. Thus the
possibility that in developing a mixed public-private innovation, the need to sometimes
act like a market entrepreneur is not critical to the successful development and launch of
such an innovation. Rather, the normal set of activities described in the literature on

policy entrepreneurs suffices to develop and launch mixed public-private innovations.

216

Proposition #3: Local policy entrepreneurs either alter the actual structure of costs
and benefits of the members of the partnership vis-a-vis participation in the
development and launch of a new program or product, or they alter the members’
perceptions of the costs and benefits of the new program or product.

Unit of Analysis: Policy Entrepreneur

It was proposed that the policy entrepreneurs would either alter the actual costs
and benefits for the potential partners necessary to the development or launch of the
program or product designed to extend health care coverage to the working uninsured or
they would use causal arguments to alter the potential partners’ perceptions of the costs
and benefits Of participation. They could also do both. It was further proposed that the
policy entrepreneurs could alter the actual costs and benefits by accessing external
sources Of funding to support the launch of the innovation, and they could be supported
by a change agent so that the change agent would foot a portion of the costs Of the
development Of the program or product. In this manner the policy entrepreneur would
change the actual costs and benefits to potential partners.

In terms of the altering of the actual costs Of the innovation to potential partners,
the aggregate score across cases did not reach medium validity. This is because the only
example of an entrepreneur altering the actual costs Of the proposed innovation to
potential partners occurred in Muskegon, where the entrepreneurial team proved to its
potential partners that a mix of state and federal funds could be brought in to cover the
costs of the innovation. What is interesting is that in the Wayne case study, this variable

did not appear. However, this is probably due to the fact that the surplus of funds from the

217

PlusCare program was a known quantity, and that perhaps it was the coercion that
ultimately brought the health system partners to cooperation.

It was further proposed that the policy entrepreneurs would make causal
arguments to change potential partners’ perceptions of costs and benefits. These causal
arguments were expected to center around the unknowns surrounding the expected
success of the program or product, especially the financial risks that the potential partners
would perceive in committing to the partnership and the launching of the program or
product.

The proposal that the potential risks to project partners signing onto the
innovations developed and launched were very real, as was confirmed by the score for
initial perceptions of the cost of any innovation to extend health care coverage to the
uninsured reveals. The aggregate score across cases was -38 (53%, high validity), from a
total of 45 appearances of the variable. These negative scores were unifome those of the
individuals and organizations that supply health care services. However, in Muskegon,
Wayne, and even Calhoun, causal arguments were used to alter the perceptions of cost to
these suppliers, as revealed by the aggregate score across cases Of 46 (63%, high validity),
from a total of 61 appearances of the variable. The negative appearances of the variable
are almost all from Calhoun, which only serves to highlight the extensive use of
argumentation in Muskegon and Wayne to alter perceptions of the cost of the actual
innovation to potential partners.

The fact that the entrepreneurial teams in Muskegon, Wayne, and tO a much lesser

extent, Calhoun, behaved as proposed in terms of altering the perceptions of cost to

218

 

 

 

convince potential partners to collaborate should not also miss the need to account for the
strong arm tactics to Obtain collaboration that were also applied in both Muskegon and
Wayne. In Muskegon, one health system signed the necessary contracts only after
realizing that the MCHP project director was willing to launch the innovation with or
without them, and the balking hospital’s CEO’s reported perception that the public
relations damage that would result from not being a part of Access Health at its launch
was unacceptably high. In Wayne the coercion was more bold, with the Wayne County
Commissioner gently reminding the two potential health system partners that their large

contracts through the Wayne County Executive could be given to others.

Exogenous Factors
PrOposition #4: The local political and market contexts constrain the ability of the
policy entrepreneur to build a winning coalition that can develop and launch an
innovative program or product.
Unit Of Analysis: County

As discussed in the literature review, both the political and market contexts were
proposed as placing constraints on the ability Of the entrepreneur to build a winning
coalition that can then develop and launch an innovative program or product. For the
political context these limits are based on institutional structures for policy making,
policy settings, and the behaviors of the groups and individuals around them. This milieu
was proposed to shape the possibilities of action by a policy entrepreneur and allows us to

develop expectations about the possibility of generating support for policy innovation.

219

For the market context these limits were proposed to stem from two contextual factors,
the pace of change (innovation) within the market, and the concentration of health care
firms within the market.

Based on the discussion presented in Chapter Three, it was expected that the bulk
of these constraints would be revealed in the coalition-building stage, prior to any effort
to develop and launch an innovative program or product. These constraints would be
represented by the availability of alternative venues for presenting the issue, as Opposed
to the closed venues of the local (and to a lesser extent the state) health care policy
subsystem. Although the literature on policy entrepreneur behavior suggests that these
two processes occur at the same time, we would expect the highly technical aspects of
program or product development to not occur until after a winning coalition had been
formed to support and participate in this part of the process. Also, the concept of guarded
dissemination suggested that policy entrepreneurs would try to make the two processes
more sequential than simultaneous, to protect information during the coalition-building
stage that is necessary and requires protection for later use in the program or product
development stage.

The variable on the existing local political context did not reach medium validity,
thus leading to the conclusion that the political context was not important in constraining
the activities of policy entrepreneurs in the development and launch of their innovations.
In retrospect, this part of the proposition seems ill-conceived, as it was also proposed (and
supporting evidence found) for the presence Of local-level health policy subsystems in

each county. Thus we should expect that, since policy subsystems develop in technically

220

complex areas where much decision making on policy is left up to industry experts, then
local politics would matter less than the views of the key actors in the policy subsystem.
Moreover, much setting of public policy, especially true in the issue area Of health care,
occurs at the state level. Thus it should have been anticipated that for this research study,
where local-level health policy subsystems were expected to exist, and in an issue area
where public resources for health care are controlled almost entirely at federal and state
levels of government, that the local political context would not have much effect on
entrepreneurial activity.

However, in support Of the original proposition, in other issue areas this may very
well not be true. Two of the scholarly works on policy entrepreneurs used in the literature
review, Mintrom (2000) and Roberts and King (1996) studied the issue area of public
education, where much of the funding is controlled at the municipal level. Thus the mis-
specification of the political context portion of this proposition for the issue area Of health
care shOuld not be expected to hold true for other issue areas.

The poor showing of the existing local political context variable was not true of
one Of the existing local market context variable. The variable for the existing market
context (at the beginning of entrepreneurial activity) had a score of —13 (71%, high
validity), from a total of 71 appearances of the variable. The market context within which
each of these attempts at innovation were begun was a topic of much discussion by nearly
all respondents. The aggregate score across cases represents a very bleak negative
showing in Calhoun and St. Clair, a positive, but mixed score from Muskegon, and a

positive score from Wayne. The Wayne score is probably seriously undervalued and with

221

the result that the score is not as representative across all contexts as it should have been.
It is likely that had the Wayne interviews been longer, and more key informants been
available for interview, the score would have been closer to zero. A zero score would
have been a wash, which is essentially what happened, two innovations that required the
active participation of actors within the local health policy subsystem were developed and
launched, and in the other two cases this did not occur. The discussion Of how context
affected the process of coalition building proposed in this proposition will be discussed in

the assessment of Proposition #5, below.

Proposition#5: The local political and market contexts must be undergoing change
in order to increase the probability of successfully building a partnership that will
develop and launch some program or product to extend health care coverage to the
uninsured.
Unit of Analysis: County

It was proposed that the political venues Of the health policy issue area are largely
closed and dominated by the suppliers of health care services. These actors have a great
deal of political power that has led to significant legal power over the supply-side of the
market. However, it was proposed that these political venues, despite remaining largely
closed, have been cracking Open. Especially noteworthy is that physicians have been
losing their grip on the authority to make diagnoses and prescribe medications to nurse
practitioners and physician assistants through changes in state law as well as the spread of

managed care arrangements. Physicians have also begun to feel financial pressures related

222

to changes in reimbursement and reimbursement structures for both Medicaid and
Medicare, and have reacted negatively to the constraints of private sector managed care
arrangements.

As was the case with Proposition #4, the local political context variable did not
Obtain medium validity. The reasons for this are assumed to be the same as presented in
the assessment of Proposition #4, that the importance of the local political context is
relatively unimportant within an issue area where public funds for health care services are
mostly under the control Of the federal government or the states.

The assessment of this proposition as it pertains to the local market context will
begin with the idea that this proposition was mis-specified, and then present the evidence
to support how it was mis-specified. Instead of proposing that the local markets for health
care needed to be undergoing change in order to Open cracks in the local health policy
subsystem, the evidence from this research suggests the proposition should rather have
stated that the local market should have already undergone change in order to open cracks
in the local health policy subsystem. Although it was proposed that a strong local-level
health policy subsystem would exist at the county level, it was also proposed that current
changes within a the market for health care services would be enough to open up cracks
that would permit the entry of a policy entrepreneur who wants to pursue innovation. The
original formulation of the proposition would seem to have underestimated the strength of
these local-level health policy subsystems, and their weakening a result of current

changes.

223

It would appear, from the evidence that will be presented below, that in fact only
in Muskegon and Wayne, where the use of managed care arrangements was an
established feature of the local health care market, were some members of the local health
policy subsystem willing to accept innovative ideas from actors outside of the system, a
strong system had developed cracks exploitable by policy entrepreneurs. If the
proposition had assumed that change would have already occurred, it would have meant
that those actors would have already spent a great deal of energy in searching for avenues
for recouping the loss in reimbursements brought on by managed care arrangements by
attempting to increase their share of the services provided within their local health care
markets. It would appear that these cracks should only have been expected to appear once
all actors that were weak had disappeared, and the remaining strong actors had played
enough versions of their own iterated Prisoner’s Dilemma game trying to eliminate or
absorb each other to arrive at the conclusion that perhaps cooperation and innovation
would offer the only remaining viable avenue for increasing revenue from the sole
remaining non-paying population that all would have to provide services to in their
emergency rooms. As well, the hospital/health system actors would have already
discovered through another process of iteration, that they had explored all Opportunities
for the shifting of the costs of providing care to the uninsured, and were now stuck with
any remaining costs for the provision of these services.

This altered view of this proposition is best supported by the evidence from
Wayne County. The absorption/elimination of unprofitable hospitals by more profitable

ones had become a fixture of the landscape by early 1994 when the innovation was

224

proposed to the two potential partner health systems. Furthermore, the private employer
purchasers in this market had long before squeezed out shifted costs from their premium
payments, especially the three automakers. Lastly, Medicaid managed care arrangements
in Michigan had debuted in Wayne County, further limiting another large source Of
reimbursements. Thus hospitals/health systems in Wayne County should have been
expected to view the HealthChoice innovation as perhaps their only remaining route for
capturing even low level reimbursements for the provision of services that currently were
bringing in no reimbursements.

The evidence from Muskegon supports this alteration of the proposition, although
the experience with managed care arrangements in Muskegon was not as deep as in
Wayne. It had only been one year since the absorption of one of Muskegon’s three
hospitals by another had occurred. Thus the iterative game Of competition-elimination-
absorption between the remaining two hospitals had not been going on for very long. One
of the remaining two hospitals had made overtures to the other to consider some form of
contractual collaboration on the sharing of services and resources, but had been rebuffed.
It was within this environment that MCHP was formed and the process of developing and
launching an innovation occurred. Because the team of entrepreneurs had no coercive
levers at hand as powerfirl as those held by the Wayne County Executive, they went
through an extensive information gathering process, development of a media strategy, the
development and use of a community venue, and the altering Of the costs and perceptions
Of the cost of the innovation to enough potential supporters that they were able to obtain

support and agreement from one of the two hospitals to sign the contracts necessary to

225

 

 

 

launch Access Health. Only after the first hospital signed the contracts, and it became
clear that Access Health would launch with or without the second hospital, did the second
hospital sign its Access Health service contracts.

This researcher was in the field collecting interview data at the point where one
hospital had agreed to sign contracts (but had not actually signed them) and the other
hospital had not. For a few weeks there appeared to be a form Of brinkmanship at work, in
the sense that the balking hospital had declared to the members of the entrepreneurial
team that they were hesitant to sign, in an apparent effort to scare the more cooperative
hospital into backing Off, for fear of bearing the potential financial risks Of Access Health
alone. Ultimately though, the cooperating hospital signed its contracts (after several
weeks of back room lobbying by the members of the entrepreneurial team), sending a
signal to the balking hospital that Access Health would launch without them.

In St. Clair County, the absence of managed care arrangements meant that the
effort to avoid what Muskegon and Wayne actors determined was a losing battle over
declining reimbursements had not even begun. Two Of the three hospitals were aware that
this environment was coming (managed care is now a feature of St. Clair’s health care
market), and had sought to forestall competition amongst themselves by merging before
the managed care storm hit. This attempted merger was disallowed by a ruling from the
Federal Trade Commission, and as soon as that decision was made, competition between
the two hospitals in Port Huron began in earnest. This competition began slightly before
the funding of the CCHMs Initiative in St. Clair County. The transcripts are replete with

descriptions of multiplying duplication of services across the two hospitals, and of the

226

unwillingness of either hospital to share ownership of efforts to extend health care
coverage to the uninsured within the county.

One example of this inability to share even the smallest effort at improving the
health of the community is illustrative of this intense competition. One hospital had a
bike safety program where it distributed free bicycle helmets to children in the city of Port
Huron. When the C-P-P Governing Board suggested that this program could be expanded
to include both hospitals, and receive funding from CCHMS Of St. Clair (which would
then receive and use the imprimatur of CCHMS of St. Clair), the offer was rebuffed,
because the hospital did not want to have any other organizational names associated with
a program that was perceived as promoting a strong public relations face for the hospital.
If agreement over bicycle helmets was unreachable in this competitive context, efforts at
obtaining agreement over an innovation to extend health care coverage to the uninsured
were doomed to failure.

The Calhoun County case provides no evidence to support or reject a revision to
the proposition, because SMHPA, once supported by potential employer partners within
Calhoun, would focus solely on securing lower health care coverage premiums from
health plans, thus leaving the other actors in the local health care system out of the
development and launch process.

Who were the active players from the local health policy subsystem across all of
the counties studied? Hospitals/health systems. The preceding discussion strongly
suggests that these players were the key partners in innovation or Opponents of change in

all four cases. These are the actors most affected by the costs of supplying health care

227

services to the uninsured because of the requirement that all persons must be seen in
emergency rooms who present there. These actors appear to be the ones that needed to
have been weakened enough by the local health care market to become willing to accept
innovative ideas from outside Of the health policy subsystem, indeed even to consider
cooperating with other hospitals within their local area with whom they were competing.

Where were the proposed physicians? In Muskegon physicians played an
important supportive role in pushing for acceptance of the innovation. References to the
weakened position of physicians within Muskegon County as a result of many years
experience with managed care reimbursement had created a context where some
physicians were eager to support an innovation where they had the Opportunity to shape
its benefits plan and reimbursement. This factor was reported as very important in the
Obtaining of physician support and participation in the development and launch of the
innovation.

In Calhoun some physicians lobbied against the acceptance of SMHPA as the
official innovation to extend health care coverage to the uninsured, to no avail. In
Calhoun, the weakness that was reported was what led some physicians tO lobby against
the acceptance of SMHPA as the accepted innovation sponsored by CCHIP. These
Opposing physicians failed to sway any votes in the single formal vote taken on
acceptance of SMHPA as CCHIP’s innovation.

Otherwise, physicians were absent. As described above, evidence of the
weakening power of physicians came out in the transcripts from both Calhoun and

Muskegon case studies, but was not mentioned in either the St. Clair or Wayne

228

 

 

interviews. In St. Clair, however, we see that all actors within the local health care system
were still operating in the reimbursement rich world of fee-for-service health care. Thus,
based on the discussion in Chapter Two, we would expect to have seen little or no
weakening of physician power in this county.

In Wayne County, no mention was ever made of physicians being involved,
however, both health systems targeted as potential partners had sufficient numbers of
physicians under contract to avoid the need to develop support for the innovation within
the physician community. In effect, these two health systems spoke both for themselves
and the physicians under contract to them.

This evidence on the role of physicians in the development and launch of
innovations runs mostly counter to what was proposed, with the important exception of
Muskegon. Thus, the evidence from this study suggests that physicians may only matter
in a context where the development of a partnering coalition is close to forming a critical
mass, but not if the formation Of such a coalition is really in doubt.

Private firms, which fund over half of the costs of health care coverage, were
almost entirely absent from the initial discussions over innovation in each county. The
inability to attract and maintain the attention of business leaders appears to be indicative
of these leaders’ ceding of decision making to the suppliers of health care services of any
search for innovation. Only in Calhoun County do we see business leaders eventually
becoming involved after a possible innovation is discovered that could enable these firms

tO realize lower health care costs, with a hoped for benefit for uninsured persons later on.

229

 

The absence of private employer purchases supports the notion that these actors are not
members of the health policy subsystem at the local level.

Another commonality across all four cases is the absence of participation in the
development and launch of innovations by health plans. In none of the cases were health
plans ever a part of the process. Speculating on their absence is difficult, as the transcripts
do not mention them at all. However, as mentioned in Chapter Two, the cachement areas
of health plans are much larger than just one county. It seems likely that no health plan
was ever interested in a search for innovation targeting the uninsured at the county level
because it would have had little financial impact on them, positive or negative, and thus
their interest was low. Additionally, Of the counties under study, only one, Wayne, had
any health plans whose headquarters were within the county. As was noted in the Wayne
case study, once the team Of policy entrepreneurs determined on their intervention, they
did not need to involve any health plans in order to develop and launch their innovation.
Muskegon’s intervention, based on the Wayne model, also did not require participation
by health plans. We may speculate, then, that because the lowest level at which health
plans operate, the multi-county level, is much broader than the county level, this makes
them relatively unimportant actors in a county-level health policy subsystem.

It was further proposed that the market context had to some extent undergone
significant change above the county level. The reduction of federal reimbursements for
Medicare claims, the state purchase of Medicaid services through managed care

organizations, the broad introduction of managed care arrangements in the private sector

230

Of the market, and the increase in purchaser demands for lower costs or at least lower
rates of increase in costs have all weakened these actors’ power within the market.

As mentioned above, only in Wayne and Muskegon counties were managed care
arrangements firmly established as the reimbursement mechanism for Medicaid. Managed
care arrangements for Medicaid were not introduced state-wide until near the end of the
CCHMS Initiative. It is probably not coincidence that these were the sole counties where
there was also evidence of a weakened health policy subsystem.

The preceding assessment of Proposition #5 permits a return to the unfinished
assessment of Proposition #4's idea concerning the effect of local market contexts on the
coalition-building process, and where in time the constraints of the market context are
felt. The proposition that the constraints of the context would be felt in the coalition
building stage appears to have been borne out by the preceding discussion. Where the
health care market had experienced negative, long term changes, the key actors had
already made every effort to find ways within their own means to address shortfalls in
reimbursements. What the preceding discussion suggests, then, is that where a policy
subsystem is strong, the members of the subsystem will not experience the proposed
weakening immediately after negative changes begin, but rather only after these actors
have attempted to ameliorate (without complete success) the negative effects Of change
over some period of time. Thus, the proposition was mis-specified in the sense that it
appears that it was not enough that change be occurring where policy subsystem actors

are strong, it is enough only when these actors have been forced to endure long periods of

231

negative change that they will then accept innovative ideas from outside the policy
subsystem.

The use of alternate venues for action occurred in Muskegon, where the lead
policy entrepreneur faced an uphill battle convincing the two key players, the two
hospitals, to agree to participate in the development and launch of an innovation. Both the
media venue and a community venue were used to help support this coalition-building
effort. It was even possible to develop a venue amongst physicians, through the strategic
use of a team of local physicians in the development of the technical aspects of the
benefits plan and reimbursement structure. Also, the ability to draw on the existing
personal connections of two members of the collective policy entrepreneur team to reach
key state agency decision makers played a role. Lastly, the recruitment of the Executive
Director of one of the hospital’s physician networks to become a member of the
entrepreneurial team provided a social connection that helped sway one of the hospital
CEOs to begin to seriously consider cooperation.

The venues used in Muskegon were not used in Wayne. The absence of this use
probably speaks to the relative bargaining strength of the lead policy entrepreneur as well
as the long established working relationships with the prospective partners of the lead
policy entrepreneur at the beginning of what was a very short coalition-building process.
Indeed, it is better to view the coalition-building process as largely complete by the time
the County Executive had the Opportunity to consider developing and launching the

innovation that became HealthChoice, as the key actors had been working together for

232

years to address shortfalls in reimbursements associated with providing care to the
indigent.

The proposed timing of coalition-building vis-a-vis innovation development in the
two cases where a mixed public-private innovation was developed and launched is a little
Off, but not by much. It was proposed that the supportive coalition (i.e., the necessary
partners) would need to be in place prior to the development Of the highly technical
aspects Of the innovation. This time ordering held true in Wayne, but only partially true in
Muskegon. The project director in Muskegon began development of the technical aspects
Of the innovation after only receiving a somewhat lukewarm agreement to support the
innovation from one of the two key hospital CEOs, and thinly disguised opposition from
the other. Rather, with lukewarm agreement from one partner in hand, the project director
took a large risk by essentially using the development of the technically complex aspects
Of the innovation by a team of local physicians to aid her in the completion of her

coalition-building process.

Generalizibility
Proposition #6: Although the cases to be proposed all exist in Michigan, the lessons
learned from these cases will yield generalizable concepts that may be applied
throughout the U.S.
Unit of Analysis: County
The health care context in Michigan is unique in several ways, e.g., all hospitals

are nonprofits, and it is one of only 12 states that has not initiated market reforms that

233

permit wider development Of individual or small group health care coverage products.
However, the main thrust of this study is on change agents and policy entrepreneurs, and
the insights gained from this research were proposed to be widely applicable to other
communities in the U.S., as well as to theories of local-level policy entrepreneurship in
other issue areas (e. g., antigrowth, environmental or education policy entrepreneurs).

The proposition would appear to hold true. The nonprofit status Of hospitals
within the studied counties certainly did not hinder their desire to compete, and thus there
appears to be no reason to believe that this unusual nonprofit environment caused the
market context to be more or less amenable to a friendly health policy subsystem. We
know health systems across the country are powerful, as they were in the four Michigan
localities studied here (Rice, 1998). The structure of local health systems share common
contextual components across the United States: private firms (especially small firms) are
passive purchasers of health care coverage with a great deal Of latent, decentralized,
market power, and only the largest private firms have begun to wield that power. The
structure of local markets vary widely in the same manner as in the four cases, some
localities have varying mixes of small and large firms, firms owned by entities outside Of
the community and sometimes outside of the state or country, and have differing
proportions of the workforce employed by public entities or employed outside of the
community within which they live. At a macro level, state Medicaid agencies across the
country are shifting from passive payers of claims to aggressive purchasers who seek to

wring more value from their federally mandated Medicaid expenditures.

234

 

More importantly, the two clearly identified entrepreneurs in Muskegon and
Wayne behaved very much in the manner of entrepreneurs as predicted by the literature.
This means that theories of policy entrepreneur behavior appear to be easily and validly

applied to an understanding of similar behavior at levels of governance below the state.

Assessment of the Decision Models

Constraints on the Assessment of the Decision Models

The assessment Of the decision models from the results Of the four cases provides
no basis for accepting or rejecting the propositions based on either model. Testing those
propositions would best be achieved by running simulations on the models and using
those results to assess the decision models’ propositions. The running of simulations on
these models was not proposed, as this research was considered exploratory, and thus
focused on the development of decision models that would articulate the relative
importance of the two key processes, coalition building and development and launch of
an innovation, on the outcome of either funding or pursuing change. The results of the
four case studies would then be compared to the decision models. Thus, the four case
studies do provide a basis for assessing what happened in each case relative to the
outcomes delineated in the decision models, and using that assessment to guide a future
research study based on the running of simulations.

The area least amenable to assessment is that of the choice to fund/not fund, or

pursue change/avoid change. For the outside-inward model, this assessment is

235

handicapped by the absence of one Of the two reports used as the basis for this decision
by WKKF staff. We thus do not have a basis for comparing the actual decisions to fund
with the decisions not to fund. The same is true for the inside-outward decision model.

We have only one case as an example, and no interview questions on the decision to

pursue change/avoid change were included in the interview instrument.

Assessing the Outside-Inward Decision Model

The three cases of outside-inward change had the following outcomes, presented
in Table 15. Of the three CCHMS cases, the most difficult to assign an outcome path to is
Calhoun. It would appear as if it should be assigned a low readiness level, based on the
statements of the WKKF staff responsible for making the decision, as well as to the fact
that its health care market was a subset of a market dominated by a neighboring county.
As well, the project director role proposed in the model was usurped by the team of
market entrepreneurs who saw an Opportunity to lower their own firms’ health care costs,
and proceeded to use CCHIP’s funds to facilitate the development and launch Of a
product that would achieve those cost reduction goals. No strong coalition supporting the
innovation can be said to have been formed, and the success of the one formal vote
required to access CCHIP funding to facilitate the development and launch Of the
innovation should not be taken as evidence that a strong coalition was formed to support
the SMHPA innovation. As proposed in Chapter Four, a weak coalition would be
anticipated to result in the development and launch Of a less innovative innovation, and

that should be viewed as the case, since the innovation had only a fuzzy clause in its

236

charter to pursue the charitable purpose of extending health care coverage to the
uninsured after SMHPA achieved some measure of profitability. Thus the payoff
achieved by the change agent’s investment in Calhoun was ac - l .58

The Muskegon case should be viewed as an unambiguous success and positive
return on investment to WKKF. The county had a high readiness level (although the
efforts required of the project director and her team of entrepreneurs were many and
varied) as it had experience with managed care in both the private and public sector,
hospital competition was an established fact, and physicians, also weakened over some
period of time, played an important role on the margins in both coalition building and in
development and launch. As noted above, there was some small amount of overlap
between the end of the coalition building stage and the start of the phase of development
and launch of the innovation. However, this overlap was short. What is not directly
captured in the decision model is the use of coercion on the part of the project director,
although we can state that the ability to convince the most cooperative hospital that it
could “go it alone” (thus pressuring the other hospital to agree and sign its contracts) is
captured in the ways in which the components of the probability pl are based in part on
the entrepreneurial skills and assets possessed by the project director. Access Health was

launched, and resulted in a payoff of 06 to the foundation.

 

’8 As footnoted above, the ultimate failure of SMHPA after launch is not a
part of this study, nor captured in the decision models. However, we

should anticipate that there was some reduction in the payoff of or - 1 as a
result of the closing of SMHPA only a year after launch.

237

The counterfactual of the importance of the project director possessing the
necessary skills and assets ascribed to policy entrepreneurs by the literature would seem
to be relatively easy to establish using the Muskegon case. Although the county should be
one perceived as at a high level readiness, with only two levels of readiness identified,
these levels then have broad parameters. It is arguable that Muskegon was perhaps at the
lower end of the high readiness level category, given the extensive use of the skills and
assets ascribed to policy entrepreneurs by the project director and her collective team of
entrepreneurs that were employed in order to build a strong coalition that then supported
the development and launch of the innovation.

If Muskegon can be defined as an unambiguous success, St. Clair must stand as an
unambiguous failure Of the decision to fund to provide a return on foundation investment.
The community’s readiness was clearly low, as detailed in its zero level of managed care
penetration at the beginning Of funding, and that competition among the two large
hospitals had only just ensued. At best we can state that a weak coalition was built, and
no innovation was ever developed and launched. This resulted in the worst possible
payoff for WKKF, 5.

There is no evidence to support the notion that the curtailment in WKKF funding
would have altered this outcome. Rather, we can argue that, once shortages in the
projected resources available for firnding were identified, WKKF staff correctly identified
that the St. Clair effort was a losing proposition, that community readiness had been low,
at least that the first two project directors had not the necessary set of skills and assets

ascribed to policy entrepreneurs, and perhaps that even if these persons had these skills

238

and assets, the effort would not have succeeded. The foundation staff therefore decided to

cut their losses on the effort and defirnd CCHMS of St. Clair.

Discussion Of the Outside-Inward Decision Model’s Propositions

Model Proposition #1
“Increases/decreases in p() are positively related to increases/decreases in the value of p,.
This is because p I is based in part Of the readiness of the community, as well as the
entrepreneurial skills/assets of the project director.”

It would appear by examining the two cases where the project directors were the
leaders of the effort, that context does influence the probability of achieving or as
proposed. These two cases provide evidence for this statement, and it is quite possible
that had St. Clair had a policy entrepreneur throughout the life of the initiative, the local
market context had not weakened the policy subsystem enough to permit the successful
building of a strong coalition, and the development and launch of an innovation.

Model Proposition #2
“The probability pl is monotonically and positively related to p3. This is because the
building of a strong coalition directly and positively affects the project director’s ability to

then develop and launch an innovation. Thus, increases in pl will result in increases in

,9

P3-

The evidence from the three case studies suggests that this proposition is true.
This is best represented by the failure to build a strong coalition in St. Clair, which

stymied any effort towards the development and launch of an innovation. However, a

239

 

weak coalition developed and launched a less innovative innovation in Calhoun, which
means that weak coalitions can succeed.

Model Proposition #3
“The probability l-p1 is monotonically and negatively related to p,,. This is because the
building of a weak coalition directly and negatively affects the ability Of the project
director to develop and launch an innovation.”

NO case provided evidence to either support or refute this proposition.

Model Proposition #4
“Increases/decreases in p2 are monotonically and positively related to p,. This is because
the building Of a strong coalition directly and positively affects the project director’s
ability to then develop and launch an innovation. Thus, increases in p2 will result in
increases in p,.”

The St. Clair case study provides support for this proposition. The inability of the
project directors to build a strong coalition appears closely relate to the fact that no effort
to develop and launch an innovation ever occurred.

Model Proposition #5
“The probability l-p2 is monotonically and negatively related to p6. This is because the
building of a weak coalition directly and negatively affects the ability of the project
director to develop and launch an innovation.”

As in the preceding proposition, St. Clair provides support for this proposition.
The formation of a very weak coalition led inevitably (in retrospect) to the absence of any

effort to develop and launch an innovation.

240

 

   

Final Conclusions from the Outside-Inward Model of Change

Because this research took the form of three case studies Of outside-inward
change, only weak evidence exists to support four of the five prOpositions related to the
outside-inward decision model. The other proposition had no evidence for support or
rejection, as no case followed the path necessary to provide evidence. The evidence from
the three CCHMS cases weakly supports the idea that it is indeed a mixture of context and
the skills/assets of the project director that lead to achieving the change agents’ preferred
outcome and payoff. Most important, none of the three cases fails to match the model.

An unanswered question is that of how WKKF staff perceived the value Of the
highest and lowest payoffs. NO interview data exists to assess what these two perceived
values were, nor is it possible to assess the perceived distance between these two values.

The biggest question remains the absence of evidence concerning the decision to
fund or not fund. Because WKKF staff were unwilling to fully share the information and
criteria used for arriving at this decision, no conclusions may be drawn on whether or not
their criteria were useful. What can be stated, though, is that, given the case of St. Clair,
funding an initiative in a county where readiness is low does appear to make the desired
outcome more difficult to achieve. As well, the ceding of the decision to hire a project
director also appears to open up the possibility, as described above in St. Clair, of a hiring
decision being made to purposely hire a person who does not have the skills and assets

necessary to improve the probability of achieving the desired outcome.

241

Assessing the Inside-Outward Decision Model

Only one case of inside-outward change was included in this research study.
Drawing conclusions from this regarding the structure of the model, its probabilities and
payoffs would be inappropriate. However, this case does present an interesting possibility
for some discussion of the inside-outward decision model, because it would appear that
iteration of the model’s processes had occurred, and the effects of this iteration can be
compared to the discussion of the potential effect of iterations from the preceding chapter.

The Wayne County Executive’s previous efforts to develop innovations to provide
some measure of coverage for their indigent population (and thus to provide
reimbursement to local health systems for the provision of services to this population) are
candidates for possible previous iterations of the same process modeled in the inside-
outward decision model. It would appear that a coalition to support the development and
launch Of the HealthChoice innovation had already been largely developed in these
previous innovative efforts, the most recent being the development and launch of
PlusCare.

This appears to be possible based on the speed with which the Wayne County
Executive determined that a window of opportunity for the development of the
HealthChoice innovation existed. PlusCare was determined to have a “real” surplus of
funding sometime in 1993, and by May of 1994 the innovation had been developed and
launched. The time needed to complete the process in Wayne County was then at most a
third of the time needed to accomplish the same innovation in Muskegon. Assuming the

model is appropriately structured, this suggests that either the context was more amenable

242

to change, the policy entrepreneur had high levels of the skills and assets ascribed to
policy entrepreneurs in the literature and effectively used them or both. It also seems
likely that iteration played a role in making the process of developing and launching
HealthChoice occur at such a rapid pace.

As mentioned in the preceding chapter, iteration could have several positive
effects. It could lead to the persistence of a strong coalition (or some portion of a strong
coalition) from a previous iteration to a future iteration, it could expand the skills/assets
of the policy entrepreneur, and it could hone the ability of the policy entrepreneur to
correctly assess the existence Of a window of opportunity, based on context and the
matching of their skills/assets to the opportunity. The evidence from the Wayne County
case, as it is only one case, is insufficient to draw conclusions on which, if any of these, is
true, and how and where the effects of iteration altered the probabilities. However, it
would seem to be the case that previous iterations of the decision model had been played
in the recent past, and that these iterations, with the very same actors being involved in
the iteration studied here, could have had some positive affect on the achieving of the best

payoff and outcome for the policy entrepreneur.

Final Conclusions on Both Decision Models

Based on the evidence from the case studies, the decision models do appear to
appropriately represent a good first attempt at modeling the processes and constraints on
the development and launch of innovations, whether sparked by the presence of an

external change agent or the decision to act Of an indigenous policy entrepreneur. Much

243

 

further research would need to be conducted to truly test these decision models, but this
exploration of the usefulness of these models at least suggests that further research would

be worthwhile.

244

Chapter Six: Conclusion and Implications for Future Research

This research attempted to understand the ways in which both foundations and
state agencies could assume the role of an external change agent to encourage innovation
at the community (county) level. This research also attempted to understand whether or
not the actions of external change agents would draw policy entrepreneurs out to lead the
effort to develop innovations at the community level. TO these ends, a two-pronged
research effort was undertaken, using four county-level change efforts within Michigan,
three funded by the W. K. Kellogg Foundation from 1993 to 2000. First, based on seven
research questions, a qualitative case study was designed and implemented that resulted
in a total Of six case studies, one of a foundation change agent, one of a state agency
change agent, and four cases of attempts at community-level innovation, with one case
being illustrative of innovation occurring in the absence Of a change agent. Second, two
decision models were designed, one of an external change agent deciding to fund/not
fund a change effort, the other a model of an indigenous local-level policy entrepreneur
deciding to pursue/not pursue a change effort. Qualitative interview data was collected
from key informants involved in all four cases Of local-level change efforts. The research
questions and the decision models were then assessed based on the qualitative interview

data collected for this study.

245

Answering the Research Questions
It is now time to revisit and attempt to answer the study questions presented in

Chapter One. The study questions upon which this research was based were:

1. How [did] a foundation play the role of change agent to encourage a local
partnership that will, through the development and launch of some program
or product, extend health care coverage to the uninsured?

A foundation change agent did imperfectly assume the role of change agent and
the assumption of this role did encourage a policy entrepreneurship team in one case and
a market entrepreneurship team in another case to develop innovations that would
irrrrnediately or in the future extend health care coverage to the uninsured. The primary
facilitator provided by the foundation change agent, due to the imperfect assumption of
the role of change agent, was that of resources and a mission that provided a broad guide
to the policy/market entrepreneurs for action, and the financial wherewithal to conduct
entrepreneurial activities that led to the development and launch of the innovations.
However, in at least one case studied, this was not enough to attract an entrepreneur and
spark the necessary entrepreneurial activity.

2. How [did] state government play the role of change agent to encourage the
development of a local partnership that will, through the development and
launch of some program or product, extend health care coverage to the
uninsured?

In no way did the state agency encourage the development of a local partnership in

any of the four cases studied. However, in two of the cases, a state agency played one of

246

 

 

 

 

 

 

 

the roles defined in the literature, that of resource linker, which was critical to the

successful development of launch of these innovations. Furthermore, one other state

agency played an important facilitative role by using its regulatory authority to interpret
regulations in a manner that favored the successfirl development and launch of both of the
aforementioned innovations.

3. What persons filled the role of local policy entrepreneur, and how did their
placement within the local policy and market contexts (i.e., their social
position and employment role) affect their ability to pressure for the
development and launch of some program or product to extend health care
coverage to the uninsured?

Of the four cases studied, the two clearly identified policy entrepreneurs were both
persons with extensive political backgrounds, one a former staffer for a U.S. Senator, the
other a long time county government Official. Both possessed and used many Of the skills
ascribed to policy entrepreneurs in the literature on that subject. Because both were
seeking to develop and launch innovations within a technically complex issue area, they
assembled collective entrepreneurship teams to support their efforts to build supportive
coalitions to develop and launch their innovations. In the other two cases studied, no
policy entrepreneur appeared as predicted, but in one case market entrepreneurs appeared

and developed a purely private sector innovation.

247

 

 

4. How [did] local policy entrepreneurs go about building collective
entrepreneurship teams to permit the development and launch of some
program or product to extend health care coverage to the uninsured?

In the two cases where clearly identified policy entrepreneurs appeared, they used
the tactics described in the literature on collective entrepreneurship to build teams that
contained persons whose talents and interests covered the set of skills necessary to
develop and launch innovations to extend health care coverage to the working uninsured.
Teams were assembled either through the hiring of qualified individuals, or the
contracting of consultants who became team members. Where necessary these skill sets
included all of the issue-specific knowledge, ability to make causal arguments,
manipulation of the media, and alter the costs of the innovations or just potential partners’
perceptions of cost in order to build a coalition that would support the development and
launch Of their innovations.

5. How [did] the local policy entrepreneur alter the structure of costs and
benefits for those members of the partnership? Did the local policy
entrepreneur alter the members’ perceptions of cost?

The two clearly identified policy entrepreneurs, where necessary, altered the
actual costs of the innovation, but otherwise used causal arguments to alter the
perceptions of the cost of their innovations to build a supportive coalition of partners for

their innovations. Both policy entrepreneurs also employed coercive tactics to Obtain

cooperation where causal argumentation was insufficient.

248

 

 

 

 

6. How [did] the political and market contexts affect the ability of local policy
entrepreneurs to build a partnership around the issue of the development
and launch of some program or product to extend health care coverage to the
uninsured?

The political contexts did not appear tO affect the prospects for the deveIOpment
and launch of any of the four innovations. However, an existing market context where
long term changes had weakened some actors within the local health policy subsystem
enough to cause them to be willing to consider innovative ideas from actors outside of the
system, did appear to be an important component in the success of the two entrepreneurs
who developed rrrixed public-private innovations. Where the market contexts had not
experienced long term change, the actors seemed far less willing to accept innovative
ideas from outside the health policy subsystem, or indeed to consider collaboration with
their competitors to be an Option.

7. Could models developed at the local level in Michigan apply elsewhere?
Models of local level policy entrepreneurship can apply elsewhere, as it appears

that the key elements to the appearance and success of local-level policy entrepreneurs are

the resources they have to engage in activities (whether available internally or provided
by an external change agent), and a market context that had created a window of
opportunity favorable for action. Although the Michigan context is dominated by
nonprofits, health services researchers have argued in the past that few differences exist
between for profit and nonproft actors in this industry. Rather, differences in their

behaviors tend to stem from the market context within which for profits/nonprofits find

249

 

 

themselves Operating (i.e., competitive versus noncompetitive and the extent of managed
care penetration). This research did find that at the same points in time, contexts,
especially market contexts, may vary greatly within an industry, and it is the market
context that may differentially weaken a policy subsystem. Thus, it is the local context,
especially the local market context, that may weaken a local health policy subsystem
enough to permit the introduction of an innovation from outside the industry. These local
market contexts can be expected to vary substantially across the nation in the same
manner as they were found to vary within Michigan. The elements of context favorability
should be expected to vary greatly from locality to locality, but this research appears to
demonstrate that where the mix of resources and context are favorable, policy

entrepreneurs will appear, take action, and develop and launch innovations.

Contributions Of this Research

One of the contributions of this research has been the explication of how a
foundation would be expected to go about assuming the role of an external change agent
to encourage innovation within a technically complex issue area. Although the results of
the foundation effort studied were mixed, an understanding of the ways in which a
foundation could go about the decision to fund or not fund such a change effort were
formalized and weak support was found for the formal decision model used to provide a
structured understanding Of this process. The formal model shed new light on two aspects
Of the decision to fund or not fund a community-level change process; one was the need

to choose a community that was ready for change in terms of its political and market

250

 

contexts, the second was that the decision to fund was intimately tied to the valuation of
the best and worst payoffs (outcomes) for the foundation acting as a change agent.

One other change agent was studied — a state agency assuming the role of a
change agent. The literature on state and federal agencies was used to inform propositions
based on how one could expect a state-level agency to assume a change agent role at the
local level. Again, as in the study of the foundation change agent, the results were
decidedly mixed. However, the results supported the findings of previous research on this
subject, as well as the literature on policy entrepreneurs that stressed the need for these
persons to have existing social connections with important actors.

This research proposed and assessed a link between external change agents and
policy entrepreneurs at the local level. This proposed link was supported by the
qualitative interview data, which also lent support to the formal model of the interaction
between a change agent’s decision to fund/not fund, and the proposed necessary
appearance of a policy entrepreneur to increase the probability for the successful
development and launch of an innovation.

An important contribution of this research has been to show that theories Of policy
entrepreneurship designed to model behavior of these individuals at the federal and state
level may be validly and easily applied to the local level. The skills and assets that
previous research on policy entrepreneurs defined as necessary for successful
entrepreneurship and innovation appear to be as necessary at the local level as they are at

the federal and state level. As well, this research has shown that these entrepreneurs need

251

not be elected politicians or bureaucrats, but may be citizens drawn out by the activities of
an external change agent.

Previous research on policy entrepreneurs has shown the importance of context as
a predictor of the success of innovative efforts lead by policy entrepreneurs. This research
effort has produced results that support this concept at the local level, and presents two
formal models that begin to incorporate the importance of context as one predictor of
successful innovative change efforts.

Another predictor of the formal modeling of the decision of an indigenous policy
entrepreneur (in the absence of an external change agent) to pursue/not pursue a change
effort is similar to the findings from the change agent decision to fund/not fund. The
decision is a function of both the entrepreneur’s perceived favorability of the context to
permit change, and the relative valuation the policy entrepreneur places on his best and
worst payoffs (outcomes). This is a particularly exciting finding, and one that deserves
further research.

The relatively recent study of collective entrepreneurship was also elucidated and
assessed in this research project. The findings support the idea that in technically complex
issue areas, we should expect that a collective team Of entrepreneurs is best able to
combine the skills and assets ascribed to individual policy entrepreneurs by the literature.
These collective teams were found to have formed in two local-level cases (the examples
of unambiguous success), and Operated much as the previous research on this subject

predicted.

252

 

 

This research also presented a new theory that entrepreneurs may find themselves
Operating across multiple contexts at the same time, both political and market. It
presented a theory for reconciling the differential use of information across these two
contexts, but found only weak support for this behavior. To recap, an entrepreneur
operating in a political environment spreads innovative ideas as widely as possible, while
an entrepreneur acting in a market environment closely guards their innovative ideas. The
reconciled behavior proposed was that an entrepreneur acting across both contexts at the
same time would guardedly disseminate their innovative ideas. However, the effort to
hide information during interviews on the part of market entrepreneurs in one case study
suggests that this proposed extension of the theory of information use by entrepreneurs
acting across both contexts may still be valid. Further research is required prior to
accepting or rejecting this theoretical extension.

Lastly, this research presented formal models that defined a two step process for
entrepreneurial activity in a technically complex issue area. Step one was the building of
a supportive coalition, and step two was the development and launch of an innovation.
The results of the qualitative case studies provided support for this proposed two step

process. This is another exciting finding of this effort that deserves further research.

Implications for Future Research
Both private foundations, and to a lesser extent state agencies, sometimes assume
the role of an external change agent to encourage innovation at the local level. As this

research intended to prove, that activity could and did spark other activity by one person

253

who exhibited the characteristics of a policy entrepreneur, as described in that literature.
One area for fiiture research would be to replicate this research across other foundations’
efforts to act as external change agents, and see if persons with the characteristics of
policy entrepreneurs appeared, and what effect these persons’ appearance (or absence)
had on the successful realization of the change agents’ goals.

The efforts of change agents to successfully encourage this kind of activity are not
guaranteed, and demand extensive planning prior to funding as well as a willingness to
fully assume all aspects of the role. Planning should focus on the specific issue area
within which innovation is desired, and the amenability of that issue area to change. The
amenability of an issue area to change should be expected to vary greatly at the local
level, as communities do not appear to experience the effects of changes within issue
areas at the same time and to the same extent. The presence of a policy subsystem
resistant to innovation should always be anticipated within technically complex issue
areas, and the need for a policy entrepreneur to lead the effort to exploit cracks, if any, in
the subsystem should be considered a necessity. Planning and assessing the amenability,
the readiness, of communities to engage in change appears to be a critical early precursor
to success.

Thus, one implication for future research would be to focus on other foundation
efforts to assume the role of change agent at the local level, and to assess the effect of
differential foundation planning prior to funding on the success of these efforts.

Perhaps the richest unmined area for research related to change agent activity is

that of federal agencies assuming the role of change agent through their grant making

254

 

efforts at the local level within the United States. DO federal agencies attempt to fully
assume the role Of change agent? Is their change agent behavior different from
foundations or state agencies? These are questions that likely have interesting answers
which could help guide federal policy in a variety of issue areas where local change is
desired.

The outside-inward decision model gives rise to the question of whether change
agents should consider not only on the amenability of the contexts of the communities
they fund, but also whether or not the change agent should be certain that the leadership
for the funded initiative possesses the necessary set Of skills to maximize the probabilities
for the building of a strong coalition and the development and launch of some innovation.
Engaging in this effort would appear to reduce the risk of poor outcomes for the change
agent, but is it an activity that change agents engage in?

As important, do change agents approach communities with the idea that their
change efforts can be iterated, assuming they can at least fund an initiative that has a high
probability of producing a strong coalition, which would increase the community’s
readiness tO support change at a future iteration? Or do change agents perceive these as
single shot situations, where they must succeed or fail on the first attempt?

Another area ripe for study is that of two-stage grant making processes. The
decision models that would reflect this process would also be anticipated to be lotteries,
with the same choice set of fund/not fund. It should be expected, though, that there would
be two lotteries, the first of funding/not funding a planning grant, with the desired goal of

locating a policy entrepreneur who successfully builds a strong coalition. Those

255

 

 

communities that are firnded asa part Of the planning process that appear to the change
agent as having achieved this goal would then decide to fund/not fund an implementation
grant, which would have the desired goal of developing and launching some innovation.

The study Of policy entrepreneurs at the local level, i.e., below the level of state
government, remains a wide Open field of study. Schneider, Teske, and Mintrom have
broken much of the necessary theoretical ground for this research, however, assessments
of the validity of theory in this area requires more testing and refinement. The findings of
this research suggest that whether in the presence or absence Of external change agents,
policy entrepreneurs appear, where necessary form collective teams of entrepreneurs, and
develop and implement innovations. This study was purely exploratory, asking the
questions of whether or not entrepreneurs operate at this level? Do they behave in the
ways predicted by the literature for the same kinds of individuals at the federal and state
levels? These findings suggest that they do exist, and they do act as predicted by the
literature. However, these findings are only a start.

The inside—outward decision model gives rise to its own questions that are not
restricted to the study of local level policy entrepreneurs. Do policy entrepreneurs assess
their own abilities, match them to the context they face, and use that assessment of the
match between their abilities and the context to develop a definition of what constitutes a
window of Opportunity? If they do, how important is the ability to assess one’s abilities
and match them to a situation to the development of a sound perception of whether or not
a window of opportunity is present? Do policy entrepreneurs iteratively face the situation

modeled here? Do past iterations influence policy entrepreneurs’ ability to correctly

256

 

assess their own skills, improve their skills, develop strong or weak coalitions that persist
into the next iteration, as well their ability to assess what constitutes a window of
Opportunity? Do policy entrepreneurs pursue change in situations where they perceive
that the probability of not achieving a complete “win,” is low, but with the belief that they
can at least build a strong coalition that will persist into future iterations, thus broadening
future windows of opportunity? These are questions that, if answered, would broaden our
understanding of policy entrepreneur behavior regardless of what level of governance is
being studied.

Across both change agents and policy entrepreneurs, this research has raised the
question of how both actors place values on their best and worst possible payoffs. How
valuable can a become? How negative a payoff can 5 be? How do these actors assess the
distance between the value of or and the value of 5? Answering these questions is critical
to understanding how these actors make a choice to act or not to act.

A goal of future research for this author will be to run simulations of both
decision models to better test the propositions related to both models presented in Chapter
Four. These simulations will hopefully begin to provide some partial answers to the

questions raised by both models.

257

 

APPENDICES

258

 

 

 

APPENDIX A

CCHMS Interview Protocol
Comprehensive Community Health Models of Michigan Initiative (CCHMS)
Project Case Study

Respondent #

 

Date/Time/Location

 

Interviewer

 

This meeting/interview is a part of the data collection strategies for the project case
studies being conducted as a part of the Cluster Evaluation of the W. K. Kellogg
Foundation’s Comprehensive Community Health Models of Michigan Initiative
(CCHMS). The purpose of the project case studies is to assess project interaction with
public, market, and organizational policies and to describe how, or if, the project will be

sustained over time.

If you agree, I would like to tape record our discussion and have a transcript made to aid
data analysis. I will be keeping the transcripts and codebook throughout the analysis and
for at least three years following that for archival purposes. When all the analysis is

complete, the transcripts and codebook will be destroyed.

259

 

You may decline to answer any questions and you may end this interview at any point.

Do I have your permission to record and make a transcript of this meeting/interview?

Yes No

If you have any questions about this evaluation, the person to contact is Dr. Jim Dearing,

517-355-1820.

TAPE ON (if approval given, remaining questions asked even if interview is not taped)

This is interview #

First I would like to discuss the use of this information. For the most part, information
will be reported in aggregate form. However, on occasion, it may be useful to include
specific quotes from your transcript, without identifying you. It is important to note that
while your name will not appear on any published material, it may appear on internal
project documents. To the extent possible, your interview will be reported internally by a

unique code known by me and a small subset Of the research team.

DO I have your permission to use information from this meeting/interview for the
purposes Of research and later publication?

Yes NO

260

 

Do I have your permission to use quotes?

Yes NO

Because we have limited time, I may on occasion have to direct the discussion, however,

if time allows, we can go back and pick up on discussion points.

DO you have any questions before we begin?

Interviewer Note:

A note about the project case studies and the C CHMs Cluster Evaluation:

The project case study component of the CCHMS Cluster Evaluation is concerned with
assessing policy and sustainability aspects of key program initiatives undertaken by the
sites. A key question we are exploring is whether a particular project can be done
sustained solely by the community or if projects require other supports and mechanisms
in order to be sustained. Therefore, the project case study will take into account the

barriers and facilitators to achieving goals of specific projects.

261

The Project Case Study is carried out through a series of semi-structured interviews,
combined with document review and participant Observation. The questions below are
grouped by type of respondent: (1) individuals in organizations responsible for
coordinating the implementation of the project under study (e.g., healthplan purchasing
alliance, project to increase access to health care, etc.); (2) paid staff and volunteers from
the local CCHMS Offices in Calhoun, Muskegon and St. Clair Counties; (3) state and
local public officials. In some instances, individuals in group 1 include paid staff and
select volunteers from the local CCHMS Offices. However, the questions for group 2 are
restricted to paid staff and volunteers from the local CCHMS Offices. The state and local
public Officials to be interviewed include those whom have been involved in various
discussions with individuals and organizations in the sites regarding CCHMs activities.
Some of these discussions specifically concern possible partnerships with the state.
These individuals are identified through document review and recommendations given by

individuals in the sites.

262

 

 

 

All Respondents

1. How did you come to be involved with the CCHMS initiative (broadly)? What is the
nature Of that involvement?

2. How would you describe the purpose or goal of (the project)? That is, why did the
county initiate this project?

3. What are some of the major things accomplished by (the project)?

State and Local Public Officials: Go to Q 4

Paid Staff & Volunteers: GO to Q6

Individuals Implementing Projects: GO to Q10

Select State and Local Public Officials

4. What, if any, actions have you (or will you) take to facilitate the goals of (X)?

5. In your view, what public policy issues need to be addressed if (X) <local CCHMS
oflice project under study> is to achieve its goals? How might this best (realistically)
be achieved?

GO to Question 24.

Paid Staff & Volunteers from the local CCHMS Office

6. How did the activities Of the governing board and work groups lead to the
development of (X)?

7. What is the nature of support for (X) in your County? (Probe: Who is supporting (X),
why, how?)

8. What is the nature of Opposition to (X) in your County? (Probe: Who is in Opposition,

why? What steps have been undertaken to work through the opposition?)

263

9. On what issues and in what way has your organization engaged the state and local-level
policy community?

GO to Question 10.

Individuals in Organizations Responsible for

Coordinating the Implementation of the Project

10. Did you seek out information on similar projects or innovations in counties outside of
(x county)?

11. What have been the major critical events that have lead to the success of (X)?
(Probe for changes in the project, the county, the staff)
12. What were some of the barriers your organization faced in trying to establish (X)?
(How were they addressed?) Prompt: What if any changes were necessary in public
policy or the private sector or your organization for the establishment of (x19?

13. What were some Of the facilitators that assisted your organization in trying to
establish (X)? Prompt: What public policy or the private sector facilitated the
establishment of ()0?

14. What do you perceive to be the benefits and costs of participation in (X) for you/your
organization?

Benefits:
Costs:

15. Do you perceive that these benefits/costs changed over time?
Changes in benefits:

Changes in costs:

264

16. What caused these changes in benefits/costs?
Causes of changes in benefits?
Causes of changes in costs?
17. What were some of the barriers and facilitators to your organization partnering with
others in the community to form (X)? (How were they addressed?)
18. What activities still need to occur for (X) to be a success? (How would you define
success?)
19. How is this project currently funded? Do you plan to change future funding of this
program? (If so, how will funding change?)
20. Do you believe that your organization will continue to champion this project in the
future?
21. Do you think there will be continued community support for this project?
22. DO you anticipate a time when (the project) will no longer be associated with
CCHMS?
23. Do you have or do you know of plans/efforts to adopt aspects of (X) in other
communities? (elaboration)
If paid staff have turned project over, questions 24-29 may be optional.
24. As you (and your organization) move forward with (X) what if any future changes in
public policy or the private sector do you anticipate and why? (Probe: For state and

local officials emphasize private sector.)

265

 

25. What, if any, actions have or will you take regarding a possible partnership/linkage
between state and local entities and project (x) (What is that partnership likely to look
like?) A

If a linkage with state or local entities is not expected, skip to question 30

If a linkage with state or local entities is expected (or underway) continue to question 26.

26. What have been some of the specific barriers and facilitators to establishing a
partnership/linkage with state and/or local entities around (X)? (How were the
barriers overcome or the facilitators used in order to move forward?)

27. If not specifically addressed in barriers/facilitators questions: What if any changes
are necessary in public policy or the private sector to establish a partnership/linkage
with state or local entities around (X)? Probe towards health.

28. What will be some of the major things accomplished by establishing a
partnership/linkage with the state around (X)? (likely time frame)

29. Who have been/will be some of the key players in trying to establish a
partnership/linkage with the state around (X) and what roles have they played? (What
roles are they likely to continue to play?)

30. Based on your experiences with (X) to date, what if anything would you do differently
and why?

31. Finally, do you think that project X will lead to health system change or changes in

the health system?

266

 

APPENDIX B

Wayne County Interview Protocol

Change Agents and Policy Entrepreneurship at the Local Level

Project Case Study

Respondent #

 

Date/Time/Location

 

Interviewer

 

This meeting/interview is a part of the data collection strategies for the project case
studies being conducted as a part of the “Change Agents and Policy Entrepreneurship
at the Local Level.” The purpose of the project case studies is to assess project
interaction with public, market, and organizational policies and to describe how, or if,

the project will be sustained over time.

267

 

If you agree, I would like to tape record our discussion and have a transcript made to aid
data analysis. I will be keeping the transcripts and codebook throughout the analysis
and for at least three years following that for archival purposes. When all the analysis

is complete, the transcripts and codebook will be destroyed.

You may decline to answer any questions and you may end this interview at any point.

DO I have your permission to record and make a transcript of this meeting/interview?

Yes NO

If you have any questions about this evaluation, the person to contact is Dr. Carol

Weissert, 517-353-3292.

TAPE ON (if approval given, remaining questions asked even if interview is not taped)

This is interview #

First I would like to discuss the use of this information. For the most part, information

will be reported in aggregate form. However, on occasion, it may be useful to include

specific quotes from your transcript, without identifying you. It is important to note

that while your name will not appear on any published material, it may appear on

268

 

 

...,4.-.

4.
.4”
,
y
4
n .

 

internal project documents. To the extent possible, your interview will be reported

internally by a unique code known by me and a small subset Of the research team.

DO I have your permission to use information from this meeting/interview for the
purposes of research and later publication?

Yes NO

Do I have your permission to use quotes?

Yes No

Because we have limited time, I may on occasion have to direct the discussion, however,

if time allows, we can go back and pick up on discussion points.

DO you have any questions before we begin?

All Respondents

1. How did you come to be involved with the HealthChoice/PlusCare initiative
(broadly)? What is the nature of that involvement?

2. How would you describe the purpose or goal Of HealthChoice/PlusCare? That is,
why did the county initiate this project?

3. What are some of the major things accomplished by (the project)?

State and Local Public Officials: Go to Q 4

Paid Staff & Volunteers: Go to Q6

269

 

Individuals Implementing Projects: GO to Q10

Select Local Public Officials
4. What, if any, actions have you (or will you) take to facilitate the goals of

HealthChoice/PlusCare?

5. In your view, what public policy issues needed to be addressed before
HealthChoice/PlusCare was able to achieve its goals? How was this achieved?

GO to Question 24.

Paid Staff & Volunteerg from the local CCHM§ Office

6. How did the activities of the governing board and work groups lead to the

development of HealthChoice/PlusCare?

. What is the nature of support for HealthChoice/PlusCare in your County? (Probe:
Who is supporting HealthChoice/PlusCare, why, how?)

. What is the nature of Opposition to HealthChoice/PlusCare in your County? (Probe:
Who is in Opposition, why? What steps have been undertaken to work through the
Opposition?)

. On what issues and in what way has your organization engaged the state and local-

level policy community?

Go to Question 10.

270

..::a'e« H

 

 

Individuals in Organizations Responsible for
Coordinating the Implementation of the Project
10. Did you seek out information on similar projects or innovations in counties outside Of

Wayne county?

11. What have been the major critical events that have lead to the success of

HealthChoice/PlusCare? (Probe for changes in the project, the county, the staff)

12. What were some of the barriers your organization faced in trying to establish
HealthChoice/PlusCare? (How were they addressed?) Prompt: What if any changes
were necessary in public policy or the private sector or your organization for the
establishment of HealthChoice/PlusCare?

13. What were some of the facilitators that assisted your organization in trying to
establish HealthChoice/PlusCare? Prompt: What public policy or the private sector
facilitated the establishment of HealthChoice/PlusCare?

14. What do you perceive to be the benefits and costs Of participation in

HealthChoice/PlusCare for you/your organization?

Benefits:

Costs:

271

 

15. DO you perceive that these benefits/costs changed over time?

Changes in benefits:

Changes in costs:

16. What caused these changes in benefits/costs?

Causes Of changes in benefits?

 

Causes of changes in costs?

17. What were some Of the barriers and facilitators to your organization partnering with
others in the community to form HealthChoice/PlusCare? (How were they

addressed?)

18. What activities still need to occur for HealthChoice/PlusCare to be a success? (How

would you define success?)

19. How is this project currently funded? DO you plan to change future funding of this

program? (If so, how will funding change?)

272

 

 

 

20. Do you believe that your organization will continue to champion this project in the

future?

21. DO you think there will be continued community support for this project?

22. Do you anticipate a time when HealthChoice/PlusCare will no longer be associated

with Wayne County Executive?

23. DO you have or do you know of plans/efforts to adopt aspects of

HealthChoice/PlusCare in other communities? (elaboration)

If paid staff have turned project over, questions 24-29 may be Optional.

24. As you (and your organization) move forward with HealthChoice/PlusCare what if
any future changes in public policy or the private sector do you anticipate and why?

(Probe: For state and local officials emphasize private sector)

25. What, if any, actions have or will you take regarding a possible partnership/linkage
between state and local entities and project HealthChoice/PlusCare (What is that
partnership likely to look like?)

If a linkage with state or local entities is not expected, skip to question 30

273

 

 

If a linkage with state or local entities is expected (or underway) continue to question 26.

26.

27.

28.

29.

30.

What have been some of the specific barriers and facilitators to establishing a
partnership/linkage with state and/or local entities around HealthChoice/PlusCare?

(How were the barriers overcome or the facilitators used in order to move forward?)

If not specifically addressed in barriers/facilitators questions: What if any changes
were/are necessary in public policy or the private sector to establish a
partnership/linkage with state or local entities around HealthChoice/PlusCare? Probe

towards health.

What were some of the major things accomplished by establishing a

partnership/linkage with the state around HealthChoice/PlusCare? (likely time frame)

Who have been/will be some of the key players in trying to establish a

partnership/linkage with the state around HealthChoice/PlusCare and what roles have

they played? (What roles are they likely to continue to play?)

Based on your experiences with HealthChoice/PlusCare to date, what if anything

would you do differently and why?

274

 

 

31. Finally, do you think that HealthChoice/PlusCare will lead to health system change or

changes in the health system?

 

275

 

APPENDIX C

Tables

Table 1

National Spending on Health Care,

1965, 1980, 1994 and 1998

 

 

 

 

 

 

1965 1980 1994 1998
National Total (millions) $41.1 $247.2 $937.1 $1 ,149.1
Percent of GDP 5.7 8.9 13.5 13.5
Per capita $202 $1,052 $3,465 $4,094

 

 

 

 

 

Source: Health Care Financing Administration website,[Online] wwwhctagov ,

September, 2000.

*This is slightly lower than the OECD figure due to variances in calculating procedures.

Percent of National Spending on Health Care, by Sector

Table 2

1965, 1980, 1994 and 1998

 

Private

Public

 

Federal

State & Local

 

1965
75.0
25.0
11.7
13.3

 

1980
57.6
42.4
29.1
13.3

 

1994
55.2
44.8
32.2
12.6

 

1998
54.5
45.5
32.8
12.7 ,

 

Source: Health Care Financing Administration website,[Online] wwwhcl‘agov ,

September, 2000.

276

 

 

Table 3

Differential Information Use Across Milieus

 

 

 

 

 

 

 

 

 

 

 

Market Milieu Political Milieu Mixed Milieus
Entrepreneur
use of Guard Disseminate Guarded Dissemination
Information
Table 4
Comparison of Change Agent Roles by Type

Type Roles
Foundation Catalyst Solution Giver Process Helper Resource Linker
Government Information Resource Linker
Agency networker

 

 

 

 

 

277

 

 

 

 

P;

‘-

2: . _...

 

ll»l.rrll\lll.lllr

 

Table 5

Count of Interviews by Case

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Foundation change agent 2 (one of these interviews was conducted
as a focus group with three
respondents)

State agency change agent 1

Calhoun County 10

Muskegon County 9

St. Clair County 4

Wayne County 4

Table 6
Timing of Interviews by Case
Case 1nterviewer(s) Time frame

WKKF First Evaluation Team Spring 1995

MDCH Author Spring 2000

Calhoun County Second Evaluation Team* Spring 1999

Muskegon County Author Spring 2000

St. Clair County Author Spring 2000

Wayne County Author Winter 2000

 

 

 

 

* The members of the second evaluation team who conducted these interviews used

questions drawn from the author’s interview instrument, in addition to their own.

278

 

Table 7
Mathematical Calculations of Simulated Values for the
Outside-Inward Decision Model

Example 1

 

Where on = 5 and 5 = 0, and the values of the probabilities are p0 = 0.47, p1 = 0.53,

p2 =0.46 , p3 = .70, p4 = .24, p5 = .47, and p6 = .21.
EV(F) = .47(.53(.70x5 + (1 - .70)(O + 2)) + (1-.53)(.24(5-1 + (1 - .24)0)) +
(1 - .47)(.46(.47x5 + (1 - .47)(0 + 2)) + (1 - .46)(.2l(5 - l) + (1 - .21)0)).

EV(F) = 3.3

EV(NF) = .47(0 + 1) + (1 - .47)(0 + 3).

EV(NF) = 2.1

Thus, EV(F) > EV(NF). The decision should be to fund the change effort.

 

279

 

 

 

Table 8
Mathematical Calculations of Simulated Values for the
Outside-Inward Decision Model

Example 2

 

 

Where at = 5 and 5 = 0, and the values of the probabilities are p0 = 0.31, p1 = 0.47,

p2 =0.28 , p3 = 0.31, p4 = 0.15, p5 = 0.30, and p6 = 0.02.
EV(F) = .31(.47(.31x5 + (1 - .31)(0 + 2)) + (1-.47)(.15(5-1 + (1 - .15)0)) +
(l - .3l)(.28(.30x5 + (l - .30)(O + 2)) + (l - .28)(.02(5 - 1) + (1 - .02)0)).

EV(F) = 2.2

EV(NF) = .31(0 + 1) + (1 - .31)(0 + 3).

EV(NF) = 2.4

Thus, EV(F) < EV(NF). The decision should be to not fund the change effort.

 

280

 

 

Table 9
Mathematical Calculations of Simulated Values for the
Outside-Inward Decision Model

Example 3

 

 

Where (X = 10 and 5 = 0, and the values of the probabilities are p0 = 0.31, p1 = 0.47,

p2 =0.28 ,p, = 0.31, p4 = 0.15, p, = 0.30, and p6 = 0.02.

EV(F) = .31(.47(.31x10 + (1 - .31)(o + 2))+(1—.47)(.15(10-1+(1-.15)0))+
(1 - .31)(.28(.30x10 + (1 - .30)(0 + 2)) + (1 - .28)(.02(10 - 1) + (1 - .02)0)).

EV(F) = 6.6

EV(NF) = .31(0 + 1) + (1 - .31)(O + 3).
EV(NF)=2.4

Thus, EV(ID > EV(NF). The decision should be to fund the change effort.

 

281

 

Table 10

Proposition #1, Variables by Case

 

 

 

 

 

 

 

 

Variable Name WKKF MDCH
Score Validity Score Validity

FCA_CAT -3 M NA NA
FCA_SOL -2 M NA NA
FCA_HLP -2 L NA NA
FCA_LNK 2 L NA NA
SCA_LNK NA NA 8

‘ SC=A NET NA NA 2 L

 

 

 

 

 

 

NOTE: “NA” is a variable not applicable to that case. “M” refers to medium validity. “L”

refers to low validity.

282

 

111111.11 ll 1

 

Table 11

Proposition #2, Variables by Case

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Cross case Calhoun Muskegon St. Clair Wayne
Name

S V S V S V S V S V

LPE_SUP -7 L x x x x -7 H x x
LPE_KNO 30 H 6 M 19 H l H 4 L
LPE_CON 30 H -5 M 24 H 0 H 11 H
LPE_CAU 45 H 6 H 29 H -1 M 11 H
LPE_COA -30 H -44 H 28 H -1 8 H 4 H
LPE_LME 27 M 19 H 6 L 0 M 2 H
LPE_ING 48 H 13 H 26 H 7 H 2 H
LPE_COL 37 H 3 L 35 H -5 . H 4 H
LPE_MED 11 M 2 L 7 M x x 2 L
LPE_PRO 25 H -12 H 35 H l M l L
lgCOM 12 M -5 M 19 H -2 L x x

 

 

 

 

 

 

 

NOTE: “S” refers to the score. “V” refers to the validity. “x” refers to a variable that
registered no appearances for that case.

283

Table 12

Proposition #3, Variables by Case

 

 

 

 

 

 

 

 

 

 

 

Variable Cross case Calhoun Muskegon St. Clair Wayne
Name

S V S V S V S V S V

PER_COS -38 H -15 H -20 H -4 1 L

PER_APC 19 M 5 M 14 H x x x x

gACO 23 L x x 21 H x x 2 L

 

 

 

 

 

 

 

NOTE: “S” refers to the score. “V” refers to the validity. “x” refers to a variable that
registered no appearances for that case.

Table 13

Proposition #4, Variables by Case

 

Variable Cross case Calhoun Muskegon St. Clair Wayne

Name

 

S V S V S V S V S

 

LCO_LPC -9 L x x x x -13 H 4 H

 

 

 

 

 

 

 

 

 

 

 

 

LCO LMC -12 H -14 H 11 H -16 H 7 H

 

 

NOTE: “S” refers to the score. “V” refers to the validity. “x” refers to a variable that
registered no appearances for that case.

284

Proposition #4, Variables by Case

Table 14

 

 

 

 

 

 

 

 

 

 

 

 

Variable Cross case Calhoun Muskegon St. Clair Wayne
Name
S V S V S V S S V
LCC_PCC -3 x x x x -3 x x
fiMCC -l 8 -4 M x x —14 H x x

 

 

 

 

 

NOTE: “S” refers to the score. “V” refers to the validity. “x” refers to a variable that
registered no appearances for that case.

Table 15

Decision Model Outcomes for the CCHMS Cases

 

 

 

 

 

 

 

Case Outcome Path WKKF Payoff
Calhoun CL, WC, DP 0t - l
Muskegon CH, SC, DP or
St. Clair CL, WC, NP 5

 

 

NOTE: See Figure 4 for an explanation of the outcome paths and payoffs.

285

 

APPENDIX D

FIGURES

286

 

 

 

Figure 1, A Nested Market View Of the Health Care System

287

 

 

 

/
l : iiiiiiiiiiiiiiiiiii 1
1 State Cachement I
l ........ A. 1'93 .
1' """"""" {(115m) 5 ,3
l 2 Cachemgnr:::::‘i:::f """""
| ' Area
‘ ......................
l
| , """"""""""" 3
\\ Cachement 1 /

__—_—”—-_—___——I——Afe”afi—

 

Figure 1

This figure refers to both the public and private demand for health
care, the creation of health care resources, and the delivery Of
health care services

288

 

 

 

Figure 2, Uninsured Families by Work Status of F arnily Head, 1999

289

 

 

 

 

Figure 3, Workers Aged 18-64 without Health Insurance, by Firm Size, 1999

291

 

 

1 00-499

500-999
3.4%

   
 

13.7% 1,000+
17.1%

Public Sector
6.8%

Self-employed
13.9%

<10
20.9%

Source: March 1999 Current Population Survey

Figure 3

292

 

 

 

 

Figure 4, Model of Outside - Inward Change

293

Enid A

23:3

 

 

 

           

FORE brdlg

 

m0
AF;

    

a _. _ . :6: 3d»

 

294

 

 

 

 

 

 

 

 

Q. Q -_. N D .. ‘ Q
ermZU” E~Ow>ECH< OWUmEZQ w><034 OwUmEZQH
0:“ 0055. 2:: Em: awn—rag P v E Q v 91. v m + w v m .3 v m + _ v m
Or“ GEEK/$.17 5mm qmmmrwmflmm l l E v E v P v E

 

 

 

 

F 5:5 3559922; ormzmo
2? 2o. 5:5 9.639322; armbmal

II. | I'll I. ll. '1!!! ll!

mo“ wEE mzdsm com—50:
Vzln@:5k<mg now—ES.

UP Ugo—ct am $58: Eon—:2
2? 2o 389.0. 995.2ch

 

 

053.45.: Moo. 3. Samoa. >:m: 0::0

 

 

 

Figure 5, Model of Inside - Outward Change

295

 

m0
cc:
0
0.5 2w Eu
cob 3.33 :5
Q. m + M Q. l _

Enid m

>3de

0: \, / 2.
:5 2.3

«60:3,.@Smxzdsz...

V .............. .

>n >n\

m+_ m+w

 

$0
2. .5

Z? Um
2-1-; SOL
m Q.

«<0
2.1%
O
21 Cw 2w
2-1.; 25% 2-th
m + M Q I . m

 

 

 

 

rmmmzon
0; 950mm 855. SE. Em: _.nm&_.mmm
or” 0:88 855 2:: 5% 8330mm

I I'll I I‘ll 5 I'll l i ll

mm” 1528 ormzma
>0” >55 055%

mm” wEE mzosm ommlzamm-
30” ESE Saar 03:30:

IIIII. III. III III II... III lllllll llll [II

D? 095.2. an FEB: .6398.

 

1x0w>w=r§< 070E220”
P V 3
E v E v E v E.

 

35403” Dam—£20”
Qv91_vm+wvm+mvm+_vm

 

 

 

 

 

2? 2o .5395. amino—con.

@0155? N2: 3, 9.2.55. >:m: 0::0

 

296

 

 

 

BIBLIOGRAPHY

 

297

 

BIBLIOGRAPHY

Bachrach, P. and MS Baratz. 1962. “Two Faces Of Power.” American Political Science
Review, 56(4), 947-52.

Baumgartner, F, and B.D. Jones. 1991. “Agenda Dynamics and Policy Subsystems.”
Journal of Politics, 53: 1044-74.

Behn, R. 1991. Leadership Counts: Lessons Learned for Public Managers from the
Massachusetts Welfare, Training and Employment Programs. Cambridge, MA:
Harvard University Press.

Blumberg, L.J. and L.M. Nichols. 1995. Health Insurance Market Reforms: What They
Can and Cannot Do. Washington, DC: The Urban Institute.

Checkoway, B. 1981. Citizens and Health Care: Participation in Planning for Social
Change. Elmsford, NY: Pergamon Press.

Cobb, R. W., and C. W. Elder. 1983. Participation in American Politics: The Dynamics
of Agenda-Building. Second Ed. Boston: Allyn and Bacon.

Covington, S. 1997. Moving a Public Policy Agenda: The Strategic Philanthropy of
Conservative Foundations. Washington, DC: National Committee for Responsive
Philanthropy.

Cunningham, M. 1972. Private Money and Public Service: The Role of Foundations in
American Society. New York: McGraw Hill.

Cyert, R. M., and J. G. March. 1963. A Behavioral Theory of the Firm. Englewood Cliffs,
NJ: Prentice-Hall.

Dabasi-Schweng. 1965. “The Influence of Economic Factors.” In Kriesberg, M., ed.,
Public Administration in Developing Countries. Washington, DC: The Brookings
Institute.

Dickinson, PG. 1962. Philanthropy & Public Policy. Washington, DC: National Bureau
of Economic Research.

298

 

 

 

 

 

 

Doig, J. and EC. Hargrove, eds. 1987. Leadership in Innovation: A Biographic
Perspective on Entrepreneurs in Government. Baltimore: Johns Hopkins University
Press.

Doty, P. 1980. Guided Change of the American Health System: Where the Levers Are.
New York: Human Sciences Press.

Dow, M. 1985. “Science, Technology and African Famine.” BOSTID Developments. Vol.
5: 16-19.

Duncan, A. 1986. “Aid Effectiveness in Raising Adaptive Capacity in the Low Income
Countries.” In Lewis J. and V. Kallab, eds., Development Strategies Reconsidered.
New Brunswick: Transaction Books.

Eyestone, R. 1978. From Social Issues to Public Policy. New York: John Wiley and
Sons

Fronstein, P. 2000. Sources of Health Insurance and Characteristics of the Uninsured:
Analysis of the March 1999 Current Population Survey. Washington, DC: Employee
Benefit Research Institute.

General Accounting Office. 1994. States Use Illusory Approaches to Shift Program Costs
to Federal Government. Report to the Chairman, Subcommittee on Oversight and
Investigations, Committee on Energy and Commerce, House of Representatives.
Washington, DC: United States General Accounting Office.

HammOnd, T.H., and J .H. Knott. 1988. “The Deregulatory Snowball: Explaining
Deregulation in the Financial Industry.” Journal of Politics, 50, pp. 3-30.

Havelock, R.G. and M. C. Havelock. 1973. Training for Change Agents: A Guide to the
Design of Training Programs in Education and Other Fields. Ann Arbor: University
of Michigan, Institute for Social Research.

Health Care Financing Administration. 2000. [Online] Available website: wwwhcfagov.
January, 2001.

 

Health Care Trends Report. 1999. Special Issue, July 1999. Washington, DC: Health
Care Trends, Inc.

Heclo, H. 1978. “Issue Networks and the Executive Establishment.” in The New

American Political System, 87-124. Washington DC: The American Enterprise
Institute.

299

 

 

 

 

Kamel, SM. 1999. The Minimum Role of External Agents in Administrative Reform: The
Case of USAID in Egypt. Dissertation for the Degree of Ph.D., Michigan State
University.

Kaufman, H. 1981. The Administrative Behavior of Federal Bureau Chiefs. Washington,
DC: The Brookings Institution.

Kellogg Foundation. 1998. Evaluation Handbook. Battle Creek, MI: WK. Kellogg
Foundation.

Kingdon, J. 1995. Agendas, Alternatives, and Public Policies. New York: HarperCollins.

Knott, J. and C. Weissert. 1995. “Foundations and health policy: identifying strategies in
health programming.” Policy Studies Review, 14 (Spring/Summer), 149-160.

Langton, S. 1981. “What’s Right and What’s Wrong With Citizen Participation.” In
Citizen Participation, edited by S. Langton. Lexington MA: Lexington Books. Pp. 8-
20.

Lin, J. 1989. “An Economic Theory of Institutional Change: Induced and Imposed
Change.” Cato Journal, 9 (1), 1-33.

Marmor, T. 2000. The Politics of Medicare. (Second Edition) New York: Aldine
DeGruyter.

Mintrom, M. 2000. Policy Entrepreneurs and School Choice. Georgetown: Georgetown
University Press. -

Moe, T.M. 1984. “The New Economics of Organization.” American Journal of Political
Science 28: 739-777.

Moe, T. M. 1990. “The Politics of Structural Choice: Toward a Theory of Public

Bureaucracy.” In Organization Theory: From Chester Barnard to the Present and
Beyond, ed. Oliver E. Williamson. New York: Oxford University Press.

Mossberger, K, and K. Hale. 1999. “Information Diffusion in an Intergovernmental
Network: The Implementation of School-to-Work Programs.” Conference Paper
delivered at the 1999 Annual Conference of the American Political Science

Association.

Mohr, L. B. 1969. "Determinants of Innovation in Organizations." American Political
Science Review 63 :1 1 1-26.

300

 

O’Connell, B. 1987. Social Change Philanthropy in Action. Washington, DC: The
Foundation Center.

O’Gorman, F. 1978. The Role of Change Agents in Development. East Lansing, MI:
Institute for International Studies in Education.

Oliver, T.R., and P. Paul-Shaheen. 1997. “Translating Ideas into Action: Entrepreneurial
Leadership in State Health Care Reform,” Journal Of Health Politics, Policy and
Law, 22(3), pp. 721-788.

Olson, M. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups.
Cambridge, MA: Harvard University Press.

Perlstadt, P., C. Jackson,, P. Freddolino, and C. Sturdevant-Reed. 1997. CCHMS Cluster
Evaluation Planning Phase Final Report. Report delivered to the WK. Kellogg
Foundation.

Perlstadt, P., Jackson-Elmoore, C., Freddolino, P., and Sturdevant-Reed, C. 1999.
“Citizen Participation in Health Planning: A Case Study of Changing Delivery
Systems.” Research in Sociology of Health Care, 16, pp. 75-98.

Rabinowitz, A. 1990. Social Change Philanthropy in America. Westport, Conn: Quorum
Books.

Rice, T. 1998. The Economics of Health Reconsidered. Chicago: Health Administration
Press.

Ricketts, M. 1987. The New Industrial Economics. New York: St. Martin’s Press.
Riker, W. 1978. The Art of Political Manipulation. New Haven: Yale University Press.

Roberts, N. and P. King. 1996. Transforming Public Policy: Dynamics of Policy
Entrepreneurship and Innovation. San Francisco: J ossey-Bass.

Rondinelli, D. 1993. Development Projects as Policy Experiments: An Adaptive
Approach to Development Administration. New York: Routledge.

Schattschneider, BE. 1975. The Semisovereign People. Hinsdale, Illinois: The Dryden
Press.

Schneider, A. L., and H. Ingram. 1997. Policy Design for Democracy. Lawrence, KS:
University of Kansas Press.

301

 

 

 

 

Schneider, M. and P. Teske. 1992. Toward a Theory of the Political Entrepreneur:
Evidence from Local Government.” American Political Science Review, 86(3), 73 7-
47.

Schneider, M. and P. Teske. 1993. “The Antigrowth Entrepreneur: Challenging the
“Equilibrium” of the Growth Machine.” Journal of Politics, 55(3), 720-36.

Schneider, M., P. Teske, with M. Mintrom. Public Entrepreneurs: Agents for Change in
American Government. Princeton, NJ: Princeton University Press.

Schumpeter, J .A. 1939. Business Cycles: A Theoretical, Historical, and Statistical
Analysis of the Capitalist Process. New York: McGraw-Hill.

Selznick, P. 1979. The TVA and the Grass Roots. Berkeley: University of California
Press.

Shaheen, P. and H. Perlstadt. 1982. “The Hill-Burton Class Action Suits: Organization
and Impact.” Indiana Law Journal, 57 (3), 385-423.

Spitz, B., D. Butler and E. Kleinau. 1993. Comprehensive Community Health Models
(CCHMS) Project: Community Profiles. Waltham, MA: Institute for Health Policy,
Florence Weller Graduate School for Advanced Studies in Social Welfare, Brandeis
University.

Starr, P. 1982. The Social Transformation of American Medicine. New York: Basic
Books.

Tye, K. 1973. “The Political Change Agent.” In T raining for Change Agents: A guide to
the Design of T raining Programs in Education and Other Fields. Edited by R.G.
Havelock and MC. Havelock. Ann Arbor: University of Michigan, Institute for
Social Research. Pp. 143-150.

Walker, J. L. 1969. “The Diffusion of Innovations Among the American States.”
American Political Science Review 63:880-899.

Wandersman, A. 1984. “Citizen Participation.” In Heller, K., R.H. Price, S Reinharz, S.
Riger and A. Wandersman, Psychology and Community Change: Challenges for the
Future. Chapter 10. Homewood, IL: Dorsey Press.

Weinstein, J. 1997. Social and Cultural Change: Social Science for a Dynamic World.
Boston: Allyn and Bacon.

302

 

 

Weissert, C., and W. Weissert. 1996. Governing Health: the Politics of Health Policy.
Baltimore: Johns Hopkins University Press.

Wennberg J. 1996. The Dartmouth Atlas of Health Care. Dartmouth: American Hospital
Press.

Wildavsky, A. B. 1979. Speaking Truth to Power: The Art and Craft of Policy Analysis.
Boston : Little, Brown.

Wilson, J. Q. 1973. Political Organizations. New York: Basic Books.

World Bank. 1992. Governance and Development. Washington, DC: World Bank.

303

 

 

 

J.H.; 1......

2:3: : :

1

\

 

111111

21259134

L

.. .. .3
1L“:
.vmtnflh

.131.

r s... . is: .5

3.: ..
, a

z
.¢ .1. . 7 .
. £5.13»...

.2»... .
. r.» :1

x. ,, . .Zwu :
. v .
. n...

x

. :x E...

i :1. . r; 55....

:54... :3525.’

ӣ34. :3: i

: flunk—AK}. .
.1... s y :2. 1. ..

a. :55}
5 ..r..._.¥.»: 2

.
a

(w 1.
:35; f.
.. L...

,.
z .2
.213

. .2.

R E:
5i.