Rethinking Punctuated Equilibrium Theory: A Public
Administration Approach to Budgetary Changes
Carla M. Flink
What stimulates policy change in organizations? Punctuated equilibrium theory (PET) posits that
over time policy moves slowly, but also experiences large, rapid changes. Explanations for
punctuations have centered on institutional friction and disproportionate information processing.
Lacking in PET literature is a theoretical understanding of policy change aside from structural and
cognitive limitations. Other organizational features can create friction to slow or accelerate the policy
process. This study utilizes both public policy and public administration theory by applying a public
administration approach to studying budgetary change. Leveraging this approach, this work analyzes
the pattern and explanations of budgetary changes. Centering on two concepts understudied in PET
literature—policy feedback and endogenous organizational change—data from hundreds of
organizations are used to demonstrate how organization performance and personnel instability
contribute to budgetary changes for core organization activity. Results indicate that high levels of
performance and low levels of personnel instability lead to incremental changes.
KEY WORDS: punctuated equilibrium theory, public policy process, public administration, public budgeting, organization performance, turnover
The ability for organizations to hold steady policies while being able to adapt to
changing needs takes a delicate balance. The policy process, as such, is both rapid
and slow. As one theory of the policy process, punctuated equilibrium theory (PET)
captures both of these dynamics by examining incremental and large, punctuated
changes. Scholars in this literature have had much success in finding the pattern of
change predicted by PET in a variety of policy settings.
Understanding the factors that contribute to the rate of policy change has been
one of the core objectives in this literature. In broad terms, scholars have identified
disproportionate information processing, institutional friction, and organizational
history as reasons why policy subsystems experience more or less punctuated
changes. Empirically, PET has been predominantly analyzed through examinations
of distributions of policy changes, although there have been works that utilize multi-
variate analyses (Robinson, Flink, & King, 2014; Robinson, Caver, Meier, & O’Toole,
0190-292X VC 2015 Policy Studies Organization
Published by Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington Road, Oxford, OX4 2DQ.
The Policy Studies Journal, Vol. 45, No. 1, 2017
2007). Traditionally, these studies use public budgeting data—federal, state, local,
and school district levels.
Throughout PET literature, much effort has been devoted to explaining how the
structure of institutions and organizations shapes policy outcomes. Features such as
centralization, veto players, organization size, and bicameralism have all been shown
to affect the rate of policy change by applying varying amounts of “friction” to the
policymaking process. Moving away from structural considerations, other factors
can slow or accelerate the policy process, although little scholarly work has been
devoted to explaining these other features that influence the rate of policy change.
Why do two organizations of the same structure experience different policy changes
over time? Part of the answer lies in looking to other literatures that can explain fac-
tors that influence organizational decision making.
The purpose and main theoretical contribution of this article is to extend explan-
ations of budgetary change in PET by applying theories from public administra-
tion—a literature devoted to explaining how bureaucrats, clientele, management,
and the environment influence performance and efficiency in organizations. Public
administration literature brings a new perspective on what causes friction in organi-
zations, and what leads to policy changes. In turn, this study makes theoretical con-
tributions to both public policy and public administration literatures by not only
examining the pattern of budgetary changes but also developing explanations for those
Two concepts understudied in PET literature—policy feedback and endogenous
organization change—are examined in this study. Policy feedback is measured by
organization performance. It indicates how well a policy is working for an organiza-
tion, hence, providing policy feedback. Endogenous organization change is how
alterations within the organization, not strictly related to policy changes, affect policy
dynamics. In this work, endogenous organization change is measured as personnel
instability (also referred to as employee turnover). Success of policies and internal
organization changes are concerns for all organizations. Understanding how both
concepts influence budgetary changes has benefits for academics and practitioners
as the public sector works within an increasingly strained resource environment.
Justification for applying public administration theories is grounded in two cri-
tiques of the conceptualization of institutional friction. For one, friction is conceptual-
ized as factors of the policy process or structure of decision making. There are
organizational features outside of the traditional idea of “structure” that can influ-
ence budgetary changes. Second, the structures of policymaking, studied as institu-
tional friction, stay relatively steady over time (Jones & Baumgartner, 2005). Policy
decisions, though, are not based solely on how institutional and organizational struc-
ture influences decision costs, but on the current environmental demands. Organiza-
tions and institutions—regardless of their design—make decisions based on these
current demands that are in constant fluctuation.
This article theorizes and tests policy feedback and endogenous organization
change as catalysts for budgetary change by using a dataset that contains budgetary,
personnel, and performance data for hundreds of school districts in Texas for an
almost 20-year period. Organizational performance and personnel instability are
102 Policy Studies Journal, 45:1
examined to assess how each influences changes in core program funding. The find-
ings demonstrate that as performance increases, there is a significant rise in the pro-
portion of incremental budgetary changes. For personnel instability, as turnover
decreases, incremental budgetary changes become more prevalent. These results are
significant while controlling for varieties of institutional friction.
In addition to theoretical advances, results also provoke discussion about two
conventional measurement choices in the literature. For one, many scholars combine
punctuated and medium size budgetary changes into one “nonincremental” cate-
gory. The findings in this study show that medium size changes fluctuate depending
on the degree of friction, while punctuated changes show little movement away
from near zero probability of occurring. More examination should be given to
medium size changes, given their frequency over punctuated changes.
Second, scholars typically combine positive and negative changes of the same
absolute size in their measurement of budgetary changes. Findings in this study indi-
cate that positive and negative changes, even of the same absolute size, are utilized
at different times over the ranges of friction. This suggests researchers should work
toward theorizing and measuring budgetary changes based on the direction of
In all, this study causes us to rethink PET in terms of the theoretical causes of
policy change, how changes are measured, and the empirical methods applied to
test PET. This study ends with suggestions for future scholarship in PET and public
2.1. Models of Policy Change
Incrementalism is part of the foundation of policy change studies. This is espe-
cially true in the field of public budgeting. The field has dedicated decades to
explaining how incrementalism applies—or does not apply—to public budgeting
(Davis, Dempster, & Wildavsky, 1966; Wanat, 1974; Wildavsky, 1964). In analysis of
budgets from every level of government, country, agency, or organization, the same
general incremental pattern is present: there are mostly small changes from year to
year, supporting incrementalism (Cornia & Usher, 1981; Davis, Dempster, & Wildav-
sky, 1974). Incrementalism, though, is not readily accepted by all budgeting scholars.
For one, the term “incrementalism” has become conceptually diffuse. Scholars have
used the term to describe a method of problem solving, a process of interaction, a
theory of organizational behavior, a theory of policy development, a shift in organi-
zational relationships, and the size of monetary change (Berry, 1990; Dempster &
Wildavsky, 1979). With this many meanings, some scholars have questioned if incre-
mentalism is still a useful term for scholarly works (Berry, 1990). However, despite
critiques noting the theoretical and empirical shortcomings of incrementalism (i.e.,
Bailey & O’Connor ; Berry ; Dempster & Wildavsky ; Natchez &
Bupp ; Tucker ), the theory could not be wholly rejected or replaced with
all new theories. Hence, it still remains relevant to budgetary studies today.
Flink: Rethinking Punctuated Equilibrium Theory 103
PET, borrowed from geological studies, emerged as another theory of the policy
process that embraced incrementalism and incorporated the expectation for large
changes—a missing element of the incremental theory (Baumgartner & Jones, 2010).
In this theory, Baumgartner and Jones (2010) relate the policy process to phenomena
from the physical sciences like earthquakes and landslides. For example, earthquakes
occur as a result of slowly building pressure from underneath the earth’s surface
that causes violent shifts of the earth’s tectonic plates. The dramatic shift of the
earth’s plates causes earthquakes.
Keeping with the bigger picture, these are slow-moving processes that eventu-
ally lead to dramatic events. Policy processes work in much the same way. Policies
typically experience modest changes. Over time though, pressure builds within the
policy subsystem until enough pressure has accumulated that a large and dramatic
policy change results. In PET, these large changes are known as punctuations. Fea-
tures of the policy process can slow or accelerate the rate of policy change.
This theory has been supported in many contexts from incarceration rates
(Schneider, 2006) to election results (Baumgartner et al., 2009), legislative actions (i.e.,
bill introductions and hearings) (Baumgartner et al., 2009), environmental policy
(Busenberg, 2004; Repetto, 2006; Salka, 2004; Wood, 2006), and education (McLen-
don, 2003). The dominant testing ground, though, has been in the field of public
budgeting. Local, state, comparative, and U.S. federal government agencies and pub-
lic organizations have all exhibited characteristics consistent with PET (Baumgartner,
Foucault, & Francois, 2006; Baumgartner et al., 2009; Breunig, 2006; Breunig & Koski,
2006; John & Margetts, 2003; Jones, Baumgartner, & True, 1998; Jones et al., 2009; Jor-
dan, 2003; Mortensen, 2005; Robinson, 2004; Robinson et al., 2007).
Determining a series’ conformity with PET has relied on assessing the shape of
the distribution of annual percentage budgetary changes. The distribution is ana-
lyzed for how closely it follows a normal distribution. Specifically, the degree of kur-
tosis (a measure of central “peakedness”) is examined. Distributions that support
PET have high values of kurtosis and are known as leptokurtic distributions—distri-
butions with significantly more central observations around the mean and in the tails
of the distribution than a normal distribution. Theoretically, this is what PET predicts
of the policy change process: mostly incremental changes with numerous sizable
changes. This leptokurtic distribution, also known as a power function, is central to
punctuated equilibrium studies. It has proven extremely robust and is considered a
general empirical law (Jones et al., 2009).
Coupled with its theoretical growth, punctuated equilibrium literature is increas-
ing in empirical complexity. Scholars are expanding their work beyond univariate
analyses of distributions to multivariate hypothesis testing that can account for other
explanations of budgetary changes. The few published works that use multivariate
empirical tests predicting budgetary outcomes have divided the distribution of
budgetary changes into categories based on their size to use either logit (Robinson
et al., 2014) or multinomial logit (Robinson et al., 2007).
This is an important direction for the literature as it advances understanding of
how various features of governments and organizations influence budgetary
changes. Traditional univariate analyses in PET have focused on the shape of the
104 Policy Studies Journal, 45:1
distribution of budgetary changes and have limited their ability to account for other
variables. This has led to a literature that has probed deeper into describing a pattern
of budgetary changes, instead of engaging in theoretical development of what trig-
gers budgetary changes. By multivariate hypothesis testing methods, scholars can
progress to understanding the causes of different sizes of budgetary changes.
2.2. Causes of Punctuations
The literature on PET identifies two broad reasons for punctuations in policy
changes: disproportionate information processing and institutional friction.
Disproportionate information processing is an artifact of the direction of policy
attention. As the name suggests, this explanation of PET attributes policy changes to
the tendency of policymakers and policymaking institutions/organizations to react
disproportionately to new information (Jones, 2001). This is in contrast to proportion-
ate information processing (Jones & Baumgartner, 2005) in which policymakers form
policy decisions proportional to information within the environment. Officials,
though, cannot adequately process all information since there is only a limited
amount of policy attention they can give (Jones & Baumgartner, 2005). As a result,
policy subsystems commonly go through periods of underresponding or ignoring
information to overreacting to it (Jones & Baumgartner, 2005; Workman, Jones, &
Jochim, 2009). The over- and under-reactions contribute to the leptokurtic shape of
the distribution of budgetary changes found broadly in policy subsystems.
Institutional friction, the other explanation for policy punctuations, occurs as a
result of the institutional or organizational barriers or decision clearance points in
the policymaking process. Friction is a term used to account for the difficulty in the
process of making policy changes. The more hurdles there are in the process, the
more friction there is within the policy subsystem. This has consequences for policy
alterations. While institutional and organizational designs with multiple actors and
decision-clearances promote checks and balances (which provide comfort to citi-
zens), they slow down the policy process. This, in turn, builds pressure within the
policy subsystem. Over time, the accumulation of pressure will yield a punctuation.
There have been many different ways in which friction has been measured. The
measure is meant to describe the concentration of power or the barriers in deciding
policy changes. Institutional friction has been measured by bureaucratization (broken
down to centralization and organization size), stage in the policy cycle, political sys-
tem (presidential or parliamentary), executive dominance, single-party governments,
bicameralism, partisan control of government, partisan distance of governing parties,
and decentralization (Baumgartner et al., 2009; Breunig, 2006; Jones et al., 2009;
Robinson, 2004; Robinson et al., 2007). Empirically, each of these factors affects the
degree of kurtosis in the distribution of budgetary changes.
While the literature on PET establishes an overall pattern of change identified
through distributions, examining distributions of budgetary change does not allow
scholars to see when punctuations occur one time point to the next—the time series
aspect of budgetary change is lost when all years are combined to form a single
Flink: Rethinking Punctuated Equilibrium Theory 105
distribution. To examine when punctuations occur while preserving them in a time
series format, Robinson et al. (2014) test how the history of punctuations influence
the probability for future punctuations. The authors develop two theoretical models
of the effect of organizational history on punctuations. In the Error Accumulation
model, the probability of an organization experiencing a punctuation is negatively
related to having one in the recent past. In this model, punctuations occur to correct
the policy subsystem to the desired level of policy. Once this correction has been
made, policies will only see incremental changes until the distance between the
actual and desired level of policy reach a critical threshold.
The other model is the “Institutional Model” of policy change that states large
policy changes occur from characteristics within the organization (like poor design
or mismanagement). Since the propensity of punctuations is tied to the organiza-
tional design, the probability of having a punctuated change is positively related to
having one in the recent past. The authors’ findings support this Institutional
Model—punctuations occur in clusters. In other words, a history of punctuations
yields a higher probability that organizations will have a punctuation in the future.
3. Untested Sources of Friction
Literature has demonstrated that institutional friction influences budgetary
changes. There are, however, common characteristics of the measures of institutional
friction that leave open many questions about other sources of friction within policy
subsystems. For one, measures have been conceptualized as the policy process.1
Indeed, this was one of the original goals of this research agenda—examining the
policy consequences of structures of policy subsystems. Prior to PET, little was
known about how organizational or institutional structure shaped policy outcomes.
Early explanations centered on factors aside from the process. For example, the polit-
ical, economic, social, and administrative environments were said to influence policy
changes (Davis et al., 1974).
Punctuated equilibrium literature, however, has progressed in explaining the
process and structural factors as well as theorizing over cognitive limitations, leaving
other factors virtually unexplored. While PET is a theory of government information
processing (Workman et al., 2009), there are other elements to organizations that can
cause friction among decision makers and in turn, affect policy changes. Public
administration literature can help bring more depth to the understandings of policy
change and provide an avenue to study concepts heavily discussed, yet not thor-
oughly tested in PET literature.
Public administration literature aims to explain how bureaucracy, management,
clientele, and various public and private organizations all interact to affect and
implement public policy. It aims to explain how well governments deliver goods
and services in light of new policies, varying management strategies, racial diversity,
conflicting goals, coordinating programs among many organizations and institutions,
and numerous other considerations. This literature has helped identify factors that
contribute to healthy and unhealthy organizations—ideas that can transfer to PET by
explaining ways in which organizations experience friction other than through
106 Policy Studies Journal, 45:1
organizational/institutional processes. Public administration literature adds a new
dimension to the PET reasoning for budgetary change by incorporating indicators of
the organization environment, uncertainty, personnel, clientele, and task difficulty as
proxies for friction within a policy subsystem. This is the main theoretical contribu-
tion of this study—understanding how these organizational features that give sub-
stance to the interactions among decision makers influence the rate of budgetary
Budgets, in short, are not simply artifacts of the policy processing structure of
organizations. This would give the impression that organizations are natural sys-
tems, with outcomes dependent on however the organization was originally
designed (leading to high or low friction). Organizations, however, are human sys-
tems—their outcomes are dependent on more than structural design. For example,
organizations of the same structure can experience different levels of budgetary
change as a result of varying amounts of friction brought about by environmental
demands or other issues surrounding the organization. Institutional friction captures
how efficiently policies are processed, but it does not give an idea to the problems
the organization is trying to address. An understanding needs to be gained about the
contextual elements that can show stress within organizations and signal a greater
need for budgetary change.
A second common feature of the institutional friction policy process measures is
their relative stability over time (Jones & Baumgartner, 2005). The fixed measures
allow for comparisons of the institutional/organizational arrangements over time—
an essential component to understanding the policy change process. However, many
aspects of organizations fluctuate over time. Instability and uncertainty within organ-
izations create friction that can jolt an organization. This lead to many theoretically
interesting questions related to how policy stability persists in light of factors that
change frequently within the organization.
These two points—structure and stability—have inhibited the study of many
organization features that can influence policy dynamics. This study addresses both
of the above critiques by examining two organizational features that are outside of
the decision-making structure and that fluctuate over time: policy feedback as organi-
zational performance and endogenous organizational change as personnel instability.
Policy feedback2 is conceptualized as the success or failure of a given policy deter-
mined by organizational performance. The feedback received by policymakers pro-
vides information on the type of adjustment—minor or major—that is needed for the
policy to be more effective for the organization. Obtaining reliable and timely feed-
back is a crucial element of policy development. Endogenous organizational change
examines how internal alterations within the organization can have broader implica-
tions for policy dynamics. The changing of personnel, managers, clients, operating
procedures, performance metrics, organization structure, and so forth, although all
internal organization issues, can have impacts on policy alterations.
Organization performance and personnel turnover are two important and salient
elements to virtually all organizations (Rainey, 2003) and are commonly studied
throughout public administration literature. They are constantly monitored and
taken into account for many organization decisions. If either of these features is less
Flink: Rethinking Punctuated Equilibrium Theory 107
than adequate, it can cause issues within the organization. Most likely, there will be
disagreement within the organization on the best way to improve the quality of out-
puts and employee retention. How each element can be a potential source of friction,
and in turn impact budgetary changes, is outlined below.
3.1. Policy Feedback: Organizational Performance
The performance of public organizations is scrutinized by citizens and public
officials. Even though they are generally characterized as underperforming (Moyni-
han, 2008; Rainey, 2003), the reason for the existence of public organizations is to pro-
vide quality goods and services to their clientele. Efforts to increase the efficiency
and effectiveness of public organizations have gained momentum. Elected officials
have developed extensive accountability systems to incentivize good and penalize
bad performance in public organizations. A weakness of these policies is their one-
size-fits-all approach. Numerous studies have demonstrated the unequal results of
these programs across organizations (Moynihan, 2008; Radin, 2006; Rainey, 2003;
Rutherford, 2014). Public organizations have unique missions, environments, and
multiple dimensions on which to measure performance—suggesting there is no one
way in which they can be motivated or evaluated.
Numerous performance initiatives implemented by government have made
attaining set standards a high-stakes endeavor. Performance is virtually the biggest
concern for any organization. Throughout academic work, this is reflected in man-
agement and organization theory’s central focus on explaining different aspects of
organization performance and effectiveness (Rainey, 2003). The literature assessing
the determinants of organizational performance is large and spans many diverse
fields. In practice, outputs are regularly monitored by public officials, managers,
employees, and service recipients. Based on what is observed, current and future tar-
get levels of performance shape organization work and direction. Over- and under-
performing organizations, however, must take different approaches to their work.
The general assumption is organizations achieve success through proper manage-
ment of their internal and external environment. When public organizations fall
below a set standard, governments must intervene to help manage the situation.
Organizations with sustained high performance have implemented successful
policies and properly managed their environment. Assuming no government inter-
ventions or other environmental shocks, changes to the organization are typically
modifications to existing routines. These organizations are more likely to benefit
from increased resources and support from government or the addition of new cli-
ents. Organizational goals then focus on maintaining current standards and possible
expansion of their work.
There are harsh consequences for underperforming public organizations. With
the push for greater accountability, these organizations are threatened by sanctions,
penalties, government interventions, and closure. Managers, employees, and clients
want to improve performance, but finding consensus on the best way to achieve that
end is difficult. The choice on what alterations to make within the organization is
108 Policy Studies Journal, 45:1
complicated by the many options available to decision makers. Resources, regula-
tion, markets, organization, and management all influence public service perform-
ance (Boyne, 2003). Isolating the parts of the organization that need to be changed
can cause conflict and friction. In light of this friction, policy changes are likely to be
large (a case can be made for either positive or negative punctuations, depending on
the situation), as a desperate attempt for performance improvement. Incremental
changes are not likely to provide the jolt needed within the organization to spark
This leads to the performance hypothesis:
Hypothesis 1: Low performance decreases the expected proportion of incremental
budgetary changes and increases the expected proportion of medium and punctu-
ated budgetary changes.
3.2. Endogenous Organizational Change: Personnel Instability
One of the core concerns of management is their organization’s workforce.
Human capital is one of the most valuable assets of any organization (O’Toole &
Meier, 2009; Rainey, 2003). Bringing in new talent, retaining workers, and enhancing
the skills of employees are essential for organizations to have quality performance.
Given the importance of human capital for organization success, personnel instabil-
ity is a relatively understudied area of public administration (Meier & Hicklin, 2008;
Raffel, 2007; Selden & Moynihan, 2000). Most works on turnover analyze its effect on
organizational performance. The leading theory is that personnel instability leads to
lower organization performance (Meier & Hicklin, 2008; O’Toole & Meier, 2003).
However, a refined version of the theory acknowledges benefits from turnover (like
the organization staying fresh, bringing in new ideas, and the dismissal of ineffective
workers) that suggest its nonlinear relationship with performance (Abelson & Bay-
singer, 1984; Meier & Hicklin, 2008; Mosher & Kingsley, 1936).
There remain many research questions on the consequences of employee turn-
over beyond performance. This study examines its effect on policy stability. As
stated before, PET focuses on procedures as sources of friction. However, even if
structures and procedures can stay constant, personnel turnover induces another
type of instability for the organization that changes the dynamics among actors
(Weber, 1946). High turnover can signal problems and dissatisfaction among
employees (Rainey, 2003). Additional stress occurs by replacing and retraining work-
ers—it can be a costly endeavor that takes a substantial amount of resources within
the organization (Griffeth & Hom, 2001; Wright & Kim, 2004). In organizations with
high turnover, there should be more friction overall, yielding more punctuated
changes. Policymakers may feel the need to enact major policy changes to retain
employees. In organizations with low turnover, the friction models suggest there is
less friction within the organization leading to less punctuation.
High turnover could also be a conscious effort by the organization to phase out
current employees for new workers. This can occur when an organization is redevel-
oping and redirecting its mission. Turnover, then, is not necessarily a voluntary
Flink: Rethinking Punctuated Equilibrium Theory 109
move by the employee to leave a poor working environment, but a planned effort by
management to bring in a new workforce and new direction for the organization. In
this case, large policy changes could be associated with friction from the evolution of
the organization stimulated by managerial decisions. Organizations that are not
experiencing a redefinition should have a more stable workforce, low friction, and
more incremental policy changes. This leads to the personnel instability hypothesis:
Hypothesis 2: High personnel instability decreases the expected proportion of incre-
mental budgetary changes and increases the expected proportion of medium and
punctuated budgetary changes.
4. Data and Methods
Data for this study come from Texas school districts. This large dataset provides
budgetary, performance, employee, and administrative information for school dis-
tricts within Texas. Given the districts’ similar policy environment, structure, and
goal of educating students, this dataset allows for a comparison of hundreds of simi-
lar organizations. In addition, it provides an adequate time span, 1993 to 2010, to
examine the dynamics of punctuated equilibrium by capturing the rare events of pol-
Similar to other studies in PET, the dependent variable in this study will be a
budgetary measure. School districts are expected to fund many diverse functions
from athletics to gifted and talented classes, to student transportation, to bilingual
education. Districts have numerous budget categories. With some limitations, school
districts are granted discretion in how they allocate their funds across programs.
Budgetary decisions are made in light of current environmental demands of the
organization. At its best, budgetary decisions are made with the intent of addressing
organization needs and allocating funds to programs that will yield the most benefits
(as viewed by decision makers, but this can be a source of friction).
For the analysis in this study, the dependent variable is the annual percentage
change in instructional spending per pupil. Instructional spending per student is one
of the core program funds for all school districts. It represents one type of educa-
tional strategy that allocates funds directly to educating students. Since this is one of
the most important functions of districts, managers will make budgetary decisions to
protect these funds from financial environmental turbulence. Meier and O’Toole
(2009) find that when the overall budget falls, instructional spending per student is
only cut a fraction compared to the overall budget change. Changes in this category
represent pointed decisions by managers—they do not solely reflect the availability
of funds for school districts from state and local sources.
Analyses follow methodology proposed in Robinson et al. (2007) by dividing the
dependent variable (annual change in instructional spending per student) into five
categories based on the size of change. The categories are: negative punctuations,
medium negative changes, incremental changes, medium positive changes, and posi-
tive punctuations. The division of categories is determined by laying a normal curve
over the histogram of budgetary changes. This leads to four intersections—two near
110 Policy Studies Journal, 45:1
the central part of the distribution and two near the tails of the distribution—that
serve as the threshold cut points between the five categories. Using this method cre-
ates a nonarbitrary way to determine incremental, medium, and punctuated
changes.3 The frequency of budgetary changes within the five categories is displayed
in Table 1.
The hypotheses cover each of these five categories (without specifying differen-
ces between positive and negative changes). Thus far, scholars have not put a lot of
effort into theorizing on the differences between medium and punctuated changes,
nor positive and negative changes. Generally, hypotheses explain incremental and
nonincremental changes and do not differentiate between positive and negative
changes. In this study, predictions are similar for each of the nonincremental catego-
ries, given that scholars have not demonstrated the uniqueness of these categories.
This study will examine if differences do exist.
As discussed above, the two explanatory variables to capture alternative
sources of friction are organizational performance and personnel instability. The
percentage of students in a district that passed the annual statewide standardized
test will be used to assess organizational performance. This is the customary mea-
sure of performance in public administration (studies that use educational data)
and education literatures. Personnel instability is measured by the percent of
teacher turnover within a school district. This is another common measure in the
education and public administration literatures (Meier & Hicklin, 2008; O’Toole
& Meier, 2003, 2009). Since performance assessment and turnover typically hap-
pen at the end of a school year, the lagged values of each of the measures are
used in the empirical model—it is more plausible the prior year influences the
current year budgetary changes.
Revisiting the two critiques of institutional friction conceptualizations (they char-
acterize only the policy process and stay relatively stable over time), both of these
variables hold up to these points. In regards to the process, both of these measures
do not refer to the structure of decision-making processes. Organizational perform-
ance should influence budgetary changes, but it is not an indicator of policy proc-
esses in the same way as conceptualizations of institutional friction. Personnel
instability is a little less clear on this issue. The measure actually assesses the stability
of teachers within a school district. In public administration terms, teachers are char-
acterized as street-level bureaucrats. While past measures of institutional friction
Table 1. Distribution of Budgetary Changes (Dependent Variable)
Change Category Frequency Percent
Negative punctuation 57 0.45 Medium, negative 1,964 15.42 Incremental 8,675 68.12 Medium, positive 1,963 15.42 Positive punctuation 75 0.59
N: 12,734. Kurtosis of distribution of budgetary changes (continuous vari- able): 41.84.
Flink: Rethinking Punctuated Equilibrium Theory 111
have considered bureaucratization, it was meant to indicate a concentration of deci-
sion making. In school districts, budgetary decisions are top-down process that typi-
cally do not involve input from all levels of the organization. Teachers do not have a
large amount of input in budgetary decision making. Thus, this measure of bureauc-
ratization is a measure of stability within an organization, not of centralization of
Fluctuations occur frequently in both of these variables, as well. For example, in
the present sample, the average annual percentage change is 2.99 for performance
and 13.45 for turnover. Also in this setting, organization size and centralization have
been examined as measures of institutional friction for other work. The average
annual percentage change for organization size is 0.75 and 0.54 for centralization.
Within this sample, there are greater alterations for the two new measures of friction
introduced in this study.
Control variables are added to the model to account for institutional friction.
Measures for centralization (percent of school district’s budget dedicated to central
bureaucracy) and centralization squared are included in the model. The squared
term is necessary to account for the rising and then declining impact of centralization
for organizations (Ryu, 2011). In other words, increasing centralization benefits
organizations only to a certain point. Organizational size (student enrollment) and
growth (percentage change in student enrollment) are also included as control varia-
bles. History of punctuations in the organization (experiencing a punctuation within
the previous five years) is accounted for in the models as well, given research by
Robinson et al. (2014) that finds a history of policy punctuations leads to a greater
probability of punctuations in the current time period (said differently, punctuations
occur in clusters). These control variables are also common to PET articles that use
this dataset (Robinson, 2004; Robinson et al., 2007, 2014). Table 2 displays the
descriptive statistics for all explanatory variables used in analyses.
This study adds to the literature that examines PET through multivariate statisti-
cal analyses (Robinson et al., 2007, 2014). The dependent variable (instructional
spending per student split into five categories) is designed to use multinomial logit
as the method of analysis. Multinomial logit is used when the dependent variable
consists of categories that are unordered and discrete.4 The method calculates the
probability of explanatory variables being in one category compared to a baseline
category. In this analysis, the baseline category is incremental changes.
The results of the multinomial logit model are shown in Table 3.5
5.1. Policy Feedback: Organizational Performance Results
Organization performance is negative and statistically significant across all cate-
gories of budgetary change in the multinomial logit model. This means that as test
performance improves in a school district, it is significantly less likely that
112 Policy Studies Journal, 45:1
organizations will experience nonincremental (medium or punctuated) budgetary
changes. In terms of friction, when a district is performing at a high level, there
seems to be less friction among decision makers that is resulting in major policy
changes. For policy feedback, this means that when the feedback is good, there are
mostly small changes. When feedback is bad, however, there is a tendency for more
dramatic policy shifts.
To help illustrate the effect of organizational performance, Figure 1 shows the
predicted probability of experiencing each of the five categories of budgetary change
over the range of pass rates for the exam. In this set of predicted probabilities, all
other variables are set to their mean values. Incremental changes see dramatic growth
Table 2. Descriptive Statistics of Explanatory Variables
Variable Mean Std. Dev. Minimum Maximum
Organizational performance (standardized test pass rate) 72.42 15.70 6 100 Personnel instability (teacher turnover) 17.47 9.51 0 100 Centralization 7.39 3.71 1 73.30 Centralization squared 68.31 125.73 1 5,372.89 Organization size (logged) 6.95 1.52 1.95 12.26 Organizational growth 0.75 7.13 263.30 117.33
Organizational history (Dummy Variable) No Punctuation: 0
11,911 Punctuation: 1
N 5 12,734.
Table 3. The Effects of Friction on the Relative Probability of Experiencing Large and Medium Versus Incremental Budgetary Changes
Size of Change Negative
Punctuation Medium, Negative
Policy Feedback Organizational performance (standardized test pass rate lagged)
20.024* 20.007* 20.034* 20.044* 23.25 23.79 220.42 26.99
Endogenous Organization Change Personnel instability (teacher turnover lagged) 0.033* 0.022* 20.002 0.013
3.91 7.86 20.79 1.56 Institutional Friction
Centralization 20.083 20.223* 20.164* 20.202* 21.29 211.44 28.39 23.46
Centralization squared 0.003** 0.005* 0.004* 0.004* 2.18 7.37 6.25 3.41
Organization size (logged) 20.725* 20.470* 20.482* 21.090* 24.99 217.80 217.95 28.43
Organizational history 2.155* 0.631* 0.191 1.247* 6.57 6.45 1.79 4.35
Organizational growth 0.014 0.032* 20.008** 0.011 1.47 9.24 22.14 1.16
Z-score below each coefficient. N 5 12,734. LR test: 1,730.76, p< 0.00. BIC: 21,301.385. Pseudo R2: 0.08. PCP: 68.75 %; PMP: 68.12 %; PRE: 1.95 %. *= p< 0.05 **= p< 0.01
Flink: Rethinking Punctuated Equilibrium Theory 113
as organization performance improves. This is consistent with Hypothesis 1 that
states that low-performing organizations will have a decrease in incremental changes.
The two categories of medium size changes offer mixed support for the hypothe-
sis (low performance increases the expected proportion of medium size budgetary
changes). For medium, positive changes the hypothesis has support—there is a sig-
nificant decrease of positive budgetary changes as performance improves. In con-
trast, medium, negative changes do not see much of a change over the spectrum of
These results are surprising and theoretically interesting considering previous
research. In analyses of PET, positive and negative changes are usually combined
into one variable representing the magnitude of the size of the budgetary change
(medium or punctuated). These studies typically contain a sentence or two in the
conclusion stating scholars should begin to theorize about the differences in positive
and negative budgetary changes (Robinson et al., 2014). The results presented here
say, yes, scholars do need to move in that direction. In Figure 1, positive and nega-
tive budgetary changes have differing slopes—one is significantly changing across
the spectrum of friction, the other is not. The findings suggest the potential for new
theoretical developments on the direction of policy change, not just the magnitude.
For punctuated changes, there is little evidence to support the hypothesis. Punc-
tuated changes, both positive and negative, are very small in their predicted proba-
bilities. Across the range of organization performance, both categories of punctuated
change remain close to zero and behave in similar manners (unlike medium
changes). Given the rarity of punctuations, it seems there should be less emphasis in
the literature in explaining their occurrence—the interesting trade-off occurs between
incremental and medium size changes.
The stark difference in the use between punctuated and medium changes has
been masked by previous works that combine both of these categories into one
“nonincremental” category. By dividing nonincremental changes between medium
and punctuated changes, a new understanding of PET is gained by seeing how little
Figure 1. Predicted Probabilities of Categories of Budgetary Changes over Organizational Perform- ance (District Student Pass Rate of Exam).
114 Policy Studies Journal, 45:1
punctuations are predicted to occur versus medium changes. Medium changes
appear to be a more popular way for the release of friction in a policy subsystem.
Scholars should make an effort to theorize on the occurrence of medium changes.
5.2. Endogenous Organizational Change: Personnel Instability Results
Personnel instability again gives mixed support for Hypothesis 2. From Table 3,
personnel instability is positive and statistically significant for the negative categories
of changes and insignificant for the positive changes. In other words, as turnover
increases, organizations are more likely to experience negative medium and negative
punctuated budgetary changes than incremental changes. There is no statistically
significant finding between incremental and positive medium or punctuated
changes. These findings demonstrate that endogenous organization change inconsis-
tently stimulates policy changes.
Similar to Figure 1, Figure 2 displays the predicted probabilities (with all other
variables held at their means) for each of the five categories of budgetary changes
over the range of turnover. From the graph, a statistically significant decrease in
incremental budgetary changes is present as turnover increases. This declining trend
supports Hypothesis 2. In relation to friction, low employee turnover signals less
stress within the organization leading to more incremental budgetary changes.
The findings for medium size budgetary changes again give mixed results.
Hypothesis 2 stated that high personnel instability increases the expected proportion
of medium size budgetary changes. In support of the hypothesis, negative medium
size changes increase as personnel instability increases. In thinking about the specific
variables, this makes sense. As more teachers are leaving the district, money is
slowly decreasing from instructional expenditures. As the classrooms become unsta-
ble, schools are pulling funds from that function.
Against the hypothesis, medium positive budgetary changes see a slight
decrease as turnover increases. The insignificance of this finding is very interesting.
Figure 2. Predicted Probabilities of Categories of Budgetary Changes over Personnel Instability (Per- cent Teacher Turnover).
Flink: Rethinking Punctuated Equilibrium Theory 115
One might expect to see larger, positive budgetary changes as turnover increases as
an effort to retain employees or as managerial effort to redevelop the organization
around new employees. These results do not support this idea—when turnover
increases it seems money is more likely to be taken away, leading to a more tense
work environment opposed to more money being spent to create a better work envi-
ronment and encourage employee retention. Like the results of performance, these
findings also support separately theorizing and empirically testing negative and pos-
Hypothesis 2 also predicts that with increased turnover comes increased friction,
thus leading to an increase in punctuations. However, there is little support for either
hypothesis. Again, positive and negative punctuations hover around the zero mark
and do not move much across the range of personnel instability. Like Figure 1, it
suggests more of the action takes place between incremental and medium size
changes, rather than punctuations. Only a small part of the budgetary change picture
is described by punctuations.
Lastly, the control variables for institutional friction are examined to check for
consistency with the literature. Every variable is significant in the expected direction
except centralization—it is statistically significant in the negative direction for three
of the categories. This is a contradiction to past studies (Robinson, 2004; Robinson
et al., 2007) that have indicated an increase in centralization yields greater probabil-
ities of nonincremental changes. However, this model contains a variable for central-
ization squared that is positive and statistically significant. This suggests a nonlinear
U-shaped relationship between centralization and budgetary changes.
The findings in this study present new theoretical insights to the PET literature.
Returning to the roots of PET, the theoretical reasons for policy changes were
extended by examining policy feedback and endogenous organization change. While
traditional conceptualizations of institutional friction measure the process of decision
making, the expanded understanding of friction presented in this study captures the
features of organizations that fluctuate, stimulate friction within the policy subsys-
tem, and give substance to the interactions of those in the organization that influence
Drawing from public administration literature, organizational performance and
personnel stability were examined as policy feedback and endogenous organization
change, respectively. These measures serve as a way to capture alternative sources of
friction in a policy subsystem. Using data from Texas school districts, it was found
that low district pass rates on the statewide standardized test as well as high teacher
turnover led to a decrease in incremental budgetary changes for instructional spend-
ing (a core expenditure of school districts). These findings give scholars a new
understanding of the causes of policy change.
Figures 1 and 2 brought new insight to separating positive and negative budget-
ary changes (something not standard to this literature). The figures show that the
predicted probabilities for positive and negative medium changes have unique
116 Policy Studies Journal, 45:1
slopes, meaning they are being utilized in different circumstances for organizations.
This study represents a first step into understanding and theorizing beyond the mag-
nitude of change, but to the direction of change. Increases and decreases in policy
outcomes can vastly impact society. An organization experiencing a medium size
budgetary change needs to know if it is a positive or negative change to take appro-
priate courses of actions. Finding out the unique stimuli for positive and negative
changes not only gives a better understanding of the policy process but also provides
valuable information to those within organizations.
The predicted probability graphs demonstrated little support for the punctuation
hypotheses, both positive and negative. In fact, the predicted probabilities for punc-
tuations remained close to zero across each of the figures. The graphs show greater
probability and fluctuations among the use of medium size changes. Perhaps schol-
ars should devote more work to theorizing on medium size changes, instead of
punctuations. This can still relay important information on how organizations and
institutions handle friction within a subsystem.
The bridging of public administration theory and PET gives ideas for future
research of other dimensions of friction or factors that can influence policy changes.
The management strategies—prospector, defender, or reactor (Miles, Snow, Meyer,
& Coleman, 1978)—used in organizations could influence policy changes. In simplest
of expectations, we would expect managers that predominantly use a defender strat-
egy to have a propensity toward incremental changes. Prospectors and reactors
would be less apt to consistently engage in incremental policy changes.
Drawing from race and ethnic politics, the racial composition of clientele and
managers could also create tension and friction in an organization. Diversity in
organizations (racial diversity in this example, but diversity can take a variety of
forms) can lead to conflicting needs or views on the direction of the organization.
Future research could also examine and compare many dependent variables of pol-
icy change (not just core functions) to see how they may fluctuate differently. What
type of friction triggers change in certain policy areas? There may be unique mecha-
nisms of policy change for minority interests versus majority interests.
Future work should also examine growth hypotheses—how the changes in per-
formance (and other factors) affect policy changes. The analyses in this study con-
sider only the absolute level of performance and level of turnover. To illustrate, in
the context of the standardized test pass rate for school districts, individual organiza-
tions have unique target levels for performance. For some schools, 75 percent student
test pass rate is acceptable. For another school, 75 percent is extremely low. Perhaps
a steady decline or growth in performance, whatever the absolute level, affects policy
Other research could center on how these unique types of friction interact to
affect policy dynamics. The strictly institutional/structural types of friction could be
conditioned by other types of friction found in performance or in organization per-
sonnel turnover. There are many possibilities in this avenue of research.
This study contributes to our understanding of policy change, but more broadly,
adds to our knowledge of how decision makers address organizational issues
through the budgetary process. Policy feedback and endogenous organization
Flink: Rethinking Punctuated Equilibrium Theory 117
change influenced the magnitude of budgetary changes, even while controlling for
structural forms of institutional friction and other organizational features. Although
there are still many questions left for future research, this study represents another
step to a better understanding of the policy process.
Carla M. Flink is an assistant professor in the Department of Public Administra-
tion at The University of Texas at San Antonio.
The author would like to thank Kenneth J. Meier; Scott Robinson; Kim Quaile Hill; Guy D. Whitten; Amanda Rutherford; the TAMU Project for Equity, Representation, and Governance; and four anony- mous reviewers for their help on this project.
1. Measures inside the policy subsystem or process would address the rules and structure of deciding policy. Measures outside the policy subsystem or process would encompass features like the environ- ment external to the organization/institution, changes in personnel or clientele, organization perform- ance, and so forth.
2. Baumgartner and Jones (2002) use the term “policy feedback” to describe policy changes. The authors consider two types of policy feedback: positive and negative. Positive feedback occurs when one policy change leads to a greater change. Negative feedback occurs as policy processes lead to stability and incrementalism in policy dynamics. The definition of policy feedback used in this work is unique from those definitions and conceptualizations.
3. The exact cut point percentage values are 233, 22, 10, and 35.5. These values are the interior and exte- rior intersections between a normal distribution overlaid on the histogram of annual percentage budg- etary changes. See Robinson et al. (2007) for further explanation.
4. Even though there is an order to the categories (positive to negative), the hypotheses are based on mag- nitude of the change (incremental to punctuated). This makes it unclear how to order positive and neg- ative changes of the same magnitude. Because of this, there is no clear way to order the categories. Thus, multinomial logit is used instead of ordered logit.
5. Diagnostic tests did not reveal multicollinearity among variables. The present study does violate the IIA assumption made for multinomial logit models. Multinomial probit produced results simi- lar to the logit models. The model was estimated with robust standard errors, district clustered standard errors, and year fixed effects, but the results were always similar. The standardized test switched from TAAS to TAKS in 2003. The results are still robust with the exclusion of that year. There is concern over the relationship between performance and personnel stability. While they are correlated at 20.3824, Wald and likelihood ratio tests indicate each variable adds significantly to the model and should be included. Total student enrollment serves as a proxy for total revenue and total expenditures in a district (enrollment is highly correlated with revenue and expenditures at 0.97 and 0.95, respectively). Inclusion of total revenue and expenditure variables in the model does not significantly alter the results.
Abelson, Michael A., and Barry D. Baysinger. 1984. “Optimal and Dysfunctional Turnover: Toward an Organizational Level Model.” Academy of Management Review 9 (2): 331–41.
Bailey, John J., and Robert J. O’Connor. 1975. “Operationalizing Incrementalism: Measuring the Muddles.” Public Administration Review 35 (1): 60–6.
Baumgartner, Frank R., and Bryan D. Jones. 2002. “Positive and Negative Feedback in Politics.” In Policy Dynamics. Chicago: University of Chicago Press.
———. 2010. Agendas and Instability in American Politics. Chicago: University of Chicago Press.
118 Policy Studies Journal, 45:1
Baumgartner, Frank R., Christian Breunig, Christoffer Green-Pedersen, Bryan D. Jones, Peter B. Mortensen, Michiel Nuytemans, and Stefaan Walgrave. 2009. “Punctuated Equilibrium in Compara- tive Perspective.” American Journal of Political Science 53 (3): 603–20.
Baumgartner, Frank R., Martial Foucault, and Abel Francois. 2006. “Punctuated Equilibrium in French Budgeting Processes.” Journal of European Public Policy 13 (7): 1086–103.
Berry, William D. 1990. “The Confusing Case of Budgetary Incrementalism: Too Many Meanings for a Single Concept.” Journal of Politics 52 (1): 167–96.
Boyne, George A. 2003. “Sources of Public Service Improvement: A Critical Review and Research Agenda.” Journal of Public Administration Research and Theory 13 (3): 367–94.
Breunig, Christian. 2006. “The More Things Change, the More Things Stay the Same: A Comparative Analysis of Budget Punctuations.” Journal of European Public Policy 13 (7): 1069–85.
Breunig, Christian, and Chris Koski. 2006. “Punctuated Equilibria and Budgets in the American States.” Policy Studies Journal 34 (3): 363–79.
Busenberg, George. 2004. “Wildfire Management in the United States: The Evolution of a Policy Failure.” Review of Policy Research 21 (2): 145–56.
Cornia, Gary C., and Charles L. Usher. 1981. “The Institutionalization of Incrementalism in Municipal Budgeting.” Southern Review of Public Administration 5 (1): 73–90.
Davis, Otto A., M. A. H. Dempster, and Aaron Wildavsky. 1966. “A Theory of the Budgetary Process.” The American Political Science Review 60 (3): 529–47.
———. 1974. “Towards a Predictive Theory of Government Expenditure: US Domestic Appropriations.” British Journal of Political Science 4 (4): 419–52.
Dempster, M. A. H., and Aaron Wildavsky. 1979. “On Change: Or, There is No Magic Size for an
Increment.” Political Studies 27 (3): 371–89.
Griffeth, Rodger W., and Peter W. Hom. 2001. Retaining Valued Employees. Thousand Oaks, CA: Sage.
John, Peter, and Helen Margetts. 2003. “Policy Punctuations in the UK: Fluctuations and Equilibria in
Central Government Expenditure since 1951.” Public Administration 81 (3): 411–32.
Jones, Bryan D. 2001. Politics and the Architecture of Choice: Bounded Rationality and Governance. Chicago:
University of Chicago Press.
Jones, Bryan D., and Frank R. Baumgartner. 2005. The Politics of Attention: How Government Prioritizes Prob- lems. Chicago: University of Chicago Press.
Jones, Bryan D., Frank R. Baumgartner, Christian Breunig, Christopher Wlezien, Stuart Soroka, Martial Foucault, Abel Francois, et al. 2009. “A General Empirical Law of Public Budgets: A Comparative Analysis.” American Journal of Political Science 53 (4): 855–73.
Jones, Bryan D., Frank R. Baumgartner, and James L. True. 1998. “Policy Punctuations: US Budget Authority 1947-1995.” Journal of Politics 60: 1–33.
Jordan, Meagan M. 2003. “Punctuations and Agendas: A New Look at Local Government Budget Expenditures.” Journal of Policy Analysis and Management 22 (3): 345–60.
McLendon, Michael K. 2003. “The Politics of Higher Education: Toward an Expanded Research Agenda.” Educational Policy 17 (1): 165–91.
Meier, Kenneth J., and Alisa Hicklin. 2008. “Employee Turnover and Organizational Performance: Test- ing a Hypothesis from Classical Public Administration.” Journal of Public Administration Research and Theory 18 (4): 573–90.
Meier, Kenneth J., and Laurence J. O’Toole. 2009. “The Dog that Didn’t Bark: How Public Managers Han- dle Environmental Shocks.” Public Administration 87 (3): 485–502.
Miles, Raymond E., Charles C. Snow, Alan D. Meyer, and Henry J. Coleman. 1978. “Organizational Strat- egy, Structure, and Process.” Academy of Management Review 3 (3): 546–62.
Mortensen, Peter B. 2005. “Policy Punctuations in Danish Local Budgeting.” Public Administration 83 (4): 931–50.
Mosher, William Eugene, and J. Donald Kingsley. 1936. Public Personnel Administration. New York: Harper and Brothers.
Flink: Rethinking Punctuated Equilibrium Theory 119
Moynihan, Donald P. 2008. The Dynamics of Performance Management: Constructing Information and Reform. Washington, DC: Georgetown University Press.
Natchez, Peter B., and Irvin C. Bupp. 1973. “Policy and Priority in the Budgetary Process.” The American Political Science Review 67 (3): 951–63.
O’Toole, Laurence J., and Kenneth J. Meier. 2003. “Plus ça Change: Public Management, Personnel Stability, and Organizational Performance.” Journal of Public Administration Research and Theory 13 (1): 43–64.
———. 2009. “The Human Side of Public Organizations Contributions to Organizational Performance.” The American Review of Public Administration 39 (5): 499–518.
Radin, Beryl. 2006. Challenging the Performance Movement: Accountability, Complexity, and Democratic Values. Washington, DC: Georgetown University Press.
Raffel, Jeffrey A. 2007. “Why Has Public Administration Ignored Public Education, and Does it Matter?” Public Administration Review 67 (1): 135–51.
Rainey, Hal G. 2003. Understanding and Managing Public Organizations. San Francisco: Jossey-Bass.
Repetto, Robert C. 2006. Punctuated Equilibrium and the Dynamics of US Environmental Policy. New Haven, CT: Yale University Press.
Robinson, Scott, Carla M. Flink, and Chad M. King. 2014. “Organizational History and Budgetary Punc- tuation.” Journal of Public Administration Research and Theory 24 (2): 459–71.
Robinson, Scott E. 2004. “Punctuated Equilibrium, Bureaucratization, and Budgetary Changes in Schools.” Policy Studies Journal 32 (1): 25–39.
Robinson, Scott E., Floun’say Caver, Kenneth J. Meier, and Laurence J. O’Toole. 2007. “Explaining Policy Punctuations: Bureaucratization and Budget Change.” American Journal of Political Science 51 (1): 140–50.
Rutherford, Amanda. 2014. “Organizational Turnaround and Educational Performance: The Impact of Performance-Based Monitoring Analysis Systems.” American Review of Public Administration 44 (4): 440–58.
Ryu, Jay Eungha. 2011. Bounded Bureacracy and the Budgetary Process in the United States. New Brunswick, NJ: Transaction Publishers.
Salka, William M. 2004. “Mission Evolution: The United States Forest Service’s Response to Crisis.” Review of Policy Research 21 (2): 221–32.
Schneider, Anne Larason. 2006. “Patterns of Change in the Use of Imprisonment in the American States: An Integration of Path Dependence, Punctuated Equilibrium and Policy Design Approaches.” Politi- cal Research Quarterly 59 (3): 457–70.
Selden, Sally Coleman, and Donald P. Moynihan. 2000. “A Model of Voluntary Turnover in State Gov- ernment.” Review of Public Personnel Administration 20 (2): 63–74.
Tucker, Harvey J. 1982. “Incremental Budgeting: Myth or Model?” The Western Political Quarterly 35 (3): 327–38.
Wanat, John. 1974. “Bases of Budgetary Incrementalism.” The American Political Science Review 68 (3): 1221–8.
Weber, Max. 1946. Essays in Sociology. HH Gerth and C. Wright Mills, Tr. New York: Oxford University Press.
Wildavsky, Aaron B. 1964. The Politics of the Budgetary Process. Boston: Little, Brown.
Wood, Robert S. 2006. “The Dynamics of Incrementalism: Subsystems, Politics, and Public Lands.” Policy Studies Journal 34 (1): 1–16.
Workman, Samuel, Bryan D. Jones, and Ashley E. Jochim. 2009. “Information Processing and Policy Dynamics.” Policy Studies Journal 37 (1): 75–92.
Wright, Bradley E., and Soonhee Kim. 2004. “Participation’s Influence on Job Satisfaction: The Importance of Job Characteristics.” Review of Public Personnel Administration 24 (1): 18–40.
120 Policy Studies Journal, 45:1