Retiring the Golden Hammer: Identifying Situational Practices for Public Strategy Implementation

While scholars and practitioners increasingly embrace contingent approaches to public strategic management, they have done so tepidly. In an increasingly perilous and turbulent governing environment, both groups must move past time-honored tools and concepts and embrace the complexity inherent to the strategy implementation process. In response, this article proposes a contingent, microorganizational process model of public strategy implementation based on Whittington’s (2017) framework of strategy as a practice and a process. Through regression analysis of 205 strategic initiatives from 43 U.S. municipalities, the study concludes that the relationships between implementation practices and proximate outcomes do indeed vary over time and across context, offering a specific list of recommended practices tailored to the intersections of implementation phase and initiative type. Public strategy implementation scholars can best aid practitioners by rejecting strategic reductivism and embracing micro-organizational implementation activity surrounding a strategic initiative, in all of its temporal and contextual splendor.

to tackle large-scale issues. In the literature, Boyne and Walker (2004) refer to these more enduring decisions to how an organization interacts with its external environment as "strategic stances" (p. 232). At the lower micro-organizational level, service-delivery program managers and specialists view strategy as initiatives or projects to "perform specific steps that an organization takes to operationalize its stance" (Boyne & Walker, 2004, p. 232). Likewise, Whittington (2017) conceptualizes strategy as both a micro-organizational practice and a macro-organizational process. In tandem, the macro-and micro-levels of the organization work symbiotically to formulate, implement, and evaluate strategy.
Despite this complexity, scholars and practitioners of public strategic management have largely opted to focus upon broader, more abstract notions of strategy. A recent review of public administration scholarship found that 93% of strategic management articles advance a solely macro-organizational definition of strategy (Mitchell, 2020). These studies largely focus on strategy formulation and evaluation, often reducing the implementation process to a single variable (George & Desmidt, 2014). Even in the instances when scholarly research concludes that managers should avoid a one-size-fits-all approach to strategy implementation, practitioners have not heeded this guidance, instead opting for their time-honored general tools and traditions (Mitchell, 2018;Nutt, 1995). This monolithic perspective ignores a rich body of micro-organizational implementation practices and activities associated with strategic initiatives George & Desmidt, 2014;Walker, 2013)-limiting the effectiveness of both scholars attempting to explain determinants of strategic success and practitioners striving to produce it.
Famed psychologist Abraham Maslow (1966) once characterized similar reductivism in his field as a 'golden hammer'; stating, "I suppose it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail." Also known as the 'law of the instrument', Maslow purports that all too often preference for a particular theoretical perspective or analytical approach drives research design, instead of the phenomenon under observation. Like a screw hammered into wood, using this golden hammer in scientific study may produce interesting short-term results, but ultimately distracts from a more effective and appropriate solution for the research question at hand.
Are public strategic management scholars and practitioners wielding a golden hammer? Through their overreliance on macro-organizational conceptions of strategy, it appears so. Why does this matter? With increasing service demands and highly constrained resources, public organizations are under intense pressure to perform; now more than ever, they must be strategic about expending resources in order to prosper in a highly scrutinized environment where faith in government is waning (Bryson, 2018;Gallup, 2021;Page, 2013). Strategy not only requires an accurate identification of a strategic issue, but also the appropriate application of a strategic solution-which is often contingent upon the micro-organizational context surrounding a particular issue and its proposed solution (Mitchell, 2019;Nutt, 2001). Particularly, strategic implementation efforts suffer from a focus on macro-organizational strategy that does not account for the micro-organizational complexity attached to a particular initiative. In other words, strategy implementation does offer plenty of nails for the organization's golden hammer to drive, but they are accompanied by just as many nuts, screws, and bolts-just because a local government is highly entrepreneurial in pursuing economic development does not mean it should take commensurate risks with implementing a new payroll system or repaving an arterial road. In an increasingly perilous and turbulent governing environment, how long can public administration managers and scholars afford to wait before expanding the toolbox?
This article presents a different path; marked by process, practice, contingency, and situation. It does not reject macro-organizational conceptions of strategy such as 'strategic stance' or 'implementation style,' but instead broadens the spotlight to also include micro-organizational strategy implementation activity and variation through Whittington's (2017) framework that sees strategy as practice and process. Micro-organizational implementation practices, often ignored in broader strategic management studies, provide richness and detail left unexplored by macro-organizational perspectives (Bryson et al., 2009;George et al., 2018). Favoring process over cross-sectional variance, the framework embraces the inherent temporal variation found in the different phases of implementation-planning, acquiring, executing, embedding, integrating (George & Desmidt, 2014;May et al., 2009;Poister & Streib, 1999;Van de Ven, 1992). Instead of seeking universal best practices, such an approach endeavors to identify best practices for the moment-tailored managerial guidance that best suits the unique context surrounding an initiative (based on its priority and complexity) given a particular implementation phase (Mitchell, 2019).
While this broader, interdependent conception of strategy provides theoretical safe harbor for micro-organizational perspectives, it is only justified if meaningful variation exists at that level. This study endeavors to demonstrate that micro-organizational activities not only have a significant effect upon implementation outcomes, but that these relationships are moderated by the situational context of the strategic initiative and the various implementation phasesin short, does the effectiveness of an implementation practice significantly vary across situational context and/or implementation phase? To test the model and hypothesis, the study examines the implementation practices of 205 strategic initiatives from 43 U.S. municipalities utilizing regression analysis for each of the five implementation phases, including three-way interactions that identify contextually appropriate practices. The analysis provides substantial evidence of micro-organizational, contingent variation that justifies a multi-level, interdependent conception of strategy; warranting expansion of contemporary public strategy implementation models. The article concludes with a specific list of implementation practices that are most impactful for a particular type of strategic initiative in a particular implementation phase-aiding municipal managers who increasingly must adapt to implementation challenges by deftly applying tools and practices that match the circumstances.

A Contingent, Micro-Organizational Process Model of Public Strategy Implementation
Beneath the surface, dynamic organizations are alive with a flurry of micro-level practices and processes all designed to effectively implement strategic initiatives. In contrast to the macroorganizational perspective built around a few dominant variables such as Miles and Snow's (1978) strategic stance, there are a "seemingly endless variety of factors" at the micro-level of an organization that influence implementation processes over time, resulting in a "complex mix" that must "make sense for a particular organization at a particular point in time" (Vinzant & Vinzant, 1996, pp. 142, 149). Whittington (2017) observes that process and institutional theories have been traditionally prone to strategic reductivism but are now moving toward incorporating more micro-organizational activity into their macro-and inter-organizational frameworks, respectively. Likewise, strategy-as-practice theorists have long championed micro-organizational strategic activity, but increasingly desire to attach it to broader constellations of theory. Therefore, Whittington (2017) sees a convergence within organizational theory that now allows for strategy to be simultaneously conceived as practice, process, and institution, presenting an excellent starting point for building a model of public strategy implementation.
In this conception, micro-organizational practitioners (those who do strategy work), practices (the tools, norms, and procedures of strategy work), and praxis (the activities and events found in strategy work) are linked to macro-organizational implementation processes that include phases and sequencing. The Whittington (2017) framework transcends the anecdotal nature of practice by allowing for greater temporal explanation of strategic practices through process analysis, more so than cross-sectional snapshots of macro-organizational strategic stances and executive traits. The strategy-as-practice perspective also offers a better epistemological approach to incorporating the contingent effect of situational context due to its emphasis on granular detail. Whittington (2017) offers his strategy framework from a private-sector context, but it is equally applicable to public organizations. Public strategic management scholars have begun the import of strategy-as-practice principles into the field by defining strategic practices in public organizations (Bryson et al., 2009), reconceptualizing public strategic management as 'strategizing' to emphasize human activity (Bryson & George, 2020;Hoglund et al., 2018), investigating how strategic management tools are applied in practice at the micro-organizational level of governments (Hansen, 2011;Mitchell, 2019), and exploring behavioral insights of strategy management participants as they strategize (George et al., 2018). The framework also assists in answering the call of Mitchell (2020) to distinguish micro-organizational public strategy implementation activity associated with particular strategic initiatives from macro-organizational strategic management and interorganizational policy implementation perspectives by developing concepts, models, and theories distinct from these broader fields.
Informed by Whittington's (2017) framework, the following sections outline how strategy operates in practice and process, culminating in a contingent, micro-organizational process model of public strategy implementation. First, strategy practices are generally described, along with how situational context can moderate their use. Second, the strategy implementation process is constructed, illustrating how strategy practices proceed in a coordinated fashion through implementation phases to produce proximate and distal outcomes. Collectively, this model depicts the micro-organizational practices that occur over time and within context to translate abstract strategy into concrete actions and outcomes.

Strategy as Practice
In public strategic management, practitioners employ dozens of practices (Jarzabkowski & Spee, 2009); some may ebb and flow frequently over time (process practices), while others are more enduring (design practices) (Mitchell, 2019). The dynamic utilization of these practices produces temporal variation across implementation phases, which is further compounded by the contingent effect of situational context. This variation provides scholars with an opportunity to examine the relationships between implementation practices and outcomes, moderated by both phase and context.

Practices
Strategy practices, also referred to as strategizing, serve as the heart and soul of this process model as they transform aspirations to capabilities (Bryson & George, 2020). Managerial practices (whether design-or process-oriented) have significant influence over organizational structure, strategy, and performance (Miles & Snow, 1978;Poister & Streib, 1999), and therefore, implementation. Design practices pertain to the employment of entrenched administrative systems and structure; although these organizational elements often remain static through an implementation process, implementation leaders can and do vary the use of them during implementation (Mitchell, 2019)-allowing for measurement of their respective implementation utility. Examples of design practices include resource availability and allocation (Poister & Streib, 1999), executive and stakeholder feedback mechanisms , project leader workload (Patanakul, 2013), strategic stance (Andrews et al., 2011), personnel stability (Andrews et al., 2016), organizational culture Fernandez & Rainey, 2006), and performance management integration (Poister & Van Slyke, 2002).
Conversely, process practices are people-oriented; the implementers and their microorganizational activities, communications, and perceptions (Mitchell, 2019). In contrast to their more enduring counterparts, process practices vary across all phases of implementation-perhaps multiple times. Examples include implementation team attributes and leadership (George, 2017;George et al., 2020), adaptation Walker, 2014), use of technology and consultants (Ahern et al., 2014;George & Desmidt, 2014), communication (Moe & Pathranarakul, 2006), stakeholder collaboration , and implementation monitoring and performance .

Situational Contexts
As managers deploy implementation practices, they do so with the belief these practices will lead to better implementation outcomes; however, these relationships are often moderated by environmental factors that center on citizen needs and political priorities (George, 2017;Nutt & Wilson, 2010) as well as the availability of technology to address those goals (Ahern et al., 2014;George & Desmidt, 2014). Additionally, an organization's structure/form of government , financial and professional capacity (George & Desmidt, 2014), culture (Walker, 2014), prior implementation performance (Nutt & Wilson, 2010), and strategic stance (Walker, 2013) can impact strategic management processes. Even at the micro-level, managerial factors related to leadership, teams, stakeholders, implementation complexity, practices, processes, and resource allocation all can interact to produce different strategic outcomes (George, 2017;George, 2021;Nutt & Wilson, 2010). Any one of these contingencies or a combination thereof comprise the situational context for strategy implementation.
In public strategic management, the large majority of scholarship has concentrated on the situational context of the organization; a sufficient unit of analysis when assessing a strategic portfolio. However, the same cannot hold true when one considers strategy implementation: The organization is not the unit being implemented, rather it is a particular strategic initiative. Thus, to fully incorporate contingency theory into public strategy implementation, one must consider the situational context of the strategic initiative.
Similar to problem structuring methods, 1 Mitchell (2019) distills contingencies down to two groups: initiative priority and implementation complexity. Relying upon a 2x2 typology (see Figure 1), initiatives with low priority and complexity are considered routine, as the implementation task is known to the organization and there is relatively little public scrutiny for the initiative. Those initiatives with low complexity and high priority are responsive efforts-these are relatively simple implementation efforts that are receiving attention from the public. Complex initiatives that are low priority are considered internal innovation, as difficult efforts with little external priority are typically driven by staff to improve the organization. Finally, complex initiatives that are high priority can be labeled as centerpiece initiatives-these are difficult efforts generally undertaken only because of immense public demand.

Strategy as Process
A process model represents a narrative epistemology-temporal in nature, driven by events, establishing sequence, and tracking variation over time ( Van de Ven, 2007); suitable for linking public strategy implementation micro-and macro-organizational activity. Strategy implementation is a process; dynamic through its phases with inherent richness and variety much like the environmental, organizational, and managerial contexts that surround it. Therefore, public strategy implementation theory should account for a strategic initiative as it progresses through micro-organizational processes, with careful consideration of how these activities are affected by the initiative's situational context. This epistemological pivot avoids

Routine Responsive
Low <=Priority=> High the strategic reductivism that limits macro-organizational approaches and offers new insight into what truly transforms strategy from plan to reality.

Phases
Scholars and practitioners of strategy implementation generally agree that phases exist but provide little definition beyond the whole of activity that occurs between strategic formulation and evaluation. As the most widely accepted professional manual for implementers, The Project Management Body of Knowledge (PMBOK) lists five different phases of implementation: Initiating, Planning, Executing, Controlling, and Closing (Project Management Institute, 2017). While offering a sound foundation, the project lifecycle does not sufficiently elevate key aspects of strategy implementation such as resource acquisition (Poister & Streib, 1999), adaptation/learning (Bryson, 2010), or integration (Vinzant & Vinzant, 1996).
The first two PMBOK phases (initiating and planning) largely align with what strategic management scholars refer to as formulation; however, some of these activities occur after strategy formulation is complete and strategic initiatives are identified-therefore within the domain of strategy implementation. Implementation planning activities include the creation of a plan for action, identification of necessary resources, and development of a timeline and budget; emulating the concepts of formulation and goals (Ahern et al., 2014), conceptualization (Van den Ende & Van Marrewijk, 2014), and design (Edkins et al., 2013)to the extent they pertain to implementation activity. Second, acquiring resources-the procurement of necessary human, financial, and physical capital-ensures that the 'who' and 'what' are in place prior to commencing with implementation. The acquiring phase includes the concepts of contracting and procurement (Edkins et al., 2013), feasibility (Van den Ende & Van Marrewijk, 2014), and organizing (Nooriafshar, 2013).
To adequately meld strategy implementation concepts such as adaptation, learning, and integration with the final three PMBOK phases, one can turn to normalization process theory (NPT) (May et al., 2009). Rooted in sociology, NPT informs public strategy implementation by categorizing related activity: 1) implementing, 2) embedding, and 3) integrating. May et al. (2009) define implementing as bringing practice into action, which fits well with the PMBOK executing phase that includes the concepts of building and creating (Edkins et al., 2013). Next, NPT refers to embedding as the process through which new practice becomes incorporated into to everyday work-that murky period after execution of planned implementation tasks when formative evaluation, learning, and adaptation continue until the strategic initiative is comfortably nestled within the existing organization, mirroring the PMBOK monitoring/controlling phase. Finally, May et al. (2009) view integrating as the process of weaving new practice into the enduring social fabric of an organization, including the concepts of summative evaluation , handover (Edkins et al., 2013), and operation (Van den Ende & Van Marrewijk, 2014). Collectively, the contributions of PMBOK, May et al. (2009), and Poister and Streib (1999) offer an innovative set of implementation phases that leverages project management, sociology, and strategic management theory to create a new series of lenses to observe the developing process narrative.

Outcomes
No single concept provides a complete picture of strategy implementation success (Atkinson, 1999). In policy implementation, the ability to meet implementation specifications and policy objectives in reality is known as fidelity (Emshoff et al., 1987). Separately, project management scholars have identified two dimensions of success: 1) proximate outcomes related to efficiency of the implementation process itself such as cost, time, and quality, and 2) distal outcomes that reflect post-implementation initiative impact toward achieving strategic goals (Baccarini, 1999), as well as broader public concerns (e.g., equity, responsiveness). When fused, the result befits public strategy implementation-fidelity to cost, time, and quality specifications serve as proximate outcomes, while meeting the associated strategic objective(s) represents the distal outcome. In this study, only proximate outcomes are evaluated as data are limited from U.S. municipalities regarding the distal outcomes stemming from particular strategic initiatives. 2 Figure 2 illustrates the contingent, micro-organizational process model of public strategy implementation once the concepts described above are assembled. A strategic initiative enters the implementation process with situational context attached (based on its relative levels of initiative priority and implementation complexity). Design and process implementation practices are applied to the strategic initiative as it progresses through the implementation phases, ultimately producing proximate and distal outcomes.

Hypothesis
The research question posits whether the four gray arrows representing the different situational contexts will produce the same outcomes when the same practices are applied at the same times. If not, that would indicate a contingent relationship between one or more implementation practices and the process outcomes. To test this, one can evaluate the efficacy of an implementation practice (or a combination thereof) to determine if this relationship varies by time and context, as follows: Hypothesis: The efficacy of a design or process implementation practice will vary by implementation phase and the situational context of the strategic initiative to which it is applied.

Sample Section and Size
The sampling frame for this study are the strategic initiatives associated with the 1,040 municipalities who were awarded the Distinguished Budget Presentation Award for FY 2014 from the Government Finance Officers Association (GFOA), 3 the most set of awardees available during the data collection window of September 2015 to May 2017. During that time, eight random samples were taken without replacement in an effort to incrementally increase sample size within existing data collection resources, ultimately totaling 459 municipalities. The research team determined if these selected organizations possessed a strategic plan that 2) The plan contains defined strategic initiatives; and 3) These strategic initiatives are finite in nature (possessing a defined beginning and end). 4 From a collaborative review by the research team, 165 (35.9%) municipalities met all three criteria. The respective chief administrative officers (CAOs) were then contacted via email to request participation in the study, with email and telephone followups as necessary-ultimately, 43 municipalities (26.1%) agreed to participate. From each participating municipality, five strategic initiatives were randomly selected retrospectively from its FY 2012 list, creating a total of 215 strategic initiatives that constitute the sample for this study.
Data collection proceeded along two avenues: perceptual surveys and information requests. The CAO or his/her designee was asked to identify a project leader from the organization for each of the five selected strategic initiatives, along with an elected official who had served since 2012. Basic implementation information was asked of each project leader regarding initiative completion, time, and cost, which was obtained for 186 of the 215 strategic initiatives (86.5%).
To elicit a variety of perspectives, a separate survey was distributed to the CAO, the identified elected official, and the project leaders asking questions about the organization's general approach to implementation and related to the initiative. Most of the survey questions ask the respondent to rate their level of agreement with a statement about the initiative's implementation on a 5-point Likert scale. To assist the respondents in differentiating between implementation phases, each was defined and described within the survey question.
Ultimately, 213 surveys were distributed to these individuals, with 179 returned (84.0%). To transform the respondent survey data to the desired unit of analysis, the survey responses for each initiative were combined to create mean response values regarding each survey question; forming the bulk of the dataset as a number of the questions are employed as independent variables in the subsequent analysis. Data for the control variables were collected from the U.S. Census, while the context variable data were produced via coding by the research team.

Dependent Variable-Implementation Proximate Outcomes via Efficiency
This study limits its examination to proximate implementation outcomes only as they are the most directly impacted by strategy implementation practices, while distal outcome data for U.S. local governments is largely unavailable. These proximate outcomes are rarely an agreedupon construct, however; even within an organization (Lim & Mohamed, 1999). At the simplest level, was the initiative completed (Okumus, 2003)? Beyond this, was the initiative completed on time and on schedule (Atkinson, 1999;Pinto & Slevin, 1988)? More subjective measures have also been used, such as the satisfaction of customers, stakeholders, and organizational leaders (Lim & Mohamed, 1999;Munns & Bjeirmi, 1996;Okumus, 2003;Poister & Streib, 2005). However, researchers have found that subjective measures of implementation success introduce bias into the data, do not have a relationship with objective implementation results, and should be avoided (Bommer et al., 1995;Liu & Walker, 1998;Olson et al., 1995;Prabhakar, 2005). Due to this widespread disagreement of the appropriate implementation success measures, the use of multiple measures to illustrate success is recommended (Jugdev & Muller, 2005;Kerzner, 1987).
This study utilizes Mitchell's (2019) implementation efficiency index (IEI) to measure a combination of proximate outcomes. IEI utilizes the traditional implementation proximate outcomes of completion, cost, and time, which ranges from 0 to 1 with incomplete initiatives receiving a 0 and completed initiatives initially receiving a 1. The completed initiative score is then multiplied by the product of the ratios of predicted/actual values for completion time and expended implementation dollars. The most efficient initiatives will receive a score of 1, as they were completed on-budget and on-time, while initiatives not completed receive a score of 0. All other initiatives receive a score somewhere between 0 and 1 (creating a continuous variable) as efficiency is moderated by the effects of delays and overspending. The IEI construct is theoretically preferred as it: 1) utilizes the traditional proximate outcomes of completion, time, and cost; 2) relies upon objective outcome data rather than subjective perceptions; and 3) utilizes multiple measures in its calculation.

Independent Variables
The independent variables in the study are divided into four groups: 1) design practices, 2) process practices, 3) situational context, and 4) controls. The first three categories stem from the strategy-as-practice discussion in the earlier model development section. The design variables represent structural, financial, and social constructs that can be leveraged within an implementation effort to improve proximate outcomes, while the process variables focus on the implementation team and its tactics and performance. Collectively, the design and process practices serve as the research variables in this study, as each is expected to significantly influence implementation efficiency in at least one phase and/or at least one situational context. The independent variable data were primarily collected via survey, asking respondents to indicate their agreement with statements indicating the presence of these strategy implementation practices.
While their individual relationships with the dependent variable are theoretically and practically valuable, it is the combination of these variables that are associated with IEI in each phase and context that provide the basis to evaluate the study hypotheses. The design practices serve a dual purpose as they, along with the control variables, mitigate potential highperforming-organization sampling bias by controlling for professionalism and resources. The detailed literature support, measurement strategy, and operationalization for each of these independent variables are found in Figure 3. Budgetary funding critical to acquiring resources necessary to execute implementation activities (Poister & Streib, 1999;Mitchell et al., 2021) Survey: "The project was adequately funded." Likert scale: 1=disagree, 5=agree Change-Conducive Culture Presence of a change-ready and changeconducive organizational culture prepared for organizational alteration (Fernandez & Rainey, 2006) Survey: "In general, do you believe that your organizational culture supports and prepares for change?"

Yes=1, No=0
Supportive Stakeholder Coalition Establishment of stakeholder coalitions that support initiatives for a particular strategic initiative (Mitchell, 2021) Survey: "Do you believe an active and supportive stakeholder coalition existed for this initiative from the idea stage through the end of implementation?"

Yes=1, No=0
Process Practices Adaptive Ability to Challenges The ability to adjust practices during implementation based on feedback increases chance of success (Bryson, 2010) Survey: "The project team effectively adapted to the challenges they encountered." Likert scale: 1=disagree, 5=agree Defined Project Leadership Continuity and order in strategic management leadership required for organizational transformation (Vinzant & Vinzant, 1996) Survey: "The project has a defined leadership structure." Likert scale: 1=disagree, 5=agree External Communication Quality Clear messages and appropriate channels are necessary to inform external stakeholders and audiences (Moe & Pathranarakul, 2006) Survey: "The project team effectively communicated with external audiences." Likert scale: 1=disagree, 5=agree Internal Communication Quality Projects are not routinized activities, requiring coordination between work units; communication among the team is key (Pinto & Prescott, 1988) Survey: "The project team effectively communicated among themselves to move implementation forward." Likert scale: 1=disagree, 5=agree

Project Leadership Quality
Strong leadership is key during organizational uncertainty, exhibiting ability to match implementation tactic to need (Vinzant & Vinzant, 1996) Survey: "The project implementation leadership was effective." Likert scale: 1=disagree, 5=agree

Contextual
Initiative Priority Prioritized initiatives 1) receive necessary resources and staff (Pinto & Prescott, 1988) and 2) higher scrutiny of implementers (Gliddon, 2004) Inclusion of initiative in FY 2014 budget message, inter-coded by research team Yes=1, No=0 Implementation Complexity Complex initiatives are typically more difficult and riskier than routine ones, accompanied by high levels of uncertainty and little past organizational experience (Faleye et al., 2011) Ordinal scale based on level of process reform and innovation, intercoded by research team 0=No reform, 1=Process reengineering, 2=New, 3=Transformation Situational Context Interaction of initiative priority & implementation complexity creates distinct situational contexts that affect relationship between implementation tools and efficiency (Mitchell, 2019) 2x2 typology formed by high and low levels of initiative priority (P) and implementation complexity (C), as depicted in Figure 2 If an organization has a "rainy day fund" or other slack resources, then it can better adapt to unforeseen circumstances and implementation cost overruns (Miller et al., 2007) Unrestricted general fund balance as a % of general fund expenditures, collected from FY 2014 financial audits

Ratio scale
The contexts are operationalized by applying the Mitchell (2019) 2x2 typology of strategic context, established by categorizing strategic initiatives in terms of their relative level of strategic priority and implementation complexity (see Figure 1). The four resulting situational contexts (Routine, Responsive, Internal Improvement, and Centerpiece) offer multiple lenses to evaluate the relationship between designated funding and implementation efficiency, uncovering any contingent effects that may exist. The priority and complexity coding was completed by the research team relying upon intercoder reliability principles; first, each team member coded the strategic initiatives individually; then, in instances where individual coding was not unanimous, the final decision was reached through team deliberation.

Data Description
The descriptive statistics are summarized in Table 1. For the 186 initiatives sampled between 2015 and 2017 with full data, 61% were fully implemented within the 3-year evaluation window. Of the completed initiatives, 87% were done so within budget, and 67% were on time.
Distinct from the proportion of completed initiatives, the mean value of the dependent variable (IEI) is 0.52-illustrating the moderating effect of time and cost upon the IEI. Possessing a mean value near midrange along with a relatively large standard error value, IEI demonstrates sufficient variation to lessen concerns regarding the narrow sampling frame of municipalities.
Regarding design practices, the average set of respondents felt that the respective strategic initiative was implemented within an "organizational culture [that] supports and prepares for change" 93% of the time. The average set of respondents for each initiative rated the initiative's level of implementation funding adequacy as 4.09 on a 5-point Likert scale, scoring just above "Somewhat Agree" and well below "Agree." Just under 80% of the mean set of respondents felt "an active and supportive stakeholder coalition existed…from the idea stage through the end of implementation".
Five different implementation process practices are evaluated in this analysis. The dataset contains an observation for each process practice in each of the five implementation phases, representing the average submitted response to a question gauging level of agreement to the presence of the practice in that particular phase. In general, the response averages range from 3.94 to 4.34 on the scale (centered around "Somewhat Agree") and tend to rise with each subsequent phase. The respondents most agreed that Project Leadership Quality was present for the respective initiative, while they least agreed that External Communications Effectiveness was present. The number of responses for each phase tended to decrease over time, representing initiatives that were abandoned before reaching later implementation phases.
Situational context variables act as moderators for the statistical relationships tested in this study, determined by low and high levels of priority and complexity. The dataset is exactly split between cases of low and high complexity, while 53% are rated as lower priority (47% rated as higher priority). This configuration results in 55 routine initiatives (26%, low priority-low complexity), 49 responsive initiatives (24%, high-low), 54 internal innovation initiatives (26%, low-high), and 49 centerpiece initiatives (24%, high-high).

Testing the Models
To complete the inferential analysis, this study utilizes five OLS regression models (one for each implementation phase), controlled for the random effects associated with each The five process practices are also evaluated for inclusion into the model, as well as their respective three-way interactions with the Priority and Complexity context variables. Threeway interactions are a preferred method for studying relationships moderated by two dichotomous variables as they can identify significant relationships and slope values for a given focal independent variable in all four of the resulting contingencies (Jaccard & Turrisi, 2003). Since the number of observations in each of the five models range from 156 to 167, the number of variables included in the model cannot exceed 16 in deference to the "no less than ten cases per independent variable" rule of thumb. Due to this limitation, only microorganizational process practices are subject to interactions since they possess more variability through the implementation process. Since each additional three-way interaction creates three new variables, only two process practices can be included into each model. Therefore, the study utilizes a stepwise approach to determine which two process practices provide the best fit for the respective model. Finally, control variables representing state-level population density and organizational fund balance are included in the model to address the sampling bias concerns discussed above by accounting for professionalism and political ideology. Thus, each model includes 16 total variables representing two process practices, three design practices, seven interactions, two situational contexts, and two controls.
A summary of all five phasic models is included in Table 2, with each producing a statistically significant relationship with implementation efficiency and explaining 16-20% of IEI variation in its respective phase. Multicollinearity and robustness checks were performed to further test the models, with no significant issues identified. The design and process practices all produce multiple significant and often contingent relationships with IEI, justifying their inclusion in the models. For all other variables, only the relationship direction and significance are reported to consolidate presentation. To demonstrate magnitude, the predicted slope values for all significant relationships are presented in Table 3. Since the values are regressed on a dependent variable with a 0-1 range, they can be interpreted as the estimated percentage improvement (or decline) in implementation efficiency for the focal variable at the intersection of a given phase and context per unit of the survey response scale (Jaccard & Turrisi, 2003).

Planning Phase
The two process practices, Adaptive Ability and Defined Project Leadership, both produce statistically significant contingent relationships with implementation efficiency during the Planning phase. Within the Routine context (low priority and complexity initiatives), the process practice of Defined Project Leadership has a positive relationship with IEI, boosting its value by 16.0% per unit of average response (for context, this variable's sample mean of the respondent average is 4.27 on a 1-5 scale). However, in the Internal Innovation context (low priority/high complexity), Defined Project Leadership reduces IEI values by 28.4% per scale unit. This is buffered by Adaptive Ability, which increases IEI by 30.9% per scale unit for initiatives within this context. These two variables do not have significant relationships with IEI in all other contexts. As for design practices, two possess strong positive relationships with IEI during the Planning phase. The perception of a Change-Conducive Culture increases IEI by 31.9%, while a perceived Funding Adequacy increases by IEI by 9.0% per scale unit. The control variables do not possess any statistically significant relationships with IEI. Overall, the Planning phasic model accounts for 18.5% of IEI variation.

Acquiring Phase
The selected process practices, Project Leadership Quality and Internal Communications Effectiveness, only produce a statistically significant relationship with implementation efficiency in the Internal Innovation context. Each scale unit increase of average agreement intensity for Internal Communications Effectiveness increases IEI by 35.5%, while Project Leadership Quality reduces IEI by 37.4%. Regarding design practices, the results largely

Discussion
Does the effectiveness of an implementation practice significantly vary across implementation phase or situational context? A cursory glance at Table 2 provides all the information necessary to answer this question in the affirmative. Each implementation phase contains a distinct set of practices that significantly affect implementation efficiency, either positively or negatively. Likewise, one can also identify a different group of recommended practices for all four types of strategic initiatives (distinguished by their situational context), driven primarily by the level of implementation complexity. Collectively, the analysis provides substantial support for the research hypothesis, along with the contingent micro-organizational process model of public strategy implementation from which it is derived.
The analysis offers practical guidance to navigating the contingent relationships of public strategy implementation, based on the situational context of the strategic initiative. In the Routine context where priority and complexity are lower, defined and effective leadership drive more efficient implementation (especially during the Planning phase), as well as effective communication with external stakeholders as implementation progresses (e.g., the Executing and Embedding phases). The ability to adapt to implementation challenges and the quality of internal communications generally appear to be negatively associated with implementation efficiency (but not to a significant degree within this sample), indicating that emphasizing these practices in this context may lead to delays and additional costs, or jeopardize the strategic initiative itself. In short, Routine initiatives benefit from an effective leader who can organize planning activities and keep stakeholders informed of implementation activities.
The converse is true for strategic initiatives in the Responsive (higher priority with lower complexity) and the Internal Innovation (higher complexity with lower priority) contexts. Adaptive ability during planning and embedding activities and internal communication quality during resource acquisition are crucial to implementation efficiency in these contexts; while emphasis on project leadership quality during acquiring, external communications during core implementation activities, and a defined project leadership structure while planning could each set back the implementation effort. Overall, these types of strategic initiatives appear to benefit from a strong, collaborative implementation team rather than rely upon defined leadership.
The implementation of strategic initiatives in the Centerpiece context where priority and complexity are higher largely mirror patterns found in the Routine context, however the effects are more pronounced. Effective communications with stakeholders during the Executing and Embedding phases promote implementation efficiency in this context, as well as effective project leadership as the initiative is integrated into the organization, while less emphasis should be paid to adapting in the Embedding phase and internal communications quality in the Integrating phase. Process practices within this context seemingly have the most effect in the latter phases of implementation, indicating a strong leader who effectively communicates with stakeholders is the key to implementation efficiency during the final push toward embedding and integrating the strategic initiative into the organization.
For practitioners, this study emphasizes the need to understand the context surrounding a particular strategic initiative, not only in terms of priority and complexity but also the current implementation phase. Public strategy implementation is a dynamic process; its management should respond in kind. While the contextual recommendations offered from this study (summarized in Figure 4) only scratch the surface of implementation contingency, public strategic practitioners should constantly assess the priority attached to a strategic initiative, the complexity of its implementation, its stage within the implementation process, and the interplay between the three; and allow for such diagnosis to drive implementation practices.
Theoretically, the study provides initial support for a contingent, micro-organizational process model of public strategy implementation, further validating Whittington's (2017) framework of strategy as a practice and process and applying it successfully to the public sector. The analysis demonstrates that relationships between implementation practices and proximate outcomes are moderated by situation and phase, establishing both temporal and contextual contingencies within strategy implementation. The findings support the notion the implementation practices differ in their influence, with process practices varying to a greater degree than design practices. Most importantly, the study provides an alternative for strategic  Due to the broad nature of the model proposed here and its multiple contingencies, its full scope could not be tested in a single study. The following questions are left for future research: 1) How do iterative processes of executing and embedding generate organizational learning? 2) How is organizational learning diffused to other governments via broader institutions? and 3) Do implementation practices affect distal outcomes contingently? If so, how? The study design also limits its generalizability in a number of meaningful ways. First, the size and composition of the dataset creates analytical challenges-especially when utilizing three-way interactions in regression. A larger, more organizationally diverse sample size would eliminate the need for stepwise regression tactics and lessen the impact of unspecified organizational effects. Second, the scope of municipalities should also be broadened beyond those with GFOA award-winning budgets and initiative-specific strategic plans to expand the prescriptions of the research to all local governments. Finally, conceptualizing strategic success in terms of implementation efficiency ignores distal outcomes that have direct impact upon communities and their citizenry; with a potential bias toward defining "successful" strategic initiatives as those that are less difficult to execute.
Ultimately, this study emphasizes the need to retire Maslow's (1966) golden hammer in public strategy implementation-a standard hammer works just fine when accompanied by wrenches, screwdrivers, and pliers. In practice, the dynamics of organizational change cannot be distilled down to just a few variables as is regularly done in macro-organizational studies.
Even when treated contingently, these broad concepts can only capture a small portion of the variation created by the rich array of implementation actors and their activities. Further, practitioners continue to show a proclivity toward one-size-fits-all solutions that limit options and frustrate progress. Both approaches set aside the immense complexity inherent to public strategy implementation, which can only be remedied by a deeper dive into microorganizational exploration. But adding more tools to the toolbox is only as effective as knowing when to appropriately use them. The situation is key, as is the ability to identify it-this study represents an early attempt to provide such guidance by identifying contingent best practices based on initiative context and implementation phase. As a field, those who practice and study public strategy implementation might quickly realize they have many more tools at their disposal to improve strategic outcomes and therefore government effectiveness-but only once they put down the hammer.

Notes
1. Problem structuring methods refer to a broad group of decision-making models that assist in understanding the context and complexity of a problem to better formulate a solution (Rosenhead, 2013). Most operate on a spectrum for a problem dimension, or multiple spectra to create a typology. A number of these tools have been applied to strategy making (Ackermann, 2012). The Cynefin framework (Snowden & Boone, 2007) is a popular decision-making model that focuses on problem complexity in terms of cause-and-effect relationships and what can be known about them. This model has similarities to the Mitchell (2019) context framework cited in this study but does not consider the organizational priority attached to problem resolution. 2. Although the use of strategic management by municipalities is rising (Poister, 2010), it is still a relatively new tool for local governments. The situation limits the study of distal outcomes because a longer evaluation period is necessary to realize if a long-term impact has occurred. This creates a paradox because as the evaluation period is lengthened (a minimum of 3-5 years post-implementation is necessary to measure full impact), one soon encounters a dearth of municipalities with an adopted strategic plan and pre-and postimplementation distal outcome data. This reality distinguishes this study from previous public strategic management work, where distal outcome data were readily available (Andrews et al., 2011;Meier et al., 2007).
3. The GFOA Distinguished Budget Presentation Award encourages state and local governments to incorporate best practices as they prepare budget documents. Over 1,600 governments have received the award. Focusing upon this group of municipalities substantially increases the convenience of data collection, as the award requires a statement of organization-wide strategic goals and strategies in budget documents (GFOA, 2005). This choice may have implications for generalization as GFOA award winners are typically better performing governments overall. Control variables and design practices included in the models account for any unexplained advantages in terms of professionalism and resources. 4. These criteria serve two purposes: 1) They reflect best practice in strategic management by creating actionable initiatives that are easily evaluated (Walter et al., 2016), and 2) They ensure the study can be conducted at the initiative level of analysis and provide proximate outcome data for the IEI dependent variable.

Disclosure Statement
The author(s) declare that there are no conflicts of interest that relate to the research, authorship or publication of this article.