Goats or wolves? Private sector managers in the public sector

Public administration reforms have propagated the use of private sector management skills in the public sector, and an increased openness to managers with a private sector background. This has created a debate between those who think private sector experience improves public institutions by bringing core managerial values such as results orientation, efficiency, or openness to innovation, and those who argue that private sector experience can damage core public sector values, such as impartiality and equity. Despite the abundant anecdotal evidence, broad empirical evidence on the effects of private sector experience on public managers' values remains limited. Using data from a survey among central government top managers in 18 European countries, we show that public managers with private sector experience have, as expected, more core managerial values. Yet, unlike the conventional view, core public values do not suffer. for customers (e.g., one stop shops), quality management systems, codes of conduct, internal steering by contract, management by objec-tives and results, benchmarking, cost accounting systems, decentralization of financial decisions, decentralization of staffing decisions, appraisal,

Office in 2014. Soon after taking office in 2007, then UK Prime Minister Gordon Brown announced his executive would be a "Government of All Talents" (GOAT), bringing respected nonpolitical professionals to top executive posts. A review of the GOAT initiative, titled "Putting Goats among the Wolves," concluded that "views about the success of these appointments were at best mixed:" many "goats" left government after a relatively short time and government's commentators were "scathing about their achievements" (Yong & Hazell, 2011, p. 10).
Summoning private sector high flyers to public offices has been popular for a long time. Woodrow Wilson appointed several business executives at the onset of the First World War, who crucially helped to adapt the economy for wartime production (Rappleye, 2017). Among the Wilson appointees were Herbert Hoover, a mining executive, and Bernard Baruch, known as "the lone wolf of Wall Street." In turn, both of them brought dozens of other business executives to assist them and, overall, they succeeded in their tasks. At the height of the Progressive Era, the appointment of private sector managers was thus considered legitimate.
Inspired by New Public Management (NPM) ideas, recent decades have witnessed a renewed interest in bringing private sector managers to public office. Especially in executive agencies, private sector experience has become more common (Horton & Jones, 1996;Van Thiel, Steijn, & Allix, 2007). In this century, an increase in switching from private to public sector has been documented (Frederiksen & Hansen, 2017), and sector switching, even if relatively new, was an increasingly normal phenomenon in the late 20th century (Light, 1999). Such "sector switchers," as they are generally called, are commonly found in technical and professional roles (Bozeman & Ponomariov, 2009;Zhang, 2017). In addition, those having a managerial occupation in the private sector have a higher probability to switch to a managerial role in the public sector (Su & Bozeman, 2009). Public managers also switch to the private sector (Frederiksen & Hansen, 2017), especially near the end of their career.
What are the effects of this switching? Which are the consequences of bringing professionals with private sector experience to the public sector? Are public managers with private sector experience different from those lacking it? And, if so, in what sense? For many, private sector experience should be, in principle, associated with "a broader horizon, human capital acquisition and thus a better preparation for management and leadership tasks in public administration" (Bach & Veit 2018, p. 260). Yet, from early on, critics have argued that "the notion that there is any significant body of private management practices and skills that can be transferred directly to public management tasks in a way that produces significant improvements is wrong" (Allison, 1986, p. 472). Is private experience the "fount of most practical management wisdom" (Painter, 2011, p. 23)? Have core private managerial values, like efficiency and effectiveness, come at the expense of core public values, like "equity, fairness, communal values" (Bozeman, 2007, p. 82)?
Despite the political rhetoric emphasizing the added value of private experience and abundant biographical storytelling about the experiences of top private managers who have switched to public sector roles, broad empirical evidence remains limited and speculation is more prominent than evidence (Bozeman & Ponomariov, 2009;Rainey & Bozeman, 2000). Additionally, many contributions on the differences between private and public values are ideological rather than descriptive (Van der Wal, de Graaf, & Lasthuizen, 2008).
This article aims to address this gap by testing the effects of private sector experience on attitudes and values of public managers. Following the literature, we hypothesize, first, that public managers with private sector experience will show more managerial values, in particular, results orientation, efficiency, and openness to innovation, and will use more management tools than their counterparts without experience in the private sector. And, second, we hypothesize that public managers with private sector experience will give, ceteris paribus, less importance to core public values such as impartiality and equity. These hypotheses are tested with a multilevel quantitative analysis utilizing the Coordinating for Cohesion in the Public Sector of the Future (COCOPS) Top Public Executive Survey (Hammerschmid, Van de Walle, Andrews, & Bezes, 2016), a survey which collects the views of 5,012 top public managers from 18 European countries.
The article is organized as follows. We first review the literature on private managers in the public sector, including the reasons typically adduced for wanting to introduce managers with private sector experience into the public sector, and evidence about the scope of the phenomenon. Then we introduce the data, the research strategy, and discuss our findings. As expected, we find that private sector experience-as well as the longer experience as a manager in the private sector-is associated with public managers' managerial values (i.e., results orientation, efficiency, and innovative attitude) and their use of more management tools (i.e., performance indicators). Yet, unexpectedly, the amount of private sector experience generally does not have a significant negative effect on the importance public managers give to impartiality and equity, except among managers with more than 20 years as a manager in the private sector. That is, previous experience in the private sector is linked to more managerial values among public managers, but not to lower support for core public values. Interestingly, results also suggest a positive impact of the years of public sector experience on core public values. Controlling for significant factors such as age, educational level, organizational type, level of politicization of the organization, and policy area, we find that the years of public sector experience are associated with public managers' emphasis on impartiality and equity. In sum, public values such as equity and impartiality do not seem to be harmed by the length of time managers spend in the private sector, but they seem to be boosted by the experience in the public sector.

| THEORY: BRINGING IN PRIVATE MANAGERS
Public administration reforms in recent decades, and not least NPM-inspired ones, have propagated the use of private sector management skills in the public sector. This has not only resulted in the transfer of management tools and practices, and in a renaming of directors and administrators in chief executives, but also in an increased openness to managers with a private sector background. Yet, what do private sector managers bring to the public sector?
There has been a long tradition in public administration research of comparing public and private sectors (Buchanan, 1975;Perry & Rainey, 1988). For some, the private sector management model is seen as superior and therefore should be imported (Farnham & Horton, 1993, p. 46). For others, there is no reason why there should be differences between the public and the private sector (Rainey & Bozeman, 2000), and there is little reason to believe that daily managerial work would be fundamentally different (Allison, 1986;Johnson & Dobni, 2016). There are also doubts about why the public/private distinction should be more important than interorganizational or sector differences . A review of existing studies finds that differences are not conclusive with some exceptions, namely, that public managers are less materialistic, have lower organizational commitment, and that public organizations are more bureaucratic (Boyne, 2002).
In another tradition, private and public sectors are fundamentally different (Box, 1999). Due to the absence of market forces and property rights in the public sector, public organizations cannot be run as private ones. This "publicness" approach (Bozeman, 1987) argues that the public and private sectors can differ on different dimensions (Scott & Falcone, 1998). Managerial positions in the private sector tend to be associated with managerial values of efficiency, whereas public positions are related to a more complex set of competing values (Clerkin, Christensen, & Woo, 2017; see also Spicer, 2015).
Empirical scholarship on sector switchers has tended to focus on switchers' job satisfaction, motivation (Hansen, 2014), job involvement (Boardman et al. 2010), experienced value differences (de Graaf & Van der Wal, 2008), careers and income changes (Bozeman & Ponomariov, 2009), or on perceptions of red tape and formalization (Chen, 2012). A lot of this work has been based on in-depth interviews with smaller samples (Apfel, 2013;, but increasingly, new data have become available that allow larger studies (Boardman, Bozeman, & Ponomariov, 2010). Yet no broad studies covering both country-and individuallevel variation on private sector experience affecting attitudes of public managers has been undertaken. This article aims to do so.
Following different strands of literature, we hypothesize, first, that private sector experience is associated with public managers who are more aware of core managerial values (i.e., results orientation, efficiency, and innovative attitude) and use more management tools (Hypothesis 1) than public managers lacking private sector experience. And, second, those managers with private sector experience consider the core public values impartiality and equity less important (Hypothesis 2).
Regarding the first hypothesis, a large literature has noted how public organizations can benefit from experience outside the public sector. Managers with private sector experience bring "managerialism": the idea that better management will make public organizations perform better and better solve social problems (Diefenbach, 2009;Painter, 2011;Pollitt, 1990). In some way, managerialism also reflects older ideas from administrative science that the management function of an organization is generic: the public or private status of an organization should not matter for managing it. In the public sector, as in the private one, this thinking created a new type of generic manager as opposed to the older model in which specialists manage organizations (Farnham & Horton, 1993). There should be, it was thought, a convergence of managerial behavior between public and private sector managers (Poole, Mansfield, & Gould-Williams, 2006).
Hence, bringing in managers having only private sector experience should not be a problem (Farnham & Horton, 1993). Quite the opposite, managers with private experience possess a more advanced knowledge of management tools and skills. And sector switchers transfer this knowledge thanks to "sector imprinting" (Boardman et al., 2010;Chen, 2012)-that is, managers hold certain norms, attitudes, skills, and behaviors that they take with them when switching sectors, making them particularly valuable in the new position. This has also been noted in private sector management scholarship that has looked at whether a new CEO is an insider or an outsider to the organization (Petrovsky, James, & Boyne, 2015). Karaevli (2007) uses the term "outsiderness" to describe the value of an outside CEO bringing in new skills, insights, or leadership styles to a company. Outsiders are not necessarily at an advantage, but they may bring different skills and tools. In particular, there are three arguments that suggest private sector experience will foster managerial values and tools in the public sector.
First, there is the radical change argument. Bringing in managers from the private sector is thought to be needed to shake things up, and to put a rusty public organization back on the tracks in order to achieve results, or to implement radical strategic or structural change (Boyne, James, John, & Petrovsky, 2010). The argument does not call for private managers in particular to take over, but suggests that a different manager will fulfill his or her role in a different, less bureaucratic way.
The second argument for introducing managers with private sector experience in public organizations refers to the values and ideas they are supposed to have, and their innovative and entrepreneurial mindset. It is a cultural argument that calls for a change of personae because of the new values they bring. This is an argument similar to that used for justifying political appointments (Mackenzie, 1987).
The third argument is cognitive. Managers with experience in the private sector possess certain knowledge and skills that are needed in the public sector. The cognitive argument is used to attract managers who are, for instance, experienced users of modern management tools, competent users of modern management jargon, or experienced managers of large organizations (George, Hammerschmid, & Van de Walle, 2019).
Nevertheless, even if public managers who come from the private sector bring management skills and tools, they have "little experience of responding to politicians, public funding, and regulatory constraints" (Petrovsky et al., 2015). The private sector "goats" may encounter the political "wolves" (Yong & Hazell, 2011, p. 10). Or, simply, there could be a problem of "publicness fit" (Petrovsky et al., 2015), because private sector managers may not fit with the values of a public organization.
Some earlier studies have explicitly looked at the effect of prior private experience on public managers' attitudes and behaviors. Fernández-Gutiérrez and Van de Walle (2019) explore the inescapable dilemma public managers may face between efficiency on the one hand and equity on the other. They do not find differences between managers with at least 1 year of experience in the private sector and those who lack that experience. When controlling for confounders such as (a business or economics) education, or having worked for the private sector, there is no significant effect on public managers' preference for efficiency over equity.
Nevertheless, there are reasons to expect private sector experience to mold the values of managers and, in particular, to weaken core public values such as impartiality, lawfulness, and neutrality (Frederickson, 1997(Frederickson, , 2005Lane, 1994). First, managers from the private sector may have more "market values" (Lane, 1994) than those in the public sector. Equity has often been presented as being in direct opposition to efficiency (Fernández-Gutiérrez & Van de Walle, 2019) especially when allocation and redistribution decisions need to be made. This tension has also been one of the main arguments for criticizing the ideas behind the NPM, ostensibly because efficiency and equity belong to different groups of administrative values (Hood, 1991). Indeed, "public administration faces a serious and seemingly irresolvable problem in continually seeking to maximize the attainment of mutually incompatible values" (Rosenbloom, 1983, p. 219). Consequently, sector switchers from the private sector may give more importance to incentive compatibility vis-à-vis broad public goals such as redistributive equity. For instance, when working in social security systems, a marketoriented professional may be inclined to "not produce or stimulate the problems that they are supposed to counteract, such as unemployment" (Rosenbloom, 1983, p. 141). Her concern will be minimizing the risks of moral hazard and adverse selection-even if this comes at the expense of equity.
Similarly, managers with private sector experience may give less preference to impartiality. For impartiality requires the rigorous observation of due processes. Strict procedures, and legal controls, are a safeguard against the abuse of power. Impartiality requires the adhesion to "legal values" (Pierre 2011), such as entitlements, uniformity, and the search of predictability, legal security, and procedural fairness. And these values may be opposed to the "managerial values" acquired by professionals in the private sector, and that are defined by needs, adaptation to clients' requirements, and the search for efficacy, effectiveness, and efficiency (Pierre 2011). This is for instance visible in requirements for hiring staff or awarding projects and contracts (Aucoin, 2012). Public values such as equal treatment, proper procedures, and predictability may figure less prominently for managers with private sector experience for, unlike the efficiency in outputs or outcomes, they cannot be quantified or measured (Lane, 1994). The public sector relies on a public service ethos that can be affected by managerial (i.e., NPM) reforms (Maesschalck, 2004).
Additionally, according to the theory of contrast effect, managers coming from the private sector may perceive things differently in public organizations. In particular, it has been found that sector switchers notice more formalized personnel rules and red tape than individuals who have not switched (Chen, 2012). This is particularly the case if the public manager has stayed in the private sector long enough for memories to become imprinted. Consequently, it is plausible that, ceteris paribus, managers with extensive "private sector imprinting," exhibit a preference for some values-for example, less preference to impartiality and equity-than those who lack it.
As a result, we hypothesize that previous private sector experience will have a dual effect on the values of public managers: it will increase managerial values, but it will hamper core public values such as impartiality and equity. Therefore, we will subsequently proceed to test these two hypotheses: Hypothesis 1 Ceteris paribus, managers with private sector experience are more likely to exhibit managerial values (i.e., achievement/results orientation, efficiency, and innovative attitude) and to use management tools than managers lacking previous private sector experience.
Hypothesis 2 Managers with private sector experience are less likely to exhibit core public values (i.e., impartiality and equity) than managers lacking previous private sector experience.

| DATA AND METHODS
While political science has long embraced cross-national comparison, the scholarly analysis of public administration and bureaucracy has a relatively modest history of comparative research (Dahlström, Lapuente, & Teorell, 2012;Dahlström & Lapuente, 2017;Suzuki & Hur, 2019). This trend has begun to change as systematic data of bureaucratic structures and bureaucrats' behavior are becoming increasingly available. This article will exploit the COCOPS Executive Survey on Public Sector Reform in Europe which tracks survey answers from 2012 to 2015 of over 9,000 senior public sector executives from 21 European countries (Hammerschmid, Oprisor, & Štimac, 2015). 1 The goal of the European Union-funded COCOPS project was to provide cross-national and quantitative data to determine the effect of NPM-style reform in Europe (Hammerschmid et al., 2015). The survey-designed by a team of cross-national public administration scholars-makes possible the comparative analysis of the experience and understanding of public sector executives with respect to the current status of management, coordination, and administrative reforms, the effects of NPMstyle reforms on performance, and the impact of the financial crisis. The survey's broad target included top public managers from all central government ministries and agencies in the countries surveyed as well as relevant regional and state ministries; the breadth of the target helps to prevent random sampling and response bias issues. 2 National research teams mapped and identified the top executive positions in central government ministries and agencies (as well as relevant regional ones). The COCOPS survey targeted 36,892 managers from the following 21 countries: Austria, Belgium, Croatia, Denmark, Estonia, Finland, France, Germany, Hungary, Iceland, Ireland, Italy, Lithuania, the Netherlands, Norway, Poland, Portugal, Serbia, Spain, Sweden, and the United Kingdom. Following data cleaning, the response rate was 25.3% and 9,333 valid responses were collected. This response rate is considered satisfactory in comparison with other public administration executive surveys. Because of missing data, Belgium, Portugal, and Poland are excluded from the sample.

| Dependent variable
For our first hypothesis we measure three types of managerial values (i.e., achievement/results orientation, efficiency, and innovative attitude) and use of management tools. The first two are based on answers to two statements: "I mainly understand my role as public sector executive as" "Achieving results" or "Ensuring efficient use of resources," where respondents could pick one from 1 (strongly disagree) to 7 (strongly agree) for both statements. Innovative attitude is measured by a survey item which asks to what extent the respondent agrees or disagrees with the statement, "Being creative and thinking up new ideas are important to me" (ordinal variable ranging from 1 [strongly disagree] to 7 [strongly agree]). 3 We also use public managers' personal use of performance indicators to capture their use of management tools. This variable is created as an index variable based on a set of questions assessing the extent of performance indicator use for a number of purposes: The respondent is asked to indicate on a scale ranging from "not at all" to "to a large extent" whether he or she uses performance indicators in their work to: "Assess whether I reach my targets"; "Monitor the performance of my subordinates"; "Identify problems that need attention"; "Foster learning and improvement"; "Satisfy requirements of my superiors"; "Communicate what my organization does to citizens and service users"; "Engage with external stakeholders (e.g., interest groups)"; and "Manage the image of my organization." We constructed a mean value variable, which indicates managers' preference for NPM-style management tools (Cronbach's α = .92). The variable is continuous, ranging from 0 to 7. Thus, we treat this as a truncated variable. Note that this survey item asks how much respondents, not organizations they work for, use the indicators. Thus, it measures managers' own use for management tools.
We use two indicators to measure adherence to core public values (Hypothesis 2): impartiality and equity. Impartiality is operationalized by a survey item concerning respondents' role understanding as a public sector executive and whether he/she understands his/her role as "Ensuring impartial implementation of laws and rules" (7-point scale ranging from strongly disagree to strongly agree). The survey item captures managers' attitudes toward impartiality in the exercise of public authority, one of the core concepts of Weberian bureaucracy. The equity variable captures the value trade-off between efficiency and equity (Fernández-Gutiérrez & Van de Walle, 2019). We use a survey item asking respondents to select their priorities between two different priorities: equity and efficiency. The survey asks respondents, "Public services often need to balance different priorities. Where would you place your own position?" ranging from 1 (equity) to 7 (efficiency). Thus, higher values mean higher priority on efficiency and lower values mean higher priority on equity.

| Independent variables
Manager's private sector experience is our independent variable. The COCOPS survey asks about the number of years of working in the private sector. The respondent is asked to select one from "None," "Less than 1 year," "1-5 years," "5-10 years," "10-20 years," and "More than 20 years." We combine the first two categories as a reference group as some public officials have a temporary job in the private sector while waiting for public service recruitment after graduation. The final variable is a 4-scale categorical variable: 0, none or less than 1 year; 1, 1-5 years; 2, 5-10 years; 3, 10-20 years; 4, more than 20 years. Figure 1 reveals differences across sample countries.
As seen from the figure, there are significant variations in the percentage of public managers having or not having private sector experience across countries. Countries such as France, Denmark, Germany, Croatia, Italy, Ireland, and Serbia record higher percentage of top public F I G U R E 1 Private sector experience by country. Samples are based on Model 1.1 in Table 2 managers reporting no private sector experience. On the other hand, Iceland, Sweden, the United Kingdom, Finland, Norway, and Estonia have higher number of managers who have at least 1 year of private sector experience. Some countries such as Iceland, the United Kingdom, Sweden, Norway, and Hungary record higher a number of managers having more than 10 years of private sector experiences. Figure 2 shows the distribution of managers with private sector experience by position. As seen from the figure, distributions of senior managers with different years of private sector experience do not significantly differ across positions.

| Control variables
This article controls for other factors that are expected to influence the dependent variables. Due to the relatively small number of countries included (N = 18), we do not have very many country-level controls. Our controls are limited, then, to significant factors that may influence the dependent variables, and then tested them in different models. The models include government effectiveness as a country-level control. Government effectiveness is an index variable composed from several items concerning perceptions of governance. We use the variable in year 2011 from Teorell et al. (2017). We expect that less effective government might desire the expertise and network of managers with private sector experience. Therefore, there is a possibility that less effective governments are likely to recruit more managers with private sector experience than more effective governments.
We also control for individual-level controls, including respondents' gender, position, age, public sector experience, educational level, organizational type, organizational size, levels of politicization at the organization, organizational goal clarity, and use of management tools in the organization. Note that we control for levels of management instruments used in respondents' organization. One of the dependent variables is public managers' personal use of management tools. This variable measures levels of managers' own, not organizational, use for management tools. Public managers working at more modern organizations that already use various management tools may have more chances to use such tools than managers working at organizations that use fewer. Besides that, organizations using more management tools may tend to have managers who are also competent at employing them. This may create selection bias. Therefore, in order to capture public managers' personal use of management tools, we F I G U R E 2 Private sector experience by respondents' positions. Samples are based on Model 1.1 in Table 2 need to control for levels of management tools used at the organizational level. This variable is created as a mean variable constructed from a survey item that asks the respondent (Cronbach's α = .88): "To what extent are the following instruments used in your organization?" ranging from 1 (not at all) to 7 (to a large extent) for 14 different NPM-style instruments. 4 We also control for levels of politicization in the respondents' organization. As we discussed, politicization of bureaucracy is associated with recruiting staff outside of the public sector. We control for organizational-level politicization as it may affect organizational decisions to hire managers with private sector experience and subsequently their values. Table 1 shows descriptive statistics. To check for muticollinearity issues, we conducted diagnostics using variance inflation factors (VIF) for our primary models. The results of these tests provided mean values of VIF of less than 1.75 for all primary models ( Table 2). The highest individual VIF score for individual variables is 4.16 (public sector experience). These VIF scores imply that our models do not suffer from any serious multicollinearity issues.

| Empirical strategy
Because of the hierarchical nature of our data set-including public sector managers (Level 1) nested within countries (Level 2)-we employ a multilevel analysis (Jones, 2008). Because the intercepts are assumed to fluctuate between countries because of fluctuating factors at the country level, we applied a random intercept model. 5 Among six dependent variables, five are ordinal variables, performance indicator excepted. We employ multilevel ordered logit regression models for the models with an ordinal variable as dependent variable (Models 1.1-1.3, 1.5, and 1.6 in Table 2). As for the remaining variable, we employ a multilevel mixed-effects tobit regression model (Model 1.4) because it is a censored continuous variable. Table 2 reports the results of the multilevel analysis. Results suggest that private sector experience matters for managerial values and use of management tools. Although there are no statistically significant effects for 1-5 years and 10-20 years of private sector experience on results orientation, 5-10 years and more than 20 years of private sector experience is positively associated with results orientation (p value <.05). As for efficiency orientation, 10-20 years of experience and more than 20 years of private sector experience are also positively associated with efficiency values (p < .01). We do not find any statistically significant effects of 1-5 and 5-10 years of private sector experience. The innovativeness model shows that private sector experience has a statistically significant and positive association with managers' innovative attitude throughout all categories of private sector experience. It is notable that private sector experience matters for innovative attitudes regardless of years of experience. We also find similar results in the use of the performance indicators model. Managers with 1-5 and 5-10 years, more than 20 years (p < .01), and 10-20 years (p < 0.05) of private sector experience are more likely to use performance indicators than those without private sector experience. Note that private sector experience is associated with innovativeness, and performance indicator use, regardless of years of private sector experience ranging from relatively short-term experience (i.e., 1-5 years) to long-term experience (i.e., longer more than 20 years) in Models 1.3 and 1.4. This suggests that exposure to the culture of the private sector itself can be positively associated with innovativeness and performance indicator use regardless of length of experience. The above results provide empirical support for the first hypothesis. However, with respect to the core public values hypothesis (Hypothesis 2), we do not find any statistically significant association between private sector experience and impartiality values and equity-efficiency trade-offs. Throughout all categories of years of private sector experience, private sector experience does not have any statistically significant association with managers' emphasis on impartiality. The equity-efficiency value trade-off model shows that private sector experience does not have any significant impact on the value trade-off, except for one category of private sector experience (i.e., more than 20 years). It is notable that private sector experience from 1-5 to 10-20 years does not seem to exert a significant effect on the equityefficiency trade-off. More than 20 years of private sector experience has a positive impact on equity-efficiency trade-off (p < .05). However, unlike the results of managerial values and performance indicator models, statistically significant impact is limited only to this (relatively small) category, failing to retain significance in other categories. Therefore, there is no convincing support for Hypothesis 2.

| RESULTS
Other significant results deserve attention. Gender has significant positive impacts on managerial values (i.e., results orientation, efficiency, and innovativeness) (p < .01). Compared to male managers, female managers tend to value results, efficiency, and innovative attitudes. On the other hand, gender is not associated with use of performance indicators. As for core public values, we find no link between gender and impartiality. Gender is negatively related to the equity-efficiency trade-off. However, statistical significance is found only at the p < .1 level. It should be also noted that managers' positions matter. Compared to managers in the third hierarchical position, managers in higher positions (i.e., top or second hierarchical level, such as secretaries-general or directors-general) tend to exhibit a higher efficiency orientation (p < .01) and have a high possibility of using performance indicators (p < .05 or .01). Managers in toplevel positions show more results orientation than those in third-level position (e.g., directors in a ministry) (p < .05). Results suggest that managerial positions do matter for the degree of managerial values and management tool use. Managers' positions do not matter with respect to impartiality. However, managers in top-level positions are likely to value efficiency over equity compared to those in a third-level position.
Notably, years of public sector experience have a significant positive impact on efficiency (p < .01) and impartiality (p < .05) and a negative impact on innovativeness (p < .05). Managers with 10-20 years and more than 20 years of public sector experience show more efficiency orientation than those with less than 10 years of experience. We find the same effect on impartiality. Managers who have a lengthy public service experience more heavily emphasize impartiality. On the other hand, 10-20 years of public sector experience has a negative influence on innovative attitudes (p < .05). The type of organizations for which managers work (i.e., ministry or agency/others) matters as well. Compared to managers working for a ministry, managers in agencies tend to show a lower emphasis on results, efficiency (p < .05), and innovativeness (p < 0.01). However, managers in agencies are likely to use more performance indicators than those in ministries. This result should be due to differences in the nature of tasks conducted. Also notable is that managers in agencies show lower emphasis on impartiality (p < .01) and equity-efficiency trade-off (p < .05) than those working for ministries.
The results of the ordered logit model are not straightforward without calculating predicted probabilities for each outcome. We visualize predicted probabilities of the models with statistically significant results. Note that our first three dependent variables (i.e., results orientation, efficiency, and innovativeness) are ordinal variables with "1" meaning disagree and "7" strongly agree. Figures 3-5 provide predicted probabilities of selecting each outcome of the dependent variable. The X-axis shows each outcome of the dependent variables and the Y-axis shows the predicted probabilities of selecting each outcome depending on years of private sector experience. Figure 3 shows the probabilities of having a results orientation. The mean value of this variable is 6.3 (Table 1). Thus, we consider Outcome 1-6 as a low emphasis on results F I G U R E 4 Predicted probability of efficient use of resources by private sector experience F I G U R E 5 Predicted probability of innovative attitude by private sector experience orientation and Outcome 7 as a high emphasis on results. Probabilities of selecting Outcome 7 are .68, .63, and .58 among managers with more than 20 years of private sector experience, with 5-10 years, and without private sector experience respectively. Thus, private sector experience increases probabilities of emphasizing results. On the other hand, probabilities of selecting an outcome of low emphasis on results such as Outcome 6 are .27, .25, and .22 among managers without private sector experience, with 5-10 years of experience, and with more than 20 years of experience respectively. Thus, private sector experience reduces probabilities of showing less results-orientation attitude. Figure 4 shows probabilities of selecting an efficiency orientation among managers with private sector experience. Following the aforementioned procedure, we consider Outcome 7 as high emphasis on efficiency and Outcomes 1-6 as low emphasis on efficiency (the mean value of efficiency is 6.16). The figure shows that private sector experience is associated with attitudes that value efficiency. Figure 5 exhibits predicted probabilities of innovativeness. Outcomes 6-7 show high innovative orientation, while Outcomes 1-5 show low innovative orientation (the mean value is 5.73). The results suggest that private sector experience increases the probability of selecting highinnovative attitudes and reduces the probability of low innovative attitudes. Finally, Figure 6 shows predicted values of using performance indicators. Recall that this variable is a censored continuous variable ranging from 1 to 7. Higher values mean more personal use of performance indicators. Managers with private sector experience, regardless of years of experience, show higher values of indicator use than those without private sector experience.
To finish, we conducted four robustness checks. First, we ran the models with two different categorizations of the private sector experience variable (two and three categories rather than five) to check whether the effects of private sector experience depend on cutoff points. We also reran the models with a policy area variable included, because managerial values may differ depending on the tasks these managers perform. Managers working in social policy may have different values from those working in defense, and the nature of their tasks may differ. A final robustness check consisted of rerunning the models without the British respondents to account for a skewed response distribution in that country. Compared to the other countries, the British sample is skewed toward lower hierarchical levels and the overall response rate was lower than that in the other countries.
We find that the results are robust across the different models. When using a different categorization of private sector experience some variables have lower statistical significance but the impact of private sector experience on managerial values and use of management tools is still present.

| CONCLUSIONS
Using a large sample of top public managers in 18 European countries, this article has analyzed the effect of having had private sector experience on public managers' managerial values (results orientation, efficiency, and openness to innovation), use of management tools, and support for core public values (impartiality and equity). It contributes to the literature on sector switchers by using a large cross-country data set focusing on top-level management functions in central government ministries and agencies. We find that public managers with private sector experience have, as expected, stronger core managerial values. Yet, unlike the conventional view, they do not exhibit less support for core public values. This suggests that managers in the private sector bring certain new values and ways of working to their new public sector environment, but their private sector background does not undermine impartiality or equity. This shows that impartiality and equity are not at odds with managerial values. Propagating the use of private sector management skills in the public sector, as NPM has done, therefore appears to add something to the public sector, rather than undermining core public values.
Apart from private sector experience, the length of public sector tenure also appears to have an impact on top managers' values. In particular, public sector tenure has a positive impact on both efficiency and impartiality values, but a negative one on innovativeness. The type of organization in which managers work, and in particular the ministry-agency distinction, is related to the values managers hold. But the data do not allow establishing whether managers (whether from the public or private sector) self-select in certain types of public organizations, or whether their values are shaped by the tasks performed in these organizations. A secondary finding is that private sector experience remains, overall, relatively rare among top public sector managers, although large country differences exist. In combination, the effect of self-selection and large country differences may mean that the type of private managers self-selecting into public employment is more or instead less representative of the wider private sector manager population in that country.
This article obviously has a number of limitations. It only looks at prior experience in the private sector, and therefore says nothing about the effect of sector switching itself on individual managers. We also do not know whether the switchers are typical for a private environment, or whether they switched because they felt closer to the public sector and self-selected into it. The operationalization of private sector experience was relatively rough. We do not know when exactly the private experience was, only how long it lasted. We do not have information on the respondents' immediate prior job, and whether this also was a job at a managerial level. This means that the reported private sector experience could have been immediately before their current job or a while before their new job. Also, the group of respondents with very long private sector experience (>20 years), the group generally receiving media attention, is very small, making some of the significant findings for this group less robust. Top public managers are a very visible group and can easily be identified even with limited data (type of function, policy sector, and age would for instance perfectly allow identification, e.g., the secretary-general of a certain ministry in a certain country). This means that, despite the decent response rate for a survey among a population of this kind, confidentiality requirements do not allow individual managers to be linked to organizations and that nonresponse follow-up and analysis was hindered by restrictions in how the survey could be fielded in some countries. Finally, we do not know whether the attention to managerialist values also translates to different behavior in the organization, or whether managers with private sector experience are merely better at talking the talk.

ACKNOWLEDGMENTS
We would like to thank participants at the annual conferences of the American Political Science Association, International Research Society for Public Management and the FGGA Research Seminar at Leiden University in 2018. This project has received funding from the research project 'Out of Control or Over Controlled? Incentives, Audits and New Public Management' (Riksbankens Jubileumsfond [the Swedish Foundation for Humanities and Social Sciences], Grant No. SGO14-1147:1).
ORCID Victor Lapuente https://orcid.org/0000-0001-6415-9422 Kohei Suzuki https://orcid.org/0000-0002-5403-4826 Steven Van de Walle https://orcid.org/0000-0003-1531-7097 ENDNOTES 1 The reduced version of the data set is available at https://dbk.gesis.org/dbksearch/sdesc2.asp?no=6599. We used a full version of the data set. 2 Top administrative-level public officials within central government ministries are included in the target. This comprises positions equivalent to that of secretary-general, director-general, or director. At central government agencies, the top two executive levels are included in the target. State-owned enterprises and audit courts are excluded. In federal countries such as Spain and Germany, the survey targeted appropriate regional and state government ministries and agencies. However, local governmental bodies and local service delivery organizations are also excluded (Hammerschmid et al., 2015). 3 No consensus has yet emerged on how to define and measure public sector innovation. However, there is an agreement that innovation must be related to a new idea (Demircioglu, 2019;Demircioglu & Audretsch, 2017). There are a number of studies which consider new ideas or creativity as factors associated with individual or organizational orientation toward innovation (Bysted & Jespersen, 2014;Damanpour & Schneider, 2008;de Vries, Bekkers, & Tummers, 2016;Fernandez & Moldogaziev, 2012). 4 They include business/strategic planning, customer/ user surveys, service points for customers (e.g., one stop shops), quality management systems, codes of conduct, internal steering by contract, management by objectives and results, benchmarking, cost accounting systems, decentralization of financial decisions, decentralization of staffing decisions, performance related pay, staff appraisal talks/performance appraisal, and risk management. 5 Results of the likelihood-ratio test suggest that a multilevel model is a better fit than a nonhierarchical model. Just in case we performed ordinary least squares regression with a country dummy variable (SE values are clustered at country level). Results do not change. We also find a random intercept model is more appropriate than random slope models as slope variance for the independent variable is very small.