Managerial characteristics and budget use in festival organizations

Festivals are an important part of popular culture and have increased in popularity in recent decades. However, they remain relatively unexplored in the accounting literature, and understanding of the use of management control tools in this context is low. This study aims to investigate the use of budgets in festivals. Informed by upper echelons theory, it investigates how individual and observable characteristics of festival managers are associated with variations in the use of budgets. The study is based on a survey of 61 festival managers from 40 festivals. The findings suggest that festival budgets are particularly important in the planning and coordination process but used less frequently for ex post evaluations. The findings also indicate a positive association between a business educational background and the use of budgets for most purposes, with the exceptions of performance evaluation and reward. This paper contributes to the literature on accounting in popular culture in general and in festivals specifically. Through its application of upper echelons theory, it also contributes to the management accounting and control literature, showing how individual characteristics of managers influence the use of budgets.


Introduction
In recent decades, there has been an increasing number of festivals (Aagedal et al. 2009;Anderton 2008), and these are an important part of popular culture (Jeacle 2012(Jeacle , 2017. Festivals can be understood as periodically recurrent, social and open occasions that last for several days and contain a series of coordinated events with a theme that is culturally significant for the local community (Falassi 1987;Getz et al. 2010;Tjora 2016). The growth in the number of festivals means that the festival sector has increased in economic and cultural significance (Aagedal et al. 2009;Tjora 2013;Toraldo and Islam 2019). In many cases, in rural towns in particular, festivals may be necessary to make places economically viable and culturally alive (Tjora 2016). The growth in festivals also means that there is greater competition to attract a large enough audience to secure the necessary revenues. This is just one of the many uncertainties that may create economic problems for festivals, including the weather, the constantly changing cultural landscape, shifting trends and the uncertainty of resource access from both public and private actors. Hence, numerous factors are involved in any explanation of the economic conditions of festivals.
Research studies and the media have often described festivals as organizations with problems of economic control (Bergamin Barbato and Mio 2007;Carlsen et al. 2010;Frey 2000), and bankruptcies and mismanaged festivals attract considerable attention (Adressa 2016;Anderton 2008;Dagbladet 2016). For example, Anderton (2008) found that, of 160 new festival events established in 2003, almost a quarter were closed down in 2007. Common reasons for closing down were poor ticket sales, difficulties in obtaining licenses, withdrawal of sponsorship support, increased competition and unfortunate weather conditions. Festival research is not an established field (Jeacle 2012;Tjora 2016), and studies on the role of management control systems in managing festivals are scarce (Getz 2010). The present study addresses this knowledge gap by investigating how Norwegian festivals use budgets as a key element of their formalized control systems (Carlsson-Wall et al. 2017;Knardal and Pettersen 2015). Moreover, this study takes the stance that the use of control systems is not imposed on organizations, but is something that top managers can decide to change and use in different ways (Malmi and Brown 2008). A number of studies have shown that these choices are affected by firm-specific characteristics and contingencies, such as the size of the organization, environmental uncertainty, industry characteristics and firm strategy (see, e.g., Chenhall 2003).
However, firm-specific variables are not the only factors that influence how control systems are used, and there have been calls to increase knowledge of the influence of individual factors (Covaleski et al. 2006). A stream of research based on upper echelons theory clearly indicates that the individual managerial characteristics of the management group have the power to explain variations in organizations' use of control systems (Hiebl 2014). The upper echelons theory hence recognizes the influence of managers' backgrounds on management control choices (Naranjo-Gil and Hartmann 2007), and this paper questions the extent to which variations in the use of budgets in festivals can be attributed not only to festival-specific characteristics but also to individual characteristics of festival managers, such as educational background, work experience and age. Introducing a managerial characteristics perspective appears especially relevant in respect to artistic organizations, as we know that festival organizations have managers with both business educational and other educational backgrounds, as well as work experience in both the private and public sectors and within and outside the artistic industry. However, our knowledge of the impact of managers' educational background and work experience on the use of control systems is limited. Accordingly, this study investigates the use of budgets in the festival field and whether and how individual observable background characteristics of managers are associated with diverse purposes of budget use.
This study makes two main contributions. First, by studying festivals, we address calls for research within popular culture. Jeacle (2009) noted that the subject of popular culture remained underexplored by accounting researchers. Although it has gone on to attract more interest since then (Carlsson-Wall et al. 2016;Jeacle 2014Jeacle , 2017Jeacle and Carter 2012;Knardal and Pettersen 2015;Lapsley and Rekers 2017), empirical evidence about the use of accounting tools such as budgets in festivals is scarce. This paper extends the findings of Hansen and Van der Stede (2004) on how budgets are used for different purposes by investigating empirically the use of budgets in festival organizations.
Second, the paper contributes to the budgeting literature by applying upper echelons theory (Hambrick and Mason 1984;Hiebl 2014;Naranjo-Gil and Hartmann 2007). This theory is less frequently used than others to address questions of budgetary control. However, the upper echelons perspective provides the opportunity to develop hypotheses on the association between individual managerial characteristics and the use of budgets. This study therefore draws on upper echelons theory to investigate the impact of managerial characteristics (educational background, work experience and age), on top of festival-specific characteristics, on the use of budgets for different purposes. This addresses the call for studies on how and why budgets are practiced (Covaleski et al. 2006), and it will offer a more holistic understanding of how a variety of managers' characteristics affect budget use.
The paper is organized as follows. First, we develop the theory and hypotheses in Sect. 2, before setting out our research method in Sect. 3. Section 4 presents the results from the survey study and data analysis. Section 5 discusses the implications of the findings and makes suggestions for further research.

The use of budgets in festival organizations
Budgets have traditionally been taken for granted as a central part of management control systems (Cools et al. 2017;Hansen and Van der Stede 2004;Hope and Fraser 2003;Libby and Lindsay 2010). Although numerous control tools have been suggested to replace or minimize the emphasis on budgets, in most firms budgets remain an integral part of the management control system (Cools et al. 2017;Davila and Foster 2007;King et al. 2010;Zor et al. 2019). For example, Ekholm and Wallin (2000) reported that 94% of their respondents use budgets, and Libby and Lindsay (2010) found that 79% of their respondents use budgets for control. 1 Research also shows that the usefulness of budgets is perceived to be high. In Libby and Lindsay's (2010) study, 88% of the respondents agreed that budgets had the potential to be extremely useful if they were used properly, and 57% reported that budgets provided "good" to "excellent" value. These findings support the conclusions of Ekholm and Wallin (2000), who claimed that the majority of organizations use budgets and find them useful; although many alternatives to budgets have been proposed in recent decades, most companies plan to keep using them (Ekholm and Wallin 2000;Libby and Lindsay 2010).
This widespread use of budgets may be due to a variety of reasons. Budgets are a very old management control tool, and Hansen et al. (2003) noted that they are used and retained because over time they have become deeply ingrained in organizations. Otley (1999, p. 370) argued that "the virtue of the budgetary control process is that it provides an encompassing framework by means of which all aspects of an organization's activity are encapsulated into a single set of financial statements against which actual outcomes can be monitored." Covaleski et al. (2006) claimed that budgets are unique in covering most aspects of management accounting, including cost accounting, responsibility accounting, performance measurement and compensation.
A budget often plays multiple roles in an organization, and there have been several attempts to generate an understanding of the multiple purposes of budgeting. Drawing on findings in the clothing and textile industry, Chapman (1998) discussed two main roles of the budget. The first role was as a tool for pre-planning and providing accountants with the opportunity to monitor activity against pre-determined courses of action. The second role was to support managers' discussions on future actions and planning for unfolding events (in other words, managing the business forward). Noting that this latter role is very different from the more distant and retrospective role of monitoring preset actions, Chapman suggested a need for a much more involved culture. Drawing on practitioners' understandings of different budget purposes Hansen and Van der Stede (2004), extended the list to include operational planning, performance evaluation, communication of goals and strategy formation as primary purposes of budgeting, adding allocation of resources and authorization of spending as supplementary purposes.
Budgets have, over the years, been subject to substantial criticism, one of which is that they are too time-consuming. There have been claims that budgeting takes up between 20 and 40% of a manager's time (Hope and Fraser 2003;Neely et al. 2003) and, therefore, that the usefulness of the budget cannot exceed the costs related to it. The budget has also been criticized for lacking flexibility and for not being able to react to changes quickly enough to remain relevant in a changing environment. This criticism has often been related to the adverse effects of infrequent updates,

3
Managerial characteristics and budget use in festival… as traditional budgeting processes are based on annual updates (Ekholm and Wallin 2000;Hope and Fraser 2003;Wallander 1999). Libby and Lindsay (2010) partly support this criticism, as their findings indicate that budgets are only somewhat effective under conditions of change.
Accounting studies that address issues in the realm of popular culture are scarce (Jeacle 2017), and there are few studies of budgeting in festivals. However, in recent years there has been some interest in the use of budgets in popular culture and other creative settings (Bergamin Barbato and Mio 2007;Cools et al. 2017;Jeacle and Carter 2012;Knardal and Pettersen 2015;Maier 2016). Bergamin Barbato and Mio (2007) provided valuable insight by describing the development of the control system of the Venice Biennale as the festival moved from being structurally and functionally bound to the Municipality of Venice to becoming a self-governing body. The Biennale became more complex over the years in terms of the number of events, including more multidisciplinary events and events co-produced with other cultural institutions. Moreover, despite a decrease in resources, the Biennale was still expected to offer the same number of events and to reach the same artistic standards. This development required new information on financial accounting, but also an increased focus on management control. Thus, the Biennale introduced cost centers, with the aim of drawing up budgets for events, not just at an overhead level. This was a complex process, but it gave the organization the opportunity to analyze costs at the event level and to pass responsibility for the management and costs of particular events and activities to individual managers.
In a case study of a large Norwegian festival, Knardal and Pettersen (2015) found that the budget was an integrated and much used control tool. The budget was designed and used so that it fitted the specific characteristics of the festival. For example, it was divided into an operating budget and a festival budget, with the festival budget being used particularly in accordance with the long planning period and pulsating nature of the festival (Carlsson-Wall et al. 2017). The budget was therefore perceived as creating stability in changing conditions and as a tool for mediating between creativity and control in the planning of the festival. Thus, although budgets have been criticized for being time-consuming and static, studies of popular culture suggest that budgets are used in that field as a control tool. The present study therefore expects to find that festivals use budgets, albeit more as a tool for planning than for ex post use such as for performance evaluation and reward setting (Hansen and Van der Stede 2004).

Managerial characteristics and budget use by festival organizations
According to upper echelons theory, managerial characteristics can explain or predict strategic choices. The pioneers of this theory, Hambrick and Mason (1984), explicitly mentioned the thoroughness of formal planning systems, budgeting and other forms of strategic choices. This has been followed up by research in management accounting and control that investigates the role of managerial characteristics in the adoption and use of control systems (Hiebl 2014). Management accounting studies informed by upper echelons theory are normally based on two important assumptions. First, it is assumed that a manager's values and cognitive base are influenced by managerial characteristics. Values and cognitive base are difficult to observe, and in most studies they are something of a black box. In upper echelons theory, they are replaced or represented by observable characteristics such as age, work experience and educational background. However, the effect of educational background is indirect. Business administrative education may change the way managers select issues (e.g., "should we change our budget system?"), analyze alternatives and implement changes.
The second assumption is that managers influence strategic choices (for example, the design of control systems). To some extent, managers may influence the adoption of a set of control systems. Whether or not to change a control system is, however, not a day-to-day matter, and not all managers will take part in the decision. Hambrick (2007) acknowledges this and introduces managerial discretion as a moderating effect; that is, if managerial discretion is high, managerial characteristics may be more important predictors of control system changes than if managerial discretion is low.
Rather than emphasizing the adoption of innovations or changes in the control system, we might look at the way in which managerial characteristics are associated with how management accounting and control systems are used. Budgets can be taken as an example. Instead of studying the design of the budgeting system (e.g., budgeting details), we examine differences in how managers use budgets (i.e., the way in which the information provided is made use of; Otley 1978). As mentioned above, Hansen and Van der Stede (2004) noted that many studies focus on the use of budget for performance evaluation, and that there has been less focus on other budget uses. Against this background, studies by Hansen and Van der Stede (2004) and others (see, e.g., Ekholm and Wallin 2011) have investigated several other uses of budgets, such as planning, coordination, resource allocation, target setting, variance analysis and reward setting.
The upper echelons perspective assumes that a manager's cognitive base and values are informed by situational contingencies and managerial characteristics. Situational contingencies are factors such as the size and profitability of the firm; by forming both the cognitive base and managers' values, these may have a direct effect on how budgets or other parts of the control system are used. The cognitive base and values are also informed by managerial characteristics such as educational background, tenure and age. This paper draws on upper echelons theory to investigate whether and how individual characteristics influence the use of budgets for different purposes in festivals. The use of budgets can also be seen as an intermediate variable between the cognitive base/values and strategic choices, rather than as a strategic choice per se. This involves shifting the focus from the design of the budget as a control system to how the budget is used. Strategic choices can, for example, be top management turnover based on performance evaluations, investments, or outsourcing decisions to meet cost budget targets.
The above discussion hence supports the suggestion that there are differences between individual managers in how they encounter administrative situations, including the use of budgets, and that these differences are functions of managers' experiences, values and personalities (Hambrick 2007;Hambrick and Mason 1984;Naranjo-Gil and Hartmann 2007). Because personal factors such as experiences, values and personalities are difficult to observe, to develop predictions of the use of budgets in a festival setting this paper follows upper echelons theory in placing the emphasis on observable managerial characteristics (Hambrick 2007). The observable characteristics applied in this paper are educational background, work experience and age (Hambrick 2007;Hiebl 2014), and four hypotheses are developed to test the association between these characteristics and the use of budgets.
One of the observable characteristics that has been central in the upper echelons perspective is educational background (Burkert and Lueg 2013;Hambrick and Mason 1984;Hiebl 2014;Naranjo-Gil and Hartmann 2007;Naranjo-Gil et al. 2009). A person's educational background indicates a certain base of knowledge and skills, and the cognitive base is hence influenced by the skills the person has obtained (Hambrick and Mason 1984). However, in addition to the instrumental factors of knowledge and skills, education can also act as an indicator of a person's values and cognitive preferences, as the choice of educational path is made at an early age and with incomplete information about what an educational choice implies (Hambrick and Mason 1984). Burkert and Lueg (2013) claimed that the theory confirming the relationship between education and management practices builds on the two pillars of mental models and cognitive preferences. Mental models change over a person's lifetime and are adapted and shaped by learning. Business education shapes mental models to comply to with the theoretical foundations and language of business. Accordingly, the way in which managers interpret and express their environment is shaped by their mental models, which are linked directly to cognition and to the individual's structuration of observations and information. The pillar of cognitive preferences refers to the inherent and more constitutive qualities of individuals (Burkert and Lueg 2013). Hambrick and Mason (1984) claimed that individuals with personality traits that generate a preference for organizing and rationalization will self-select for studies in business. Naranjo-Gil and Hartmann (2007) found that CEOs with a primarily administrative (business-related) background were positively associated with higher use of financial information, whereas CEOs with a non-business-related background were associated with higher use of non-financial information. In this context, it is to be expected that festival managers with a business administrative educational background will be more positively associated with the use of budgets than festival managers without a business administrative education. The first hypothesis is therefore as follows: H1 There is a positive association between business educational background and the use of budgets for different purposes.
There is evidence that decision making is based not only on previous training and education but also on the patterns of knowledge gained in work experience (Carpenter et al. 2004). Hambrick and Mason (1984) emphasized tenure as an observable characteristic. Work experience is related to tenure, but as tenure is assessed by applying quantitative terms, work experience also includes qualitative aspects. A more complex picture can be obtained by acknowledging that both past and current experiences continuously develop and shape knowledge, skills, attitudes, ambitions, beliefs and behaviors (Tesluk and Jacobs 1998). Defining work experience in terms of parameters of observable background, Naranjo-Gil and Hartmann (2007) found differences in the use of management information systems between hospital managers with a clinical background and those with an administrative background. More specifically, they found that managers with a clinical background were positively associated with the use of non-financial information and negatively associated with the use of financial information. In contrast, managers with an administrative background were positively associated with the use of financial information and had no association with the use of non-financial information. An explanation for these findings is that clinicians with a dominant medical background are closer to the core activities of the hospital, and thus more concerned with making decisions that meet the needs of individual patients and promote the quality of care (Schultz et al. 2004).
Two hypotheses regarding festival managers' work experience are tested in this study, the first concerning private versus public sector work experience and the second concerning an artistic versus a non-artistic background. The first hypothesis is developed with the expectation that festival managers with work experience in the private sector are more associated with the use of budgets than festival managers with work experience in the public sector. More than managers from the private sector, public sector managers face great complexity in the form of conflicting demands from a variety of stakeholders (Metcalfe 1993), and they must in some cases deprioritize the budget in order to meet other demands. Boyne (2002) has also shown that public sector managers are less materialistic than their counterparts in the private sector and are accordingly more concerned with contributing to society than with financial achievements. These considerations lead to the first hypothesis on work experience:

H2a
There is a positive association between work experience in the private sector and the use of budgets for different purposes.
The second hypothesis on work experience mirrors the reasoning of Naranjo-Gil and Hartmann (2007) that hospital managers with a clinical background are more in tune with the core activities of the organization. In line with their findings, we expect managers with work experience from the artistic industry to be more concerned with making decisions that benefit and prioritize the artistic expression of the festival more than its budgetary goals. These considerations give rise to the second hypothesis on work experience:

H2b
There is a negative association between work experience in the artistic industry and the use of budgets for different purposes.
The final hypothesis is developed with reference to the observable characteristic of age. Age as a characteristic has received sustained attention in studies based on upper echelons theory (Hambrick and Mason 1984;Hiebl 2014;Naranjo-Gil et al. 2009).

Managerial characteristics and budget use in festival…
Naranjo-Gil et al. (2009) found a positive association between age and the use of innovative management accounting systems; younger CFOs made more extensive use of such systems. This is explained by the general observation that age and innovativeness are often negatively related, as flexibility decreases with age and rigidity and resistance to change increase. Naranjo-Gil et al. (2009) also pointed out that older managers have been subject to a more traditional accounting education and have been involved in more traditional application of management accounting tools. As the budget is one of the most traditional management accounting tools, to the extent that it is taken for granted in most organizations (Hansen and Van der Stede 2004;Libby and Lindsay 2010), it can be argued that older festival managers can be expected to make more comprehensive use of a traditional management accounting tool such as the budget.
The last hypothesis to be tested is therefore as follows: H3 There is a positive association between age and the use of budgets for different purposes.

Sample
The festivals in the dataset were identified from several online sources, in particular, membership lists from the two largest interest organizations for festivals (Norway Festivals and Norwegian Live Music Association). An initial search identified 53 festivals with more than MNOK 5 (approx. EUR 461 000) 2 in total revenues. These were all festivals that took place within a pre-defined period and had major similarities in how they were organized. E-mails were sent with a short explanation of the study and a question asking them to confirm if they were interested in participating. Festivals that did not respond to the e-mail or the reminder e-mail were contacted by telephone. Most of the festivals in the study are music festivals. However, other types of festivals, such as film and food festivals, are also included in the data set. We checked for differences between music festivals and the other types of festivals, but found no differences.
The target respondents for the study were senior festival managers. Each festival manager's rank in the organization was identified via the festival's web page. The highest ranking managers were chosen, since in most festivals they are likely to have the most knowledge and can influence both the design and use of the management control system. As the number of relevant festivals was low, we tried to increase the number of respondents, and hence the statistical power of the study, by identifying two respondents in each festival. However, in some cases this was not possible, as the festival administrations are small and do not include several employees with relevant knowledge about the use of the budget. The identification of employees with relevant knowledge and the exclusion of other employees was carried out using the job titles of the employees in each festival. Employees with titles such as Artistic Director or Producer were considered to have limited knowledge about the use of the budget and were therefore excluded as respondents.
Four festivals that withdrew from the study and one that went bankrupt were removed from the dataset. The total target population was thus 74 respondents from 48 festivals. In all the festivals with two respondents, there was a variation in managerial characteristics within the festival. This presented a methodological challenge, as the use of two respondents from one festival may increase variation in managerial characteristics but may not increase variation in the use of budgets. This may lead to type II error, whereby the effect of differences in managerial characteristics is underestimated. To control for this problem, a dataset with only one respondent (the first to answer) from each festival was applied. Although similar results (the same significant variables) were obtained, we still cannot exclude the possibility of type II error in our analysis. The method of selecting the sample hence provided a non-probability sample (or convenience sample) ( Van der Stede et al. 2005). A non-probability sample is less desirable than a probability sample, and conclusions from the sample can only be viewed as indicative. However, in the current case, a non-probability sample was the only option; the population was very small and, in addition to obtaining a high response rate, our focus was on increasing the strength of the findings by targeting the most appropriate respondents.
Data were collected through an online questionnaire (see Appendix 1) during spring and autumn 2015, and to increase the construct validity the survey was pretested on a person from one of the large festivals and on academic colleagues (Van der Stede et al. 2005). A total of 61 respondents from 40 festivals completed the survey, a response rate of 82% (at the individual level). Diamond (2000, p. 239) suggested that a response rate between 75 and 90% normally yields reliable results, and the current study is within these limits. Only 18% of the targeted population did not reply, which reduced the problem of non-response bias. Moreover, a simple test of differences between the first 10 and the last 10 respondents did not indicate a nonresponse bias. Question 4 in the survey was a yes/no question as to whether the festival uses budgets, and the questionnaire was programmed to stop respondents who answered 'no' to this question from answering the remaining questions. However, all the respondents answered 'yes' to this question.
Financial data was collected from the Brønnøysund Register Centre, a government body responsible for publicly registering financial reports from all Norwegian enterprises, and builds on reports from 2012 to 2014.

Use of budgets
The use of budgets for different purposes was measured using seven items on a Likert scale, ranging from 1 (not at all) to 5 (to a very large extent). The seven items were based on the following dimensions of budget use (Ekholm and Wallin 2000; Hansen and Van der Stede 2004): planning, coordination, resource allocation, target setting, variance analysis, performance evaluation and reward.
A factor analysis was conducted to investigate whether the observed variables were related and to create compound variables for use in the statistical analysis. The results are shown in Table 1.
The factor analysis resulted in the creation of two composite factors. The first factor includes the items planning, coordination, and resource allocation and is labeled PCRA . The items included in this factor reflect an ex ante use of the budget. The second factor includes the items performance evaluation and reward and represents an ex post use of the budget. This factor is labeled PER. Target setting (TS) and variance analysis (VA) did not load into the two factors and were applied as individual measures. Target setting represents the use of budgets for setting targets for individuals and units, and variance analysis concerns the control and monitoring of actual vs. budgeted performance, both during the planning of the festival and after the festival has ended. Cronbach's Alpha values are 0.808 for PCRA and 0.801 for PER, indicating that the internal consistency of the composite measures is good. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy gives a value of 0.740, indicating that factor analysis was appropriate. The descriptive statistics for these variables are provided in Table 2.

Managerial characteristics
The independent variables in the study are business educational background, work experience in the private sector, work experience in the artistic industry and age. Business educational background was measured using an open question in the survey which asked the respondents to describe their educational background. Both the authors later systematized these responses into a dummy variable, where 0 denotes no business educational background and 1 denotes a business educational background. Work experience was measured using the same open question in the questionnaire. The respondents described their professional background, and this was later categorized by both the authors into the two dummy variables. Work experience in the artistic industry can be experience as a performing artist and/or experience within management. There was agreement between the authors on the codification. Age was measured as a continuous variable by an open question in the questionnaire. The descriptive statistics of the managerial characteristics are provided in Table 3.

3
Managerial characteristics and budget use in festival…

Control variables
The control variables in the study are size, profit margin, and public support ratio. The control variables were based on financial accounting data from 2012 to 2014. Size was chosen as a control variable, as large festivals have more opportunities to employ specialized people in both administrative and artistic functions. Size was measured as the average total revenues from 2012 to 2014. Profit margin was included as a control variable. Profit margin can influence the use of control mechanisms such as budgets, as it is expected that the need for tight budget control may decrease as profit margins increase. Profit margin was measured as operating profit/ total revenues. The final control variable, public support ratio, was included for the same reason as profit margin. Public support refers to funding received by the government, and high public support may provide a festival with a good financial environment and hence reduce its focus on the use of budgets. The public support ratio was calculated as public support/total revenues. Both profit margin and public support ratio were based on an average of the years 2012-2014. Descriptive statistics for managerial characteristics and control variables are shown in Table 3.

Empirical results
This section presents and discusses the empirical findings. First, the descriptive statistics for all the variables are given, and then we present the results of the analysis of the influence of managers' characteristics on the use of budgets for different purposes. Table 2 shows the descriptive statistics for budget use. The use ranges from 1 to 5, with 1 indicating that the budget is not used at all and 5 that it is used to a very large extent. Table 2 shows that budgets are mostly used prior to the festivals, for planning, coordination and resource allocation (PCRA), with a mean of 4.19, and target setting (TS), with a mean of 3.23. Budgets are also used ex post to analyze variance between budgets and actuals (VA) (mean = 4.34), but less often for performance evaluation and reward (PER) (mean = 2.04). These findings show a contrast in budget use with other types of organizations. Hansen and Van der Stede (2004) claimed that most studies have focused on the use of budgets for performance evaluation and have more or less neglected other uses, such as operational planning. In a study of Danish companies, Sandalgaard (2013) found that budgets were often used for performance evaluation, and Olsen (2012) also found that budgets were used for performance evaluation in both higher education institutions, health organizations and banks. These studies found notably higher levels of use for performance evaluation than this study.

Descriptive statistics
The age range is from 25 to 62 years, with a mean age of 45 years. Most festival managers had no business educational background (63.9%). Of the respondents, 60.7% had work experience in the private sector and 39.3% in the public sector. Relatively few had work experience in the artistic industry: 11.5% compared to 88.5% with no work experience in the artistic industry. For the control variables, size (measured as total revenue) ranged from MNOK 5.4 (approx. EUR 497,000) 3 to MNOK 59 (MEUR 5.4). 4 The net profit margin ranged from − 13.6 to 18.2%, with a mean of 1.7%, and public support ranged from 0 to 91.7%, with a mean of 31.5%. Table 4 shows the Pearson correlations between the variables. All dependent variables (different uses of budgets) are positively correlated, indicating that managers who use budgets for one purpose are more likely to use them for other purposes as well. For the independent variables, managers with a business educational background are associated with more work experience from the private sector. This is not surprising, since students of business administration mainly start working in the private sector. The level of public support is positively associated with managers with work experience in the artistic industry, and negatively associated with work experience in the private sector and with profit margin. This may indicate that public support increases with financial stress (negative profit margins), and that managerial characteristics influence public support; however, the direction of cause and effect is not obvious.

Hypothesis testing
Tables 5, 6, 7 and 8 show the results of the regression analyses for the influence of the characteristics variables on the use of budgets for different purposes.
The R 2 values show that the regression models better explain variations in the use of budgets for planning, coordination, resource allocation (PCRA) and variance analysis (VA). These purposes are typical financial planning and control activities. It is also these activities that have the highest levels of use, with mean levels above 4, albeit with significant differences (see Table 2). For performance evaluation and reward (PER), and for target setting (TS), the levels of use are lower and may thus be seen as less important. The R 2 values indicate that the model explains a minor part of the variation in the use of budgets for these purposes. Thus, the results show that individual characteristics make a difference for budget-related activities, but not for all types of uses. The effects of different individual characteristics were tested using a two-tailed test in the regression analysis. Findings are shown in Tables 5, 6, 7, and 8.
Hypothesis H1 investigates the association between business educational background and the use of budgets for different purposes. H1 is supported for all purposes, although not significantly on a two-tailed test for performance evaluation and 1 3 Managerial characteristics and budget use in festival… The strong support for business educational background is consistent with studies showing that managers with that background are more confident using financial measures than managers from a non-business-related field (Finkelstein and Hambrick 1996;Naranjo-Gil and Hartmann 2007;Naranjo-Gil and Hartmann 2006). However, in contrast to prior studies, the findings suggest that managers with a business educational background are not more likely to use budgets for performance evaluation and reward (Hansen and Van der Stede 2004). The next pair of hypotheses (H2a and H2b) examines the association between work experience and the use of budgets for different purposes. H2a expects a positive association between work experience in the private sector and the use of budgets for different purposes. The findings show a negative association, but not at a significant level. For work experience in the artistic industry, H2b expected the opposite, that is, a negative association between a more artistically creative background and the use of budgets. Again, the results were not statistically significant. These findings contrast with research by Naranjo-Gil and Hartmann (2007) and the theoretical underpinnings suggested by Schultz et al. (2004). H3 expects a positive association between the age of festival managers and the use of budgets for different purposes. The findings show no indications of an association, and H3 is not supported. Thus, the findings contrast with the implications of prior studies (see, e.g., Naranjo-Gil et al. 2009).
The analysis of the effect of the control variables shows that size is significantly positively associated with the use of budgets for planning purposes. The findings also suggest that profit margins are significantly negatively correlated with the use of budgets for planning purposes.

Discussion and concluding remarks
This paper investigates the use of budgets in Norwegian festivals. The point of departure is that festival managers can influence how budgets are used, and that budget use is a result of festival managers' cognitive base and values. Upper echelons theory was therefore applied to investigate the effect on the use of budgets of the observable background characteristics of educational background, work experience and age (Hiebl 2014).
The descriptive statistics suggest that budgets are used prior to festivals for the purposes of planning, coordination, resource allocation and target setting. Budgets are also used for variance analysis and, to a lesser extent, for performance evaluation and reward. These results elucidate one of the important characteristics of festivals, shared by other pulsating organizations: the importance of planning in terms of a short and intense execution phase (Carlsson-Wall et al. 2017). Accordingly, the findings suggest that the budget is a tool for providing the necessary a priori stability (Knardal and Pettersen 2015), and that ex post use of budgets (for performance evaluation and reward) is less common for festivals. Hansen and Van der Stede (2004) claimed that this has traditionally been an important purpose of budgets, and their view was supported by, for example, Olsen (2012). The contrary findings in this study may be related to the specific characteristics of festivals. Festival production requires collaborative team effort; for evaluation and reward purposes, festival performance or team performance may therefore be more relevant than employee performance. Another explanation may be that festival organizations use tools other than budget for performance evaluation and reward. However, little is known about these aspects, and further research is necessary.
In terms of the four hypotheses of this paper, the findings suggest that business educational background is positively associated with the use of budgets for most purposes, with the exception of performance evaluation and reward. H1 is thus partly confirmed. As the upper echelons theory assumes that managers' cognitive base and values are influenced by educational background, it is therefore expected that managers will have an impact on the use of control systems (Hambrick and Mason 1984). In this context, it is not surprising to find that a business educational background increases the use of budgets in festivals for most purposes.
It was, however, surprising that a business educational background does not increase the ex post use of budgets for performance evaluation and reward.
This study also investigated the association between work experience and the use of budgets. Following Hambrick and Mason's (1984) suggestion that tenure is influential on cognitive base, we added work experience as a factor that develops and shapes knowledge, skills, attitudes, ambitions, beliefs and behaviors (Naranjo-Gil and Hartmann 2007;Tesluk and Jacobs 1998). However, the study found no significant evidence of the influence of work experience on the use of budgets, and consequently H2a and H2b were not confirmed. However, in both cases, the signs were the opposite of what we expected (albeit the results were not significant). The findings are therefore not consistent with prior research on the association between work experience and the use of management accounting (Metcalfe 1993;Boyne 2002;Naranjo-Gil and Hartmann 2007). The low number of respondents with experience from the artistic industry (7/11.5%) may explain the insignificance of the result; however, the sign is in the opposite direction for all parts of the analysis. This may prompt a rethinking of the supposed association between artistic background and "lack of control." In terms of private sector background, the results may indicate that budgets are different from the management tools investigated by previous studies (e.g., Naranjo-Gil et al. 2009). There may be an experience effect from particular systems in the private sector, although budgets are a public sector tool as much as a private sector tool. Another explanation may be that the difference between private and public sector managers is not relevant to the background effect of managers in festivals; in other words, the typical private sector manager may not take a job as a festival manager, or may act differently if she does.
The last hypothesis tested concerns the association between age and the use of budgets, with an expectation that older festival managers would make more extensive use of budgets (Naranjo-Gil et al. 2009). The analysis showed no significant association between age and use of budgets for any of the budget purposes, and H3 was therefore not confirmed.
The analysis of the effect of the control variables (size, profit margin and public support ratio) shows a significant positive association between size and the use of budgets for planning purposes. This may be related to the fact that larger festivals need a relatively high degree of planning in general, including financial planning. The findings also suggest that profit margins are negatively correlated with the use of budgets for planning. This could be explained by festival managers perceiving the need for budgets and tight budget control as lower when profit margins are increasing. Conversely, the need for budgets may increase when profit margins are low.
The current study makes two main contributions. First, it improves understanding of the use of management control tools within the realm of popular culture (Jeacle 2009(Jeacle , 2012(Jeacle , 2017 and, more specifically, for festivals, a relatively neglected field within the management control literature. The findings suggest that budgets are used to a rather large extent in festivals, in particular for planning and variance analysis prior to festivals. Since this contrasts with studies in other sectors that have found a high use of budgets for performance evaluation (Hansen and Van der Stede 2004;Olsen 2012;Sandalgaard 2013), the findings help to develop an understanding of the use of budgets in this specific sector.
Second, the paper contributes to the management control and budgeting literature by applying upper echelons theory (Hambrick and Mason 1984;Hiebl 2014;Naranjo-Gil and Hartmann 2007). Upper echelons theory has not often been used to address questions of budgetary control, but this theoretical perspective provides the opportunity to develop a better understanding of the association between multiple managerial characteristics and the use of budgets. The findings identify a positive association between business educational background and the use of budgets for most purposes. This is as predicted, as festival managers with a business educational background are expected to be more confident and familiar with the use of accounting tools, including budgets. It is more surprising, however, that the findings on the use of budgets for performance evaluation and reward are not significant, in contrast to prior studies. It is also surprising that work experience is not associated with the use of budgets for different purposes. We expected that higher age would increase the use of a traditional accounting tool such as the budget, but in this area no association was found. Although the results for H2a, H2b and H3 were not significant, this paper addresses the call for studies on individual characteristics (Covaleski et al. 2006) by providing a more fine-tuned and holistic understanding of festival management and festival managers' characteristics and how these can affect the use of budgets.
The limitations of this study should be borne in mind when interpreting its results. Some of the limitations are inherent to the survey method ( Van der Stede et al. 2005), such as the possibility of a respondent's answers being biased. Podsakoff et al. (2003) warned that respondents may, for example, have a consistency motif, that they may respond in accordance with what they perceive as socially acceptable or that they may be influenced by their state of mood. Another source of potential bias is that the present study captures the perception of budget use rather than actual budget use; however, measuring perceptions is unavoidable here, since it is the users (the festival managers) who use the budgets. A specific limitation in this study is the relatively small number of respondents, as a larger dataset would have increased its statistical power. Unfortunately, it was not possible in this case to collect more data; nonetheless, the response rate is high, and against this background the study sample can be considered representative of the population. A further limitation is that the dataset includes two respondents from some of the festivals in the sample. There is therefore a risk of type II error, whereby the effect of differences in managerial characteristics are underestimated. Finally, the survey was conducted in Norway with a relatively small sample, and this may limit the generalizability of the findings.
Despite these limitations, this study provides valuable insights, but more research is needed. The findings in this paper suggest that the use of budgets in festivals is different from many other sectors. Future research should explore the use of budgets and other accounting tools in festival contexts, but also within the field of popular culture in general (Jeacle 2012). The upper echelons theory provides an opportunity to study how individual managerial characteristics can explain strategic accounting choices, and this study shows that it is not only sector-specific variables that influence the use of budgets but also festival managers' characteristics. Future research should therefore investigate further the role of managerial characteristics in the adoption and use of management control systems.
This study indicates the potential benefits of having managers with a business educational background, since they use budgets more and in different ways. Marginson and Ogden (2005, p. 435) have argued that budgeting offers "a source of structure and certainty" to practice, and more structure may be helpful for festivals, as demonstrated by Knardal and Pettersen (2015). An alternative would be to increase the knowledge of constructive use of budgets by training festival managers.
However, we do not know the best mix of qualifications in a management team. Future studies should develop a better understanding not only of individual managers' characteristics but also of the mix of individual characteristics in top management teams (Hambrick 2007;Hiebl 2014;Naranjo-Gil and Hartmann 2006). This study shows that background matters, but we do not know much about the dynamics of differences, and a more process-orientated view may not fit with a survey design. The question remains as to how managers with different backgrounds interact in the decision-making process, and more in-depth case studies are required to provide greater insight into this issue.
Funding Open Access funding provided by NTNU Norwegian University of Science and Technology (incl St. Olavs Hospital -Trondheim University Hospital) Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.
No. 5) Enter the degree to which budgets are used for the following activities: Not at all-to a very limited degree-to a small degree-to a large extent-to a very large extent (1-5).
Planning (create a comprehensive plan for the festival).
Coordination (coordination of the various processes within the festivalassigning responsibility).
Resource allocation (prioritization of program posts and other priorities). Communication of goals (setting goals for individuals and units/departments within the festival).
Variance analysis (during the budget period and/or after the budget period (after the end of the festival)).
Performance evaluation (employee performance is measured by how well they have met certain budget targets).
Reward (employees are rewarded on the basis of budget targets).