Budgeting Practices and Performance in Small Healthcare Businesses

worry account look into 21 (2010) 4055 Contents lists available at ScienceDirect Management be interrogation journal homepage www. elsevier. com/locate/mar reckoning practises and act in sm either well(p)ness charge transmission linees Robyn King a , Peter M. Clarkson a,b,? , Sandra W aloneace c a b c UQ condescension School, The University of Queensland, Brisbane 4072, Australia Faculty of Business Administration, Simon Fraser University, Burnaby, Canada V5A 1S6 De dampenment of Accounting and BIS, The University of Melbourne, Victoria 3010, Australia a r t i c l e i n f o a b s t r a c tWe put evidence linking autochthonic health c ar credit line characteristics, ciphering be m others, and pedigree masterceeding. Based on a try of revenue rejoinders from a succeed of members of the Australian Association of physical exertion Managers (AAPM), we ? nd that itemors identi? ed by misfortune-establish inquiry ar customful for calculateing a backups cypher ing charges. Speci? c solelyy, we ? nd the acceptation of write work outs to be related to surface of it and social coordinate, and for vexationes apply indite work outs, the conclusion of intent is related to seam structure, dodge and sensed surroundal equivocalty.Finally, we ? nd evidence of a family descent amongst figureing radiation diagram and performance. Here, we initially ? nd a nones performance to be tyrannically associated with the call of pen figures. More re? ned tests of the ? t amidst business misadventure actors and intent of operational cypher de borderination harmonizely picture evidence of a validatory joining amongst the accomplishment of ? t and performance. Crown Copyright 2009 Published by Elsevier Ltd. All rights reserved. Keywords Budgeting SME health c atomic occur 18 businesses . Introduction This chew over suss outs the blood among contextual ciphers identi? ed from misfortune- found research, the conduct ion and completion of aim of cyphers, and business performance in spite of appearance the Australian firsthand quill healthc ar place scope. 1 We focus on figures beca sub bout they be apportioned to be ace of the main focussing affirm dodges This field of view is based biggerly on Robyn Kings Honours thesis padd in the UQ Business School at the University of Queensland.The authors would like to thank the editor and the ii anon. referees, as well as workshop participants at Monash and Swinburne Universities, the 2007 AFAANZ Annual Conference, and particularly Aldonio Ferreira, Axel Schultz, Shannon Anderson and Julie Walker for comments on an earlier version of the manuscript. ? Corresponding author at UQ Business School, The University of Queensland, Brisbane 4072, Australia. Tel. +61 7 3346 8015 fax +61 7 3365 6788. netmail address P. email&160protected uq. edu. au (P. M. Clarkson). Primary health rush is the initial c atomic name 18 of a patient as an outp atient excluding diagnostic testing tertiary healthc be is that provided in a hospital background knowledge. (MCS) in organisations, deal been entrap to be the earliest MCS that a business feigns, and draw out to pull together signi? sternt tutelage in the research lit and in t to each oneing material (e. g. , Davila and Foster, 2005, 2007 Sandino, 2007). We appoint the Australian primitive healthc ar sector as our experimental isthmusting both beca engross of its importance socially and economicalally, and beca role it is plausibly to be comprised of businesses that vary s deficiencyly in their budgeting directs. possibility-based research proposes that on that point is no single MCS suitable for all businesses. Instead, the suit baron of a particular MCS is palisaded to be contingent upon characteristics of a business including its surface, system, structure, and also cargons perceptions of the uncertainty of the environment inside which the business operates . We begin by examining the relationship mingled with a businesss budgeting entrusts and these four contextual work outs. In so doing, we watch out the s oftenment of a budgeting place as consisting of 2 stages, the initial finality regarding take upion and the consequent termination regarding extent of use.Here, the term adoption re? ects the decision by a business to use a ceremonious carry with to confinement its incoming 1044-5005/$ see front matter. Crown Copyright 2009 Published by Elsevier Ltd. All rights reserved. doi10. 1016/j. mar. 2009. 11. 002 R. King et al. / Management Accounting research 21 (2010) 4055 41 ?nancial performance (Davila and Foster, 2005). Alternatively, the term extent of use refers to both the number of different types of budgets the business uses and the frequency of their use. In our analysis, we develop arguments for, and confideigate, these two stages separately.We wherefore turn to consider the relationship amongst a business s budgeting practice and its performance. The comparative ? t of the businesss MCS with its calamity computes is argued to intrusion on performance, with performance increase with the degree of ? t (Chenhall, 2003). Thus, ceterus paribus, businesses exploitation a practice which does not ? t, whether by over-budgeting or under-budgeting, ar contained to experience weaker performance. We argue that not all of our precedent businesses are plausibly to exhibit shell budgeting practice because of the dif? ulties associated with identifying and implementing best practice, and the discontinuous character of upgrades (Luft, 1997). We interpret the relation between ? t and performance exploitation the method proposed by Ittner and Larcker (2001) and classi? ed as a Cartesian/Contingency approach (Gerdin and Greve, 2004). Degree of ? t is pass judgmentd as the difference between the extent of budget use and that call ined by the businesss contingency reckons. This approach as sumes that at any point, not all businesses bequeath in fact excite implemented their optimal practice.To conduct our investigation, a scripted survey of 988 members of the Australian Association of Practice Managers (AAPM) was compressn. In brief, we ? nd that considerabler, to a greater extent than de centralize healthcare businesses are to a greater extent(prenominal)(prenominal) likely to adopt written budgets. Further, for the subset of businesses that use written budgets, we ? nd that the extent of budget use is positively associated with structure (decentralisation) and strategy (cost leaders), and negatively chargedly associated with perceived environmental uncertainty ( vigour). Finally, we inventory a relationship between survival of budgeting practice and performance.Here, we initially ? nd performance to be positively associated with the use of written budgets. More re? ned tests consequently provide evidence of a positive association between the degree of ? t and performance. Our study consumes several(prenominal) contri to a greater extentoverions. First, we present evidence that contingency factors do indeed provide insights into both the adoption of budgets and the extent of their use for our take of teensy Australian principal(a) healthcare businesses. Interestingly, the results suggest that size and structure capture the businesss initial decision to adopt a formal budgeting practice.However, once a business has adopted a formal practice, strategy, structure, and perceived environmental uncertainty appear to be the primal determinants inherent the subsequent decision regarding the extent of budget use. We also present evidence that the ? t of our take businesss budgeting practices is associated with performance. To our knowledge, in that respect has been comparatively little existential evidence on this relationship documented in the literary works to date. Second, contingency-based research has predominantly been cond ucted in the large business sector.We eliminate this work by examining a half-size business setting. We argue that our setting has the advantage of allowing for an examination not unaccompanied of the extent of budget use but also of the initial decision to adopt a budgeting practice. In con- junction, it also provides an opportunity to discover more closely the different underlying theoretical constructs of size that the two most comm all utilise proxies, gross(a) fees and full-time equivalent employees, whitethorn be capturing. Finally, from a practical positioning the healthcare sector is under continuing pressure to increase its ef? iency. This study tots by examining the contexts in which the use of budgets is associated with upgraded performance in primary quill healthcare. The results should be of bene? t to both practitioners and those who advise practitioners on MCS object. The remainder of this write up is structured as notes plane section 2 presents backgrou nd material sectionalization 3 describes our experimental setting Section 4 presents the hypotheses, Section 5 the method and Section 6 the results and Section 7 provides a abridgment and conclusions. 2. Background MCS induce been de? ed as all devices and systems that businesses use to promise that the behaviours and decisions of their employees are consistent with the organisations accusings and strategies (Malmi and Brown, 2008). We focus on a businesss decision regarding its budgeting practices. Budgets are considered an MCS because they can in? uence the behaviours and decisions of employees by translating a businesss objectives into plans for action, communicating the objectives, and providing a benchmark against which to respect performance.We view budgets as both an important and appropriate focus disposed(p) over that they are considered to be one of the main MCS in organisations, pay been raise to be the earliest system adopted in startup businesses, and continu e to receive signi? cant attention in the research literature and in teaching material (Davila and Foster, 2005, 2007 Sandino, 2007). Within the MCS literature, the term budgeting is utilize to refer to a broad range of coronateics (see Chapman et al. (2007) for a review). In this study, we de? e a budget as a forward looking set of numbers which projects the future ? nancial performance of a business, and which is useful for evaluating the ? nancial viability of the businesss chosen strategy or deciding whether changes to the overall plan are prayd (Davila and Foster, 2005). Budgets puddle been identi? ed as geting a number of usances which include reservation goals explicit, coding breeding, facilitating co-ordination, promoting taradiddleability, facilitating reign, and contracting with out-of-door parties (Davila et al. 2009). Bene? ts of budgeting include increasing ef? ciency through with(predicate) mean and co-ordination, supporting both control and learning thro ugh the comparison of substantial results with plans, and more globally the ability to weave together all the disparate threads of an organisation into a comprehensive plan that serves many purposes (Hansen and Otley, 2003). prone these various roles and emf bene? ts, one might expect all businesses to adopt a formal budgeting practice.In fact, this view appears to underlie frequently of the empirical MCS research predominantly conducted in a large business setting, as it is often assumed that large businesses impart already have formal MCS that can be readily lookd (Chenhall, 2003). Here, the focus has typically 42 R. King et al. / Management Accounting Research 21 (2010) 4055 been on the adoption and performance implications of speci? c MCS innovations like activity based costing (ABC) (Ittner et al. , 2002). However, it is argued in the MCS literature that a rational adoption decision should require an military rating of the associated cost as well as the bene? s (Davila an d Foster, 2005). Costs of a formal MCS include the substantially flierd out of pocket cost associated with implementing and operating the system (Hansen and Otley, 2003 Hansen and train der Stede, 2004). Other costs that are not so easily rhythmd are the possibility that budgets create inflexibleness at that placeby limiting co-operation and creative solvent, over-emphasise short-term cost control and top down authorisation, encourage gaming, and de-motivate employees (Hansen and Otley, 2003 Hansen and Van der Stede, 2004). Further, arguably the bene? s and costs associated with adopting a formal budgeting practice exit not be the same for every(prenominal) business but leave dep culmination on business-speci? c contextual factors. Thus, ex ante, it is not clear that adoption of a formal budgeting practice is necessarily a rational outcome for all businesses. Our study distinguishes itself from the mass of the MCS literature by foc exploitation on the miniscule business setting, speci? cally the primary healthcare business sector. Within this sector, we argue that it is likely that for some businesses, the costs of budgeting ordain outweigh the perceived bene? s whereas the converse will likely be true for others. Thus, we argue that this setting provides the opportunity to gain insights into both the decision regarding the adoption of budgets as well as the subsequent decision regarding the extent of their use. We investigate the contextual factors that delineate the costs and bene? ts associated with budgeting from a contingency framework perspective. Our initial focus is on the contextual factors that differentiate adopters from non-adopters.We then consider those factors that drive the decision regarding the extent of budget use and conclude with an investigation into the relationship between the ? t of a businesss budgeting practice with its contextual factors and its performance. In adopting a contingency framework perspective, we get lai d that research referring to contingency theory has been subject to the criticism that contingency is a normal idea preferably than a theory as on that point is no a priori intuition of its own as to what the pertinent factors are and as to their likely consequences (Spekle, 2001).In this study, we do not consider contingency as a theory but rather as a framework for investigating identi? ed factors for which we have a priori intuition based on other organisational, economic and sociological theories. A supercharge criticism of the contingency-based literature is its simplistic nature of investigating one contextual factor or MCS at a time (Fisher, 1995). In the face of this criticism, some studies have begun investigating multiple contextual variables simultaneously (Hansen and Van der Stede, 2004 Cadez and Guilding, 2008). We also consider multiple contextual factors, speci? cally those identi? ed in Chenhalls (2003) review of the MCS literature since 1980. This review con? rm s environment, technology,3 structure and size as the descriptors of the fundamental generic elements of context. Strategy is also include as it emerged in the mid-eighties as an important factor that in? uences the design of MCS (Lang? eld-Smith, 1997). Finally, contingency-based research has its grow in sociology. The underlying premise of sociology is that humans are boundedly rational and satis? cing (March and Simon, 1958).Bounded moderateness can impede decision-making as not all attainable alternatives are known with certainty at a devoted point in time. One identi? ed role of MCS is to look managers in decision-making (Lawrence and Lorsch, 1967). However, the decision on the optimal MCS is itself restricted by bounded rationality, as well as the personal incentives of the manager. Thus, art object conceptually organisations whitethorn be expected to use the optimal MCS, this is not always possible. Businesses liner the same contextual factors may at that placefore c hoose different MCS, with the differences re? cted in their performance. In our study, we investigate the effect of a mismatch between the contextual variables and the extent of use of budgets on performance. 3. Experimental setting We adopt the Australian primary healthcare sector as our experimental setting for two reasons. First, we view it as an inherently interesting research setting in its own right, prone its importance both socially and economically. Second, we believe it to be an ideal setting within which to conduct an investigation into budgeting because as argued below, it is a sector within which at that place are likely both signi? ant incentives and disincentives to budget. Thus, this setting provides the advantages of the controls that burn up from working within a single industry while at the same time, one within which athletics in budget use can reasonably be expected. 4 In detail, the Australian primary healthcare sector plays a uniquely important role in dam age of both the function it provides and its place in the economy. In wrong of function, it represents the gateway through which patients most typically enter the health system. initial contact with For a sample of 57 organisational units, all with budgeting practices in place, Hansen and Van der Stede (2004) undertook an preliminary study focusing on four contextual factors (strategy, structure, environment, and size) as possible antecedents to identi? ed reasons to budget. Alternatively, for a sample of large businesses, mostly manufacturing from Slovenia, Cadez and Guilding (2008) ? nd that superior performance results from an appropriate match between the contingent factors strategy, size, and food market orientation, and strategic management accounting applications. We initially considered technology as an additive contextual variable but decided against its inclusion given our choice of experimental setting. Recent advances in checkup technology have force on diagnosti c specialties such as pathology and radiology (specialties not included in our study) to a much greater degree than primary healthcare. With the elision of three opthalmologists, the only specialists included in the survey were those that provide outpatient go from insular populate and like GPs, still largely rely on their skill and basic instruments such as the stethoscope.From an administrative and medical records perspective, the use of technology is widespread, with computers being used in 89. 8% of all GP practices and 94. 5% of all specialist practices in 2002 (ABS, 2002). 4 In this regard, Merchant (1981) suggested that a desirable extension of budgeting studies guided by contingency frameworks would be to collect data from samples chosen to magnify the chromosomal mutation on the dimensions of interest while controlling for the many possible interacting factors which obscure or distort the ? ndings. R. King et al. Management Accounting Research 21 (2010) 4055 43 the he alth system is through a superior customary medical practitioner (GP). For specialised care, patients are then referred to specialist medical practitioners. GPs and specialists provide this primary care from secret consulting rooms and refer on to other providers for diagnostic tests. 5 economicalally, the sector contributed 1. 71% of GDP in 2007 (AIHW, 2008). 6 Faced with rising costs, doctor shortages and increasing waiting times, primary healthcare businesses are increasingly under pressure to become more ef? ient (Department of health and Ageing, 2005). Here, budgeting has been identi? ed as a management accounting fauna that enhances ? nancial performance and improves ef? ciency (Davila and Foster, 2007). Further, the majority of the existing management accounting research has been focused on in-hospital care so little is known about MCS in the outpatient setting (Abernethy et al. , 2007), with the exception of a recent U. S. -based study of the relation between performa nce-based compensation and ownership of primary healthcare businesses (Ittner et al. 2007). Taken together, these facts reinforce our view that the primary healthcare sector within Australia is an important and potentially bountiful setting for the conduct of MCS research customaryly, and research into budgeting practices more narrowly. More instantly to the current investigation, we seek an experimental setting with variation in budget use, including the presence of both adopters and non-adopters. We believe that a number of factors conspire to found the primary healthcare sector a reasonable choice, ex ante.First, the small business sector has been argued to have get take aims of formal planning and control (Chenhall and Lang? eld-Smith, 1998). As such, it might also be expected to include non-adopters. 7 In Australia, primary healthcare is largely provided by private businesses own by the doctors working in the business as sole traders, in partnership, or through a company. Management of these businesses has tralatitiously been by the owner, although on that point now appears to be a trend towards the delegation of management to practice managers. 8These primary healthcare businesses typically have fewer than 50 employees and would thereby most often be classi? ed as small businesses according to the OECD de? nition (Holmes and Kelly, 1989). 9 Second, prior research also ? nds that rapid growth small-to-medium enterprises (SME) provide more extensive future-oriented ? nancial drawing than matched non-growth concerns (McMahon, 2001). Arguably, small primary healthcare businesses are less likely to be rapidly growing as there is currently an undersupply of primary healthcare workers cod to an ageing workforce and restrictions on training places. 0 These limited organic growth prospects further our expectations of ? nding non-adopters. Third, ? rms within the service sector typically do not quest to account for stock, thereby eliminating one driving force behind the use of civilise MCS. Finally, research on small and family business supports the view that necessary management skills are compulsory before planning can be initiated (Gibb and Scott, 1985). Since there is a subvert likelihood that primary healthcare owners, the majority of whom are doctors, have formal training in MCS, non-adoption is even more likely relative to other service sectors.Conversely, there are also economic incentives to adopt budgets. As noted, primary healthcare businesses are under increasing pressure to become more ef? cient. The National health Performance deputation has adopted a framework speci? cally designed for measuring healthcare system performance, with one of the identi? ed components being ef? ciency. In response, a number of institutions and private management consultants now offer education on managing primary healthcare practices that includes training on budgeting. 1 In conjunction, there is the added incentive to undertake the se education programs in that continuing education is a requirement of the accreditation process for GP practices. Further, only veritable practices can entrance fee Practice motivator Payments (PIP) from the Federal Government which can represent signi? cant additional revenue. 12 5 thither is currently no Medicare funding for diagnostic service provided within the primary healthcare setting and so, these businesses do not invest in the associated technology. The specialist practices included in the study were the private rooms of orthopaedic surgeons, opthalmologists, ermatologists, gynaecologists, and gastroenterologists. These specialists conduct initial consultation and post operative follow up from their private rooms but perform procedures in hospitals or twenty-four hour period surgeries for which they have visitation rights, using equipment supplied by the hospital/day surgery where a fee is charged right off to the patient for its use. 6 Overall, the healthcare indust ry contributed 8. 98% of GDP in 20062007, with 19% of repeated health expenditure on medical services provided by GPs and specialists. In 20062007, Medicare paid $4029. million for GP services, representing an average 4. 93 items per capita (AIHW, 2008). 7 In contrast, international studies of formal budget use have focused on large businesses, ? nding the vast majority use yearbook ? xed budgets (Horngren et al. , 2006). Australian evidence is consistent, with 97% of large businesses found to use budgets (Chenhall and Lang? eld-Smith, 1998). 8 The exact extent to which management is being delegated to practice managers is uncertain given a lack of studies into the prevalence or role of practice managers (Department of Health and Ageing, 2005). There were 9600 private GP practices operating in Australia at the end of June 2002. Of these, 68. 5% were single practitioner practices employing 2. 9 persons on average, and only coulomb employed more than 10 practitioners. At the same ti me, there were 9864 private specialist practices, 89. 7% of which were solo specialist practices employing an average 3. 2 persons and only 19. 2% had greater than 10 practitioners (ABS, 2002). 10 There has been a value unfavourable long-term trend since 1999 towards an increasing percentage of primary care practitioners aged 55 years or over.In conjunction, the World Health makeup (WHO) call ins a global workforce shortage of 4. 25 million health workers over the next decade (Cresswell, 2007). 11 The 20082009 Federal Government Budget proposal includes administered Program 5. 3Primary Care Policy, Innovation and Research which among other things, funds initiatives to improve service delivery and help GPs access current best business practice. As a part of their response, the Australian Medical Association (AMA) and the AAPM have made available speci? cally tailored business education programs for healthcare managers. 2 The Royal Australian College of General Practitioners (RACG P) standards for general practice include Our administrative staff can describe (and there is evidence of) training undertaken in the past 3 years that is pertinent to their role in our practice. The practice manager is speci? cally mentioned in the standard, as is the term practice management training. In order to access Practice incentive Payments, GP practices must have complied with the RACGP standards. In 2007, 80% of GP practices were accredited (AIHW, 2008).By way of context, an accredited practice with 44 R. King et al. / Management Accounting Research 21 (2010) 4055 However, notwithstanding, for some businesses, given their lack of size and sophistication, these incentives are unlikely to outweigh the costs of budgeting which include the initial investment in software, skills, and the added labour hours. 4. Hypothesis development 4. 1. Overview This study investigates the relationship between factors identi? ed from contingency-based research, the adoption and extent of use of budgets, and business performance. The speci? contextual factors we consider are size, structure, strategy, and perceived environmental uncertainty (Chenhall, 2003). We argue, based on how each identi? ed factor is expected to daze both a businesss postulates for and thereby the bene? ts it derives from budgeting, and also its ability to tackle the costs of a budgeting practice, that the four contingency factors play different roles relative to the two stages of the budgeting decision. Speci? cally, we predict that a businesss adoption decision primarily related to its size and to a lesser extent, its structure (decentralisation).For businesses that make the doorsill adoption decision, we predict that those that are more decentralised, employ a cost leadership strategy, and for which management perceives a lower level of environmental uncertainty will use budgets to a greater extent. Quite clearly, however, the roles contend by the various factors ultimately remain an e mpirical question and as such, we give consideration to each when we empirically model the two stages of the budgeting decision. Finally, we predict that the match between the contextual factors and extent of budget use will be re? ected in business performance. These predictions are formalised below. . 2. Determinants of budgeting practice 4. 2. 1. Size The construct of size has frequently been viewed as re? ecting two dimensions, complexity and availability of resources, with both argued to be increasing with size (Fredrickson and Mitchell, 1984 Mintzberg, 1994). spot small, single-business organisations can often be controlled with largely informal mechanisms such as direct supervision and oral parleys, larger organisations require more formal controls as the change magnitude complexity associated with a larger number of employees creates problems in social control, talk, and co-ordination (Lawrence and Lorsch, 1967).Here, Davila (2005) argues, following Levitt and March (198 8), that to regain ef? ciency in managing the organisation, co-ordination and control mechanisms are formalised with the objective of coding and documenting organisational learning and cut the demand that routine activities impose on the management teams time. Further, in term of a businesss ability to invest in a formal budgeting practice, it is widely accepted that larger businesses are better positioned given their greater resources, ? nancial and otherwise.Larger businesses not only have the resources required to acquire software and skills but they can also more ef? ciently achieve these administrative tasks through economies of surpass and the greater technical specialisation of their employees (Merchant, 1981). Chenhall (2003) ? nds that size has been considered as a contextual variable in only a few MCS studies as most examine comparatively large businesses. Such a ? nding ? ts well with Banbury and Nahapiets (1979) argument that there should only be a relationship betwee n resource availability and the introduction of formal systems in organisations of relatively small size.Consistent with these types of arguments, small business studies fall in size as in? uencing the acquisition and preparation of accounting information including budgets (Holmes and Nicholls, 1989 McMahon, 2001). Further, evidence from longitudinal studies of startup businesses suggests that size in? uences the decision to adopt operating budgets, with larger ? rms adopting the budgets earlier (Davila and Foster, 2005, 2007). They ? nd that when the business is small, control and co-ordination happens through frequent informal interactions but that the ef? iency of an informal system requiring direct contact with employees rapidly decreases with increasing size, thereby making it more ef? cient to use a formal control system. In the primary healthcare setting, we view the ? xed costs associated with the adoption of a budgeting practice to be signi? cant and thus, following the a rguments of Davila and Foster (2005), propose size as a determining contingency factor underlying the adoption decision. In this sector, businesses are required to comply with substantive red tape that places trying demands on their resources (Productivity Commission, 2003).As such, it is likely that only larger businesses have both the acquire for and the resources to devote to budgeting. Smaller practices are unlikely to be able appropriate resources away from their primary revenue generating clinical activities. Thus, we predict a positive relationship between business size and use of budgets. However, we also argue consistent with Banbury and Nahapiet (1979) that once a business has reached a critical size and uses a budget, size is unlikely to play a signi? cant further role in the determination of budgeting practice. Thus, our ? st hypothesis, express in the alternative, is H1 . The adoption of written budgets by primary healthcare businesses is positively associated with business size. 4. 2. 2. Structure The structure of a business relates to the formal speci? cation of roles for organisational members or tasks for groups to ensure that the activities of the organisation are carried out (Chenhall, 2003). musical composition two components, preeminence and integration, have been identi? ed in the literature, we focus only on differentiation because of the small size of our sample businesses. Differentiation is de? ed as the extent to which managers act as quasi-owners and is achieved through decentralisation of authority (Lawrence and Lorsch, 1967). A centralised busi- 4 FTE urban GPs would receive $60,000 per annum from PIP (Medicare, 2009). R. King et al. / Management Accounting Research 21 (2010) 4055 45 ness structure is characterised by decision-making that is restricted to owners and upper management whereas a decentralised business delegates decision-making to lower levels of management and operating(a) staff. inclined the closer links bet ween the ownership and control of the business, ecision-making in centralised businesses should require relatively fewer MCS. Herein, existing evidence reveals centralised businesses as having relatively few administrative controls and less sophisticated budgets while decentralised businesses have more formal controls (Bruns and Waterhouse, 1975 Merchant, 1981). We thus argue that structure has the potential to play a role in a businesss initial decision to adopt a budgeting practice, although we view its role as secondary to size since it is unlikely that the businesss ability to project the ? ed costs associated with a budget practice will be directly related to its structure. Further, we argue that for small healthcare businesses that have reached the doorway size and use budgets, structure also has the potential to play a role in its subsequent decision as to the extent of use. As the business becomes more differentiated, decentralisation increases and thereby so does the need for formal MCS (Lawrence and Lorsch, 1967 Merchant, 1981). Our second hypothesis, expressed in two parts and in the alternative, is then13 H2a .The adoption of written budget by primary healthcare businesses is positively associated with business structure (decentralisation). H2b . The extent of written budget use by primary healthcare businesses which opt to use written budgets is positively associated with business structure (decentralisation). 4. 2. 3. Strategy Business strategy, de? ned as how a business chooses to compete within its particular industry (Lang? eld-Smith, 1997), has been the focus of much of the research on MCS as opposed to corporate or operable strategy (Chenhall, 2003).While there are a number of different typologies of business strategy, we use Porters typology which focuses on cost leadership and harvest-home differentiation strategies. Porters cost leaders are characterised by warlike prices, consistent quality, ease of purchase, and a relatively restri cted yield selection. In contrast, differentiators offer the market something perceived as unique. Different types of MCS will be suited to different strategies due to their differing information and feedback requirements. Cost leadership strategies are argued to require speci? operating goals and budgets to facilitate cost containment at an operable level (Chenhall and Morris, 1995). Alternatively, product differentiator strategies would require more outward focussed, broadscope, MCS to collect infor- mation on competitors for planning purposes (Simons, 1987). Since primary healthcare businesses have constraints on the totality number of services they can provide such as open up hours and the number of medical practitioners, a cost leadership strategy should require that tighter cost controls be in place in order to entertain overall pro? ability. In contrast, a product differentiator strategy operating with gameyer margins under the same constraints should require fewer con trols. While we do not expect strategy to be determinative of the threshold decision to use budgets as it is unlikely to either affect the businesss ability to meet the initial ? xed costs or contribute suf? ciently to the businesss primary need for a budgeting practice, based on the arguments above we do expect it to impact on the desire to invest in marginal costs associated with a greater extent of budget use.Thus, formally, our third hypothesis, expressed in the alternative, is H3 . For small healthcare businesses, which opt to use written budgets, those following a cost leadership strategy will use budgets to a greater extent than those following a product differentiation strategy. 4. 2. 4. Perceived environmental uncertainty (PEU) PEU is de? ned as a attitude where managers perceive elements of the environment to be uncertain, with uncertainty distinguished from risk as uncertainty de? nes situations in which probabilities are not attached (Chenhall, 2003).In a general sense, PEU is seen to be an important contextual factor in the design of MCS because increased PEU makes managerial planning and control more dif? cult (Lawrence and Lorsch, 1967). PEU is, however, a general term and a number of researchers have provided more speci? c classi? cations of the environment (Waterhouse and Tiessen, 1978 Ouchi, 1979). In this study, we focus on the two most usually researched elements of PEU, the projectile nature of the environment ( vim) and the level of competition (hostility).Contingency-based research in large business has found that greater wholeheartedness is associated with a need for more externally focussed, broad scope and timely information (Chenhall and Morris, 1995). Planning becomes more dif? cult in more dynamic conditions as probabilities cannot as easily be attached to future events and controls such as static budgets may quickly become outdated. Thus, greater informal communication is required for stiff decision-making and formal controls are less bene? cial or desirable (Chapman, 1997). Alternatively, large business research focussed n hostility has found that businesses facing greater competition rely on more formal controls and emphasise budgets (Khandwalla, 1972). Thus, in addition to con? rming PEU as an important contextual factor, these ? ndings also reveal the importance of specifying the dimension of interest (Chenhall, 2003). Consistent with the research on dynamism in large business but in contrast with that on hostility, Matthews and Scott (1995) ? nd for small businesses, the sophistication of planning decreases with both increased dynamism and increased hostility.They argue from an economic perspective that for small businesses, the more uncertain 13 While it could be argued that it would be impractical for a small business consisting of a single medical practitioner and few administrative staff to have a decentralised structure, given the heavy demands of clinical work on the medical practitioners tim e, decentralisation is possible to the extent that operational and ? nancial decision-making is delegated to employees. This was con? rmed in discussions in the pilot study. 46 R. King et al. Management Accounting Research 21 (2010) 4055 the environment the less likely the manager is to expend scarce resources on budgets with an unproven effect on performance. The rational manager trying to meet the fundamental goal of making a pro? t will weigh up the bene? ts against the costs associated with budgeting. Based on these arguments and ? ndings, we propose that as with strategy, neither dimension of PEU is likely to impact the managers threshold decision to adopt a budget as it does not directly impact on the ability to meet the costs.However, for businesses that have already identi? ed the need and ability to budget, both dynamism and hostility will impact on the decision to incur the added marginal costs of increased budget use. Consistent with Matthews and Scott (1995) and the larg e organisation literature, given the relatively small nature of our sample businesses and the likely resource constraints that they face, we predict a negative association between the PEU dimension of dynamism and the extent of use. 14 However, contrary to the large rganisation literature but consistent with Matthews and Scott (1995), we also predict a negative association between the PEU dimension of hostility and the extent of use. Our fourth hypothesis, expressed in the alternative, is then H4 . The extent of written budget use by primary healthcare businesses which opt to use written budgets is negatively associated with the PEU elements of dynamism and hostility. 4. 3. Performance and budgets Budgets have been proposeed for planning, monitoring, and controlling business activities, with each thought to assist businesses to achieve pro? ability (Horngren et al. , 2006). However, the effect of budgets on pro? tability has not as yet been clearly demonstrated in the literature (M cMahon, 2001). There is evidence of a positive association between the use of budgets and performance as proxied by growth in small and medium enterprises (Gorton, 1999). even without extensive empirical evidence, planning and the use of appropriate budgets are promoted by donnishs, educators and accounting practitioners as a means of enhancing ? ancial performance (Hansen and Otley, 2003 Gorton, 1999). Thus, we might expect that primary healthcare businesses using budgets experience better performance than those that do not. More carefully, according to contingency-based research, a state of equilibrium in the relationship between the contingency factors and the type of MCS is best described by ? t (Covaleski et al. , 2003). Fit occurs when the organisation designs its practices in such a way that it has a positive impact on performance relative to alternative practices.Thus, there will be no universally effective extent of budget use, as each combination of contingency factors wi ll ? t with different practices. The positive impact on performance of attaining ? t is due to the ef? ciencies that result from using the most suitable MCS. When there is a lesser ? t between the extent of budget use and the contingency factors, performance will be impaired. Further, mis? t will be associated with lower performance irrespective of whether it arises from over-budgeting or under-budgeting. Thus, our ? th hypothesis, stated in the change form, is H5 . A businesss performance is positively associated with the degree of ? t between the extent of budget use and its contingency factors. The inef? ciencies arise because the need for a budgeting practice is incongruent with the adopted practice. If the business over-commits to budgeting, it is likely to have expended scarce human and ? nancial resources without enjoying commensurate bene? ts. Conversely, if it under-commits, its performance will likely bewilder because of control and/or co-ordination problems.To illustrat e, consider the various contextual factors identi? ed above. First, regarding size, a relatively small business that uses an extensive budgeting practice will have unnecessarily expended resources implementing and operating the practice when in fact informal communication is practical and likely preferred. Conversely, a relatively large business without a formal budgeting practice will likely ? nd both communication and co-ordination problematic given the complexity associated with a larger number of employees.In a similar fashion, a relatively centralised business with an extensive budgeting practice has likely expended resources on a level of control that is greater than required to encourage employees to make decisions that are in keeping with the organisational objectives. In terms of strategy, a cost leadership strategy requires more speci? c controls than a differentiation strategy. Thus, the adopted business strategy will likely be less effective if an incompatible budgeting practice is implemented to support the strategy. Finally, high levels of PEU make it much more dif? ult to plan with certainty, thereby reducing the need for and advantages of budgeting. Finally, notwithstanding equilibrium-type arguments, we believe that there will be variation in the degree of ? t in our setting given the dif? culties that primary healthcare businesses likely face in identifying and implementing their best budgeting practices. Given their size and medical focus, it is likely that many of our sample businesses approach budgeting from a relatively unsophisticated perspective and/or view it as a lower priority. This makes the process of identifying a practice that ? s relatively slow and involves rivulet and error type of learning. Here, Luft (1997) argues that while static equilibrium theories can predict the techniques the ? rm should end up with they cannot predict how long it will take to complete the process or what the path to the solution will be. Thus, it i s likely that there will be a lag between the need for and the use of a particular budgeting practice. Milgrom and Roberts (1992) also argue that organisations are dynamically learning and moving towards an optimal level of management accounting practice. The problem of implementing a budget practice that ? s is further complicated by the fact that implementation of new budget practices is likely to 14 The variation in PEU of managers from different industries has typically been the focus of forward research. In this study, industry is a constant but arguably PEU is still of interest, as it has also been found to vary among the managers of businesses within the same industry (Boyd et al. , 1993). R. King et al. / Management Accounting Research 21 (2010) 4055 47 occur in a lumpy fashion because when increasing the extent to which they use budgets, businesses are forced to do so in larger rather than small increments.As Luft (1997) argues, changes in information systems are often sh arply discontinuous. 5. manner 5. 1. Sample frame and description For this study, a cross-section(a) research design is used and the quantitative measurement tool is a mail questionnaire. recognised problems associated with the implementation of survey-based studies include the initial dif? culty of identifying and accessing appropriate respondents, and then of achieving acceptable response rates (Dillman, 2000). In light of this, since budgets are considered to be a traditional management tool and there is an identi? d trend towards delegating management responsibilities to practice managers, practice managers were chosen as the target subjects for the survey (Department of Health and Ageing, 2005). The further problems of contacting practice managers via a cost effective means and further participation were addressed by approaching the Australian Association of Practice Managers (AAPM) for support. The AAPM is the only recognised professional body for practice managers in Austra lia and consists of managers of dental, medical, and assort health businesses.Currently, there are 1200 members of the AAPM from medical practices, representing approximately 6% of the small private medical businesses operating in Australia. Membership in the AAPM is unpaid subject to an annual subscription fee. There are a number of bene? ts associated with social status including discounts for management education courses and national conferences. Thus, it is likely that members of the AAPM are interested in staying informed about current management trends, wish to become part of a professional network, and have the means to pay the membership fee.While no demographic data currently exist for practice managers who join the AAPM and those who do not, member businesses appear slightly skewed towards larger practices relative to the community of GP businesses. For 20052006, 83% of GP businesses had between one and ? ve GPs, with the remaining 17% having six-spot or more (PHCRIS, 2008). In comparison, for our sample practices, 66. 4% had between one and ? ve GPs and the remaining 33. 6%, six or more. However, notwithstanding this potential bias, the advantages of accessing the AAPM practice manager mailing list and having the AAPM recommend participation are considered to dominate.The questionnaire was initially developed from the existing literature and then pilot tested on a sample of 20 members of the AAPM and ? ve academic researchers. Based on this testing, a number of changes were made to wording and layout to enhance understandability in this setting. The ? nal questionnaire consisted of 35 questions presented in 10 sections and was estimated to require between 20 and 30 min to complete. Questions relating to each of the relevant constructs discussed in Section 4 were presented in dedicated and clearly labelled sections.The ? nal questionnaire was sent to a random sample of 1000 of the medical practice members of the AAPM. Of the 1000 surveys distrib uted, 12 were returned unopened. From the remaining 988 questionnaires, receipts complete and usable responses were get, representing a 14. 6% response rate which is comparable with those achieved in other studies of small businesses (Dennis, 2003). 15 request demographic data reveal 112 of the practices to be GP practices and 32 to be specialist practices, and that they employed between 2 and 42 full-time equivalents (FTE).Further, 98. 6% show up that they use computers in some capacity. The average age of the practices is 23. 94 years, and for the 114 organisations that responded to the question, their average gross fees were $1,553,919, ranging from $206,816 to $11,000,000. 5. 2. Budgeting practice 5. 2. 1. Empirical model The ? rst stage of this study seeks insights both into the factors underlying a businesss decision to adopt a budgeting practice and into its subsequent decision as to the extent of budget use.To do so, it appeals to contingencybased research to identify fou r contextual factors (size (SIZE), structure (STRUC), strategy (STRAT), and perceived environmental uncertainty (PEU)) argued to drive the decisions, although in different combinations. Given this framework, we employ the following common empirical model to formally examine each of these two decisions BUDGi = + 1 lnSIZEi 0 + 2 STRUCi + 3 STRATi + 4 PEUi + 5 TYPEi +? (1) where the various measures are described below. For the decision to adopt, based on H1 and H2a , 1 and 2 are predicted to be positive.For the decision as to the extent of use, based on H2b , H3 , and H4 , 2 and 3 are predicted to be positive and 4 negative. Practice type (TYPE) has been included in the model to control for potential structural differences (Hair et al. , 2006). Speci? cally, identi? ed differences in the pricing (higher) and supply (lower) of services by specialist versus general practices suggest that the market for specialist services may be relatively more heterogeneous (Department of Health and Ag eing, 2005). We measure TYPE as a dichotomous variable, set equal to 1 for general practices and 0 for specialist practices. . 2. 2. Dependent variable measurement To examine the decision to adopt, we measure BUDG as a dichotomous variable set equal to 1 if the business indicates, in response to an explicit yes/no question, that it 15 The mailing was restricted to 1000 questionnaires due to ? nancial constraints. normal techniques to admonish non-response were employed including a personalised cover letter from the AAPM, promised con? dentiality, transitoriness of questions, the inclusion of a reply-paid envelope, a follow-up e-mail reminder, and a promise to make results available to participants (Dillman, 2000).Testing for non-response bias, undertaken by comparing 15 responses received in the ? rst month to the ? nal 15 responses received, revealed no statistically signi? cant differences. 48 R. King et al. / Management Accounting Research 21 (2010) 4055 uses a written budget and 0 otherwise. We base this analysis on all 144 respondents and run Eq. (1) as a logistic regression. For this and all subsequent analysis, reported p-values are one-tailed since we have predictions relating to each of the contingency factors. To examine the extent of budget use, we restrict our tests to the 65 respondents using written budgets.Here, the survey questionnaire asked respondents to indicate on a 5-point Likert scale how systematically their business used operating budgets of various durations (annual, half p.a., quarterly, monthly, and/or hebdomadally), as well as gold ? ow, ? exible, rolling, long-term, or other budgets. These questions capture both the types of budgets used and the extent of their use, and are an adaptation of the alternative measurement approaches used in Chenhall and Lang? eldSmith (1998) and Jankala (2005). 16 The Cronbachs alpha is 93. 8%. Panel A of send back 1 presents descriptive statistics for the responses relating to the custom of eac h type f budget. To construct the extent of budget use measure, we apply exploratory common factor analysis with wise rotation to the responses. Two uncorrelated factors with eigenvalues of 3. 845 and 1. 234 are extracted, explaining 64. 14% of the total variance. The factor loadings are presented in the ? nal two columns of Panel A of Table 1. The ? rst factor aligns with operating budgets and the second with other types of budgets. As such, we consider two sub-categories (operating budgets and other budgets) and measure BUDG for each as the average summated budget usage score across the relevant budgets in the subcategory. 7 Here, Eq. (1) is run using OLS. Given a consistent lack of signi? cance, we do not report or discuss results based on our analyses of the other budgets measure. 5. 2. 3. Contingency factor measurement avocation the majority of contingency-based MCS studies, we measure size (SIZE) as the number of full-time equivalent (FTE) employees (Chenhall, 2003). Respond ents were asked to identify the number of FTE employees as administrative/ receipt staff, practice manager, nursing/allied health, medical, and other. Table 2 reveals that the mean (median) number of FTE employees for our sample businesses is 11. 31 (10. 500). For sensitivity purposes, we alternatively consider gross fees as a measure of size. Data on gross fees were provided by 114 of the businesses, with a mean (median) value of $1,553,919 ($1,322,359). For the remaining three contingency factors, the measures are based on responses across 7-point Likert scales to dedicated questions in the questionnaire survey. Panel B of Table 1 presents the questions and descriptive statistics for the responses. For STRUC and PEU which involve multiple questions, exploratory common factor analysis is then applied o develop the empirical measures. The factor loadings are presented in the ? nal two columns of Panel B. 18 In detail, the measure of structure (STRUC) is based on responses to six que stionnaire items asking the extent to which decision-making authority has been delegated within the business and at which level operating decisions are made. The six items, originally developed by Gordon and Narayanan (1984), have been subjected to considerable scrutiny and empirical testing for reliability and validity in previous research (Chenhall, 2003). The Cronbach alpha is 82. 7%.Application of exploratory common factor analysis to the response hit leads to the extraction of only one factor with an eigenvalue of 3. 261, explaining 53. 81% of the variance. Thus, STRUC is measured as the average summated scores across the six items. Organisational strategy (STRAT) is based on Porters classi? cation scheme (Porter, 1980) and measured by the response to a single question drawn from Govindarajan (1988). This question asks the respondents to indicate their belief as to the best description of the businesss strategic emphasis, ranging from product differentiation to cost leadership .This was found in the pilot testing to be the only question from previous research applicable to our setting, Finally, we initially measure perceived environmental uncertainty (PEU) based on the responses to nine questions developed by Gordon and Narayanan (1984) to capture the intensity of competition, the dynamic and unpredictable nature of the external environment, and the potential elements of change in the environment. Based on the correlations among the responses, only ? ve items were eventually used with a Cronbach alpha of 64. %, as the responses to four items had correlations of less than the 30% level recommended for inclusion in factor analysis (Hair et al. , 2006). 19 Consistent with previous research (Gordon and Narayanan, 1984), application of exploratory common factor analysis with incorporeal rotation led to the extraction of two factors with eigenvalues of 2. 327 and 1. 122 explaining 67. 70% of the total variance. Following the literature, we label these factors as PEU hostility (PEUhost ) and PEU dynamism (PEUdyn ).PEUhost loads on the two questions relating to the competitiveness of the business environment whereas PEUdyn loads on the three questions relating to the predictability of the external environment. We include both in our empirical model, measuring each as the average summated response scores across the relevant questions. 16 Jankala (2005) prefers this measure of systematic use as a more reliable and dead measure of a businesss commitment to the use of budgets, rather than the more subjective scales used by, for example, Chenhall and Lang? eld-Smith (1998) that measure bene? t derived. 7 The yearly operating budgets did not load on either factor as a large majority of the businesses indicated that they used yearly operating budgets on a systematic basis. 18 The reported factor loadings, eigenvalues, and percentage variation explained are based on the full sample of 144 respondents. When the exploratory common factor analysis i s repeated based only on the 65 respondent businesses using budgets, all measures are qualitatively identical. 19 The four items removed were competition for manpower, new services marketed, ability to predict preferences for customers, and change in legal, political, and economic environment.Their removal is perhaps not strike that as there was little variation in the responses received with the sample small businesses drawn from the same industry. R. King et al. / Management Accounting Research 21 (2010) 4055 Table 1 Descriptive statistics and factor loadings for survey questionnaire responses. Question Descriptive statistics Mean Panel A Budgeting BUDG 1 Which of the following budgets are prepared and how consistently? (5-point scale 1 = not used, 2 = seldom, 3 = at times, 4 = often, 5 = systematically) a. Operating budget, yearly b. Operating budget, half-yearly c.Operating budget, quarterly d. Operating budget, monthly e. Operating budget, weekly f. Cash ? ow budget g. Flexibl e budget h. Rolling budget i. long-term budget Med SD Factor loadings 1 2 49 2. 690 2. 064 2. 064 2. 092 1. 578 2. 079 1. 701 1. 704 1. 795 1 1 1 1 1 1 1 1 1 1. 899 1. 663 1. 659 1. 662 1. 315 1. 626 1. 317 1. 400 1. 473 0. 746 0. 614 0. 533 0. 668 0. 472 0. 149 0. 317 0. 168 0. 255 0. 322 0. 034 0. 346 0. 508 0. 823 0. 606 0. 715 Panel B Contingency factors Structure (STRUC) STRUC 1 To what extent has authority been delegated to the manager or employee for each of the following decisions? Please indicate actual rather than stated authority) (7-point scale 1 = no delegation, 7 = total delegation) a. Initiate ideas for new services b. Hiring and ? ring of personnel c. picking of large investments d. Budget allocations e. Pricing decisions STRUC 2 Most operation decisions are made at what level? (7-point scale 1 = owner level, 7 = manager level) 4. 999 5. 250 3. 173 3. 980 4. 311 4. 349 5 6 3 5 5 4 1. 517 1. 923 1. 963 2. 123 1. 900 1. 870 0. 531 0. 610 0. 732 0. 803 0. 761 0. 581 How would you best describe your practices strategic emphasis? 7-point scale 1 = product differentiation 7 = cost leadership) Perceived environmental uncertainty (PEU) PEU 1 How stable/dynamic is the external environment (economic and technological) facing your practice? (7-point scale 1 = very stable, 7 = very dynamic) a. Economic environment b. Technological environment PEU 2 How would you classify the market activities of your competitors (i. e. , other healthcare practices) in the past 3 years? (7-point scale 1 = becoming more predictable, 7 = becoming less predictable) How intense is each of the following in your industry, the healthcare profession? 7-point scale 1 = negligible, 7 = intense) a. Bidding for purchases b. Price competition Strategy (STRAT) STRAT 1 2. 983 3 1. 127 n/a n/a 4. 134 4. 761 3. 691 4 5 4 1. 603 1. 596 1. 122 0. 968 0. 755 0. 369 0. 176 0. 185 0. 118 PEU 3 2. 446 3. 553 2 4 1. 352 1. 502 0. 147 0. 168 0. 676 0. 625 Panel C Performance (PERF) PERF 1 W hich best describes your response to the following statements over the past 3-year period? Compared to key competitors, my practice (7-point scale 1 = strongly disagree, 7 = strongly agree) a. Is more competitive b. Has more patients c. Is growing faster d. Is more pro? table e.Is more innovative f. Has more doctors 5. 082 5. 353 5. 105 5. 210 5. 320 4. 094 5 6 5 5 6 4 1. 607 1. 619 1. 644 1. 593 1. 643 2. 204 0. 660 0. 711 0. 849 0. 667 0. 592 0. 502 For Panel A, the exploratory factor analysis was conducted based on the 65 businesses that produce written budgets. For Panels B and C, the exploratory common factor analysis was based on the full sample of 144 respondents. Items deemed to load on the identi? ed factor appear in bold. 50 R. King et al. / Management Accounting Research 21 (2010) 4055 Table 2 Descriptive statistics for contingency-based model variables.Measure Operating budgets Mean medial Standard disagreement Min max Size (SIZE) Mean normal Standard variance Min max Structure (STRUCT) Mean median value Standard deviation Min max Strategy (STRAT) Mean median Standard deviation Min max PEUhost Mean Median Standard deviation Min max PEUdyn Mean Median Standard deviation Min max Performance (PERF) Mean Median Standard deviation Min max Full sample (n = 144) 1. 948 1. 000 1. 376 15 11. 531 10. 500 6. 583 2 42 4. 344 4. 333 1. 381 17 1. 990 2. 000 1. 122 17 2. 987 3. 000 1. 199 16 4. 201 4. 333 1. 179 1 6. 6 4. 027 4. 083 1. 248 17 Written budget (n = 65) 3. 085 3. 000 1. 50 15 12. 893 12. 000 6. 367 3. 5 31 4. 862 5. 000 1. 212 1. 67 6. 83 1. 860 2. 000 1. 014 16 3. 231 3. 000 1. 183 16 4. 241 4. 333 1. 129 1 6. 67 4. 228 4. 333 1. 302 17 No written budget (n = 79) n/a p-Value 10. 410 9. 250 6. 586 2 42 3. 918 3. 833 1. 372 17 2. 090 2. 000 1. 200 17 2. 785 2. 500 1. 181 1 5. 5 4. 169 4. 333 1. 224 1 6. 67 3. 861 3. 833 1. 184 1 6. 5 0. 024 0) or under-budgets (? < 0), the impact of lack of ? t on ? nancial performance sho uld be the same. Formally, the model we employ, illustrated using ? , has the following form PERFi = 0 we also requested objective measures of pro? ability from our sample businesses (Govindarajan and Gupta, 1985). Unfortunately, less than one-third of our sample provided the data. Thus, the use of objective measures for robustness purposes is also impractical. 6. Results and analysis 6. 1. Preliminaries Descriptive statistics for measures used in tests of our hypotheses are presented in Table 2, both for the overall sample of 144 respondents and for the sample partitioned on the basis of whether a written budget is adopted, along with tests for differences in mean values between partitions. As revealed, there is considerable cross-sectional variation in each measure.Further, there appear to be several signi? cant differences between respondents adopting written budgets and those not. Speci? cally, businesses using budgets have more FTE employees (SIZE, p = 0. 024), are more decentr alised (STRUC, p < 0. 001), face a more hostile enviro