Professional norms, public service motivation and economic incentives What motivates public employees? Lotte Bøgh Andersen, University of Aarhus, paper for Study Group III, EGPA 2007 Abstract The theories of professions, public service motivation, and economic incentives explain the behaviour of the producers of publicly financed services differently. They emphasize professional norms, sector, and economic incentives, respectively. The few existing attempts to integrate these theories have, however, indicated that these factors interact. Using interviews, surveys and registers, the paper investigated how professional norms, economic incentives and sector affected the behaviour of Danish dentists and physicians. It was found that when strong professional norms existed, economic incentives were unimportant for both public and private employees. In contrast, when no firm professional norm applied, economic incentives affected behaviour. Controlling for different economic incentives, sector does not seem to affect the behaviour much. The results imply that the economic and professional perspectives should be combined in the analysis of behaviour among health professionals. Introduction The motivation of the producers of publicly financed services is important due to its major consequences for the quality and costs of the services. If general practitioners reduce visits, and if hospital physicians try to limit their time with the patients, the quality of services suffers. Likewise, the costs of the services accelerate, if private dentists only think about fees, and if public dentists systematically try to maximize their budgets. Simple economic incentive theory would expect this. Fortunately, the producers of publicly financed services are not only motivated by leisure and remuneration. There is, however, little consensus on the contents and causes of non-economic motivation. In the literature on public service motivation (PSM), it has been argued that motivation depends on the sector. Specifically, public employees are expected to have more altruistic motives than private employees (Rainey 1982; Perry and Wise 1990; Crewson 1997). In other words, PSM expects motivation (and therefore behaviour) to differ between sectors (Wright 2001). In contrast, the sociology of professions (Mosher 1968; Roberts & Dietrich 1999; Freidson 2001) argues that the relevant values (professional norms) are developed and sanctioned within the occupation. Professionals from the same occupation are expected to share motivation regardless of sector. The PSM and the sociology of professions thus disagree about the relative importance of sector and occupation, but both are value-oriented departures from the simple economic incentive theory. 1 However, economic incentives cannot be totally dismissed. Under some circumstances financial incentives can induce public employees to increase their work effort (Burgess, Propper & Wilson 2002; Meier & O´Toole 2002). For example, health care professionals produce more when they are paid on fee-for-service rather than on a fixed salary basis (Krasnik, Groenewegen, Pedersen et al. 1990; Donaldson & Gerard 1993; Taylor-Gooby, Sylvester, Calnan et al. 2000; Gosden, Forlan, Kristiansen et al. 2001). This indicates that more than one type of motivation is important. Thus, the interesting issue is how economic, professional, and public service motives interact. Few attempts have been made to integrate the findings from the PSM, the sociology of professions and the economic incentive theory. Investigating cesarean section surgeries in hospitals with different ownership structures, Goodrick and Salancik (1996) found that economic incentives – for both public and private employees – are unimportant when strong professional standards are in place. In contrast, when no firm professional norm applies, they found that economic incentives affect behaviour. These findings are very interesting, but we need more empirical evidence to determine the general relationships between behaviour and different types of motives. This paper investigates how professional norms, economic incentives and sector affect the behaviour of Danish health care producers. Differentiating between situations with and without professional norms within the same occupations, the paper investigates the motives and behaviour of dentists and physicians from the private and the public sector. I have interviewed public physicians with and without incentives to increase the production, and the incentives to induce demand also varied between the interviewed private physicians due to their different list sizes. The interviewed private dentists had an incentive to prefer treatment over prevention, while this choice did not have economic consequences for the interviewed public dentists. In addition to the qualitative interviews, a survey of dental workers has been conducted. In both interviews and survey, the health care producers were asked about services governed by firm professional norms and services without professional norms. The interviews and the survey provide good descriptions of norms and motives, but the results cannot be generalized statistically, and the action indicators are rather subjective. This calls for more general, objective indicators. Fortunately, health care registers cover the entire Danish population and contain information about both treatments and prevention. This enables me to analyze how the motivational differences identified in the interviews affect behaviour. The register analysis differentiates between private and public employees, between producers with different economic incentives, and between services with and without firm professional norms. The paper thus brings all three theories into play. After a discussion of the three theories, the paper puts forward its theoretical model of motivation and behaviour and proposes six hypotheses. Next, it presents the research design and methods. The analysis is structured around the hypotheses. Finally, the paper discusses the findings and draws conclusions. 2 Theoretical framework Publicly financed services are often characterized by output and outcome which is difficult to observe (Wilson, 1989). At the same time, the line of command is seldom simple, and each public employee must often satisfy multiple principals (Miller, 2005). If public employees only tried to increase their remuneration and decrease their workload, we would expect a serious agency loss (ibid.). Other selfish motives (e.g. budget maximization (Niskanen, 1971) and bureau shaping (Dunleavy, 1991)) have been suggested, but these motives can hardly stand also. Public employees probably have, like Downs (1961) claimed more than 40 years ago, have both selfish and unselfish motives – and the mixture might vary from situation to situation. Although studies of public sector employees’ motivation are plentiful (Wright 2001), no consensus exists on this mixture between selfish and unselfish motivation – or as Le Grand (2003) would say – between the situations in which public employees are knight and knaves. The literature on public service motivation (PSM) has tested whether public employees have a special motivation (different from private employees), but the concept of ‘public service motivation’ is not unequivocal in the literature. Rainey (1982) equates PSM with a wish to ‘engage in meaningful public service’. Crewson (1997) defines PSM as intrinsic work attributes like a sense of usefulness to society and a wish to help others. Gabris and Simo (1995: 35) conceive PSM as a ‘strive to harness the technical efficiency of bureaucratic organization for the purpose of advancing democratic values’. Perry and Wise (1990) seek to clarify the nature of Public Service Motivation and argue that PSM may be grounded in three ways: Rationally, affectively and based on norms. The rationally motivated public sector employee is drawn to government to make good policies or to advocate for special interests. The affectively motivated public employee personally identifies with a public program and is genuinely convinced of its social importance. Finally, PSM may be conceived as a norm-based desire to serve the public interest or as a sense of loyalty to duty and government. Despite the different understandings of PSM, most of the literature agrees that PSM exists in the public sector, and that higher levels of PSM are found in the public sector than in the private sector (Wright 2001). Public sector employees are – or consider themselves to be – committed to provide good and adequate service to the community and their clients (Steele 1999; Graham & Steele 2001; Vrangbæk 2003). This may also be the case when the aim of providing adequate and high quality services conflicts with budgetary constraints (Crilly & Le Grand 2004). In this sense, public employees are knights (Le Grand, 2003). 3 This does not, however, mean that financial rewards for performance do not affect public employees. Perry and Wise (1990) write that ‘public organizations that attract members with high levels of PSM are likely to be less dependent on utilitarian incentives to manage individual performance effectively’ (Perry & Wise 1990: 371). This is based on Rainey’s (1982) comparison of public and private managers, where public managers perceived the relationship between performance and extrinsic rewards to be weaker than private managers did. However, Rainey (1982: 297) cautions that financial rewards are also highly valued by many public employees. Even though Crewson (1997) concludes that public employees are less likely than private to be driven by financial rewards, his empirical findings do not indicate that there is a significant difference between public and private sector employees’ wish for higher pay. In fact, his results show that wages are a very important motivator for eight of ten employees, both in the public and private sectors. Gabris and Simo (1995), who criticize the notion of PSM, also find that wages are an equally strong motivation for public and private sector employees. Based on a score of articles Perry (1988) concludes that employees are motivated by the merit pay system. Merit pay has, however, failed to establish a clear link between pay and performance, and it has not improved the performance of government agencies. Contrary to this evidence, other studies show that financial incentives induce public employees to increase their work effort (Burgess, Propper & Wilson 2002) and that higher paid public employees produce better organizational outcome (Meier & O´Toole 2002). Specifically, a number of health care personnel studies suggest that these employees produce more when they are paid on fee-for-service rather than on a fixed salary basis (Krasnik, Groenewegen, Pedersen et al. 1990; Donaldson & Gerard 1993; Taylor-Gooby, Sylvester, Calnan et al. 2000; Gosden, Forlan, Kristiansen et al. 2001). Still, under certain circumstances financial incentives seem to be totally unrelated to the behaviour of health care employees in one particular case: When strong professional norms prescribe a given behaviour. Professional norms are prescriptions commonly known and used by the members of an occupation. These prescriptions refer to which actions are required, prohibited or permitted in specific situation (Andersen, 2005: 71-73; Ostrom 1986: 4). The existence of such norms is a vital part being a profession. Although the sociology of professions has reached agreement on one definition, most of the literature (e.g. Freidson, 2001; Robert & Dietrich, 1999; Goode 1969) agrees that a profession (at least) has a specialized theoretical knowledge and strong professional norms (Andersen, 2005: 22). The theoretical connection between these two variables can be modelled as 4 follows (the argument draws on Andersen, 2005: 22-31): Assuming that some services require a special expertise, an asymmetry in information between providers (the relevant occupation with this expertise) and users of these services exists. It is hardly controversial to claim that if the expertise is based on specialized knowledge, this asymmetry increases, and the same is probably the case for theoretical knowledge, because such knowledge cannot be codified and thereby controlled (Freidson, 2001). Especially if significant externalities are involved in the transactions, broader social constraints are warranted on the behaviour of the relevant occupation (Roberts & Dietrich 1999: 989). The public can accomplish this by entering into a professionalism contract with the occupation which thereby gains professional status (recognition of their knowledge and high standards/norms). In such an implicit bargain, the profession promises to keep their house in order, that is, to uphold a certain standard. In return, the profession can enjoy a higher status and receives higher pecuniary rewards (Day og Klein 1987:19; Watson, 2003: 192). Still, the profession (as a collective actor) must uphold the promised standards. Despite the information asymmetry, an occupation can hardly keep its status in the long run if sloppy practices are widespread. Individual professionals must therefore provide proper services (as promised in the implicit contract). But why should the individuals contribute to the professional status, which is a collective good (Olson 1965: 14) as any member of the profession consumes it, and it cannot feasibly be with held from any of the professionals. Therefore, the profession enforces a number of formal and informal professional norms, and the urge to do a job creditably in the eyes of one’s professional peer (Miller 2000: 307) is expected to be a very strong determinant for behaviour. This argument (and the sociology of professions in general) explains why financial incentives seem to be totally unrelated to the behaviour of health care employees, when strong professional norms prescribe a given behaviour. This seem to be the case for individual behaviour (Andersen & Blegvad 2006) as well as for health care provider institution activities (Goodrick & Salancik 1996). This does not, however, imply that sector and incentives cannot also be important for behaviour. Instead of taking side, the paper claims that both sector, economic and professional incentives play a role. The relevant literatures (Public Service Motivation, economic theory for example in the form of Principal Agent models and the sociology of professions) have lived separate lives, but the last decade has seen several attempts to integrate the findings (Le Grand 2003; Brehm and Gates 1999; Miller & Whitford 2007). The increased interest in non-economic types of motivations (e.g. Brehm and Gates 1999) is related to the puzzling rarity of economic incentive contracts in the public sector. This might be due to low efficacy of the agents (Miller & 5 Whitford 2007), measurement problems (Wilson 1989: 159) or team production externalities (Miller 1992 & 2000), but it might also reflect that economic incentives do not matter that much compared to other types of motivation. Hypotheses In order to integrate the literatures, we need specific expectations about the relationship between sector, financial incentives and professional norms. Based on existing studies of professionalised services (Andersen & Blegvad, 2006; Iversen & Lurås, 2000; Goodrick & Salancik, 1996), the paper suggests that professional norms (if they exist) are always important, while incentives are importance if no professional norm dictates behaviour. Depending on the conceptionalization of the public and private sectors, the paper expects public employees to differ (at least in motivation) from public employees. Further, based on the Public Service Motivation literature (Rainey 1982; Perry and Wise 1990; Crewson 1997) I expect that economic incentives matter more in the private sector (compared to the public sector), and that more (and stronger) performance-related financial incentives exist in the private sector. Figure 1 illustrates these causal expectations. Figure 1: Theoretical model Professional norms Financial incentives Behavior Sector The next issue is how this theoretical model can be tested. Utilizing the fact that professional norms only exist for part of the behavior (even for occupations with a high degree of professionalism) this paper analyzes the behavior of public and private dentists and physicians in situations with and without professional norms. Further, these health care providers have different financial incentives, and this enables us to test the relative importance of and the interaction between sector, economic incentives and professional norms. Specifically, Danish dentists face widely different financial incentives regarding their choice between prevention and treatment of children’s caries. Some dentists are rewarded highly for making fillings and poorly for using preventive measures, while the remuneration of others is independent of this choice. As no professional norm regulates the use of fissure sealings (lacquer sealing of the masticating surfaces), the use of this measure is expected to depend on the incentives. 6 Hypothesis 1: Dentists for whom prevention is not lucrative use fewer fissure sealings compared to dentists whose choice of treatment is not linked to their remuneration The obvious objection is, however, that exactly private dentists are paid on a fee-per-item basis (per filling for example) while public dentists get a fixed salary. In Denmark, some municipalities contract with public dentists concerning the delivery of child dental care, while other municipalities employ their own dentists. Anyhow, the services are free of charge for the children. As sector and type of economic incentive co-vary perfectly, this part of the research design is indeterminate with regard to the relative strength of sector compared to incentives. This is often the case as incentives are far more widespread in the private sector compared to the public sector. Still, the test of hypothesis 1 tells us whether the chain illustrated below (which is part of the theoretical model) is plausible. Figure 2: Illustration of the claim in hypothesis 1 Sector Financial incentives (H1: to more use treatment and less prevention) Behavior (H1: actual use of fissure sealings) To see whether economic incentives affect behaviour or whether the differences are only due to the different sector affiliation, we need to hold the sector constant (and vary the incentive). Studies of public hospital doctors and private GPs (general practitioners) in Denmark enable us to do that. First, we can test whether public doctors working in a ward, where an organizational incentive to maximize production was introduced (the budget of the ward per operation became dependent on the number of operations) differs from the behaviour of public doctors without such an incentive. This should be especially pronounced for services such as hip operations where the waiting lists are always long. The incentive is collective (at the ward level) and the test of its effects on individual behaviour is thus conservative. Second, the literature on supplier-induced demand (Evans 1974; Richardson & Peacock 2006; Serritzlew & Andersen, 2006) claims that GPs with a short list have a stronger incentive to increase the number of services per patient, especially time consuming, well paid services. In the Danish GP system, which combines per capital payment (25 % of the average income of GP) with fee per item (the remaining 75 %), the talk therapy consultation is such a service. Thus, the effect of incentives on behavior can be tested in two hypotheses which are also illustrated in figure 3. Hypotheses 2: A public doctor whose ward is paid per operation produces more hip operations than a public doctor whose ward is not paid per operation. Hypothesis 3: There is a negative relationship between the number of therapy consultations per patient and the number of listed patients per GP. 7 Figure 3: Illustration of the claim in hypothesis 2 and 3 Financial incentives (H2: Pay per operation or not) (H3: Number of patients on list) Behavior (H2: Number of therapy consultations) (H3: Number of hip operations) Adding the professional norms to the model, we need to compare services with and without professional norms. Such services clearly exist for both GPs (Serritzlew & Andersen, 2006) and dentists (Andersen & Blegvad, 2002). According to six qualitative interviews with GPs, no professional norm regulates the use of the before-mentioned talk theory (tested in hypothesis 2), while a firm norm requires that the GP must give some kind of service if a patient makes a contact. Thus, the total number of services per patient is not expected to depend on the number of patient on the list. This is tested in hypothesis 4 which concern only private employees (all GPs are private in Denmark). Hypothesis 5 supplements this with a test of both private and public employees. For preventive services with strong professional norms (e.g. instruction in good tooth brushing habits) I expect neither sector nor economic incentives to matter: Everybody is expected to comply with the professional norm. Hypothesis 4: Dentists give tooth brush instruction regardless of their incentives and sector. Hypothesis 5: There is no relationship between the total number of consultations per patient and the number of listed patients per GP. Figure 4: Illustration of the claim in hypothesis 4 and 5 Professional norms (H4: Give service if contact) (H5: Use tooth brush instruction) Financial incentives (H4: Induce demand if short list) (H5: Cut down prevention if not lucrative) Behavior (H4: Total number of consultations per patient) (H5: Actual use of tooth brush instruction) Sector (H4: Only private) (H5: Private and public) The last hypothesis concerns the direct relationship between sector and behavior. As the PSM literature primary talk about motivation, it would be interesting to compare the subjective account about motivation, the perceived differences between the public and private sectors and the reasons for choosing the private and the public sector, respectively. Although an association between sector and motivation does not necessarily have behavioral implications, it is interesting none the less. Using the Crewson (1997) conceptualization of PSM, the hypothesis is: Hypothesis 6: Public health care professional are more often than private health care professionals motivated by sense of usefulness to society and a wish to help others, whereas private health care professionals are more often motivated by pecuniary motives. Data and methods The research design combines many data sources in order to test the theoretical model. For practical reasons, all the data are from Denmark, but the causal relationships are expected to be general. The paper is predominantly based on statistical analysis of register data, but I also included qualitative interviews and a survey. The two most important registers are the Danish Odontological Register (DOR) and the Danish Health Insurance Register (DHIR). Additionally, I use a lot of register information to control for spuriousness etc., but this paper will focus on the causal relationships in the hypothesis and the used control variables will not be commented on (of course, all the analyses include the relevant controls). The Danish Odontological Register contains data on dental health and the most important measures used for all individual Danish children. Each time, a child receives dental services, the dentist fill in a registration form which is scanned into the register. As we have registered for each child whether he or she attended a private or a public dentist, the units of analysis are the individual children. As indicated in hypothesis 1, the paper focuses on fissure sealings, but differences in the dental health measured as the average number of caries infected tooth surfaces in the permanent teeth are also of interest as an indirect indicator of behavior. The Danish Health Insurance Register contains the number and type of services provided by GPs. Based on this register, we investigated services provided in April and May 2006 in the County of Aarhus. As the incentives vary for different GP practices (depending on the number of listed patients), the units of analysis were 257 practices. For the group practices, the number of patients was divided by the number of GPs working in the practice. We investigated only day-time services, because patients visit different doctors at night. The total number of base services per listed patient is an example of a measure strongly governed by professional norms: If a patient contacts the GP, the standards within the occupation say that he must give a service. But the standards do not always prescribe which service the GP must choose. Especially if the patient has social or psychological problems, the GP can freely choose between cognitive therapy and the ordinary consultation (which might lead to a referral to a specialist). To supplement the register information, I conducted a post delivered survey (together with dentist Marianne Blegvad) with employees in the county of Ringkjøbing (n=186). The questionnaire asks if the employees used a number of preventive measures. Some of these measures were regulated by professional norms, while others were not. Finally, I conducted 18 semi-structured interviews with health care professionals: 6 interviews with GPs with varying numbers of patients on their lists, 3 interviews with private dentists, 3 interviews with public dentists, 3 interviews with hospital doctor with incentives to increase productions, and 3 interviews with hospital doctors who had no such incentive. The interviews are analyzed, using NVivo 7 (the categories from the focused coding can be seen in the appendix). Firstly, I use the interviews both to see, how the health care professionals perceive their own motivation and behaviour. Secondly, the interviews helped me find out which services are lucrative and which are not. Finally, the interviews are part of the investigation of the professional norms in the profession. The other parts of this investigation were three elite interviews and analyses of textbooks from the relevant educations. Results The first hypothesis claims that dentists for whom prevention is not lucrative use fewer fissure sealings compared to dentists whose choice of treatment is not linked to their remuneration. In the Danish context, this means that we expect private providers of publicly financed dental care for children to provide significantly fewer fissure sealing per child than public providers of the same services (the municipalities choose whether they will use private or public dentists). The test is conservative in the sense that the municipalities with private providers face slightly worse initial dental care (due to lower average fluoride contents in the drinking water and lower level of education in the population). This should, ceteris paribus, lead to a higher use of fissure sealing. If the results thus show that fissure sealings are used less, we can be more certain that it is due to the difference between the providers. The hypothesis tests whether economic incentives or/and sector matters for health care professional when no firm professional norm applies. The most reliable measure is the register data on fissure sealings. Table 1 shows that children attending public dentists get twice as many fissure sealings compared to children attending private dentists. Table 1: Pooled analysis (1996-2001) for fissure sealings and ownership form Private Public Difference 0.76 (n=37,513) 1.59 (n=278,322) -0.83 12-year-olds: Average number of fissure sealings on the 6’s and 7’s 1.02 (n=34,833) 2.14 (n=255,347) -1.12 15-year-olds: Average number of fissure sealings on the 6’s and 7’s Source: Andersen & Blegvad (2002: 43) Significance p<0.0005 p< 0.0005 The survey of dental workers and the semi-structured interviews show the same tendency: 93% of the public employees with assess to a fissure sealing apparatus (n=50) used it, the corresponding proportion of private dental workers (n=22) was only 78%. Even though the Municipality of Ringkøbing demands that all children receive fissure treatment, only one of the interviewed private dentists used it systematically. One private dentist did not use it at all, while another used it on indication. Explaining his behaviour, the dentist who did not use fissure sealing referred directly to the (lacking) norm in this area. All the interviewed public-sector dentists used fissure sealing systematically. Thus, all three data sources (register data on individual children, employee survey and interviews) confirm that private dental workers use much less fissure sealing than public dental workers. This confirms hypothesis 1, but we cannot determine whether this is due to the different sector or the different economic incentives. We thus turn to the hypothesis 2 and 3 which test whether economic incentives affect behaviour within the public and private sector. Hypotheses claims that public doctors whose ward is paid per operation produce more hip operations than public doctors whose ward is not paid per operation. This hypothesis is much harder to test than hypothesis 1: Hospital services are much more complex than fissure sealings of teeth. Comparing a ward where an incentive was introduced to a ward without incentive to increase production, we can explore how the orthopaedists themselves talk about their motivation and actions, and we can use the growth in the number of surgeries as a rough validation of the interviews. Still, the incentive, if any, is weak and collective. One of the interviewees expresses it thus: “When we increase activity, it trickles from the region to the hospital and to the ward”. The interviews also show that the interviewees generally do not think much about the budgetary dependence on production. Still, the orthopaedists from the ward with the incentive seem to pay a little more attention to the production than the other orthopaedists. A couple of quotations illustrate the point. Everybody say that economic incentives do not matter, and I do not think that it is correct….doctors often put themselves on at a pedestal, shouting that their patient must have the best, no matter what it costs. This is, I think, unrealistic in this system (orthopaedist in the ward with production incentive) We make lots of surgeries which are economically irresponsible, because the payment – diagnose related groups – is based on a standard patient, and only a little part get through to the ward (orthopaedist in the ward without/with very little production incentive) As shown in figure 5, both wards increased their production between 2003 and 2005 (where more activity based incentives were introduced), but the different growth should be very carefully interpreted due to the small number of cases and the complex character of the service. Figure 5: The growth in the number of surgeries at two Danish hospital wards with varying incentives to increase production. 2003-2005. 250% Index (2003=100) 200% With incentives 150% Without incentives 100% 50% 2003 2004 2005 Year In order to test the effect of economic incentives for a less complex service, we now turn to the general practitioner whose rather standardized services are typically produced by one person. Some of these services are not regulated by professional norms, and the use of these services is expected to depend on the economic incentives of the GP. The Danish list system where each GP has a list of patient who must attend him (for day time services) allows us to compare the number of services delivered per patient. As an implication of the Supplier Induced Demand literature, hypothesis 3 expects a negative relationship between the number of therapy consultations per patient and the number of listed patients per GP. In this test, the list size is a proxy for the economic incentive of the GP to induce demand. An alternative proxy could be whether the practice is open for more patients (that is, the GP wants more patient). Table 2 shows the result of a multivariate regression (the socio-democratic characteristics of the patients were included and had no effect). Table 2: Regression analyses of the number of talk therapies pr. 100 patients in April and May 2006 (OLS). Unstandardized regression coefficients. Multivariate model Constant 3.4 Number of patients per GP (list size) -0.00065 *** Open practice -0.0054 Average age of the GPs -0.024 ** R2 0.065 N 257 Note: * p<0.10 ** p<0.05 *** p<0.01. Analyses with the 19 practices which did not use cognitive therapy in the investigated two month show similar results. The regression does, as expected, show a negative correlation between the number of patients and the use of talk therapy. This indicates that economic incentives, at least for services without professional norms produced in the private sector, do affect behaviour. But as many health care services are regulated by professional norms, we need to turn to these services. The firmness of professional norms should be seen as a continuum from no norm to a very firm norm. Within the dental profession, a very firm norm says that children (especially in high risk groups) must have brush instruction. Accordingly, hypothesis 4 expects that all dentists (regardless of their incentives and sector) give tooth brush instruction. Table 3 indicates that this is true. It shows the percentage of public and private dental workers who uses a number of preventive measures (including brush instruction) to normal and high risk children. Table 3: Percentage of clinical personnel in private and public dental care who give high-risk and average children brush instruction. Percentage giving brush For high-risk For average children instruction children Public sector 95% (n=114) 89% (n=114) Private sector 86% (n=44) 77% (n=44) Source: Andersen & Blegvad 2002, 41 The data implies that almost every dental worker uses the measure for high risk children, and that this is true for both public and private dentists. The proportion using the measure for average children is a little lower, which should be expected as the text books and elite interviews show that the norm is strongest for high-risk children. Further, private dental workers seem to use the measure a little less than public employees. This difference is not (unlike for fissure sealings) statistically significant. If we look at the relative incidence of caries among children attending dentists from the two sectors, these statistics also indicate that the behaviour related to the most important preventive measures are rather similar: The two groups have almost equal dental health. The only difference (if any – it is not statistically difference, although consistent over time) is that the children with much caries tend to have even more in the private solution: The percentage of children without caries is exactly the same (Andersen & Blegvad, 2006). This indicates that if strong professional norms apply, neither sector nor economic incentives can overrule the norm. A similar test (of the claim that economic incentives do not affect behaviour if firm professional norms apply) is performed for GP in hypothesis 5. Whereas the list size was expected to affect behaviour for a service without professional norms (talk therapy), no such association should exist if a norm regulates the behaviour. This should be the case for the total number of services – a GP must give some kind of service if his patients contact him. Table 4 shows that no statistically significant relation exists between the list size and the number of base services (different kinds of consultations). Table 4: Regression analyses of the number of base services pr. 100 patients in April and May 2006 (OLS). Unstandardized regression coefficients. Multivariate model Constant 124 Number of patients per GP (list size) -0.0013 Open praxis 1.1 Average age of the GPs -0.31 2 R 0.0093 N 257 Note: * p<0.10 ** p<0.05 *** p<0.01. So far, we have analyzed behaviour in different situations, utilizing different measures. We now turn to the concept of motivation to test the last hypothesis, which says that public health care professional are more often than private health care professionals motivated by sense of usefulness to society and a wish to help others, whereas private health care professionals are more often motivated by pecuniary motives. Generally, it cannot be confirmed. The public employees talk a lot about the high wages in the private sector and about their private colleagues being motivated by pecuniary motives, but the private interviewees talk just as much about the patients and doing good as the public interviewees. Further, many of the (private) GPs mention that they try to maximize public value for patient as well as for society. Doing things right according to the standards within the profession and having interesting tasks play a bigger role for the interviewed professionals, at least according to themselves. They might twist the truth a little bit to place themselves in a favourable light, but no difference in motivation can be identified based on the interviews. Discussion The fact that both private and public health professional only seem to be motivated by pecuniary motives to a small degree – when no firm norms apply – is probably due to the fact that they are professionals. Occupational norms are a defining characteristic of professions, and the design does, in this respect, represent a least likely case for the association between sector, economic incentives and behaviour. Even for the professionals, economic incentives affect behaviour sometimes, and sector thus often has an indirect effect because different (often more performance related) incentives exist in the private sector. This conditional effect of economic incentives agrees with many of the findings in the literature (Goodrick & Salancik 1996; Iversen & Lurås 2000; Brehm & Gates 1997). Professional norms (if they exist) appear to be the stronger motive for physicians and dentists in their daily work, and as many of the core functions of many welfare states are performed by professionals, this is an important message. Physicians and dentists do, however, have very high professional status, and most public sector occupations are between the extremes of this high status and no status at all. It is therefore most relevant whether the findings can be generalized to other occupations. They can hardly be transferred directly: As the strength of the professional norms decrease, economic and sector probably have stronger effects on behaviour. But the findings indicate that professional norms, as well as economic incentives and sector, are important for behaviour. The presented evidence is Danish, thus holding contextual factors such as the type of the health care system constant. Although it should be tested in a comparative study, there is no reason to believe that the causal mechanism should be fundamentally different in other countries. The medical and dental professionals are, to a high degree, global, and many professional norms are based on international, scientific evidence. Still, the high levels of social capital in Denmark (Bjørnskov & Svendsen 2002) might increase the importance of professional norms as the main element in social capital (the level of trust) can be seen as a necessary condition for professionalism contracts to develop. This trust goes both ways: The professionals must trust that they will be rewarded in term of status and recognition, and the public must trust the professionals to keep their part of the deal – to uphold a minimum standard of the services. Conclusion The question in this paper is how professional norms, economic incentives and sector affect the behaviour of Danish health care producers. This was made more specific by testing five hypotheses about motives and behaviour. First, we found that dentists for whom prevention is not lucrative use fewer fissure sealings compared to dentists whose choice of treatment is not linked to their remuneration. This indicates that economic incentives (and the sector behind the differences in incentives) can affect behavior if no professional norm applies. Holding the sector constant, the next two analyses also showed (small) effects of the incentives: The negative association between the number of therapy consultations per patient and the number of listed patients per GP indicates that GPs with short list (and therefore an incentive to induce demand) do in fact give more services, and public doctors whose ward was paid per operation seemed to care more about production. Still, hip operations showed to be a too complex service to directly test the association between incentives and production. The abovementioned decisions are not governed by professional norms. Talk therapy and fissure sealing were not at the time of investigation accepted in the professions, and although a norm regulates whether a patient should have a new hip, it does not say how many hips a doctor should produce. Some services are, however, governed by professional norms. A norm requires that the GPs give some service if a patient contacts them (their discretion lies in the choice of service), and dental workers are required to instruct in tooth brushing. As expected, neither sector nor economic incentives matter in these situations. Professional norms thus almost dictate the behaviour of Danish health care professionals at some areas, and if a firm norm exists, members of the same occupation – with different incentives – behave similarly in the private and public sector. But not all behaviour is regulated by such norms. If no firm professional norm applies, economic incentives play role. Whether sector matters, is more difficult to say based on the evidence in this article. While the public sector interviewees clearly think that money matters more in the private sector, no clear sector effect can be identified. This can, however, be due to the research design – and to general high correlation between sector and incentives which makes it difficult to separate the two factors. Still, as incentives have been shown to matter in analyses holding the sector constant, the main conclusion of the paper is: If a firm professional norm exists, it dominates the behaviour of health care professional, but if no such norm applies, economic incentives also matter. Much does, however, remain to be done: This paper has shown the importance of professional norms for the ‘most likely case’ for this to have an effect. Further research should investigate occupations with less professional status, and another step might be to investigate the role of professional norms in a county with low social capital. Appendix Table A1: Categories and sub-categories from the analysis of semi-structured interviews Main category Subcategories (if any) Motivation Autonomy Exciting tasks Research Leisure Career Colleges Patients Money Professional norms Sanctioning of norms Norms within the medical specialty Norms in Denmark compared to Sweden Scientific evidence Choice of treatment Economic considerations DRG (diagnosis related group) Private-public comparisons References Andersen, Lotte Bøgh & Marianne Blegvad (2002). Private or Public Service Provision? Economic and professional incentives in Danish dental care for children. Working paper. http://www.ps.au.dk/lotte/Baggrundsrapport.pdf Andersen, Lotte Bøgh (2005). Offentligt ansattes strategier. Århus: Politica. Bjørnskov, Christian & Gert Tinggaard Svendsen (2002). Why Does the Northern Light Shine So Brightly? Decentralisation, Social Capital and the Economy. Working Paper 02-15, Aarhus School of Business (http://www.hha.dk/nat/WPER/02-15_gts.pdf accessed 250907) Brehm, John & Scott Gates (1997). Working, Shirking and Sabotage. A Bureaucratic Response to a Democratic Public, USA: University of Michigan Press. Burgess, Simon, Carol Propper & Deborah Wilson (2002). Does Performance Monitoring Work? A Review of the Evidence from the UK Public Sector Excluding Health Care, Bristol: University of Bristol. Crilly, Tess & Julian Le Grand (2004). ‘The motivation and behavior of hospital trusts’, Social Science and Medicine, 58, 1809-1823. Crewson, Phillip E. (1997). "Public-Service Motivation: Building Empirical Evidence of Incidence and Effect", Journal of Public Administration Research and Theory, 7, 499-518. Day, Patricia & Rudolf Klein (1987). Accountability. Five public sectors. London and New York: Tavistock Publications. Donaldson, Cam & Karen Gerard (1993). Economics of Health Care Financing: The visible hand, Houndmills & London: Macmillan. Downs, Anthony (1961). “Why the Government Budget is too small in a Democracy”, World Politics, pp. 541-563. Dunleavy, Patrick (1991). Democracy, Bureaucracy and Public Choice: Economic Explanations in Political Science. London: Harvester Wheatsheaf. Evans, Robert G. (1974). “Supplier-induced demand: Some empirical evidence and implications” pp. 162-173 in Mark Perlman (ed.) The economic of health and medical care, London & Basingstoke: Macmillan. Freidson, Eliot (2001). Professionalism. The Third Logic, Cambridge: Polity Press. Goode, William J. (1969). “The Theortical Limits of Professionalization” i Amitai Etzioni (ed.), The Semi-professions and Their Organization. New York: The Free Press. Goodrick, Elizabeth & Gerald R. Salancik (1996). “Organizational Discretion in Responding to Institutional Practices: Hospitals and Cerarean Births”, Administravie Science Quarterly, 41, pp. 1-28. Gosden, Toby, Frode Forlan, Ivar Kristiansen, Matthew Sutton, Brenda Leese, Antonio Giuffrida, Michelle Sergison & Lone Pedersen (2001). ‘Impact of Payment Method on Behaviour of Primary Care Physicians: A Systematic Review’, Journal of Health Services Research and Policy, 6, 44-55. Graham, Alison & Jane Steele (2001). Optimising Value: The Motivation of Doctors and Managers in the NHS, London: Public Management Foundation. Iversen, Tor & Hilde Lurås (2000). ”Economic motives and professional norms: the case of general medical practice” Journal of Economic Behavior & Organization. 43: 447-470. Krasnik, Allan, Peter P. Groenewegen, Poul A. Pedersen, Peter v. Scholten, Gavin Mooney, Adam Gottschau, Henk A. Flierman & Mogens T. Damsgaard (1990). ‘Changing remuneration systems: effects on activity in general practice’, British Medical Journal, 300, 1698-1701. Le Grand, Julian (2003). Motivation, Agency, and Public Policy. Of Knights and Knaves, Pawns and Queens, Oxford: Oxford University Press. Meier, Kenneth J. & Laurence O’Toole (2002). ‘Public Management and Organizational Performance: The Impact of Managerial Quality’, Journal of Policy Analysis and Management, 21, 629-643. Miller, Gary J. (2000). “Above Politics: Credible Commitment and Efficiency in the Design of Public Agencies.” Journal of Public Administration Research and Theory. 10 (April): 289-328. Miller, Gary J. (2005). "The Political Evolution of Principal-Agent Models", Annual Review of Political Science, 8, pp. 203-25. Miller, Gary J. & Andrew B. Whitford (2007). “The Principal's Moral Hazard: Constraints on the Use of Incentives in Hierarchy” Journal of Public Administration Research and Theory, 17: 213233. Miller, Gary J. (1992). Managerial dilemmas: The political economy of hierarchy. Cambridge: Cambridge University Press. Mosher, Frederick C. (1968). Democracy and the public service. New York: Oxford Univ. Press. Niskanen, William A. (1971). Bureaucracy and Representative Government. Chicago: W.A. Aldine-Atherton. Olson, Mancur (1971). The Logic of Collective Action. Cambridge: Harvard University Press. Ostrom, Elinor (1986). “An Agenda for the Study of Institutions”, Public Choice, 48, pp. 3-25. Perry, James L. & Lois R. Wise (1990)."The Motivational Bases of Public Service", Public Administration Review, 50, 367-373. Perry, James L. (1988).’Making Policy by Trial and Error: Merit Pay in the Federal Service’, Policy Studies Journal, 17, 389-406. Rainey, Hal G. (1982)."Reward Preferences Among Public and Private Managers: In Search of a Service Ethic, American Review of Public Administration, 16, 288-302. Richardson, Jeff & Stuart Peacock (2006). Reconsidering theories and evidence of supplier induced demand. Research paper 2006 (13). Centre for Health Economics, Monash University, Australia. http://www.buseco.monash.edu.au/centres/che/pubs/rp13.pdf. Roberts, Jennifer & Michael Dietrich (1999). "Conceptualizing Professionalism: Why Economics needs Sociology", American Journal of Economics and Sociology, 58: 977-998. Serritzlew, Søren & Lotte Bøgh Andersen (2006). "Økonomiske incitamenter i praksissektoren. Lægens tre fristelser", Politica, vol. 38 no. 4, pp. 392-409. Steele, Jane (1999). Wasted Values: Harnessing the Commitment of Public Managers, London: Public Management Foundation. Taylor-Gooby, Peter, Stella Sylvester, Mike Calnan & Graham Manley (2000). ‘Knights, Knaves and Gnashers: Professional Values and Private Dentistry’, Journal of Social Policy, 29, 375-95. Vrangbæk, Karsten (2003). ‘Værdilandskabet i den offentlige sektor. Resultaterne fra en survey’, pp. 105-133 in Carsten Beck Jørgensen (ed.), På sporet af en offentlig identitet, Århus: Aarhus Universitetsforlag. Watson, Tony J. (2003) (1980). Sociology, work and industry, 4. udgave, London: Routledge. Wilson, James Q. (1989). Bureaucracy. What Government Agencies Do and Why They Do it. New York: BasicBooks. Wright, Bradley E. (2001). “Public-Sector Work Motivation: A Review of the Current Literature and a Revised Conceptual Model” Journal of Public Administration Research and Theory. 4: 559-86.