کشش عرضه کار پرستاران: نقش حسابداری برای حاشیه های گسترده

We estimate a multi-sector model of nursing qualification holders’ labour supply in different occupations. A structural approach allows us to model the labour force participation decision, the occupational and shift-type choice, and the decision about hours worked as a joint outcome following from maximising a utility function. Disutility from work is allowed to vary by occupation and also by shift type in the utility function. Our results suggest that average wage elasticities might be higher than previous research has found. This is mainly due to the effect of wages on the decision to enter or exit the profession, which was not included in the previous literature, rather than from its effect on increased working hours for those who already work in the profession.

Maintaining an adequate supply of nurses is a key ongoing policy issue in many countries. The World Health Organisation (2006) estimated global shortages of health workers of 4.3 million. Shortages are important in the light of evidence, largely from the US, of the impact of the level of nurse staffing on patients’ health outcomes (Lankshear et al., 2005, Kane et al., 2007 and Cook et al., 2010). It is clear that shortages are common across many low and high income countries, which means that research on the determinants of nursing labour supply remains a priority. Even where overall shortages are less of an issue, shortages in specific sectors, disease areas, and geographic areas highlight the role of maldistribution. Although the budgetary pressures from the global financial crisis may have reduced demand for nurses employed in the public sector in some countries, the ‘need’ for health care, and therefore for nurses, continues to grow as the population ages and as developing countries continue to struggle to provide basic health services. Shortages are a long-term issue since labour markets, especially for nurses who work in the public sector, are often slow to adjust to changes in demand and supply conditions partly due to centralised wage bargaining. At the same time, the nursing workforce is also ageing and a high number of ‘baby boomer’ nurses are leaving the profession as they retire. This trend is aggravated by relatively low numbers of young graduates who enter the profession. The stock of trained nurses who could in principle work in the profession but are currently not doing so is relatively high. For example in Australia, around 17% of registered or enrolled nurses were not working as nurses in 2005 (Australian Institute of Health and Welfare (AIHW), 2008). Raising wages in order to attract them back to the profession may be one possible option to tackle the problem, but increasing wage offers will be an optimal strategy for the demand side of the market (i.e. public and private health care providers) only if the response of the supply side is strong enough. Otherwise, the cost of inducing any substantial change in labour supply will be too high and have an even greater dampening impact on demand. Most previous research suggests that the labour supply elasticity with respect to wage is fairly low among nurses at around 0.3 (see Antonazzo et al., 2003 and Shields, 2004 for a review). However, a major drawback of these studies is that they do not model the decision to exit or enter the nursing profession versus other occupations. The elasticity of labour supply in nursing at the extensive margin might differ considerably from that at the intensive margin. This is of particular importance as the nursing workforce primarily consists of women. Not only is their general labour force attachment lower than that of men, but their career paths, moving out of and back into the labour force around childbirth, also lower the opportunity costs of changes in occupation and thus make career changes more likely. In order to obtain reliable estimates of the total labour supply elasticity in nursing, it is thus important to model not only individuals’ labour force participation decisions – which has been done in previous research using a broad range of methods – but also their occupational choices. This paper seeks to investigate the viability of wage policies to increase nurses’ labour supply through a multi-sector labour supply model. We extend the literature by considering a sample of nursing qualification holders, rather than those who work as registered nurses only. This means that our sample contains potentially available nurses regardless of their labour force participation and regardless of the occupation they are currently working in. Thus our sample enables us to model the extensive margin not only in terms of workforce participation ( Creedy and Duncan, 2002), but also in terms of choice of non-nursing occupations. This allows us to address the following questions: (1) how responsive is the decision to enter the nursing profession to wages in nursing?, (2) how responsive is the decision to exit the profession to wages in non-nursing occupations?, and (3) is the estimated total elasticity of labour supply in nursing higher than suggested by previous research, once transitions between nursing and other occupations are allowed for in the model? The answers to these questions will indicate whether increasing wages in nursing could be a more promising policy to tackle the problem of nurse staffing shortages than previously thought. Additionally, we examine the choice for shift type (distinguishing regular day shifts, regular night shifts and irregular shifts) as part of nurses’ labour supply. In our structural approach employment in different shift types enters the utility function as separate arguments. This leads to our fourth research question: (4) does the responsiveness of labour supply to wage changes differ by shift type? A further novelty of our research is that we investigate the importance of family circumstances for nurses’ labour supply, as these are expected to affect the utility derived from income and the disutility derived from working hours in nursing jobs with different shift types or a non-nursing occupation. This leads to our final research question: (5) to what extent does the responsiveness to wage changes depend on a nurse's family circumstances? To give a preview of our findings, the results suggest that nurses’ labour supply might be more responsive to changes in wages than earlier research concluded, as we find an average elasticity of hours worked in nursing with respect to nursing wages of 1.3. This higher elasticity is due to changes in the entry into and exit from the nursing profession in response to wage changes. Earlier research has ignored the possibility of entry from or exit to other occupations. Work in night shifts turns out to be associated with a higher disutility from work than work in other shift patterns, and thus the responsiveness to wages is considerably smaller. Estimating labour supply elasticities for different groups of nurses with different family circumstances, it is found that the labour supply elasticity with respect to other non-labour income is considerably higher for nursing qualification holders with children and for those with a higher qualification. For the labour supply elasticity with respect to nursing wages, it is found that lower qualified, childless, and older nursing qualification holders respond more strongly to a given wage increase than their counterparts. The paper continues as follows. Section 2 briefly discusses previous literature, and Section 3 the proposed estimation strategy. This is followed in Section 4 by a description of the data, including some summary statistics. The results are presented in Section 5. Section 6 concludes.

Can wage policy help to maintain an adequate level of nurse staffing? How responsive are nurses to relative wages in nursing versus other occupations in their decision to exit and enter the profession? Our evidence shows nurses’ labour supply elasticity to be higher than previous research suggested. Wage increases thus appear to be a more promising policy to increase labour supply than was implied by previous research. We have estimated a structural model of labour supply for nursing qualification holders that distinguishes explicitly between working in different shift types and in different occupations. The responsiveness of nursing qualification holders with respect to wages in nursing jobs is high at an elasticity of 1.37, and this result is mainly due to our explicit inclusion of the decision to enter and exit the occupation, rather than examining only changes in the number of hours worked of those already working in nursing. Our results thus show that nursing qualification holders who currently do not work in nursing are an important group to target from a policy perspective. At the same time, wage increases for non-nursing jobs are expected to draw individuals out of nursing occupations to almost the same extent as increases in nursing wages might attract nursing qualification holders, who currently work in another occupation, back to nursing. The high responsiveness of the occupational choice implies not only that wage increases in nursing are a promising policy for governments to increase labour supply in the public sector, but also that it is crucial to ensure that wages in nursing do not fall behind the development of wage levels offered in the alternative occupations for nursing qualification holders, as this will reduce labour supply in nursing substantially. Naturally, nurses’ observed hours of work are not solely determined by the supply side of the nurses market. The demand side influences nurses’ observed hours of work as well, e.g. through restricting the number of combinations of shift type and hours worked on offer. We have investigated the sensitivity of results to this by exploring how results would change if only nurses at their optimal (or preferred) labour supply are included in the analysis. It turns out that this would reduce the elasticity from 1.37 to 1.21, which remains a substantial elasticity, still indicating a potential role for wage policy. The majority of nursing qualification holders are at their preferred hours of work (72%), and perhaps those who are not quite at their preferred hours are not far away from them, so that the impact of including those at sub-optimal hours in the analysis is not large. From the observed hours of work and shift type combinations (Table 1), it appears that a broad range of choices are available to employees in this sector, making it more likely that they work at their preferred hours. A further finding is that changes in wages for certain shift types are predicted to result in considerable shifts in labour supply between shift patterns, which is in line with previous research. If policymakers increase wages for a specific shift type only, they need to be aware that such a policy will lead to a decrease in labour supply in other shift types to a substantial extent, as about half of the labour supply response is caused by the current nursing workforce changing shift type, rather than by attracting nursing qualification holders who are currently not working in nursing, or by increasing of supplied hours of nurses who work in the shift type already. Increased labour in one shift type thus comes at the cost of decreased labour supply in another, if wages are increased only for certain shift patterns. Such a policy may therefore be useful to adjust the distribution across shift types when there is a mal-distribution across shift types, but it is unlikely to provide a solution when there is a shortage in one shift type without oversupply in other shift types. Comparing wage and income elasticities for different groups of nurses, shows that wage elasticities are higher for low qualified, older and childless nursing qualification holders. At the same time, available non-labour income plays a particularly important role for nursing qualification holders with young children. This indicates that wage policies are not going to be equally effective amongst all individuals who hold a nursing qualification. Some groups will be far more responsive than other groups. This needs to be considered in designing wage policies.