Research

Publications

Moscelli G, Nicodemo C, Sayli M, and Mello M. (2024). Trends and determinants of clinical staff retention in the English NHS: a double retrospective cohort study BMJ Open 14:e078072 

Research Reports

Einarsdóttir, A., Mumford, K., Birks, Y., Lockyer, B., and Sayli, M. Understanding LGBT+ Employee Networks and How to Support Them. Final report for ESRC funded grant award ES/N019334/1. October 2020, 55 pp.

Einarsdóttir, A., Mumford, K., Birks, Y., Aguirre, E., Lockyer, B., and Sayli, M. LGBT+ Employee Networks Within the NHS: Technical Report and Data Addendum. Supplementary report for ESRC funded grant award ES/N019334/1. October 2020, 149 pp.

Datasets

Mumford, K., Einarsottir, A., Birks, Y., Aguirre, E., Lockyer, B. and Sayli, M. (2021). The Concise National Health Service Human Resources and Equality Diversity Survey, 2018-2019. [Data Collection]. Colchester, Essex: UK Data Service. 10.5255/UKDA-SN-855320

Einarsdottir, A., Mumford, K., Birks, Y., Lockyer, B., Sayli, M.  and Sudthasiri, S. (2021). LGBT+ Networks, 2017-2020. [Data Collection]. Colchester, Essex: UK Data Service. 10.5255/UKDA-SN-855322


Interdisciplinary Publications

Einarsdottir, A., Mumford, K. A., Siriviriyakul, S.,  and Sayli, M. (2022).  Do I have to say I’m gay? Using a video booth for public visibility and impact. Qualitative Research.




Working Papers

Abstract Excessive turnover can significantly impair an organization’s performance. Using high-quality linked administrative data and staggered difference-in-differences empirical strategies, we evaluate the impact of a programme that encouraged public hospital organizations to increase staff retention by providing data and guidelines to improve the non-pecuniary aspects of nursing jobs. We find that the programme has decreased the nurse turnover rate by 5.29%, decreased exits from the public hospital sector by 5.68%, and reduced mortality rates within 30 days from hospital admission by 3.40%, preventing about 7,813 deaths. Our results are consistent with a theoretical model in which information is provided to managers of multi-unit organizations, who trade-off coordinating decisions across units and adapting them to local conditions

This work is a revised version of the earlier version of the working paper, which can be accessed here.

[In Media: Press Release, UK Today News, Medical Xpress, News Medical Life-Sciences]

[Online posts: Will the NHS long-term workforce plan solve the current crisis? Economics Observatory & CentrePiece Autumn 2023]


Abstract  Retention of skilled workers is essential for labor-intensive organisations like hospitals, where an excessive turnover of doctors and nurses can reduce the quality and quantity of services provided to patients.  Exploiting a unique and rich panel dataset based on employee-level payroll and staff survey records from the universe of English NHS hospitals, we empirically investigate the role played by two non-pecuniary job factors, staff engagement and the retention of complementary coworkers, in affecting employee retention within the public hospital sector.  We estimate dynamic panel data models to deal with reverse causality bias and validate these estimates through unconditional quantile regressions with hospital-level fixed effects. Our findings show that a one standard deviation increase in nurse engagement is associated with a 16\% standard deviation increase in their retention; and also that a 10\% increase in nurse retention is associated with a 1.6\% increase in doctor retention, with this coworkers' complementarity spillover effect driven by the retention of more experienced nurses. Nurse and doctor engagement is positively associated with managers who have effective communication, involve staff in the decision-making process, and act on staff feedback; in particular, older nurse engagement is responsive to managers caring for staff health and well-being.

This work is a revised version of an earlier version, which can be accessed here.

[In Media: The Guardian, Nursing Times,  Nursingnotes, RCNi, Healthcare in Europe, Medical Xpress ]



Abstract Retention of skilled workers is critical for the delivery of public services in high-stakes environments such as hospital care.  We study how contractual pay terms affect the retention of trainee doctors in the English NHS and the relationship between trainee doctors' attrition and hospital quality.  Our setting is a nationwide reform that reduced unsocial working hours pay rates. Using a longitudinal sample and a novel linkage of administrative datasets, our quasi-difference-in-difference strategy leverages the pre-reform exposure of each trainee doctor to unsocial working hours and suggests that the implementation of the new pay terms led to a 6.7\% increase in the annual number of trainee doctors leaving the English NHS.  As possible mechanisms, we show that the reform was detrimental to pay satisfaction and disproportionately affected female trainee doctors.  We also exploit the effect of the reform to document a positive association between trainee doctors' attrition and hospital mortality.



Work in Progress



[Online coverage:  Royal Economic Society and LSE Business Review]
AbstractThis paper examines the interdependency between a woman's labour market attachment and her partner's non-employment in the UK in a causal fashion. Additionally, it aims to identify the sources of the potential endogeneity in partners' labour market outcomes a la Lin and Wooldridge (2017). As couples do not come together at random, this paper presents the observed sorting patterns and provides a measure of assortative mating based on the unobserved factors in partners' labour market outcomes. Typically, the identification would come from policy changes or exogenous shocks, which are not available in the data. Hence, this study uses the male partner's parental employment when he was 14 years old, and measures of his health as separate instruments. On the contrary of Mavromaras and Zhu (2015) who use a similar identification strategy, the partner's parental employment turns out to be a 'weak' instrument, whereas the partner's self-assessed health measures are better instruments. An average woman, with a workless partner, is around 30 percentage points less likely to be active than an average woman whose partner is employed. This result is robust to different model specifications and sample restrictions. When the unobserved permanent couple characteristics are controlled for, the same effect drops to 13 percentage points. Nevertheless, the effect of a woman's partner's non-employment on her participation probability remains significantly negative. The findings indicate that the endogeneity in this framework stems from the unobserved permanent factors rather than idiosyncratic shocks. The unobserved factors in partners' labour market outcomes are positively correlated, which is interpreted as a negative assortative mating based on unobserved factors common to both partners.


Abstract 

An increase in the labour supply of a couple member in a given household provides additional insurance against lower income due to the other member's non-employment. In this paper, I examine whether and how the inactive married or cohabiting woman’s participation decision is influenced by her partner's different labour market activities, i.e. the added worker effect, and investigate how the time she spends away from the labour market affects her labour supply decision. Using a panel of couples' monthly labour market histories that I constructed from the British Household Panel Survey 1991-2009, I find significant negative duration dependence in woman's participation, which is strongest in the first three years of her inactivity. A woman with an unemployed partner is 23% less likely to enter the labour market than a woman whose partner is employed. On the other hand, a woman's labour supply decision depends on her partner's labour market activity, and a woman with an inactive partner is more likely to participate in the labour force than a woman whose partner is unemployed. The duration dependence and the added worker effect do not vary by the way a woman enters the labour force, i.e. via job-finding or via job-search, however, claiming income support or unemployment benefit within an interview year has destination-specific effects as it increases the probability of a woman's participation via job-search rather than job-finding.  

This dataset contains monthly labour market information of matched partners using all available waves of the British Household Panel Survey 1991-2009 (BHPS).  Such a dataset is not available in the UK Data Archive (UKDA), and this paper provides a guide to the complex task of combining the BHPS records to construct a dataset on pairs of individuals, as cohabiting or legally married couples, and their labour market histories. It discusses the existing datasets in the UKDA and ways to incorporate these into the construction of couples' monthly labour market histories. Constructing the labour market histories benefits the set of principles and codes presented in Mare (2006, 2015). The partnership histories from Nazio (2010) and Pronzato (2011) are synthesized and updated with BHPS annual records, and the data of couple members are matched to each other. The final dataset consists of  7025 women and 7021 couples who are observed in 7261 uninterrupted couple spells between September 1990 and April 2009.