Academic research
Ongoing research
Assortative mating and wealth inequality in Great Britain
Testosterone and labour market earnings (joint with Peter Eibich and Alexander Plum)
2024
Relabelling of Individual Early Retirement Pension in Finland: application and behavioural responses using Finnish register data. (joint with Satu Nivalainen and Noora Järnefelt) Forthcoming, Journal of Economic Behaviour and Organisation
Using rich Finnish population level registers, we examine the impact of fusing a flexible early retirement pathway with a more stringent pathway, without changing eligibility conditions, so-called ‘relabelling’, on individual application behaviour. Our findings show that among affected cohorts the likelihood of applying for (successfully claiming) disability-related early retirement declined by 1.8 (1.5) percentage points equivalent to a relative drop of approximately 37% (39%) following the reform. Individuals with below tertiary level education and stronger lifetime labour market attachment exhibit a stronger behavioural response to the reform. We find tentative evidence of programme substitution to early retirement pathways designed to keep individuals in the labour market albeit on a part time basis. Our findings suggest that social norms and lack of awareness associated with early retirement pathways can strongly influence application behaviour even when eligibility conditions remain unchanged, offering policymakers novel ways to extend working lives.
2023
Socio-economic inequalities and trajectories of a novel multidimensional metric of Active and Healthy Ageing: the English Longitudinal Study of Ageing (joint with Olivia Malkowski and Max Western) (Nature- Scientific Reports)
Healthy ageing research largely has a unidimensional focus on physical health, negating the importance of psychosocial factors in the maintenance of a good quality-of-life. In this cohort study, we aimed to identify trajectories of a new multidimensional metric of Active and Healthy Ageing (AHA), including their associations with socio-economic variables. A latent AHA metric was created for 14,755 participants across eight waves of data (collected between 2004 and 2019) from the English Longitudinal Study of Ageing (ELSA), using Bayesian Multilevel Item Response Theory (MLIRT). Then, Growth Mixture Modelling (GMM) was employed to identify sub-groups of individuals with similar trajectories of AHA, and multinomial logistic regression examined associations of these trajectories with socio-economic variables: education, occupational class, and wealth. Three latent classes of AHA trajectories were suggested. Participants in higher quintiles of the wealth distribution had decreased odds of being in the groups with consistently moderate AHA scores (i.e., ‘moderate-stable’), or the steepest deterioration (i.e., ‘decliners’), compared to the ‘high-stable’ group. Education and occupational class were not consistently associated with AHA trajectories. Our findings reiterate the need for more holistic measures of AHA and prevention strategies targeted at limiting socio-economic disparities in older adults’ quality-of-life.
Link to paper:
https://www.nature.com/articles/s41598-023-33371-0
Labour supply and expected retirement age among employees in the UK: evidence from the United Kingdom Household Longitudinal Study (joint with Adriaan Kalwij) Revise and Resubmit, Journal of Pension Economics and Finance
We examine individuals’ retirement behaviour in response to changes in the State Pension eligibility age introduced in various Pension Acts in the UK. Our findings show that the probability of retirement reduces significantly in response to a one-year increase in State Pension eligibility age, by 16 pp and 13 pp for men and women respectively. We find women adjust their expected retirement age downwards in response to an increase in the SP eligibility age. Our findings suggest that whilst changes in the State Pension eligibility age are important for individual’s actual retirement, they do not induce individuals and especially women to revise their expected retirement age and this could result in suboptimal retirement planning. The latter can be problematic for those who rely disproportionately on State Pension as their main source of income and, arguably, targeted communication campaigns are needed to improve retirement planning.
Parental homeownership and education: The implications for offspring wealth inequality in Great Britain (joint with Paul Gregg) Journal of Social Policy
The rapid widening of wealth inequalities has led to sharp differences in living standards in Great Britain. Understanding whether and separately the rate at which individuals accumulate particular types of wealth by family background is important for improving wealth and social mobility. We show offspring wealth inequality is driven by housing wealth, and holding such wealth is becoming increasingly associated with early life circumstances relating to parental housing tenure and education, even after controlling for adult offspring’s own characteristics. Importantly, we find adult offspring whose parents hold a degree and are homeowners are no less likely to report homeownership and housing wealth compared to older cohorts from the same background. Our findings infer the intergenerational rank correlation in housing wealth is set to double in approximately three decades.
Link to paper:
https://doi.org/10.1017/S0047279423000442
2022
In And Out Of Unemployment –Labour Market Transitions And The Role Of Testosterone (Joint with Peter Eibich, Julien Schmied and Alexander Plum) (Economics and Human Biology)
Biological processes have provided new insights into diverging labour market trajectories. This paper uses population variation in testosterone levels to explain transition probabilities into and out of unemployment. We examine labour market transitions for 2,004 initially employed and 111 initially unemployed British men from the UK Household Longitudinal Study (“Understanding Society”) between 2009 and 2015. We address the endogeneity of testosterone levels by using genetic variation as instrumental variables (Mendelian Randomization). We find that for both initially unemployed men as well as initially employed men, higher testosterone levels reduce the risk of unemployment. Based on previous studies and descriptive evidence, we argue that these effects are likely driven by differences in cognitive and non-cognitive skills as well as job search behaviour of men with higher testosterone levels. Our findings suggest that latent biological processes can affect job search behaviour and labour market outcomes without necessarily relating to illness and disability.
Link to paper:
https://www.sciencedirect.com/science/article/pii/S1570677X22000193
Testosterone Effects as Men Age (joint with Allan Mazur, Julien Schmied and Alexander Plum) (The Aging Male)
The literature on testosterone (T) in men reports diverse correlates of T, some with minimal empirical support and most with little indication of how they change with advancing age. Here we test eight putative correlations against a large sample of British men, replicating nearly all. Most of the examined correlations change across men’s life courses. One instance of age attenuation is the diurnal cycle of T, which is considerably weaker among older than younger men. Single men have higher T than married men of the same age; however, this difference lessens as men get older. Also, elevated T among smokers is less pronounced as men age. On the other hand, the inverse relationship between obesity and T is sustained across the adult age range. Of particular interest is the lessening of T with age, which is well established in the literature. However, there is disagreement about the course of decline. We find T having a steep decline around age 30, after which the levels remain roughly constant. Correlations involving health become stronger among older men. After age 30 or 40, the inverse relationships between T and HbA1c, diabetes, and metabolic syndrome all become increasingly significant, though not necessarily strong in magnitude. There is a basis here for the generalization that among older men, those healthy have higher T than those who are not, but not a lot higher.
Link to paper:
https://www.tandfonline.com/doi/full/10.1080/13685538.2021.2023493
Intergenerational wealth transmission in Great Britain (joint with Paul Gregg) (Review of Income and Wealth)
The aim of this paper is to document the intergenerational persistence of wealth between adult offspring and their parent’s using the Wealth and Assets Survey for Great Britain. As parental wealth it is not directly observed it is assessed as mean values based on age, home ownership and education from retrospective questions. Estimates are then derived employing a commonly used two stage estimator. For offspring aged around 44 and parents aged around 74, the oldest where wealth can reliably be observed in the sample, the intergenerational wealth elasticity (IWE) is 0.4 and the rank-rank elasticity 0.3. However, wealth is a stock accumulated over a person’s working life and then dissaving takes place in retirement. Thus, peak wealth holding occurs around the age of 64 and this represents a proxy measure of life-time wealth accumulation. Under certain assumptions about parental wealth holding we explore wealth persistence for older offspring up to age 64. Importantly, we find at these older ages wealth persistence is generally lower than for those currently aged in their 30s and early 40s, though rank based estimates are broadly stable. The average IWE is 0.35 (ages 28-64) and rank equivalent 0.3 in 2012. For those in their 30s however, the IWE is 0.4, even though the short panel suggests a strong life cycle bias where wealth persistence is lower at ages below 64. Exploration of this contradiction shows that those who have a relatively high wealth among older cohorts came from more typical backgrounds than in younger ones. The six year panel data also shows that intergenerational wealth elasticity is 3.8 percentage points higher when comparing people with those the same age six years previously. There is, thus, very strong evidence of higher wealth intergenerational persistence in younger age cohorts. As it was already higher than for older cohorts and has risen rapidly, standing at 0.44 (ages 32-44) by 2018.
Link to published paper:
https://onlinelibrary.wiley.com/doi/full/10.1111/roiw.12620