Research

Abstract: I study how parental Alzheimer’s disease and related dementias (ADRD) affect adult children’s employment, living arrangements, and caregiving, using longitudinal data from the Health and Retirement Study spanning over two decades. I employ an event-study design with two treatment definitions: the first reported parental ADRD diagnosis, and the first transition to Mild Cognitive Impairment or Dementia (MCI/dementia) under the Langa-Weir classification, the latter capturing first signs of cognitive deterioration observable years before formal diagnosis. Employment falls by 1.7 percentage points at diagnosis and by 5.9 pp six years later, with no recovery, driven by full-time work. Despite the gradual progression of the disease, employment shows no clear anticipation. Under the MCI/dementia definition, effects emerge only years after the first signs of cognitive impairment, mirroring those observed around diagnosis and consistent with a labor market response driven by disease severity. By contrast, caregiving, measured as personal and financial care, responds earlier, rising steadily after the first signs of cognitive impairment and intensifying at diagnosis, with a peak at around 7-8 pp. Finally, living arrangements show limited effects overall, though daughters exhibit a slightly higher tendency to co-reside with the ill parent. Taken together, the results highlight the value of using both treatment definitions to fully assess the burden and dynamics of ADRD on adult children. 

Abstract: Cognitive measures collected in aging surveys are used widely, yet their reliability is rarely tested directly. SHARE-HCAP offers a near-ideal test-retest setting: the same respondents, in five European countries, complete the same four cognitive tests once in the regular SHARE Wave 9 interview and again in SHARE-HCAP, five months later, on average. We build a standardized Cognitive Score combining the four tests common to both instruments, and find that its reliability, measured using the intra-class correlation coefficient (ICC, 0.742) is markedly higher than that of any single test (0.52-0.55, with the exception of animal naming). We then examine which factors are associated with measured cognitive performance. Women and respondents with tertiary education score higher, while the time between interviews and the time of day are unrelated to the performance. A change of interviewer between the two occasions, instead, is associated with a lower cognitive score, an association concentrated in the lower tail of the distribution; consistent with this, the ICC is systematically higher among respondents who kept the same interviewer throughout. These results carry direct implications for how cognitive assessments are administered and collected in surveys.

Abstract: Building on Mazzonna and Peracchi (2024), who find that U.S. respondents experiencing severe cognitive decline, but unaware of it, are more likely to suffer wealth losses over a two-year period, this study examines whether this protective role of awareness persists over a longer horizon. Using SHARE data on European respondents followed over eight years, I find that what actually matters over such a longer time horizon is having a severe cognitive decline, regardless of being aware of it. This pattern holds for both U.S. and European respondents when compared over the same time horizon. Severe cognitive decline also reduces the ability to pay for unexpected expenses, again, regardless of awareness, an effect that is further compounded when severe cognitive decline co-occurs with depression. Beyond these objective outcomes, I show that a pessimistic view of one's own cognitive health, rather than objective decline, worsens the perception of everyday financial strain, a pattern confirmed when proxying such a negative view using depression, and predicts reliance on informal financial support. These findings highlight the complex relationship between cognitive decline, mental health, and financial stability, emphasizing the need for specialized support for cognitively and mentally fragile individuals.

Abstract: The COVID-19 pandemic exposed older individuals to both the direct consequences of infection and the indirect ones of the restrictions adopted to contain it. Using three waves (2015, 2019/2020, 2021/2022) of the Survey of Health, Ageing and Retirement in Europe (SHARE), covering seventeen countries, I study whether cognitive decline among elderly Europeans accelerated during the pandemic. Across four cognitive domains, the evidence points to a domain-specific acceleration, specifically for orientation in time, concentrated among men and older individuals. Splitting the sample by individual- and country-level pandemic exposure suggests a similar pattern, supporting the hypothesis that the COVID-19 pandemic played a role in accelerating cognitive decline.