How financial strain affects health: Evidence from the Dutch National Bank Household Survey
Thoughts
Head empty
Annotations
prentice2017 - p. 2
Also, this literature has indicated that both impulsivity and stress are risk factors for negative health behaviours but there has been little work on understanding the causal mechanism from stress to time discounting to worse health-related behaviour especially for economic stresses.
prentice2017 - p. 2
We use the Dutch National Bank 15 Household Survey (DNB) which provides us with a large nationally-representative sample of over 40000 observations over a twenty-year period covering the years before and after the financial crisis.
prentice2017 - p. 3
Using longitudinal data in our study, we test for a temporal ordering that indicates causation from stress to health in our structural model. To avoid concerns about unmeasured confounders driving the relationship between financial strain and ill-health, we use prior levels of the dependent variable in models of all the key variables
prentice2017 - p. 3
A measure of financial strain is constructed using two questions on the household’s financial situation.
prentice2017 - p. 4
DNB is a panel survey that has been active since 1993. The survey, collected by CentER (Tilburg University, the Netherlands) gathers information annually from approximately 2000 households in separate questionnaires on general information, household and work, ac commodation and mortgages, health and income, assets and liabilities and economic and psychological concepts. This study uses all waves of data from years 1996 to 2015 as data relevant to our models were particularly sparse in the earliest years of the survey.
prentice2017 - p. 5
We expect that financial strain increases time discounting leading to a increased incidence of poor health-related behaviours.
prentice2017 - p. 6
The final measure is the Consideration of Future Consequences (CFC) scale, a psychological construct used to measure an individual’s future orientation (Strathman et al., 1994). 25 Respondents indicate to what extent they agree with the 11 statements on attitudes referring to the trade-off between the present and the future using a 7-point scale
prentice2017 - p. 7
Our measure of financial strain comes from two survey questions on the household’s current financial situation. When asked ‘How well can you manage on the total income of your household?’ respondents chose from five possible answers : ‘1-it is very hard’,‘2-it is hard’, ‘3-it is neither hard nor easy’,‘4-it is easy’ and ‘5-it is very easy’. A second question asked ‘How is the financial situation of your household at the moment?’ with possible responses: ‘1-there are debts’,‘2-need to draw upon savings’,‘3-it is just about manageable’, ‘4-some money is saved’ and ‘5-a lot of money can be saved’. A dichotomous financial strain variable was constructed which was one if the response in either of these questions was 1 or 2 and zero otherwise.
prentice2017 - p. 7
A number of covariates were used as control variables. The respondent’s gender, highest education level completed, whether they had a partner present in the household and whether they lived in owner-occupied accommodation were represented by binary variables. The twelve categories of response to primary occupation were reduced to a dichotomous indicator of whether employed or not to reduce the number of controls in the model. Age and the number of children in the household entered the model as continuous variables. The survey records net household income as a categorical variable. This was recoded taking the midpoint of each category and upper (lower) interval boundaries for the lowest (highest) category. The average net household income response within household was taken where responses differed
prentice2017 - p. 8
The basic three wave model is represented in figure 1. The health outcome for person i at time t (Yit) is affected by financial strain from two periods before (Si,t−2) both directly and indirectly through changes in the mediator (Mi,t−1). Path ab represents the indirect effect via the mediator while path c represents the direct effect from financial stress to the health outcome not via time discounting. This part of the model is given in equation form below : Mit−1 = δM Mi,t−2 + βaSi,t−2 + γM′ Xi,t−1 + M i,t−1 Yit = δY Yi,t−1 + βbMi,t−1 + βcSt−2 + γY ′ Xit + Y it
prentice2017 - p. 10
Although the strength of our study design is that the temporal ordering of financial strain and health consequences in our model controls for potential reverse causation, the cross-wave structure may not capture short-term periods of financial strain which can also impact on health
prentice2017 - p. 11
The mediated effect of strain on being overweight through changes in difficulty controlling expenditure is statistically significant but also small
prentice2017 - p. 12
at 0.007 (0.177*0.040, p=0.000)