132
Dynamics of Health Care Expenditures and Income
Dynamics of Health Care Expenditures and Income
Tuesday, June 25, 2019
Exhibit Hall C (Marriott Wardman Park Hotel)
Catastrophic health care expenditures are correlated over time such that the accumulated effects threaten to deplete
savings for a US household. Understanding patterns of health care expenditures
at the household level is of interest also in countries with full insurance,
such as Denmark. Moreover, health care expenditures and income have a
simultaneous relation, including a negative social gradient, which is of
interest regardless of insurance system. That is, higher health care
expenditures reflect poorer health which decreases individual ability to work
and earn income. On the other hand, a higher income enables households to
consume more health promoting goods which decreases health care expenditures.
In this paper, I set up a bivariate model with contemporaneous and dynamic effects between health care expenditures and income. Using data at the individual level from the Health and Retirement Study, HRS, and Danish administrative data a two step estimation procedure is applied. In the first step, contemporaneous effects are identified with a system instrumental variable estimation. Then a VARMA(p,q) model is fitted to the covariance structure of the residuals from the first step. This expands existing models with an income dimension which opens a pathway through which castrophic healthcare expenditures can affect income. Thus, health shocks might be more severe for the household than previously assessed.
% Previous models have neglected the dynamic relationship between health care expenditures and income. Instead, focus have solely been on health care expenditures even though income was found to be endogenous. Thus, health shocks might be more severe for the household than previously assessed since health expenditure shocks can (i) increase health care expenditures and (ii) decrease income.
For estimates with HRS data a contemporaneous 1% increase in income decreases health care expenditures by 0.07% and a contemporaneous 1% increase in healthcare expenditures decreases income by 0.49%. When the error terms follow a VARMA(1,1) process, shocks to healthcare expenditures increase healthcare expenditures and decreases income. A one standard deviation shock to healthcare expenditures increases healthcare expenditures by 234% after 10 periods and decreases income by 1378% in aggregation. A one standard deviation income shocks increase income by an accumulated amount of 109% after 10 periods and decreases healthcare expenditures by 12%.
In the Danish case a 1% contemporaneous income increase decreases health care expenditures by 0.25% and a 1% contemporaneous healthcare expenditure increase increases income by 0.31%. Using a VARMA(1,1) process of the error terms a one standard deviation shock to healthcare expenditures has an aggregated effect of 205% increase in healthcare expenditures and 172% decrease in income after 10 periods. The dynamic effects of a shocks to income are negligible.
In a univariate setup, previous authors have estimated the effects of a one-standard-deviation shock to health care expenditures to accumulate to 482% which significantly downplays the financial distress of the household. These results stress the importance of intertemporal income effects when households experience catastrophic healthcare expenditures.
savings for a US household. Understanding patterns of health care expenditures
at the household level is of interest also in countries with full insurance,
such as Denmark. Moreover, health care expenditures and income have a
simultaneous relation, including a negative social gradient, which is of
interest regardless of insurance system. That is, higher health care
expenditures reflect poorer health which decreases individual ability to work
and earn income. On the other hand, a higher income enables households to
consume more health promoting goods which decreases health care expenditures.
In this paper, I set up a bivariate model with contemporaneous and dynamic effects between health care expenditures and income. Using data at the individual level from the Health and Retirement Study, HRS, and Danish administrative data a two step estimation procedure is applied. In the first step, contemporaneous effects are identified with a system instrumental variable estimation. Then a VARMA(p,q) model is fitted to the covariance structure of the residuals from the first step. This expands existing models with an income dimension which opens a pathway through which castrophic healthcare expenditures can affect income. Thus, health shocks might be more severe for the household than previously assessed.
% Previous models have neglected the dynamic relationship between health care expenditures and income. Instead, focus have solely been on health care expenditures even though income was found to be endogenous. Thus, health shocks might be more severe for the household than previously assessed since health expenditure shocks can (i) increase health care expenditures and (ii) decrease income.
For estimates with HRS data a contemporaneous 1% increase in income decreases health care expenditures by 0.07% and a contemporaneous 1% increase in healthcare expenditures decreases income by 0.49%. When the error terms follow a VARMA(1,1) process, shocks to healthcare expenditures increase healthcare expenditures and decreases income. A one standard deviation shock to healthcare expenditures increases healthcare expenditures by 234% after 10 periods and decreases income by 1378% in aggregation. A one standard deviation income shocks increase income by an accumulated amount of 109% after 10 periods and decreases healthcare expenditures by 12%.
In the Danish case a 1% contemporaneous income increase decreases health care expenditures by 0.25% and a 1% contemporaneous healthcare expenditure increase increases income by 0.31%. Using a VARMA(1,1) process of the error terms a one standard deviation shock to healthcare expenditures has an aggregated effect of 205% increase in healthcare expenditures and 172% decrease in income after 10 periods. The dynamic effects of a shocks to income are negligible.
In a univariate setup, previous authors have estimated the effects of a one-standard-deviation shock to health care expenditures to accumulate to 482% which significantly downplays the financial distress of the household. These results stress the importance of intertemporal income effects when households experience catastrophic healthcare expenditures.
