Lecture 5- Income and the Utilization of Long-Term Care Services (Goda, et al 2010) Flashcards
Who wrote Income and the Utilization of Long Term Care Services and when?
Goda et al. 2010
What is the main goal of the paper?
To analyse the effect that an increase in social security (retirement) income has on the type and amount of long -term care used (nursing home usage vs formal home care)
Why might there be omitted variable bias when using social security income as a determinant of the amount of long term care used?
Because there are factors that can impact both retirement income and amount of long term care, that haven’t been included in the regression.
E.g. health as healthier individuals will earn a higher retirement income and also most likely need less long term caree.
What did Goda do to prevent omitted variable bias in the regression?
He used “Notch” as an instrument variable. This was the variation in retirement income that people would receive based on when they were born
Further describe the Notch
In 1972 the US Congress made a mistake and workers were given a higher social security payment relative to what those born before and after them.
Where was the data from?
They took data from longitudinal surveys from AHEAD gleaning information from 1993-1995. It had info on over 6000 households
A follow up survey was conducted in 1995
What type of study was done? ( RCT, dif-in-dif, natural experiment)
It was a natural experiment as the notch that gave the discontinuity in social security income was out of the control of the researchers.
What was the main regression?
Uhi= B I +δXh+uh
U shows the usage of long-term care by individual I in household h.
I shows the amount of social security (retirement) income received each year.
X is a vector of background statistics -race, age
What was the regression with the Instrumental Variable?
Ih = λN+γ Xh+uh
I means income as before
N is a dummy variable equalling to 1 if the respondent was born in the Notch year so if they were part of the treatment group
X is a vector of background statistics
What is the main assumption of the Instrumental Variable
That the notch (the instrument) is correlated with income but not with the choice of long term care ( instrumental variable must be correlated with x but not impact y directly can only impact y indirectly through x)
What does previous literature say about incomes influence on the amount/type of long term care used?
What is the issue with this study
There is not a large amount of research on income and its effect on long term care.
Ettner in 1994 showed that increases in income had no statistically significant effect on nursing home numbers, however it did cause an increase in formal home care.
However, this study is most likely biased as other variables that impact both income and use of long-term care (such as health) were not taken into account.
What were the main results concerning an increase in income and the amount of long term care used?
When taking the Instrumental Variable into account and not just the OLS, an increase in income led to a statistically insignificant increase in long term care of about 4.3%
What were the main results concerning an increase in income and the type of long term care used?
Formal home care increased with income
Nursing home care fell as income increases.
These two opposite effects explain why the overall level of long term care is unchanged
What are the main explanations for the results?
People prefer formal home care to nurse homes so if they can afford it they substitute nursing homes for formal home care?
OR it could be that as income increases, the elderly are in a better health condition so have less need for nursing homes and are better able to be taken care of at home.
The first explanation is the most likely
